An Ester Bond Linking a Fragment of a Serine Proteinase to Its Serpin Inhibitor † Rikke Egelund, ‡ Kees W. Rodenburg, ‡ Peter A. Andreasen,* ,‡ Margit S. Rasmussen, § Roy E. Guldberg, § and Torben E. Petersen § Laboratory of Cellular Protein Science and Protein Chemistry Laboratory, Department of Molecular and Structural Biology, UniVersity of Aarhus, 10 GustaV Wied’s Vej, 8000 C Aarhus, Denmark ReceiVed December 11, 1997; ReVised Manuscript ReceiVed February 25, 1998 ABSTRACT: Most known members of the serpin superfamily are serine proteinase inhibitors. Serpins are therefore important regulators of blood coagulation, complement activation, fibrinolysis, and turnover of extracellular matrix. Serpins form SDS-resistant complexes of 1:1 stoichiometry with their target proteinases by reaction of their P 1 -P 1 ′ peptide bond with the active site of the proteinases. The nature of the interactions responsible for the high stability of the complexes is a controversial issue. We subjected the complex between the serine proteinase urokinase-type plasminogen activator (uPA) and the serpin plasminogen activator inhibitor-1 (PAI-1) to proteolytic digestion under nondenaturing conditions. The complex could be degraded to a fragment containing two disulfide-linked peptides from uPA, one of which included the active site Ser, while PAI-1 was left undegraded. By further proteolytic digestion after denaturation and reduction, it was also possible to degrade the PAI-1 moiety, and we isolated a fragment containing 10 amino acids from uPA, encompassing the active site Ser, and 6 amino acids from PAI-1, including the P 1 Arg. Characterization of the fragment gave results fully in agreement with the hypothesis that it contained an ester bond between the hydroxyl group of the active site Ser and the carboxyl group of the P 1 Arg. These data for the first time provide direct evidence that serine proteinases are entrapped at an acyl intermediate stage in serine proteinase-serpin complexes. The serpins constitute a family of extracellular glycopro- teins from animals, plants, and viruses. Most known mem- bers of the serpin superfamily are serine proteinase inhibitors (1). The first amino acid sequence of a serpin serine pro- teinase inhibitor to be reported was that of antithrombin III (2). The sequence showed no similarity to the sequences of the by that time well-known small “standard-mechanism” inhibitors. Therefore, the sequence alone gave no clues with respect to the inhibition mechanism. Since then, biochemical studies have shown that serpin serine proteinase inhibitors form SDS-resistant complexes with a 1:1 stoichiomtetry with their target proteinases by reaction of their P 1 -P 1 ′ bond (reactive center peptide bond) with the active site of the proteinases (1). The X-ray crystal structures of serpins have shown that they are globular proteins with nine R-helices and three -sheets. The P 1 -P 1 ′ bond is localized in a surface-exposed, approximately 20 amino acid long peptide loop, the reactive center peptide loop (the RCL). The RCL is linked C-terminally to strand 1 of -sheet C and N- terminally to strand 5 of -sheet A (3). Serpins are able to undergo large conformational changes, of which the most conspicuous one is brought about by proteolytic cleavage of the RCL, either with nontarget proteinases or by cleavage of P 1 -P 1 ′ at in vitro dissociation of serpin-target proteinase complexes. The cleavage leads to insertion of the part of the RCL N-terminal to the cleavage site as strand 4 in -sheet A. Cleaved serpins have a higher thermodynamic stability than native serpins (3). At least partial strand insertion also occurs during complex formation (4). However, the three- dimensional structure of a serine proteinase-serpin complex has not yet been determined, and despite many biochemical studies (1), the state of the P 1 -P 1 ′ bond in the complex has remained controversial. We have now employed a direct protein chemical approach for characterization of the association between the active site of the proteinase and the P 1 -P 1 ′ bond of the serpin. MATERIALS AND METHODS Materials. Human uPA 1 was purchased from Wakamoto Pharmaceutical Co. (Tokyo, Japan). Anhydro-uPA was prepared as described by Wun et al. (5). Latent human PAI-1 was purified from serum-free conditioned medium of dexa- methasone-treated HT-1080 cells (6, 7). Latent PAI-1 was converted into the active conformation by denaturation with guanidinium chloride, followed by extensive dialysis against 10 mM NaH 2 PO 4 , pH 7.4, 140 mM NaCl (7). To produce PAI-1 in complex with uPA, guanidinium chloride-activated PAI-1 was incubated with an equal amount of uPA for 90 min at 37 °C in 10 mM NaH 2 PO 4 , pH 7.4, 140 mM NaCl. The complex was separated from noncomplexed uPA and PAI-1 by immunoaffinity chromatography on two different Sepharose-4B columns, one coupled with an anti-human † This work was supported by the Danish Cancer Society, the Danish Medical Research Council, the NOVO-Nordisk Foundation, FELFO, and the Danish Biotechnology Program. * Corresponding author. Telephone: +45 8942 5080. Fax: +45 8612 3178. E-mail: pa@mbio.aau.dk. ‡ Laboratory of Cellular Protein Science. § Protein Chemistry Laboratory. 1 Abbreviations: PAI-1, type-1 plasminogen activator inhibitor; RCL, reactive center loop; uPA, urokinase-type plasminogen activator. 6375 Biochemistry 1998, 37, 6375-6379 S0006-2960(97)03043-2 CCC: $15.00 © 1998 American Chemical Society Published on Web 04/15/1998