ELUCIDATION OF STRUCTURAL REQUIREMENTS ON PLASMINOGEN-ACTIVATOR INHIBITOR-1 FOR BINDING TO HEPARIN

Plasminogen activator inhibitor 1 (PAI-1), a member of the serpin superfamily of proteins, has been demonstrated previously to interact functionally with the glycosaminoglycan heparin (Ehrlich, H. J., Keijer, J., Preissner, K. T., Klein Gebbink, R., and Pannekoek, H. (1991) Biochemistry 30, 1021-1028). Heparin specifically enhances the rate of association between PAI-1 and thrombin about 2 orders of magnitude, whereas no effect is detected with other serine proteases (e.g. factor Xa). For the heparin-dependent serpins antithrombin III and heparin cofactor II, basic amino acid residues in and around the helix D subdomain were proposed to be involved in the binding of glycosaminoglycans. Here we employed site-directed mutagenesis of full-length PAI-1 cDNA to identify the amino acid residues that mediate heparin binding. To that end, 15 single-point mutants of PAI-1, each having individual arginyl, lysyl, or histidyl residues replaced by a neutral (alanyl) residue ("ala-scan"), and one double mutant were constructed, expressed in Escherichia coli, and purified to apparent homogeneity. The purified biologically active proteins were subjected to the following analyses: (i) heparin-dependent inhibition of thrombin; (ii) heparin-dependent formation of sodium dodecyl sulfate-stable complexes with thrombin; and (iii) binding to and elution from heparin-Sepharose. Based on the data presented, we propose that the amino acid residues Lys65, Lys69, Arg76, Lys80 and Lys88 constitute major determinants for heparin binding of PAI-1. These residues are located in and around the helix D domain and are conserved in the other heparin-dependent thrombin inhibitors, antithrombin III and heparin cofactor II.