PB1138 - Computational structural analysis of coagulation Factor VIII maturation to form Xase complex

Background: The maturation process of coagulation faactor VIII (FVIII) during Xase complex formation with activated coagulation factor IX (FIXa) and coagulation factor X (FX) remains poorly understood.

Aims: This study employs advanced computational tools, specifically Alphafold 2, guided by machine learning and artificial intelligence, to predict the structural dynamics of FVIII during its maturation within the coagulation cascade.

Methods: Monomer structures of FVIII, FIXa, FX, and von Willebrand factor (VWF) were predicted with the Alphafold 2. Five models with the highest confidence score from each protein were selected for subsequent analysis. Alphafold 2-multimer analysis, RoseTTAFold, and DeepAssembly algorithms refined the structure of protein complexes, binding sites, and interactions among the proteins.

Results: B domain of FVIII shielded A1-A2-A3-C1-C2 domains in the naive FVIII structure. B domain removal by furin cleavage exposed C1, C2, and A3 domains and facilitated binding FVIII to VWF's D’D3 domains. To analyze the Xase complex, Alphafold2-multimer predicted 25 FX binding models without altering the FVIII structure. On the other hand, FIXa binding induced FVIII structural bending in some models, particularly in the A2 domain. Finally, we assembled serine protease domains of FIXa and FX in front of FVIII, based on the reported structural analysis with cryo-electron microscopy.

Conclusion(s): The employment of machine learning and artificial intelligence for the structural prediction of FVIII will deepen the understanding of structural dynamics and the role of FVIII in hemostasis. Furthermore, these approaches would be useful to understand