Characterization of Recombinant Human Single Chain Urokinase-type Plasminogen Activator Mutants Produced by Site-specific Mutagenesis of Lysine 158"

The cDNA encoding full-length single chain uroki-nase-type plasminogen activator (scu-PA) was cloned and sequenced, and the recombinant scu-PA (rscu-PA) was expressed in Chinese hamster ovary cells. Two mutants, constructed by in vitro site-specific mutagenesis of Lys''' in rscu-PA to Gly'" (rscu-PA-GlylSs) or to Glu'" (mu-PA-Glu"), were also expressed in Chinese hamster ovary cells. Wild type and mutant rscu-PAS were purified to homogeneity by immunoadsorption on an insolubilized monoclonal antibody raised against natural scu-PA (nscu-PA), followed by gel filtration. The specific activity of the mutant scu-PAS on fibrin plates is very low (~1,000 IU/mg) compared to that of the wild type rscu-PA (44,000 IU/mg). The mutants, in contrast to the wild type rscu-PA, are not converted to amidolytically active two chain u-PA (tcu-PA) by plasmin and do not cause lysis of a "'I-fibrin-labeled plasma clot immersed in citrated plasma. However, in a purified system, both mu-PA-Gly'" and rscu-PA-Glu16s acti- vate plasminogen

The intrinsic plasminogen activating potential of scu-PA has been explained by its high affinity for plasminogen (low Michaelis constant) which compensates for its low catalytic rate constant (18,19). This high affinity of scu-PA for plasminogen is not mediated via the lysine binding site(s) of plasminogen (20) and is independent of the NH2-terminal region of scu-PA (21).
In the present study we have constructed mutants of recombinant scu-PA (rscu-PA) by site-specific mutagenesis of Lys'= to Gly'= or to Glu15' , in order to further investigate the significance of the conversion of scu-PA to tcu-PA for the activation of plasminogen. The mutants were expressed in a mammalian cell system, purified from the conditioned medium, and their structural and functional properties evaluated.
Three overlapping cDNA clones were spliced together to produce a contiguous cDNA sequence encoding full-length scu-PA. This cDNA was HindIII linkered (5'-GAAGCTTC-3', Boehringer Mannheim) and cloned into pUC18 (27) to yieldpULscu-PA. cDNA restriction endonuclease fragments of this construct were then subcloned into M13mp18 or M13mp19 (27) and sequenced by the dideoxynucleotide chain-termination method (28). Restriction enzymes and DNA modifying enzymes were from New England Biolabs, Boehringer Mannheim, or Gibco/BRL (Gent, Belgium).
Mutagenesis-Site-directed mutagenesis was performed essentially as described (29) using a Sau3AI fragment of scu-PA cDNA (nucleotides 399-734, encoding amino acids S e P through Ilem), cloned in M13mp18, as template.
100 ng of phosphorylated deoxyoligonucleotide of either 5'-CCCCAATAATWAAAGCGGGGCC-3'(Gly1"-mutant primer) or 5'-CCCCAATAAl"l'Tl'TAAAGCGGGGCC-3'(Glu1"-mutant primer) (synthesized by New England Biolabs), which are complementary to the template, and 20 ng of this M13 template DNA were heated at 65 "C in 50 mM Tris-HC1 buffer, pH 7.8, containing 10 mM MgCIz and 20 mM dithiothreitol for 5 min and annealed by stepwise cooling to room temperature and 0 "C. ATP was added to a concentration of 0.4 mM and dNTPs to a concentration of 0.05 mM in a final volume of 50 pl. 400 units of T, DNA Ligase (New England Biolabs) and 5 units of Klenow fragment of E. coli DNA polymerase I (Boehringer Mannheim) were added. The DNA obtained after 1-h incubation at 14 'C was used to transfect E. coli JMlOl cells (27).
