Selection and characterization of a new class of peptide exosite inhibitors of coagulation factor VIIa.

A new series of peptide inhibitors of human Factor VIIa (FVIIa) has been identified and affinity matured from naive and partially randomized peptide phage libraries selected against the immobilized tissue factor x Factor VIIa (TF x FVIIa) complex. These "A-series" peptides contain a single disulfide bond and a 13-residue minimal core required for maximal affinity. They are exemplified by peptide A-183 (EEWEVLCWTWETCER), which binds at a newly identified exosite on the FVIIa protease domain, described in the accompanying report [Roberge, M., Santell, L., Dennis, M. S., Eigenbrot, C., Dwyer, M. A., and Lazarus, R. A. (2001) Biochemistry 40, 9522-9531]. A-183 was obtained from a trypsin digest of A-100-Z, a recombinant protein comprising A-183 and the Z domain of protein A. Surprisingly, A-183 was a very potent inhibitor of TF x FVIIa, inhibiting activation of Factor X (FX) and Factor IX and amidolytic activity of Chromozym t-PA with IC50 values of 1.6 +/- 1.2, 3.5 +/- 0.3, and 8.5 +/- 3.5 nM, respectively. Kinetic analysis revealed that A-183 was a partial (hyperbolic) mixed-type inhibitor of FX activation having a Ki of 200 pM as well as a partial competitive inhibitor of amidolytic activity. The A-series peptides were also specific and potent inhibitors of TF-dependent clotting as measured in a prothrombin time (PT) clotting assay and had no effect on the TF-independent activated partial thromboplastin time. At saturating concentrations of peptide, the maximal extent by which A-183 and A-100-Z inhibited the rate of FX activation was 78 +/- 3 and 89 +/- 6%, respectively. The degree of inhibition of the rate of FX activation correlated with a maximum fold prolongation in the PT assay of 1.8-fold for A- 183 and 3.3-fold for A-100-Z. The A-series peptides represent a new class of peptide exosite inhibitors that are capable of attenuating, rather than completely inhibiting, the activity of TF x FVIIa, potentially leading to anticoagulants with an increased therapeutic window.

A novel exosite on coagulation factor VIIa and its molecular interactions with a new class of peptide inhibitors.

A new inhibitory peptide binding exosite on the protease domain of coagulation Factor VIIa (FVIIa) has been identified. A novel series of peptide inhibitors of FVIIa, termed the "A-series" peptides, identified from peptide phage libraries and exemplified by peptide A-183 [Dennis, M. S., Roberge, M., Quan, C., and Lazarus, R. A. (2001) Biochemistry 40, 9513-9521], specifically bind at a site that is distinct from both the active site and the exosite of another recently described peptide inhibitor of FVIIa, E-76 [Dennis, M. S., Eigenbrot, C., Skelton, N. J., Ultsch, M. H., Santell, L., Dwyer, M. A., O'Connell, M. P., and Lazarus, R. A. (2000) Nature 404, 465-4701. Peptide A-183 prolonged TF-dependent clotting in human, but not rabbit plasma. Thus, a panel of human FVIIa mutants, containing 70 of the 76 rabbit sequence differences in the protease domain, localized the binding site to residues in the 60s loop and the C-terminus. The location of the exosite was refined by a series of FVIIa alanine mutants, which showed that proximal residues Trp 61 and Leu 251 were critical for binding. Kinetic and equilibrium binding constants for zymogen FVII, FVIIa and TF x FVIIa were determined using immobilized N-terminal biotinylated A-183 by surface plasmon resonance. No peptide binding to nine other human serine proteases was observed. Key residues on the peptide were determined from binding to FVIIa and inhibition of FX activation using a series of alanine mutants of A-183 fused to the Z domain of protein A. Analysis of the mutagenesis data is presented in the context of a crystal structure of A-183 in complex with a version of zymogen FVII [Eigenbrot, C., Kirchhofer, D., Dennis, M. S., Santell, L., Lazarus, R. A., Stamos, J., and Ultsch, M. H. (2001) Structure 9, 627-636]. The shape and proximity of this exosite to the active site may lend itself towards the design of new anticoagulants that inhibit FVIIa.

