Effect of Zymogen Domains and Active Site Occupation on Activation of Prothrombin by von Willebrand Factor-binding Protein*

Background: Conformational prothrombin activation by vWbp from Staphylococcus aureus causes fibrin formation. Results: Binding of vWbp with highest affinity to prothrombin species with occupied active sites is strengthened by the COOH-terminal half of vWbp. Conclusion: Affinity of vWbp for prothrombin is allosterically modulated by the presence of substrate and zymogen domains. Significance: vWbp exploits the physiological mechanisms of thrombin regulation to initiate abnormal coagulation. Prothrombin is conformationally activated by von Willebrand factor-binding protein (vWbp) from Staphylococcus aureus through insertion of the NH2-terminal residues of vWbp into the prothrombin catalytic domain. The rate of prothrombin activation by vWbp(1–263) is controlled by a hysteretic kinetic mechanism initiated by substrate binding. The present study evaluates activation of prothrombin by full-length vWbp(1–474) through activity progress curve analysis. Additional interactions from the COOH-terminal half of vWbp(1–474) strengthened the initial binding of vWbp to prothrombin, resulting in higher activity and an ∼100-fold enhancement in affinity. The affinities of vWbp(1–263) or vWbp(1–474) were compared by equilibrium binding to the prothrombin derivatives prethrombin 1, prethrombin 2, thrombin, meizothrombin, and meizothrombin(des-fragment 1) and their corresponding active site-blocked analogs. Loss of fragment 1 in prethrombin 1 enhanced affinity for both vWbp(1–263) and vWbp(1–474), with a 30–45% increase in Gibbs free energy, implicating a regulatory role for fragment 1 in the activation mechanism. Active site labeling of all prothrombin derivatives with d-Phe-Pro-Arg-chloromethyl ketone, analogous to irreversible binding of a substrate, decreased their KD values for vWbp into the subnanomolar range, reflecting the dependence of the activating conformational change on substrate binding. The results suggest a role for prothrombin domains in the pathophysiological activation of prothrombin by vWbp, and may reveal a function for autocatalysis of the vWbp·prothrombin complexes during initiation of blood coagulation.

molecules, but it also impacts binding of exogenous ligands such as the inhibitor hirudin (2).
The human pathogen Staphylococcus aureus secretes two proteins that can each interact with proexosite I on ProT, staphylocoagulase and von Willebrand factor-binding protein (vWbp) (11)(12)(13). Both are potent procoagulant cofactors capable of triggering rapid clotting of human plasma through direct cleavage of fibrinogen into fibrin by the activator⅐ProT complexes, but this activity is not reliant on proteolysis of the activation loop of ProT. Instead, conformational activation of the zymogen occurs through direct binding interactions with the activator and, ultimately, insertion of the NH 2 terminus of either staphylocoagulase or vWbp into the NH 2 -terminal binding cleft in the catalytic domain of ProT (13,14), forming the salt bridge with Asp 194 (chymotrypsinogen numbering) that is characteristic of proteolytic activation (3). In contrast to the usually minor effects seen with other ligands, binding of staphylocoagulase or vWbp to ProT generates what could be described as a definitive allosteric transition, where an inactive zymogen precursor is completely converted to an active species through a nonproteolytic alternative to its common activation pathway.
We previously described the molecular mechanism of vWbp procoagulant activity, and in addition, identified the role for a substrate-induced hysteretic kinetic mechanism for activation of ProT by vWbp, distinct from staphylocoagulase (13). Hysteresis is classically defined for a number of regulatory metabolic enzymes as a slow transition or conformational change initiated by various processes, including polymerization and ligand displacement (15,16). The key initiator of hysteretic behavior in vWbp-mediated ProT activation is a tight-binding substrate that alters the slow conformational equilibrium between inactive and active forms of the vWbp⅐ProT complex (13), which represents the first example of such hysteretic control in serine proteinases.
In addition to the established effect of substrate, preliminary kinetic and binding data indicated that the presence of the fragment domains of ProT, in particular fragment 1, imparts either a steric or an allosteric impediment to fully productive binding of vWbp. The present study defines the contribution to affinity of F1 and F2 and the catalytic domain of ProT and the predominantly unstructured COOH-terminal half of vWbp, as well as the magnitude of influence that active site occupation by a structural substrate mimic has on the binding behavior of vWbp. The preferential binding of certain ProT proteolysis products to vWbp likely has a significant influence on the pathophysiological behavior of vWbp and may be essential to understanding its role as a staphylococcal virulence factor.
Kinetic The progress curves of p-nitroaniline formation were analyzed by simultaneous nonlinear least-squares fitting in KinTek Global Kinetic Explorer using the hysteretic kinetic model described previously (13,24,25), assuming diffusion-controlled rapid equilibrium binding steps.