The deoxyoligonucleotides encoding the Gly'" and Glu'" mutations were radiolabeled using 100 pCi of [d"P]ATP (New England Nuclear, 7000 Ci/mmol) to a specific activity of l@ cpm/pg, in 70 mM Tris-HC1 buffer, pH 8.0, containing 10 mM MgC12, 5 m M dithiothreitol, and 10 units of TI-polynucleotide kinase in a 20-pl reaction mixture which was incubated for 30 min at 37 "C. The labeled deoxyoligonucleotides were then used for plaque hybridization to the mutagenized phage DNA, essentially as described (26).
Eukaryotic Expression of scu-PA cDNA-A HindIII DNA restriction endonuclease fragment containing 88 base pairs of 5"untranslated DNA, the scu-PA coding sequence, and 83 base pairs of 3'untranslated DNA (nucleotides -22-1448) was cloned into the expression vector pSV328DHFR (Fig. 1). This vector is a derivative of pSV328A+ (31), containing mouse DHFR cDNA in a separate expression module, Le. between an SV40 early promoter/enhancer and a 3' region of the rabbit @-globin gene. To construct this derivative pSV5DHFR (32) (American Type Culture Collection No. 37148) and pSV328A+ (kindly provided by M. Van Heuvel, Erasmus Institute, Rotterdam, the Netherlands) were each digested with HindIII and the resulting protruding ends filled-in with Klenow fragment of E. coli DNA polymerase I (5 units in 20-4 reaction volume) in the presence of 500 pM dNTPs. The linearized and filled-in pSV5DHFR was then digested with BglII, and a 650-base pair DNA restriction endonuclease fragment, containing the complete mouse DHFR coding sequence was isolated. The linearized and filled-in pSV328A+ was digested with BamHI, and this vector was used to accept the DHFR coding sequence by ligation. The resulting plasmid, pSVDHFR thus contains the mouse DHFR coding sequence flanked by the SV40 early promoter and the rabbit @-globin gene 3'-end. PuuII digestion, followed by SalI-linker (B'XGTCGACG-3') addition and SalI digestion resulted in a SalI DNA restriction endonuclease fragment containing all the necessary genetic elements for expression of the DHFR gene. This fragment was then ligated into the S d site in the pBR328 portion of pSV328A+ resulting in pSV328DHFR (Fig. 1). Final construction of the expression vector for the cDNA of natural scu-PA pSV5DHFR (32) was digested with HindIII, similarly filled-in, and further digested with BglII (Bg). The blunt-end to BgnI restriction endonuclease fragment of pSV5DHFR containing the DHFR sequence (stippled segment) was then ligated into the linearized bluntend to HindIII pSV328A+ vector to yield pSVDHFR. pSVDHFR was digested with PUuII (P), and SalI linkers were added followed by digestion with SalI (S). The DHFR containing restriction endonuclease fragment of pSVDHFR was ligated into SalI digested pSV328A+ to yield pSV328DHFR. E, EcoRI; EP, SV40 early promoter; Amp, ampicillin resistance gene. pSV328A+ contains pBR328 sequence (line), SV40 early promoter (solid black segment, EP) and rabbit @globin 3'sequence (open segment). pSV5DHFR contains pBR322 sequence (line), SV40 early promoter (solid black segment, EP), mouse DHFR cDNA sequence (stippled segment) and another segment of 3"untranslated sequence (solid black segment) containing both SV40 and Polyoma T-antigen sequences (32). A polylinker region is located between the HindIII site and the EcoRI site downstream from the SV40 early promoter. DHFR deficient Chinese hamster ovary cells (33) were a gift from W. Fiers, University of Gent, Belgium and were maintained in minimal essential a medium (Gibco/BRL, Gent, Belgium), supplemented with 10% fetal calf serum (Flow Laboratories, Irvine, Scotland) in air supplemented with 5% COz. The cells were transfected, using the calcium phosphate coprecipitation method (34) either by cotransfection with 0.5 pg pSV5DHFR (32) and 5 pg of either pSVscu-PA-Gly' " or pSVscu-PA-Glu'", or as a single transfection using 5 pg of pSVscu-PADHFR. Selection