Structure-Function Relationships in Staphylokinase as Revealed by “Clustered Charge to Alanine” Mutagenesis*

Eighteen mutants of recombinant staphylokinase (SakSTAR) in which clusters of two or three charged residues were converted to alanine (“clustered charge- to-alanine scan”) were characterized. Fifteen of these mutants had specific plasminogen-activating activities of (cid:62) 20% of that of wild-type SakSTAR, whereas three mutants, SakSTAR K11A D13A D14A (SakSTAR13), Sak- STAR E46A K50A (SakSTAR48), and SakSTAR E65A D69A (SakSTAR67) had specific activities of 3%. Sak- STAR13 had an intact affinity for plasminogen and a normal rate of active site exposure in equimolar mix- tures with plasminogen. The plasmin-SakSTAR13 complex had a 14-fold reduced catalytic efficiency for plas- minogen activation but was 5-fold more efficient for conversion of plasminogen-SakSTAR13 to plasmin-Sak- STAR13. SakSTAR48 and SakSTAR67 had a 10–20-fold reduced affinity for plasminogen and a markedly re- duced active site exposure; their complexes with plasmin had a more than 20-fold reduced catalytic efficiency toward plasminogen. Thus, plasminogen activation by catalytic amounts of SakSTAR is dependent on complex formation between plasmin(ogen) and SakSTAR, which is deficient with SakSTAR48 and SakSTAR67, but also on the induction of a functional active site configuration in the plasmin-SakSTAR measured at different Initial activation generated plasmin parameters ( K m and k cat ) were determined from Lineweaver-Burk plots by linear regression analysis. With wild-type SakSTAR, SakSTAR13, SakSTAR48, and SakSTAR67, similar kinetic was with low M r plasminogen. In experiments, the of the conversion of single chain rPlg-Ala 741 two rPli-Ala 741 either free or in an equimolar complex with SakSTAR13 (final concentration, 3 (cid:109) M by addition of preformed plasmin-SakSTAR13 complex (final concentration, 20 n M was monitored as a function of time. samples removed from the incubation mixtures at different time points (0–35 min) and subjected to SDS-PAGE under reducing conditions followed by quantitation of generated rPli-Ala 741 by densitometric scanning.

Staphylokinase, produced by certain strains of Staphylococcus aureus, activates the human plasma fibrinolytic system indirectly (1,2). It forms a 1:1 stoichiometric complex with plasmin, which activates plasminogen (3,4). Recombinant staphylokinase (SakSTAR) 1 was shown to induce fibrin-specific clot lysis in a human plasma milieu in vitro (5,6), in animal models of thrombosis (7), and in patients with acute myocardial infarction (8,9). This fibrin specificity has been explained by specific molecular interactions between SakSTAR, plasmin(ogen), ␣ 2 -antiplasmin, and fibrin. The plasmin-SakSTAR complex is rapidly inhibited by ␣ 2 -antiplasmin (6,10), resulting in dissociation of active SakSTAR from the complex and recycling to other plasminogen molecules (11). The inhibition rate of the complex by ␣ 2 -antiplasmin (second-order rate constant Ն 10 6 M Ϫ1 ⅐s Ϫ1 ) is, however, reduced more than 100-fold in the presence of fibrin (12). Thus, in the absence of fibrin, ␣ 2 -antiplasmin inhibits the activation of plasminogen by SakSTAR by preventing generation of active plasmin-SakSTAR complex. Fibrin stimulates plasminogen activation by SakSTAR via a mechanism involving the lysine-binding sites of plasminogen, probably by facilitating the generation of plasmin-SakSTAR complex and by delaying its inhibition at the surface of a clot (13).
