Plasminogen Activation by Streptokinase via a Unique Mechanism*

The mechanism of human plasminogen (HPlg) activation by streptokinase (SK)-type activator was investi-gated with recombinant truncated SK peptides. An en-zyme-substrate intermediate of HPlg z SK z HPlg ternary complex was demonstrated by a sandwich-binding experiment. Formation of the ternary complex was saturable, HPlg-specific, and inhibited by 6-aminocaproic acid. Three interaction sites between SK and HPlg were demonstrated. SK-(220–414) bound to HPlg with two binding sites: one to the micro-HPlg region, the catalytic domain of HPlg, and one to the kringle 1–5 region, with K d values of 1.50 3 10 2 7 and 2.44 3 10 2 6 M , respectively. SK-(16–251) bound to a single site on the kringle 1–5 region of HPlg with a K d of 4.09 3 10 2 7 M . SK-(220–414) and SK-(16–251) competed for binding on the same or nearby location on the human kringle 1–5 domain. Combination of SK-(220–414) and SK-(16–251), but not either peptide alone, could effectively activate HPlg. In addition, SK-(16–251) dose-dependently enhanced the activation of HPlg by SK-(16–414), while the HPlg activation by SK-(16–414) was inhibited by SK-(220–414). We conclude midlog isopropyl-1-thio- b - D -galactopyranoside. 3 h, expressing cells target homogeneity exchange Sandwich Binding Assay— The assay was designed to detect the formation of the Plg SK z Plg complex. Wells of a strip plate (Costar) were coated with 0.2 ml HPlg or BPlg at a concentra- tion of 10 m g/ml in M carbonate/bicarbonate h at 4 °C and blocked with 1% BSA in phosphate-buffered saline (PBS) for 1 h. Native SK (1 m M ) was added to the wells as a bridge and incubated for 30 min at 4 °C. After washing with PBS containing 0.05% Tween 20, the corresponding 125 I-labeled Plgs, which were labeled by the IODO-GEN method as described previously (27), at various concentrations were added to each well and incubated for another 30 min at 4 °C. After washing with the same buffer, the radioactivity and the amount of 125 I-Plg was determined with a LKB g -counter. Nonspecific binding was determined by using BSA instead of SK protein. This was subtracted from the total binding to obtain the amount of specific binding. The binding of 125 I-HPlg and SK 1:1 complex, preformed in solution, to the HPlg-precoated plate was also measured. To test the effect of EACA on the binding of SK and HPlg, EACA at a final concentration of 10 m M was added to 125 I-labeled HPlg, and experiments were performed with the same procedures. The amount of immobilized Plgs coating on the surface of each well and the amount of bound SK as a bridge were estimated by two separate groups of experiments. We used 125 I-Plgs in place of unlabeled Plgs for a coating plate in one assay and 125 I-SK in place of the unlabeled SK as a bridge in the separate experiments. After conducting the same procedures as in the sandwich binding assay, we determined

Streptokinase (SK), 1 a potent plasminogen (Plg) activator, is a single-chain secretory protein produced by strains of ␤-hemolytic Streptococcus (1)(2)(3). The SK and human plasminogen (HPlg) complex can activate Plg to plasmin (Plm) from different mammalian species (3)(4)(5)(6)(7)(8)(9)(10). Plm thus produced in the blood circulation in turn catalyzes the hydrolysis of fibrin and dissolution of blood clots. SK has been used for more than 20 years as a thrombolytic agent in the treatment of thromboembolic blockages in the blood vessels such as acute myocardial infarction, although the mechanism of its function as a Plg activator remains unclear.
SK or HPlg alone does not have enzymatic activity. However, HPlg in interacting with SK becomes a virgin enzyme, which has amidolytic activity, although the activating peptide Arg 560 -Val 561 remains intact (5,6,8,11). Shortly after complex formation, HPlg⅐SK is converted to HPlm⅐SK (12), both of which can act as a Plg activator to catalyze the hydrolysis of the activating peptide bond of other Plgs (7,13,14). Plgs from many different mammalian species have been tested for their sensitivity to activation by SK, but only those from human, monkey, and cat are found to be sensitive (15), while those from cow, bovine, sheep, pig, mouse, and rat are not (3). The regions of amino acid sequences of Plgs that may account for the species specificity were determined with HPlg mutants (16).