for DHFR+ cells was in Ham's F-12 medium (Gibco/BRL) lacking glycine, thymidine, and hypoxanthine, and supplemented with 8% dialyzed fetal calf serum (500 ml of heat tubing, Spectrum Medical Instruments, Los Angeles, with a M, exclusion limit of 3500). Isolated DHFR+ cells were monitored for secretion of rscu-PA, rscu-PA-G1ylB, or rscu-PA-Glu'" with a monoclonal antibody based ELISA (35) or on fibrin plates (36). For large scale production, cells were grown at 37 "C in 850 cmz roller bottles (Becton Dickinson, Oxnard, CA) in 200 ml of minimal essential (Y medium supplemented with 10% fetal calf serum and 25 mM Hepes buffer, pH 7.3. At confluency 1 g (0.48 m2 surface area) of microcarrier beads (Superbeads, Flow Laboratories) were added per roller bottle, and the volume of medium was increased to 350 ml. The cells reached confluency on the beads in approximately 3 days, after which the medium was replaced by serum-free medium consisting of equal volumes of Ham's F-12 medium and Dulbecco's modified Eagle's medium (Gibco/ BRL) containing 4.5 g/liter glucose. This medium was further supplemented with 50 mM Hepes buffer, pH 7.3,lO pg/ml insulin (Gibco/ BRL), 5 pg/ml transferrin (Gibco/BRL), 7.5 pg/ml glycine, 1 X nonessential amino acids (Gibco/BRL), 2 ng/ml linoleic acid (Sigma), 50 units/ml penicillin/streptomycin (Gibco/BRL), 2 mM L-glutamine (Gibco/BRL), and 10 kallikrein inhibitor units/ml aprotinin (Bayer, Leverkusen, Federal Republic of Germany). Conditioned medium was harvested after a 3-day incubation period.
Purification of scu-PAS-All scu-PA moieties characterized in the present study were purified by immunoadsorption on insolubilized monoclonal antibody 4DlE8 (21), followed by gel filtration on Ultrogel AcA 44. Therefore, 500-1,000-ml batches of conditioned medium from CALU-3 cells (9) or from transfected Chinese hamster ovary cells were concentrated 3-&fold by ultrafiltration on PM-30 membranes (Amicon, Danvers, MA). The concentrate was applied at 4 "C at a flow rate of 7 ml/h to the 0.9 X 3-cm immunoadsorption column (6 mg of antibody/g of Sepharose 4B) which was equilibrated with TNTA buffer and was eluted with the same buffer containing 2 M KSCN. u-PA-related antigen was monitored by ELISA (35). Urokinase-containing fractions were pooled, concentrated to 3 ml with polyethylene glycol 20,000, dialyzed, and applied to an Ultrogel AcA 44 column (1.6 X 90 cm) equilibrated with TNTA buffer at a flow rate of 10 ml/h. Before use, all scu-PA preparations were filtered through a small column of benzamidine-Sepharose to remove tcu-PA (15) and washed on a Centricon 30 microconcentrator (Amicon) to remove aprotinin. Protein concentrations were determined by amino acid analysis as described below.
Other Proteins and Reagents-Plasminogen with NH2-terminal glutamic acid (Glu-plasminogen), a2-antiplasmin, and fibrinogen were prepared from pooled normal human plasma as described (37). t-PA was purified from conditioned medium of human melanoma cells (Bowes) (38). Fibrinogen was labeled with lZ5I using the iodogen method (39). The plasminogen preparations were over 95% activatable to plasmin as measured by active-site titration withp-nitrophenylp'-guanidinobenzoate (40) after treatment for 2 h with streptokinase (1500 IU/mg of protein) in 0.1 M phosphate buffer, pH 7.30, containing 25% glycerol at 0 "C. When required, plasminogen was treated with aprotinin-Sepharose until no residual plasmin was detectable with S-2251 (lower detection limit less than 0.1%) (15). Plasmin-Sepharose was obtained by overnight incubation of plasminogen-Sepharose (2.8 mg of protein/ml of settled gel volume, prepared as described below) in 0.1 M phosphate buffer, pH 7.30, containing 25% glycerol at 4 "C with tcu-PA (molar ratio of urokinase to plasminogen 150). Plasmin activity was measured in a 1:200 diluted suspension with S-2251 (final concentration 0.3 mM).
