Structure-Function Analysis of the Streptokinase Amino Terminus (Residues 1-59)*

Transcription

1 Structure-Function Analysis of the Streptokinase Amino Terminus (Residues 1-59)* Lakshmi V. Mundada, Mary Prorok, Melanie E. DeFord, Mariana Figuera, Francis J. Castellino, William P. Fay From the Research Service, Ann Arbor Veterans Affairs Hospital, and the Department of Internal Medicine, University of Michigan Medical School, Ann Arbor, MI (LVM and WPF), and from the Department of Chemistry and Biochemistry, University of Notre Dame, Notre Dame, IN (MP, MED, MF, FJC). *This work was supported by a Merit Review Award from the Department of Veterans Affairs (WPF) and by NIH grants ΗL (WPF) and HL13423 (FJC). To whom correspondence should be addressed. Mail: University of Michigan Medical Center, 1150 W. Medical Center Dr., Ann Arbor, MI , USA; Tel.: (734) ; Fax: (734) ; wfay@umich.edu. Running Title: Functional Analysis of the Streptokinase N-Terminus

2 SUMMARY Streptokinase (SK) binds to plasminogen (Pg) to form a complex that converts substrate Pg to plasmin (Pl). Residues 1-59 of SK regulate its capacity to induce an active-site in bound Pg by a non-proteolytic mechanism and to activate substrate Pg in a fibrin-independent manner. We analyzed 24 SK mutants to better define the functional properties of SK Mutations within the α β 1 strand (residues 17-26) of SK completely prevented non-proteolytic active-site induction in bound Pg and rendered SK incapable of protecting Pl from inhibition by α 2 -antiplasmin (AP). However, when fibrin-bound, the activities of α β 1 strand mutants were similar to that of wild-type (WT) SK and resistant to α 2 -AP. Mutation of Ile 1 of SK also prevented non-proteolytic active-site induction in bound Pg. However, unlike α β 1 strand mutants, the functional defect of Ile 1 mutants was not relieved by fibrin, and complexes of Ile 1 mutants and Pl were resistant to α 2 -AP. Plasmin enhanced the activities of α β 1 strand and Ile 1 mutants, suggesting that SK-Pl complexes activated mutant SK-Pg complexes by hydrolyzing the Pg Arg 561 -Val 562 bond. Mutational analysis of Glu 39 of SK suggested that a salt bridge between Glu 39 and Arg 719 of Pg is important, but not essential, for non-proteolytic active-site induction in Pg. Deleting residues 1-59 rendered SK dependent on Pl and fibrin to generate plasminogen activator (PA) activity. However, the PA activity of SK in the presence of fibrin was markedly reduced compared to WT SK. Despite its reduced PA activity, the fibrinolytic potency of SK was greater than that of WT SK at higher (but not lower) SK concentrations due to its capacity to deplete plasma Pg. These studies define mechanisms by which the SK α domain regulates rapid active-site induction in bound Pg, contributes to the resistance of the SK-plasmin complex to α 2 -AP, and controls fibrin-independent Pg activation. 2

3 Streptokinase (SK), 1 a protein secreted by several streptococcal species, binds and activates human plasminogen. Streptokinase may play an important role in streptococcal virulence by generating proteolytic activity at the bacterial cell surface, thereby facilitating invasion of host tissues (1). Streptokinase is also used as a clot-dissolving agent to treat thrombotic diseases (2). Unlike mammalian plasminogen activators, which activate plasminogen by limited proteolysis, streptokinase lacks intrinsic enzymatic activity. Rather, SK binds to plasminogen and induces within it an active-site by a non-proteolytic mechanism (3). The SK-plasminogen complex converts free plasminogen molecules to plasmin. Plasminogen within the activator complex also is converted to plasmin by a poorly understood proteolytic mechanism, and the SK-plasmin complex is resistant to inactivation by α 2 - antiplasmin (4). Streptokinase consists of 440 amino acids, including a 26 amino acid N-terminal signal peptide that is cleaved during secretion to yield the mature 414 amino acid protein of M r 47,000 (2). Streptokinase contains 3 domains, denoted α (residues 1-150), β (residues ), and γ (residues ) (5). Each domain binds plasminogen, though none can activate plasminogen independently (5-8). Several lines of evidence suggest that residues 1-59 of the SK α domain play a key role in plasminogen activation. The Lys 59 -Ser 60 peptide bond of SK is cleaved very rapidly upon incubation of SK with plasminogen, and the resulting N-terminal peptide remains associated with the SK-plasminogen complex (8;9). SK deletion mutants lacking residues 1-59 form complexes with plasminogen that exhibit markedly reduced plasminogen activator (PA) activity, despite the fact that these mutant SK-plasminogen complexes can generate near normal amidolytic activity after prolonged incubation (7;10-12). 3

