Interaction of streptokinase and plasminogen. Studied with truncated streptokinase peptides.

The interaction of streptokinase (SK) with human plasminogen (HPlg) was investigated using truncated SK peptides prepared by gene cloning techniques. SK(16-414) and SK(16-378) could activate HPlg as efficiently as the authentic SK. SK(60-414), which had been preincubated with SK(1-59), could also activate HPlg. SK(91-414), SK(127-414), and SK(158-414), at a concentration of one-tenth of HPlg, all failed to activate HPlg. However, the truncated SK peptides in complexes with equimolar HPlg could form amidolytically active virgin enzymes that slowly converted to human plasmin (HPlm) after a lag period of 15 min. SK(16-316) could not activate HPlg. No virgin enzyme was detected when SK(16-316) was incubated with equimolar HPlg, but the HPlg in the complex was modified to HPlm after reaction for 20 min. SK(220-414) and SK(16-251) had no ability to transform HPlg to virgin enzyme or to HPlm in equimolar complex with HPlg, although they could bind to HPlg. The functions of five regions in the SK molecule (a, Ile1-Lys59; b, Ser60-Asn90; c, Val158-Arg219; d, Tyr252-Ala316; e, Ser317-Ala378) in interaction with HPlg are deduced. Region a is important in stabilizing the conformation of the SK molecule, and region b is essential for HPlg activation. Region c is required for induction of the conformational changes of HPlg to virgin enzyme. Regions c and d are required for the conversion of HPlg to HPlm in the HPlg.SK equimolar complex. Coordination of regions c, d, and e of SK is essential for a virgin enzyme formation, and coordination of regions b, c, d and e is required for an effective SK-type HPlg activator.

Streptokinase (SK) 1 is a single-peptide secretory protein of 414 (1) or 415 (2) amino acid residues produced by various strains of ␤-hemolytic Streptococcus (1)(2)(3). The SK and human plasminogen (HPlg) can form an equimolar activator complex that catalyzes the conversion of plasminogens (Plgs) to plasmins (Plms) of different mammalian species (4 -11). Plm is a potent protease that in turn catalyzes the hydrolysis of fibrin, which causes the dissolution of blood clots. SK, therefore, is used as a thrombolytic agent clinically to relieve thromboembolic blockages in blood vessels, such as acute myocardial infarction.
Studies of the activation of HPlg by SK suggest that more than one intermediate step is involved in the entire process. In the first step, a 1:1 stoichiometric complex of HPlg and SK is formed, and the conformation of the catalytic domain of HPlg is altered to expose its enzyme-active center (6,7,9). The HPlg moiety in the complex has been named virgin enzyme since it has similar catalytic activity to human plasmin (HPlm), but the activating peptide bond of Arg 560 -Val 561 is not cleaved (12). In the second reaction step, the HPlg⅐SK complex is converted to HPlm⅐SK, and then in the final reaction step, the HPlm⅐SK catalyzes the hydrolysis of the specific activating peptide bond of Arg 560 -Val 561 on substrate HPlg, resulting in the formation of HPlm (5,8,(13)(14)(15).
The exact interaction sites of SK with HPlg and their functions, however, have not been determined. Proteolytic SK fragments obtained in the reactions of SK with human, rabbit, and dog Plg(m)s have been used to study the functions of the SK molecule (15)(16)(17)(18). A 36 and a 25.7-kDa fragment obtained in the reaction with HPlm and dog Plg, respectively, can activate HPlg (15)(16)(17). A 17-kDa SK fragment consisting of Val 143 (Glu 148 ) to Arg 289 (Lys 293 ) obtained in the reaction with Sepharose-immobilized HPlg (HPlm) is the smallest SK fragment that can bind to HPlg (18). A SK fragment, SK-o, Ser 60 -Lys 333 (or Ser 60 -Lys 332 according to the numbering used in this paper), is essential for minimal SK activator activity (19). Obviously, the COOH-terminal peptide of SK-p, Ala 334 -Lys 387 (or Ala 333 -Lys 386 ), is required for strong binding with HPlm. The NH 2 -terminal 59-amino acid peptide is important in maintaining the proper conformation of SK for full activator activity (19). These studies imply that domain-like structures may exist in the SK molecule that exert various functions in HPlg activation. This study was undertaken to elucidate the functions of various domains of SK and to determine more precisely the interaction sites between SK and HPlg using a series of truncated SK peptides lacking NH 2 and/or COOH-terminal amino acid residues.