The factor VII zymogen structure reveals reregistration of beta strands during activation.

BACKGROUND: Coagulation factor VIIa (FVIIa) contains a Trypsin-like serine protease domain and initiates the cascade of proteolytic events leading to Thrombin activation and blood clot formation. Vascular injury allows formation of the complex between circulating FVIIa and its cell surface bound obligate cofactor, Tissue Factor (TF). Circulating FVIIa is nominally activated but retains zymogen-like character and requires TF in order to complete the zymogen-to-enzyme transition. The manner in which TF exerts this effect is unclear. The structure of TF/FVIIa is known. Knowledge of the zymogen structure is helpful for understanding the activation transition in this system. RESULTS: The 2 A resolution crystal structure of a zymogen form of FVII comprising the EGF2 and protease domains is revealed in a complex with the exosite binding inhibitory peptide A-183 and a vacant active site. The activation domain, which includes the N terminus, differs in ways beyond those that are expected for zymogens in the Trypsin family. There are large differences in the TF binding region. An unprecedented 3 residue shift in registration between beta strands B2 and A2 in the C-terminal beta barrel and hydrogen bonds involving Glu154 provide new insight into conformational changes accompanying zymogen activation, TF binding, and enzymatic competence. CONCLUSIONS: TF-mediated allosteric control of the activity of FVIIa can be rationalized. The reregistering beta strand connects the TF binding region and the N-terminal region. The zymogen registration allows H bonds that prevent the N terminus from attaining a key salt bridge with the active site. TF binding may influence an equilibrium by selecting the enzymatically competent registration.

Improved activity of an actin-resistant DNase I variant on the cystic fibrosis airway secretions.

In cystic fibrosis (CF), actin and DNA originating from inflammatory cells contribute to the thickness of airway secretions. Actin can bind to DNA-rich fibers and potently inhibit the enzymatic activity of rhDNase. The in vitro effects of the actin-resistant rhDNase variant (A114R) were analyzed and compared with those of the wild-type rhDNase. Frozen and thawed CF airway secretions were incubated for 30 min with different concentrations (0.1, 0.5, 1, 5, or 10 microg/ml) of either actin-resistant rhDNase or wild-type rhDNase. We observed that both the wild-type and the actin-resistant rhDNase significantly decreased (p < 0.05 and p < 0.001, respectively) the airway secretion viscosity. The decrease in airway secretion viscosity was significant even at low concentrations (0.1 microg/ml) of the actin-resistant variant. Incubation with the actin-resistant variant resulted in a significant decrease (p < 0.02) of the airway secretion contact angle and cough transport. A significantly higher (p < 0.01) increase in contact angle and cough transport of airway secretions was observed at 10 microg/ml with the actin-resistant variant as compared with the wild-type rhDNase. The present study had demonstrated that the actin-resistant rhDNase variant (A114R) has an enhanced capacity to improve the physical properties and cough transport of airway secretions from patients with cystic fibrosis.

Recombinant decorsin: dynamics of the RGD recognition site.

Decorsin is an antagonist of integrin alphaIIbbeta3 and a potent platelet aggregation inhibitor. A synthetic gene encoding decorsin, originally isolated from the leech Macrobdella decora, was designed, constructed, and expressed in Escherichia coli. The synthetic gene was fused to the stII signal sequence and expressed under the transcriptional control of the E. coli alkaline phosphatase promoter. The protein was purified by size-exclusion filtration of the periplasmic contents followed by reversed-phase high-performance liquid chromatography. Purified recombinant decorsin was found to be indistinguishable from leech-derived decorsin based on amino acid composition, mass spectral analysis, and biological activity assays. Complete sequential assignments of 1H and proton bound 13C resonances were established. Stereospecific assignments of 21 of 25 nondegenerate b-methylene groups were determined. The RGD adhesion site recognized by integrin receptors was found at the apex of a most exposed hairpin loop. The dynamic behavior of decorsin was analyzed using several independent NMR parameters. Although the loop containing the RGD sequence is the most flexible one in decorsin, the conformation of the RGD site itself is more restricted than in other proteins with similar activities.