SakSTAR consists of 136 amino acids, of which 45 are charged, in a single polypeptide chain without disulfide bridges; it consists of two widely separated domains of similar size linked by a flexible helix (14). To investigate the structurefunction relationships of SakSTAR, which determine its interaction with plasminogen in more detail, charged amino acids were mutagenized to alanine in clusters of 2 or 3 residues (clustered charge to alanine scan) (15). Replacement of clusters of charged amino acids in the SakSTAR regions comprising amino acids 11-14, 46 -50, and 65-69 resulted in a markedly reduced plasminogen activating capacity. Investigation of the mechanism of plasminogen activation by wild-type and mutant SakSTAR moieties supported a model involving formation of an equimolar complex of SakSTAR with traces of plasmin, which converts plasminogen to plasmin and, more rapidly, inactive plasminogen-SakSTAR to active plasmin-SakSTAR.
T4 DNA ligase, Klenow Fragment of E. coli DNA polymerase I and alkaline phosphatase were obtained from Boehringer Mannheim. The oligonucleotide-directed mutagenesis system and the pMa/c plasmids were kindly provided by Corvas (Ghent) (21). The expression vector pMEX602SAKB was constructed as described previously (9). M13KO7 helper phage was purchased from Promega (Leiden). Luria Broth growth medium was purchased from Life Technology (Merelbeke).
Laboratory Techniques-Plasmid DNA was isolated using a QIA-GEN-purification protocol (provided by Westburg N.V., Leusden, The Netherlands). Transformations of Escherichia coli were performed with the calcium phosphate procedure (22). DNA sequencing was performed using the dideoxy chain termination reaction method and Automated Laser Fluorescent A.L.F. (Pharmacia Biotech Inc.).
Protein concentrations were determined according to Bradford (23). SDS-PAGE was performed with the Phast System (Pharmacia) using 10 -15% gradient gels and Coomassie Brilliant Blue staining. Reduction of the samples was performed by heating at 100°C for 3 min in the presence of 1% SDS and 1% dithioerythritol. SakSTAR-related antigen was determined with a specific enzyme-linked immunosorbent assay (24), calibrated with the respective purified SakSTAR moieties. NH 2terminal amino acid sequence analysis was performed on an Applied Biosystems 477A protein sequencer with identification of amino acids by high performance liquid chromatography.
The specific plasminogen-activating activities of SakSTAR moieties were determined with a plasminogen-coupled chromogenic substrate assay, and were expressed in home units by comparison with an inhouse standard of natural staphylokinase, which was assigned an activity of 100,000 home units/mg of protein as determined by amino acid analysis (25).
Association rate constants (k a ), dissociation rate constants (k d ) and affinity constants (K a ϭ k a /k d ) for the interactions between different staphylokinase moieties and rPlg-Ala 741 or VFK-plasmin were determined by real time biospecific interaction analysis using the BIAcore instrument (Pharmacia) as described elsewhere (26). The kinetic parameters for the hydrolysis of S-2403 (final concentration, 0.05-1 mM) by plasmin-SakSTAR complexes (final concentration, 5 and 10 nM) were determined by Lineweaver-Burk analysis. Inhibition of plasmin-Sak-STAR complexes (final concentration, 5 nM) by ␣ 2 -antiplasmin (final concentration, 25 nM) was monitored continuously in the presence of S-2403 (final concentration, 0.3 mM), and the apparent second-order inhibition rate constants were calculated as described previously (6).