EXPERIMENTAL PROCEDURES
Materials-Enzymes used in DNA manipulation were purchased from Boehringer Mannheim, Life Technologies, Inc., or Promega and were used according to the Cold Spring Harbor Manual or the recommendations of the suppliers. The two oligonucleotides used in SK gene amplification by polymerase chain reaction (PCR) were custom synthesized by Pan Asia Hospital Supply Co. (Taiwan, Republic of China). The oligonucleotide for sense primer was 5Ј-GGAGGGATCCATGAAAAAT-TACTTATCT-3Ј (nucleotide position 809 -836 by the numbering convention for the PstI fragment containing SK) and for antisense primer was 5Ј-AAGAGGATCCTTATTTGTCGTTAGGGTT-3Ј (nucleotide position 2151-2124) (1). For cloning convenience, BamHI recognition sequences (underlined) were created by replacing some of the nucleotides. Blue-Sepharose CL 6B, Lys-Sepharose, and DEAE-cellulose were from Pharmacia Biotech Inc. NH 2 -D-Val-Leu-Lys-p-nitroanilide (S-2251) was obtained from Sigma. SK antibody was prepared in our laboratory from mice. All other chemicals were of the highest grade commercially available.
Proteins and Enzymes-HPlg was prepared from pooled human plasma by a modification of the Deutsch and Mertz method (20). Forms 1 and 2 of native HPlg were separated by chromatography on Lys-Sepharose column (21). Form 2 of HPlg was used throughout this study. HPlm was prepared by activating HPlg with Sepharose-bound urokinase as described previously (22). Native SK (Behringwerke AG, Marburg, Germany) was further purified by passing it through a Blue-Sepharose CL 6B column to remove serum albumin (15).
Construction of SK and Truncated SK Genes-The SK gene was amplified by PCR by the standard procedure from Streptococcus equisimilis H46A (ATCC 12449). The 1.3-kilobase DNA fragment was then cloned into the BamHI site of the multicloning region of pGEM-3Z vector (Promega) and propagated in Escherichia coli JM109. Unidirectional deletion of the SK gene was carried out with exonuclease III using the Erase-a-Base system (Promega) to construct truncated SK genes from the NH 2 or COOH terminus. Nucleotide sequences of the SK and truncated genes were determined by dideoxy sequencing method (23,24).
Expression and Purification of Recombinant Truncated SK Peptides-Truncated SK genes were subcloned in-frame into the BamHI site of the overproducing plasmid pET-3 (Novagen), in which the cloned genes were inducibly expressed under the control of T7 promoter in E. coli strain BL21(DE3)pLysS. Bacterial cells were grown to mid-log phase at 37°C, and target gene expression was induced after shifting the incubation temperature to 33°C and adding 1 mM isopropyl-1-thio-␤-D-galactopyranoside. 3 h later, cells were harvested, washed, and disintegrated by French press, then the proteins were concentrated by ammonium sulfate precipitation. The precipitated proteins were dialyzed and loaded onto a DEAE cellulose column. After elution with salt gradient, the active fractions were detected by HPlg activation with S-2251. Because the HPlg activator activities of some short peptides of SK were too low to detect, the fractions containing SK peptides were also analyzed by SDS-PAGE and Western blotting. Quantitation of the activities of truncated SK peptides in each purification step was conducted in flat bottomed 96-well plates by activating HPlg to hydrolyze S-2251. A serially diluted native SK (93,750 IU/mg) solution was used as standard.
