The β‐domain of streptokinase affects several functionalities, including specific/proteolytic activity kinetics

Streptokinase (SK) is a plasminogen activator which converts inactive plasminogen (Pg) to active plasmin (Pm), which cleaves fibrin clots. SK secreted by groups A, C, and G Streptococcus (SKA/SKC/SKG) is composed of three domains: SKα, SKβ and SKγ. Previous domain‐swapping studies between SK1/SK2b‐cluster variants revealed that SKβ plays a major role in the activation of human Pg. Here, we carried out domain‐swapping between skcg‐SK/SK2‐cluster variants to determine the involvement of SKβ in several SK functionalities, including specific/proteolytic activity kinetics, fibrinogen‐bound Pg activation and α2‐antiplasmin resistance. Our results indicate that SKβ has a minor to determining role in these diverse functionalities for skcg‐SK and SK2b variants, which might potentially be accompanied by few critical residues acting as hot spots. Our findings enhance our understanding of the roles of SKβ and hot spots in different functional characteristics of SK clusters and may aid in the engineering of fibrin‐specific variants of SK for breaking down blood clots with potentially higher efficacy and safety.

Streptokinase (SK) is a plasminogen activator which converts inactive plasminogen (Pg) to active plasmin (Pm), which cleaves fibrin clots. SK secreted by groups A, C, and G Streptococcus (SKA/SKC/SKG) is composed of three domains: SKa, SKb and SKc. Previous domain-swapping studies between SK1/SK2b-cluster variants revealed that SKb plays a major role in the activation of human Pg. Here, we carried out domainswapping between skcg-SK/SK2-cluster variants to determine the involvement of SKb in several SK functionalities, including specific/proteolytic activity kinetics, fibrinogen-bound Pg activation and a 2 -antiplasmin resistance. Our results indicate that SKb has a minor to determining role in these diverse functionalities for skcg-SK and SK2b variants, which might potentially be accompanied by few critical residues acting as hot spots. Our findings enhance our understanding of the roles of SKb and hot spots in different functional characteristics of SK clusters and may aid in the engineering of fibrin-specific variants of SK for breaking down blood clots with potentially higher efficacy and safety. Streptokinase (SK), a plasminogen activator (PA) secreted by groups A, C and G streptococci (GAS, GCS and GGS, respectively), converts the inactive plasminogen (Pg) to the active plasmin (Pm) which cleaves the fibrin clots. Despite being considered as a virulence factor (especially in GAS pathogenesis), traditionally, a nonfibrin-specific SK, isolated from the less virulent GCS (H46A or ATCC9542), was widely used as a fibrinolytic drug [1,2]. PA activity of SK is accomplished in two pathways. First, it binds to Pg and forms a 1 : 1 binary SK-Pg* activator (amidolytic) complex, which converts the free Pg substrate to Pm (nonfibrin-specific pathway I). Subsequently, the generated Pm binds SK to form SK-Pm proteolytic activator complex which converts Pg molecules to Pm (fibrin-specific pathway II) [2,3].
The 414 amino acid, SK, is composed of three distinct structural domains: a, b and c spanning residues: 1-146, 147-290 and 291-414, respectively. Protein engineering studies indicated the importance of all three domains and the potential role of several critical amino acids (hot spots), such as Ile1 [3], Lys256, Lys257 [4] and recently Ile33, Asn228 and Phe287 [5], for SK functionality. The attained information was used to improve the fibrinolytic characteristics of SK for enhanced PA potency, fibrin specificity and resistance to the inhibitory effect of plasma a 2 -antiplasmin (a 2 -AP). Concurrently, heterogeneity of SKs at the gene (sk) and protein (SK) levels in different strains (even the same group) of streptococci (specifically for GAS) and its relation to functional differences was shown [2]. Studies indicated the highest sequence diversity of b-domain compared to a and c, particularly in a distinct hypervariable region (sk-V 1 ; residues 147-218). Accordingly, the sk-V 1 was suggested as the main source of sk allelic variations, and consequently, phylogenetic analysis of the sk-V1 nucleotide sequences was used to classify the GAS-SK (ska) alleles into two main clusters; SK1 and SK2, in which SK2 was further subdivided into subclusters SK2a and SK2b [6,7]. These clusters successfully classified GAS strains into those that contain (a) a Pg/Pm direct binding M-like protein, 'PAM' and usually induce invasive skin infections (SK2b), (b) a fibrinogen (Fg) binding M1 protein that does not directly interact with Pg and usually induce upper respiratory tracts (UTR) infections (SK2a) and (c) a M protein that does not interact with either Pg or Fg (SK1) and optimally activates Pg in solution. Although presence of Fg generally enhances the PA activity of all SK types, the specific characteristic of SK1 for optimal Pg activation in solution is in contrast to the SK2 groups (specially SK2b) which strictly require the presence of Fg to display PA activity [6-9]. Interestingly, GCS/GGS-SKs (skcg), despite expressing Pg-binding proteins different from PAM and other GAS-M proteins and displaying high Pg activation in solution (similar to SK1), are clustered in SK2a section of the phylogenetic tree [6]. Indeed, most of the ska alleles in SK2a cluster are homologous to skcg than SK1 or SK2b. Moreover, complexes of Pm with SK2a and skcg-SK display higher resistance to inhibition by a 2 -AP than SK1 or SK2b [1]. Therefore, skcg-SKs display some characteristics specific for either SK1 or SK2a clusters and are thus interesting candidates for comparative studies.
Attempts to address the role of b-domain heterogeneity for functional characteristics of SK clusters/ subclusters started with a study on exchange and swapping the major polymorphic regions between SK1 b-domain (SK1b) and SK2ab [10]. However, apparently due to the similar PA potencies of the used SK1 and SK2a variants, this study failed to uncover any effect on Pg activation kinetics of the chimeric and parental SKs. Recently, two other studies addressing the domain-exchange strategies between a SK1 (with high PA activation rate) and SK2b indicated the major role of the b-domain in the PA activity, which might be further assisted by a-domain [11]. But how bdomain exchange might alter the kinetics of the amidolytic/proteolytic pathways, Fg-bound-Pg activation or the resistance to inhibition by a 2 -AP, specially between skcg-SK and SK2b, are other concerns that never addressed. Recently, SK from a newly isolated GGS (SKG88) with high PA activity was introduced [12]. In the present report, using SKG88 and two other SKs belonging to SK2a and SK2b and employing domain-exchange approaches, these concerns are addressed.