Construction of SakSTAR Mutants-In total, 20 SakSTAR mutants were constructed in which clusters of two or three charged residues were converted to alanine (Fig. 1). Eleven SakSTAR mutants (Table I) were constructed at the Center for Molecular and Vascular Biology in Leuven, Belgium, by site-directed mutagenesis, using the pMa/c vector and the repair-deficient E. coli strain WK6/MutS. Therefore, a 453base pair EcoRI-HindIII fragment containing the entire coding region of SakSTAR was cut out of the plasmid pMEX602SAKB and cloned into the EcoRI-HindIII sites of the pMc plasmid, which lacks the promotor sequence (yielding pMc-SakSTAR). For in vitro site-directed mutagenesis, single-stranded DNA was prepared by transformation of the pMc-SakSTAR construct in E. coli and infection of a fresh culture with helper phage M13KO7. Six h after infection, phages were isolated from the medium by precipitation with polyethylene glycol, and the single stranded DNA was extracted with phenol/chloroform treatment. Subsequently, single-stranded pMc-SakSTAR was hybridized with pMa (EcoRI-HindIII digested and isolated using the Prep-a-Gene protocol (Bio-Rad)) and the appropriate synthetic oligonucleotide. Oligonucleotides for mutagenesis consisted of 28 -44 base pairs with a (silent) mutation creating or deleting a restriction site (Table I). Extension reactions were carried out with the Klenow fragment of DNA polymerase and T4 DNA ligase, as described previously (22). After transformation of E. coli WK6/MutS cells and selection with ampicillin, colonies were grown on nitrocellulose membranes and denaturated in situ. DNA was hybridized overnight at room temperature using the respective radiolabeled mutant oligonucleotides (1.5 ϫ 10 8 cpm of [␥ 32 P]ATP for labeling of 20 -30 ng of oligonucleotide). Filters were washed at 42°C with solutions containing 0.1% SDS and 2 ϫ SSC, 1 ϫ SSC, 0.2 ϫ SSC, and 0.1 ϫ SSC (22). Positive clones were grown in a 3-ml volume, DNA was prepared by miniscreen and analyzed by restriction enzyme digestion. Larger scale preparations (250 ml) were prepared from positive clones, and mutations were confirmed by sequencing of the complete coding region for SakSTAR.
In addition, 9 mutants (Table I) were constructed at the Institute for Molecular Biotechnology in Jena, Germany. Mutants SakSTAR34, Sak-STAR58, SakSTAR94, SakSTAR97, SakSTAR100a, and SakSTAR109 were generated by cloning of couples of tailor-made sakSTAR encoding fragments generated by polymerase chain reaction in pMEX6. These fragments were produced using primers derived from the "left" and "right" flanking regions of the pMEX6 vector outside the sak gene (l-primer, CGTATAATGTGTGGAATTGTGAGCGG and r-primer, GGCTGAAAATCTTCTCTCATCCGCC) and "upward" or "downward" polymerase chain reaction primers in the sak gene (u-primer and d-primer) ( Table I). The l-and u-primers were used to generate the NH 2 -terminal SakSTAR-encoding fragment (N-fragment). The d-primer carrying the coding sequence for the charge to Ala mutation and the r-primer were used to generate the modified COOH-terminal SakSTAR-encoding fragment. The fragments were trimmed by appropriate restriction enzymes and cloned into the EcoRI-HindIII-digested expression vector pMEX6.
To construct mutants SakSTAR118 and SakSTAR120, the NH 2terminal SakSTAR encoding fragments were generated by polymerase chain reaction in pMEX6 using the l-and u-primers listed in Table I. The COOH-terminal SakSTAR encoding fragments were generated by annealing the complementary oligonucleotides (l-1 and l-2), encoding the charge to Ala modification and extending upstream to a protruding end compatible with the XmaI site induced in the codon for Pro 114 with the u-primer and downstream to the StyI site of SakSTAR. The EcoRI-XmaI digested NH 2 -terminal SakSTAR-encoding fragment and the COOH-terminal SakSTAR-encoding oligonucleotides were ligated in EcoRI-StyI-digested pMEX602SAKB. Mutant SakSTAR135 was generated by digestion of pMEX602SAKB with StyI and PstI, and ligation of the annealed complementary oligonucleotides l-1 and l-2 (Table I), with protruding StyI and PstI compatible ends, into the linearized vector.
Expression and Purification of SakSTAR Mutants-The mutated HindIII-EcoRI fragment was ligated back into the pMEX602SAKB expression vector with the pTaq promotor. The mutant proteins were produced intracellularly in soluble form in E. coli WK6 cells transformed with this vector. SakSTAR mutants, as well as wild-type protein, were purified from the sonicated bacterial extracts by cationexchange chromatography on Sephadex SP-50 and hydrophobic interaction chromatography on phenyl-Sepharose (9). Alternatively, mutants with a significantly decreased specific activity as compared with wild-type SakSTAR were also prepared with the freezing/thawing method, as described previously (27).