Protein Concentration-Protein concentrations of SK peptides were determined with Folin reagent (Merck) by the Lowry method (25).
Amino Acid Sequence Analysis-The amino acid sequence of the truncated SK peptide was determined by Edman degradation in an Applied Biosystems Sequencer (model 477A).

SDS-PAGE Analysis and Western
Blotting-Protein samples were subjected to SDS-PAGE according to the method of Laemmli (26). The electrophoresed proteins on the gel were transferred onto Immobilon-P transfer membrane (Millipore), and Western blotting was carried out (27).
Steady-state Kinetic Parameters of Activation of HPlg by SK and Truncated SK Peptides-A one-stage assay as described previously was used to measure the kinetics of HPlg activation by SK and truncated SK peptides (28,29). Briefly, HPlg at final concentrations ranging from 0.04 to 4 M was incubated with 0.5 mM S-2251 in an assay cuvette containing 150 l of 0.05 M Tris buffer (pH 7.4) and 0.1 M NaCl. Activation was initiated by adding an SK peptide of fixed concentration, and the change in absorbance at 405 nm was monitored at 37°C in a Hitachi 330 spectrophotometer. The increments of absorbance between 10 and 300 s after addition of SK peptides were used to measure the initial rate of HPlg activation. Initial reaction rates were determined from the slopes of plots of absorbance versus t 2 , and double-reciprocal plots were then constructed. HPlg activation parameters, K plg (the apparent Michaelis constant for the HPlg substrate) and k plg (the catalytic rate constant of activation), were calculated as described by Wohl et al. (28). The ⑀ 1 M at 405 nm employed for p-nitroanilide was 9559.
Discontinuous Assays of Amidolytic Activity for Determination of the Maximal Formation of HPlg⅐SK-Peptide Complexes-Incubations of equimolar HPlg and truncated SK peptides (final concentration, 2 M) were carried out at 25°C in 10 mM HEPES/NaOH, pH 7.4. Aliquots were removed at various intervals to assay the amidolytic activity. The amidolytic activity was measured by adding aliquots of the HPlg⅐SKpeptides (final concentration, 0.2 M) in an assay cuvette containing 0.5 mM S-2251 in 0.05 M Tris buffer (pH 7.4) and 0.1 M NaCl. The absorb-ance at 405 nm was monitored between 10 and 90 s. The initial reaction rate was calculated, and the duration to achieve maximal amidolytic activity was determined as described by Chibber et al. (30).
Continuous Assays of Amidolytic Activity for HPlg⅐SK-Peptide Virgin Enzyme Complexes-Incubations of equimolar HPlg and truncated SK peptides (final concentration, 2 M) at 25°C in 10 mM HEPES/NaOH, pH 7.4, for a time interval determined by the above method to achieve maximal active-site formation. The amidolytic activity was measured by adding aliquots of the HPlg⅐SK-peptides (final concentration, 0.2 M) in an assay cuvette containing 0.5 mM S-2251 in 0.05 M Tris buffer, pH 7.4, and 0.1 M NaCl. The absorbance at 405 nm was monitored.
Amidase Parameters of the HPlg⅐SK-Peptide Virgin Enzyme Complexes-HPlg (final concentration, 2 M) was incubated with equimolar SK peptides in 10 mM HEPES/NaOH, pH 7.4, at 25°C for 4 -7 min to achieve the maximal active-site formation. The amidolytic activity of the stoichiometric complex (final concentration, 0.2 M) was measured with various concentrations of S-2251 ranging from 0.2 to 4 K m , at 37°C and in 0.05 M Tris-HCl, pH 7.4, 0.1 M NaCl. The changes in absorbance at 405 nm were monitored. The initial rate and substrate concentration data were analyzed on a Lineweaver-Burk plot.
