Warfarin alters vitamin K metabolism: a surprising mechanism of VKORC1 uncoupling necessitates an additional reductase

Abstract

The anticoagulant warfarin inhibits the vitamin K oxidoreductase (VKORC1), which generates vitamin K hydroquinone (KH₂) required for the carboxylation and consequent activation of vitamin K-dependent (VKD) proteins. VKORC1 produces KH₂ in 2 reactions: reduction of vitamin K epoxide (KO) to quinone (K), and then KH₂ Our dissection of full reduction vs the individual reactions revealed a surprising mechanism of warfarin inhibition. Warfarin inhibition of KO to K reduction and carboxylation that requires full reduction were compared in wild-type VKORC1 or mutants (Y139H, Y139F) that cause warfarin resistance. Carboxylation was much more strongly inhibited (∼400-fold) than KO reduction (two- to threefold). The K to KH₂ reaction was analyzed using low K concentrations that result from inhibition of KO to K. Carboxylation that required only K to KH₂ reduction was inhibited much less than observed with the KO substrate that requires full VKORC1 reduction (eg, 2.5-fold vs 70-fold, respectively, in cells expressing wild-type VKORC1 and factor IX). The results indicate that warfarin uncouples the 2 reactions that fully reduce KO. Uncoupling was revealed because a second activity, a warfarin-resistant quinone reductase, was not present. In contrast, 293 cells expressing factor IX and this reductase activity showed much less inhibition of carboxylation. This activity therefore appears to cooperate with VKORC1 to accomplish full KO reduction. Cooperation during warfarin therapy would have significant consequences, as VKD proteins function in numerous physiologies in many tissues, but may be poorly carboxylated and dysfunctional if the second activity is not ubiquitously expressed similar to VKORC1.

Graphical abstract

Figure 1.

The Y139F VKORC1 mutant fully reduces KO to KH₂to drive carboxylation. (A) Cycling between oxidized (KO) and reduced (KH₂) vitamin K drives VKD protein carboxylation by the γ-glutamyl carboxylase. VKORC1 fully reduces KO to the quinone intermediate (K) and then to KH₂; an unknown warfarin-resistant quinone reductase can also perform the second reaction. (B) Purified human VKORC1 bearing the Y139F mutation was tested for KO reduction to KH₂ by performing all steps under nitrogen to block KH₂ oxidation. (C) Microsomes containing only the carboxylase (Carb) or also expressing wt or Y139F VKORC1 were tested for KO or K supported carboxylation, as assessed by [¹⁴C]-CO₂ incorporation into the peptide FLEEL. (D) The microsomes in panel C were also tested for KH₂-supported carboxylation. (E) Insect cells containing tagged (Y139F_flag) and untagged Y139F or only the individual forms were subjected to immunopurification with anti-FLAG antibody, followed by western analysis using antibody against VKORC1, as before. Mock indicates uninfected cells.

Figure 2.

Warfarin inhibits full VKORC1 reduction and consequent carboxylation much more than expected from the inhibition of the 2 individual reactions. (A) Y139F was assayed under nitrogen for KO to KH₂ reduction in the absence (−) or presence (+) of warfarin. (B-D) Microsomes containing the carboxylase and Y139F or Y139H or wt VKORC1 were assayed for KO reduction to K, using high-performance liquid chromatography to monitor vitamin K forms, and for carboxylation by measuring [¹⁴C]-CO₂ incorporation into FLEEL. (E-F) The K to KH₂ reaction was analyzed by first assaying microsomes containing the carboxylase and wt VKORC1 (E) or Y139F (F) for KO to K inhibition, which gave the indicated concentrations of K. Carboxylation that depends on K to KH₂ reduction was then monitored for warfarin inhibition with these K concentrations, or with KO (65 µM).

Figure 3.

Warfarin uncouples the 2 VKORC1 reactions that fully reduce KO to KH₂. (A) r-wt VKORC1/r-fIX BHK cells were incubated without vitamin K (−K) or with KO or K (both at 2 µM) in the absence (−) or presence (+) of warfarin (2 µM). Secreted fIX was analyzed in westerns for carboxylation (anti-Gla) or total fIX (anti-fIX). The differences in fIX migration are a result of glycosylation. (B) r-fIX BHK and r-wt VKORC1/r-fIX BHK cells incubated with KO (2 µM) were monitored for KO reduction by isolation of intracellular vitamin K, followed by high-performance liquid chromatography analysis to separate and quantitate individual forms. fIX carboxylation was monitored by western analysis with anti-Gla antibody. (C) r-wt VKORC1/r-fIX BHK cells were incubated with KO (2 µM) and varying amounts of warfarin. KO reduction and fIX carboxylation were monitored as in (B). The KO to K curve is only slightly affected by subsequent K reduction, as only ∼15% of the K resulted in Gla product. (D) The K to KH₂ reaction and consequent carboxylation was analyzed using low K levels that result from inhibition of the KO to K reaction. Cells were incubated with warfarin (50 nM) and KO (2 µM) or K (0.4-2 µM). A range of K concentrations was used to identify cells with the appropriate intracellular K concentration for comparison with cells incubated in KO. Warfarin decreased KO to K reduction 2.5-fold, giving an intracellular K level of 1 nmol for 10⁷ cells. Left: warfarin (50 nM) inhibition of carboxylation in cells containing this intracellular K level. Right: warfarin inhibition of cells incubated with KO (2 µM).

Figure 4.

Warfarin uncouples full KO to KH₂reduction in cells expressing warfarin-resistant mutants. (A) Western analysis with anti-VKORC1 antibody was performed on r-fIX BHK cells expressing r-VKORC1 variants or untransfected cells (C). (B) Specific activities were determined by assaying the cells as in Figure 3B and normalizing activity to protein expression. (C-D) Warfarin sensitivity was monitored as in Figure 3C. (E) Warfarin inhibition of carboxylation dependent on the Y139F K to KH₂ reaction was tested as in Figure 3D.

Figure 5.

Y139H is only more active than wt VKORC1 in the presence of warfarin. (A) The response of VKORC1-dependent fIX carboxylation to warfarin was monitored for wt and Y139H VKORC1. (B) The relative activities were determined by normalizing the fIX carboxylation values to protein levels determined by western analysis with anti-VKORC1 antibody. (C) The same samples were tested in a clotting assay.

Figure 6.

293 cells that have warfarin-resistant quinone reductase activity do not show differences in inhibition of KO reduction and fIX carboxylation. (A) r-FIX 293 cells were generated by transfection of human r-fIX in pCMV6-A-Puro (Origene) and selection with puromycin (1 µg/mL). A clonal isolate was tested for warfarin inhibition as in Figure 3A, revealing the presence of warfarin-resistant quinone reductase activity. (B) Warfarin sensitivity of KO reduction and fIX carboxylation were analyzed as in Figure 3C.

Figure 7.

How is full reduction of KO to KH₂accomplished during warfarin therapy? (A) VKORC1 normally performs 2 reactions to generate the KH₂ cofactor required for VKD protein carboxylation. (B) During warfarin therapy, a warfarin-resistant quinone reductase may cooperate with VKORC1 to produce KH₂.