Streptokinase triggers conformational activation of plasminogen through specific interactions of the amino-terminal sequence and stabilizes the active zymogen conformation.

Cleavage of Arg(561)-Val(562) in plasminogen (Pg) generates plasmin (Pm) through a classical activation mechanism triggered by an insertion of the new amino terminus into a binding pocket in the Pg catalytic domain. Streptokinase (SK) circumvents this process and activates Pg through a unique nonproteolytic mechanism postulated to be initiated by the intrusion of Ile(1) of SK in place of Val(562). This hypothesis was evaluated in equilibrium binding and kinetic studies of Pg activation with an SK mutant lacking Ile(1) (SK(2--414)). SK(2--414) retained the affinity of native SK for fluorescein-labeled [Lys]Pg and [Lys]Pm but induced no detectable conformational activation of Pg. The activity of SK(2--414) was partially restored by the peptides SK(1--2), SK(1--5), SK(1--10), and SK(1--15), whereas Pg(562--569) peptides were much less effective. Active site-specific fluorescence labeling demonstrated directly that the active catalytic site was formed on the Pg zymogen by the combination of SK(1--10) and SK(2--414), whereas sequence-scrambled SK(1-10) was inactive. The characterization of SK(1--10) containing single Ala substitutions demonstrated the sequence specificity of the interaction. SK(1--10) did not restore activity to the further truncated mutant SK(55-414), which was correlated with the loss of binding affinity of SK(55--414) for labeled [Lys]Pm but not for [Lys]Pg. The studies support a mechanism for conformational activation in which the insertion of Ile(1) of SK into the Pg amino-terminal binding cleft occurs through sequence-specific interactions of the first 10 SK residues. This event and the preferentially higher affinity of SK(2--414) for the activated proteinase domain of Pm are thought to function cooperatively to trigger the conformational change and stabilize the active zymogen conformation.

Pg. The studies support a mechanism for conformational activation in which the insertion of Ile 1 of SK into the Pg amino-terminal binding cleft occurs through sequence-specific interactions of the first 10 SK residues. This event and the preferentially higher affinity of SK  for the activated proteinase domain of Pm are thought to function cooperatively to trigger the conformational change and stabilize the active zymogen conformation.
The fibrinolytic proteinase plasmin (Pm) 1 is formed by pro-teolytic cleavage of plasminogen (Pg) at Arg 561 -Val 562 through a mechanism that is similar for nearly all serine proteinase zymogens (1,2). The classical mechanism based on the studies of trypsinogen (3)(4)(5) and chymotrypsinogen (6,7) involves an insertion of the newly formed Val 562 amino terminus into a specific binding cleft on the proteinase. This interaction triggers the conformational change that completes the folding of the proteinase and activates the enzyme by the formation of the substrate binding site and the oxyanion hole (3)(4)(5)(6)(7). In the case of Pg activation, additional atypical changes occur in the positions of the catalytic residues and Trp 761 , which are also required for activation (8,9).
The activation of Pg by the thrombolytic drug streptokinase (SK) represents a rare and poorly understood exception to the standard mechanism, because the binding of SK to the catalytic domain of Pg results in the expression of the active catalytic site without proteolysis (10 -12). The transition of trypsinogen toward a trypsin-like conformation can be achieved without proteolysis through strongly linked binding of a ligand to the catalytic site and binding of the amino-terminal dipeptide Ile-Val to the binding cleft (4,5). Thus, the inactive conformation of trypsinogen and homologous zymogens is thought to be in highly unfavorable but reversible equilibrium with an active conformation (4). On this basis, it was hypothesized that a "molecular sexual" encounter of the amino terminus of SK with the binding cleft of Pg could induce the activation of the zymogen (5). Experimental evidence consistent with this molecular sexuality hypothesis was provided in the recent findings that the deletion (13) or substitution (14) of Ile 1 of SK was correlated with the loss of activity in conformational activation. This mechanism, however, is in apparent disagreement with the reported high activity of recombinant SK species in which the first 16 residues were deleted or the amino terminus was blocked by fusion to a maltose-binding protein (15,16). Moreover, it has been reported subsequently that this SK fusion protein is inactive in conformational activation (14). Although recent evidence demonstrates that Ile 1 is important, the possible role of the amino-terminal 2-15 residue sequence of SK is unknown. An alternative mechanism for conformational activation has also been proposed in which a conformational change accompanying SK binding reorients Lys 698 of Pg to occupy the amino-terminal binding cleft in place of Val 562 (17). Recent studies, however, implicate Lys 698 of micro-Pg in SK binding affinity (14) while also finding that this residue is not essential for conformational activation. Finally, a comparison of the affinities of SK for Pg and Pm demonstrated a 3100 -3500-fold higher affinity of SK for the activated proteinase domain of Pm, which has been proposed to participate in the conformational activation of Pg by stabilizing the active conformation of the zymogen (18,19).
