Abstract
Therapeutic heparin concentrations selectively inhibit the intrinsic tenase complex in an antithrombin-independent manner. To define the molecular target and mechanism for this inhibition, recombinant human factor IX (FIX) with alanine substituted for solvent-exposed basic residues (H92, R170, R233, K241) in the protease domain was expressed in HEK 293 cells, activated by FXIa, and characterized with regard to enzymatic activity, heparin affinity, and inhibition by low molecular weight heparin (LMWH). The recombinant FIX proteins were purified to homogeneity by SDS-PAGE analysis and exhibited indistinguishable chromatographic behavior when eluted from a Resource Q column with a calcium gradient. FIX was activated with human FXIa (150:1 molar ratio) at 4 °C for 2–6 hr, incubated with anti-FXIa polyclonal antisera crosslinked to Affi-gel, and FIXa active sites were determined by a modified antithrombin III titration. These mutations had only modest effects on chromogenic substrate hydrolysis and the kinetics of factor X activation by FIXa. The kcat/KM for factor X activation by FIXa-phospholipid (5 nM FIXa, 50 μM PC:PS vesicles, 30% ethylene glycol) was similar for all recombinant proteases except FIXa R233A, which was modestly lower due to a 1.4-fold increase in KM(app). In a functional binding assay, FIXa H92A and K241A exhibited apparent FIXa-FVIIIa affinity similar to FIXa wild type (WT) (KD(app)= 2.2, 1.9, and 1.7 nM, respectively). FIXa R170A had markedly increased FIXa-FVIIIa affinity (KD(app)= 0.4 nM), and consistent with previous results, dramatically increased coagulant activity (372%) relative to FIXa WT (J. Chang et al. JBC, 1998). FIXa R233A had significantly reduced cofactor affinity (KD(app)= 4.4 nM) and coagulant activity (59%). FIXa R233A also had reduced ability to stabilize the in vitro half-life of FVIIIa relative to FIXa WT (2-fold faster degradation), even at increased protease concentration. Thus, this mutation disrupts interaction with the A2 domain, suggesting that this critical cofactor interactive site extends from the established c165–170 α-helix to the proximal portion of the C-terminus α-helix on FIXa. Using KD(app) to calculate the concentration of FIXa-FVIIIa, no significant differences in the kcat/KM for factor X activation by FIXa-FVIIIa-phospholipid were observed between proteases(1–2 nM FVIIIa, 0.1 nM FIXa, 50 μM PC:PS vesicles). Relative heparin affinity of FIXa was assessed by direct and competition binding to immobilized LMWH, detected by surface plasmon resonance. Mutant FIXa (250 nM) demonstrated moderately reduced (H92A, R170A, K241A) or minimal (R233A) binding to immobilized LMWH, relative to FIXa WT. Solution competition demonstrated that the EC50 for LMWH was increased less than 2-fold for FIXa H92A and K241A, but over 3.5-fold for FIXa R170A, indicating that relative heparin affinity was WT>H92A/K241A>R170A>>R233A. Kinetic analysis of the inhibition of factor X activation by intrinsic tenase demonstrated that relative affinity for LMWH was WT>K241A>H92A> R170A>>R233A, correlating with heparin affinity. Notably, FIXa R233A demonstrated minimal binding to immobilized LMWH, and almost complete resistance to inhibition by LMWH in the intrinsic tenase complex. Thus, LMWH inhibits intrinsic tenase by interacting with the heparin-binding exosite on the FIXa protease domain. The extensive overlap between heparin and FVIIIa interactive sites on the protease domain suggests that oligosaccharide disrupts critical interactions with the cofactor A2 domain.
Author notes
Corresponding author