The detailed three-dimensional structure of SK or Plg⅐SK complex based on x-ray crystallography has not been solved. According to amino acid sequence alignment, two internal homologous domains, SK-(1-173) and SK-(254 -415) were suggested (2). NMR and CD spectroscopy studies suggest that SK consists of at least three to four independent domains linked by mobile segments of the protein chain (17,18). Therefore, truncated SK proteins obviously would contain some intact structure of these domains. In the previous report, we demonstrated that most of the truncated SK peptides obtained by gene cloning techniques had secondary structures similar to those of the corresponding regions on native SK moieties, and they could be used to study the functions of the SK domains independently (19). Based on the study, five functional regions (a, Ile 1 -Lys 59 ; b, Ser 60 -Asn 90 ; c, Val 158 -Arg 219 ; d, Tyr 252 -Ala 316 ; e, Ser 317 -Ala 378 ) in the SK molecule for interaction with HPlg were deduced. From the results, we concluded that coordination of SK region Val 158 -Ala 378 is essential for a virgin enzyme formation and that coordination of SK region Ser 60 -Ala 378 is required for an effective SK-type HPlg activator. Additionally, the smaller SK fragments, SK-(16 -251) and SK-(220 -414), can bind to HPlg significantly, although they have neither virgin enzyme activity in a stoichiometric complex with HPlg nor HPlg activator activity. Multiple binding sites of HPlg and SK molecules have been proposed (20 -23). Nevertheless, how the separate HPlg-binding sites in the SK sequence interact with HPlg during HPlg activation remains a fundamental question to be resolved. In the present study, experiments were performed to elucidate the functions of NH 2 -and COOH-terminal SK binding regions. A more detailed mechanism of HPlg activation by SK was proposed.

EXPERIMENTAL PROCEDURES
Materials-Enzymes used in DNA manipulation were purchased from Boehringer Mannheim, Bethesda Research, or Promega Laboratories and were used according to the Cold Spring Harbor Manual or the recommendations by the suppliers. The full-length SK gene was obtained from Streptococcus equisimilis H46A (ATCC 12449) by the amplification with polymerase chain reaction as previously reported (19). Lys-Sepharose, CNBr-activated Sepharose 4B, and DEAE-cellulose were from Pharmacia Biotech Inc. Soybean trypsin inhibitor and NH 2 -D-Val-Leu-Lys-p-nitroanilide (S-2251) were obtained from Sigma. Bis-(sulfosuccinimidyl) suberate was purchased from Pierce; 6-aminocaproic acid (EACA) was from Merck. All chemicals were of the highest grade commercially available. HPlg and SK antiserum were prepared in our laboratory from mice.
Expression and Purification of Recombinant SK Truncated Peptides-SK peptides used in this study were obtained by recombinant gene cloning techniques as previously reported (19). Briefly, the fulllength SK gene was unidirectionally deleted by exonuclease III from the NH 2 or COOH terminus. Then the fragments coding for the truncated peptides were subcloned in frame into the overproducing plasmid pET-3a or pET-3b or pET-3c (Novagen) at the BamHI site. Transformed bacteria cells, Escherichia coli strain BL21(DE3)pLysS, were grown to midlog phase, and target gene expression was induced by adding 1 mM isopropyl-1-thio-␤-D-galactopyranoside. After 3 h, the expressing cells were harvested, washed, and disintegrated. Then the target proteins were concentrated by ammonium sulfate precipitation and purified to homogeneity by anion exchange chromatography methods with a DEAE-cellulose column.
