Structure of the non-covalent complex of prothrombin kringle 2 with PPACK-thrombin.

Abstract Prothrombin fragment 2 (the second kringle) has been co-crystallized with PPACK (D-Phe-Pro-Arg)-thrombin and the structure of the non-covalent complex has been determined and refined ( R = 0.16) at 3.2 A resolution using X-ray crystallographic methods. The kringles interact with thrombin at a site that has previously been proposed to be the heparin binding region. The latter is a highly electropositive surface near the C-terminal helix of thrombin abundant in arginine and lysine residues. These form salt bridges with acidic side chains of kringle 2. Somewhat unexpectedly, the negative groups of the kringle correspond to an enlarged anionic center of the lysine binding kringles such as plasminogen K1 and K4 and TPA K2. The anionic motif is DGDEE in prothrombin kringle 2. The corresponding cationic center of the lysine binding site region has an unfavorable Arg71Phe substitution but Lys35 is conserved. However, the folding of fragment 2 is different from that of prothrombin kringle 1 and other kringles: the second outer loop possesses a distorted two-turn helix and the hairpin β-turn of the second inner loop pivots at V64 and D70 by 60°. The Lys35 is located on a turn of the helix, which causes it to project into solvent space in the fragment 2-thrombin complex, thereby devastating the cationic center of the lysine binding site. Since fragment 2 has not been reported to bind lysine, it most likely has a different inherent folding conformation for the second outer loop, while the movement of the V64-D70 β-turn is most likely a conformational change accompanying complexation, both of which reveal a new heretofore unsuspected flexibility in kringles. The fragment 2-thrombin complex is the first cassette module-catalytic domain interaction determined for a multi-domain blood protein and only the second domain-domain interaction to be described among such proteins, the other being Ca +2 prothrombin fragment 1 (Gladomain and kringle 1). It has been possible to propose a reasonable model for the four domain prothrombin structure using these two double domain structures.