Abstract
Blood coagulation factor V (FV) has little or no procoagulant activity and is activated by thrombin (IIa) to FVa following proteolytic removal of a central glycosylated B-domain (residues 710-1545; 836 a.a.). Discrete proteolysis is necessary for activation; however, recent data indicates that it is incidental to the mechanism by which cofactor function is realized. It was found that shortening the B-domain in the absence of bond cleavage to 155 amino acids also yielded an active cofactor form that assembles and functions in prothrombinase. These observations support the conclusion that macromolecular binding sites, which govern the eventual function of FVa, are concealed on FV by B-domain sequences. At present, it is unknown which elements of the B-domain contribute to maintaining the procofactor form. To begin to investigate this, we have designed eight recombinant derivatives of FV with variable amounts of the B-domain: FV-810 (des811–1491), FV-866 (des867–1491), FV-902 (des903–1491), FV-956 (des957–1491), FV-1033 (des1034–1491), FV-1053 (des1054–1491), FV-1106 (des1107–1491), and FV-1152 (des1153–1491). These constructs were stably expressed in BHK cells and purified to homogeneity in high yield (1–2 mg/L of media). The proteins migrated as single bands on SDS-PAGE, and were completely processed to FVa following incubation with IIa. Using either a purified prothrombinase assay with limiting amounts of (pro)cofactor or a one-stage PT-based clotting assay, we found that most of the derivatives were grouped into two categories: procofactor-like and cofactor-like. Factor V-810, FV-866, and FV-902 exhibited activity profiles consistent with the cofactor-like form and, as expected, bound with high affinity to FXa-membranes (Kd = 0.2–0.4 nM; n = 1). In contrast, FV-1033, FV-1053, FV-1106, and FV-1152 had very little activity and exhibited functional properties consistent with the procofactor-like form. Direct binding fluorescent measurements revealed that unlike FV, these constructs did appear to bind FXa-membranes, albeit with a much lower affinity compared to FVa. Interestingly, an outlier to either of the groups was FV-956 (301 a.a. B-domain). This derivative bound FXa-membranes with moderate affinity (Kd = 5 nM; n = 1) and it exhibited functional properties intermediate between the procofactor and cofactor forms in both the PT-based clotting and purified component assays. Control experiments indicated that all of the derivatives exhibited full cofactor activity and high affinity binding to FXa-membranes following treatment with IIa. Collectively, these data demonstrate that the single-chain cofactor-like forms can be shifted to the procofactor-like form by relatively small, systematic increases in the length of the B-domain. This transition was accompanied by a marked decrease in functional activity and FXa binding. Surprisingly, these data also show that elimination of >50% of the B-domain (1034–1491; 458 out of 836 a.a. deleted), a region that contains 31, nine amino acid tandem repeats and 16 out of 25 potential N-linked carbohydrate sites, has little, if any, influence on maintaining the procofactor form. Taken together, our data are consistent with the interpretation that a length of at least 375 amino acids or specific sequences contained within residues 902–1033 is sufficient to suppress cofactor activity of single chain FV.
Author notes
Corresponding author