Coupling of conformational and proteolytic activation in the kinetic mechanism of plasminogen activation by streptokinase.

Binding of streptokinase (SK) to plasminogen (Pg) induces conformational activation of the zymogen and initiates its proteolytic conversion to plasmin (Pm). The mechanism of coupling between conformational activation and Pm formation was investigated in kinetic studies. Parabolic time courses of Pg activation by SK monitored by chromogenic substrate hydrolysis had initial rates (v(1)) representing conformational activation and subsequent rates of activity increase (v(2)) corresponding to the rate of Pm generation determined by a specific discontinuous assay. The v(2) dependence on SK concentration for [Lys]Pg showed a maximum rate at a Pg to SK ratio of approximately 2:1, with inhibition at high SK concentrations. [Glu]Pg and [Lys]Pg activation showed similar kinetic behavior but much slower activation of [Glu]Pg, due to an approximately 12-fold lower affinity for SK and an approximately 20-fold lower k(cat)/K(m). Blocking lysine-binding sites on Pg inhibited SK.Pg* cleavage of [Lys]Pg to a rate comparable with that of [Glu]Pg, whereas [Glu]Pg activation was not significantly affected. The results support a kinetic mechanism in which SK activates Pg conformationally by rapid equilibrium formation of the SK.Pg* complex, followed by intermolecular cleavage of Pg to Pm by SK.Pg* and subsequent cleavage of Pg by SK.Pm. A unified model of SK-induced Pg activation suggests that generation of initial Pm by SK.Pg* acts as a self-limiting triggering mechanism to initiate production of one SK equivalent of SK.Pm, which then converts the remaining free Pg to Pm.

Binding of streptokinase (SK) to plasminogen (Pg) induces conformational activation of the zymogen and initiates its proteolytic conversion to plasmin (Pm). The mechanism of coupling between conformational activation and Pm formation was investigated in kinetic studies. Parabolic time courses of Pg activation by SK monitored by chromogenic substrate hydrolysis had initial rates (v 1 ) representing conformational activation and subsequent rates of activity increase (v 2 ) corresponding to the rate of Pm generation determined by a specific discontinuous assay. The v 2 dependence on SK concentration for [Lys]Pg showed a maximum rate at a Pg to SK ratio of ϳ2:1, with inhibition at high SK concentrations.
[Glu]Pg and [Lys]Pg activation showed similar kinetic behavior but much slower activation of [Glu]Pg, due to an ϳ12-fold lower affinity for SK and an ϳ20-fold lower k cat /K m . Blocking lysine-binding sites on Pg inhibited SK⅐Pg* cleavage of [Lys]Pg to a rate comparable with that of [Glu]Pg, whereas [Glu]Pg activation was not significantly affected. The results support a kinetic mechanism in which SK activates Pg conformationally by rapid equilibrium formation of the SK⅐Pg* complex, followed by intermolecular cleavage of Pg to Pm by SK⅐Pg* and subsequent cleavage of Pg by SK⅐Pm. A unified model of SK-induced Pg activation suggests that generation of initial Pm by SK⅐Pg* acts as a self-limiting triggering mechanism to initiate production of one SK equivalent of SK⅐Pm, which then converts the remaining free Pg to Pm.
Streptokinase (SK) 1 is a protein from Streptococcus equisimilis that converts the zymogen, plasminogen (Pg), into the fibrin clot-dissolving serine proteinase, plasmin (Pm) (1). The domain organization of the circulating form of Pg ([Glu]Pg) consists of a 77-residue N-terminal peptide, five kringle domains, and a serine proteinase catalytic domain that is activated by cleavage at Arg 561 -Val 562 (2,3). Pm cleavage of [Glu]Pg releases the N-terminal peptide, yielding [Lys]Pg, which is more readily activated by plasminogen activators, including SK⅐Pg and SK⅐Pm (4 -11). The mechanism of Pg activation by SK is unique in that SK binding to Pg results in expression of an active catalytic site on the zymogen by a conformational change, without the normally required peptide bond cleavage (12,13). Conformational activation is followed by proteolytic conversion of Pg to Pm by cleavage of Arg 561 -Val 562 , and SK bound to Pm propagates proteolytic Pg activation. Recognition of Pg as a substrate of SK⅐Pm is mediated by expression of a specific substrate recognition exosite on the SK⅐Pm complex (14,15).
