Von Willebrand factor-binding protein is a hysteretic conformational activator of prothrombin

Abstract

Von Willebrand factor-binding protein (VWbp), secreted by Staphylococcus aureus, displays secondary structural homology to the 3-helix bundle, D1 and D2 domains of staphylocoagulase (SC), a potent conformational activator of the blood coagulation zymogen, prothrombin (ProT). In contrast to the classical proteolytic activation mechanism of trypsinogen-like serine proteinase zymogens, insertion of the first 2 residues of SC into the NH₂-terminal binding cleft on ProT (molecular sexuality) induces rapid conformational activation of the catalytic site. Based on plasma-clotting assays, the target zymogen for VWbp may be ProT, but this has not been verified, and the mechanism of ProT activation is unknown. We demonstrate that VWbp activates ProT conformationally in a mechanism requiring its Val¹-Val² residues. By contrast to SC, full time-course kinetic studies of ProT activation by VWbp demonstrate that it activates ProT by a substrate-dependent, hysteretic kinetic mechanism. VWbp binds weakly to ProT (K_D 2.5 μM) to form an inactive complex, which is activated through a slow conformational change by tripeptide chromogenic substrates and its specific physiological substrate, identified here as fibrinogen (Fbg). This mechanism increases the specificity of ProT activation by delaying it in a slow reversible process, with full activation requiring binding of Fbg through an exosite expressed on the activated ProT*·VWbp complex. The results suggest that this unique mechanism regulates pathological fibrin (Fbn) deposition to VWF-rich areas during S. aureus endocarditis.

The tightly regulated balance among procoagulant, anticoagulant, and fibrinolytic pathways in blood is kinetically controlled by the rates of proteolytic activation of zymogens of the chymotrypsinogen family. The classical proteolytic activation mechanism is initiated by cleavage at Arg¹⁵-Ile¹⁶ (chymotrypsinogen numbering), followed by insertion of the Ile¹⁶ NH₂ terminus into the NH₂-terminal-binding cleft, and formation of a critical salt bridge with Asp¹⁹⁴ (1–3). This triggers folding of segments of the catalytic domain, which forms the substrate-binding site and oxyanion hole required for activity. The mechanism is fundamentally conformational in that zymogens are in unfavorable equilibrium between conformationally active forms that vary in equilibrium constants from 10⁸ (trypsinogen) (4) to ≈7 (single-chain tissue-type plasminogen activator) (5). Dipeptides that mimic the conserved NH₂ termini (Ile-Val or Val-Val) induce structural changes in trypsinogen toward the active proteinase conformation (1, 3), and tight-binding inhibitors like bovine pancreatic trypsin inhibitor convert trypsinogen fully to the conformation of trypsin (3).

By contrast, staphylocoagulase (SC) from Staphylococcus aureus activates the coagulation zymogen prothrombin (ProT) nonproteolytically, and formation of the tightly bound ProT·SC complex initiates direct cleavage of fibrinogen (Fbg) into fibrin (Fbn) (6, 7). The ProT·SC(1-325) complex expresses a new exosite that mediates specific substrate recognition of Fbg, whereas full-length SC(1-660) also interacts with Fbg as an adhesion protein through COOH-terminal repeat sequences (8, 9). This benefits the pathogen by facilitating escape from the host defense system through formation of protective Fbn–platelet–bacteria vegetations in acute bacterial endocarditis (10).

Von Willebrand factor-binding protein (VWbp), also secreted by S. aureus, was first identified by phage-display of an S. aureus (Newman) DNA library screened against von Willebrand factor (VWF) (11), leading to its classification as an adhesion protein. Binding to both immobilized and soluble VWF is mediated by a 26-aa region within VWbp (11), but the physiological significance of this interaction has not been established. VWbp was further recognized as a putative member of the bifunctional zymogen activator and adhesion protein (ZAAP) family for which SC is the prototype (12, 13). It was subsequently postulated to be an SC homolog from its 25% sequence identity with SC and clotting of human plasma (14).