The unresolved roles of SK binding to Pg and Pm and interactions of the SK amino-terminal sequence with the Pg aminoterminal binding site in the conformational activation mechanism were evaluated in the present fluorescence and kinetic studies of Pg interactions with an SK mutant lacking Ile 1 (SK 2-414 ). SK  bound to fluorescein-labeled [Lys]Pg and [Lys]Pm with affinities equivalent to native SK, but this tight binding alone was insufficient to induce conformational activation of Pg. Compelling evidence for an additional critical requirement for interactions of the SK amino terminus was obtained in the novel demonstration that the loss of activity of SK 2-414 could be restored partially by the specific binding of peptides based on the amino-terminal 10-residue sequence of SK. The results support a mechanism of conformational activation in which the amino-terminal sequence of SK beginning with Ile 1 interacts in a sequence-specific manner with the amino-terminal binding cleft of Pg and through extended binding sites on the SK 2-414 ⅐Pg complex to induce a transition toward the active conformation. Evidence for the additional role of preferential binding affinity of SK for Pm compared with Pg was obtained in the finding that conformational activity of the further SK truncation mutants SK 55-414 and SK 78 -414 could not be restored by SK [1][2][3][4][5][6][7][8][9][10] . The loss of activity was associated with a selective loss of affinity of SK 55-414 for labeled [Lys]Pm, whereas the affinity for [Lys]Pg was not significantly affected. On this basis, conformational activation is concluded to occur through the amino-terminal interaction in cooperation with the stabilization of the active conformation by the enhanced binding affinity of the remainder of SK for the active form of the zymogen.

Protein Purification and Characterization-[Glu]Pg and [Lys]
Pm carbohydrate forms 1 and 2 were purified and characterized by the procedures published previously (18 -20). Pm preparations were 71-84% active as determined by active site titration (18 (18,19,22,23), which yielded probe incorporation stoichiometries of 0.9 -1.2 mol of fluorescein/mol of Pg or Pm. Native SK was obtained from Diapharma and purified by affinity chromatography (18,19,23). Wild-type SK, SK 55-414 , and SK 78 -414 were described by Fay and Bokka (24) and purified essentially as described previously (18,24). The polymerase chain reaction was used to generate SK lacking the codon for Ile 1 (SK 2-414 ). The 5Ј and 3Ј primers were TCATCTATCATAT-GGCTGGACCTGAGTGGCTG and AAGGTTACCGAACCATC, respectively. Template DNA was pET-3a/SKC (24). The resulting product was digested with NdeI and AflII and then ligated into similarly restricted pET-3a/SKC. An analysis of a single clone (pET-3a/SK 2-414) confirmed the absence of unintended mutations in the SK sequence bounded by the NdeI and AflII sites. SK 2-414 was expressed in BL21(DE3)pLysS and purified as described previously (24). Protein sequence analysis confirmed that the amino terminus of the recombinant protein was alanine.
Ile-Val, Phe-Val, and SK 1-2 were purchased in Ͼ99% purity from Bachem, Schweizerhall, and Sigma, respectively. Peptides were dissolved in buffer or water, and the concentrations were determined by weight. The solubility of SK 1-10 and SK 1-15 limited the concentrations of these peptides in the reactions to 3.5 and 1 mM, respectively.