4 Incubation of peptides containing amino acids 1-59 of SK with inactive deletion mutants lacking these residues restores PA activity (8;9;13). Deletion of Ile 1 of SK markedly inhibits its capacity to induce an active-site in plasminogen, which supports the hypothesis that establishment of a salt-bridge between Ile 1 of SK and Asp 740 of plasminogen is necessary for SK to induce an active-site in plasminogen by a non-proteolytic mechanism (11;14). Residues 1-59 also appear to play an important role in determining SK s capacity to efficiently activate plasminogen in the absence of fibrin, a property that distinguishes SK from tissue-type plasminogen activator (t-pa). Whereas full-length SK efficiently activates plasminogen in the absence or presence of fibrin, the very low PA activity of SK is greatly enhanced by fibrin (15). Several questions regarding the function of the amino terminal fragment (residues 1-59) of the SK α domain remain unanswered. First, what residues play key roles in mediating the functional effects of the amino terminal fragment on plasminogen activation? Although the role of Ile 1 of SK in plasminogen activation has been examined in some detail (11;14), less is known about the function of other specific residues (16-19). Second, what is the mechanism by which complexes composed of plasminogen and SK mutants lacking portions of the amino terminus attain normal or near-normal amidolytic activity, but only after a prolonged lag phase? And third, what are the mechanisms underlying the fibrin specificity exhibited by SK mutants lacking residues 1-59? The purpose of this study was to address these issues by generating and characterizing an extensive panel of recombinant SKs carrying mutations within residues

5 EXPERIMENTAL PROCEDURES Reagents: Glu-plasminogen was purified from human plasma by lysine-sepharose chromatography (20). Purified plasminogen was incubated with D-Phe-Phe-Arg chloromethylketone dihydrochloride (Calbiochem-Novabiochem), a plasmin inhibitor, and dialyzed extensively before use. Plasmin was prepared by incubating plasminogen (2 mg/ml) with urokinase (0.5 U/mL, American Diagnostica). Chromogenic plasmin substrate (S-2251) was from Chromogenix. Human α 2 -antiplasmin and soluble fibrin fragments (Desafib-X) were from American Diagnostica. Recombinant human plasminogen resistant to hydrolysis at Arg 561 (Pg R561A) was generated and purified as described (21). Human α-thrombin was from CalBiochem. Generation and purification of recombinant SK: Mature (i.e. 414 amino acids, N-terminal isoleucine), wild-type SK (derived from S. equisimilis H46A) was generated and purified as described previously (10). Point mutations were introduced into the streptokinase coding sequence by using the pselect mutagensis vector (Promega) according to the manufacturer s instructions. Ile 1 and all clusters of charged amino acids within residues 1-59 of SK were mutated to alanine residues, as described previously (22). SK mutants lacking Ile 1 (i.e. SK ) or residues 1-59 (i.e. SK ) were generated by a PCR-based strategy, as described (23). All mutations were confirmed by DNA sequencing. Mutant streptokinases were subcloned into pet-3a, expressed in E.coli strain BL21(DE3)pLysS as mature-length proteins, and purified to homogeneity as described (10). Protein sequence analysis of Ile 1 point and deletion mutants confirmed the presence of the intended N-terminal residue. 5

6 Functional assays: The capacity of streptokinase to generate amidolytic activity when in complex with plasminogen was determined by incubating SK (180 nm), plasminogen (60 nm), and S (300 µm) in 0.01 M Tris-HCl, 0.14 M NaCl, ph 7.5 (TBS) at room temperature and monitoring the absorbance of reaction mixtures at 405 nm (A405) for 15 minutes. The amidolytic activity of each complex, expressed as % wild-type, was calculated by dividing the A405/min of mutant complex by the A405/min of wild-type SK complex, then multiplying by 100. Standard curves constructed from different concentrations of wild-type SK-plasminogen complex revealed that A405/min was a linear function of complex concentration. SK plasminogen activator activity was studied by incubating SK (3.5 nm), plasminogen (700 nm), and S-2251 (300 µm) in TBS at room temperature and measuring the A minutes later. The plasminogen activator activity of each mutant, expressed as % wild-type, was calculated by dividing the A405 of mutant SK by the A405 of wild-type SK, then multiplying by 100. Standard curves of wild-type SK revealed that A405 was a linear function of SK concentration over the range of nm. To study the role of the SK α domain in the resistance of the SKplasmin complex to inhibition by α 2 -antiplasmin, SK (50 nm) and plasmin (10 nm) were incubated for 5 minutes at room temperature, then α 2 -antiplasmin (50 nm) was added. Five minutes later, residual amidolytic activity of complexes was determined by hydrolysis of S Effects of fibrin on activity of SK-plasminogen complex: To study the effects of fibrin on the amidolytic activity of SK-plasminogen complexes, SK (240 nm), Pg R561A (120 nm), and S (500 µm) were incubated in TBS with 1 mg/ml BSA in the absence or presence of Desafib X (desaa fibrinogen X, 1.2 µm), and the A405 of the reaction mixture was monitored. Effects 6