Protein Concentration-The protein concentrations were determined spectrophotometrically using the following ⑀ 1% 280 nm values and molecular weights, respectively: Plgs, 17.0 and 94,000; micro-HPlg, 16 Sandwich Binding Assay-The assay was designed to detect the formation of the Plg⅐SK⅐Plg complex. Wells of a radioimmunoassay strip plate (Costar) were coated with 0.2 ml of HPlg or BPlg at a concentration of 10 g/ml in 0.05 M carbonate/bicarbonate buffer, pH 9.6, for 36 h at 4°C and blocked with 1% BSA in phosphate-buffered saline (PBS) for 1 h. Native SK (1 M) was added to the wells as a bridge and incubated for 30 min at 4°C. After washing with PBS containing 0.05% Tween 20, the corresponding 125 I-labeled Plgs, which were labeled by the IODO-GEN method as described previously (27), at various concentrations were added to each well and incubated for another 30 min at 4°C. After washing with the same buffer, the radioactivity and the amount of 125 I-Plg was determined with a LKB ␥-counter. Nonspecific binding was determined by using BSA instead of SK protein. This was subtracted from the total binding to obtain the amount of specific binding. The binding of 125 I-HPlg and SK 1:1 complex, preformed in solution, to the HPlg-precoated plate was also measured. To test the effect of EACA on the binding of SK and HPlg, EACA at a final concentration of 10 mM was added to 125 I-labeled HPlg, and experiments were performed with the same procedures. The amount of immobilized Plgs coating on the surface of each well and the amount of bound SK as a bridge were estimated by two separate groups of experiments. We used 125 I-Plgs in place of unlabeled Plgs for a coating plate in one assay and 125 I-SK in place of the unlabeled SK as a bridge in the separate experiments. After conducting the same procedures as in the sandwich binding assay, we determined the radioactivities and the amounts of bound proteins.
Concentration-dependent Binding Assay of 125 I-SK Peptides to HPlg and Determination of the Binding Affinity-Wells of a radioimmunoassay strip plate were coated with HPlg and blocked with BSA as above described. 125 I-SK peptides at various concentrations were added to the plate followed by incubation for 1 h at 4°C. After washing with PBS containing 0.05% Tween 20, the amount of 125 I-SK peptide was determined by an LKB ␥-counter. Nonspecific binding was measured by the addition of a 30-fold molar excess of each unlabeled peptide, and the value was subtracted from the total binding to obtain the amount of specific binding. The binding affinities of these 125 I-SK peptides to HPlg were determined by Scatchard transformation. To determine the binding affinity of 125 I-SK peptides to human kringle 1-5 or BPlg, the same concentration-dependent binding assay was performed by using human kringle 1-5 or BPlg instead of HPlg in the coating step.
Competitive Binding Assay-Competition between different HPlg moieties or different SK peptides for SK and HPlg binding was determined by adding excess unlabeled peptide fragments. To assay the competition between different Plg moieties, the amounts of 125 I-SK-(16 -251) or 125 I-SK-(220 -414) binding to HPlg with excess unlabeled HPlg fragments as competitors were analyzed. Wells of a radioimmunoassay strip plate were coated with HPlg and blocked with BSA as described above. Unlabeled human kringle 1-5, micro-HPlg, or BSA at various concentrations were then added to the HPlg-precoated wells for 5 min at 4°C. Subsequently, 125 I-SK-(16 -251) or 125 I-SK-(220 -414) was added to the wells and incubated for 1 h at 4°C. The binding procedures were performed as described above, and the amount of bound labeled SK-peptides was determined. The percentage of labeled SK peptide binding was determined by the fractional binding of the labeled SK peptides to HPlg in the presence of a given competitor in comparison with that occurring in the absence of any unlabeled competitor. To evaluate the competition between different SK peptides for binding to HPlg, the same binding procedure was carried out by adding the various unlabeled SK peptides as competitors prior to adding 125 I-SK-(16 -251) or 125 I-SK-(220 -414). The binding percentages in the presence of competitors were also determined.
Cross-link of Human Kringle 1-5 and Native SK-Human kringle 1-5 and native SK at a concentration of 0.3 mg/ml were mixed at room temperature for 3 min and followed by reaction with 0.1 mM bis-(sulfosuccinimidyl) suberate as cross-link reagent. After 60 min, the reaction was stopped by adding Tris buffer, pH 7.5, to a final concentration of 20 mM. Parallel samples containing either human kringle 1-5 or native SK were use as controls. The samples were then subjected to SDS-PAGE in 10% acrylamide according to the method of Laemmli (30). The electrophoresed proteins on the gel were transferred onto Immobilon-P transfer membrane (Millipore Corp.), and Western blotting was carried out with HPlg or SK antiserum (31).
