Structure-function studies of plasminogen activator inhibitor-1

Sammanfattning: Plasminogen activator inhibitor-I (PAI-1) is an important physiological inhibitor of tissue-type plasminogen activator (tPA) and urine-type plasminogen activator (uPA). High plasma levels of PAI-1 correlate to thrombotic disease, and low levels can be connected to a bleeding tendency. The native PAI-1 molecule is functionally active, but it spontaneously transforms into an inactive "latent" form with a halflife of about 2 h at physiological conditions. To investigate the relationship between structure and function of PAI-1, we used site-directed mutagenesis to produce more than forty PAI-1 mutants. Most of them carry a single amino acid change and are from three different regions in the PAI-1 molecule. These PAI-1 mutants, as well as wild type PAI-1 (wtPAI-1), have been expressed in an E. coli expression system and the mutant proteins were purified by heparin-Sepharose and anhydrotrypsin agarose chromatographies. Using this method, active PAI-1 proteins have been purified and subsequently characterized regarding PAI-1 inhibitory activity, interactions with tPA and vitronectin, and stability. The transformation of active PAI-1 to its latent form is accompanied by the insertion of the reactive center loop (RCL) into the A ß-sheet of the PAI-1 molecule. The study of the Phe113 - Asp138 Stretch indicated that, even if this stretch is located at the opposite end of RCL on the PAI-1 molecule, it is still important for functional activity of PAI-1. This stretch seems to be involved in PAI-1 stability (Asp125, Arg133), substrate behavior when incubated with tPA (Asp125, Phe126, Arg133) and heparin binding (Arg115, Arg118). In the three-dimensional structure, the B ß-sheet in PAI-1 I is located beneath the A ß- sheet, and the s2B and s3B strands are situated in the vicinity where RCL is believed to be inserted. Therefore, the role of the amino acids in the s2B and s3B strands for PAI-1 stability was extensively studied. Mutations of the residues in the central portions of both strands caused a significant decrease in stability with half-lives of about 10 - 25% as compared with that of wtPAI-1. However, mutations at both sides of the central portion in each of the two strands frequently resulted in an increased PAI-1 I stability, by a factor up to 7-fold. This demonstrated that the residues on s2B and s3B are of major importance for PAI-1 stability, most likely by affecting the reactive center loop insertion rate into the A ß-sheet of the molecule, either in a positive or a negative direction. The increased stability of active PAI-1 at slightly acidic pH suggests that one or more histidine residue(s) may be involved in stabilizing the active form. Our study indicated that His229 indeed might be involved in this process, since substitutions of this residue significantly decreased the pH dependence of PAI-1 I stability. Mutations at His143 gave results suggesting that this residue may also play a role regarding the pH dependence of PAI-1 I stability. Our data also suggested that Tyr221 and Tyr228 may interact with the histidine(s) since the mutants of these two tyrosines were very stable at neutral pH but less stable at acidic PH. Two patients with a bleeding tendency and low plasma PAI-1 concentrations were investigated for mutations in the PAI-1 structural gene. In both patients, a mutation in the PAI-1 gene (1334G -> A) was observed. This mutation caused an exchange of the alanine residue at -9 of the PAI-1 propeptide to threonine. However, it was demonstrated subsequently that this constituted a common polymorphism, with an allele frequency of about 15%. Thus, it is not likely that the polymorphism is directly involved in the cause of the bleeding tendency.