7 of fibrin on PA activities of complexes of SK and wild-type plasminogen were determined by incubating SK ( nm), Glu-plasminogen (700 nm), Desafib X (1 µm), and S-2251 (300 µm) and monitoring A405. To study the amidolytic and PA activities of SK-plasminogen complexes bound to insoluble fibrin matrices, fibrin-coated microtiter plate wells (96 wells/plate) were prepared by incubating fibrinogen (1.4 µm) and thrombin (0.5 U/mL) in 100 µl of PBS (0.05 M sodium phosphate, 0.15 M NaCl, ph 7.5) and air drying plates overnight at 37 C. Wells were blocked for 1 hour with 3% BSA containing D-Phe-Pro-Arg chloromethylketone dihydrochloride, a thrombin inhibitor (10 µm, CalBiochem), and washed 5 times with rinsing buffer (0.9% NaCl, 0.05% Triton X-100). Wild-type plasminogen or Pg R561A ( nm in 100 µl TBS containing 1 mg/ml BSA) was added to wells. One hour later wells were washed extensively, then SK (0-150 nm) was added. To measure the amidolytic activity of complexes, S-2251 (300 µm) was added to wells in the absence or presence of α 2 - antiplasmin (240 nm) and the A405 of wells was monitored. To measure the PA activity of fibrin-bound complexes, wells were washed extensively 30 minutes after adding SK, then 100 µl of wild-type plasminogen (1 µm) was added to wells. After 20 minutes at room temperature 25 µl were removed and transferred to wells containing S-2251 (final concentration 300 µm) to determine plasmin formation. Plasma clot lysis: Pooled, citrated human plasma (50 µl), containing a trace amount of 125 I- human fibrinogen (Amersham Pharmacia), was clotted in the presence of a plastic rod (to facilitate clot washing and transfer) by addition of thrombin (1 U/mL) and CaCl 2 (5 mm). After one hour at 37 C, clots were washed 5 times with TBS containing hirudin (1 µg/ml), then suspended in citrated plasma (450 µl) containing streptokinase (10-50 nm) and hirudin (1 7

8 µg/ml). After minutes of gentle rocking at 37 C, % clot lysis was calculated by comparing radioactive counts in the plasma to those present in the washed clot prior to its placement in plasma. Plasma fibrinogen concentrations were measured as described (15). Other methods: Protein concentrations of purified proteins were determined with the BCA Reagent (Pierce). DNA sequencing was performed by the Sanger method. SDS-PAGE was performed with the PhastSystem (Pharmacia). Gels were stained with Coomassie Brilliant Blue R. Differential scanning calorimetry was performed as described (21). Circular dichroism (CD) spectra analyses were performed at 25 C with a JASCO J-600 spectropolarimeter with a 1.0 nm bandwidth and a 0.5 nm step resolution. 8

9 RESULTS Effects of point mutations on SK PA activity. We generated an extensive panel of point mutations within residues 1-59 of SK. Mutations targeted by a charged-to-alanine strategy did not significantly decrease SK PA activity (Table 1), suggesting that none of the clusters of charged residues in this region was essential for normal activator activity. Of note, a 3-fold increase in PA activity was observed for SK K36A. Mutation of Ile 1 of SK to an alanine caused a moderate decrease in PA activity (to 23+7% that of wild-type SK) that was characterized by a lag phase in the appearance of plasmin formation, after which the PA activity of SK I1A was indistinguishable from that of wild-type SK (Fig. 1). A relatively long series of hydrophobic amino acids (residues 11-26), which is missed by a charged-to-alanine mutagenesis strategy, is located within the SK amino terminal fragment, and deletions into this region are associated with a marked reduction in PA activity (9;10;24). Therefore, we generated a series of neutral-to-charged mutations between positions of SK. Several mutations within this region decreased PA activity to approximately 1% that of wild-type SK (i.e. L18D, V20D, V20K, and V22D), while the PA activities of SK V19D and SK G24D were reduced to % and 2.5+2% that of wild-type SK, respectively. Differential scanning calorimetry of SK S16D, L18D, and G24D revealed that they exhibited a single temperature of maximum heat capacity (T m ) of approximately 45 C, while wild-type SK and SK K36A exhibited two T m of approximately 45 C and C (Table 2). However, the secondary structural characteristics of mutants exhibiting abnormal calorimetry profiles (i.e. SK S16D, L18D, and G24D), as evaluated by CD spectroscopy in the near- ( nm) and far- ( nm) UV regions, did not differ from those of wild-type SK (data not shown). These results 9

10 suggested that the mutations introduced between positions altered the tertiary structure of the α domain of SK. Effects of point mutations on SK amidolytic activity. Charged-to-alanine mutations induced at most only minor reductions in the amidolytic activities of SK-plasminogen complexes (Table 1). The amidolytic activity of plasminogen in complex with SK K36A was approximately 2- flold greater than that of wild-type complex (Table 1, Fig. 2). The amidolytic activity observed during a 15 minute incubation of SK I1A with Glu-plasminogen was severely reduced, though gradually increasing rates of S-2251 hydrolysis were observed after a prolonged lag phase (Fig. 2). Similarly, the cluster of mutations between positions of SK markedly inhibited the amidolytic activity of the activator complex. However, in 2-stage assays (in which delayed generation of amidolytic activity during incubation of equimolar SK and plasminogen was confirmed by hydrolysis of S-2251 [stage 1], then a 200-fold molar excess of plasminogen was added to the reaction mixture to assess PA function [stage 2]), the PA activity of SK I1A was indistinguishable from that of wild-type SK, while the PA activities of position mutants were not detectable or markedly reduced (data not shown). These results suggested that once an active-site was formed, complexes containing mutations at position 1 of SK processed substrate plasminogen normally, whereas complexes containing mutations at positions of SK were severely defective in converting substrate plasminogen to plasmin. Analysis of mechanisms underlying the lag in the generation of amidolytic activity by SK α- domain mutants. We studied the capacity of recombinant SKs (240 nm) to induce amidolytic activity in Pg R561A (120 nm), a mutant zymogen incapable of conversion to plasmin (21). 10