The structure of the NH₂-terminal half of SC, SC(1-325), is unique, with a characteristic elbow-like fold between 2 domains (D1–D2) comprised of 3-helix bundles (13). Secondary structure modeling of the corresponding D1–D2 domains of VWbp(1-263) revealed a high degree of similarity (13). Structural and kinetic studies of SC demonstrated insertion of the first 2 NH₂-terminal residues of SC into the NH₂-terminal-binding cleft in the ProT catalytic domain (13). The NH₂ terminus forms a salt bridge with Asp¹⁹⁴ (13), triggering the conformational change that activates the catalytic site. This “molecular sexuality” mechanism (1) has only been demonstrated for SC and the plasminogen activator streptokinase (15), whereas essentially nothing is known about the mechanism of plasma clotting by VWbp.

The specific interactions responsible for the procoagulant activity of VWbp have not been defined, and the mechanism of its activation of ProT is unknown. In the present studies, we defined the molecular basis and unusual hysteretic kinetic mechanism of ProT activation by VWbp and identified Fbg as its specific substrate, which are postulated to play a role in the pathology of S. aureus infection.

Results

Conformational Activation of ProT by VWbp(1-263) and VWbp(1-474).

Whether VWbp(1-263) and full-length VWbp(1-474) activate ProT through a conformational change was examined by using 3-step active site-specific labeling: (i) inactivation of active species in ProT-VWbp mixtures with N^α-[(acetylthio)acetyl]-d-Phe-Pro-Arg-CH₂Cl (ATA-FPR-CH₂Cl), (ii) generation of a free thiol with NH₂OH, and (iii) covalent attachment of 5-(iodoacetamido)fluorescein (5-IAF) (16). SDS-gel electrophoresis of samples from reactions including all 3 steps showed specific, covalent incorporation of fluorescein into ProT only (Fig. 1, lanes 4), with no proteolysis of the zymogen for either VWbp(1-263) or VWbp(1-474). Stringent controls omitting one of the steps of the labeling scheme confirmed that all of the labeling steps were necessary for probe incorporation (Fig. 1, lanes 3 and 5–7).

Fig. 1.

Active site-specific labeling of ProT·VWbp(1-263) and ProT·VWbp(1-474) complexes assessed by SDS-gel electrophoresis. Fluorescence (A and C) and protein-stained (B and D) SDS gels for reactions containing 15 μM ProT and 50 μM VWbp(1-263) (A and B) or VWbp(1-474) (C and D). Reduced samples of ProT (lane 1) and VWbp (lane 2), ProT-VWbp inactivated with ATA-FPR-CH₂Cl (lane 3), and inhibited complex incubated with NH₂OH and 5-IAF (lane 4). Control samples were ProT incubated without VWbp (lane 5), omission of the NH₂OH step (lane 6), or the complex blocked with excess FPR-CH₂Cl before sequential incubation with ATA-FPR-CH₂Cl and 5-IAF (lane 7).

Activation of Pre 1 by the NH₂-Terminal Insertion Mechanism.

To ascertain whether VWbp activates ProT through the molecular sexuality mechanism, VWbp(1-474) was compared in kinetic assays of prethrombin 1 (Pre 1; ProT lacking the fragment 1 domain) activation to the NH₂-terminal deletion mutants, VWbp(2-474) and VWbp(3-474). Initial rates of ProT activation could not be used quantitatively in this analysis because of their inherent curvature, whereas initial rates of Pre 1 activation by VWbp(1-474) were nearly linear. Deletion of the first NH₂-terminal Val residue of VWbp reduced its activity at saturation ≈78%, whereas deletion of both Val¹ and Val² completely inactivated VWbp (Fig. 2). Qualitatively similar results were obtained with ProT, with no detectable activity for VWbp(3-474).

Fig. 2.