Fluorescence Studies-Fluorescence titrations were done with an SLM 8100 spectrofluorometer following methods described previously (18,19) with 500 nm excitation and 516 nm emission (8  Pg was preincubated for 5 min before initiation with the SK mutant. The slow background rate obtained with Pg alone was subtracted, and the progress curves, measured by the absorbance increase (⌬A 405 nm ), were fit by the parabolic Equation 1 (28,29) as follows.
This analysis gave the initial rate of chromogenic substrate hydrolysis at the beginning of the reaction (V 1 ) because of the conformationally activated SK⅐Pg* (where Pg* represents conformationally activated Pg) complex and the rate of the increase in activity with time due to Pm formation (V 2 ). For reactions in the presence of 6-AHA that were linear,   , and 3 mM SK 1-10 when present. After incubation for 5 min with ATA-FFR-CH 2 Cl and removal of excess inhibitor, labeling was performed as described above. Samples were analyzed by 4 -15% SDS polyacrylamide gel electrophoresis.

RESULTS AND DISCUSSION
Binding of SK  to Pg and Pm-The binding of SK 2-414 to fluorescent analogs of Pg and Pm that were specifically labeled at the catalytic site with ATA-FFR-CH 2 Cl and 5-(iodoacetamido)fluorescein (18,19) (Fig. 1). The binding of recombinant wildtype SK was similarly indistinguishable from native SK with a dissociation constant for [Lys]Pg of 77 Ϯ 12 nM (data not shown) and the previously determined value for [Lys]Pm of 33 Ϯ 6 pM (18). The maximum fluorescence change observed for labeled Pm upon binding SK 2-414 was Ϫ18 Ϯ 2%, which was considerably less than the Ϫ52 Ϯ 1% observed for native SK and Ϫ49 Ϯ 1% for wild-type SK (18), indicating significant differences in the active site environments of the SK⅐Pm and SK 2-414 ⅐Pm complexes. In contrast to the results for Pm, SK 2-414 binding to labeled Pg resulted in a maximum change in fluorescence of Ϫ24 Ϯ 1%, which was indistinguishable from that of native SK (Ϫ28 Ϯ 1%). These results showed that the lack of activity previously observed for SK   (13,14) was not attributed to the loss of affinity for [Lys]Pg or [Lys]Pm, and that the amino-terminal Ile residue did not contribute significantly to binding affinity. These results were in approximate agreement with the finding of ϳ5-fold reduced affinity for SK 2-414 compared with native SK, estimated from the kinetics of [Glu]Pg activation (13). However, Ile 1 was required to induce the full perturbation in the active site of Pm that accompanied the binding of SK as judged by the difference in the maximum fluorescence change.
Conformational Activation of Pg by SK  and SK 1-10 -The activity of SK 2-414 in plasminogen activation via the conformational change in Pg and the effect of synthetic peptides based on the SK amino-terminal sequence were examined in the kinetic studies of [Lys]Pg activation. Reactions were measured by continuous monitoring of the hydrolysis of the plasmin substrate H-D-Val-Leu-Lys-p-nitroanilide. As previously demonstrated (28,29), the addition of native SK to 5-20 nM Pg resulted in a parabolic increase in product concentration with time, which could be analyzed by the fitting of Equation 1 (see under "Experimental Procedures") to extract the initial rate of chromogenic substrate hydrolysis (V 1 ) because of rapid conformational activation of Pg and the subsequent slower rate of proteolytic conversion of Pg to Pm (V 2 ) from the acceleration phase. The dependence of V 1 on chromogenic substrate concentration at saturating levels of SK showed Michaelis-Menten kinetics with parameters for the activated SK⅐Pg* complex of 37 Ϯ 11 s Ϫ1 for k cat and 3.0 Ϯ 1.1 mM for k m (data not shown).
As observed previously (13), the reactions of Pg with saturating levels of SK 2-414 yielded progress curves with no initial activity followed after several minutes by rapid acceleration where the full-time course was not well described by Equation 1. Although this apparent change in mechanism is not completely understood, the acceleration phase has been previously shown to involve Pm formation (13) and may involve traces of Pm in Pg preparations that binds to SK 2-414 and proteolytically activates Pg. This activation pathway has been suggested (13) to account for previous reports that SK species modified at the amino terminus retained activity (15,16). Investigation of this secondary process was not pursued further, and the studies were focused on V 1 , which was the rate of interest in conformational activation of Pg.