Materials and methods
Bacterial strains and reassessment of the SK clusters

Construction of the parental and b-domainexchanged SK-encoding plasmids
The detailed steps for cloning of sk into pET26b vector to construct the parental plasmids (pET26b-SK G88 , pET26b-SK ALAB49 and pET26b-SK STAB902 ) are illustrated in Fig. S2.
For construction of the b-domain-exchanged SKs, the region corresponding to nucleotides 375-699 (residues 125-233) from the parental vectors was digested by BstEII/ BsiWI restriction enzymes and the digested fragments (327 bp) were cross ligated between SK G88 and two other SKs (SK ALAB49 and SK STAB902 ) (Fig. 1B). All the molecular methods were based on the standard protocols [15].

Expression, purification and characterization of SK proteins
Escherichia coli Rosetta (Novagen, USA) was used for protein expression via IPTG induction, and expressed SKs were purified under native conditions using nickel-nitriloacetic acid (Ni-NTA) affinity chromatography (Qiagen, USA) according to the manufacturer's protocols (QIAexpressionist 2002; Qiagen). Protein concentrations were determined by Bradford assay. Expression and the purity of the purified SKs were assessed by 12% (w/v) SDS/ PAGE and confirmed by western blotting. Protein characterizations assays are described in Figs S5 and S6.

Determination of SK-specific activity (SA*)
For evaluation of the SA* in the presence/absence of Fg, the standard colorimetric assay using the chromogenic substrate (S-2251; Sigma, USA) was used throughout this study, as previously described [7]. The detailed procedure for the assay, construction of the calibration curve and calculation of the SA* are provided in Fig. S7, Figs 2 and 3.