Both conformational and proteolytic activation reactions are known to contribute to SK-induced Pg activation, but different interpretations of the mechanism of conformational activation and its coupling to subsequent proteolytic Pm formation have been postulated. Quantitative equilibrium binding and kinetic analysis of conformational activation in the companion paper (16) demonstrated that it can be described by rapid and reversible SK binding and activation of the catalytic site. The activation equilibrium is more favorable for Pg species with an extended conformation, [Lys]Pg and [Glu]Pg at low chloride concentration (4,5,17), compared with [Glu]Pg in the compact conformation (11,18,19), due to the differential involvement of lysine-binding site interactions.
Previous steady-state kinetic studies of SK-initiated Pm formation concluded that Pm was formed initially by intramolecular proteolytic activation of Pg within the conformationally activated SK⅐Pg* complex (20). The subsequent findings that SK⅐Pm did not catalyze a mixture of SK, Pg, and bovine trypsin inhibitor to Pm and the resistance of diisopropyl fluorophosphate-inactivated SK⅐Pg* to Pg activators also supported the intramolecular cleavage hypothesis (21,22). More recent studies (14,23,24), however, provide evidence for an SK⅐Pg*⅐Pg ternary complex, suggesting that intermolecular cleavage of Pg bound to SK⅐Pg* in a substrate mode contributes to initial Pm formation. Equilibrium binding studies demonstrated that the reaction product [Lys]Pm binds SK with ϳ50,000and ϳ11,000-fold higher affinity to fluorescein-labeled Pg analogs (15,(25)(26)(27), and native Pg (16), respectively, compared with the initial substrate, [Glu]Pg. The large difference in affinity may result in displacement of Pg from the SK⅐Pg* complex by Pm, forming the very stable proteolytic Pg activator complex SK⅐Pm. The pathway of Pm generation from SK⅐Pg* and the source of SK⅐Pm are unclear, in part because of conflicting evidence for the intramolecular and intermolecular mechanisms.
The role of lysine-binding sites on the kringle domains of Pg and Pm in the mechanism of SK-induced Pg activation is also only partly understood. The lysine analog, 6-aminohexanoic acid (6-AHA), partially reduces binding affinity of SK for [Lys]Pg in parallel with inhibition of conformational activation, but does not affect [Glu]Pg (16,25,26), and inhibits overall formation of Pm (28,29). It appears that the affinity of SK binding to Pg and subsequent proteolytic formation of Pm are both inhibited by 6-AHA, but the relationship between these sources of inhibition and the basis for the differential effect on [Glu]Pg and [Lys]Pg activation are not clear.
The goal of the present studies was to apply the results of previous equilibrium binding and kinetic studies of SK-Pg and SK-Pm interactions in combination with initial rate kinetic analysis to define the coupling of conformational Pg activation to the proteolytic activation pathway of SK-induced Pm formation. The studies address the unresolved mechanism of proteolytic Pm formation, the relationship between catalysis by SK⅐Pg* and SK⅐Pm complexes, and regulation of these events by lysine-binding site interactions. The kinetics of proteolytic Pg activation and inhibition of Pm formation by saturating SK concentrations observed here is explained by a mechanism in which Pm is formed by intermolecular cleavage of Pg by the SK⅐Pg* complex. The binding and conformational activation equilibrium for [Lys]Pg, characterized in the companion paper (16), is partially facilitated by lysine-binding site interactions, whereas subsequent Pm formation is greatly dependent on these interactions. By contrast, the independence of [Glu]Pg conformational and proteolytic activation on lysine-binding sites accounts for the higher substrate specificity of the SK⅐Pg* complex for [Lys]Pg compared with [Glu]Pg. The combined equilibrium binding and kinetic data do not support a significant role for intramolecular Pm formation within the SK⅐Pg* complex. The studies provide the first quantitative kinetic analysis of the pathway of Pm formation by SK and demonstrate the critical role of differences in SK affinity for Pg activation species and the differential roles of lysine-binding site interactions in initiating and propagating the activation mechanism. A unified kinetic mechanism of Pg activation suggests that the activation process at concentrations of Pg Ͼ Ͼ SK is triggered by a self-limiting initial cycle of Pm formation catalyzed by SK⅐Pg* that results in sequestering of SK in the SK⅐Pm complex, which ultimately catalyzes full conversion of free Pg to Pm.

EXPERIMENTAL PROCEDURES
Protein Purification and Characterization- [Glu]Pg, [Lys]Pg, Pm, and SK were prepared and their concentrations determined as described in the companion paper (16).