The activity of SK  was tested further at high sensitivity in kinetic studies at 100 nM [Lys]Pg and 500 nM SK 2-414 in the absence and presence of the SK peptide SK 1-10 (Fig. 2). Limiting the assays under these conditions to the initial phase of Pg activation resulted in progress curves that were well described by Equation 1 and allowed V 1 to be determined accurately (Fig.  2). SK 2-414 alone had no detectable activity in conformational activation under these conditions as shown by the intercept of the linearly transformed data, corresponding to a value for V 1 that was indistinguishable from zero (Fig. 2, inset). An addition of 1 mM SK 1-10 , however, increased V 1 dramatically indicating that the activity of SK 2-414 in conformational activation was partially restored by the peptide (Fig. 2).

Effect of Aminoterminal Peptides on Conformational Activation of Pg by SK 2-414 -
The specificity of the binding site for amino-terminal peptides was first characterized in similar kinetic studies of the effects of various peptides on conformational activation of [Lys]Pg by SK 2-414 . SK 1-2 , SK 1-5 , SK 1-10 , and SK 1-15 increased V 1 linearly with increasing peptide concentration (Fig. 3A), indicating that the regained function in conformational activation was associated with a weak binding of the peptides to the SK 2-414 ⅐Pg complex. The effectiveness of the peptides increased with increasing length past the aminoterminal dipeptide sequence through at least SK 15 (Fig. 3A). On the basis of the kinetic parameters determined for the SK⅐Pg* complex, SK 1-10 and SK 1-15 at 1 mM supported a regain of 3.4 Ϯ 1% and 11 Ϯ 1% of the activity of native SK, respectively. Peptides corresponding to the Pg sequence, which normally inserts into the amino-terminal binding pocket upon proteolytic conversion to Pm, Pg 562-563 , Pg 562-565 , and Pg 562-569 (C566S), showed less activity than the SK peptides and no significant increase with increasing length (Fig. 3A). Pg 562-569 (C566S), for example, wasϽ5% as effective as SK [1][2][3][4][5][6][7][8][9][10] in restoring conformational activation. A sequence-scrambled peptide of identical composition to SK 1-10 induced no increase in V 1 , indicating the sequence specificity of the interaction of SK 2-414 ⅐Pg with SK 1-10 (Fig. 3B). In other control experiments, SK 1-10 had no significant effect on the kinetics of Pg activation by native SK or on the chromogenic substrate activity of the SK⅐Pm complex (data not shown).
Direct Demonstration of Activation of the Pg Catalytic Site by SK  and SK 1-10 -Active site-specific fluorescence labeling was used to confirm the identity of the active species formed in reactions of Pg with SK 2-414 and SK [1][2][3][4][5][6][7][8][9][10] . Reaction mixtures containing [Lys]Pg and SK 2-414 in the absence or presence of 3 mM SK 1-10 or the presence of the same concentration of scrambled SK 1-10 were inactivated with ATA-FFR-CH 2 Cl. The thiol group subsequently generated on the incorporated inhibitor was specifically labeled with 5-(iodoacetamido)fluorescein, and the reaction products were examined by SDS gel electrophoresis (Fig. 4). Covalent active site-specific labeling of Pg was observed in the presence of SK 1-10 as evidenced by the intensely fluorescent band corresponding to Pg in the reduced sample (Fig. 4, A and B, lanes 5), whereas a less intense band was also present that corresponded to the labeled light chain of Pm formed during the experiment (Fig. 4 A and B, lanes 6 and  7). Although the control reactions showed a significantly lower fluorescein incorporation, essentially all of the Pg in these reactions were converted to Pm at the high protein concentrations necessary for the experiment (Fig. 4B). Therefore, similar reactions were performed in the presence of 50 mM 6-AHA to inhibit proteolytic activation (30). It was first confirmed that under these conditions SK 1-10 activated SK 2-414 with indistinguishable potency as that seen in the absence of 6-AHA and also that the scrambled peptide was inactive (Fig. 3B). Pg was not significantly converted to Pm under these conditions, and the Pg zymogen was intensely labeled in the presence of SK 1-10 as before, whereas control reactions with scrambled SK 1-10 or no peptide resulted in no significant labeling (Fig. 4, C and D). These results clearly identified the intact Pg zymogen as the initially formed active species in the absence and presence of saturating 6-AHA and demonstrated that conformational activation required both SK 2-414 and SK [1][2][3][4][5][6][7][8][9][10] .