Complex Formation of SakSTAR Mutants with Plasminogen-The time course of active site generation in equimolar complexes of plasminogen with SakSTAR was monitored by titration with NPGB at room temperature, as described elsewhere (4, 28, 29). Concentrated stock The stability of equimolar plasmin-SakSTAR complexes was analyzed by incubation of 0.5 ml of a 3 M solution in 0.1 M phosphate buffer, pH 7.4, for 1 h at 4°C with 75 mg of suction-dried lysine-Sepharose gel. After excessive washing, the gel was eluted with 50 mM 6-aminohexanoic acid in 0.1 M phosphate buffer, pH 7.4, and the distribution of SakSTAR between the unbound fraction and the eluate was monitored by enzyme-linked immunosorbent assay and SDS-PAGE.
For kinetic analysis of plasminogen activation, equimolar plasmin-SakSTAR complexes (final concentration, 2 M) were prepared by incubation of plasminogen with the SakSTAR mutants at 37°C for 5-30 min in 0.1 M phosphate buffer, pH 7.4, containing 25% glycerol; the mixtures were then stored on ice. Plasmin-SakSTAR complex (final concentration, 10 -20 nM) was incubated with plasminogen (1-33 M) at 37°C in 0.1 M phosphate buffer, pH 7.4, and generated plasmin was measured at different time intervals (0 -4 min) with S-2403. Initial activation rates were obtained from linear plots of the concentration of generated plasmin versus time. Kinetic parameters (K m and k cat ) were determined from Lineweaver-Burk plots by linear regression analysis. With wild-type SakSTAR, SakSTAR13, SakSTAR48, and SakSTAR67, a similar kinetic analysis was performed with low M r plasminogen.
In separate experiments, the rate of the conversion of single chain rPlg-Ala 741 to two chain rPli-Ala 741 either free or in an equimolar complex with SakSTAR13 (final concentration, 3 M) by addition of preformed plasmin-SakSTAR13 complex (final concentration, 20 nM) was monitored as a function of time. Therefore, samples were removed from the incubation mixtures at different time points (0 -35 min) and subjected to SDS-PAGE under reducing conditions followed by quantitation of generated rPli-Ala 741 by densitometric scanning.

Production of SakSTAR Mutants
Mutant SakSTAR moieties were purified from E. coli WK6 cell culture medium with yields of 2-250 mg/l (Table I), representing recoveries of 17-88% of SakSTAR-related antigen. Wild-type SakSTAR was obtained in the same way with a yield of 83 mg/liter, representing a recovery of 15%. SakSTAR100a could not be obtained because of a very low expression level, and SakSTAR100b could not be purified because of apparent instability of the expressed protein, as shown by SDS-PAGE of crude cell extracts (data not shown). NH 2 -terminal amino acid sequence analysis of wild-type SakSTAR (220 pmol of starting material) revealed one homogeneous sequence (with yields in parenthesis) corresponding to Ser (184) Bradford (23), and yields are expressed in mg/liter of culture medium. b This mutant could not be obtained because of apparent protein instability as shown by SDS-PAGE of crude cell extracts. c This mutant could not be obtained because of a very low expression level. rectly processed in the expression system.
All purified proteins were homogeneous as shown by SDS-PAGE using 10 -15% gradient gels with or without reduction, and migrated with an apparent molecular mass of 18 kDa (not shown).

Functional Characterization of SakSTAR Mutants
Specific Activity-The specific activities of wild-type Sak-STAR and of 18 mutants, determined with a plasminogencoupled chromogenic substrate assay, are summarized in Table II. Two mutants (SakSTAR9 and SakSTAR97) had a specific activity of 20 -25% of that of wild-type SakSTAR and three mutants (SakSTAR13, SakSTAR48, and SakSTAR67) had a specific activity of Յ3% of that of wild-type, whereas the specific activities of the remaining 13 mutants were less than 2.5-fold different from the wild-type molecule.