Gel Electrophoresis-Pg activation reactions were initiated at 25°C by addition of SK in 50 mM Hepes, 0.125 M NaCl, 1 mM EDTA, 1 mg/ml polyethylene glycol 8000, pH 7.4, and inactivated with 100 M (D-Phe)-Phe-Arg-CH 2 Cl for 2 min at room temperature before preparation of gel samples and electrophoresis on SDS 4 -15% polyacrylamide gels. Gels were stained with Sypro Orange (Molecular Probes) and protein bands visualized with a 300-nm transilluminator.
Plasminogen Activation Kinetics-Activation of Pg by SK was measured by continuous monitoring of the increase in absorbance of D-Val-Leu-Lys-pNA (VLK-pNA) at 405 nm (⌬A 405 nm ). The parabolic progress curves of pNA formation were fit by Equation 1 (30,31) to obtain the initial rate of substrate hydrolysis at the beginning of the reaction (v 1 ) and the rate of increase in activity with time (v 2 ).  (15,31). Discontinuous Plasmin Assay-Discontinuous assays of the initial rates of Pm generation in reactions of Pg and SK were performed by incubation of SK and [Lys]Pg 1 for various times and quenching the reaction by addition of 1 M FFR-Pm and 10 mM 6-AHA to dissociate the SK⅐Pg* complex and to inhibit additional formation of Pm. Following 2 min of incubation, chromogenic substrate assays of quenched samples were initiated by addition of 50 M VLK-pNA, and the concentrations of Pm were calculated from slopes of the linear rates. The substrate concentration was chosen such that free Pm and SK-bound Pm had equal contributions to the rate, independent of SK concentration (15). Discontinuous assays of Pm generation were linear with time at up to 60% Pg depletion, and data were fit by linear least squares to determine the rates. Pre-quenched reactions in which 6-AHA and FFR-Pm were present in the incubation resulted in rates of essentially zero. The activities of quenched reactions were stable with time, exhibiting Ͻ0.5% increase in activity over 20 min after quenching. The assay yielded the Pm concentration accurately in mixtures of Pg or SK containing known levels of Pm.
Kinetic Model of Plasminogen Activation by Streptokinase-The dependences of v 1 and v 2 on SK concentration can be described by the mechanism in Schemes 1 and 2. In Scheme 1, SK binds to Pg with dissociation constant K A to form the conformationally activated SK⅐Pg* complex, which catalyzes chromogenic substrate (S) hydrolysis with Michaelis-Menten parameters K m and k c . Binding of Pg as a substrate of SK⅐Pg* with dissociation constant K S forms the ternary Michaelis complex (SK⅐Pg*⅐Pg) that generates Pm with the catalytic rate constant k 0 . As illustrated in Scheme 2, the SK⅐Pm complex can also catalyze Pg activation to Pm, and SK⅐Pm hydrolyzes the chromogenic substrate with Michaelis constants KЉ m and kЉ c . The extremely high affinity of SK for Pm ( , where [Pg] free is given by the cubic equation described previously (16). Under conditions where [Pg] free Ͻ Ͻ K S and [S] 0 Ͻ Ͻ K m , which were SCHEME 1 SCHEME 2 found to apply (16), Equation 3 can be reduced to Equation 4, and the proteolytic activation step can be treated as bimolecular with rate constant k I,g , where [Pg] free is defined by the quadratic equation in [Pg] 0 and [SK] 0 described in the companion paper (16). v 2 rates expressed as [Pm] s Ϫ1 were calculated by using KЉ m ϭ 300 M and kЉ c ϭ 34 s Ϫ1 as determined for the SK⅐Pm complex (15). The v 2 dependences were obtained from activation progress curves as a function of SK and Pg concentrations analyzed by fitting the simplified Equation 4 and the quadratic binding equation.
As will be shown, the concentrations of Pg were also much lower than KЈ S for activation of Pg as a substrate of the SK⅐Pm complex, indicating that the proteolytic reaction in Scheme 2 could also be represented as bimolecular, with rate constant k I,m . With the above assumptions for both proteolytic reactions in Schemes 1 and 2, Reactions 1-6 were simulated by numerical integration of the differential rate equations using Kinetics (ARSoftware).
