18Vitamin K epoxide reductase (VKOR) drives the vitamin K cycle, activating vitamin K-dependent 19 blood clotting factors. VKOR is also the target of the widely used anticoagulant drug, warfarin 20Despite VKOR's pivotal role in coagulation, its structure and active site remain poorly understood. 21In addition, VKOR variants can cause vitamin K-dependent clotting factor deficiency 2 or alter 22 on abundance and activity. While most variants show wild type-like activity, 25% show low 83 abundance or activity, which could confer warfarin sensitivity or cause disease in a homozygous 84 context. Finally, we analyzed warfarin resistance variants and found that they span a range of 85 abundances, indicating that increased abundance is an uncommon mechanism of warfarin 86 resistance. 87 88 RESULTS 89
Multiplexed measurement of VKOR variant abundance using VAMP-seq 90To measure the abundance of VKOR variants, we applied Variant Abundance by 91Massively Parallel sequencing (VAMP-seq), an assay we recently developed (Matreyek et al., 92 2018). In VAMP-seq, a protein variant is fused to eGFP with a short amino acid linker. If the 93 variant is stable and properly folded, then the eGFP fusion will not be degraded, and cells will 94 have high eGFP fluorescence. In contrast, if the variant causes the protein to misfold, protein 95 quality control machinery will detect and degrade the eGFP fusion, leading to a decrease in eGFP 96 signal (Fig. 1a). mCherry is also expressed from an internal ribosomal entry site (IRES) to control 97 for expression. Differences in abundance are measured on a flow cytometer using the ratio of 98 eGFP to mCherry signal. To determine whether VAMP-seq could be applied to VKOR, we fused 99 eGFP to VKOR N-or C-terminally and found that both orientations had high eGFP signal ( Figure 100 1-figure supplement 1). We compared N-terminally tagged wild type (WT) VKOR to R98W, a 101 variant that ablates a putative ER retention motif and reduces abundance (Czogalla et al., 2014), 102 and to TMD1Δ, a deletion of residues 10-30 which comprise the putative first transmembrane 103 domain (TMD1; Fig. 1b). Both reduced-abundance variants exhibited much lower eGFP:mCherry 104 ratios than WT, demonstrating that VAMP-seq could be applied to VKOR. 105We constructed a barcoded site-saturation mutagenesis VKOR library that covered 92.5% 106 of all 3,240 possible missense variants. To express this library in HEK293T cells we used a Bxb1 107 recombinase landing pad system we previously developed (Matreyek et al., 2017). In this system,