Affinity for Binding to Plasmin(ogen)-Association and dissociation rate constants (k a and k d ) and apparent affinity constants (K a ) for binding of SakSTAR moieties to rPlg-Ala 741 , as measured by biospecific interaction analysis, are summarized in Table II. The K a value of SakSTAR48 and SakSTAR67 was, respectively, 22-and 12-fold lower than that of wild-type Sak-STAR, as a result of a more than 4-fold lower k a and an approximately 3-fold higher k d . K a values of the other mutants ranged between 1 ϫ 10 8 and 6 ϫ 10 8 M Ϫ1 , as compared with 2.4 ϫ 10 8 M Ϫ1 for wild-type SakSTAR. SakSTAR13, with a low specific activity, had an intact affinity for plasminogen (K a ϭ 2.2 ϫ 10 8 M Ϫ1 ).
Binding of SakSTAR13 or SakSTAR67 to VFK-plasmin occurred with a 3.5-or a 2.3-fold lower affinity than binding of wild-type SakSTAR. Binding of SakSTAR13 was characterized by (mean Ϯ S.D.; n ϭ 3) k a ϭ 4.4 Ϯ 0.12 ϫ 10 5 M Ϫ1 ⅐s Ϫ1 and k d ϭ 19 Ϯ 0.9 ϫ 10 Ϫ4 s Ϫ1 , yielding K a ϭ 2.4 Ϯ 0. Active Site Exposure in Equimolar Mixtures with Plasminogen-In equimolar mixtures of plasminogen and SakSTAR, the active site, as monitored by titration with NPGB, was rapidly exposed with wild-type SakSTAR and all its mutants, with the exception of SakSTAR48 and SakSTAR67 (Fig. 2). Indeed, active site exposure in mixtures of plasminogen (final concentration, 3.6 M) and wild-type SakSTAR (final concentration, 4 M) occurred with a lag phase (determined as the abscissa intercept of the linear phase of plots of generated plasmin-SakSTAR complex versus time) of 7 min, and with a rate (determined from the maximal slope) of 42 nM⅐s Ϫ1 , resulting in quantitative active site exposure. With SakSTAR13, corresponding values were 1 min for the lag phase and 21 nM⅐s Ϫ1 for the rate, resulting in 97% active site exposure. SakSTAR48 showed a delayed active site exposure, with a lag phase of 20 min and a rate of 15 nM⅐s Ϫ1 , resulting in quantitative active site exposure. In contrast, with SakSTAR67, no significant complex formation was observed after 60 min (Fig. 2). All other mutants were comparable with wild-type SakSTAR (data not shown). SDS-PAGE under reducing conditions (Fig. 2, inset) confirmed nearly quantitative conversion of plasminogen to plasmin in the mixtures with wild-type SakSTAR, SakSTAR13, and Sak-STAR48 but not with SakSTAR67. Active site exposure was also associated with conversion of SakSTAR to the corresponding SakSTAR-⌬10 moiety. When NPGB was added to the plasminogen solution before addition of SakSTAR, no active site generation was observed with wild-type SakSTAR or with the mutants SakSTAR13, SakSTAR48, or SakSTAR67 (data not shown).
Stability of Plasmin-SakSTAR Complexes-Adsorption of preformed equimolar mixtures of plasmin(ogen) and wild-type SakSTAR, SakSTAR13, SakSTAR48, or SakSTAR67 to lysine-Sepharose followed by elution with 6-aminohexanoic acid re- b Association rate constants (k a ), dissociation rate constants (k d ) and affinity constants (K a ) for binding to rPlg-Ala 741 were determined by biospecific interaction analysis. vealed the presence of stable complexes, as monitored by SDS-PAGE under nonreducing conditions (data not shown). Densitometric scanning confirmed that the ratio of plasminogen to SakSTAR in the eluates was comparable with that in the equimolar mixtures at the start. The relative areas under the curves were 92%:8% versus 90%:10% for wild-type SakSTAR, 88%:12% versus 90%:10% for SakSTAR13, 87%:13% versus 92%:8% for SakSTAR48, and 89%:11% versus 91%:9% for Sak-STAR67. Stability of the complexes was further confirmed by quantitation of SakSTAR-related antigen by enzyme-linked immunosorbent assay in the unbound fractions and in the eluates (96, 77, 69, and 83% recovery of the start material in the eluates for wild-type SakSTAR, SakSTAR13, SakSTAR48, and SakSTAR67, respectively).