Determination of kinetic constants for amidolytic and proteolytic activities
For analysing amidolytic kinetics, first stoichiometric concentrations of Pg and SK (5.5 µM SK and 5 µM Pg) were mixed and incubated for 5 min to produce the SK-Pg* activator complex. Subsequently, an aliquot of the complex (100 nM) was transferred to the assay buffer along with various concentrations of S2251 (0.1-1.5 mM) in a total volume of 100 µL [12].
For analysing proteolytic kinetics, 100 nM of SK was added to assay buffer containing, '0.1 mM S2251 and varying concentrations of Pg (0.3-5.0 µM)' and changes in absorbance at 405 nm were monitored for 30 min. The data were plotted as velocity/substrate concentration, and kinetic parameters of Pg activation were determined from Michaelis-Menten (V vs S) and inverse (1/V vs 1/S) Lineweaver-Burk plots using GRAPHPAD PRISM 6 (GraphPad Software, La Jolla, CA, USA) [12].

Inhibition by a 2 -antiplasmin
Stoichiometric complexes of SK-Pm (400 nM SK and 200 nM Pm) were incubated for 5 min. The complex was diluted to 20 nM in assay buffer containing a 2 -AP (final concentration: 100-400 nM). The mixtures were incubated for 15 min, then S-2251 (500 µM) was added to the reaction, and residual activity of complex was measured by change in absorbance at 405 nm [1].

Statistical analyses
Unpaired, two-tailed Student's t test with 95% confidence intervals was used for analysis of SK-PA activities and kinetics using SPSS software version 22.0 (SPSS Inc., USA). All linear regressions were by GRAPHPAD PRISM 6, and P-values < 0.05 were considered significant.

Contribution of the SKb in specific activity (SA*)
For calculation of the SA*, the time-course activity profiles (the change in absorbance at 405 nm as a function of time) were measured ( Fig. 2A,B), and subsequently, SA* was calculated (Fig. 2C,D and Table 1), from the slope of the linear portion of the curve (Fig. 3) in which serial dilutions of commercial/standard SK (Streptase) were used as the reference for calibration (Fig. S7). As shown in Table 1, the SA* of SK2a G88 (760.82 9 10 3 IUÁmg À1 ) was about 28-fold and 22-fold higher than that of SK2b ALAB49 (26.64 9 10 3 IUÁmg À1 ) and SK2a STAB902 (36.50 9 10 3 IUÁmg À1 ), respectively. Prior studies reported over 10-fold higher PA activity for SK1 compared to SK2b [8,11] which further supports the similarity of SK1 and skcg-SK (SK2a G88 ) for optimal PA activity in solution [11]. Indeed, SK1b is the most divergent among all SK clusters, and the divergence between SK1 and SK2b might even exceed 40% [6,7], which might further support the determining role of b-domain for functional characteristics between SK1 and SK2b clusters [11,16]. But sequence alignments ( Fig. S8 and Table 3) indicated that the exchanged b-domains between SK2a G88 and SK2b ALAB49 (intracluster; Fig. 1B) were 89% similar, while aand c-domains exhibited 82% and 86% similarity, respectively. Therefore aand c-domains might have more contribution in functional characteristics of the skcg/SK2b domain-exchanged SKs in our study (SK C1 /SK C2 ) than that of SK1/SK2b in the prior report [11]. In contrast, exchanging the SK2a G88 b and SK2a STAB902 b, for making the two intra-subcluster constructs (SK C3 ; 2a G88 -2a STAB -2a G88 and SK C4 ; 2a STAB -2a G88 -2a STAB ; Fig. 1B) led to less alterations in the SA* values for SK C3 /SK C4 compared to SK2a G88 and SK2b ALAB49 (Fig. 2D and Table 1). Thus, our results, consistent with a prior study on SK1b and SK2ab exchanged domains, could not uncover any major effects on PA potencies [10]. However, in the prior study, despite sharing less than 50% identity between exchanged SK1b  Fig. 2A,B) were used. Serial dilutions of Streptase â (CSL) were used as reference for preparation of the standard curve (Fig. S7) and calibration of international unitsÁmg À1 protein (specific activity) in the samples. SK2a G88 , SK2b ALAB49 and SK2a STAB902 are parental constructs. SK C1 (2a G88 -2b ALAB -2a G88 ) and SK C2 (2b ALAB -2a G88 -2b ALAB ) denote the intracluster chimeras. SK C3 (2a G88  and SK2ab, the parental SKs had relatively similar SA* [10], while despite clustering as SK2a, the SK2a G88 and SK2a STAB902 in our study show highly different SA* (Table 1). Indeed, the exchanged b-domains (residues 128-233) of SK2a G88 and SK2a STAB902 were around 97% identical (corresponding to only three residue substitutions out of 108; K138S, I151V, E161K; Table 3) while their aand c-domains exhibited 85% and 88% similarity, respectively (Fig. S8). Therefore, it might be the presence of only few scattered residues acting as hot spots rather than accumulated altered residues in a specific domain that counts for the highly different SA* activities, as recently claimed [5]. Having shown that b-domain exchange between SK2a intra-subclusters (SK C3 /SK C4 ) had little contribution to SA*, the rest of the experiments were only performed for SK C1 /SK C2 .