In this model, Pm produced by the SK⅐Pg* complex with bimolecular rate constant k Pg is identified as Pm A , whereas Pm formed by the SK⅐Pm complex with rate constant k Pm is Pm B . The binding steps (Reactions 1, 3, and 5) were assumed to be rapidly established relative to the irreversible bimolecular proteolytic Reactions 2, 4, and 6. Diffusion-controlled values of k 1 and

RESULTS
Assignment of v 2 to the Rate of Pm Formation-As shown previously, continuous assays of Pg activation were parabolic, with the initial rate (v 1 ) representing substrate turnover by the conformationally activated SK⅐Pg* complex, followed by an in-crease in rate (v 2 ) thought to represent proteolytic Pm formation (16). A discontinuous assay specific for Pm was developed to verify that v 2 could be assigned to the rate of Pm formation. The assay was specific for Pm and was not affected by the presence of Pg or SK⅐Pg* as demonstrated by control experiments. Fig. 1A shows a comparison of Pm formation from a reaction of 10 nM [Lys]Pg 1 and 50 nM SK determined by the discontinuous and continuous methods. Pm formation as measured by the discontinuous Pm assay was linear for up to 500 s with a slope of 0.0072 Ϯ 0.0003 nM Pm s Ϫ1 . The initial Pm concentration was essentially zero (Ͻ1% of total Pg), and no lags or bursts in Pm formation were observed, indicating the absence of kinetically controlling intermediates of Pm generating reactions. The v 2 values from continuous assays of reactions of 10 nM [Lys]Pg 1 and 50 nM SK collected as a function of fixed levels of 50 -400 M VLK-pNA, when transformed as described under "Experimental Procedures," were linear with slopes equal to v 2 . The dependence of v 2 on chromogenic substrate concentration was consistent with hyperbolic inhibition of Pm formation with an apparent inhibition constant of 330 Ϯ 80 M. When the progress curves were corrected for inhibition by substrate, the average slope of Pm formation was 0.0059 Ϯ 0.0002 nM Pm s Ϫ1 , in good agreement with the rate measured by the discontinuous assay. Rates of Pm generation determined by the discontinuous and continuous assays in the presence of 200 M VLK-pNA for activation of 10 nM [Lys]Pg 1 by SK showed excellent agreement over a wide SK concentration range (Fig. 1B). The results demonstrated indistinguishable rates of Pm formation from the Pg activation reactions measured by the continuous and discontinuous assays.
Inhibition of Pm Formation by Excess SK-To characterize further the correspondence between v 2 and Pm formation and inhibition of the reaction by excess SK, the products of reac- Specific cleavage of SK to form the SK degradation product SKЈ (where SKЈ is the two-fragment complex generated by cleavage of streptokinase at Lys 59 ) (32) at low SK concentration and nonspecific degradation at high SK concentrations were consistent with previous results (15). These results demonstrated clearly that v 2 represented the rate of proteolytic Pm formation, which was inhibited by high SK concentrations.
Analysis of the Kinetics of SK-induced Pm Generation-Previous kinetic studies of Pg activation by SK suggested that Pm was initially formed either by intramolecular cleavage of Pg within the SK⅐Pg* complex (20 -22) or by intermolecular cleavage of Pg by SK⅐Pg* (14,23,24). To evaluate whether the intermolecular model of Pg activation (Scheme 1), without consideration of SK⅐Pm-catalyzed Pg activation (Scheme 2), fitted the observed kinetics, SK kinetic titrations of several fixed concentrations of Pg were performed and analyzed according to the simplified two-step form of this model shown by the upper reactions in Scheme 3 (Fig. 3).
In this scheme, SK binds to Pg with the dissociation constant K A , forming the conformationally activated SK⅐Pg* complex, which subsequently binds and cleaves Pg to Pm with bimolecular rate constant k Pg . Results in the companion paper (16) demonstrated that representation of the formation of SK⅐Pg* as a single binding and activation equilibrium, and the SK⅐Pg*catalyzed proteolytic reaction as bimolecular were justified.  (16). Simultaneous analysis of the v 2 data resulted in a value of 2.3 Ϯ 0.2 nM for K A , which was ϳ5-fold lower than the value obtained from the v 1 data and k Pg of 0.79 Ϯ 0.02 M Ϫ1 s Ϫ1 (Fig. 3A). Comparable values were obtained for reactions of 5-20 nM [Lys]Pg 2 with SK (Table I) Table I). These fits demonstrated weaker binding of SK to [Glu]Pg compared with [Lys]Pg, as indicated in direct binding studies (16), and a 20-fold reduction in Pm generation measured by k Pg (Table I).
