Factor VIII C1 Domain Residues 2092–2093 Participate in Membrane Binding

Background: Activated factor VIII (FVIIIa) assembles with factor IXa (FIXa) on the membranes of activated platelets and on synthetic phosphatidylserine(PS)-containing membranes. This membrane-bound complex catalyses the conversion of the zymogen, factor X, to factor Xa. Established membrane-binding amino acids of FVIII are in the C2 domain. However, the C2 domain alone binds with a lower affinity to membranes than intact FVIII or the FVIII light chain, suggesting a role of the A3 and/or C1 domains in membrane binding. Moreover, a possible role for the C1 domain in platelet binding has been recently reported. The position of the C domains in the FVIII crystal structure suggests a possible role for residues from surface loop K2092-S2094 of the C1 domain in membrane binding. The present study addresses the role of this loop in membrane binding.

Methods: The role of the C1 surface loop K2092-S2094 was assessed by competition studies using KM33. This is a scFv fragment cloned from the antibody repertoire of a hemophilia A patient, with an epitope that comprises residues 2092–2094. In addition, FVIII mutants incorporating yellow fluorescent protein in place of the B domain and with K2092/F2093 changed to alanine (FVIII_YFP and FVIII_{YFP K2092A/F2093A}) were expressed and purified. Binding of recombinant FVIII labelled with fluorescein-maleimide (FVIII_fl), FVIII_YFP and FVIII_{YFP K2092A/F2093A} to phospholipid membranes (4% or 15% PS/20% PE/PC as balance) supported by glass microspheres and purified platelets was measured by flow cytometry. Lower affinity, non-equilibrium binding of sonicated vesicles to immobilized factor VIII was measured in a microtiter plate assay. The cofactor function of FVIII was measured in a factor Xase assay with limiting phospholipid.

Results: KM33 inhibited >95% of FVIII binding to phospholipid membranes containing 15% PS, indicating that the C1 domain epitope is important for membrane binding. The affinity of FVIII_{YFP K2092A/F2093A} for the same membranes was reduced 3-fold compared with FVIII_YFP (Kd’s of 91 ± 6 vs. 31 ± 2 nM). KM33 decreased the overall activity for the factor Xase complex by 95% on vesicles with 15% PS and >99% on vesicles with 4% PS. The implied membrane affinity for FVIII_{YFP K2092A/F2093A} in the factor Xase complex was decreased 3-fold for vesicles with 15% PS but the Vmax was equivalent to FVIII_YFP. The implied affinity of FVIII_{YFP K2092A/F2093A} was reduced approximately 40-fold for vesicles with 4% PS confirming the importance of the C1 domain epitope for full factor VIII function. In the microtiter assay, mAb BO2C11, against the C2 domain, blocked approx. 80% of binding to vesicles containing 15% PS, KM33 blocked 5% of binding and both antibodies together blocked ~95% of binding. Binding to 4% PS vesicles was inhibited 70% by KM33 alone and B02C11 alone blocked all binding. Thus, the two membrane-binding motifs are required for detectable binding to membranes with 4% PS but can independently support some binding to membranes with 15% PS. KM33 inhibited approx 90% of FVIII binding to platelets. The binding of FVIII_{YFP K2092A/F2093A} to platelets stimulated with calcium ionophore A23187 was reduced 50% compared to FVIII_YFP; however binding to platelets stimulated with thrombin receptor activating peptide (TRAP) was comparable to FVIII_YFP. The cofactor function of FVIII_{YFP K2092A/F2093A} was reduced approximately 80% on platelets stimulated with either TRAP or A23187.

Conclusion: The present study demonstrates that the FVIII C1 domain contributes to membrane binding and residues K2092 and/or F2093 participate in this interaction. The relative importance of these residues for membrane binding is dependent on the amount of PS present in synthetic membranes. On platelets K2092 and/or F2093 are necessary for full cofactor function of FVIII.