Contribution of the SKb in the kinetics of amidolytic/proteolytic activity
Amidolytic/proteolytic activity of the SKs was studied by measuring the steady-state kinetic constants of the S2251 hydrolysis including substrate affinity (K m ), catalytic activity (K cat ) and the constant of catalytic efficiency (K cat /K m ; efficiency of the Pg conversion into Pm). As shown in Table 2, K m and K cat values did not alter significantly between SK C1 and SK2a G88 (0.39 mM and 1.39 s À1 vs 0.41 mM and 1.39 s À1 , respectively) leading to almost similar catalytic efficiency (K cat /K m ).
For SK C2 compared to SK2b ALAB49 , the K cat raised by 2.5% (0.88 vs 0.86 s À1 ) and the K m reduced by 17% (0.34 vs 0.41 mM) leading to an overall 24% increase in catalytic efficiency (2.59 9 10 3 vs 2.10 9 10 3 s À1 ÁM À1 ) ( Table 2). Interestingly, evaluation of the kinetic parameters for proteolytic activity indicated that the catalytic efficiency of SK C1 declined by 47% compared to SK2a G88 (229.27 9 10 3 vs 428.57 9 10 3 s À1 ÁM À1 ), which was mainly due to threefold increase in K m value (0.77 vs 2.05 µM). For SK C2 compared to SK2b ALAB49 , the K m declined by 57% (3.45 vs 7.92 µM) and the K cat increased by 40% (0.22 vs 0.16 s À1 ) leading to more than threefold augmented values for catalytic efficiency (63.77 9 10 3 vs 20.20 9 10 3 s À1 ÁM À1 ). These results indicated the determining role of the skcgb (SK2a G88 b) on enhancement of proteolytic activity (Table 2), mainly due to the augmentation of the K m values (increased substrate affinity) which is in accordance with the SA* results (Table 1). Our results are consistent with a prior report on the importance of the SKb for strong binding of Pg substrate to the SK-Pm proteolytic complex and its efficient conversion to Pm [17].

Contribution of the SKb on Fg-bound-Pg activation
The Pg activation rate of various SKs in the presence/ absence of Fg was measured by monitoring the absorbance at 450 nm and calculated by linear regression  SK C1 , SK C2 , SK2b ALAB49 ) raised significantly (2.00, 2.00, 0.83, 0.92 9 10 À2 vs 0.57, 0.19, 0.14, 0.10 9 10 À2 , respectively), but the effect is more  reflective for SK2bb containing constructs (SK C1/ SK2b ALAB49 ) than SK2a G88 /SK C2 (10.80/9.20 vs 3.51/5.93-fold enhancement of activation rates, respectively). . These results are consistent with prior reports for higher influence of Fg on enhancement of SK2b activity compared to SK1 [7,8] and that of the SK2a compared to a skc-SK [9], indicating more similarity of the skcg (SK2a G88 ) to SK1 variants for activation of Fg-bound-Pm. As shown in Fig. 4A and consistent with a recent study [12], SK2a G88 showed high intrinsic Fg-bound-Pg activation. This higher Fg-bound-Pg activation (twofold higher than SK2b ALAB49 ; 2.00 vs 0.92) is completely retained in SK C1, while in the absence of Fg, PA potency of SK C1 is three times lower than the parental SK2a G88 (0.19 vs 0.57). Of note, these characteristics of SK C1 might be of interest for development of a fibrin-specific version of SK for targeted fibrinolysis [3]. Interestingly, SK C2 retained the same (and low) activity as SK2b ALAB49 in the absence of Fg, while its Fg-bound-Pg activation showed 42% (0.83 vs 2.00) and 90% (0.83 vs 0.92) of the parental activity (SK2a G88 and SK2b ALAB49 , respectively). Collectively, while these results support the major contribution of the SKb for the Fg-bound-Pg, but in agreement with prior reports also implies the potential contribution of other domains for this characteristic [3].