Binding of Truncated SK Peptides to HPlg-SK or truncated SK peptide was labeled with 125 I by the method described previously (31,32). The binding analysis was carried out by coating a radioimmunoassay strip plate (Costar) with 0.2 ml of HPlg (2 mg/ml) in 0.05 M carbonate/bicarbonate buffer, pH 9.6 for 36 h at 4°C, and blocked with 1% bovine serum albumin in phosphate-buffered saline (PBS). After washing, 125 I-labeled SK or truncated SK peptide at increasing concentrations was added to the wells and incubated at 4°C for 1 h. The unbound peptide was thoroughly washed out with PBS containing 0.05% Tween 20, and the radioactivity was determined with a LKB ␥-counter. Nonspecific binding of the tested peptide was measured in the presence of a 30-fold excess of each unlabeled peptide and was subtracted from the total binding.

RESULTS
The SK gene of 1.3 kilobases was amplified by PCR with standard procedures from S. equisimilis H46A and was constructed into pGEM-3Z and pET3 plasmid. Unidirectional deletion of SK gene from either end was carried out with exonuclease III. Nine truncated SKs were prepared and designated SK (16 - (1,33). In sequencing the PCR-amplified full-length SK DNA, one silent mutation at Lys 61 (AAA 3 AAG) and one mutation at Arg 401 to Pro (CGT 3 CCT) were detected as compared with the SK gene sequences (1). The slightly modified SK DNA was used for preparation of truncated SK peptides without correction of these mutations. Apparently the alteration had no effect on the HPlg activator activity of recombinant SK since purified SK(16 -414) had a specific activity of 118,755 IU/mg (Table I), which is comparable with purified commercially available native SK. The expressed recombinant SK and its truncated peptides were also sequenced and found to be identical to the published sequence (1, 2) except that an additional fusion peptide of 14 -18 amino acid residues derived from the pET-3 plasmid was attached at their NH 2 termini. The SK(16 -414) DNA was used to prepare COOH-truncated SK, since SK(16 -414) protein had the same HPlg activator activity as native SK, and the SK peptides without NH 2 -terminal 15 amino acids were more easily overexpressed in the E. coli cells than the full-length SK. In general the NH 2 -and COOH-truncated SK peptides overexpressed in the E. coli system consisted of more than 70% of the total amount of proteins in the crude extracts of the E. coli cells. Homogeneous SK peptides were obtained after ammonium sulfate precipitation and DEAE column chromatography as shown in SDS-PAGE (Fig. 1). All of the SK peptides could be detected by a polyclonal antiserum raised against the native SK. The recovery and specific activities of SK peptides during each purification step are summarized in Table I. The specific HPlg activator activities of purified SK(16 -414) and SK(16 -378) were higher than 100,000 IU/mg and were comparable with that of the purified commercial native SK. SK(60 -414), SK(91-414), SK(127-414), SK(158 -414), and SK(16 -316) had low specific HPlg activator activities ( Table I). The HPlg activator activities of the two short peptides, SK(16 -251) and SK(220 -414), were too low to detect.