6-AHA decreased v 2 for [Lys]Pg activation dramatically (Fig.  4), by ϳ50-fold compared with the maximum rate, resulting in more linear progress curves at all SK concentrations. The v 2 values resulting from activation of [Glu]Pg, in contrast to [Lys]Pg activation, were not affected significantly by 6-AHA and were comparable with those of [Lys]Pg in 6-AHA (Fig. 4,  inset). These results indicated that inhibition of Pm formation by 6-AHA was due to both partial reduction of the affinity of SK for [Lys]Pg and greater inhibition of subsequent proteolytic activation. By contrast, the affinity of SK for [Glu]Pg and its proteolytic conversion to Pm were not significantly lysine-binding site-dependent. Simulation of the Reaction Mechanism-The major assumption in the above analysis was that SK⅐Pm-catalyzed Pg activation (Scheme 2, see "Experimental Procedures") did not contribute significantly to the initial rate of Pm formation. This was examined further by simulation of the coupled SK⅐Pg* and SK⅐Pm reactions in Scheme 3 by numerical integration of the differential rate equations. The kinetic parameters for SK⅐Pmcatalyzed cleavage of [Lys]Pg (KЈ S and kЈ 0 in Scheme 2) were determined in v 2 measurements of [Lys]Pg activation by SK⅐Pm formed with an excess of Pm (data not shown), where Pm alone does not activate Pg (15,31). KЈ S and kЈ 0 for this reaction were 270 Ϯ 60 nM and 0.31 Ϯ 0.02 s Ϫ1 (Table I). KЈ S was Ͼ13-fold higher than the highest Pg concentration used experimentally (Fig. 3A), allowing the proteolytic reaction to be treated as bimolecular with rate constant k Pm as shown in Scheme 3. For the simulations, the bimolecular rate constants k Pg and k Pm and the dissociation constant for the SK⅐Pm complex (KЈ A ) were fixed at their determined values, whereas K A was assigned a value of 4.5 nM, comparable with values determined for [Lys]Pg from analysis of v 1 (16) and here for v 2 (2-12 nM). These parameters gave the dashed line in Fig. 1A for the time course of the increase in total Pm concentration, which agreed well with the rates measured by the continuous and discontinuous v 2 assays. Simulation of the data under the conditions of Fig. 1A showed that the linear rate was a very sensitive function of K A , with the best fit by a value 4.5 nM.
Given the sequential nature of the SK⅐Pg-and SK⅐Pm-catalyzed Pg activation reactions, it was unclear why the progress curves of Pm formation were apparently linear over a wide range of SK and [Lys]Pg concentrations and for up to 50 -60% substrate depletion. Simulation of the time courses of the appearance and disappearance of the reaction species at 10 nM [Lys]Pg and substoichiometric 5 nM SK with the above parameters gave the results shown in Fig. 5A, where a represents free Pm A formed by SK⅐Pg* and b represents SK⅐Pm A . The total Pm A concentration, free and SK-bound (Fig. 5A, aϩb), increased to a steady maximum, whereas the total Pm B concentration (cϩd) showed a lag, with an increase to a near-linear rate that was initially accounted for primarily by formation of SK⅐Pm B (Fig. 5A, d), consistent with the sequential reactions. The total Pm concentration increased, initially following the rate of Pm A formation, and curved upward slightly as Pm B was produced (Fig. 5A, Pm T ). Linear least squares analysis of the simulated initial rate of Pm T formation demonstrated that it represented the rate of total Pm A formation, as expected from the sequential reactions. Similar linear analysis of the rate of Pm T formation at [SK] 0 Ͻ [Pg] 0 for up to 50% Pg depletion gave rates that were overestimated by a maximum of 2-fold compared with the initial rates calculated with Equation 4. These rates corresponded closely to the values of v 2 determined experimentally (Fig. 3A).