Activation of Plasminogen-Catalytic amounts of wild-type SakSTAR induced rapid activation of plasminogen to plasmin (Fig. 3), with a lag phase (determined as the abscissa intercept of the linear phase of plots of generated plasmin concentration versus time) of 2.5 min and a rate (determined from the slope of the linear phase of these plots) of 1.4 nM⅐s Ϫ1 , resulting in approximately 80% plasminogen activation within 15 min. In contrast, SakSTAR13, SakSTAR48, and SakSTAR67 did not induce measurable plasmin generation within 50 min (detection limit Յ 0.03 nM⅐s Ϫ1 ). Addition of traces of plasmin to the plasminogen solution prior to the addition of SakSTAR, Sak-STAR13, SakSTAR48, or SakSTAR67 resulted in an enhanced activation rate of plasminogen, comparable or somewhat more rapid than that obtained by the addition of equimolar preformed plasmin-SakSTAR complexes (data not shown). All other SakSTAR mutants induced extensive plasminogen activation (71-108% of maximal), with lag phases ranging between 1.5 and 17 min and rates ranging between 0.77 and 2.3 nM⅐s Ϫ1 (data not shown). SDS-PAGE under reducing conditions (Fig.  3, inset) confirmed quantitative conversion of plasminogen to plasmin with wild-type SakSTAR, but no conversion with Sak-STAR13, SakSTAR48, or SakSTAR67.
Activation of plasminogen by preformed plasmin-SakSTAR complexes obeyed Michaelis-Menten kinetics, as revealed by linear double-reciprocal plots of the initial activation rate ver-sus the plasminogen concentration (data not shown). The K m ranged between 5 and 17 M for the mutants, including Sak-STAR13 (K m ϭ 11 M), SakSTAR48 (K m ϭ 7 M), and Sak-STAR67 (K m ϭ 5 M), as compared with 12 M for wild-type SakSTAR. In contrast the k cat values of SakSTAR13 (k cat ϭ 0.11 s Ϫ1 ), SakSTAR48 (k cat ϭ 0.03 s Ϫ1 ), and SakSTAR67 (k cat ϭ 0.03 s Ϫ1 ) were significantly lower than that of wild-type Sak-STAR (k cat ϭ 1.7 s Ϫ1 ) or of the other mutants (k cat ranging between 0.9 and 2.3 s Ϫ1 ). As a result, the catalytic efficiencies (k cat /K m ) for plasminogen activation of complexes of plasmin with SakSTAR13 (0.01 M Ϫ1 ⅐s Ϫ1 ), SakSTAR48 (0.005 M Ϫ1 ⅐s Ϫ1 ) and SakSTAR67 (0.006 M Ϫ1 ⅐s Ϫ1 ) were 14 -28-fold lower than that of wild-type SakSTAR (0.14 M Ϫ1 ⅐s Ϫ1 ), whereas k cat /K m of the other mutants was comparable with wild-type SakSTAR (0.11-0.24 M Ϫ1 ⅐s Ϫ1 ). The catalytic efficiencies of the mutant plasmin-SakSTAR complexes for plasminogen, relative to wild-type SakSTAR, are represented in Fig. 4.
Conversion of rPlg-Ala 741 (final concentration, 3 M) to rPli-Ala 741 by catalytic amounts (20 nM) of preformed plasmin-SakSTAR13 complex occurred with an initial rate of 0.37 nM⅐s Ϫ1 and reached a maximum of 270 nM after 12 min. In contrast, conversion to rPli-Ala 741 in the rPlg-Ala 741 ⅐SakSTAR13 complex was much more efficient, occurring with an initial rate of 1.4 nM⅐s Ϫ1 to reach a maximum of 1,700 nM after 20 min. Plasmin did not significiantly convert rPlg-Ala 741 to rPli-Ala 741 (Fig. 5).