Aprotinin-Sepharose, benzamidine-Sepharose, and plasminogen-Sepharose were obtained by coupling aprotinin (Trasylol@, Bayer), p-aminobenzamidine or plasminogen to CNBr-activated Sepharose 4B which was purchased from Pharmacia P-L Biochemicals. The International Fkference Preparation for urokinase (66/46) was obtained from the National Institute for Biological Standards and Control (London, United Kingdom).
Treatment of Lys" Mutants of scu-PA with Plasmin-scu-PA preparations (nscu-PA, rscu-PA, rscu-PA-Gly'", and mu-PA-Glu'") were incubated with plasmin in TNT buffer at 37 "C. At timed intervals (0-20 min), 10-pl samples were removed and generated urokinase activity measured following addition to 800 pl of TNT buffer containing the chromogenic substrate s-2444 at a final concentration of 0.3 mM. Urokinase activity was expressed in IU by comparison with the International Reference Preparation (66/46).
Conversion of single chain to two chain u-PA was also monitored on 12% SDS-PAGE under reducing conditions (42). Therefore, samples were removed from the incubation mixtures after 20 min, mixed with 0.062 M Tris-HC1 buffer, pH 6.80, containing final concentrations of 10% glycerol, 1% SDS, and 1% dithioerythritol, and incubated at 100 "C for 3 min.
Activation of Plasminogen by Lysl" Mutants of scu-PA in Purified Systems-Activation of plasminogen (final concentration 1.5 pM) in TNT buffer at 37 "C in the presence of a2-antiplasmin (final concentration 1.6 p~) by the scu-PAS was measured using final concentrations of 18 nM for nscu-PA, 40 nM for rscu-PA, 53 or 530 nM for rscu-PA-Gly'", and 42 or 420 nM for rscu-PA-Glu'", respectively. At different time intervals (0-60 min), 5-pl samples were removed from the mixtures and incubated for 10 min at 37 "C in 300 p1 of TNT buffer containing 1000 IU of streptokinase. Under these conditions, plasmin is rapidly and quantitatively neutralized by a,-antiplasmin, whereas the streptokinase-plasmin(ogen) complex, which is not inhibited by a2-antiplasmin (43), can be quantitated by addition of 700 p1 of S-2251 solution (final concentration 0.6 mM) (15). The residual plasminogen, measured as streptokinase-plasmin(ogen) complex was then expressed in percent.
Kinetic parameters for the activation of plasminogen by the scu-PAS were determined essentially as previously described (18,19). nscu-PA (22 nM)  Other Laboratory Methods-Double immunodiffusion analysis was performed according to Ouchterlony (45). Amino acid analysis was performed with a 119 CL Beckman amino acid analyzer after sample hydrolysis with 6 M HC1 in uacuo at 100 "C for 20 h.