Sequence Specificity of the SK 1-10 Interaction-Previous studies (4) of trypsinogen demonstrated the specificity of the amino-terminal binding cleft for dipeptides in which the natural Ile-Val sequence bound most tightly and the extension of the peptides to three residues had no further effect. As shown by the results in Fig. 3A, the specificity of the analogous interaction of [Lys]Pg with dipeptides was characterized by very low but reproducible enhancements in the activity of comparable magnitude for Ile-Val and Ile-Ala (SK 1-2 ). Ac-Ile-Val and Phe-Val had no detectable activity, as predicted from the studies of trypsinogen, because of the lack of a free amino terminus and steric hindrance, respectively (4). Interestingly, the natural sequence in [Lys]Pm (Val-Val) was also inactive compared with Ile-Val, similar to the large difference in affinity observed for trypsinogen (4). These results supported the involvement of the amino-terminal binding site of Pg in the activation of the SK 2-414 ⅐Pg complex by SK 1-10 and suggested that the site had a trypsinogen-like specificity for dipeptides.
To address the specificity of the interaction of SK 1-10 in more detail, wild-type and nine single Ala-substituted SK 1-10 peptides were compared in conformational activation in the ab- sence and presence of near-saturating 6-AHA. The potencies of the peptides were assessed as described above from the slopes of linear plots of V 1 as a function of peptide concentration. As shown in Fig. 5, Ile 1 and Gly 3 were essential for peptide activity, whereas Trp 6 and Arg 10 contributed significant but smaller effects. Interestingly, the substitution of Asp 9 by Ala had a potency-enhancing effect of 2-fold, suggesting a negative role for this residue in the native sequence. The results were indistinguishable for experiments in the presence and absence of 6-AHA with one exception. The substitution of Pro 4 with Ala in SK 1-10 had little effect on the reactivation of SK 2-414 in the absence of 6-AHA but showed a large decrease in potency in the presence of the lysine analog (Fig. 5). The activity of this peptide in restoring conformational activation was apparently enhanced through lysine binding site interactions.
Effect of SK 1-10 on Pg Activation by the Truncation Mutants SK 55-414 and SK 78 -414 -The further truncated SK mutants SK 55-414 and SK 78 -414 were also inactive in the conformational activation of [Lys]Pg, but unlike SK 2-414 , they showed no detectable reactivation by SK [1][2][3][4][5][6][7][8][9][10] (Fig. 3B). The affinity of SK 55-414 for [5F]FFR-[Lys]Pg was determined from fluorescence titrations (Fig. 6) and compared with the previously determined affinity of this mutant for [5F]FFR-[Lys]Pm (18) and the corresponding affinities of fluorescein-labeled Pg and Pm for native and wild-type SK (18,19). SK 55-414 was shown previously to bind labeled [Lys]Pm with a dissociation constant of 4 Ϯ 1 nM, 360-fold more weakly than native SK (11 Ϯ 2 pM) (18). By contrast, SK 55-414 bound to [Lys]Pg with a dissociation constant of 79 Ϯ 24 nM, indicating a Ͻ2-fold lower affinity compared with the dissociation constants of 44 Ϯ 9 nM for native SK (18,19) (Fig. 6) and 77 Ϯ 12 for wild-type recombinant SK. The maximum fluorescence change for SK 55-414 binding to labeled [Lys]Pg of Ϫ35 Ϯ 3% was similar to the amplitude for native SK of Ϫ28 Ϯ 1% (19). Thus, the near total loss of reactivation of SK 55-414 in conformational activation by SK 1-10 was correlated with a differential loss of affinity for the activated catalytic domain of Pm of ϳ360-fold, whereas the binding to [Lys]Pg was affected Ͻ2-fold. This demonstrated a critical role for residues 3-54 of the SK ␣-domain in the preferentially high affinity of SK for Pm compared with Pg in conformational activation.
