Resolution of conformational activation in the kinetic mechanism of plasminogen activation by streptokinase.

Streptokinase (SK) activates plasminogen (Pg) by specific binding and nonproteolytic expression of the Pg catalytic site, initiating Pg proteolysis to form the fibrinolytic proteinase, plasmin (Pm). The SK-induced conformational activation mechanism was investigated in quantitative kinetic and equilibrium binding studies. Progress curves of Pg activation by SK monitored by chromogenic substrate hydrolysis were parabolic, with initial rates (v(1)) that indicated no transient species and subsequent rate increases (v(2)). The v(1) dependence on SK concentration for [Glu]Pg and [Lys]Pg was hyperbolic with dissociation constants corresponding to those determined in fluorescence-based binding studies for the native Pg species, identifying v(1) as rapid SK binding and conformational activation. Comparison of [Glu]Pg and [Lys]Pg activation showed an approximately 12-fold higher affinity of SK for [Lys]Pg that was lysine-binding site dependent and no such dependence for [Glu]Pg. Stopped-flow kinetics of SK binding to fluorescently labeled Pg demonstrated at least two fast steps in the conformational activation pathway. Characterization of the specificity of the conformationally activated SK.[Lys]Pg* complex for tripeptide-p-nitroanilide substrates demonstrated 5-18- and 10-130-fold reduced specificity (k(cat)/K(m)) compared with SK.Pm and Pm, respectively, with differences in K(m) and k(cat) dependent on the P1 residue. The results support a kinetic mechanism in which SK binding and reversible conformational activation occur in a rapid equilibrium, multistep process.

Streptokinase (SK) 1 is used clinically as a thrombolytic drug to activate plasminogen (Pg) to plasmin (Pm), the serine proteinase responsible for dissolution of fibrin clots (1). Native [Glu]Pg is a multidomain zymogen consisting of a 77-residue N-terminal peptide, five kringle domains, and a serine proteinase catalytic domain that is activated by cleavage of Arg 561 -Val 562 (2,3). [Glu]Pg is in a compact conformation, maintained by intramolecular interactions between the N-terminal peptide and lysine-binding sites in kringles 4 and 5 (4 -6). Release of the N-terminal peptide by Pm cleavage generates [Lys]Pg, which adopts an extended conformation and is cleaved by plasminogen activators at a faster rate (5,(7)(8)(9)(10)(11)(12)(13). SK possesses no intrinsic catalytic activity but interacts with Pg and Pm, converting both the zymogen and active proteinase into specific proteolytic Pg activators (14 -19). SK binding to Pg results in conformational expression of an active catalytic site on the zymogen without the usual strict requirement for peptide bond cleavage (14,16,17). Pm is generated subsequently by proteolytic cleavage of Arg 561 -Val 562 , and the SK⅐Pm complex propagates Pg activation through expression of a substrate recognition exosite (20,21).
It is well established that both conformational and proteolytic activation contribute to SK-induced Pg activation, but there are a number of unresolved questions concerning the mechanism of conformational activation and its coupling to subsequent proteolytic Pm formation. Early studies (14,16,17) demonstrated that interaction of SK with Pg produced the activated Pg catalytic site in the SK⅐Pg* complex. Subsequent kinetic studies indicated that Pg activation involved an initially formed SK⅐Pg* activation complex and an isomerized form of the complex (22,23). The isomerization was time-, chloride ion-, and fibrinogen-dependent, and the active complexes interconverted slowly under certain experimental conditions. In other studies, a species of Pg isolated from a mixture of SK and Pg was reported to be an active "virgin enzyme" form of free Pg*, suggesting irreversible conformational activation (24). The relationship between SK-Pg binding and conformational activation and its reversibility have not been clearly established.
Previous studies of the mechanism of Pg activation by SK have not taken into account the binding affinities of SK for Pg and Pm species, primarily because consistent and reliable equilibrium binding constants have not been available. A number of studies have examined the binding interactions by various experimental approaches but have resulted in a very broad range of dissociation constants, varying from 28 pM to 220 nM for [Glu]Pg (25)(26)(27)(28)(29)(30)(31)(32), for example, where the higher affinities may result from Pm generation in SK/Pg mixtures. Our equilibrium binding studies have employed active site-labeled fluorescent Pg and Pm analogs to quantitate SK binding in the absence of proteolytic reactions (21,28,33,34). The results for the active site-labeled proteins support a relatively low affinity of SK for [Glu]Pg, a 13-16-fold enhanced affinity for [Lys]Pg due to the expression of lysine-binding site interactions and a 3100 -3500-fold further enhancement accompanying activation of the Pg catalytic domain. These changes in affinity account for a remarkable, overall ϳ50,000-fold higher affinity of SK for the reaction product, [Lys]Pm, compared with the initial substrate, [Glu]Pg (21,33).