11 Neither SK I1A nor any of the position mutants was able to generate amidolytic activity during a 1 hr incubation at 37 C (data not shown). All other point mutants induced amidolytic activity comparable to that of wild-type SK during incubation with Pg R561A, except for SK E39A, D41A, which exhibited an activity of % that of wild-type SK. These results suggested that mutations at position 1 or position of SK prevented rapid, proteolysisindependent formation of an active-site in complexed plasminogen. Addition of plasmin (1.6 nm) to reaction mixtures of wild-type plasminogen (80 nm) and SK I1A, SK L18D, or SK G24D (each at 240 nm) accelerated S-2251 hydrolysis to a much greater extent than could be explained simply by the amidolytic activity of added plasmin (Fig 3B-D), but had no significant effect on wild-type SK (Fig. 3A). However, addition of trace plasmin did not enable SK I1A, SK L18D, or SK G24D to generate amidolytic activity in Pg R561A (data not shown), suggesting that plasmin catalyzed active-site formation by a mechanism that involved hydrolysis of the plasminogen Arg 561 -Val 562 bond. Consistent with these data, SDS-PAGE analysis of equimolar reaction mixtures of SK I1A and wild-type plasminogen revealed that the appearance of amidolytic activity correlated with the delayed conversion of plasminogen within the activator complex to plasmin, whereas the appearance of amidolytic activity of complexes composed of wild-type SK and wild-type plasminogen was immediate and preceded conversion of plasminogen to plasmin (Fig. 4). Fibrin dependence of SK α-domain point mutants. Wild-type SK or SK point mutants were incubated with Pg R561A in the presence or absence of soluble fibrin fragments and amidolytic activity was monitored. Whereas SKs containing mutations between position did not generate significant amidolytic activity in the absence of fibrin, their activities in the presence 11

12 of fibrin were comparable to that of wild-type SK (Fig. 5). However, fibrin had no significant effect on the markedly reduced capacity of SK I1A to generate amidolytic activity in Pg R561A. Similarly, the PA activities of fibrin-bound complexes containing Pg R561A and SKs with mutations at residues were comparable to those containing wild-type SK, while the PA activity of fibrin-bound complexes containing SK I1A was markedly reduced, and the PA activity of fibrin-bound complexes containing SK was undetectable (data not shown). These results indicated that 1) point mutations within the β 1 strand of the α domain generated SKs that exhibited markedly reduced amidolytic and PA activities in the absence of fibrin, but normal or nearly normal activities in the presence of fibrin, and 2) even in the presence of fibrin, Ile 1 was required for SK to efficiently induce active-site formation in plasminogen by a non-proteolytic mechanism. Inhibition of activator complexes by α 2 -antiplasmin and modulation by fibrin: To study the role of the SK α domain in the resistance of the SK-plasmin complex to inhibition by α 2 - antiplasmin, we incubated SK (50 nm) and plasmin (10 nm) for 5 minutes, then added α 2 - antiplasmin (50 nm). Five minutes later, residual amidolytic activity of complexes was determined by hydrolysis of S No significant inhibition of wild-type SK-plasmin complex by antiplasmin was observed. In contrast, marked inhibition of S-2251 hydrolysis by antiplasmin was observed during incubation of plasmin with the following SK mutants: S16K, L18D, V19D, V20K, V20D, V22D, G24D, and SK Complexes of plasmin and all other point mutants listed in Table 1 were not inhibited by α 2 -antiplasmin (data not shown). Consistent with these results, antiplasmin (240 nm) essentially completely inhibited the generation of amidolytic activity during a 1 hour incubation of SK L18D or SK G24D (each at 12

