NH2-terminal Structural Motifs in Staphylokinase Required for Plasminogen Activation*

Staphylokinase (Sak) forms an inactive 1:1 stoichiometric complex with plasminogen which requires both conversion of plasminogen to plasmin and hydrolysis of the Lys10-Lys11 peptide bond of Sak to become a potent plasminogen activator (Schlott, B., Guhrs, K.-H., Hartmann, M., Rocker, A., and Collen, D. (1997) J. Biol. Chem.272, 6067–6072). Exposure of a positively charged NH2-terminal amino acid after hydrolysis of Sak is a major determinant of the plasminogen-activating potential, but in itself is neither necessary nor sufficient. Here, the structural motifs of the NH2-terminal region Lys11-Gly-Asp-Asp-Ala-Ser16-Tyr-Phe-Glu of processed Sak, required for plasminogen activating potential, were studied by deletion and substitution mutagenesis. Expression in Escherichia coli of variants with deletion of 11, 14, 15, or 16 NH2-terminal amino acids yielded correctly processed but inactive molecules. Expression of their homologues with the NH2-terminal amino acid substituted with Lys-generated derivatives from which the NH2-terminal initiation Met was no longer removed, yielding inactive (≤ 10%) Sak42DΔN11(M),G12K, active (>50%) Sak42DΔN14(M),A15K and Sak42DΔN15(M),S16K, and inactive Sak42DΔN16(M),Y17K. Lys variants without NH2-terminal Met, generated from fusion proteins in which a His6 tag and a factor Xa recognition sequence were linked to the NH2 terminus of the Sak variants, were indistinguishable from their NH2-terminal Met-containing counterparts. All variants studied had intact affinities for plasminogen as measured by biospecific interaction analysis. The activity of Sak42DΔN11(M),G12K could be restored by additional substitution of both Asp13 and Asp14 with Asn, yielding active Sak42DΔN11(M),G12K, D13N, D14N, whereas substitution in Sak42DΔN16(M),Y17K of Phe18 and Glu19 with Asn yielded inactive Sak42DΔN16(M),Y17K,F18N,E19N. These data, in combination with the recent finding that the 20 NH2-terminal amino acids of Sak lack secondary structure, suggest that the NH2-terminal region of Sak is not required for binding to plasmin/plasminogen, but that a positively charged amino acid in the ultimate or penultimate NH2-terminal position corresponding to amino acids 11–16 of this flexible region participates in the reconfiguration of the active site of the plasmin molecule to endow it with plasminogen-activating potential.

Staphylokinase, a protein secreted by certain Staphylococcus aureus strains, is known to be able to dissolve fibrin clots since the 1940s (1). Following cloning of the gene and isolation of large amounts of recombinant protein (2), its therapeutic potential for thrombolysis was recently demonstrated in patients with acute myocardial infarction (3) or with peripheral arterial occlusion (4). The sak gene encodes a mature protein of 136 amino acids in a single polypeptide chain without disulfide bridges. Three natural Sak 1 variants differ at amino acid positions 34, 36, and 43 (5-7) and in thermal stability (8) but have very similar kinetic parameters of plasminogen activation and fibrin-dissolving potency (2).
Sak is not an enzyme, but a cofactor that initially forms an inactive 1:1 stoichiometric complex with plasminogen, which requires both conversion to Sak-plasmin to expose an active site in plasmin and NH 2 -terminal processing of Sak to reconfigure this active site for the specific cleavage of single chain plasminogen into active two-chain plasmin (9). The exposure of a positively charged amino acid at the new NH 2 terminus after cleavage of the Lys 10 -Lys 11 (P 1 -P 1 Ј) peptide bond is a critical step to unveil the plasminogen-activating potential of Sak, but additional structural motifs may be necessary to endow the Sak molecule with plasminogen-activating potential (9).
The tertiary structure of Sak was recently elucidated by crystallography and x-ray diffraction analysis (10) and the secondary structure independently by NMR (11). In both studies, no secondary structure could be assigned to the NH 2terminal 20 amino acids, which suggests that the NH 2 -terminal region may be rather flexible.
In the present study, the NH 2 -terminal structural elements of staphylokinase required to confer plasminogen-activator specificity to its complex with plasmin were studied by deletion and substitution of amino acids extending from Gly 12 to Glu 19 . These deletion mutants lack plasminogen-activating potential but (some of them) can be functionally rescued by substituting their NH 2 -terminal amino acid with a positively charged Lys, notwithstanding the fact that the Escherichia coli expression system no longer removed the NH 2 -terminal initiation Met from the Lyscontaining variants. The shortest variant with intact plasminogen-activating potential was found to be Sak42D⌬N15(M),S16K. * This work was supported by Bundes Ministerium fü r Bildung und Forschung Grant 0311015. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