Lysine-binding sites on the kringle domains of Pg and Pm participate in SK-induced Pg activation by mechanisms that are only partly understood. Conversion of Pg to Pm is inhibited by the lysine analog, 6-aminohexanoic acid (6-AHA) (35,36). In our studies, 6-AHA reduced binding affinity of SK for [Lys]Pg and [Lys]Pm only partially but did not affect the affinity of SK for [Glu]Pg (21,28,33), whereas others (36) have reported weakening of [Glu]Pg binding by 6-AHA. The role of lysinebinding site interactions in modulating conformational activation of [Glu]Pg and [Lys]Pg is not clear.
The goal of the present studies was to apply, for the first time, the combination of quantitative equilibrium binding and kinetic analysis required to define the mechanism of SK-induced Pg activation. Resolution of the conformational activation steps in the kinetic mechanism is characterized in the present paper, and coupling of the conformational and proteolytic activation pathway is reported in the companion paper (37). The studies presented here address the relationship between SK-Pg binding and conformational activation, the nature of Pg*, the unknown role of differences in affinities of SK for [Glu]Pg and [Lys]Pg, and the roles of lysine-binding sites. The results demonstrate that Pg activation can be kinetically resolved into consecutive conformational and proteolytic activation steps at pH 7.4, I 0.15 M, and 25°C. Conformational activation is induced by rapid equilibrium binding of SK to Pg and an accompanying reversible conformational change to form SK⅐Pg*. The mechanism of SK binding to fluorescein-labeled Pg involves at least two steps that can be resolved on the stopped-flow time scale. Kinetic separation of conformational and proteolytic activation allowed the substrate specificity of the transiently formed SK⅐Pg* complex to be characterized. The SK⅐Pg* complex exhibited significant decreases in substrate specificity for tripeptide p-nitroanilide substrates compared with that demonstrated previously for SK⅐Pm or Pm (21), indicating a unique active site specificity for Pg* in the SK⅐Pg* complex. The combined equilibrium binding and kinetic data under our experimental conditions did not reveal slowly formed, functionally distinguishable activator complexes and did not support formation of a virgin enzyme. These studies provide the first quantitative equilibrium binding and kinetic analysis of conformational Pg activation by SK, and demonstrate the critical role of differences in SK affinity for [Glu]Pg and [Lys]Pg and the differential roles of lysine-binding site interactions in regulating the activation mechanism.

Protein Purification and Characterization-[Glu]
Pg was purified from human plasma and separated into carbohydrate variants Pg 1 and Pg 2 by minor modifications of published procedures (38,39). Pm was prepared by activation of [Glu]Pg as described previously (28,40). D-Phe-Phe-Arg-CH 2 Cl-blocked Pm (FFR-Pm) was prepared by incubation of 16 M Pm with 80 M D-Phe-Phe-Arg-CH 2 Cl (FFR-CH 2 Cl) for 1 h (Ͼ99.9% inhibition) in 5 mM Hepes, 0.3 M NaCl, 10 mM 6-AHA, 1 mg/ml PEG, pH 7.0 at 25°C, followed by dialysis against Ͼ2000 volumes (with three changes) of 50 mM Hepes, 0.125 M NaCl, pH 7.4. When added to a 1 nM Pm assay containing 200 M D-Val-Leu-Lys-pNA (VLK-pNA), 1.0 M FFR-Pm was found to affect the rate of hydrolysis Ͻ3%, indicating no residual FFR-CH 2 Cl present in the preparation.
[Lys]Pg was prepared by using a modified procedure from that described previously (33), in which [Lys]Pg was chromatographed on soybean trypsin inhibitor-Sepharose in 10 mM Hepes, 0.15 M NaCl, 20 mM 6-AHA, 1 mg/ml PEG, pH 7.4, at 4°C to remove Pm and further purified on aminohexylagarose in 0.1 M KH 2 PO 4 , 0.2 M NaCl, pH 6.0, at 4°C eluted with buffer containing 10 mM 6-AHA. SK was obtained from Diapharma as outdated therapeutic material and purified by chromatography on Pm-Sulfolink (28). All proteins were quick-frozen in dry ice/isopropyl alcohol and stored at Ϫ70°C. Protein concentrations were determined by absorbance at 280 nm using the following absorption coefficients ((mg/ ml) Ϫ1 cm Ϫ1 ) and molecular weights: 0.95 and 47,000 for SK (41,42) and 1.70 and 84,000 for [Lys]Pg 2 and Pm 2 (13,39,43). Pm concentrations were determined by active site titration with fluorescein mono-pguanidinobenzoate (44).