13 240 nm) with plasminogen (80 nm) and trace plasmin (1.6 nm Fig. 3C-D). However, complexes formed by incubating plasminogen and SK L18D or SK G24D in fibrin-coated microtiter plate wells were resistant to inhibition by α 2 -antiplasmin (Fig. 6). These results suggested that the SK α domain contributes to the resistance of the SK-plasmin complex to inhibition by α 2 -antiplasmin, and that mutations within the β 1 strand of the α domain generated mutants with fibrin-dependent resistance to antiplasmin. Effects of fibrin on the activity of SK : The PA activity of SK is severely reduced, yet is markedly enhanced by fibrin (15). We performed a series of experiments to identify potential mechanisms underlying the effect of fibrin on SK SK (240 nm) did not generate amidolytic activity during incubation with Pg R561A (120 nm), either in the absence or presence of soluble fibrin fragments (1.2 µm, Fig. 5), suggesting that fibrin could not substitute for residues 1-59 of SK to provide structural elements necessary to induce an activesite in plasminogen by a non-proteolytic mechanism. However, fibrin significantly enhanced the capacity of complexes of SK and wild-type plasminogen to generate PA activity, although the specific activity of SK in the presence of fibrin was markedly less than that of wild-type SK (Fig. 7A). Addition of trace plasmin potently accelerated the generation of PA activity during incubation of SK and wild-type plasminogen in the presence of fibrin, but not in its absence (Fig. 7B). These results suggested that activation of plasminogen by SK required not only fibrin, but also SK plasmin complex (as opposed to SK-plasminogen complex), with the plasmin provided as a trace contaminant in the plasminogen preparation. However, the rate of plasminogen activation by SK plasmin, even in the presence of fibrin, was markedly slower than that induced by wild-type SK. 13

14 Previous studies suggested that SK exhibits a greater fibrinolytic potency than wild-type SK (15). We compared the capacity of wild-type SK and SK to lyse plasma clots. At a concentration of 10 nm, wild-type SK exhibited much greater fibrinolytic potency than SK (Fig. 8A). At a concentration of 50 nm, wild-type SK was superior to SK at early time points (15-30 min), but at later time points SK induced more extensive clot lysis than did wild-type SK (Fig. 8B). Plasma fibrinogen depletion during clot lysis with wild-type SK was relatively mild at a concentration of 10 nm, but was severe at a concentration of 50 nm (Fig. 8C). In contrast, minimal depletion of plasma fibrinogen was observed with either concentration of SK These results suggested that SK exhibits greater fibrinolytic potency than wild-type SK only under conditions in which plasma concentrations of wild-type SK are high enough to induce extensive activation of plasminogen in plasma, thereby depleting plasma fibrinogen and plasminogen. Consistent with this hypothesis, addition of purified plasminogen to plasma during clot lysis potentiated the fibrinolytic efficacy of wild-type SK, but had no detectable effect on SK (Fig. 8B). 14

15 DISCUSSION None of the mutants generated by a charged-to-alanine mutagenesis strategy exhibited a significant defect in amidolytic or PA activity during incubation with wild-type plasminogen. However, when the panel of charged-to-alanine mutants was screened with Pg R561A (i.e. under conditions in which active-site formation can occur only via a non-proteolytic mechanism) the capacity of SK E39A, D41A to induce amidolytic activity was moderatelyseverely reduced. Residues of SK are located in the β 2 strand of the α domain. These results suggest an important role of the α β 2 strand in establishing an active-site in plasminogen and support the hypothesis that a salt bridge between Glu 39 of SK and Arg 719 of plasminogen plays an important, but not essential, role in the formation of an active complex (5). We also observed that mutation of Lys 36, located immediately adjacent to the β 2 strand, resulted in a significant increase in SK amidolytic activity. Within the activator complex, Lys 36 of SK is in close proximity to the active-site of plasmin (5). It is possible that mutations at or close to this residue produce steric or charge effects that facilitate substrate access to the active-site. These results support the feasibility of using mutagenesis to increase SK function. Mutations in the α β 1 strand (residues 17-26) of SK markedly reduced the formation of an active complex during incubation with plasminogen. Mutations between residues may have altered the position or mobility of the SK N-terminus and prevented Ile 1 from assuming a conformation required for it to interact with Asp 740 of plasminogen to induce active-site formation by a non-proteolytic mechanism (14). However, complexes of plasmin and SKs harboring mutations between positions were severely defective in PA activity, 15

16 whereas complexes of SK I1A or SK and plasmin exhibited normal PA activity. These results suggested that 1) although Ile 1 of SK plays a key role in inducing an active-site in complexed plasminogen (11;14), Ile 1 is not required for normal processing of substrate plasminogen by the complex, and 2) the loss of PA function of position mutants was not mediated simply by steric effects on Ile 1. Since residues reside within the hydrophobic core of the α domain, mutations in this region may have disrupted protein folding and substrate turnover by the activator complex. Consistent with this hypothesis, mutations in this region were associated with apparent structural changes in SK, as assessed by differential scanning calorimetry. However, the lack of significant changes in the CD spectra of these mutants, and the observation that fibrin restored the capacity of position mutants (but not SK ) to induce amidolytic activity in Pg R561A, suggested that their potential alterations in α domain structure were not irreversible, and probably minor. The SK α β 1 strand has extensive contact with residues of plasminogen, with Arg 719 forming a van der Waals contact with Val 19 of SK (5). Our data support the hypothesis that contact between residues of plasminogen and the β 1 strand of the SK α-domain is essential for rapid formation of a fully functional complex. Mechanism of delayed generation of amidolytic activity by SK mutants lacking Ile 1. Although Ile 1 of SK appears to be required for SK to induce an active-site in complexed plasminogen (11;14), complexes of plasminogen and SK N-terminal deletion mutants lacking Ile 1 can, after a lag phase, generate nearly normal amidolytic activity (7;10-12). Our studies suggest that active (i.e. SK I1A-plasmin) complexes can activate inactive (i.e. SK I1A-plasminogen) complexes by hydrolyzing the latter s Arg 561 -Val 562 plasminogen bond, as hypothesized by Wang et al (14). It 16