Activation of HPlg by Catalytic Amounts of SK Peptides-A one-stage assay as described previously was used to measure HPlg activation by SK peptides (32,33). HPlg was activated by incubation with a catalytic amount of NH 2 -terminal SK peptide alone, COOH-terminal SK peptide alone, or combination of both SK peptides, at 37°C in 50 mM Tris, 0.1 M NaCl, pH 7.4, containing S-2251 as a substrate. The change in absorbance at 405 nm was monitored with a Hitachi 330 spectrophotometer. Competition assays were also performed to determine the effects of SK peptides on the activation of HPlg by SK- (16 -414). HPlg was incubated with a catalytic amount of SK-(16 -414) alone or in the presence of various concentrations of NH 2 -or COOH-terminal SK peptides at 37°C in 50 mM Tris, 0.1 M NaCl, pH 7.4, containing S-2251. The rate of HPlm production was monitored with a spectrophotometer.

RESULTS
Sandwich binding assays were performed to study the formation of the HPlg⅐SK⅐HPlg complex. In this experiment, a microtiter plate was first coated with HPlg, and native SK was added to allow HPlg⅐SK complexes to develop. 125 I-HPlg was then added to the HPlg⅐SK-coated plate. The amount of 125 I-HPlg bound to the HPlg⅐SK-coated plate, which represented that of the HPlg⅐SK⅐HPlg complex, was determined by measuring the radioactivity. Fig. 1 shows the increasing formation of HPlg⅐SK⅐HPlg complex in a dose-dependent and saturable manner and shows that 4.69 Ϯ 0.30 pmol of 125 I-HPlg was bound in the saturated condition. No binding of 125 I-HPlg was observed when BSA was used in place of SK. The binding of 125 I-HPlg and SK 1:1 complex, preformed in solution, to the HPlg-precoated plate was similar to the binding of 125 I-HPlg to the HPlg⅐SK-coated plate (Fig. 1). The level of bound 125 I-HPlg decreased in the presence of 10 mM EACA (Fig. 1). Only limited amount of ternary complex formed if BPlg was tested instead of HPlg (Fig. 1). In separate experiments, the 125 I-labeled proteins were used to determine the molarity of Plgs coated on the plate surface and the SK that bound to the Plgs on the surface. The amount of HPlg and BPlg coating on the surface of each well was 4.97 Ϯ 0.07 and 3.91 Ϯ 0.06 pmol (n ϭ 6), respectively. The amount of SK that bound to the immobilized HPlg and BPlg was 5.01 Ϯ 0.07 and 3.08 Ϯ 0.08 pmol (n ϭ 5), respectively. Therefore, it appeared that the binding of SK to either of these two Plgs coated on the well surface is very close to 1:1 stoichiometry. It also confirmed that a nearly 1:1:1 of HPlg⅐SK⅐HPlg ternary complex existed in our sandwich assay when referred to the amount of HPlg that was bound on the HPlg⅐SK-coated plate in the saturated condition (Fig. 1). However, in case of BPlg, only the BPlg⅐SK complex, not the BPlg⅐SK⅐BPlg ternary complex, was detected.
Recombinant truncated SK peptides of COOH-terminal or Competition between human kringle 1-5 or micro-HPlg with intact HPlg for SK binding sites was evaluated by a competitive binding assay. The binding of 125 I-SK-(16 -251) with HPlg was effectively competed for by human kringle 1-5, not by micro-HPlg (active site domain of HPlg) (Fig. 3A). Thus, SK-(16 -251) obviously could only interact with human kringle 1-5. On the other hand, both human kringle 1-5 and micro-HPlg could compete with HPlg for the binding of 125 I-SK-(220 -414) (Fig.  3B). At a high concentration of competitor peptide, 40 -60% of the binding was inhibited (Fig. 3B). These data confirmed that SK-(220 -414) could bind to HPlg through interaction with kringle domain and catalytic domain independently. The result of the cross-link experiment also confirmed that native SK can react with kringle domain of HPlg. Two high molecular weight protein bands were observed in the SDS-PAGE of cross-linked products of human kringle 1-5 and native SK (Fig. 4, lane 3). These bands contained both SK and HPlg proteins, since both bands showed positive reaction with anti-HPlg and anti-SK polyclonal antibodies (Fig. 4, lanes 4 and 5). The protein bands Competition between different SK moieties for binding to HPlg was also studied. The binding of SK- (16 -251) to HPlg was completely inhibited by SK-(220 -414) (Fig. 5A), and the binding of SK-(220 -414) to HPlg was partially inhibited by SK-(16 -251) as well as SK-(16 -191) (Fig. 5B). Therefore, the bind- Wells of a radioimmunoassay strip plate were coated with HPlg and blocked with BSA as in Fig. 1. 125 I-SK peptides at final concentrations ranging from 0 to 8 M were added to the plate, followed by incubation for 1 h at 4°C. The wells were then washed, and the radioactivity was determined. Nonspecific binding was measured by the addition of 30fold molar excess of each unlabeled peptide, and the value was subtracted from the total binding to obtain the amount of specific binding.