SK (16 -414) and SK(16 -378) at a catalytic concentration (0.02 M) could effectively catalyze the conversion of HPlg (2 M) to HPlm, and an increasing rate of substrate hydrolysis was observed as more HPlm was produced ( Fig. 2A). The conversion of HPlg to HPlm was completed in 10 min as confirmed by SDS-PAGE. The bands of HPlg (94 kDa) disappeared, and both the heavy (65 kDa) and light (26 kDa) chains were detected (Fig. 2B). The second-order rate constants of HPlg activation, k plg /K plg , of native SK, SK (16 -414), and SK (16 -378) were similar (Table II). SK(60 -414) had low HPlg activator activity (Table I). However, its activator activity could be dosedependently enhanced by incubating with a catalytic amount of SK(1-59), which was prepared by limited digestion of native SK with immobilized HPlm and purified by HPLC ( Fig. 2A). After 10 min of incubation, about half of HPlg was activated by SK(1-59) modified SK(60 -414)*, while complete activation occurred after 30 min of incubation (Fig. 2B). This result suggests that SK (1-59)    with HPlg in a one-to-one equimolar complex. The equimolar HPlg and SK(91-414) had a maximal amidolytic activity after reaction for 7 min (Fig. 3A). However, no cleavage of the activating peptide bond, Arg 560 -Val 561 , of HPlg was observed at up to 13 min determined at intervals as shown by SDS-PAGE analysis (Fig. 3B), suggesting that a virgin enzyme, which consisted of intact HPlg and SK(91-414) equimolar complex, was induced. However, after reaction for more than 15 min, some HPlg was hydrolyzed, heavy and light chains of HPlm were detected, and SK(91-414) was degraded (Fig. 3B, lanes e  and f). The cleavage of the activating peptide bond was also confirmed by NH 2 -terminal amino acid determination of the HPlm light chain. Virgin enzymes could also be induced in stoichiometric complexes of HPlg and other truncated SK peptides, although different durations of incubation were required to reach maximal amidolytic activities. For example, to achieve the maximal amidolytic activities, it took 4 min for HPlg and SK (16 -414) (Table III). The virgin enzymes of HPlg⅐SK(91-414), HPlg⅐SK(127-414), and HPlg⅐SK(158 -414) had lower reaction constants (Table  III). Initially, no HPlm conversion was observed in each HPlg⅐SK-peptide virgin enzyme complex, since only HPlg and the corresponding SK peptide were detected by SDS-PAGE (Fig. 4B). However, after incubation for 15 min, HPlm formation was also observed in the equimolar complexes of HPlg⅐SK(127-414) and HPlg⅐SK(158 -414) as that of HPlg⅐SK(91-414), suggesting that the activation peptide bond in these complexes was also cleaved.
The COOH-truncated peptide, SK(16 -316) at a concentration one-tenth of HPlg, had no HPlg activator activity. It also failed to induce a virgin enzyme formation in an equimolar stoichiometric complex with HPlg under the same conditions described previously. However, enzyme activity of HPlm con-  verted from HPlg in HPlg⅐SK(16 -316) equimolar complex could be detected after incubation for 20 min and reached its maximum after 30 min (Fig. 5A). Cleavage of the activating peptide of HPlg in the HPlg⅐SK(16 -316) complex was observed after 20 min of incubation as analyzed by SDS-PAGE (Fig. 5B). The SK (16 -316) in the complex was degraded as the HPlm appeared (Fig. 5B). The truncated peptides, SK(220 -414) and SK (16 -251), could neither induce catalytically active virgin enzyme in the stoichiometric complexes with HPlg nor activate HPlg to HPlm in the complexes. However, the isotope-labeled SK(220 -414) and SK(16 -251) could bind to HPlg (Fig. 6).

DISCUSSION
The secondary structures of truncated SK peptides used in this study were determined by circular dichroism (CD) spectroscopy and were all found to be similar to that of the corresponding regions of the native SK moiety reported (34) except for SK(60 -414). This result suggests that the recombinant truncated SK peptides were suitable for the study of structurefunction relationship of SK. SK(60 -414) had a majority of disordered secondary structure according to the CD spectroscopy. It also had a lower HPlg activator activity compared with the SK-p, Ser 60 -Lys 387 (or Ser 60 -Lys 386 according to the numbering used in this study) obtained by limited digestion of native SK (19). The fusion peptide of 14 amino acid residues in the NH 2 -terminal of SK(60 -414) might interrupt the proper conformation of SK(60 -414). However, after incubation with SK(1-59), SK(60 -414) was shifted to SK(60 -414)*, with elevation of HPlg activator activity. This might be due to a conformational change in SK(60 -414) induced by SK(1-59). Therefore, SK(1-59) was important in maintaining the proper conformation of the core region of SK (19).
