Fig 2: Localization and characterization of FSP1 as vitamin K reductase.a Fluorescence confocal microscope imaging of FSP1-sfGFP fusion (green) with cell organelle markers (red) of ER, mitochondrion, and Golgi. FSP1-sfGFP and the cell organelle marker were transiently co-transfected into HEK293 cells for 48 h. b Effect of FSP1 myristylation site mutation (G2A) and mutations of residue responsible for the dinucleotides (NDAH and FAD) binding on VKR activity. Wild-type FSP1 or its mutants was transiently expressed in TKO cells for VKR activity assay. Wild-type FSP1 activity was normalized to 100% (indicated by dotted line). Bottom: Immunoblotting of FSP1, its mutants, and GAPDH (loading control) in the corresponding cells. c AlphaFold model of FSP1 (gray) with FAD (cyan), NADH (gold) and vitamin K (green) based on ligand positions in the yeast NDH-2 enzyme Ndi1 (PDBcode 4G73). The GxGxxG motifs are colored purple (Gly18-Gly23) and orange (Gly149-Gly154) corresponding to interactions with the nucleotide motifs of FAD and NADH respectively. Residues Asp41 and His174 located near the ribose hydroxyls from the two nucleotides are colored pink as is Phe21 from the first GxGxxG region. d FSP1 residues surrounding the isoalloxazine of FAD, nicotinamide of NADH and quinone of vitamin K in the model that were mutated in this study are colored pink. e Effect on FSP1 activity due to mutations surrounding the proposed vitamin K binding site. VKR activity was determined as described above (Fig. 4b). Bottom: Immunoblotting of FSP1, its mutants, and GAPDH (loading control) in the corresponding cells. f Proposed activation mechanism of the activity-based fluorescent probe of vitamin K (VK-ASM) towards FSP1. The probe is inactive until the vitamin K moiety is reduced to hydroquinone, and the rearrangement of the reduced intermediated releases the strong fluorescent tag. g Effect of FSP1 mutations on the reduction of vitamin K and CoQ10. Wild-type FSP1 and its mutants were transiently expressed in TKO cells for forty-eight hours, cell lysate was used for activity assay using the activity-based fluorescent probe of vitamin K (VK-ASM) and CoQ10 (NIR-ASM) as the substrate. Fluorescence intensity of the wild-type FSP1 for each probe was normalized to 100% (indicated by the dotted line). Data are presented as mean ± SD of three independent experiments (n = 3) in Fig. 4b, e, g. Similar results were observed at least twice as shown in Fig. 4b, e.

Fig 3: Characterization of vitamin K reduction in different reporter cells.a Kepo and b vitamin K concentration titrations in different reporter cells. Reporter cells were incubated with increasing concentrations of Kepo or vitamin K for 24 h. VKD carboxylation of the reporter-protein FIXgla-PC were determined using ELISA. Carboxylation activity of HEK293 cells at the highest Kepo or vitamin K concentration was normalized to 100%. HEK293: FIXgla-PC/HEK293 reporter cell; FSP1 KO: FIXgla-PC/HEK293 cells with their endogenous fsp1 gene knocked out; DKO: FIXgla-PC/HEK293 with their endogenous vkor and vkorl genes knocked out. c Half-maximal effective concentration of vitamin K (EC50) determined in HEK293, DKO, and FSP1 KO reporter cells from Fig. 3b. d Warfarin inhibition of VKD carboxylation in DKO and FSP1 KO cells with vitamin K as the substrate, and in FIXgla-PC/HEK293 cells with Kepo as the substrate. Increasing concentrations of warfarin with 11 µM vitamin K or 5 µM Kepo were incubated with the corresponding reporter cells for 24 h for VKD carboxylation activity assay. Carboxylation activity of each cell line without warfarin was normalized as 100%. e VKD carboxylation efficiency in FSP1 KO and TKO cells using vitamin K as the substrate. Carboxylation activity was presented as concentration of carboxylated reporter-protein in the cell culture medium (ng/mL) determined by ELISA. f HPLC-based conventional VKR in vitro activity assay of the selected NAD(P)H-dependent oxidoreductases in TKO cells. Candidate proteins were transiently expressed in TKO cells for 48 h and VKR activity of reducing vitamin K to KH2 was determined. VKR activity of FSP1 was normalized to 100%. Insert: Immunoblotting of the expression of the tested oxidoreductases and similar immunoblotting results were observed at least twice. DHRS, dehydrogenase/reductase SDR family protein; CBR, carbonyl reductase. (g) Warfarin inhibition of NQO1 reducing vitamin K to KH2 by HPLC-based VKR in vitro activity assay. Final concentration of warfarin in the reaction mixture was 100 µM. h, i reduction of menaquinone-4 (MK4) (h) and menaquinone-7 (MK7) (i) by FSP1 determined by monitoring the production of their corresponding hydroquinone by HPLC-based VKR in vitro activity assay. HEK293: cell lysate of HEK293 cells; HEK293 + FSP1, cell lysate of HEK293 cells overexpressing FSP1. A same number of HEK293 and HEK293 + FSP1 cells were used for the activity assay. Data are presented as mean ± SD of three independent experiments (n = 3) in Fig. 3a–g.

Fig 4: Effect of the CoQ10-associated ferroptosis suppression pathways on VKD carboxylation.a Scheme of the FSP1-CoQ10-NADH and DHODH-CoQ10-DHO pathways for ferroptosis suppression. CoQ10 can be reduced to CoQ10H2 by either FSP1 or DHODH using the reducing equivalent from NADH or DHO, respectively. Reduced CoQ10 traps lipophilic radicals to suppress ferroptosis. NADH reduced nicotinamide adenine dinucleotide, NAD+ oxidized nicotinamide adenine dinucleotide, DHO dihydroorotate, OA orotic acid, PLOOH phospholipid hydroperoxides, PLOO· phospholipid peroxyl radical. b Cell-based VKR activity of FSP1 and DHODH. FSP1 and DHODH were transiently expressed in TKO reporter cells for 24 h for carboxylation activity assay. Wild-type FSP1 activity was normalized to 100%. c Effect of FSP1 inhibitors on VKD carboxylation. Increasing concentrations of FSP1 inhibitors, HQNO or iFSP1, with 11 µM vitamin K were incubated with DKO reporter cells for 24 h for VKD carboxylation activity assay. Carboxylation activity of cells incubated with 11 µM vitamin K without inhibitor was normalized as 100%. d Effect DHODH inhibitors on VKD carboxylation. Increasing concentrations of DHODH inhibitors, BRQ and LFM, with 11 µM vitamin K were incubated with DKO reporter cells for 24 h for VKD carboxylation activity assay. Carboxylation activity of cells incubated with 11 µM vitamin K without inhibitor was normalized as 100%. e Lineweaver–Burk plot of VKR activity in DKO cells as a function of vitamin K concentration in the absence and presence of HQNO. DKO reporter cells were cultured with increasing concentrations of vitamin K containing 0, 0.625, 1.25, 2.5, 5.0, and 10 µM HQNO for the activity assay. Data were plotted using Graphpad software. Data are presented as mean ± SD of three independent experiments (n = 3) in Fig. 5b–d.