THE CONSTRUCTION AND EXPRESSION OF CHIMERIC UROKINASE-TYPE PLASMINOGEN-ACTIVATOR GENES CONTAINING KRINGLE DOMAINS OF HUMAN PLASMINOGEN

A series of chimeric urokinase-type plasminogen activator (uPA) genes, which contain combinations of kringle domains of human plasminogen (HPg) in place of the uPA kringle (K(uPA)), has been constructed and expressed. Some of the resulting recombinant (r) variant uPA chimeras contain modules that potentially mediate the macroscopic binding of HPg to its activation effectors, fibrin(ogen) and 6-aminohexanoic acid (EACA). Such binding sites are not possessed by K(uPA), but are present in certain of the HPg kringles, viz., kringle 1 (K1(HPg)), kringle 4 (K4(HPg)), and kringle 5 (K5(HPg)). The recombinant (r) chimeras constructed included molecules with replacements of K(uPA) with K1(HPg) (r-[K(uPA) → K1(HPg)]uPA), and with K(uPA) replaced by double kringle combinations of K1(HPg)K4(HPg)(r-[K(uPA) → K1(HPg)K4(HPg)]uPA), K2(HPg)K3(HPg)(r-[K(uPA) → K2(HPg)K3(HPg)]uPA), and K4(HPg)K5(HPg) (r-[K(uPA) → K4(HPg)K5(HPg)]uPA). All of these variant genes, along with their wild-type (wt) r-uPA counterparts, were expressed in human kidney 293 cells. In cases wherein EACA-binding kringles from HPg have been placed in uPA, this property has been retained in the chimeric molecule and employed as an essential part of the purification procedures for the variants. The steady state amidolytic activity of two-chain (tc) wtr-uPA toward the chromogenic substrate, H-D-pyroglutamyl-Gly-L-Arg-p-nitroanilide (S2444), is characterized by a k(cat)/K(M) (pH 7.4, 37°C) of 120 s-1 mM—1. This value ranges from 92 s-1 mM-1 (ter-[K(uPA) → K1(HPg)]uPA) to 166 s-1 mM-1 (ter-[K(uPA) → K1(HPg)K4(HPg)]uPA) for each of the variants, demonstrating that the catalytic efficiency of the active site is altered only in a small way by changes in the noncatalytic domain of uPA. Small differences are also observed in the abilities of these ter variants to interact with the fast-acting plasma inhibitor of uPA, viz., plasminogen activator inhibitor-1 (PAI-1). The second-order rate constant for the interaction of PAI-1 with ter-uPA, 0.46 x 107 M-1s-1 (pH 7.4, 10°C), ranges from 0.29 x 107 M-1s-1(ter-[K(uPA) → K1(HPg)K4(HPg)]uPA) to 1.08 x 107 M-1s-1 (ter-[K(uPA) → K4(HPg)K5(HPg)]uPA), for the ter-chimeric variants. Neither wtr-uPA nor any of its chimeric r-variants interacted macroscopically with a fibrin clot under conditions that allowed binding of 74% of single-chain r-tissue-type plasminogen activator. However, the ter- chimeric uPA variants provided HPg-enriched clot lysis times between 0.2 (r- [K(uPA) → K1(HPg)K4(HPg)]uPA) and 2.4 (r-[K(uPA) → K2(HPg)K3(HPg)]uPA) relative to that of wtr-uPA. These results demonstrate that no perceptible relationship exists between macroscopic fibrin binding and the clot lytic properties of the variants. In addition, under conditions wherein the HPg activation by tissue-type plasminogen activator was significantly stimulated by fibrinogen, fibrin, and CNBr fragments of fibrinogen, this same activation was not similarly stimulated by these effectors when wtr-uPA or any of the chimeric variants were employed as the activators. These results demonstrate that despite the retention of the EACA binding properties of K1(HPg), K1(HPg)K4(HPg), and K4(HPg)K5(HPg), a factor that supports the independent domainal character of HPg kringle domains, neither fibrin-binding capabilities nor fibrin-selectivity for HPg activation was engineered into uPA when the domains were inserted into uPA in place of K(uPA). © 1993 Academic Press, Inc.