Cloning, Sequencing and Reconstruction of scu-PA cDNA
Three overlapping cDNA clones (Fig. 3) were isolated from two Xgtll-based cDNA banks. AUK4 and AUK23 originated from an oligo(dT)priming of poly(A+) from CALU-3 cells and were identified with the 49 deoxyoligonucleotide-long probe corresponding to nucleotides 931-979. The combined sequence of these two cDNAs encoded amino acids 40-411 of scu-PA and most of the 3"untranslated sequence. The missing amino-terminal sequence was identified in a second cDNA library made by specifically priming the poly(A+) mRNA of CALU-3 cells with the U2-synthetic deoxyoligonucleotide. Plaque hybridization with a [32P]-labeled nick-translated NcoI-EcoRI DNA restriction endonuclease fragment from XUK23, identified ASPUK17, which contains 102 nucleotides of 5"untranslated sequence, the initiation codon, and the sequence coding for amino acids 1-163. The complete coding sequence was reconstructed in pUC18 to yield pULscu-PA, including nucleotides -22-1448 (Fig. 3). The numbering is according to Holmes et al. (lo), supplemented at the 5'-end with additional sequence identified by negative numbers. The complete CALU-3 mRNA derived cDNA sequence, which we determined by the dideoxy nucleotide chain-termination method (28), revealed a total of six differences as compared with the published sequence of the cDNA from Detroit 562 human pharyngeal carcinoma cells (10). These include a GC insert in the 5"untranslated sequence between nucleotides 15 and 16 and deletion of guanosine at position 43. Both differences were also observed by Riccio et al. (46). There are four third base "silent" substitutions, three of which were also found in the genomic sequence (46). One C to T substitution in the Asn254 codon, one A to G substitution in the codon, one C to A substitution in the Pro345 codon and one A to G substitution in the Gln346 codon.

Mutagenesis and Expression
The scu-PA cDNA was mutagenized at the nucleotide region specifying amino acid 158, substituting a glycine codon or a glutamic acid codon for the natural lysine codon. This resulted in rscu-PA-Gly15' and rscu-PA-Glu" as described under the "Materials and Methods" section. The desired mutations were confirmed by DNA sequencing of the entire fragment employed for site-specific mutagenesis, and no other mutations were found. The natural cDNA was integrated in the selectable eukaryotic expression vector pSV328DHFR harboring DHFR. The mutants were built into a nonselectable plasmid, pSV328A+ and cotransfected with pSV5DHFR.
Chinese hamster ovary DHFR-cells were transfected, and DHFR+ colonies were isolated and screened for u-PA-related antigen production. The wild type or mutant scu-PA-producing cells were cultured to obtain conditioned medium for purification of the expressed proteins.

Purification and Physicochemical Characterization of Lysl=
Mutants of scu-PA u-PA-related antigen concentrations in the conditioned media used for purification of the various molecular forms of scu-PA were 0.5 pg/ml, 0.5 pg/ml, 0.8 pg/ml, and 1.6 pg/ml for nscu-PA, rscu-PA, rscu-PA-Gly'", and rscu-PA-Glu1.5R, respectively. After immunoadsorption and gel filtration, final yields of about 35, 60, 50, and 70% were obtained for nscu-PA, rscu-PA, rscu-PA-GlyIw, and rscu-PA-Glu'=, respectively. In order to exclude the copurification of an endogenous plasminogen activator from Chinese hamster ovary cells, we have cotransfected Chinese hamster ovary cells with pSV5 DHFR and a different SV40 early promotor-based plasmid expressing a cDNA unrelated to either u-PA or t-PA. Conditioned medium was carried through the purification procedure on insolubilized monoclonal antibody 4DlE8 as described. In the concentrated eluate we were unable to find any significant amount of plasminogen activator antigen (by ELISA as described) or activity (on fibrin plates or by direct incubation with 50 p~ plasminogen). Fig. 4 shows that on SDS-PAGE, the purified scu-PAS migrate as single chain polypeptides with a M, of approximately 54,000 under both nonreducing and reducing conditions. Immunodiffusion analysis (Fig. 5) confirmed that all forms of scu-PA were immunologically identical when reacted with a rabbit antiserum raised against tcu-PA, whereas none of them reacted with rabbit antiserum against t-PA. The amino acid compositions of the four molecular forms of purified scu-PA were indistinguishable (data not shown).