17 appears that SK mutants lacking normal function of Ile 1, due to mutation of that residue, or due to steric effects of other mutations, require trace amounts of plasmin, which binds to SK to generate active complexes that trigger progressive activation of other complexes, resulting in an activation mechanism similar to that of staphylokinase (25). Role of residues 1-59 of SK in fibrin-independent plasminogen activation. Deletion of residues 1-59 of SK renders it fibrin-dependent by a poorly defined mechanism (15). We found that fibrin did not enable SK to generate amidolytic activity in Pg R561A, suggesting that fibrin cannot substitute for the N-terminal fragment of SK to enable formation of an activator complex by a proteolysis-independent mechanism. However, our results suggest that once plasmin is generated, fibrin can partially substitute for the N-terminal fragment of SK to enable a reduced turnover rate of substrate plasminogen, while free SK plasmin complexes are essentially devoid of PA activity. The environment of the fibrin clot is ideally suited to provide plasmin to prime SK , since fibrin enhances plasminogen activation by t-pa, and fibrin-bound plasmin is resistant to inhibition by α 2 -antiplasmin (26). Our data suggest that the fibrin selectivity of SK is accounted for by its requirement of both fibrin and plasmin to generate a complex with PA activity, thereby targeting plasminogen activation to the clot surface. Reed et al. found that SK was a superior fibrinolytic agent to wild-type SK at all concentrations studied (15). In contrast, we found that SK induced less clot lysis than wild-type SK at low concentrations, but more lysis than wild-type SK at higher concentrations. The fibrinolytic superiority of wild-type SK over SK at lower concentrations is consistent with our kinetic data. The superiority of SK over wild-type SK at higher concentrations appears to be due to the capacity of wild-type SK, but not SK , to activate plasminogen in 17

18 plasma, thereby leading to plasminogen steal and an arrest of clot lysis (27). Consistent with this hypothesis, addition of purified plasminogen to plasma accelerated clot lysis by wild-type SK, but not SK Since SK N-terminal deletion mutants have been proposed as potential therapeutic agents (15), it is important to realize that their fibrinolytic efficacy depends on their plasma concentration. The effect of fibrin on the amidolytic and PA activities of SK α β 1 strand point mutants was similar to that observed with SK , although more potent. In the presence of fibrin, the activities of SK α β 1 strand point mutants were significantly greater than that of SK and comparable to that of wild-type SK. However, even in the presence of fibrin, Ile 1 is required to efficiently induce an active-site in plasminogen by a non-proteolytic mechanism. We hypothesize that the fibrin dependence of α β 1 strand mutants is explained by conformational changes induced by fibrin that restore the ability of Ile 1 of SK to establish a salt bridge with Asp 740 of plasminogen, which appears necessary for non-proteolytic active-site induction (14). Role of SK α domain in resistance of the activator complex to α 2 -antiplasmin. Our studies suggest an important role of the SK α domain in mediating the resistance of the SK-plasmin complex to inhibition by antiplasmin. Previous studies demonstrated that the affinity of SK for plasmin is markedly reduced compared to wild-type SK (28), suggesting that residues 1-59 of SK play a key role in maintaining the cohesiveness of the SK-plasmin complex. It is possible that deletion of residues 1-59 or mutation of residues within the α β 1 strand enable antiplasmin to displace SK from the SK-plasmin complex and inhibit plasmin. 18

19 In summary, our studies identify key residues within residues 1-59 of SK that mediate its functional properties, and they provide insights into the mechanisms by which the SK α domain coordinates the rapid assembly of a functional activator complex, facilitates fibrinindependent plasminogen activation, and contributes to the resistance of the activator complex to inhibition by α 2 -antiplasmin. Further studies are indicated to examine the potential use of SK α domain mutants as fibrin-specific plasminogen activators. 19

20 REFERENCES 1. Boyle MDP and Lottenberg R (1997) Thromb Haemost. 77, Reddy KN (1988) Enzyme 40, Bajaj AP and Castellino FJ (1977) J Biol Chem. 252, Cederholm-Williams S, De Cock F, Lijnen H, and Collen D (1979) Eur J Biochem. 100, Wang X, Lin X, Loy J, Tang J, and Zhang X (1998) Science 281, Conejero-Lara F, Parrado J, Azuaga AI, Dobson CM, and Ponting CP (1998) Protein Sci. 7, Rodriguez P, Fuentes P, Barro M, Alvarez JG, Munoz E, Collen D, and Lijnen HR (1995) Eur J Biochem. 229, Shi G-Y and Wu, D.-H. (1994) Biochem J. 304, Nihalani D, Kumar R, Rajagopal K, and Sahni G (1998) Protein Sci. 7, Fay WP and Bokka LV (1998) Thromb Haemost. 79, Wang S, Reed G, and Hedstrom L (1999) Biochemistry 38, Young K-C, Shi G-Y, Chang Y-F, Chang L-C, Lai M-D, Chuang W-J, and Wu H-L (1995) J Biol Chem. 270,