Activation of low M r plasminogen by preformed plasmin-SakSTAR complexes also obeyed Michaelis-Menten kinetics. The kinetic parameters of the plasmin-SakSTAR13 complex (K m ϭ 4.0 M and k cat ϭ 0.08 s Ϫ1 ) and of the wild-type plasmin-SakSTAR complex (K m ϭ 5.3 M and k cat ϭ 1.6 s Ϫ1 ) for activation of low M r plasminogen were comparable with those for the activation of intact plasminogen, as reported above. In contrast, activation of low M r -plasminogen by the plasmin-Sak-STAR48 and plasmin-SakSTAR67 complexes was characterized by higher K m values (25 M and Ն60 M, respectively) and comparable or higher k cat values (0.2 s Ϫ1 and Ն0.2 s Ϫ1 ) than those for activation of intact plasminogen. Thus, the catalytic efficiencies for activation of low M r plasminogen were about 2-fold higher for wild-type plasmin-SakSTAR, plasmin-Sak-STAR13, and plasmin-SakSTAR48 and about 2-fold lower for plasmin-SakSTAR67, as compared with those for activation of intact plasminogen.
The activity of preformed plasmin-SakSTAR complexes toward S-2403, as determined by Lineweaver-Burk analysis, was comparable for wild-type SakSTAR and all the mutants. K m values for the mutants ranged between 120 and 390 M, as compared with 240 M for wild-type SakSTAR, and k cat values ranged between 50 and 100 s Ϫ1 , as compared with 63 s Ϫ1 for wild-type SakSTAR, yielding catalytic efficiencies of 0.16 -0.62 M Ϫ1 ⅐s Ϫ1 for the mutants, as compared with 0.26 M Ϫ1 ⅐s Ϫ1 for wild-type SakSTAR (data not shown).
Inhibition of Plasmin-SakSTAR Complexes by ␣ 2 -Antiplasmin-The second-order inhibition rate constants (k 1(app) ) for the inhibition of preformed mutant plasmin-SakSTAR complexes by ␣ 2 -antiplasmin ranged between (mean Ϯ S.D.; n ϭ 3 or 4) 1.6 Ϯ 0.05 ϫ 10 6 and 2.6 Ϯ 0.08 ϫ 10 6 M Ϫ1 ⅐s Ϫ1 , as compared with 2.0 Ϯ 0.01 ϫ 10 6 M Ϫ1 ⅐ s Ϫ1 for the wild-type plasmin-SakSTAR complex (data not shown). DISCUSSION In the present study, structure-function relationships in SakSTAR, which determine its interaction with plasminogen, were investigated by construction of mutants in which clusters of two or three charged amino acids were mutagenized to alanine. A clustered charge to alanine scan approach has previously been used to study structure-function relationships in other plasminogen activators, e.g. tissue-type plasminogen activator (30) and urokinase-type plasminogen activator (31).
Twenty mutants were designed, two of which could not be obtained because of a very low expression level or of protein instability (SakSTAR100a with E99A and E100A substitution, and SakSTAR100b with E99A, E100A, and K102A substitution). Out of 18 mutants that were studied in detail, only three (SakSTAR13, SakSTAR48, and SakSTAR67) were markedly different from wild-type SakSTAR with respect to their interaction with plasmin(ogen). This was revealed by a specific activity of Յ3% of that of wild-type, the absence of measurable plasminogen activation by catalytic amounts of the mutants, and a 10 -20-fold lower catalytic efficiency of preformed complexes with plasmin for the activation of plasminogen. Furthermore, two of these mutants (SakSTAR48 and SakSTAR67) had a 10 -20-fold reduced affinity for binding to plasminogen as compared with wild-type SakSTAR, and one mutant (Sak-STAR67) did not induce active site exposure in equimolar mix-tures with plasminogen.