Simulation of the reaction time course at excess SK concentration (50 nM) with the same parameters gave the results in Fig. 5B. Under these reaction conditions, the rate of total Pm formation (Fig. 5B, Pm T ) was quite linear and agreed well with the initial rate calculated for the increase in total Pm A concentration (aϩb) from Equation 4. Analysis of the reaction products with increasing SK concentration indicated that inhibition of the rate by excess SK was due to depletion of the free Pg concentration by formation of the SK⅐Pg* complex, with inhibition of the SK⅐Pm-catalyzed reaction for the same reason. It was apparent from these results that the modest upward curvature in the rate of total Pm generation from [Lys]Pg at substoichiometric levels of SK would not have been detectable in the experimental progress curves. It was also clear that to within a factor of 2-fold, the linear approximation of the rate at all of the concentrations studied corresponded to the initial rate of Pm formation catalyzed by the SK⅐Pg* complex. This was apparently due to the sequential reactions, the bimolecular reaction conditions of the proteolytic steps, their similar rate constants, differing only by 1.5-fold, and the common cause of inhibition by excess SK. This accounts for the consistent fit of the v 2 dependence for [Lys]Pg activation on SK concentration (Fig. 3A). Simulation of progress curves for this experiment showed that the lower value of K A ϳ2 nM fit the data better at

TABLE I
Binding and kinetic parameters determined for conformational and proteolytic Pg activation Dissociation constants for SK-Pg binding (K D ), kinetically determined apparent dissociation constants obtained from analysis of v 1 (K A (v 1 )) and from analysis of v 2 (K D (v 2 )), are listed for activation of the indicated forms of Pg. K D and K A (v 1 ) were taken from Ref. 16. Bimolecular rate constants determined for SK⅐Pg-catalyzed Pg activation (k Pg ), Michaelis-Menten constants, kЈ S , and kЈ 0 , and the bimolecular rate constant (k Pm ) for SK⅐Pm-catalyzed Pg activation are also listed. Experiments were performed, and the results are analyzed as described under "Experimental Procedures" and the text.  (Table I).
A model in which all Pm was generated by intramolecular cleavage of SK⅐Pg*, as proposed previously (20 -22), did not fit the observed v 2 data, particularly for [Lys]Pg, because it predicted no inhibition in v 2 with increasing SK concentration. Analysis of the data with an expanded version of Scheme 3, including the possibility of intramolecular proteolytic conversion of SK⅐Pg* to SK⅐Pm, did not fit the v 2 dependence on SK concentration as well and required a 10 -20-fold higher rate of intermolecular cleavage to account for the data. Including this first-order generation of SK⅐Pm resulted in higher rates of Pm formation in excess SK because the generation of SK⅐Pm was not inhibited by depletion of free Pg by SK binding. These results ruled out intramolecular cleavage as a significant reaction in activation of [Lys]Pg by SK.

DISCUSSION
The results of quantitative equilibrium binding and kinetic studies presented here and in the preceding paper (16) support a unified hypothesis for the mechanism of SK-induced Pg activation to Pm. As illustrated in Fig. 6, rapid equilibrium binding of SK to Pg is accompanied by reversible conformational activation of the catalytic site in the Pg*⅐SK complex.
[Lys]Pg binds SK with ϳ12-fold higher affinity than [Glu]Pg due to the contribution of lysine-binding site interactions (not illustrated) and the absence of such interactions for [Glu]Pg. When lysinebinding sites are blocked with 6-AHA, SK binds and conformationally activates [Glu]Pg and [Lys]Pg with essentially equivalent affinity (16). The initially formed Pg*⅐SK complex with Pg* bound in the catalytic mode binds free Pg in a substrate mode to form the Pg*⅐SK⅐Pg ternary Michaelis complex. Binding of substrate Pg is likely mediated by a specific recognition exosite expressed on the Pg*⅐SK complex, as has been shown for Pm⅐SK (15). If true, this feature of the model would favor free Pg binding to the Pg*⅐SK complex in the substrate mode over binding in nonproductive complexes because the exosite is not expressed on free SK, Pg, or Pm (15). This suggests further that the productive Pg*⅐SK⅐Pg ternary complex would be as-sembled as illustrated, by sequential binding of Pg in the catalytic mode followed by Pg binding as a substrate.
Intermolecular proteolytic cleavage of Pg in the Pg*⅐SK catalytic cycle (Fig. 6) generates Pm, which dissociates from the substrate/product site. Pg substrate binding and catalysis measured by the bimolecular rate constant was independent of 6-AHA for [Glu]Pg, whereas [Lys]Pg activation was inhibited ϳ50-fold. Catalytic turnover of [Lys]Pg was reduced to a rate indistinguishable from that for [Glu]Pg activation when lysinebinding sites were blocked. On this basis, we conclude that the preference for activation of [Lys]Pg over [Glu]Pg by Pg*⅐SK is accounted for primarily by lysine-binding site interactions, and when these interactions are blocked, [Lys]Pg and [Glu]Pg are essentially equivalent in both conformational activation and in the initial proteolytic activation reaction.