Fluorescence Binding Studies-Fluorescence equilibrium binding experiments were performed with an SLM 8100 spectrofluorometer in the ratio mode using acrylic cuvettes (Sarstedt) coated with polyethylene glycol 20,000 to minimize protein adsorption (47). Excitation and emission wavelengths were 500 and 516 nm, respectively (8 nm bandpasses). The fractional change of the initial fluorescence intensity was expressed as (F obs Ϫ F o )/F o ϭ ⌬F/F o from data collected from several cuvettes encompassing overlapping segments of the titration curves. Pg in the presence of the respective native Pg species resulted in a rapid fluorescence change that was stable for Ͼ30 s, which was taken as the rapidly established competitive equilibrium value, followed by slower changes over several minutes due to proteolysis. To obtain the most reliable measurements, only one addition of SK per cuvette was used. Competitive titration data collected as a function of total native Pg concentration were fit simultaneously with the competitive cubic binding equation (48,49) using SCIENTIST (MicroMath) to obtain the maximum fluorescence change (⌬F max /F o ), dissociation constant for SK binding to labeled Pg, and competitive dissociation constant for binding of SK to native Pg. Parameters are reported as Ϯ 2 S.D.
Chromogenic Substrate Kinetics-Initial rates of substrate hydrolysis (Ͻ5% substrate depletion) were measured at 405 nm and 25°C in 50 mM Hepes, 0.125 M NaCl, 1 mM EDTA, 1 mg/ml PEG, pH 7.4, in polyethylene glycol 20,000-coated cuvettes. All substrates were from commercial sources except Pro-Gly-Arg-pNA, which was synthesized by SynPep (Dublin, CA) and was 98% pure. Concentrations of peptide-pNA substrates were determined from the absorbance at 342 nm by using an absorption coefficient of 8,266 M Ϫ1 cm Ϫ1 , and product concentrations were calculated by using an absorption coefficient of 9,933 M Ϫ1 cm Ϫ1 (50).
Plasminogen Activation Kinetics-Activation of Pg by SK was measured by continuous monitoring of the increase in absorbance of VLK-pNA at 405 nm (⌬A 405 nm ). The parabolic progress curves of pNA formation were fit by Equation 1 (15,51).
The analysis gave 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 ). Linear transformation of the continuous assay time courses was accomplished by dividing ⌬A 405 nm by t (52). Kinetic Model of Plasminogen Activation by Streptokinase-The observed dependence of v 1 on SK concentration can be described by the mechanism shown in Scheme 1.

SCHEME 1
In this model, SK binds to Pg with the dissociation constant K A to form a reversible, conformationally activated SK⅐Pg* complex, which can bind chromogenic substrate (S) with Michaelis constant K m and generate product with catalytic rate constant k c . As will be shown, the assumption in this model that SK⅐Pg* formation is reversible is valid. In addition, SK⅐Pg* can bind Pg as a substrate with a dissociation constant K S to form the ternary Michaelis complex (SK⅐Pg*⅐Pg) and generate Pm with the catalytic rate constant k 0 . This study characterizes Pg binding to SK, resolves the processes involved in conformational activation of Pg resulting in formation of the activated SK⅐Pg* complex, and quantifies the kinetic parameter K S for Pg binding. The companion paper (37) describes the coupled proteolytic mechanism of Pm generation by the activated SK⅐Pg* complex and formation of the SK⅐Pm complex that catalyzes Pg activation to Pm.
The observed progress of activity formation is given by Equation 1, where v 1 and v 2 represent the rate of chromogenic substrate turnover by SK⅐Pg* and the rate of generation of Pm, respectively. From the Michaelis-Menten equation and the equilibrium binding expression for SK binding to Pg, v 1 for the model in Scheme 1 is given by and

substitution of Equation 3 into Equation 2 gave Equation 4
.