ing sites of SK-(16 -251) and SK-(220 -414) on the kringle domain should be at an overlapping location.
Enzymatic assays were performed to determine the functions of the two significant HPlg-binding SK peptides. Neither SK-(220 -414) nor SK-(16 -251) alone at a catalytic concentration could activate HPlg (Fig. 6A). However, when these two SK domains were added to HPlg simultaneously, HPlg was activated (Fig. 6A). In the parallel SDS-PAGE analysis, the conversion of HPlg to HPlm by the combination of SK-(220 -414) and SK-(16 -251) was completed in 30 min, as the band of HPlg (94-kDa) disappeared and both the heavy (66-kDa) and light (26-kDa) chains were detected (Fig. 6B, lane 4). To determine if SK-(220 -414) and SK-(16 -251) have effects on the activation of HPlg by full-length SK, HPlg was incubated with a catalytic amount of SK-(16 -414) alone or in the presence of various concentrations of SK fragments. SK-(16 -251) in a nanomolar concentration range could dose-dependently amplify the activation of HPlg by SK- (16 -414). Increasing the concentrations of SK-(16 -251) to micromolar range caused more rapid enhancement, and the extent of amplification reached a plateau (Fig. 7A). In contrast, nanomolar concentrations of SK-(220 -414) had no effect on the HPlg activation, and increasing amounts of SK-(220 -414) to micromolar concentrations progressively inhibited the activation of HPlg by SK-(16 -414) (Fig. 7B). DISCUSSION We have previously demonstrated that several regions in the SK molecule were essential for its functions as a plasminogen activator (19). In this study, we have used peptide fragments of SK and HPlg to further elucidate the interaction between these two molecules. It has been clearly demonstrated and widely accepted that formation of a 1:1 complex of SK and HPlg, as the catalytically active unit, is the first step in Plg activation (5,6,12). In the process of free HPlg activation, an enzyme-substrate intermediate, in the form of HPlg⅐SK⅐HPlg, might be present temporarily during activation. In our study, the possible existence of a HPlg⅐SK⅐HPlg ternary complex was demonstrated by a sandwich-binding experiment (Fig. 1). We also demonstrated that preformed HPlg⅐SK complex could bind to HPlg in an approximately 1:1 ratio. In contrast, in the report of Summaria et al. (20), a HPlm B-chain⅐SK⅐HPlg ternary complex, but not a HPlg⅐SK⅐HPlg ternary complex, was proposed based on the results of agarose double diffusion analysis. Two binding sites for the SK and HPlg interaction have been proposed, and when SK is bound to intact HPlg, the two interaction sites of SK are occupied by one single HPlg molecule. In the HPlm B-chain complex, only one of the two sites on SK is occupied, leaving the other one free for interacting with a second HPlg molecule. However, no explanation for the precipitin zones around the wells containing the HPlg and SK mixture was given (Fig. 1, A  (well 1) and B (well 4) of Ref. 20). One possible explanation is that HPlg and SK might form insoluble polymers in wells of agarose, which can not migrate into the agarose gel. The precipitin reaction of SK and HPlg, in a manner analogous to an antigen-antibody reaction, might be the result of formation of insoluble polymers in which each molecule can bind to more than one reacting molecule to form an insoluble network (20). Similar high molecular weight polymers were also observed in sedimentation velocity analysis and native gel electrophoresis of HPlg⅐SK complexes (34,35). Results of the sandwich binding assays give direct evidence to support the hypothesis that one SK molecule can interact with two HPlg molecules simultaneously.