The finding that mutations in the regions of amino acids 11-14, 46 -50, and 65-69 affected the interaction with plasmin(ogen) in different aspects suggested that mutants Sak-STAR13, SakSTAR48, and SakSTAR67 may be useful to study different steps in the mechanism of the interaction of SakSTAR with plasminogen. The following kinetic model has been proposed for the activation of plasminogen by SakSTAR (4).
Plasminogen (P) and SakSTAR (S) produce an inactive 1:1 stoichiometric complex (P⅐S), which does not activate plasminogen, as demonstrated by titration with NPGB. The activation reaction appears to be initiated by trace amounts of contaminating plasmin (p), which generates p⅐S, which converts P to p and P⅐S to p⅐S. According to this model, SakSTAR48 and Sak-STAR67, with a 10 -20-fold reduced affinity for binding to plasminogen, have an impaired formation of p⅐S and P⅐S complexes, and thus of subsequent conversion of P to p and of P⅐S to p⅐S. This is confirmed by our findings that no active site exposure occurred in equimolar mixtures of plasminogen with SakSTAR67, whereas with SakSTAR48, active site exposure was markedly delayed as compared with wild-type SakSTAR (Fig. 2). With SakSTAR13, which had a normal affinity for plasminogen, formation of p⅐S and of P⅐S was apparently normal, as was conversion of P⅐S to p⅐S, but conversion of P to p was markedly impaired.
Preformed complexes of SakSTAR13, SakSTAR48, or Sak-STAR67 with plasmin (p⅐S) had a comparable affinity for the plasminogen substrate as wild-type SakSTAR (K m values for plasminogen activation of 5-12 M) but a much lower catalytic rate constant for plasminogen activation (k cat of 0.03-0.11 s Ϫ1 , as compared with 1.7 s Ϫ1 ), resulting in 14 -28-fold lower catalytic efficiencies. These findings explain why catalytic amounts of SakSTAR13 (normal formation of p⅐S complex, but low enzymatic activity) or of SakSTAR48 and SakSTAR67 (delayed p⅐S formation) do not induce measurable plasminogen activation.
The finding of rapid and quantitative formation of p⅐S in equimolar mixtures of plasminogen and SakSTAR13, despite the low catalytic efficiency of the p⅐S complex of this mutant for activation of P, can be explained by the observation that the conversion of P⅐S to p⅐S by catalytic amounts of preformed p⅐S is much more efficient than conversion of P to p. The finding that the rate of plasminogen activation by catalytic amounts of SakSTAR48 and SakSTAR67 is enhanced by the addition of traces of plasmin, from undetectable to a rate similar to that observed with their preformed complexes with plasmin, suggests that, under these conditions, active p⅐S is formed, not as a result of conversion of P⅐S to p⅐S but of direct formation of the p⅐S complex between added plasmin and SakSTAR. However, the rate of plasmin generation is still much lower than that observed in control experiments with wild-type SakSTAR because of the low catalytic efficiency of the mutant p⅐S complexes. The affinities of SakSTAR48 and SakSTAR67 for VFKplasmin are 10 -20-fold higher than for plasminogen, which is confirmed by the observation that preformed complexes of plasmin with SakSTAR13, SakSTAR48, and SakSTAR67 are stable following adsorption onto lysine-Sepharose.
Taken together, these data further demonstrate that the running concentration of active p⅐S determines the rate of conversion of P to p and of P⅐S to p⅐S. In the presence of excess p inhibitor (e.g. NPGB), P⅐S and P cannot be converted because of lack of active p⅐S. In the absence of inhibitor, the reaction most likely is initiated by trace amounts of p that form p⅐S that converts both P to p and, more rapidly, P⅐S to p⅐S. Indeed, a contamination of P with 30 ppm of p is sufficient to explain the kinetics of activation of P by SakSTAR (4).
In summary, mutagenesis in the regions 11-14, 46 -50, or 65-69 of SakSTAR resulted in impairment of its interaction with plasmin(ogen). These mutants have allowed to identify different steps of the interaction between plasminogen and SakSTAR in functional isolation.