Solving the mass conservation equation for [Pg] free gave the cubic Equation 5-9 as an exact solution in terms of the total concentrations, dissociation, and kinetic constants.
Under conditions where [Pg] free Ͻ Ͻ K S which were found to apply, Equation 4 can be reduced to Equation 10. Equation 10 is reduced to Equation 11.
In this situation, [Pg] free is defined by the quadratic Equation 12, Dissociation of the SK⅐Pg* Complex with Active Site-blocked Pm-The chromogenic substrate activity of the SK⅐Pg* complex, formed by incubation of 10 nM [Lys]Pg 1 with 50 nM SK for 2 min in pH 7.4 buffer containing 50 mM 6-AHA at 25°C, was measured by monitoring the increase in absorbance at 405 nm after addition of 200 M VLK-pNA. Dissociation of the complex by competitive binding of FFR-Pm was achieved by addition of varying concentrations of FFR-Pm after the formation of SK⅐Pg*. Following 2 min of incubation, substrate was added to measure the remaining active SK⅐Pg*. Alternatively, reactions were initiated by addition of SK to mixtures of Pg, FFR-Pm, and substrate that had been pre-equilibrated for 5 min. All progress curves in buffer containing 50 mM 6-AHA were nearly linear and gave v 1 , the rate of chromogenic substrate turnover by SK⅐Pg*, as a function of FFR-Pm concentration. The results were fit by the cubic binding equation for competitive binding of Pg and FFR-Pm to SK (48,49), with the dissociation constant for FFR-Pm and SK fixed at the independently determined value of 0.22 nM (33).

Dependence of the Rates of Pg Activation on SK Concentration-
The dependence of Pg activation on SK concentration was investigated in continuous activity assays in which Pg was activated with SK in the presence of 200 M VLK-pNA, and the reaction was monitored by the increase in absorbance at 405 nm with time. Progress curves at low SK concentrations were parabolic, becoming increasingly linear at higher SK concentrations (Fig. 1A). No lags were observed in the absorbance change over time, linear transformations of the progress curves, indicating rapid formation of active species. Analysis of the progress curves in a model-independent manner by fitting a second order polynomial (Equation 1) resulted in the rates v 1 and v 2 , which will be shown below and in the companion paper (37) to represent the rate of hydrolysis of chromogenic substrate by the rapidly formed SK⅐Pg* complex and rate of proteolytic Pm generation, respectively. The fitted initial ⌬A 405 nm at zero time was essentially zero (Ϯ0.002). The v 1 dependence of activation of 10 nM [Lys]Pg 1 on SK concentration is shown in Fig. 1B. The dependence of v 1 on SK was hyperbolic and well described by the quadratic binding equation with an apparent dissociation constant of 10 Ϯ 3 nM and a maximum rate of 0.020 Ϯ 0.001 M s Ϫ1 . By contrast, the dependence of v 2 on SK was very unusual, increasing sharply to a maximum of 2 ϫ 10 Ϫ4 M s Ϫ2 at an SK concentration approximately equal to one-half that of Pg and decreasing to essentially zero at high SK concentrations.
The dependence of v 1 for [Lys]Pg 1 and [Glu]Pg 1 activation at 2, 5, 10, 15, and 20 nM, and 5, 10, 15, and 20 nM, respectively, on increasing SK concentration in the presence of 200 M VLK-pNA is shown in Fig. 2, A and B. The data were fit both by Equations 4 -9 for the full mechanism in Scheme 1 (see "Experimental Procedures") without assumptions and by Equations 11 and 12 with the assumptions, [Pg] free Ͻ Ͻ K S and [S] 0 Ͻ Ͻ K m . The K m value for hydrolysis of VLK-pNA by the activated SK⅐[Lys]Pg* complex was 3000 M, determined separately as described below. On this basis, the maximum rates at 200 M substrate were taken to represent the bimolecular rate constants (k c /K m ). Whereas the fitted values for k c /K m and K A obtained from both the explicit and simplified equations were indistinguishable, the values for K S from fitting by the explicit equations were undefined, indicating that indeed [Pg] free Ͻ Ͻ K S and that the proteolytic reaction could be treated as bimolecular. The fitted v 1 dependence yielded k c /K m of 12 Ϯ 1 nM Ϫ1 s Ϫ1 and K A 12 Ϯ 3 nM for [Lys]Pg 1 activation by SK, k c /K m of 11 Ϯ 1 nM Ϫ1 s Ϫ1 , and K A 143 Ϯ 12 nM for [Glu]Pg 1 activation (Table I).  (Fig. 4). Analysis of titrations of v 1 with SK using Equations 11 and 12 gave an apparent dissociation constant of 50 Ϯ 5 nM for [Lys]Pg 1 , which was a 4-fold decrease in affinity from that determined in the absence of 6-AHA (Table I) Table I).