Pre 1 activation by NH₂-terminal truncation mutants of VWbp. Initial velocities (v_obs) of hydrolysis of 100 μM d-Phe-Pip-Arg-pNA are shown for mixtures of 1 nM Pre 1 as a function of VWbp(1-474) (filled circles), VWbp(2-474) (open circles), or VWbp(3-474) (filled triangles) concentration. The lines represent the least-squares fits of the quadratic binding equation.

Kinetic Analysis of ProT Activation.

The rates of tripeptide-p-nitroanilide (pNA) chromogenic substrate hydrolysis by ProT activated by either VWbp(1-263) or VWbp(1-474) showed upward curvature over time. Although this could be a result of a slow equilibration time for ProT·VWbp complex formation, varying the preincubation time of the reaction before substrate addition up to 1 h had very little effect on the curvature [see supporting information (SI)]. Because of the apparent effect that binding of substrate by ProT·VWbp(1-263) had on the activation kinetics, progress curves collected as a function of substrate and VWbp(1-263) concentration were truncated at ≤10% substrate depletion and analyzed individually with the classical hysteresis equation (Eq. 1, Materials and Methods) (17, 18). The excellent nonlinear least-squares fits indicated that the curvature initiated by substrate was a single exponential process. The observed rate constants increased hyperbolically with increasing VWbp(1-263) concentration and decreased with increasing substrate concentration. This ruled out a mechanism involving an unfavorable preexisting conformational equilibrium between active and inactive forms of ProT in which VWbp(1-263) bound only the active form that also bound substrate. This mechanism predicts a decrease in k_obs with increasing VWbp(1-263) concentration, rather than the observed increase. The decease in k_obs with increasing substrate concentration suggested that substrate binding followed the slow conformational change.

To evaluate the mechanism further, 2 chromogenic substrates (d-Phe-Pip-Arg-pNA and Tosyl-Gly-Pro-Arg-pNA) having a ≈5-fold difference in K_m for thrombin were used in full time-course activation progress curve analysis. Progress curves collected as a function of substrate and VWbp(1-263) concentration were analyzed simultaneously by numerical integration of the differential rate equations for candidate mechanisms combined with nonlinear least-squares fitting with DYNAFIT (19). The hysteretic mechanism shown in Scheme 1 produced a consistent fit to the data for both substrates (Fig. 3), with the parameters listed in Table 1. To obtain the best fit, product inhibition was included in the mechanism as a single binding step (not included in Scheme 1). Including a conformational equilibrium succeeding product binding did not improve the fit. In Scheme 1, the initial inactive ProT·VWbp(1-263) complex is formed with low affinity (K_D 2.2–2.4 μM) in a rapid equilibrium step. ProT·VWbp(1-263) is in an unfavorable, slow equilibrium with active ProT*·VWbp(1-263), and tight binding of substrate to the ProT*·VWbp (1-263) complex shifts the equilibrium toward the active form. The slow shift in equilibrium increases the rate of substrate hydrolysis, responsible for the upward curvature of the progress curves. The fitted parameters showed that both substrates exhibited 2- to 4-fold-lower K_m for the ProT*·VWbp(1-263) complex compared with thrombin (Table 1). The analysis also revealed an unfavorable equilibrium constant (K_con 9 ± 1 and 9 ± 6) for the slow conformational change, and excellent agreement between the individual rate constants (k_C1, k_C2) to form ProT*·VWbp(1-263) for the 2 substrates (Table 1). In addition, the apparent affinity (K_D) of ProT for formation of the initial inactive ProT·VWbp(1-263) complex was nearly identical with either substrate (2.4 μM and 2.2 μM) (Table 1). Results obtained with other substrates with K_m in the 100- to 300-μM range were not fit well by the mechanism in Scheme 1 because of the importance of obtaining near-saturation within the experimentally accessible concentration range. A more complex mechanism involving significant activity for the inactive complex failed to produce consistent fits or reasonable parameters. The possibility that autocatalytic cleavage of ProT to Pre 1, previously reported for SC (6) and SC(1-325) (20), accounted for the hysteresis was evaluated by Western blot analysis under conditions identical to those used for collecting the progress curves (see SI). At the highest VWbp(1-263) concentration (10 μM) in the kinetic studies, 23 ± 9% (mean and range) of ProT was converted into Pre 1 over the whole time course for both substrates. At 1 μM VWbp(1-263), Pre 1 formation was reduced to 1–10%. Analysis of individual progress curves at 10 μM VWbp showed little effect on their shape, demonstrating that Pre 1 formation did not account for the observed hysteresis.

