ABSTRACT: Alpha-1 antitrypsin (AAT, also known as alpha-1 proteinase inhibitor or SERPINA1) is the most abundant member of the serpin superfamily found in human plasma. The naturally occurring variant AAT M358R, altered at the P1 position of the critical reactive center loop (RCL), is re-directed away from inhibition of AAT's chief natural target, neutrophil elastase, and toward accelerated inhibition of thrombin (FIIa), kallikrein (Kal), and other proteases such as factor XIa (FXIa). FXIa is an emerging target for the development of antithrombotic agents, since patients with FXI deficiency are protected from thromboembolic disease and do not exhibit a strong bleeding tendency. Previously, we used phage display, bacterial lysate screening, and combinatorial mutagenesis to identify AAT-RC, an engineered AAT M358R with additional changes between RCL positions P7-P3', CLEVEPR-STE [with changes bolded and the P1-P1' (R358-S359) reactive center shown as R-S]. AAT-RC was 279- and 16-fold more selective for FXIa/IIa or FXIa/Kal than AAT M358R; the increased selectivity came at a cost of a 2.3-fold decrease in the rate of FXIa inhibition and a 3.3-fold increase in the stoichiometry of inhibition (SI). Here, we asked which alterations in AAT-RC were most important for the observed increases in selectivity for FXIa inhibition. We back-mutated AAT-RC to AAT-RC-1 (P7-P3' FLEVEPRSTE), AAT-RC-2 (P7-P3' FLEAEPRSTE), and AAT RC-3 (P7-P3' FLEAIPR-STE). Proteins were expressed as cleavable, hexahistidine-tagged glutathione sulfotransferase fusion proteins in E. coli and purified by proteolytic elution from glutathione agarose, with polishing on nickel chelate agarose. Selectivity for FXIa over Kal of AAT-RC-1, -2, and -3 was 14, 21, and 2.3, respectively. AAT-RC-2 inhibited FXIa 31% more rapidly than AAT M358R, with the same SI, and enhanced selectivity for FXIa over Kal, FXa, FXIIa, activated protein C, and FIIa of 25-, 130-, 420-, 440-, and 470-fold, respectively. Structural modeling of the AAT-RC-2/FXIa encounter complex suggested that both E (Glu) substitutions at P3 and P3' may promote FXIa binding via hydrogen bonding to K192 in FXIa. AAT-RC-2 is the most selective and active AAT variant reported to date for FXIa inhibition and will be tested in animal models of thrombosis and bleeding.