Since SK (16 -414) and SK (16 -378) were as active as native SK in HPlg activation, the peptides Ile 1 -Asn 15 and Ser 379 -Lys 414 were of little functional importance for SK. The study in which two SK gene products, cSK L , Ile 1 -Ala 384 (or Ile 1 -Ala 383 according to the numbering used in this study), and cSK S , Ile 1 -Leu 383 (or Ile 1 -Leu 382 according to the numbering used in this study), were cloned suggested that Ala 384 (or Ala 383 ) was essential for activator activity of SK (35), since Ile 1 -Ala 384 (or Ile 1 -Ala 383 ) had full activator activity, while Ile 1 -Leu 383 (or Ile 1 -Leu 382 ) retained only 25% activity. However, SK (16 -378) prepared in this study was as active as native SK in HPlg activation and also had a properly folded secondary structure. The low activator activity of cSK S might be due to reasons other than Ala 384 (or Ala 383 ) directly involved in activation of HPlg.
In this report, we defined five important regions in SK molecule as a HPlg activator. These regions were as follows: a, Ile 1 -Lys 59 ; b, Ser 60 -Asn 90 ; c, Val 158 -Arg 219 ; d, Tyr 252 -Ala 316 ; e, Ser 317 -Ala 378 as shown in Scheme I.
Region a functioned to stabilize the conformation of SK in maintaining its full activator activity. SK(60 -414), which contained regions b, c, d, and e, was a competent HPlg activator, although the activation rate was slower than those of SK (16 -414) and SK(16 -378) ( Fig. 2A). SK(91-414), at a molar ratio of one-tenth of HPlg, could not activate free HPlg. However, an equimolar complex of HPlg and SK(91-414) was amidolytically active. The complex of HPlg and SK(127-414) or SK(158 -414) had properties similar to SK(91-414). Therefore, the SK peptide, which contained regions c, d, and e, had the ability to form a so-called virgin enzyme complex with HPlg, and region b was essential for HPlg activation. The conversion of HPlg to HPlm in these one-to-one complexes of SK peptides and HPlg was detected after a lag period of 15 min (Fig. 3B). The reason of this slow HPlg conversion to HPlm in the complexes remains unclear. It is possible that the conformation of HPlg is transformed in the complex and the activating peptide bond is more vulnerable for hydrolysis by the less effective activators. In regard to the formation of virgin enzyme complex with HPlg, SK(220 -414), which contained regions d and e, lost this ability. Therefore, the region c, Val 158 -Arg 219 , was thought to possess one of the essential interaction cores for virgin enzyme formation. The COOH-terminal truncated SK(16 -378) had all the essential regions to form a virgin enzyme complex with HPlg and to catalyze HPlg activation. The SK(16 -316), which consisted of regions a, b, c, and d but not e, could not form virgin enzyme and could not catalyze the activation of HPlg into HPlm as a typical SK-type HPlg activator. However, the HPlg moiety was slowly converted to HPlm in equimolar HPlg⅐SK (16 -  which SK(16 -251) contained only regions a, b, and c. Therefore, region d apparently was involved in the induction of the conformational changes in HPlg, so that it could be activated in the complex.
In conclusion, by studying with the truncated SK peptides, five regions of defined functions in SK molecules were deduced. The functional studies of the truncated SK peptides provided the evidence that more than one region on SK could interact with HPlg. By comparing the functions of the truncated SK, we were able to define the function of each region of SK. These findings were consistent with the results of NMR and CD spectroscopy studies of SK in which a flexible structure of SK with existence of at least three or four domains was proposed (36,37). The elucidation of the function of each of these specific regions of SK molecule is very important for understanding the molecular mechanism of its interaction with HPlg. Further studies on the functions of specific point mutation of SK in different regions might provide more critical information needed for the delineation of the intriguing interaction between SK and HPlg.