Scheme 1.

Fig. 3.

Full time-course analysis of the kinetics of ProT activation by VWbp(1-263). (A) Progress curves for hydrolysis of d-Phe-Pip-Arg-pNA by mixtures of VWbp(1-263) and ProT, at 200 μM substrate (black points), contained 1 nM ProT and 0.025, 0.075, 0.15, 0.3, 0.6, 3, or 10 μM VWbp. Substrate depletion assays (red points) contained 1 nM ProT, 1 μM VWbp, and concentrations of 7, 20, 40, 78, or 116 μM substrate. (B) Progress curves for hydrolysis of 200 μM Tosyl-Gly-Pro-Arg-pNA (black points) contained 1 nM ProT and 0.1, 0.3, 0.5, 0.75, 1, 3, or 10 μM VWbp. Substrate depletion assays (red points) contained 1 nM ProT, 5 μM VWbp, and 25, 50, 75, 100, 150, or 500 μM substrate. Representative examples of the whole dataset are shown for each substrate, which contained 17–24 progress curves over 0.025–10 μM VWbp and 5–500 μM substrate. Data are shown as every tenth point, and the fit with the mechanism in Scheme 1 and parameters in Table 1 is represented by solid lines.

Table 1.

Summary of kinetic parameters (±2 SD) for ProT activation by VWbp(1-263) from the hysteretic mechanism shown in Scheme 1, where K_I is the dissociation constant for competitive product inhibition

Parameters	d-Phe-Pip-Arg-pNA	Tosyl-Gly-Pro-Arg-pNA
Km (thrombin)	2.5 ± 0.2 μM	11.3 ± 0.3 μM
kcat (thrombin)	77 ± 1 s−1	164 ± 1 s−1
KD	2.4 ± 0.1 μM	2.2 ± 0.2 μM
Kcon (kC2/kC1)	9 ± 1	9 ± 6
kC1	0.0052 ± 0.0001 s−1	0.0053 ± 0.0004 s−1
kC2	0.049 ± 0.005 s−1	0.047 ± 0.028 s−1
Km	0.6 ± 0.2 μM	6 ± 4 μM
kcat	95 ± 1 s−1	50 ± 1 s−1
KI	2.2 ± 0.3 μM	30 ± 20 μM

Competitive Binding of VWbp(1-263) and [5-F]Hir(54-65) to ProT.

To investigate the binding interactions involved in ProT activation by VWbp, independent of substrate-mediated effects, competitive binding titrations of the proexosite I-specific probe [5-F]Hir(54-65) and VWbp(1-263) with ProT were performed (Fig. 4). The decrease in fluorescence of [5-F]Hir(54-65) titrated with ProT in the absence of VWbp(1-263) gave a K_D consistent with previous results (K_P 2.8 ± 0.3 μM) (21). Analysis of the fluorescence changes in the presence and absence of VWbp(1-263) as a function of native ProT concentration gave a K_D of VWbp(1-263) for proexosite I on ProT of 2.5 ± 0.3 μM, nearly identical to the values determined kinetically. Furthermore, the ability of VWbp(1-263) to compete with the labeled peptide for binding to ProT indicates that VWbp binds to proexosite I (13).

Fig. 4.

Competitive binding of native ProT to [5-F]Hir(54-65) and VWbp(1-263). Fractional change in fluorescence (ΔF/F_o) of 48 nM [5-F]Hir(54-65) as a function of total native ProT concentration ([ProT]_o) at 0 (filled circles), 10 (open circles), and 20 μM (filled triangles) VWbp(1-263). The lines represent the simultaneous fit by the cubic equation with the parameters given in the text.