Contribution of the SKb on resistance to a 2 -AP inhibition
As shown in Fig. 4B, all four SK-Pm complexes resisted the inhibitory effect of a 2 -AP (400 nM) by retaining more than 50% of their activity. This observation is consistent with the long-known phenomena for resistance of the SK-Pm complex to the major physiological plasmin inhibitor 'a 2 -AP' [18]. However, for SK2a G88 and SK C2 (containing SK2a G88 b), retained activity was about 80%, while for that of SK2b ALAB49 and SK C1 (containing SK2b ALAB49. b), it was about 50% (Fig. 4B). Thus, our results indicated that skcg-SK was more resistant to inhibition by a 2 -AP than SK2b, which is in agreement with recent findings for Pm-complexed with either SK-H46A (skc) or SK2a variants [1]. Therefore, our results indicated the resemblance of skcg-SK to SK2a variants for 'a 2 -AP resistance' and the determining role of SKb in these characteristic. Although this finding is consistent with prior suggestion on the contribution of SKb to the interaction of SK with inhibitors [19], the role of other domains, specially residue 1-59 of a-domain for resistance to a 2 -AP, was also suggested [3].

Potential contribution of the substituted residues (hot spots) in SK functionalities
As emphasized earlier, a recent study on a new isolate of skcg-SK (GGS-132) indicated that presence of only three altered residues (Ile33Phe, Asn228Lys and Phe287Ile) that probably acted in a synergic mode as hot spots might induce enhanced proteolytic/Fgbound-Pg activation compared to a SKC (GCS-SKC9542) [5,12]. Accordingly, it was also shown that SK2a G88, in the present study, exhibited enhanced proteolytic/Fg-bound-Pg activation compared to the same SKC, while only seven residues scattered within domains of the two SKs were substituted (98% similarity) [12]. As shown in Table 3, the exchanged segments between SK2a G88 b and SK2b ALAB49 b differ by 12 residues. Among these altered residues, 'V160I, E161R and K209E', consistent with a recent report [17], might have potentially acted as hot spots for the induced functionalities of the domain-swapped SKs (SK C1 /SK C2 ). The precise insight on the effect of these substitutions might be gained via site-directed mutagenesis experiments.
In conclusion, to the best of our knowledge, we reported the first domain-exchange study for skcg and cluster 2-ska alleles to elucidate the contribution of SKb for a broad range of functional characteristic including kinetics of specific/ proteolytic activity, fibrinogen-bound Pg activation and a 2 -antiplasmin resistance. Results pointed to the 'minor to determining' contribution of SKb in these functionalities which might be potentially accompanied by a few critical residues acting as hot spots. Our findings indicated the (a) similarity of coclustered, skcgb and SK2ab Table 3. The altered residues in the exchanged SKb domain of SK2a G88 compared to SK2a STAB902 and SK2b ALAB49 . Conserved (identical) residues are indicated by dots. Residue position   132  134  138  151  153  154  160  161  176  178  209  210 213 variants (only three residues alteration) and minor contribution of their SKb for highly different SA* between these two alleles; (b) similarity of skcg to cluster 1-ska alleles (SK1) for optimal PA activity in solution and activation of Fg-bound-Pm compared to that of the SK2 variants and major contribution of SKb in this characteristic; (c) major role of the SKb on enhancement of proteolytic activity between skcg-SK and SK2b that is mainly due to the augmentation of the K m values (increased substrate affinity); and (d) similarity of skcg-SK to SK2a variants for 'a 2 -AP resistance' and the determining role of SKb in this characteristics. These findings might assist in better understanding of the roles displayed by SKb and hot spots for different functional characteristic of SK clusters and engineering fibrin-specific versions of SK.