Competitive Fluorescence Titrations with Active ProT Derivatives-Continuous fluorescence intensity measurements were taken in 50 mM HEPES, 110 mM NaCl, 5 mM CaCl 2 , 1 mg/ml PEG 8000, pH 7.4, at 25°C. Fluorescence was monitored for a buffer blank, [5F]Hir(54 -65) alone, after the addition of vWbp(1-263) or vWbp , and immediately after the addition of competitors of fluorescent peptide binding (Pre 1, Pre 2, T, or MzT). Data were collected with either an SLM 8100 or a PTI QuantaMaster spectrofluorometer at ex ϭ 491 nm and em ϭ 520 nm. The fractional change in fluorescence was calculated as (F obs Ϫ F o )/F o ϭ ⌬F/F o , and the data for each experiment were fit globally with the cubic equation for tight competitive binding to obtain the dissociation constants and the stoichiometric factor for each competitor (17). The Gibbs free energy of binding (⌬G binding ) was calculated, using the equation ⌬G binding ϭ RTlnK D , where R ϭ 1.98 cal ϫ mol Ϫ1 ϫ degree Ϫ1 and T ϭ 298.15 K (25°C). In

Kinetic Analysis of ProT Activation by vWbp(1-474)-To
confirm the role of hysteresis in the functional activity of fulllength vWbp, vWbp(1-474) was employed in ProT activation assays. Although the degree of curvature in the rate of substrate cleavage seen with vWbp(1-474) is less than that of vWbp(1-263), the data were fit very well by the established hysteretic kinetic mechanism (Scheme 1) (Fig. 1). The calculated K D for the initial formation of the vWbp⅐ProT complex was 25.3 Ϯ 0.1 nM, ϳ100-fold tighter binding than vWbp(1-263) (13), verifying that the COOH-terminal half of vWbp contributes greatly to the affinity of vWbp for ProT. The overall equilibrium constant for the conformational change between active vWbp⅐ProT* and inactive forms of the vWbp⅐ProT complex was also over 3-fold more favorable for the active form (K con 3.1 Ϯ 1.8), primarily due to ϳ3-fold slower reverse rate constant (k C2 0.0185 Ϯ 0.0005 s Ϫ1 ) with a similar forward rate constant (k C1 0.00592 Ϯ 0.0004 s Ϫ1 ) for the slow conformational change (13). The remaining parameters for chromogenic substrate hydrolysis (K m 3.61 Ϯ 0.01 M; k cat 63 Ϯ 1 s Ϫ1 ) showed an ϳ6-fold increase in K m and a 34% decrease in k cat when compared with those determined for vWbp(1-263), giving only a slightly lower (6-fold) specificity constant (k cat /K m ) (13).
Binding of vWbp  or vWbp  to Active Siteblocked ProT Derivatives-To isolate the effect of active site occupation on the affinity of ProT and its derivatives for vWbp, both the zymogen and the active enzyme forms were labeled with the covalent, active site-specific inhibitor FPR-CH 2 Cl. The conformational changes associated with labeling with the inhibitor mimic the catalytic transition state of a bound substrate (26), a process predicted to substantially increase the affinity of vWbp for ProT due to the influence of substrates on the kinetic mechanism of activation. The resulting higher affinities for the active site-blocked analogs precluded the use of the same experimental approach used for the active derivatives. The fluorescent hirudin peptide could not effectively compete with either vWbp construct for binding because of the massively higher affinity of the FPR derivatives. To address this problem, [TMR]FPR-ProT was employed as a competitive ligand with the FPR analogs due to its similar high affinity for vWbp.  Fig. 6).

DISCUSSION
The present results reveal a substantial role for both steric and allosteric changes in the mechanism of binding and activation of ProT by vWbp. Previous work characterized vWbp as  Table 1. Binding experiments were performed and analyzed as described under "Experimental Procedures."  Table 1. Binding experiments were performed and analyzed as described under "Experimental Procedures." interacting with proexosite I on ProT due to its ability to compete for binding with the COOH-terminal peptide of hirudin, Hir(54 -65) (13). Similarly to many exosite I-specific ligands, vWbp demonstrates increased affinity for ProT derivatives with a proteinase-like conformation, with the COOH-terminal region of vWbp also contributing to binding affinity and complex activity. The preference of vWbp for active site-occupied binding partners emphasizes the importance of substrate in the hysteretic activation mechanism. Together, these results indicate that vWbp is capable of exploiting the physiological mechanisms that govern thrombin recognition of other ligands.