Clotting of Fbg by ProT·VWbp(1-263).

We examined human Fbg as a potential substrate by monitoring the clotting of Fbg by mixtures of ProT and VWbp(1-263), as detected by the increase in turbidity (Fig. 5). The capacity of VWbp(1-263) to trigger formation of a Fbn clot demonstrates that Fbg is the first substrate identified for the ProT·VWbp(1-263) complex. The complex displayed efficient clotting of Fbg at 10 nM ProT, and in contrast to immediate cleavage of Fbg by 10 nM thrombin or ProT·SC(1-325) complex, which showed a brief ≈30-sec lag, all of the ProT·VWbp(1-263) assays showed a much longer, reproducible lag in the time course (Fig. 5). No increase in turbidity was seen until ≈220 sec after addition of ProT at 1 μM VWbp(1-263), implicating the hysteretic mechanism in recognition of the natural substrate. Although an exhaustive survey of other potential substrates was not done, kinetic assays revealed no activation of the thrombin substrates, protein C, factor V, or factor XI, or 2 other coagulation zymogens (factor IX and factor X) by ProT·VWbp(1-263). In addition, the thrombin-selective serpins antithrombin and heparin cofactor II did not inhibit the activity of the ProT·VWbp(1-263) complex in the presence or absence of heparin.

Fig. 5.

Clotting of Fbg by mixtures of ProT and VWbp(1-263). Increase in turbidity for mixtures of 0.5 mg/mL Fbg and 10 nM thrombin (blue), 10 nM ProT (black), 10 nM ProT·SC(1-325) (red), or 10 nM ProT and VWbp concentrations of 0.1 (a), 0.3 (b), 0.5 (c), or 1 μM (d) (green).

Discussion

Our results support the conclusion that VWbp is a previously uncharacterized nonproteolytic ProT activator that utilizes molecular sexuality and employs a unique substrate-activated hysteretic kinetic mechanism. Active site-specific labeling of ProT with a fluorescent probe in mixtures with VWbp(1-263) or VWbp(1-474) provided a direct demonstration of nonproteolytic activation of the zymogen. These experiments also indicate that the SC(1-325)-homologous D1 and D2 domains of VWbp(1-263) are sufficient for ProT activation. Total loss of activity accompanying deletion of the first 2 NH₂-terminal residues of VWbp shows that it activates the ProT derivative, Pre 1, and ProT through the molecular sexuality mechanism.

Our studies and previous reports indicate that SC(1-325) has the same maximal chromogenic substrate activity as SC(1-660) (22). In contrast, ProT activation by full-length VWbp(1-474) displays higher activity than VWbp(1-263), although VWbp(1-474) also shows hysteresis. The finding that Pre 1 activation by VWbp(1-474) shows less hysteresis than ProT implicates the fragment 1 domain in the mechanism. The COOH-terminal region of VWbp is associated with the higher activity of the ProT·VWbp complex, possibly through additional interactions that increase substrate affinity and/or hydrolysis. Kinetic analysis of ProT activation by VWbp demonstrates a hysteretic model of enzyme activation, where there is a slow response to an abrupt stimulus that controls the observed rate of catalysis (17). The hysteretic kinetic concept is applicable to a number of enzymes involved in metabolic regulation and can result from isomerization, ligand displacement, or enzyme polymerization (17). Although slow hysteretic transitions have been examined for several monomeric enzymes (18, 23), our results demonstrate hysteresis in serine proteinase zymogen activation.