Binding of SK to Native and Fluorescently Labeled [Lys]Pg and [Glu]Pg-To
Competitive fluorescence binding studies for native Pg in the presence of 10 mM 6-AHA showed that SK bound native [Lys]Pg 1 12-fold more weakly, with a dissociation constant of 115 Ϯ 32 nM (Fig. 3A)  Pg was comparable with that reported previously for the fluorescent labeled Pg species (28,33). These results demonstrated quantitative correspondence between the kinetically determined dissociation constants for conformational activation and the dissociation constants determined in the competitive equilibrium binding studies in the absence of lysine-binding site interactions.
The effect of the active site label on the affinity of SK for [Lys]Pg and [Glu]Pg was indistinguishable for [Lys]Pg 1 and [Glu]Pg 1 , which exhibited 4.2-and 5.3-fold weaker binding to SK compared with the native proteins, respectively, in the absence of 6-AHA, and 4.6-and 6.5-fold weaker binding, respectively, in 10 mM 6-AHA (Table I). This was consistent with the 6.6-fold decrease reported previously (28) for [Glu]Pg 1 labeled with 2-((4Ј-iodoacetamido)anilino)naphthalene-6-sulfonic acid, thus indicating a consistent effect of occupation of the active site by the probe and linking peptide on SK affinity and conformational activation. The results of the native protein competitive titrations established that the labeled proteins bound SK with ϳ5-fold lower affinity, but importantly, these effects were similar for [Lys]Pg and [Glu]Pg. The effect of 6-AHA on the interaction of SK for the native and labeled proteins was not altered by this ϳ5-fold difference in affinity.
The role of the compact to extended conformational change in [Glu]Pg activation was investigated by comparison of activation of [Glu]Pg with SK in buffer containing 125 mM Cl Ϫ , which stabilizes the compact conformation of [Glu]Pg with buffer in which Cl Ϫ had been replaced with acetate, which allows [Glu]Pg to adopt the extended conformation (4,11,12,54). Dependence of v 1 on SK at 1 and 10 nM [Glu]Pg 1 indicated a K A ϳ2 nM (results not shown), consistent with the previously determined K D of 11 nM for binding of fluorescein-labeled [Glu]Pg (33) and an ϳ5.5-fold lower affinity for the labeled zymogen. As observed by equilibrium binding and by kinetics, SK preferentially bound and conformationally activated the extended conformation of [Glu]Pg.
Presteady-state Kinetics of Conformational Activation-Rapid-reaction kinetic studies were performed to investigate SK-Pg interactions in the fast conformational activation process forming SK⅐Pg*. As shown in Fig. 5 (Fig. 5). The TABLE I Equilibrium binding and kinetic constants for SK-plasminogen interactions Dissociation constants (K D ) are listed for SK binding to the indicated labeled and native Pg species obtained from direct and competitive fluorescence titrations in the absence and presence of 10 mM 6-AHA. Also listed are kinetically determined dissociation constants (K A (v 1 )) for native SK-Pg interactions and the specificity constants (k c /K m ) for chromogenic substrate hydrolysis by the SK-saturated complexes in the absence and presence of 10 mM 6-AHA. Kinetic and binding studies were performed, and the data were analyzed as described under "Experimental Procedures."