EXPERIMENTAL PROCEDURES
Materials-Enzymes and reagents for gene construction experiments were purchased from Boehringer Mannheim (Germany), sequenase was from Amersham (Amersham Buchler, Braunschweig, Germany), and Ni-NTA resin from Quiagene (Hilden, Germany). Factor Xa protease was supplied by New England Biolabs (Schwalbach, Germany). Other reagents were of the highest quality commonly available. Oligonucleotides were synthesized using a DNA/RNA-synthesizer model 394 (Applied Biosystems, Foster City, CA). Semiphosphorylated linker cassettes were obtained by chemical phosphorylation of the respective oligonucleotides at the end of the synthesis (Glen Research, Sterling, VA). Native human plasminogen was purified from human plasma according to Deutsch and Mertz (12). The synthetic plasmin substrate S-2251 was purchased from Chromogenix (Essen, Germany).

Construction of Expression Plasmids and Purification of Sak42D
Variants-The variants of Sak42D (6) were initially derived from the expression plasmid pMEX602Sak42D (2). The expression plasmids for variants Sak42D⌬N10(M) and Sak42D⌬N11 and Sak42D⌬N14 were described previously (9) as was the strategy for construction of the expression plasmids for the other deletion mutants (2,9,13). Primers and templates used for the generation of the PCR fragments are listed in Table I. For all PCR reactions the same "right"-flanking primer (r-primer) was used (13). The PCR reactions with primers d5 to d7 were carried out with template DNA pMEXSak42D⌬N14(M),A15K. The PCR fragments were trimmed at the 5Ј-end by NdeI and at the 3Ј-end by HindIII, respectively, and cloned into the unique NdeI and HindIII sites of pMEX602Sak42D replacing the wild type sak42D gene as described elsewhere (14). The plasmids were preselected by digestion with MluI or SfuI and the desired mutants identified by DNA sequencing. Expression and purification of Sak42D proteins and the preparation of enzymatically processed Sak42D (Sak42D proc ) was achieved as described previously (9).
The expression plasmids pMEXHXasak42D⌬N11,G12K, pMEXHX-asak42D⌬N14,A15K and pMEXHXasak42D⌬N16 were constructed using the cassette methodology as described in (9). The nucleotide boxes covering the coding sequence from the NdeI site to the MluI site in the acceptor plasmid pMEXsak42D(⌬5-13) (9) and from the NdeI site to the SfuI site in the expression plasmid of variant Sak42D⌬N14(M),A15K, respectively, were substituted by the linker pairs L1/L2, L3/L4 and L5/L6, respectively. By this manipulation, a metal ion binding site (His 6 -tag) followed by the substrate recognition site of factor Xa was generated in front of the truncated Sak42D variants. The variants were obtained by expression of the proteins, and cleavage with factor Xa. Briefly, cell extracts prepared by sonication of E. coli TG1 cells transformed with the appropriate expression plasmids were cleared by centrifugation (20,000 rpm for 1 h) and applied to a Ni-NTA column (4 ml of resin/200 ml of cell culture volume). Bound proteins were eluted with 0.25 M imidazole and dialyzed extensively against 20 mM Tris-HCl buffer, pH 8.0, containing 100 mM NaCl, 2 mM CaCl 2 . The dialyzed eluate from 200 ml of culture volume was treated with 60 l of factor Xa solution for 9 h at 37°C. The separated NH 2 -terminal part and uncleaved material were removed by passage through a Ni-NTA column. The flow-through was concentrated by precipitation with (NH 4 ) 2 SO 4 at 85% saturation, and the redissolved material was further purified using a Superdex 75 HR 16/60 TM column (Pharmacia, Freiburg, Germany).
Analytic Methods-The NH 2 -terminal amino acid sequences of all purified Sak42D variants were determined on the Applied Biosystems model 476A (Applied Biosystems, Foster City, CA). For the Sak species generated in equimolar (4 M) mixtures with plasminogen after 5 min of incubation at 37°C, high resolution electrophoresis was carried out with the Mini-Protean II system (Bio-Rad, Munich, Germany) using 16% T (total acrylamide), 3% C (cross-linking agent) gels in the discontinuous Tris-Tricine buffer system according to Schagger and von Jagow (15). Samples were applied to the gels after reduction by heating at 100°C for 5 min in the presence of 1% SDS and 0.4% ␤-mercaptoethanol. Following gel separation, the proteins were blotted onto polyvinylidene difluoride membranes (Millipore, Eschborn, Germany) and then subjected to amino acid sequencing.