As predicted from the kinetic model (Scheme 1), the K_D for formation of the inactive ProT·VWbp(1-263) complex and the rate constants for the slow conformational change were independent of the structures and kinetic parameters for 2 tripeptide-pNA substrates. The relatively low affinity of ProT·VWbp(1-263) complex formation revealed an additional disparity between the behavior of VWbp(1-263) and SC. Whereas SC(1-325) binds to native ProT with extremely high affinity (K_D 17–72 pM) (24), the initial affinity of ProT and VWbp(1-263) was 2.2–2.4 μM from the kinetic analysis and 2.5 μM from competitive binding experiments. The kinetic mechanism suggests that VWbp(1-263) binds the native ProT zymogen with a weaker affinity than the zymogen with a substrate occupying the active site. The weak affinity of VWbp(1-263) for ProT and the hysteretic generation of proteolytic activity represent a unique regulatory mechanism in which the initially low affinity prevents premature activation of ProT, and the hysteresis delays the response to substrate, restricting Fbn formation. The hysteretic mechanism thus functions to increase the specificity of VWbp for Fbn formation compared with SC.

The NH₂-terminal insertion pocket, catalytic site, and certain regulatory exosites on coagulation proteinases are allosterically linked (25, 26). In the classical zymogen activation mechanism, these sites are rapidly expressed as part of the zymogen-to-proteinase conformational transition. Based on the linkage between these sites, the hysteretic conformational change between inactive and active forms of ProT·VWbp complex represents these events as a slow transition, unique among zymogen activation mechanisms. The need for substrate binding to complete the conformational change indicates that either NH₂-terminal insertion is an unfavorable intramolecular equilibrium or that the VWbp Val-Val NH₂ terminus inserts normally but does not result in an optimal structure to effect full activation. Mutation of VWbp(1-474) Val¹ to Ile to mimic the SC NH₂-terminal dipeptide did not detectably affect the hysteretic behavior, indicating that the difference in the dipeptides is not responsible. For α-chymotrypsin, interconversion of an inactive conformation at high pH to the active form at neutral pH is controlled by protonation of the α-amino group of Ile¹⁶ (27). In pH-jump experiments, the rate constants for formation of the chymotrypsin active site correspond to k_C1 and k_C2 of 3 s⁻¹ and 0.6 s⁻¹, respectively, and K_con of 0.2 (27). If the values for chymotrypsin are taken as representative of the rate of NH₂-terminal insertion and activation of the catalytic site for ProT, the values for ProT·VWbp(1-263) are 590-fold (k_C1) and 12-fold (k_C2) slower. This suggests that the activating conformational change for ProT·VWbp(1-263) is not limited by the rate constants for insertion and conformational activation, but by additional interactions of ProT with VWbp(1-263) that slow the activation process.

Screening of other coagulation zymogens for direct activation by VWbp indicates that it does not activate protein C, factor IX, factor X, or factor XII. The finding that VWbp binds to the low-affinity precursor form of exosite I on ProT, and the established role of exosite I in thrombin interactions with protein substrates (including Fbg), inhibitors, and regulatory proteins (25) indicates that ProT·VWbp will exhibit highly restricted substrate and inhibitor specificity, preventing interactions with thrombomodulin, factor V, factor VIII, heparin cofactor II, and protease activated receptor-1 (25). Fbg substrate recognition by activation of proexosite I in the ProT·VWbp complex is not possible because this site is blocked by VWbp. Neither VWbp nor ProT alone bind Fbg, indicating that Fbg substrate recognition by the ProT·VWbp complex is mediated by expression of a new exosite. The occupation of exosite II in ProT by the fragment 2 domain restricts further the substrate specificity of VWbp by blocking heparin-accelerated inhibition by antithrombin, as well as exosite II-dependent binding of factor V, factor VIII, and the platelet receptor GPIbα (25). The studies show that Fbg is the only established substrate of the ProT·VWbp(1-263) complex that binds with sufficient affinity to shift the hysteretic conformational equilibrium to the active complex. The lag observed in the time course of Fbg clotting assays initiated with ProT-VWbp(1-263) mixtures suggests that the dependence of the hysteretic kinetic mechanism on a protein substrate restricts activation of ProT by VWbp to areas rich in Fbg. Localized depletion of Fbg in an endocarditis vegetation may result in shutting off ProT·VWbp by the more rapid reversal of the conformational change. Additional results with VWbp(1-474) show no significant effect of the plasma concentration of VWF (150 nM) on the hysteretic mechanism of ProT activation, for either untreated VWF or the extended conformation induced by vortexing VWF under conditions that enhance VWF cleavage by ADAMTS13 (28).