Pg species
Ϫ6-AHA ϩ6-AHA  Dissociation of the SK⅐Pg* Complex with Active Site-blocked Pm-To determine whether conformational activation of Pg by SK was due to a reversible conformational change or to irreversible generation of conformationally activated Pg* virgin enzyme via the plasminogen activator complex, SK⅐Pg* was formed rapidly and then incubated with Pm blocked with D-Phe-Phe-Arg-CH 2 Cl (FFR-Pm), which binds very tightly to SK (21,33). Incubation of 10 nM [Lys]Pg 1 with 50 nM SK in buffer containing 50 mM 6-AHA for 2 min resulted in formation of the activated SK⅐Pg* complex, which could be quantitated from the initial rates of chromogenic substrate hydrolysis. Subsequent addition of FFR-Pm resulted in decreased activity of the complex, consistent with rapid competitive binding of SK to [Lys]Pg and FFR-Pm and dissociation of SK⅐Pg* to free SK and inactive zymogen (Fig. 6). The reversibility of complex formation was further demonstrated in assays in which SK was added to a pre-quenched mixture of FFR-Pm, 10 nM [Lys]Pg, and substrate, with indistinguishable results (Fig. 6). Analysis of the data with the cubic binding equation for competitive binding of [Lys]Pg and FFR-Pm to SK gave a dissociation constant of 23 Ϯ 21 nM for SK binding to native [Lys]Pg and a maximum activity of 0.013 Ϯ 0.005 M s Ϫ1 for the SK⅐Pg* complex. These parameters were in agreement with previously determined values of 50 Ϯ 5 nM and 0.017 Ϯ 0.001 M s Ϫ1 (Fig.  4 and Table I). These results demonstrated rapid reversibility of SK⅐Pg* complex formation and provided no evidence for formation of an irreversibly activated virgin enzyme species.
Kinetic Parameters of Tripeptide-p-Nitroanilide Substrate Hydrolysis by the SK⅐Pg* Complex-Resolution of conformational activation from proteolytic reactions in Pg activation allowed for the first time a detailed investigation of the kinetic properties of the transiently formed SK⅐Pg* complex. Kinetic parameters were determined from v 1 rates of [Lys]Pg 2 activation in the presence of saturating SK for hydrolysis of an array of 1 dipeptide and 11 tripeptide substrates, including a peptide with the sequence of the plasminogen activation cleavage site (PGR-pNA) ( Table II). The previously reported parameters for hydrolysis of the same substrates by the SK⅐Pm complex are shown for comparison (21). The results demonstrated reduced k c /K m values representing losses of 5-18-and 10 -130-fold in substrate specificity of SK⅐Pg* when compared with SK⅐Pm (Table II) or free Pm, respectively (21). Differences in kinetic parameters for hydrolysis by SK⅐Pm and SK⅐Pg* were dependent on the identity of the substrate P1 residue. The driving force behind the decrease in specificity of SK⅐Pg* compared with SK⅐Pm for substrates with Lys in the P1 position was a 650 -1500% increase in K m accompanied by modest effects of Ϫ12 to 60% on k c (Fig. 7). In contrast, substrates with Arg at P1 exhibited 43-97% decreases in k c and small effects on K m ranging from a 55% decrease to a 400% increase, with the exception of pyro-EPR-pNA, which exhibited an 1180% increase (Table II and Fig. 7). These results demonstrated significant loss of substrate specificity between SK⅐Pm and SK⅐Pg*, and suggested that these changes were dependent on differences in the interactions of the S1 specificity subsite of the enzymes with the P1 residue of the substrates. The decreased specificity of SK⅐Pg* and SK⅐Pm for PGR-pNA relative to Pm (21) demonstrated that changes in the S1-S3 specificity subsites play little or no role in the specificity for conversion of Pm or Pg into proteolytic Pg activators. DISCUSSION Resolution of conformational activation from subsequent proteolytic reactions in the kinetic mechanism of SK-induced Pg activation, in conjunction with analysis of SK binding to native [Glu]Pg and [Lys]Pg, allowed the quantitative relationship between SK-Pg binding and conformational activation to be established for the first time. Conformational activation of Pg was readily reversible, and the results provided no evidence for irreversible formation of the free, activated virgin form of Pg* postulated previously (24). SK binding to Pg and conformational expression of the Pg catalytic site in the SK⅐Pg* complex occurred rapidly and without rate-controlling intermediates on the seconds time scale. The slow (minutes) isomerization of the SK⅐Pg* complex reported in previous studies (22,23) with [Glu]Pg at low temperature and/or low chloride concentrations was not observed in our studies. This does not mean that this conformational change does not occur but likely represents differences in the experimental conditions, where this event apparently occurs rapidly under our conditions.