Although two interaction sites on SK for HPlg were suggested in the previous report (20), the specific locations have not been identified. A straight line was obtained in the Scatchard plot of the binding of the NH 2 -terminal domain of SK to HPlg, indicating only one binding site was involved. On the other hand, a curved line was obtained in the similar plot of the binding of COOH-terminal domains of SK to HPlg, indicating in this case that two binding sites were involved (Fig. 2). The binding of the NH 2 -terminal domain of SK with HPlg was competed for by human kringle 1-5, but not by micro-HPlg, while the binding of COOH-terminal domains of SK with HPlg could be competed for by both kringle 1-5 and micro-HPlg (Fig.  3). Therefore, we conclude that SK had three binding sites for HPlg. One binding site located at the NH 2 -terminal domain of SK is specific for the interaction with the kringle domain of HPlg. There are two other binding sites on the COOH-terminal domains of SK; one specifically interacts with the catalytic domain, and the other interacts with the kringle domain of HPlg. Since the binding of the COOH-and NH 2 -terminal domains of SK is mutually exclusive for binding to the HPlg kringle domain (Fig. 5), these two binding sites of SK cannot interact with the same HPlg molecule simultaneously. On the basis of the binding constants we determined, we conclude that the order of binding strength was as follows: catalytic domain⅐SK-(220 -414) Ͼ kringle domain⅐SK-(16 -251) Ͼ kringle domain⅐SK-(220 -414). Taken together, the binding of COOH-terminal domains of SK to the catalytic and kringle domains of the first HPlg molecule and the binding of the NH 2 -terminal domain of the same SK to the kringle domain of a second HPlg molecule would result in the formation of the HPlg⅐SK⅐HPlg ternary complex as demonstrated in the sandwich binding assay (Fig. 1). The result of the cross-linking experiment (Fig. 4) also supports the existence of the ternary complex. The hypothetical binding mode could also explain the presence of high molecular weight complexes of these two molecules (34,35).
Formation of the ternary complex is inhibited by 10 mM EACA (Fig. 1), since it might interfere with the interaction of SK and lysine-binding sites on the kringle domain of HPlg. A similar explanation could also be applied to the observations that high molecular weight complexes of SK and HPlg shifted to a 1:1 complex in 100 mM EACA (34,36). Activation of HPlg by SK was substantially inhibited by EACA as described in a previous report (37) and our own observation. 2 The inhibitory effects of EACA could be partially due to its interference on the formation of the HPlg⅐SK⅐HPlg complex. When SK is used as a thrombolytic agent, with HPlg in excess and no EACA, the condition would be in favor of the formation of the HPlg⅐SK⅐HPlg ternary enzyme-substrate complex. The ternary complex might represent the intermediate of the two-stage activation process proposed by Reddy and Markus (6). It was clearly demonstrated that HPlg⅐SK as well as the HPlm⅐SK complex can activate Plgs with cleavage site-resistant HPlg (38). The HPlg⅐SK complex is remarkable in that SK induces a typical serine protease active site in the HPlg molecule. Finding the interaction sites between SK and HPlg should be valuable in revealing the mechanism and the species specificity of the active complex formation. Two peptide segments, Arg 579 -Met 584 and Arg 625 -Lys 627 (by the numbering convention for the full-length HPlg) were believed to be involved in a reaction with SK, by using the human-bovine hybrid micro-HPlg (39). In this report, we demonstrated that COOH-terminal, not NH 2 -terminal, domains of SK had strong affinity for micro-HPlg. Therefore, we assume that the COOH-terminal domains play a major role in interaction with micro-HPlg and are essential for induction of the active site in HPlg. The inertness of SK toward BPlg was also attributed to the specific interaction sites. The weak binding between BPlg and COOH-terminal domains of SK is consistent with the incapacity of SK to form activator with BPlg. However, the binding of BPlg and COOH-terminal domains was observed, although weaker than HPlg. There should be still some unidentified protein segments in HPlg that are involved in the interaction between Plgs and SK.