Peptide exosite inhibitors of factor VIIa as anticoagulants.

Potent anticoagulants have been derived by targeting the tissue factor-factor VIIa complex with naive peptide libraries displayed on M13 phage. The peptides specifically block the activation of factor X with a median inhibitory concentration of 1 nM and selectively inhibit tissue-factor-dependent clotting. The peptides do not bind to the active site of factor VIIa; rather, they work by binding to an exosite on the factor VIIa protease domain, and non-competitively inhibit activation of factor X and amidolytic activity. One such peptide (E-76) has a well defined structure in solution determined by NMR spectroscopy that is similar to the X-ray crystal structure when complexed with factor VIIa. These structural and functional studies indicate an allosteric 'switch' mechanism of inhibition involving an activation loop of factor VIIa and represent a new framework for developing inhibitors of serine proteases.

Ca2+-dependent activity of human DNase I and its hyperactive variants.

We have recently constructed hyperactive human deoxyribonuclease I (DNase I) variants that digest double-stranded DNA more efficiently under physiological saline conditions by introducing positively charged amino acids at eight positions that can interact favorably with the negatively charged DNA phosphates. In this study, we present data from supercoiled DNA nicking, linear DNA digestion, and hyperchromicity assays that distinguish two classes of DNase I hyperactive variants based upon their activity dependence on Ca2+. Class A variants are highly dependent upon Ca2+, having up to 300-fold lower activity in the presence of Mg2+ alone compared to that in the presence of Mg2+ and Ca2+, and include Q9R, H44K, and T205K, in addition to wild-type DNase I. In contrast, the catalytic activity of Class B variants, which comprise the E13R, T14K, N74K, S75K, and N110R hyperactive variants, is relatively Ca2+ independent. A significant proportion of this difference in Ca2+-dependent activity can be attributed to one of the two structural calcium binding sites in DNase I. Compared to wild-type, the removal of Ca2+ binding site 2 by alanine replacements at Asp99, Asp107, and Glu112 decreased activity up to 26-fold in the presence of Mg2+ and Ca2+, but had no effect in the presence of Mg2+ alone. We propose that the rate-enhancing effect of Ca2+ binding at site 2 can be replaced by favorable electrostatic interactions created by proximal positively charged amino acid substitutions such as those found in the Class B variants, thus reducing the dependence on Ca2+.

Expression and characterization of a DNase I-Fc fusion enzyme.

Recombinant human deoxyribonuclease I (DNase I) is an important clinical agent that is inhaled into the airways where it degrades DNA to lower molecular weight fragments, thus reducing the viscoelasticity of sputum and improving the lung function of cystic fibrosis patients. To investigate DNases with potentially improved properties, we constructed a molecular fusion of human DNase I with the hinge and Fc region of human IgG1 heavy chain, creating a DNase I-Fc fusion protein. Infection of Sf9 insect cells with recombinant baculovirus resulted in the expression and secretion of the DNase I-Fc fusion protein. The fusion protein was purified from the culture medium using protein A affinity chromatography followed by desalting by gel filtration and was characterized by amino-terminal sequence, amino acid composition, and a variety of enzyme-linked immunosorbent assays (ELISA) and activity assays. The purified fusion contains DNase I, as determined by a DNase I ELISA and an actin-binding ELISA, and an intact antibody Fc region, which was quantified by an Fc ELISA, in a 2:1 stoichiometric ratio, respectively. The dimeric DNase I-Fc fusion was functionally active in enzymatic DNA digestion assays, albeit about 10-fold less than monomeric DNase I. Cleavage of the DNase I-Fc fusion by papain resulted in a specific activity comparable to the monomeric enzyme. Salt was inhibitory for wild type monomeric DNase I but actually enhanced the activity of the dimeric DNase I-Fc fusion. The DNase I-Fc fusion protein was also less Ca2+-dependent than DNase I itself. These results are consistent with a higher affinity of the dimeric fusion protein to DNA than monomeric DNase I. The engineered DNase I-Fc fusion protein described herein has properties that may have clinical benefits.

