The objective of this project was to determine how an allosteric switch in antithrombin that is activated by the cofactor, heparin, is regulated by the interaction of the reactive center loop hinge P14 residue with a binding pocket in sheet A. To accomplish this objective, we employed a saturation mutagenesis approach wherein a series of antithrombin variants were engineered in which 14 of the 19 possible alternative P14 residues were substituted for the wild-type serine residue.

Recombinant antithrombins mutated in the P14 residue were expressed in baculovirus-infected insect cells and purified by heparin-Sepharose chromatography. Wild-type and variant antithrombins were compared by analyzing the kinetics of inhibition of factor Xa (FXa) and thrombin with and without heparin, the kinetics and affinity of heparin binding and the protein fluorescence changes induced by this binding, titrations of inhibition stoichiometry, and the thermostability.

Fourteen mutants were expressed and characterized. The inhibition stoichiometries, heparin affinities, and basal FXa inhibition rates were increased and melting temperatures decreased by charged, polar, and large nonpolar residues relative to wild-type serine, indicative of varying extents of conformational activation of antithrombin without heparin. Such variants showed altered kinetics of heparin binding, indicating heparin binding to both native and activated forms in a preequilibrium with activation of the former through an induced conformational activation mechanism. By contrast, small nonpolar residues did not greatly affect inhibition stoichiometry and showed a decreased heparin affinity and lower basal FXa inhibition rates, indicating stabilization of the native state reactive loop-sheet A interaction. The kinetics of heparin binding to such mutants reflected an induced conformational activation mechanism. Heparin normalized the overall kinetics of protease inhibition by mutant and wild-type antithrombins despite variable inhibition stoichiometries, consistent with heparin inducing similar activated states in which the reactive loop was expelled from sheet A.

Our studies establish the size and hydrophobicity requirements for optimal interaction of the P14 residue of antithrombin with the sheet A binding pocket and demonstrate that some P14 residue mutations conform to a two-state conformational equilibrium mechanism whereas other suggest that activated states are possible.