Although several previous studies have sought to quantitate binding of SK to Pg/Pm species, the results have not yielded consistent affinities, and in no studies has equilibrium binding of SK been related directly to nonproteolytic Pg activation. The approach based on active site-labeled fluorescent Pg analogs enabled quantitation of the dissociation constants for native [Glu]Pg and [Lys]Pg by resolving the rapid competitive binding equilibria measured by fluorescence from slower proteolytic reactions. The results demonstrate the quantitative corre-   (21)) by competitive binding indicate ϳ800and ϳ11,000-fold higher affinity for Pm compared with [Lys]Pg and [Glu]Pg, respectively. The fluorescein-labeled Pg analogs demonstrated a consistent, ϳ5-fold lower affinity for SK. These results were in very good agreement with the dissociation constant determined previously for [Glu]Pg by competitive binding with 2-anilinonaphthalene-6-sulfonic acid-labeled [Glu]Pg and the 6.6-fold lower affinity of the labeled analog (28). The magnitude of the ratio of affinities is ϳ5-fold lower than determined for the fluorescein-labeled Pg/Pm analogs (ϳ3300-and ϳ50,000-fold (33)) because of the absence of a detectable effect of labeling on Pm binding (21). This may reflect the experimental difficulty of resolving the relatively small effect of labeling on the background of the extremely high affinity of SK for Pm. It is clear from these results that the presence of the fluorescence probe and linking D-Phe-Phe-Arg-chloromethyl ketone in the catalytic site reduce the affinity of SK similarly for [Glu]Pg and [Lys]Pg. Although one might expect the presence of a transition state analog inhibitor in the catalytic site to stabilize the active conformation of Pg and thereby increase affinity for SK, the opposite effect was observed. This may be explained by the reduced tripeptide substrate binding affinity, approximated by K m , associated with SK binding to Pm (21) and the similarly high Michaelis constants seen here for SK⅐Pg*. These observations indicate that SK affinity is negatively linked to substrate binding to the catalytic site. The lower affinity for Pg containing the fluorescence probe-tripeptide label covalently attached to the catalytic histidine residue may be due to such linkage. An additional possibility is that active site labeling affects the equilibrium constant for a conformational change on the activation pathway that normally contributes to overall affinity of SK for the native zymogen but is less favorable for the labeled zymogen. Such quantitative differences are to be expected, given the conformational formation of the substrate binding subsites and oxyanion hole induced by SK binding. Regardless of the source of the ϳ5-fold lower affinity of the fluorescent Pg analogs, as shown here, they reproduce the properties of the native proteins with indistinguishable differences in affinity between [Glu]Pg and [Lys]Pg and in the magnitude of the effects of 6-AHA on these interactions.
Examination of the effects of near-saturating concentrations of 6-AHA on SK binding and conformational activation demonstrated major differences between the role of lysine-binding sites in conformational [Glu]Pg and [Lys]Pg activation. As observed in previous equilibrium binding studies with the fluorescent Pg analogs (33), the higher affinity of [Lys]Pg for SK compared with [Glu]Pg was accounted for by the contribution of lysine-binding site interactions to [Lys]Pg binding but not to [Glu]Pg binding. The affinities of SK binding and conformational activation of native [Lys]Pg were both ϳ12-fold higher than [Glu]Pg, and this difference was lost when lysine-binding sites were blocked with 6-AHA. The kinetically measured SK affinities for [Glu]Pg and [Lys]Pg activation in the presence of 6-AHA were equivalent and in good agreement with the affinities determined by competitive binding. The source of the differential contribution of lysine-binding sites to conformational activation of [Glu]Pg and [Lys]Pg is thought to be the different conformations of these Pg forms and the associated differences in accessibility of lysine-binding sites. At physiological chloride ion concentration, [Glu]Pg is in a compact conformation, maintained by interactions between the N-terminal 77-residue peptide and lysine-binding sites in kringles 4 and 5 (4 -6, 54). Cleavage of the N-terminal peptide by Pm to form [Lys]Pg is accompanied by a shift to the extended conformation and exposure of unoccupied lysine-binding sites (5,7,13). In the interaction of SK with [Glu]Pg, 6-AHA has no significant effect because lysine-binding sites are inaccessible or blocked in both the compact conformation and the extended conformation produced by 6-AHA binding, respectively. The increased affinity of SK for [Lys]Pg and the accompanying increased affinity of conformational activation was reduced to that of [Glu]Pg in the presence of 6-AHA. Thus, the compact-extended conformational change in [Glu]Pg by itself has apparently little effect on SK binding or conformational activation but indirectly enhances affinity and activation through increased lysine-binding site interactions. This was substantiated by the observation here that SK bound and conformationally activated [Glu]Pg in the extended conformation formed at low chloride concentrations with high affinity, which is due to lysine-binding site interactions with the extended form (33). On the basis of the large differences in affinity between the compact and extended forms of [Glu]Pg, previous studies (33) concluded that SK binding is accompanied by conversion of [Glu]Pg to the extended form in the SK⅐[Glu]Pg* complex. The partial reduction in the affinity of SK for activating [Lys]Pg by 6AHA, the lack of an effect for [Glu]Pg, and the similar chromogenic substrate activities of the SK⅐Pg* complexes in the absence and presence of 6-AHA demonstrate that lysine-binding site interactions linked to the compact-extended equilibrium enhance but are not required for Pg binding and conformational activation.