21 13. Parrado J, Conejero-Lara F, Smith RAG, Marshall JM, Ponting CP, and Dobson CM (1996) Protein Sci. 5, Wang S, Reed G, and Hedstrom L (2000) Eur J Biochem. 267, Reed GL, Houng AK, Liu L, Parhami-Seren B, Matsueda LH, Wang S, and Hedstrom L (1999) Proc Natl Acad Sci. 96, Lee SH, Jeong ST, Kim IC, and Byun SM (1997) Biochem Molecl Biol Intl. 41, Kim D, Lee S, Kim I, Kim S, and Byun S (2000) Thromb Res. 99, Lee BR, Park SK, Kim JH, and Byun SM (1989) Biochem Biophys Res Commun. 165, Liu L, Sazonova IY, Turner RB, Chowdhry SA, Tsai J, Houng AK, and Reed GL (2000) Journal of Biological Chemistry 275, Deutsch D and Mertz E (1970) Science 170, Nilsen SL and Castellino FJ (1999) Protein Expr Purif 16, Bennett WF, Paoni NF, Keyt BA, Botstein D, Jones AJS, Presta L, Wurm FM, and Zoller MJ (1991) J Biol Chem. 266, Boxrud PD, Verhamme IMA, Fay WP, and Bock PE (2001) J Biol Chem. 276, Jackson KW and Tang J (1978) Thromb Res. 13,

22 25. Lijnen HR, Van Hoef B, DeCock F, Okada K, Ueshima S, Matsuo O, and Collen D (1991) J Biol Chem. 266, Lee AYY, Fredenburgh JC, Stewart RJ, Rischke JA, and Weitz JI (2001) Thromb Haemost. 85, Torr SR, Nachowiak DA, Fujii S, and Sobel BE (1992) J Am Coll Cardiol. 19, Boxrud PD, Fay WP, and Bock PE (2000) J Biol Chem. 275,

23 FOOTNOTES 1 The abbreviations used are: SK, streptokinase; PA, plasminogen activator; TBS, Tris-buffered saline; PBS, phosphate-buffered saline. 23

24 FIGURE LEGENDS Fig. 1. Plasminogen activator activity of SK I1A. Plasminogen (700 nm) was incubated alone or with wild-type SK or SK I1A (3.5 nm), and plasmin formation was monitored by hydrolysis of S-2251 (300 µm). Experiment was repeated in triplicate with indistinguishable results. Fig. 2. Amidolytic activities of SK-plasminogen complexes. Plasminogen (80 nm) was incubated alone or with SK (240 nm), and generation of amidolytic activity was monitored by hydrolysis of S-2251 (300 µm). Experiment was performed > 3 times with similar results. Fig. 3. Effects of plasmin and α 2 -antiplasmin on the amidolytic activities of activator complexes. Streptokinase (240 nm) was incubated with plasminogen (80 nm) and generation of amidolytic activity was monitored by hydrolysis of S-2251 (300 µm). Plasmin (PL, 1.6 nm) with or without α 2 -antiplasmin (AP, 240 nm) was added to some reaction mixtures, as indicated. A. Wild-type SK. Dashed line represents control reaction consisting of plasmin (1.6 nm) and S Open circles represent incubation of plasminogen (80 nm) and S-2251 in the absence of SK. B. SK I1A. C. SK L18D. D. SK G24D. Fig. 4. SDS-PAGE analysis and amidolytic activity of SK I1A-plasminogen complexes. Wild-type SK or SK I1A (7.8 µm) and wild-type plasminogen (7.8 µm) were incubated at 15 C. At the indicated time points, samples of the reaction mixture were removed and subjected to SDS-PAGE analysis under reducing conditions (A. Wild-type SK; B. SK I1A) and 24

25 measurement of amidolytic activity (after 100-fold dilution) by incubation with 300 µm S-2251 (C). Molecular weights (in kd) are shown at right of panels A-B. Plg, plasminogen; Pl-HC, plasmin heavy chain, PL-LC, plasmin light chain, SK, native and proteolytically modified forms of streptokinase. Experiment was performed 3 times with similar results. Fig. 5. Effect of fibrin on the amidolytic activities of complexes of Pg R561A and SK. Pg R561A (120 nm) was incubated with wild-type SK (240 nm), the indicated SK mutants (240 nm), or buffer (No SK) in the absence (white bars) or presence (black bars) of soluble fibrin fragments (1.2 µm). Amidolytic activity was measured by hydrolysis of S Data represent mean + 1 S.D. of 3-4 experiments. Fig. 6. Resistance of fibrin-bound activator complexes to inhibition by α 2 -antiplasmin. Plasminogen (50 nm) was incubated in fibrin-coated wells for 1 hour, then wells were washed and SK (150 nm) and S-2251 (300 µm) were added in the absence or presence of α 2 - antiplasmin (240 nm). Control fibrin-coated wells contained Pg R561A, but no SK. Amidolytic activity of fibrin-bound complexes was determined by measuring the rate of S hydrolysis. Data represent mean + 1 S.D. of 3 experiments. Fig. 7. Effect of fibrin and plamin on the plasminogen activator activity of SK (A) Effect of fibrin. Wild-type SK or SK (10 nm) were incubated with plasminogen (Plg, 300 nm) and S-2251 (300 µm) in the absence or presence of soluble fibrin fragments (Fn, 1 µm), and plasmin formation was determined by measuring substrate hydrolysis. Control reactions lacking SK are shown. (B) Effect of plasmin. SK (10 nm) was incubated with 25

