Porcine models are, among other animal models, very suitable for in vivo investigations in the vascular field especially with respect to the possible relationship between atherosclerosis and thrombosis. In order to use this model to define the in vivo role of PAI-1, the characterization of porcine PAI-1 and its availability for the generation of immunological tools are a prerequisite. Porcine plasminogen activator inhibitor-1 (poPAI-1) cDNA was isolated from a cDNA library prepared from cultured porcine aortic cells and characterized in comparison with PAI-1 cDNA's from other species including human, bovine, rabbit, rat and murine. Subsequently the DNA sequence coding the mature protein was cloned into an appropriate vector for expression in Escherichia coli and recombinant porcine PAI-1 was purified and characterized. On SDS-PAGE the apparent molecular weight was estimated to be 45 kDa, identical to the molecular weight of human PAI-1. The purified recombinant porcine PAI-1 (rpoPAI-1) had a specific activity of 508,800 +/- 800 U/mg (mean +/- SD, n = 3) towards human tissue-type plasminogen activator (ht-PA) and a functional half-life in vitro of 2.1 +/- 0.8 h (n = 3). Incubation with a two fold molar excess of ht-PA (n = 3) or human urokinase-type plasminogen activator (hu-PA, n = 2) followed by analysis by SDS-PAGE revealed reaction products corresponding to active (71 +/- 7% resp. 96 +/- 3.6%), latent (12 +/- 0.4% resp. 2.6 +/- 2.4%) and substrate (16.6 +/- 6.8% resp. 1.5 +/- 1.3) forms. Inactivated samples of porcine PAI-1 could be reactivated with guanidinium chloride up to 52% of its original specific activity towards t-PA and u-PA. The second order rate constant of inhibition of ht-PA was 1.64 +/- 0.37 10(7)M-1 s-1 (n = 9). In gel filtration rpoPAI-1 in buffer eluted at a volume corresponding to 24 kDa, whereas in the presence of porcine plasma, the molecular form containing PAI-1 activity eluted at a volume corresponding to 330 kDa, presumably as a consequence of binding of active PAI-1 to vitronectin. Taken together, these data demonstrate that no obvious functional differences exist between human and porcine PAI-1.