Generation of Amidolytic Activity in Equimolar Mixtures of Sak42D
Variants and Plasminogen-Amidolytic activity was quantitated with the chromogenic plasmin substrate S-2251 (1 mM final concentration) and monitored at 405 nm for up to 12 min using a Spectramax 250 microplate reader (Molecular Devices, Sunnyvale, CA). Plasminogen (3.6 M final concentration) was incubated with the Sak moieties (4 M final concentration) at 37°C in 100 mM phosphate buffer, pH 7.4, containing 0.01% Tween 80 (activation buffer). Samples were withdrawn at 2-min intervals and diluted 40-fold in buffer containing chromogenic substrate S-2251, and the change in absorbance at 405 nm (⌬A 405 ) was recorded.
Activation of Plasminogen by Catalytic Amounts of Sak42D Variants-The activation of plasminogen (1.5 M final concentration) by Sak42D moieties (5 nM final concentration) was assayed at 37°C in activation buffer. At different time intervals up to 30 min, generated plasmin was measured with the S-2251 substrate. Sak42D and Sak42D⌬N14(M),A15K were also evaluated at lower temperatures to detect differences in activation energies.

Kinetic Constants of Plasminogen Activation by Catalytic Amounts of Preformed Equimolar Complexes of Sak42D Variants with Plasmin-
The kinetic constants of plasminogen activation by plasmin-Sak complexes were derived from Lineweaver-Burk plots. Therefore, equimolar mixtures of Sak42D variants and freshly prepared plasminogen were preincubated for 5 min at 37°C in activation buffer containing 25% glycerol (v/v) and kept on ice. These preformed activator complexes (10 -40 nM final concentration) were then mixed with plasminogen (0.25-10 M final concentration), and generated plasmin was measured at 37°C from the change in absorbance at 405 nm with S-2251. Accurate Lineweaver-Burk plots could not be constructed with variants displaying catalytic efficiencies Ͻ20% of that of wild type. For these variants, catalytic efficiencies were determined from the apparent second order rate constant at a plasminogen concentration of 5 M.
Binding of Sak42D Variants to Plasminogen-Association rate constants (k a ) and dissociation rate constants (k d ) for the interaction between Glu-Plg or rPlg(S741A) and the Sak variants were determined by real-time biospecific interaction analysis using the BIAcore instrument (Biacore Inc) (16). Glu-Plg or rPlg(S741A) were immobilized on the surface of sensor chip CM5 using the amine coupling kit (Biacore Inc), as recommended by the manufacturer. Immobilization was performed from protein solutions at 10 g/ml in 10 mM sodium acetate buffer, pH 5.0, at a flow of 5 l/min for 6 min. The Sak42D variants were injected in solution over the sensor. The concentration of free analyte was kept constant by maintaining a continuous flow of solution at 20°C past the sensor surface. At least three concentrations of each analyte (25-100 nM CGAATTTTCGACCTTCAATGTGATGATGATGGTGATGCA a R stands for "AϩG" according to the IUB nomenclature.

Purification and NH 2 -terminal Sequence of Sak42D
Variants-The Sak42D variants were purified from E. coli TG1 cells transformed with the respective plasmids with yields of 70 -500 g/liter culture volume. SDS-PAGE (data not shown) displayed single bands with electrophoretic mobilities corresponding to the extent of NH 2 -terminal deletion. Homogeneity of the purified proteins was further revealed by NH 2 -terminal sequencing as illustrated in Table II. The mutants Sak42D⌬N11, Sak42D⌬N14, and Sak42D⌬N15, like the wild type Sak42D, were correctly processed in the E. coli TG1 expression system with removal of the NH 2 -terminal initiation Met. The variants with Lys substitution in the NH 2 -terminal position were expressed with a Met residue at the NH 2 terminus, added to initiate expression in E. coli, but which was not removed when followed by Lys.