26 plasminogen (Plg, 300 nm) and S-2251 (300 µm). Soluble fibrin fragments (Fn, 1 µm) and/or plasmin (PL, 6 nm) were added to some reactions, as indicated, and a control reaction containing plasminogen, plasmin, and fibrin, but lacking SK , was included. Plasmin formation was determined by measuring S-2251 hydrolysis. Fig. 8. Induction of plasma clot lysis by wild-type SK and SK Radiolabeled fibrin clots were suspended in hirudin-anticoagulated plasma. Wild-type SK or SK was added, and clot lysis was monitored by measuring radioactive counts in the plasma at the indicated time points. (A) Concentration of SK was 10 nm. A control reaction lacking SK is also shown. (B) Concentration of SK was 50 nm. As indicated, additional plasminogen (Pg, 0.43 nmole) was added to some reaction mixtures at each time point at which clot lysis was measured. (C) Capacity of wild-type SK and SK to deplete plasma fibrinogen. Wild-type SK or SK (each at a concentration of 10 nm or 50 nm) were added to plasma, and 2 hours later plasma fibrinogen concentration was measured. Results are standardized to control reactions, which lacked SK. Data represent mean of triplicate experiments. 26

27 Table 1. Amidolytic and Plasminogen Activator Activities of Streptokinase Mutants in Complex with Plasminogen Amidolytic Activity of SK- Plasminogen Complex (% of wild-type SKplasminogen complex) Plasminogen Activator Activity SK-Plasminogen Complex (% of wild-type SK-plasminogen complex) Mutant I1A < E5A D9A, R10A V13D V13K S16D S16K L18D V19D V20D < V20K V22D < G24D V26D V26K E27A D32A K36A E39A, D41A R45A, H48A K51A, E53A K59A, K61A Data are mean + 1 S.D. of triplicate experiments. 27

28 Table 2. Temperature(s) of maximum heat capacity (T m ) of recombinant streptokinases. SK T m Wild-type 47.4, 54.9 SK S16D 45.3 SK L18D 45.7 SK G24D 46.1 K36A 48.1, 60.5 Each T m corresponds to the midpoint of the peak of the raw thermogram. Values shown represent the mean of 2 separate determinations, which varied by <2%. 28

29 A WT I1A No SK Minutes Fig. 1 29

30 A Minutes WT K36A G24D I1A No SK Fig. 2 30

31 A B WT WT+PL WT+PL+AP No SK Plasmin Minutes Minutes I1A I1A+PL I1A+PL+AP Plasmin 0 A405 A C D L18D L18D+PL L18D+PL+AP Plasmin Minutes Minutes G24D G24D+PL G24D+PL+AP Plasmin A405 A405 Fig. 3

32 Plg Pl-HC A SK 43 Pl-LC sec 10 min 30 min 1 hr 2 hr Plg B 97 Pl-HC 68 SK 43 Fig. 4 Pl-LC Amidolytic Activity (mod/min) sec 10 min 30 min 1 hr 2 hr C 29 SK I1A WT SK 32

33 60 A405/min (mod/min) (-) Fibrin (+) Fibrin Fig. 5 0 WT S16K L18D V19D V20D V22D G24D I1A No SK 33

34 A405/min (mod/min) (-) Antiplasmin (+) Antiplasmin Fig. 6 0 WT L18D V19D G24D No SK 34

35 A WT SK+Plg WT SK+Plg+Fn SK Plg SK Plg+Fn Plg Plg+Fn Minutes A Fig. 7 A SK Plg+Fn SK Plg+Fn+PL SK Plg SK Plg+PL Plg+PL+Fn Minutes B 35

36 80 60 WT SK SK No SK A % Clot Lysis Hours 120 B 100 % Clot Lysis Plasma Fibrinogen, % Control nM 50nM WT SK Hours WT SK WT SK + Pg SK SK Pg SK C Fig. 8 36

37 Structure-function analysis of the streptokinase amino terminus (residues 1-59) Lakshmi V. Mundada, Mary Prorok, Melanie E. DeFord, Mariana Figuera, Francis J. Castellino and William P. Fay J. Biol. Chem. published online April 18, 2003 Access the most updated version of this article at doi: /jbc.M Alerts: When this article is cited When a correction for this article is posted Click here to choose from all of JBC's alerts