Cloning and characterization of an actin-resistant DNase I-like endonuclease secreted by macrophages.

We have cloned human and murine DNase I-like cDNAs, termed LS-DNase, which are expressed at high levels in liver and spleen tissues. LS-DNase expression is highly specific to macrophage populations within these and other tissues. Mature LS-DNase from both species is a secreted, non-glycosylated protein containing 285 residues, with a calculated molecular mass of 33 kDa and a basic isoelectric point. Human and murine LS-DNase are highly conserved and share 83% identity. Sequence analysis reveals that LS-DNase shares 46% amino acid sequence identity with DNase I. However, several residues identified as important for interaction of human DNase I with actin are not conserved in both human and murine LS-DNase. Consistent with this observation, recombinant human LS-DNase possesses a DNA hydrolytic activity which, unlike DNase I, is not inhibited by G-actin. The existence of a family of DNase I-like molecules that have tissue-specific expression patterns and the possible role of a macrophage specific DNase are discussed.

Improved potency of hyperactive and actin-resistant human DNase I variants for treatment of cystic fibrosis and systemic lupus erythematosus.

The ability of recombinant human DNase I (DNase I) to degrade DNA to lower molecular weight fragments is the basis for its therapeutic use in cystic fibrosis (CF) patients and its potential use as a treatment for systemic lupus erythematosus (SLE). To increase the potency of human DNase I, we have generated and characterized three classes of mutants: (a) hyperactive variants, which have from one to six additional positively charged residues (+1 to +6) and digest DNA much more efficiently relative to wild type, (b) actin-resistant variants, which are no longer inhibited by G-actin, a potent inhibitor of DNase I, and (c) combination variants that are both hyperactive and actin-resistant. For DNA scission in CF sputum where the DNA concentration and length are large, we measured a approximately 20-fold increase in potency relative to wild type for the +3 hyperactive variant Q9R/E13R/N74K or the actin-resistant variant A114F; the hyperactive and actin-resistant combination variant was approximately 100-fold more potent than wild type DNase I. For digesting lower concentrations of DNA complexed to anti-DNA antibodies in human serum, we found a maximal enhancement of approximately 400-fold over wild type for the +2 variant E13R/N74K. The +3 enzymes have approximately 4000-fold enhancement for degrading moderate levels of exogenous DNA spiked into human serum, whereas the +6 enzyme has approximately 30,000-fold increased activity for digesting the extremely low levels of endogenous DNA found in serum. The actin resistance property of the combination mutants further enhances the degree of potency in human serum. Thus, the human DNase I variants we have engineered for improved biochemical and pharmacodynamic properties have greater therapeutic potential for treatment of both CF and SLE.

Hyperactivity of human DNase I variants. Dependence on the number of positively charged residues and concentration, length, and environment of DNA.

Human DNase I, an enzyme used to treat cystic fibrosis patients, has been engineered to more effectively degrade double-stranded DNA to lower molecular weight forms by introducing positively charged amino acids at positions that can interact favorably with the proximal negatively charged phosphate groups of the DNA. A series of combination mutants having from one to six additional basic residues compared with the wild type has been constructed, expressed in human 293 cells, and characterized. The degree of hyperactivity for the mutants was highly dependent upon the conditions in various assays, including the concentration and length of the DNA substrate and the salt and divalent metal ion concentrations. The level of hyperactivity was inversely proportional to both DNA concentration and DNA length, consistent with the processive nicking mechanism for the hyperactive variants. Salt was inhibitory for wild type DNase I but actually enhanced the activity of the hyperactive variants. Under optimal conditions for wild type, variants with one additional positive charge possessed the highest activity, which was only severalfold greater than that for wild type. However, in the presence of low DNA concentrations and molecular weights, no Ca2+, and 150 mM NaCl, the variant with six engineered basic residues was most active, having >10,000-fold higher activity than the wild type enzyme. Therefore, any potential increase in potency for the hyperactive variants in vivo will be determined by the concentration, length, and environment of the DNA.

