PF4 can potentially dimerize β2GPI, favoring the recognition of the complex by autoantibodies. (A) The binding of PF4 to β2GPI involves exclusively residues of chains a and c (shown as ball/sticks in the color of their parent chain). The molecular interaction interface of β2GPI is shown as transparent surface colored by its electrostatic potential. Markedly, no residues belonging to chains b (green) and d (cyan) are used for this interaction. (B) Chains b and d of PF4 tetramer can bind an additional (to that interacting with chains a and c) β2GPI molecule, leading to dimerization of the latter. Each β2GPI monomer is shown as ribbon in different color (orange and yellow). Its molecular surface is colored according to its electrostatic potential and displayed in wire representation (for the membrane binding regions) and as dotted surface for the rest of the molecule. (C) A plausible model for the mode of antibody binding to β2GPI dimer, which is formed on the interaction with the PF4 tetramer. The immunoglobulin structure, shown in light blue (heavy chains) and cyan (light chains), was adapted from the published structure by Padlan40 (based on PDB structures 2IG2 and 1FC2). White arrows indicate the epitopes that thrombosis-associated antibodies recognize on domain I of β2GPI.41 Notably, these epitopes are arranged in a geometry that is compatible with their readily and bivalent recognition by antibody's antigen-binding sites. Blue arrows indicate the positively charged regions in domain V of β2GPI, which are involved in interaction with negatively charged membranes. The interaction interface of PF4 is displayed as dots and is colored by its electrostatic potential.