The mechanism explains the inhibition of Pm formation by high SK concentration and the larger inhibition seen for [Lys]Pg compared with [Glu]Pg. The first proteolytic cycle is inhibited by high SK concentration because of the dual role of Pg as the catalytic component of Pg*⅐SK and as substrate for this complex. At high SK concentration, proteolytic Pg activation is inhibited because the free Pg concentration is depleted by formation of the Pg*⅐SK complex. As predicted by this model for the kinetics of [Lys]Pg activation, for which the affinity for The upper catalytic cycle is initiated by Pg-SK binding and conformational activation in formation of Pg*⅐SK, which binds Pg in the substrate mode, forming the ternary Michaelis complex, Pg*⅐SK⅐Pg, from which initial Pm is produced. Pm generated by Pg*⅐SK binds tightly to free SK in the catalytic mode, forming Pm⅐SK in competition with Pg*⅐SK. The tightly bound Pm⅐SK complex initiates the lower catalytic cycle, binding free Pg in the substrate mode to form Pm⅐SK⅐Pg, from which free Pm is generated. SK is relatively high, the maximum rate occurs at ϳ2 mol Pg/mol SK, reflecting the concentrations required for optimum formation of the productive Pg*⅐SK⅐Pg ternary Michaelis complex (Fig. 6). For [Glu]Pg, the reaction is also inhibited by high SK concentration but inhibition is weaker because of the lower affinity of SK for [Glu]Pg. The results show that the rate of Pm formation from [Lys]Pg approaches zero at high SK concentrations. A mechanism involving generation of Pm by intramolecular cleavage within the Pg*⅐SK complex proposed previously (20 -22) does not fit the data because the first-order generation of Pm would not be inhibited by SK. A mixed intramolecularintermolecular mechanism also predicts incomplete inhibition by SK. These results support the conclusion that Pm is formed initially by intermolecular cleavage of free Pg by Pg*⅐SK. The essentially complete inhibition of Pm formation observed for [Lys]Pg also indicates that proteolytic formation of Pm by Pg*⅐SK acting on other Pg*⅐SK complexes is at most a very slow process. The kinetics of Pm formation for both [Glu]Pg and [Lys]Pg as a function of SK and Pg concentration were well described by the intermolecular model, which predicts complete inhibition at saturating SK concentration.
An initial assumption was necessary for analysis of the rates of Pm generation measured by continuous monitoring of chromogenic substrate hydrolysis. The measured rates were inhibited hyperbolically by D-Val-Leu-Lys-pNA with an apparent inhibition constant of 330 Ϯ 80 M. The inhibition constant was essentially the same as KЉ m for the Pm⅐SK complex (Scheme 2) (15), suggesting that inhibition of the contribution of Pm⅐SKcatalyzed Pm formation to the measured rate of total Pm generation could be responsible. It is also possible, however, that negative linkage between chromogenic substrate binding to Pg*⅐SK and the affinity of SK binding and conformational activation (K A ) could account for the observed inhibition (16). Simulation of the reactions showed that the rate of Pm generation was very sensitive to small changes in K A . Because this issue could not be resolved unambiguously, experiments were performed at a constant substrate concentration, and the measured v 2 rates were corrected for 1.4-fold inhibition at this concentration. The appropriateness of this correction was supported by the normalization of the continuously and discontinuously measured rates with the above inhibition constant, where the latter were not subject to inhibition by substrate. Because this effect was small and the correction was constant for all of the rates, it does not change the interpretation of the large differences in the kinetic constants demonstrated or affect the main conclusions regarding the mechanism.
Pm generated and released in the first catalytic cycle (Fig. 6) binds free SK in the catalytic mode with extremely high affinity. Results of equilibrium binding studies for the native proteins yielded dissociation constants of 130 Ϯ 76 nM for [Glu]Pg, 10 Ϯ 3 nM for [Lys]Pg, and 0.012 Ϯ 0.004 nM for Pm (15,16), representing 800-and 11,000-fold higher affinity of SK for Pm compared with [Lys]Pg and [Glu]Pg, respectively. As expected from these large differences in affinity, and as shown in the companion paper (16), Pm readily displaces Pg from the Pg*⅐SK complex. This indicates that the initial Pm generated will sequester SK in the tightly bound Pm⅐SK complex.