The specificity of conformationally activated Pg in the transiently formed SK⅐Pg* complex for tripeptide-pNA substrates was distinctly different from the specificity of free Pm and SK⅐Pm characterized previously for the same panel of substrates (21). In terms of specificity constants, Pm was the most specific; the specificity of SK⅐Pm was unaffected or reduced 2.6 -10-fold for the majority of substrates (21), and SK⅐Pg* was the least specific, with reductions of 18 -130-fold compared with Pm. As shown previously for SK⅐Pm, increases in K m contribute substantially to the loss of specificity with SK binding. This may be the result of reduced access of substrates to the catalytic site, which is in a deep cavity surrounded by SK in the SK⅐micro-Pm structure (20,55). However, binding of SK to Pm also results in a uniquely different specificity for substrate residues in the P2 position, whereas the marginally higher specificity of Pm for substrates with Lys compared with Arg at P1 is unaffected (21). By contrast, comparison of SK⅐Pg* and SK⅐Pm demonstrated differences in specificity dependent on P1. Substrates with Lys at P1 were primarily different in SK⅐Pg* because of increases in K m , with modest changes in k cat . K m for substrates with Arg at P1, with one exception, was not greatly affected, whereas k cat was consistently decreased. These results indicate that the S1 specificity subsite in SK⅐Pg* and SK⅐Pm is distinctly different, characterized by a loss of apparent substrate affinity for P1 Lys substrates and decreased rates of catalytic turnover of P1 Arg substrates. The uniquely different specificities of SK⅐Pm and the conformationally activated catalytic site in SK⅐Pg* were not correlated with a preference of either complex for cleavage of the Pro-Gly-Arg activation sequence in Pg, consistent with exosite-mediated Pg substrate binding dominating Pg substrate recognition by both complexes (21). The difference in substrate specificity of SK⅐Pg* and SK⅐Pm suggests the possibility that the catalytic site is incompletely formed in the conformationally activated zymogen. However, it is also possible that this difference is due to differences between the enzyme and zymogen in their interactions with SK domains that affect the catalytic site.
Three components have been postulated to contribute to the mechanism of conformational Pg activation. Studies demonstrating a critical role of the N terminus of SK in conformational activation support the "molecular sexuality" mechanism (34,56,57). In this mechanism, insertion of the SK N terminus into the N-terminal binding cleft in the Pg catalytic domain and salt bridge formation with Asp 194 (chymotrypsin numbering) triggers the activating conformational change. Other studies support a role for interactions of the SK ␥-domain with Pg resulting in reorientation of Lys 156 to form the critical salt bridge (20,55,58). The contribution of these two mechanisms to conformational activation has not been resolved. In either case, the ϳ800-fold higher affinity of SK for the activated catalytic domain of Pm compared with [Lys]Pg is also thought to contribute to nonproteolytic activation by stabilization of the activated conformation (21,34). The present studies demonstrate that the molecular events in conformational activation are initiated by rapid and reversible SK binding. Rapid-reaction kinetics of SK binding to fluorescein-labeled [Glu]Pg and [Lys]Pg demonstrated biphasic reactions, indicating that at least two binding or conformational change reactions occur and can be resolved on the conformational activation pathway. The effect of 6-AHA on the kinetics of SK binding to [Glu]Pg and [Lys]Pg showed that the kinetics for [Glu]Pg were unaffected, whereas the rate of the fast phase was decreased for [Lys]Pg to that observed for [Glu]Pg. These results parallel the effects of 6-AHA on SK binding and conformational activation and indicate that the fast phase of SK binding to [Lys]Pg is linked to interactions with lysine-binding sites on Pg. One of these fast reactions may correspond to the isomerization of the SK⅐Pg* complex at low chloride concentrations described previously (22). The results indicate that further rapid-reaction kinetic studies should enable resolution of the individual molecular events on the pathway of conformational Pg activation by SK. The mechanism of coupling of the conformational activation process characterized here to proteolytic Pg activation is addressed in the companion paper (37).