Although vWbp is capable of binding (pro)exosite I and potentially altering the specificity of ProT, a thorough analysis of the kinetics of ProT activation by vWbp(1-474) indicates that other regions of the zymogen interact with the COOHterminal half of vWbp. Although the NH 2 -terminal half of vWbp (vWbp(1-263)) is predicted to have a secondary structure composed of two ␣-helical bundles (27), multiple prediction analyses on the COOH-terminal half of vWbp assign almost no regular structure, apart from a helical segment (ϳ50 residues) at the distal end (not shown). The higher activity and ϳ100-fold higher affinity of vWbp(1-474) in forming the initial vWbp⅐ProT complex revealed in the kinetic analysis likely results from additional binding interactions, which may affect the conformation of proexosite I or facilitate insertion of the NH 2 terminus of vWbp into the activation pocket on ProT. The unstructured region of vWbp may sterically alter the fragment domains to increase the overall affinity of vWbp for the zymogen. Whether this process is mediated by the largely disordered region or the smaller COOH-terminal helical portion of vWbp remains unknown.
The fragment domains of ProT present a potential hindrance to binding by a macromolecular ligand such as vWbp, but release of these domains is normally strictly regulated during ProT activation. When it is not bound to a membrane surface through Ca 2ϩ -mediated interactions between ␥-carboxyglutamic acid residues and phosphatidylserine (28), the orientation of F1 relative to the rest of the zymogen is largely undefined, although crystal structures of F1, F2, Pre 1, and MzT(-F1) have been solved (28 -31). Release of F1 through cleavage at Arg 155 is greatly inhibited in the presence of Ca 2ϩ (32,33), and F1 typically remains covalently linked to F2 in fragment 1.2 (F1.2) (34). Loss of F1 or F1.2 not only relieves potential blocking interactions for binding of vWbp, but also produces conformational changes within the activation domain of ProT itself (35,36) that could favor ligand-exosite association.
A role for both steric and allosteric processes is indicated by the results of the competitive binding studies using ProT analogs with unmodified active sites, which revealed substantial increases in the affinity of vWbp as F1 and F2 are removed from ProT. The greatest enhancement results from loss of F1, with ϳ30-fold higher affinity for vWbp(1-263) and a 6-fold change with vWbp(1-474). It is worth noting that although vWbp(1-474) is capable of binding to all of the ProT derivatives with higher affinities than vWbp(1-263), there is still an effect from the loss of F1 despite the significant contribution of the COOHterminal region of vWbp to activity and affinity. A direct comparison of the ⌬G values shows a 12-26% increase in binding free energy upon loss of F1 alone (Pre 1) for both vWbp constructs, but only a 4 -6% increase upon subsequent loss of F2 (Pre 2). The same pattern is not seen upon examination of the three active enzyme forms, T, MzT, and MzT(-F1). vWbp  binds indistinguishably to all three species (K D 1.3-1.6 nM), suggesting that the existence of a proteinase-like activation domain promotes an equally favorable structure for ligand binding. An identical relationship can be seen with vWbp(1-474), where although it is capable of binding the active enzymes with ϳ5-fold tighter K D values than vWbp(1-263), the presence of the fragment domains has no obvious effect. These results are consistent with previous studies on the influence of F1 and F2 on the expression of exosite I, where loss of F1 in Pre 1 increases the affinity of the exosite 7-fold for a fluorescently labeled hirudin peptide, but any modulating effect of F1 is canceled out by proteolytic activation of the zymogen (35).
What does this binding behavior indicate within the context of the hysteretic mechanism of activation by vWbp? An optimal ProT conformation that supports initial high affinity binding of vWbp (conformational selection) would seem to be implied from the binding studies alone (37). In contrast, the key step in the model of activation by vWbp occurs upon binding of a substrate into what is presumed an imperfectly formed, inactive active site within a relatively low affinity vWbp⅐ProT complex. Substrate serves to generate a more catalytically competent, high affinity complex, consistent with substrate-and cofactormediated induced fit mechanisms of activation. The activity of the complement protease factor D is regulated by binding of its substrate, C3b-complexed factor B (38,39). Similarly, factor VIIa exists in a state of incomplete activation with a conformation resembling an intermediate stage between zymogen and protease (40 -42). Requirement for binding its cofactor tissue factor (43) shows parallels to the unfavorable equilibrium that exists for the vWbp⅐ProT complex in the absence of substrate (Scheme 1).