Mutational analysis of human DNase I at the DNA binding interface: implications for DNA recognition, catalysis, and metal ion dependence.

Human deoxyribonuclease I (DNase I), an enzyme used to treat cystic fibrosis patients, has been systematically analyzed by site-directed mutagenesis of residues at the DNA binding interface. Crystal structures of bovine DNase I complexed with two different oligonucleotides have implicated the participation of over 20 amino acids in catalysis or DNA recognition. These residues have been classified into four groups based on the characterization of over 80 human DNase I variants. Mutations at any of the four catalytic amino acids His 134, His 252, Glu 78, and Asp 212 drastically reduced the hydrolytic activity of DNase I. Replacing the three putative divalent metal ion-coordinating residues Glu 39, Asp 168, or Asp 251 led to inactive variants. Amino acids Gln 9, Arg 41, Tyr 76, Arg 111, Asn 170, Tyr 175, and Tyr 211 were also critical for activity, presumably because of their close proximity to the active site, while more peripheral DNA interactions stemming from 13 other positions were of minimal significance. The relative importance of these 27 positions is consistent with evolutionary relationships among DNase I across different species, DNase I-like proteins, and bacterial sphingomyelinases, suggesting a fingerprint for a family of DNase I-like proteins. Furthermore, we found no evidence for a second active site that had been previously implicated in Mn2+-dependent DNA degradation. Finally, we correlated our mutational analysis of human DNase I to that of bovine DNase I with respect to their specific activity and dependence on divalent metal ions.

Engineering hyperactive variants of human deoxyribonuclease I by altering its functional mechanism.

Human deoxyribonuclease I (DNase I), an enzyme used to treat cystic fibrosis patients, has been engineered to more effectively degrade double-stranded DNA to lower molecular weight fragments by altering its functional mechanism from the native single-stranded nicking pathway to a much more efficient one which results in increased double-stranded scission. By introducing positively charged amino acids at DNase I positions that can interact favorably with the proximal negatively charged phosphate groups of the DNA, we have created a hyperactive variant with approximately 35-fold higher DNA-degrading activity relative to wild type. This enhancement can be attributed to both a decrease in Km and an increase in Vmax. Furthermore, unlike wild-type DNase I, the hyperactive variants are no longer inhibited by physiological saline. Replacement of the same positions with negatively charged amino acids greatly reduced DNA cleavage activity, consistent with a repulsive effect with the neighboring DNA phosphates. In addition, these variants displayed similar activities toward a small synthetic substrate, p-nitrophenyl phenylphosphonate, suggesting that the difference in DNA cleavage activity is due to the interaction of the engineered charged residues with the DNA phosphate backbone rather than any change in catalytic machinery. Finally, experiments involving the repair of DNase I digested DNA with T4 DNA ligase and the Klenow fragment of DNA polymerase I suggest that single-stranded gaps are introduced by the hyperactive variants. Thus, the increased functional activity of the hyperactive variants may be explained in part by a shift toward a processive DNA nicking mechanism, which leads to a higher frequency of double-stranded breaks.

Potent bifunctional anticoagulants: Kunitz domain-tissue factor fusion proteins.