NH 2 -terminal Sequences of Sak Moieties in Equimolar Complexes with Plasminogen and Generation of Amidolytic
Activities-As expected, wild type Sak42D was processed at the Lys 10 -Lys 11 peptide bond as described earlier (9). The variants with deletion of 11, 14, or 15 amino acids, with or without substitution of the next amino acid with Lys, maintained unaltered NH 2 -terminal sequences after complex formation. Variants with deletion of 11 amino acids in combination with substitution of Gly 12 with Lys and of Asp 13 with Ile or Asn were partially processed to derivatives hydrolyzed after the substituted Lys peptide bond (Table II).
In line with previous observations (9), amidolytic activity toward the synthetic plasmin substrate S-2251 was detected in all equimolar mixtures of plasminogen with the Sak variants (not shown). Mutants whose complexes with plasmin did not efficiently activate plasminogen, expressed up to 3-fold higher amidolytic activity than complexes of wild type Sak42D with plasmin. This was not unexpected in view of the known higher specific activity of the active site of plasmin toward the chro-mogenic substrate, as compared with that of the active site of the plasminogen-activating plasmin⅐Sak complex (9).
Activation of Plasminogen by Catalytic Amounts of Sak Variants-Catalytic amounts of Sak42D induced activation of plasminogen to plasmin with a lag phase of approximately 5 min, reaching a maximal rate (determined from the slope at the inflection point of these curves) after 5-15 min. The results obtained with active complexes are illustrated in Fig. 1. As described previously (9), the Sak42D⌬N10(M) and Sak42Dproc variants displayed activation curves similar to those of wild type Sak42D (not shown). Sak 42D⌬N14 and Sak42D⌬N15 were inactive but could be rescued by replacing their respective NH 2 -terminal amino acid by a lysine residue, yielding functionally active Sak42D⌬N14(M),A15K and Sak42D⌬N15(M),S16K (Fig. 1). Sak42D⌬N11 was also inactive but could only be rescued by substitution of Gly 12 with Lys and in addition of both Asp 13 Y17K,F18N,E19N, did not restore plasminogen-activating capacity (not shown).
To evaluate the influence of the unprocessed NH 2 -terminal Met residues in the variants with NH 2 -terminal Lys substitutions, "methionine-free" Sak42D⌬N11,G12K, Sak42D⌬N14, A15K, and Sak42D⌬N16,Y17K variants were prepared. Table  II demonstrates that Sak42D⌬N14,A15K still activated plas- minogen, whereas the other variants, like their Met-containing counterparts, were inactive. Temperature Dependence of Functional Properties of Sak Moieties-To evaluate possible differences in activation energies for the induction of plasminogen-activating potential between unprocessed wild type Sak42D and the "preprocessed" deletion mutant with rescued plasminogen-activating potential, Sak42D⌬N14(M),A15K, the effect of temperature on the activation of plasminogen by catalytic amounts of Sak variant were studied (Fig. 2). At 37°C, plasmin generation progressed at similar rates, whereas at 10°C no plasmin generation occurred within 30 min. At intermediate temperatures (15-25°C), plasmin generation proceeded significantly faster with Sak42D⌬N14(M),A15K than with wild type Sak42D.
Binding of Sak Variants to Plasmin/Plasminogen-The association and dissociation rate constants (k a and k d ) and the  apparent affinity equilibrium constants (K a ) for binding of Sak42D moieties to native Glu-Plg in the presence of excess plasmin inhibitor or to rPlg(S741A) are summarized in Table  III. The k a , k d , and K a values of all Sak42D variants studied were very similar to those of wild type Sak42D. The affinity constants for binding to native Glu-Plg were approximately 1000-fold lower than those for binding to rPlg(S741A) as a result of 100-fold higher k a and 10-fold lower k d values.