Functional Characterization
Treatment of Lysl= Mutants of scu-PA with P h m i n -Specific activities measured on fibrin plates amounted to 55,000 IU/mg for nscu-PA, 44,000 IU/mg for rscu-PA, 800 IU/mg for rscu-PA-Gly'=, and 530 IU/mg for rscu-PA-Glu".  The background amidolytic activity on S-2444 of all scu-PA moieties used was very low: 400,410,250, and 150 IU/mg for nscu-PA, rscu-PA, rscu-PA-GlyIm, and rs~u-PA-Glu'~~, respectively. These activities correspond to an equivalent of less than 0.4% of urokinase in the samples used for analysis. Plasmin causes a time-dependent conversion of nscu-PA and of rscu-PA to tcu-PA (Fig. 6). Maximal conversion results in an amidolytic activity corresponding to a specific activity of about 80,000 IU/mg of protein for nscu-PA and 50,000 IU/ mg of protein for rscu-PA. With the mutant proteins rscu-PA-GIy"* and rs~u-PA-Glu'~", no significant generation of tcu-PA activity was observed under the same experimental conditions, even when using a 10-fold higher plasmin concentration (Fig. 6 ) . SDS-PAGE under reducing conditions (Fig.  4C) shows that incubation of nscu-PA or rscu-PA with plasmin (molar ratio 1/140) for 20 min at 37 "C is associated with virtually complete conversion of the single chain molecule (M, 54,000) to the two chain form, containing a NHp-terminal chain (A-chain) of M, 20,000 and a COOH-terminal chain (Bchain) of M, 34,000. Using a 10-fold higher molar ratio of plasmin to scu-PA, no significant conversion to a two chain molecule was observed with rscu-PA-GluISR, whereas a partial degradation of rscu-PA-Gly" was observed.
Activation of Plasminogen by Lys'" Mutants of scu-PA in Purified System- Table I shows that, in mixtures containing 1.5 p M plasminogen and 1.6 p M cup-antiplasmin, plasminogen is activated to plasmin by nscu-PA and by rscu-PA, whereas no significant activation occurs with rscu-PA-Gly" and rscu-PA-GlP' at comparable enzyme to substrate ratios. However, at very high concentrations (scu-PA to plasminogen ratio about 1:3), significant conversion of plasminogen to plasmin was observed. Kinetic analysis revealed that plasminogen is activated to plasmin by all four scu-PA moieties, following Michaelis-Menten kinetics as shown by linear double-reciprocal plots of the activation rate (u) uersw the plasminogen concentration ( [ P w ) (Fig. 7 ) . For nscu-PA and rscu-PA, the activation rate of plasminogen, measured in the presence of 1 mM S-2251, increased with the plasminogen concentration in the range between 0.25 and 1.50 p~ (Fig. 7A). The kinetic constants, obtained by linear regression analysis, were very similar to those previously reported (19): K , = 0.8 p M and 1.0 pM and k2 = 0.001 s" and 0.002 s" for nscu-PA and rscu-PA, respectively. With rscu-PA-GlylS and rscu-PA-Glu", no significant activation was observed at these plasminogen concentrations. At higher plasminogen concentrations (10-75 p~) , however, a concentration-dependent activation was observed (Fig. 7 B ) which yielded the following kinetic constants: K, = 62 p~ and 87 p~, and kp = 0.011 s" and 0.010 s-l for rscu-PA-Gly168 and rscu-PA-Glu'm, respectively (Table 11).