A strategy to design potent antagonists of human coagulation factor VIIa (FVIIa) by linking two proteins that independently inhibit activity and bind at separate, nonoverlapping sites is presented. A bifunctional inhibitor (KDTF5), comprising a Kunitz-type domain engineered to inhibit the FVIIa active site and a soluble tissue factor (TF) variant that is defective as a cofactor for factor X (FX) activation, was developed from structure-based modeling of a ternary FVIIa-Kunitz domain-TF complex. KDTF5 inhibited FVIIa-dependent FX activation with a Ki* of 235 +/- 45 pM, a 193-fold and 398-fold increase in potency compared to the TF variant and Kunitz domain individually. Similarly, KDTF5 was a more potent anticoagulant in vitro compared to either inhibitory domain alone. The results demonstrate the harnessing of a macromolecular chelate effect by fusing two inhibitory ligands that bind a target at spatially distinct sites.

Engineering actin-resistant human DNase I for treatment of cystic fibrosis.

Human deoxyribonuclease I (DNase I), an enzyme recently approved for treatment of cystic fibrosis (CF), has been engineered to create two classes of mutants: actin-resistant variants, which still catalyze DNA hydrolysis but are no longer inhibited by globular actin (G-actin) and active site variants, which no longer catalyze DNA hydrolysis but still bind G-actin. Actin-resistant variants with the least affinity for actin, as measured by an actin binding ELISA and actin inhibition of [33P] DNA hydrolysis, resulted from the introduction of charged, aliphatic, or aromatic residues at Ala-114 or charged residues on the central hydrophobic actin binding interface at Tyr-65 or Val-67. In CF sputum, the actin-resistant variants D53R, Y65A, Y65R, or V67K were 10-to 50-fold more potent than wild type in reducing viscoelasticity as determined in sputum compaction assays. The reduced viscoelasticity correlated with reduced DNA length as measured by pulsed-field gel electrophoresis. In contrast, the active site variants H252A or H134A had no effect on altering either viscoelasticity or DNA length in CF sputum. The data from both the active site and actin-resistant variants demonstrate that the reduction of viscoelasticity by DNase I results from DNA hydrolysis and not from depolymerization of filamentous actin (F-actin). The increased potency of the actin-resistant variants indicates that G-actin is a significant inhibitor of DNase I in CF sputum. These results further suggest that actin-resistant DNase I variants may have improved efficacy in CF patients.

A comparison of the effect of decorsin and two disintegrins, albolabrin and eristostatin, on platelet function.

Naturally-occurring fibrinogen receptor antagonists and platelet aggregation inhibitors that are found in snake venom (disintegrins) and leeches share many common features, including an RGD sequence, high cysteine content, and low molecular weight. There are, however, significant selectivity and potency differences. We compared the effect of three proteins on platelet function: albolabrin, a 7.5 kDa disintegrin, eristostatin, a 5.4 kDa disintegrin in which part of the disintegrin domain is deleted, and decorsin, a 4.5 kDa non-disintegrin derived from the leech Macrobdella decora, which has very little sequence similarity with either disintegrin. Decorsin was about two times less potent than albolabrin and six times less potent than eristostatin in inhibiting ADP-induced human platelet aggregation. It had a different pattern of interaction with glycoprotein IIb/IIIa as compared to the two disintegrins. Decorsin bound with a low affinity to resting platelets (409 nM) and to ADP-activated platelets (270 nM), and with high affinity to thrombin activated platelets (74 nM). At concentrations up to 685 nM, it did not cause expression of a ligand-induced binding site epitope on the beta 3 subunit of the GPIIb/IIIa complex. It did not significantly inhibit isolated GPIIb/IIIa binding to immobilized von Willebrand Factor. At low doses (1.5-3.0 micrograms/mouse), decorsin protected mice against death from pulmonary thromboembolism, showing an effect similar to eristostatin. This suggested that decorsin is a much more potent inhibitor of platelet aggregation in vivo than in vitro, and it may have potential as an antiplatelet drug.

Potent and selective Kunitz domain inhibitors of plasma kallikrein designed by phage display.