The results of this study characterize the coagulation-specific molecular mechanisms behind a potential virulence factor from S. aureus. As the first additional member of the ZAAP family of bifunctional conformational zymogen activators and adhesion proteins, elucidation of the role of VWbp in activation of coagulation during endocarditis may aid in the development of mechanism-based therapies. Further clarification of the interactions of VWbp with both ProT and plasma adhesion proteins will offer insight into how VWbp may serve to specifically localize ProT activation in the course of S. aureus infection. This raises the question of why are there 2 ProT activators, VWbp and SC, secreted by S. aureus. All S. aureus strains that contain the gene for VWbp also contain the gene for SC (11). The evolutionary advantage responsible for maintaining both activators is likely because of the differences in the adhesion protein targets and the kinetic mechanisms of ProT activation. In the context of endocarditis, the initial vascular injury of the endothelium on heart valves at high shear rates activates coagulation, with VWF mediating the early stages of platelet accumulation by tethering through exposed subendothelial collagen and the platelet GPIbα receptor, followed by conversion of Fbg to Fbn within the thrombus (29). Additional binding of platelets by VWF occurs after platelet activation by thrombin and depends on the platelet integrin α_IIbβ₃ (30). S. aureus concurrently adheres to multiple extracellular matrix proteins, including fibronectin, collagen, and Fbn, through microbial surface components recognizing adhesive matrix molecules (MSCRAMMs) (31). Through initial association with VWF at transient sites of vascular injury, VWbp could restrict pathologic Fbn deposition to focal points in the vasculature, establishing new locations for bacterial dissemination. Thus, VWbp may catalyze the first assault in Fbn deposition, followed by rampant SC-mediated Fbn deposition and vegetation growth.

Materials and Methods

Protein and Peptide Purification.

Human ProT, Pre 1, thrombin, Fbg, Y⁶³-sulfated fluorescein-labeled hirudin (54-65) [[5-F]Hir(54-65)], and ATA-FPR-CH₂Cl were prepared as described (16, 24, 32).

Cloning, Expression, and Purification of VWbp Constructs.

The VWbp(1-474) gene was amplified from S. aureus Mu50 strain (ATCC) genomic DNA and cloned into a modified pET30b(+) vector (Novagen) (24) by using NcoI and XhoI restriction sites. VWbp was expressed with an NH₂-terminal His₆-tag and tobacco etch virus (TEV) proteinase cleavage site in Rosetta 2 (DE3) pLysS Escherichia coli (Novagen) with isopropyl-d-thiogalactopyranoside induction. Recombinant VWbp protein was extracted from inclusion bodies by the use of 3 M NaSCN buffer (pH 7.4) and purified by Ni²⁺-iminodiacetic acid chromatography. The His₆-tag was removed by overnight incubation with a 1:10 molar ratio of TEV proteinase to fusion protein similar to the procedure described for recombinant streptokinase (33). Mutants of VWbp(1-474) (VWbp(2-474), VWbp(3-474), and VWbp(1-263)) were generated by QuikChange site-directed mutagenesis (Stratagene). NH₂-terminal sequencing of all of the VWbp constructs confirmed the correct Val-Val-Ser-Gly-Glu sequence. VWbp concentration was determined from the 280-nm absorbance by using the following calculated absorption coefficients [(mg/mL)⁻¹ cm⁻¹] (34) and molecular weights: VWbp(1-263), 0.582, 30,700; VWbp(1-474), 0.488, 55,000.

Active Site-Specific Labeling of ProT.