During these measurements, the concentration of generated plasmin was always less than 1% of the total plasminogen Concentration. Control experiments revealed that under the conditions used, spontaneous generation of plasmin at a plasminogen concentration of 50 p M was about 0.01% after 20 min. However, at the end of the incubation period, the molar ratio of plasmin to either rscu-PA-GlylS or rscu-PA-GlulS increased to as much as 7 to 1. Therefore, in order to confirm the lack of conversion of scu-PA-Gly" or scu-PA-Glu'68 to amidolytically active u-PA, in the presence of such a molar excess of plasmin, control experiments whith insolubilized plasmin were carried out as described under "Materials and Methods." With concentrations of insolubilized plasmin of Binding to Fibrin-In a purified system, neither nscu-PA, rscu-PA, nor the two Lys15' mutants bound specifically to fibrin, whereas specific and concentration-dependent binding of t-PA to fibrin was confirmed (data not shown).  (Fig. 8B). DISCUSSION scu-PA, with a very low activity towards low molecular weight amide substrates (2,13,14), is converted to tcu-PA with a high amidolytic activity, by specific cleavage with plasmin of the Ly~'~'-Ile'~~ peptide bond (3,5,6,10,12,13). We have reported previously that scu-PA nevertheless directly activates its natural substrate plasminogen (15,18,19), whereas others (7) have reported that scu-PA is inactive towards plasminogen. Investigation of the activation mechanism of plasminogen by scu-PA is complicated by the fact that generated plasmin may convert scu-PA to tcu-PA. Therefore, in order to investigate the significance of this conversion of scu-PA to tcu-PA for the activation of plasminogen, we have constructed mutants of recombinant scu-PA by sitespecific mutagenesis of Lys15' to Glylm or to Glu15' . These mutants were expressed in a mammalian cell system (Chinese hamster ovary cells) and purified on an insolubilized murine monoclonal antibody directed against urokinase. Both mutants (rscu-PA-G1ylm and r~cu-PA-Glu'~') were compared with the wild type recombinant scu-PA (rscu-PA) and with scu-PA purified in the same way from conditioned medium of CALU-3 cells (nscu-PA) (9).
Both mutants are found to be indistinguishable from rscu-PA and nscu-PA on the basis of their M,, their reactivity with an anti-u-PA antiserum, and the lack of specific binding to fibrin. Functional characterization, however, revealed that, in contrast with rscu-PA and nscu-PA, r~cu-PA-Gly'~~ and rscu-PA-Glu15' have a very low specific activity on fibrin plates and are not converted to amidolytically active tcu-PA upon treatment with plasmin. In addition, at physiological plasminogen concentrations (1.5-2 p M ) in purified systems, a greatly reduced rate of plasminogen activation is observed with these mutants. Finally, they do not induce fibrin clot lysis in a plasma milieu at concentrations up to 150 nM. Kinetic analysis, however, reveals that r~cu-PA-Gly'~~ and r s c u -P A -G l~'~~ activate plasminogen following Michaelis-Menten kinetics with a catalytic rate constant (b) which is about 5-times higher than that of rscu-PA but with an approximately 100-fold higher Michaelis constant (K,,,). The catalytic efficiency (k2/K,,,) of both mutants, therefore, is about 20-fold lower than that of wild type rscu-PA. This probably accounts for their lack of plasminogen activation at physiological concentrations of the substrate.
From these results we conclude that generation of tcu-PA is not a prerequisite for the activation of plasminogen by scu-PA. The positively charged Lys residue at position 158, however, seems to be important for the high affinity of scu-PA for plasminogen. Indeed, substitution of this Lys residue in scu-PA by a neutral (Gly) or negatively charged (Glu) amino acid reduces its affinity about 100-fold.
The K,,, of r~cu-PA-Gly"~ and rscu-PA-Glu'm for the activation of plasminogen in a purified system is similar to that reported for t-PA (65 p~) (47). Whereas the efficient and fibrin-specific plasminogen activation by t-PA in plasma has been explained by binding of both t-PA and plasminogen to fibrin, resulting in an increased stability of the Michaelis complex, the mechanism for the fibrin-specific thrombolytic properties of scu-PA is less well understood. The fibrinspecificity of scu-PA has been explained by the presence of a competitive inhibitor in human plasma, neutralizing scu-PA in the absence of fibrin (15). Alternatively, a conformational change in Glu-plasminogen upon binding to fibrin, rendering it more susceptible to activation, may contribute to the fibrin specificity of scu-PA (20,48).
In addition, it has been reported that scu-PA and t-PA act synergically on clot lysis in vivo both in animal models (49) and in humans (50), although no synergism can be demonstrated in an in vitro plasma clot lysis system (15,51). These observations indicate that as yet undefined regulatory interactions may modulate the fibrin-specific thrombolytic properties of scu-PA in vivo. Therefore, the findings of the present study do not a priori preclude that rscu-PA-Gly15' or rscu-P A -G~U '~~ might induce fibrin-specific thrombolysis in uivo.