Phage displaying APPI Kunitz domain libraries have been used to design potent and selective active site inhibitors of human plasma kallikrein, a serine protease that plays an important role in both inflammation and coagulation. Selected clones from two Kunitz domain libraries randomized at or near the binding loop (positions 11-13, 15-19, and 34) were sequenced following five rounds of selection on immobilized plasma kallikrein. Invariant preferences for Arg at position 15 and His at position 18 were found, whereas His, Ala, Ala, and Pro were highly preferred residues at positions 13, 16, 17, and 19, respectively. At position 11 Pro, Asp, and Glu were favored, while hydrophobic residues were preferred at position 34. Selected variants, purified by trypsin affinity chromatography and reverse phase high performance liquid chromatography, potently inhibited plasma kallikrein, with apparent equilibrium dissociation constants (Ki*) ranging from approximately 75 to 300 pM. From sequence and activity data, consensus mutants were constructed by site directed mutagenesis. One such mutant, KALI-DY, which differed from APPI at 6 key residues (T11D, P13H, M17A, I18H, S19P, and F34Y), inhibited plasma kallikrein with a Ki* = 15 +/- 14 pM, representing an increase in binding affinity of more than 10,000-fold compared to APPI. Similar to APPI, the variants also inhibited Factor XIa with high affinity, with Ki* values ranging from approximately 0.3 to 15 nM; KALI-DY inhibited Factor XIa with a Ki* = 8.2 +/- 3.5 nM. KALI-DY did not inhibit plasmin, thrombin, Factor Xa, Factor XIIa, activated protein C, or tissue factor. Factor VIIa. Consistent with the protease specificity profile, KALI-DY did not prolong the clotting time in a prothrombin time assay, but did prolong the clotting time in an activated partial thromboplastin time assay > 3.5-fold at 1 microM.

Analysis of the factor VIIa binding site on human tissue factor: effects of tissue factor mutations on the kinetics and thermodynamics of binding.

Surface plasmon resonance (SPR) measurements on a BIAcore instrument have been used to measure the effects of mutations in human tissue factor (TF), the initiator of blood coagulation, on the kinetics and affinity of binding to human FVIIa. TF mutant proteins were produced in soluble form by expression of the extracellular domain (sTF) in Escherichia coli followed by immunoaffinity purification. Mutants were designed and analyzed on the basis of the structure of sTF recently determined by X-ray crystallography [Muller et al. (1994) Biochemistry 33, 10864-10870]. Wild-type sTF binding to immobilized FVIIa has k(on) = 3.4 +/- 0.8 x 10(5) M-1 s-1 and k(off) = 2.1 +/- 0.1 x 10(-3) s-1 with a calculated KD of 6.3 +/- 1.2 nM and delta G of -11.2 +/- 0.1 kcal mol-1. Calorimetric measurements indicate that binding occurs with a favorable delta H of -32 kcal mol-1, an unfavorable delta S of -70 cal K-1 mol-1, and a delta Cp of -730 cal K mol-1. The value of delta Cp is consistent with burial of a large nonpolar surface area upon binding. Five residues on TF, Lys20, Trp45, Asp58, Tyr94, and Phe140, make a large contribution (delta delta G = 1-2.5 kcal mol-1) to FVIIa binding, a set of 17 mutations result in modest decreases in affinity (delta delta G = 0.3-1 kcal mol-1), and 40 mutations have delta delta G smaller than the experimental uncertainty (+/- 0.3 kcal mol-1). Mutations at four sites result in small (0.3-0.5 kcal mol-1) increases in affinity. Decreases in affinity result primarily from increased rates of dissociation. These data define a putative FVIIa binding site on one face of the TF structure with most of the contacts contributed by the N-terminal fibronectin type III domain. The critical binding residues are found on beta-strands. An additional set of residues located on the surface of the C-terminal fibronectin type III domain opposite the FVIIa binding site have a role in the procoagulant activity of sTF but are not involved in FVIIa binding.