Supporting information
Additional supporting information may be found online in the Supporting Information section at the end of the article. Fig. S1. Multiple DNA Sequence alignment of skb-V1 fragments. Multiple Sequence alignment of skb-V1 fragments (hypervariable region1 of the b-domain)from nucleotide "445 to 655" of sk gene, corresponding to the amino acid residues "147 to 218" of the SK of the strains used in this study (G88, STAB902 and ALAB49; Genbank Accession numbers: HM390000.1, CP007041.1 and AY234134, respectively) together with the wellknown strains that were employed for construction of phylogenetic tree in Fig. 1A  expression was induced at OD 600 of 0.5-0.6 by isopropyl-b-D-thio-galactoside (IPTG) to a final concentration of 1 mM for 3 hours at 37°C. Finally, cells were harvested by centrifugation and stored at -20°C for purification steps. Purification of His-tagged SK proteins from induced E.coli Rosetta cells was performed under native conditions, using nickel-nitriloacetic acid (Ni-NTA) affinity chromatography and according to manufacturer's protocol (QIAexpressionist, 2002, Qiagen company website). Briefly, the cell pellets were resuspended in binding buffer (50 mM NaH 2 PO 4 , 300 mM NaCl, 10 mM imidazole) with 0.5 mg/ml lysozyme at 2-5 ml per gram wet weight. After incubation on ice for 30 min, the cells were disrupted by sonication, and supernatant was collected after centrifugation at 10,000 g for 20-30 min at 4°C. After addition of 1ml resin Ni-NTA to the clear lysate, the mixture was shaken at 4°C for 60 minutes, loaded on column and washed 4 times with 4 ml wash buffer (50 mM NaH 2 PO 4 , 300 mM NaCl, 20 mM imidazole) then 4 times with 0.5 ml elution buffer (50 mM NaH 2 PO 4 , 300 mM NaCl, for expression and purification produced the same results (not shown). Fig. S6. Confirmation of the expressed proteins by western blotting. Western blotting was performed according to the standard protocols [15]. Briefly, proteins were transferred from SDS-PAG to the nitrocellulose membrane and the membrane was blocked by 5% BSA. Mouse anti-His monoclonal antibody (Qiagen, USA) was used as the primary antibody and goat anti-mouse IgG conjugated to HRP (Horse Radish peroxidase) (Qiagen, USA) as the secondary (tracking) antibody. Detection of the bands was by 3, 3diaminobenzidine (DAB) (Qiagen, USA). Western-blot analysis for SK2a G88 and SK C1 proteins are shown. Lane 1; molecular weight marker (SM7012, Cinnagen Co), lanes 2 and 4; crude lysis of E. coli Rosetta cells after induction by IPTG (1mM), expressing SK2a G88 and SK C1 , respectively, lane 3; crude lysis of E. coli Rosetta cells before induction (no band was observed). Lanes are spliced together to remove an intervening lane and the vertical dotted line is at the location of the spliced lanes. The arrow indicates the location of 47 kDa (SK). Fig. S7. Calibration curve for standard SK. Serial dilutions of StreptaseÒ (CSL, Behring, Germany), a commercially available standard SK, were used to prepare the standard calibration curve based on Hydrolysis of S-2251 by Pg, as explained in Fig. 2 and Fig. 3. Fig. S8. Amino acid sequence alignment of SK proteins corresponding to reference strain SK 9542 (S. equisimilis, ATCC9542, the commercial source for production of SK), SK2b ALAB49 , SK2a G88 , and SK2a STAB902 . The alignment was created using MEGA6 software [14]. Conserved (identical) amino acids in the alignment are indicated by dots. The exchanged fragments are highlighted.