The Pm⅐SK complex initiates the lower catalytic cycle in Fig.  6, which consists of exosite-mediated Pg binding as a substrate and its proteolytic activation to generate additional Pm (15 [Glu]Pg as a substrate, whether this reflects a difference in lysine-binding site interactions, which seems likely, has not yet been determined. Despite the sequential nature of the Pg*⅐SK-and Pm⅐SKcatalyzed reactions, the experimentally determined progress curves of total Pm formation under our conditions did not reveal the nonlinearity possible for these reactions. Moreover, the dependence of v 2 on Pg and SK concentration was reproduced by a model including only the Pg*⅐SK catalytic cycle and excluding the subsequent conversion of Pg to Pm by Pm⅐SK. To investigate the reasons for this, the full reaction scheme (Scheme 3) was simulated with the simplifying assumptions that both proteolytic reactions were bimolecular over the range of Pg concentration studied. Applying this assumption was justified for Pg*⅐SK-catalyzed Pg cleavage from the analysis of the rates of conformational activation (v 1 ) as a function of SK and Pg concentration (16). The Pm⅐SK-catalyzed reactions could also be treated as bimolecular, based on the Pg concentrations being much lower than the determined KЈ S . With these assumptions and the determined kinetic parameters, simulation of the time courses of the reaction species demonstrated the behavior expected for the sequential reactions. The bimolecular rate constants for [Lys]Pg activation by Pg*⅐SK and Pm⅐SK were similar, differing by only 1.5-fold. The combination of first-order conditions for the proteolytic reactions with similar rate constants and the common mechanism of inhibition of the rate by depletion of free Pg resulted in the near-linear progress curves and good fit by the mechanism including only Pg*⅐SK-catalyzed Pg activation. The situation for [Glu]Pg activation was less complicated, and the parameters determined from analysis of v 1 and v 2 by fitting of the model containing only the Pg*⅐SK reaction were in good agreement. This was due to the lower affinity of SK for binding and conformational activation of [Glu]Pg and the ϳ13-fold lower bimolecular rate constant for catalysis by Pm⅐SK compared with Pg*⅐SK, resulting in a lesser contribution of the Pm⅐SK-catalyzed cycle to total Pm formation.
Comparison Simulation of the time courses of the appearance and disappearance of the reaction species in the mechanism including both Pg*⅐SK and Pm⅐SK catalytic cycles (Scheme 3 and Fig. 6) revealed a fascinating and unique property of the overall mechanism. The time courses for [Glu]Pg activation with the determined kinetic parameters are shown in Fig. 7A. The simulation was performed at 5 nM SK and 1500 nM [Glu]Pg, representing ϳ10 ϫ K A for the SK-[Glu]Pg conformational activation equilibrium, such that SK was initially saturated with [Glu]Pg. Under these conditions, just over one turn of the Pg*⅐SK catalytic cycle results in a burst of Pm formation (Fig. 7A, curve c) representing slightly more than one SK equivalent, which is sufficient to bind all of the SK in the Pm⅐SK complex (curve b) and reduce the concentration of the Pg*⅐SK complex (curve a) to essentially zero. Simulation of [Lys]Pg activation by 5 nM SK and 100 nM [Lys]Pg (ϳ8 ϫ K A ) showed similar behavior, where the Pg*⅐SK catalytic cycle generated ϳ1.6 SK equivalents of Pm before all of the SK was sequestered in Pm⅐SK complexes due to the sum of Pm generated in both catalytic cycles (Fig.  7B). Compared with [Glu]Pg activation, generation of concentrations of Pm in larger excess of SK was required in [Lys]Pg activation to bind all of the SK. This was due to the 800-and 11,000-fold higher affinity of SK for Pm compared with [Lys]Pg and [Glu]Pg, respectively, and the ϳ380-fold slower rate of Pm⅐SK-catalyzed activation of [Glu]Pg compared with [Lys]Pg. These results support the conclusion that Pm formation initiated by conformationally activated Pg* in the Pg*⅐SK complex acts as a triggering process to produce enough Pm to bind all of the SK, which shuts off the initial catalytic cycle. The very tightly bound Pm⅐SK complex generated then propagates full conversion of the remaining Pg to Pm. Rapid and reversible binding and conformational activation of Pg by SK and an initial round of intermolecular Pg activation catalyzed by Pg*⅐SK is thus a self-limiting process that produces one SK equivalent of the irreversibly activated Pm⅐SK complex, which becomes the sole catalyst in the full conversion of Pg to Pm.