Membrane binding constitutes a key step in the activation of coagulation factors. This process largely hinges on specialized “floating” domains, most prominently the GLA and the C1/C2 domains found in almost all coagulation proteins. Currently, the details of how these domains interact with the membrane are largely unknown, due primarily to the lack of atomic models describing the proteins' membrane-bound forms. In sharp contrast to integral membrane proteins, studying insertion of membrane floating proteins has proven extremely challenging, mainly due to the unknown depth of membrane penetration. Although several computational studies have attempted to study the process using conventional models of the membrane, these methods are often prohibitively costly and inefficient due, in part, to the slow diffusion of the lipid molecules, which is required to accommodate the insertion of the floating domain. Here, we present a study on the membrane-bound forms of several important hemostasis factors, namely, human factor IX (hfIX), human factor × (hfX), human protein C (hPrC), human protein S (hPrS), and human protein Z (hPrZ), which utilize the GLA domain for membrane binding, as well as human factor V (hfV) and human factor VIII (hfVIII) in which the C1/C2 domain plays this role. To make the binding event accessible to molecular dynamics (MD) simulations, we have developed a novel membrane mimetic system, composed of a highly mobile hydrophobic core and explicit head groups that allow one to study specific interactions between the floating domains and lipid molecules. These interactions determine not only the specificity of binding to certain biological membranes, but also the molecular basis of the variable binding affinity of coagulation factors. These two factors are of utmost importance in physiology of blood coagulation, in designing mutations of altered affinity, and in developing novel pharmacological agents for thrombotic disorders. The floating domains were placed in the membrane-mimetic system and simulated under physiological conditions. Given the efficiency of the method, we have been able to capture spontaneous binding of these coagulation factors to the membrane (Figure 1) multiple times, despite starting from various initial orientations ensuring that each coagulation factor converged to the same final, membrane-bound structure over multiple trials. From these simulations, we have been able to elucidate the residues important for association with the membrane, as well as several unique phosphatidylserine (PS) specific binding sites in hfIX, hfX, hPrC, hPrS, and hPrZ. Using the membrane-bound structures of these coagulation factors together with free energy calculations, we have been able to elucidate important differences between the binding of high-affinity and low-affinity coagulation factors to the membrane.
No relevant conflicts of interest to declare.
Asterisk with author names denotes non-ASH members.