BackgroundGla-domainless factor (F) Xa (GD-FXa) was proposed as a trap to endogenous anticoagulant tissue factor pathway inhibitor (TFPI) to restore thrombin generation in hemophilia. Using computational chemistry and experimental approaches, we previously showed that S195A GD-FXa also binds TFPI and restores ex vivo coagulation in plasma obtained from person(s) with hemophilia. MethodsTo design a GD-FXa variant with improved anti-TFPI affinity, we performed molecular dynamics simulations and identified suitable sites for mutagenesis. The calculations identified residues R150FXa and K96Fxa as cold-spots of interaction between GD-FXa and the K2 domain of TFPI. In the three-dimensional model, both residues face toward TFPI hydrophobic residues and are thus potential candidates for mutagenesis into hydrophobic residues to favor an improved protein–protein interaction. ResultsCatalytically inactive GD-FXa variants containing the S195A mutation and the additional mutations K96Y, R150I, R150G, R150F, and K96YR150F, were produced to experimentally confirm these computational hypotheses. Among these mutants, the R150FFXa and the K96YR150FFXa were slightly more effective than S195A GD-FXa in restoring coagulation in FVIII deficient plasmas. However, in surface plasmon resonance experiments, they showed TFPI binding affinities in the same range and acted similarly as S195A GD-FXa in FXa/TFPI competition assays. In contrast, the R150 mutants completely lost their interactions with antithrombin as observed in the surface plasmon resonance experiments. ConclusionsWe therefore conclude that their antithrombin resistance is responsible for their improved thrombin generation, through an extension of their half-lives.