VWbp(1-263) or VWbp(1-474) (50 μM) were incubated with ProT (15 μM) and ATA-FPR-CH₂Cl (125 μM) to covalently inactivate the ProT catalytic site formed in the active complex. The inhibited complex was incubated with 0.1 M NH₂OH in the presence of 33 μM 5-IAF to label the generated thiol. Probe incorporation (0.84–0.86 mol probe/mol ProT) was determined as described (16).

Pre 1 Activation Kinetic Titrations.

Titrations of Pre 1 activity as a function of VWbp(1-474), VWbp(2-474), or VWbp(3-474) concentration were measured by the increase in the initial rate of hydrolysis of 200 μM d-Phe-Pip-Arg-pNA at 405 nm and 25 °C. Pre 1 was incubated with VWbp for 20 min in 50 mM Hepes, 110 mM NaCl, 5 mM CaCl₂, 1 mg/mL polyethylene glycol (PEG) 8000 (pH 7.4) before substrate addition. The maximum velocity was determined by nonlinear least-squares analysis of the hyperbolic titrations.

ProT Activation Kinetics.

Progress curves for hydrolysis of 2 thrombin-specific chromogenic substrates (d-Phe-Pip-Arg-pNA and Tosyl-Gly-Pro-Arg-pNA) by mixtures of ProT and VWbp(1-263) were measured in the buffer described above. Full time-course assays were performed under both saturating and limiting substrate concentrations, with continuous data collection for 2 h, or until A_{405 nm} = 1.0. Proteins were incubated for 20 min at 25 °C before initiating the reactions by addition of substrate. Analysis of individual progress curves was performed on data truncated to ≤10% substrate depletion by fitting the integrated equation for the classical hysteresis mechanism,

where P_t is the product formed at time t, v_f, and v₀ are the final and initial velocities, and k_obs is the observed first-order rate constant (17). This equation is applicable only under conditions of minimal substrate depletion and product inhibition. Analysis of progress curves from full time-course assays under saturating and substrate depletion concentrations, and as a function of VWbp(1-263) concentration, were analyzed by simultaneous nonlinear least-squares fitting of the numerically integrated differential rate equations for candidate mechanisms with DYNAFIT (19). The rate constants for initial binding of ProT and VWbp and binding of substrate were assumed to be diffusion controlled (10⁸ M⁻¹s⁻¹) rapid equilibrium steps.

Fluorescence Titrations.

Competitive binding of [5-F]Hir(54-65) and VWbp(1-263) to ProT was measured in titrations of mixtures of 48 nM labeled peptide and VWbp as a function of ProT concentration in the same buffer used for the kinetics. Continuous time-course measurements were taken for each point in the titration that included a buffer blank, [5-F]Hir(54-65) alone, addition of VWbp(1-263), and addition of native ProT. This separated the fluorescence change measured in the first 2 min, representing the rapidly established competitive binding equilibria, from slower proteolysis of the ProT·VWbp complex. Fluorescence (8-nm slits) at 520 nm with excitation at 491 nm was measured with an SLM 8100 spectrofluorometer at 25 °C, in PEG 20,000-coated acrylic cuvettes. The fluorescence changes expressed as (F_obs − F_o)/F_o = ΔF/F_o in the absence and presence of 2 fixed concentrations of VWbp(1-263) (10 and 20 μM) were fit simultaneously by the cubic equation for tight competitive binding to obtain the K_Ds for [5-F]Hir(54-65) binding to ProT (K_P) and VWbp(1-263) binding to ProT (K_C), with the stoichiometry fixed at 1 (32).

Fbn Turbidity Assays.

Cleavage of Fbg by either thrombin, ProT·SC(1-325), or ProT·VWbp(1-263) complexes was monitored from the increase in turbidity at 450 nm at 25 °C in the above-described buffer by using a microtiter plate reader. Fbg was added (0.5 mg/mL) to various concentrations of VWbp(1-263), 10 nM SC(1-325), or buffer, and reactions were started immediately by addition of 10 nM ProT or thrombin.