The concept of substrate binding as a fundamental amplifier of ProT activation by vWbp is further strengthened by the equilibrium binding results with FPR-blocked ProT derivatives. Structural studies have shown that inhibition of thrombin by FPR-CH 2 Cl gives a similar orientation of residues in the substrate-binding cleft as is seen with binding of substrates, but distinctly different from changes seen with binding of hirugen to exosite I (44). Thus, corresponding FPR analogs of all the enzyme and zymogen derivatives of ProT were produced through either direct inhibition with FPR-CH 2 Cl or formation of a reversible, conformationally activated complex of the zymogen with staphylocoagulase (27), allowing incorporation of the inhibitor into the induced active site on the zymogenactivator complexes. The substrate-bound structural mimicry of the FPR analogs allowed vWbp(1-263) and vWbp(1-474) to bind with 4 -6-fold higher affinity, even with the zymogens containing F1/F2. Therefore, the vWbp⅐ProT complex does not  Table 1. Binding experiments were performed and analyzed as described under "Experimental Procedures."  Table 1. Binding experiments were performed and analyzed as described under "Experimental Procedures." simply require an "enzyme-like" orientation of the activation domain between the catalytic residues, oxyanion hole, and substrate specificity site, but it also calls for occupation of the S1 site by an appropriate substrate, consistent with the hysteretic mechanism of activation.
The involvement of substrate in the mechanism is clear, but does the presence of F1/F2 have any detectable effect on the binding of vWbp to derivatives containing an active catalytic domain? The design of the binding assays with the active ProT derivatives allowed observation of changes in fluorescence immediately upon the addition of the ProT analog to the [5F]Hir(54 -65) and vWbp ligand mixture, and most of the assays showed very rapid equilibrium between the competing proteins. The only discrepancies were witnessed upon either the addition of active MzT or in situ activation of Pre 1 to MzT(-F1), where a noticeably longer time period was required to reach equilibrium and a stable fluorescence change. Fig. 7A shows two representative fluorescence traces for MzT binding, with the assay containing vWbp(1-474) requiring an additional ϳ50 s to reach equilibrium. The fact that vWbp(1-263) does not show a lag suggests that the COOH-terminal half of vWbp has opposition from F1.2 in MzT and that the kinetics of binding are affected but not the fundamental equilibrium dissociation constant.
A different mechanism may be at work during conversion of Pre 1 to MzT(-F1) (Fig. 7B), where vWbp(1-263) and vWbp(1-474) both alter the kinetics of activation by ecarin to different degrees when compared with Pre 1 in the presence of [5F]Hir(54 -65) alone. Whether this is a consequence of partial blockage of areas on the zymogen required for interaction with ecarin is unknown, but it could also be influenced by conformational changes in Pre 1 induced by vWbp that may alter accessibility of the activation bond or orientation of the F2 domain. These findings suggest that F1 and F2 introduce some degree of hindrance to vWbp, even with an active catalytic domain and an exosite I conformation that favors high affinity binding.
Although vWbp can form an active complex with ProT exclusively through NH 2 -terminal insertion and substrate-dependent changes in conformation, previous studies of both vWbp (13) and staphylocoagulase (11) have identified a pattern of autocatalysis consistent with cleavage at two thrombin-sensitive sites. In the presence of vWbp, production of Pre 1 from ProT has been detected in reaction mixtures at both high and low concentrations of ProT, with formation of a species called Pre 2Ј after extended incubations at high ProT concentrations (13). The sequence of cleavage in vWbp-ProT mixtures corresponds with what is known about the mechanisms of feedback proteolysis in MzT and MzT(-F1), where both intramolecular and intermolecular events contribute to formation of proteol-  Table 1). Experiments were performed and analyzed as described under "Experimental Procedures." ysis products (45). This indicates that the proteolytic activity of vWbp⅐ProT* is consistent with thrombin-like behavior.
If vWbp can form a procoagulant complex with ProT that is not only capable of cleaving fibrinogen into fibrin, but also contains the potential to rid itself of its regulatory zymogen domains, what implications would this have for the pathological capacity of vWbp in a blood-borne staphylococcal infection? The most vital function of F1 for ProT is to mediate binding of the zymogen to the activated platelet or endothelial cell membrane (46), permitting association with factor Xa and factor Va in the prothrombinase complex for thrombin production. If vWbp can bind to membrane-bound ProT, as well as free ProT in the blood, premature release of high affinity procoagulant vWbp⅐Pre 1* complexes from membranes could occur in a pathological dissemination mechanism. This would have the amplified effect of liberating serpin-insensitive protein complexes to downstream locations where they could freely associate with both von Willebrand factor and fibrinogen, setting up new foci for fibrin deposition and bacterial colonization. Indirect effects related to other regulatory ligands would also occur, including blockage of all exosite I-specific ligands by the presence of vWbp and reduction of exosite II interactions from the presence of either F1.2 or F2. The end result would be vWbp⅐ProT* or vWbp⅐Pre 1* complexes with thrombin-like activity that are nearly impervious to inhibition or down-regulation, but that exhibit high specificity for the single substrate fibrinogen due to the driving force of the substrate-mediated mechanism of vWbp zymogen activation.