DISCUSSION
Sak is a highly fibrin-selective and potent agent for thrombolytic therapy in patients with thromboembolic arterial disease (3,4). Fibrin selectivity is due to the fact that Sak has a low affinity for plasminogen in plasma, but a high affinity for plasminogen or for traces of plasmin associated with fibrin clot (17). Furthermore, the plasminogen⅐Sak complex is inactive and both conversion of plasminogen to plasmin and NH 2 -terminal processing of Sak to Sak42⌬N10 by hydrolysis of the Lys 10 -Lys 11 peptide bond are required to generate active plasmin⅐SakЈ complex at the fibrin surface, whereas removal of Lys 11 results in loss of plasminogen-activating potential. The aim of the present study was to characterize structural motifs in the NH 2 -terminal region of Sak possibly required for binding to plasminogen and for reconfiguration of the active site of plasmin in the plasmin⅐Sak complex from that of fibrin degrading plasmin to that of plasminogen-activating plasmin⅐Sak complex. This was studied by evaluation of the NH 2 -terminal processing, the catalytic efficiency for plasminogen activation, and the binding affinity to insolubilized plasminogen of deletion/substitution variants of the Lys 11 -Gly-Asp-Asp-Ala-Ser 16 -Tyr-Phe-Glu region of Sak.
The results indicate that removal of K, KGDD, or KGDDA, yielding Sak42D⌬N11, Sak42D⌬N14, or Sak42D⌬N15, did not affect the association and dissociation rate constants of the binding of these variants to plasminogen, but reduced the catalytic efficiency for plasminogen activation by more than 1 order of magnitude. Substitution of the NH 2 -terminal amino acids by Lys in Sak42D⌬N14 or Sak42D⌬N15, yielding Sak42D⌬N14,A15K and Sak42D⌬N15,S16K, rescued the plasminogen-activating activity, whereas Sak42D⌬N16,Y17K and Sak42D⌬N11,G12K were inactive. Further substitution of Asp 13 and Asp 14 in Sak42D⌬N11,G12K with Asn, yielding Sak42D⌬N11,G12K,D13N,D14N, rescued the plasminogen-activating activity, but a similar substitution of Phe 18 and Glu 19 in Sak42D⌬N16,Y17K with Asn, yielding Sak42D⌬N16,Y17K, F18N,E19N did not endow the latter variant with plasminogen-activating potential. Finally, all variants studied had similar association and dissociation rate constants for low affinity binding to insolubilized native plasminogen and for high affinity binding to insolubilized active site-substituted recombinant plasminogen.
No differences in plasminogen-activating potency were observed between Lys-substituted deletion variants with or without the presence of the NH 2 -terminal initiation Met. Nevertheless, although the catalytic efficiencies of Sak42D⌬N14(M), A15K and Sak42D⌬N15(M),S16K were similar to that of wild type Sak42D, they had 10-to 20-fold higher k cat and K m values. The mechanism underlying these differences remains unclear. It is possible that the exposed NH 2 -terminal Lys primarily influences the active site configuration via its ⑀-amino group and that masking of the NH 2 -terminal ␣-amino group might yield a more efficient active center.
These findings indicate: 1) that binding to plasmin/plasminogen and reconfiguration of the active site of plasmin are mediated by different structures in the Sak molecule; 2) that a positive amino acid in the NH 2 -terminal position of Sak, in the region corresponding to positions 11-16 is required for reconfiguration of the active site of plasmin; and 3) that one or two negatively charged amino acids following the positive NH 2 -terminal amino acid prevent reconfiguration of the active site of plasmin.
The variant Sak42D⌬N16,Y17K marks the limit for rescuing, by introduction of a Lys at the NH 2 terminus, the plasminogen-activating potency of NH 2 -terminal trimmed Sak molecules. This finding is in agreement with the recent finding (10,11) that Sak lacks secondary structure in the region spanning the first 20 NH 2 -terminal amino acids. This region may have a flexible conformation located at the surface of Sak, which after hydrolysis and exposure of the positively charged amino acid may interact with the active site of plasmin to reconfigure its catalytic specificity, converting it from a fibrin-degrading to a plasminogen-activating proteinase.
As reported previously, NH 2 -terminal processing of wild type Sak42D does not appear to be the rate-limiting step in plasminogen activation by Sak at 37°C (18). Consequently, no differences in activation efficiency were found at 37°C between wild type Sak42D, which needs to be processed NH 2 -terminally, and the "preprocessed" deletion variants. However, at lower temperatures, activation of plasminogen with catalytic amounts of Sak42D⌬N14(M),A15K proceeded more rapidly compared with that of wild type Sak42D. This finding suggests that the activation energy of the processing of Sak42D is higher than that of the reconfiguration of the active site.
The Lys residue unmasked by cleaving off the NH 2 -terminal peptide from Sak42D may be directly involved in the reconfiguration of the active site of plasmin. Recent data concerning the molecular interactions modulating the catalytic efficiency and specificity of serine proteinases support this assumption. Thus, a Lys residue in single-chain tissue plasminogen activator stabilizes the active conformation of this enzyme (19). Moreover, in a series of proteases, Lys may substitute His as the general base in the catalytic triad (20 -22). Definitive proof of the involvement of Lys residues of Sak in complexes with plasmin that specifically activate plasminogen may require elucidation of the tertiary structure of the complex.