A phosphatidylserine binding site in factor Va C1 domain regulates both assembly and activity of the prothrombinase complex

Abstract

Tightly associated factor V(a) (FVa) and factor X(a) (FXa) serve as the essential prothrombin-activating complex that assembles on phosphatidylserine (PS)-containing platelet membranes during blood coagulation. We have previously shown that (1) a soluble form of PS (C6PS) triggers assembly of a fully active FVa-FXa complex in solution and (2) that 2 molecules of C6PS bind to FVa light chain with one occupying a site in the C2 domain. We expressed human factor V(a) (rFVa) with mutations in either the C1 domain (Y1956,L1957)A, the C2 domain (W2063,W2064)A, or both C domains (Y1956,L1957,W2063,W2064)A. Mutations in the C1 and C1-C2 domains of rFVa reduced the rate of activation of prothrombin to thrombin by FXa in the presence of 400 muM C6PS by 14 000- to 15 000-fold relative to either wild-type or C2 mutant factor rFVa. The K(d')s of FXa binding with rFVa (wild-type, C2 mutant, C1 mutant, and C1-C2 mutant) were 3, 4, 564, and 624 nM, respectively. Equilibrium dialysis experiments detected binding of 4, 3, and 2 molecules of C6PS to wild-type rFVa, C1-mutated, and C1,C2-mutated rFVa, respectively. Because FVa heavy chain binds 2 molecules of C6PS, we conclude that both C2 and C1 domains bind one C6PS, with binding to the C1 domain regulating prothrombinase complex assembly.

Figure 1

Amino acid residues in human FVa that contribute to PS binding. The structure of bovine FVa_i solved by Adams et al is shown as rendered by the Cn3D software package. Selected amino acids are labeled using the corresponding residues in human FVa. Amino acid residues previously implicated in FVa binding to PS membranes are highlighted in yellow. The indole side chains of Trp²⁰⁶³ and Trp²⁰⁶⁴, located in the FVa C2 domain, penetrate the lipid bilayer and contribute a majority of the free energy associated with FVa binding to PS membranes. The hydrophobic side chains of Tyr¹⁹⁵⁶ and Leu¹⁹⁵⁷ are located in a structurally analogous position within the FVa C1 domain and provide a modest contribution to the free energy associated with FVa binding to PS membranes. Glycosylation of Asn²¹⁸¹ (green) results in the FVa₁ glycoform. FVa₂ is not glycosylated at Asn²¹⁸¹.

Figure 2

Prothrombin activation in the presence of C6PS. Initial rates of human prothrombin activation were determined in reaction mixtures containing 50 mM Tris, pH 7.4, 150 mM NaCl, 5 mM CaCl₂, 1 nM human FXa, 5 nM human rFVa₂, 1 μM prothrombin, and different concentrations of C6PS as described in “Methods.” The rate of prothrombin activation was determined in the presence of 0 to 400 μM C6PS and rFVa₂ having either an intact (A, circles) or a mutant ²³C1 domain (A,B, triangles). Data obtained with rFVa₂ having an intact C2 domain are shown in red (wild-type, circles in A; C1 mutant, triangles in panel B), whereas data with a mutant C2 domain are shown in green (C2 mutants, circles in A; and C1-C2 mutant, triangles in panel B). The lines drawn through the data in panel A are hyperbola, corresponding to a simple, single-site binding model, with parameters K_d^app = 3.8 (± 0.5) μM and saturating activity 179 (± 7) nM/sec (wild-type) and K_d^app = 3.7 (± 0.8) μM and saturating activity 177 (± 12) nM/sec (C2 mutant). Hyperbolic fits to the data up to 50 μM of C6PS are shown as dashed lines in the inset to panel A, with one of these fits (wild-type) also plotted as a dashed line in panel A. The data for rFVa₂ with a mutated C1 domain were also well described by hyperbola (B), but with K_d^app's = 60 (± 4) μM and 56 (± 3) μM and saturating activities of 0.0079 (± 0.0003) and 0.0078 (± 0.0002) nM IIa/sec for the C1 and C1-C2 mutants, respectively.

Figure 3

Binding of C1 mutant human rFVa₂ to soluble C6PS. The intrinsic fluorescence intensities of 0.2 μM of C1 mutant rFVa₂ in 50 mM Tris, 150 mM NaCl, 5 mM CaCl₂, 0.6% polyethylene glycol, pH 7.5, was measured as a function of C6PS concentration at 24°C to follow C6PS binding. The data were analyzed according to a simple binding model as described in “Binding of C6PS to human rFVa₂,” with results shown in Table 1. Error bars represent SEM.

Figure 4

Binding of human rFVa₂ to DEGR-FXa in the presence of C6PS. Binding of factor rFVa₂ to DEGR-FXa was detected by the change in fluorescence intensity of DEGR, which is covalently bound to the active site of FXa. Small aliquots of wild-type rFVa₂, C2 mutant, C1 mutant, and C1-C2 mutant were added to DEGR-FXa (1 nM in 5 mM Ca²⁺, 50 mM Tris, 150 mM NaCl, 0.6% polyethylene glycol, 400 μM C6PS, pH 7.5). The lines through the data were obtained by least-squares regression to a simple single-site binding model with the best fit K_d values for wild-type rFVa₂, C2 mutant, C1 mutant, and C1-C2 mutant being 3.1 (± 0.4) nM, 4.0 (± 0.6) nM, 564 (± 35) nM, and 624 (± 40) nM, respectively.

Figure 5

Activity of human rFVa C1 and C2 domain mutants in the presence of C6PS. The cofactor activity of rFVa mutants was determined as described in the equation 1 in “Lipid-dependent cofactor activity.” Reaction mixtures contained 1 μM prothrombin, 1 nM FXa, 5 nM rFVa, 5 μM DAPA in 50 mM Tris, 150 mM NaCl, 5 mM Ca²⁺, pH 7.5. Shown are means (± SD) of 4 measurements. *P < .002 compared with wild-type rFVa.

Figure 6

Proposed model of prothrombinase assembly for wild-type human FVa, and for C1- and C2-mutated FVa. This model is based on the well-known fact that the free energy of binding of extrinsic membrane proteins to membranes often reflects multiple interactions, and on the results presented in this paper.