The results of these studies support the molecular sexuality hypothesis (5,13) and extend the understanding of the conformational activation mechanism in the conclusion that the combination of sequence-specific interactions of SK 1-10 with the amino-terminal binding cleft of Pg and enhanced binding af-finity of SK 2-414 for the activated conformation of the Pg catalytic domain are both critical elements of the mechanism. As expected from previous studies of trypsinogen and Ile-Val (4,5), the combination of SK 1-10 and SK 2-414 was less effective in Pg activation than native SK. This was consistent with the apparently low intrinsic affinity of the amino-terminal interaction and the loss of proximity and orientation available to the SK sequence in the intact molecule. Unlike previous studies of this site in trypsinogen, however, the SK peptides increased in potency with increasing length beyond the dipeptide and were significantly more effective than comparable Pg peptides. Moreover, an analysis of SK 2-414 reactivation by each of wildtype and nine Ala-substituted SK 1-10 peptides demonstrated that not only Ile 1 but also Gly 3 and, to a lesser extent, Pro 4 , Trp 6 , Asp 9 , and Arg 10 played significant roles in determining the specificity of the peptide interaction. On this basis, the results are concluded to reflect the specificity of the interactions of the SK 1-10 sequence with sites that extend beyond the amino-terminal dipeptide binding site. It is important to consider the possible effect of the 2-10 residue sequence present on SK 2-414 in activation by the peptides. The removal of Ile 1 in SK 2-414 that normally anchors the amino-terminal binding interaction may result in weakening the binding of the remaining SK 2-10 sequence to Pg or disordering of the SK 2-10 segment. The latter possibility is suggested by the disordered structure of this sequence when SK is bound to micro-Pm where Val 562 occupies the amino-terminal binding site and Ile 1 does not interact (17). In either case, the results would reflect the specificity of the amino-terminal binding site on Pg in the SK 2-414 ⅐Pg complex, although the possibility that the peptides interact with the SK component of the complex cannot be ruled out completely. The SK 1-10 sequence specificity of the aminoterminal interaction is suggested to contribute to the narrow specificity of SK for the activation of Pg among homologous zymogens.
The fluorescence probe in [5F]FFR-[Lys]Pm is thought to be in or near the S4 substrate specificity subsite. Previous studies with similar tripeptide fluorescence probes and chromogenic substrates demonstrated that the probes did not report changes affecting S1 but reported effects on the specificity of the S2 subsite, and these changes contributed to those reported by the probes located in the vicinity of S4 (18,19). Thus, the  (18,19). Fluorescence titrations were performed and analyzed as described under "Experimental Procedures." sources of the fluorescence changes include changes in the environment of the S2-S4 subsites, if not other events as well. In this context, the dissimilar maximum fluorescence changes observed with SK 2-414 and native SK binding to [5F]FFR-[Lys]Pm but not for [5F]FFR-[Lys]Pg are thought to represent Ile 1 -dependent differences in the subsite perturbations of Pg and Pm. Further studies will be needed to resolve the net fluorescence changes accompanying SK binding into the individual sources of the perturbations.
The inability of SK 1-10 to complement the activity of SK 55-414 and SK 78 -414 compared with SK 2-414 was correlated with a 360-fold reduced affinity of SK 55-414 for labeled [Lys]Pm (18), whereas the affinity for labeled [Lys]Pg was not significantly affected. This indicated that the interactions of residues 3-54 of the SK ␣-domain were not essential for Pg binding but were essential for both preferentially tighter binding of SK to Pm and for conformational activation of Pg. Together, the results support a mechanism in which the insertion of Ile 1 of SK into the amino-terminal binding cleft acts as the trigger to initiate the transition of Pg toward the active conformation. This is thought to occur in cooperation with preferential stabilization of the active zymogen species as a result of the higher affinity of SK for the activated conformation of the Pg catalytic domain.