According to the reaction model proposed by Markus et al. (40), the role of the SK moiety in HPlg⅐SK⅐HPlg ternary complex may have two components: (a) SK may open the active center on one HPlg, thereby inducing it to serve as an activator HPlg; (b) SK may modulate the geometry near the Arg 560 -Val 561 bond of the substrate HPlg, and the activating peptide bond was brought toward the active center of the activator HPlg to form the transient Michaelis complex. The COOHterminal domains of SK Val 158 -Ala 378 were previously demonstrated to be responsible for the formation of enzymatic center in complexing with HPlg (19). The present study showed that HPlg could be effectively activated by combination of the two functionless SK peptides, SK-(220 -414) and SK- (16 -251). We assumed that the NH 2 -terminal domain SK-(16 -251) might provide the region of polypeptide 158 -220 in complement with SK-(220 -414) to form an activator with HPlg. In addition, SK-(16 -251) might also function as a substrate modulator, since it could amplify the activation of HPlg by SK-(16 -414) (Fig. 7A). The most plausible explanation for this observation is that HPlg, by interacting with the NH 2 -terminal domain of SK with its kringle 1-5 domain, becomes a better substrate for activation by the HPlg⅐SK activator complex. The hypothesis is also in consistent with the following observations: (a) a series of NH 2 -terminal truncated SK peptides still retains the comparable activities of virgin enzyme complex with HPlg but has lower rates of HPlg activation (19); (b) micro-HPlg, which lacks kringle domain, is activated by SK less effectively than Glu-HPlg (29). However, further studies are required to reveal the specific binding sites on the NH 2 -domain of SK and kringle of HPlg as well as their reaction mechanism. It is possible that binding of SK- (16 -251) to the kringle domain of HPlg induces the conversion of Glu-HPlg to the Lys-HPlg conformation, which might be attributed to the stimulating effect of the NH 2 -terminal domain of SK, since SK-(16 -251) cannot enhance activation of Lys-HPlg by SK. 2 However, it is also possible that the NH 2 -terminal domain of SK could induce the conformational changes near the activating peptide bond (Arg 560 -Val 561 ) of HPlg through a different mechanism. On the other hand, the HPlg activation by SK-(16 -414) is inhibited by SK-(220 -414). It could be due to the fact that excess SK-(220 -414) would compete with SK-(16 -414) for binding to kringle 1-5 as well as to the catalytic domain of HPlg and inhibit the formation of the HPlg activator (Fig. 7B). This observation is in agreement with the previous finding that SK-(244 -352) dosedependently inhibited the generation of an active site by fulllength SK and HPlg (41) .
The activation scheme of HPlg by SK is proposed, based on the findings on multiple functions of SK as shown in Scheme 1.
The COOH-terminal domains of SK can bind to the kringle and the catalytic domains of HPlg. The binding of the catalytic domain of HPlg to the COOH-terminal domains of SK is most important in inducing a conformational change in the active center of HPlg (Scheme 1, part a) and, therefore, forming an activator of HPlg, which can specifically hydrolyze the activating peptide bond Arg 560 -Val 561 of the substrate HPlg. The NH 2 -terminal domain of SK, by interaction with the kringle domain of HPlg, might alter the geometry of substrate HPlg especially near the region of the activating peptide bond (part b) and make it more effectively converted to HPlm (part c). The ternary complex in Scheme 1, HPlg⅐SK⅐HPlg, which was demonstrated in the sandwich binding assays, might also represent the transient Michaelis complex proposed by Markus et al. (40). The multiple functions of SK domains in reacting with HPlg render SK a very efficient HPlg activator. In the case of BPlg, the binding of its catalytic domain to the COOH-terminal domains of SK is significantly weaker than HPlg, and no ternary complex of BPlg and SK formed. This might be one of the reasons why BPlg cannot be activated directly by SK (3).
The reaction of SK and HPlg is a unique example of proteinprotein interaction that is very important in the regulation of various biological functions and deserves further delineation. SK, as a simple protein cofactor in the activation of HPlg, may provide a unique opportunity to study the mechanism of reaction between the protein cofactors and enzymes. The detailed mechanism, when elucidated, may be exploited to design a new generation of better SK as a thrombolytic agent.