生物谷報道：下一期Nature（February 05, 2004)將報道一種新的蛋白VKOR,香豆素抗凝血劑（同芐丙酮香豆素一樣）通過抑制維生素-K環(huán)氧化物還原酶（VKOR）復合物來減少凝血活性,。盡管研究工作已進行了60年,，但這種復合物的成分仍未識別出來?，F(xiàn)在，兩個研究小組已經識別出了一種關鍵的VKOR蛋白,，其手段一是通過在芐丙酮香豆素抗性患者身上,、在患有一種罕見出血失調癥的患者身上和在芐丙酮香豆素抗性大鼠身上證實突變的存在，二是通過用siRNA使基因表達沉寂,。這是對凝血過程及胚胎形成來說關鍵的維生素-K循環(huán)過程的第一種已知的分子成分,。這項工作可幫助人們了解因香豆素治療所產生的變化，并幫助研究人員找到可能的新藥物目標,。該期封面所示為該VKOR蛋白所在的內質網（紅色）,。

Nature 427, 537 - 541 (05 February 2004); doi:10.1038/nature02214 
 
 
Mutations in VKORC1 cause warfarin resistance and multiple coagulation factor deficiency type 2

SIMONE ROST1,2,*, ANDREAS FREGIN1,*, VYTAUTAS IVASKEVICIUS3, ERNST CONZELMANN4, KONSTANZE HÖRTNAGEL2, HANS-JOACHIM PELZ5, KNUT LAPPEGARD6, ERHARD SEIFRIED3, INGE SCHARRER7, EDWARD G. D. TUDDENHAM8, CLEMENS R. MÜLLER1, TIM M. STROM2,9 & JOHANNES OLDENBURG1,3

1 Department of Human Genetics, University of Würzburg, Biozentrum, Am Hubland, 97074 Würzburg, Germany
2 Institute of Human Genetics, GSF National Research Center, Ingolstädter Landstrasse 1, 85764 München-Neuherberg, Germany
3 Institute of Transfusion Medicine and Immune Haematology of the DRK Blood Donor Service, Sandhofstrasse 1, Johann Wolfgang Goethe-Universität, 60526 Frankfurt, Germany
4 Department of Physiological Chemistry II, University of Würzburg, Biozentrum, Am Hubland, 97074 Würzburg, Germany
5 Federal Biological Research Center for Agriculture and Forestry, Institute for Nematology and Vertebrate Research, Toppheideweg 88, 48161 Münster, Germany
6 Department of Medicine, Nordland Hospital, 8092 Bodo, Norway
7 Center of Internal Medicine, Johann Wolfgang Goethe-Universität, Theodor-Stern-Kai 7, 60528 Frankfurt, Germany
8 MRC Clinical Sciences Centre, Imperial College, London W12 ONN, UK
9 Institute of Human Genetics, Klinikum rechts der Isar, Technical University, 81675 München, Germany
* These authors contributed equally to this work

Correspondence and requests for material should be addressed to J.O. ([email protected]). The sequences are deposited in GenBank under accession codes AY423042, AY423043, AY423044, AY423045, AY423046, AY423047 and AY423048

Coumarin derivatives such as warfarin represent the therapy of choice for the long-term treatment and prevention of thromboembolic events. Coumarins target blood coagulation by inhibiting the vitamin K epoxide reductase multiprotein complex (VKOR)1. This complex recycles vitamin K 2,3-epoxide to vitamin K hydroquinone, a cofactor that is essential for the post-translational -carboxylation of several blood coagulation factors2, 3. Despite extensive efforts, the components of the VKOR complex have not been identified4-8. The complex has been proposed to be involved in two heritable human diseases: combined deficiency of vitamin-K-dependent clotting factors type 2 (VKCFD2; Online Mendelian Inheritance in Man (OMIM) 607473), and resistance to coumarin-type anticoagulant drugs (warfarin resistance, WR; OMIM 122700). Here we identify, by using linkage information from three species, the gene vitamin K epoxide reductase complex subunit 1 (VKORC1), which encodes a small transmembrane protein of the endoplasmic reticulum. VKORC1 contains missense mutations in both human disorders and in a warfarin-resistant rat strain. Overexpression of wild-type VKORC1, but not VKORC1 carrying the VKCFD2 mutation, leads to a marked increase in VKOR activity, which is sensitive to warfarin inhibition.

We previously mapped the locus for VKCFD2 to a 20-megabase (Mb) region on chromosome 16p12–q21 (ref. 9). We noted that a linkage group of genes on 16p11 is orthologous to genes around the warfarin resistance loci Rw in rats10, 11 and War in mice12. This homology led us to propose that VKCFD2 and warfarin resistance may be caused by allelic mutations in the same gene. If so, this would narrow down the critical interval in humans to a region of roughly 4.0 Mb on the short arm of chromosome 16. According to the human genome assembly (build 33, April 2003), this region contains 129 Ensembl genes with roughly 1,100 exons (http://www.ensembl.org/).

Using genomic DNA from two VKCFD2 and four WR probands, we carried out a systematic mutation screen by directly sequencing genes from this region and found missense mutations in a gene of unknown function in all investigated VKCFD2 and WR probands (Fig. 1). This gene (IMAGE3455200) extends over a genomic region of 5,126 base pairs (bp) and comprises three exons encoding a protein of 163 amino acids with a calculated relative molecular mass of 18,000 (Mr 18K). Both non-related VKCFD2 index patients and their affected siblings carried the same homozygous point mutation in the third exon (292C  T), whereas the parents were found to be heterozygous (Fig. 1). The mutation caused the replacement of Arg 98 by tryptophan. The families were of Lebanese and German origin. Their haplotypes in the region of homozygosity encompassing the mutated gene were different. Furthermore, cytosine 292 is part of a CpG dinucleotide, which is a known mutation hotspot. Thus, the two mutations most probably arose independently.

 Figure 1 VKORC1 mutations in individuals with VKCFD2 and WR.   Full legend
 
High resolution image and legend (112k)
 

Four different heterozygous mutations were detected in the individuals with WR, as expected for the autosomal dominant inheritance of this phenotype. Patient C showed a valine to leucine substitution (Val29Leu), patient D a valine to alanine substitution (V45A), patient E an arginine to glycine replacement (Arg58Gly), and patient F an exchange of leucine to arginine (Leu128Arg). The Arg58Gly mutation was shared by the two affected brothers of patient E. Sequencing the whole coding region in 384 control chromosomes detected no amino acid substitutions but two synonymous single nucleotide polymorphisms (129C  T, Cys43Cys; 358C  T, Leu120Leu).

Because we had thought that warfarin resistance in rats might be due to mutations in the orthologous gene, we also screened this gene in the rat strain that was used to map Rw to chromosome 1 (ref. 10). DNA samples of ten sensitive and ten resistant rats were sequenced. We detected a heterozygous missense mutation, Tyr139Cys (416A  G), in ten resistant rats and in one sensitive rat that was absent from the other nine sensitive animals. Warfarin resistance determination by the blood-clotting response test (see Methods) is known to have false-positive and false-negative error rates of at least 0.01 (ref. 13). Under these conditions, linkage between mutation and Rw phenotype was still significant with a log likelihood ratio (lod) score of 3.9 at a recombination rate of zero.

We also sequenced 16 resistant and 5 sensitive wild-caught rats14. The sensitive rats carried a homozygous tyrosine at position 139, whereas 12 and 4 resistant rats were homozygous and heterozygous, respectively, for the mutation Tyr139Cys. The identification of seven independent mutations with two distinct phenotypes in humans and rat, and the absence of mutations from 384 human control chromosomes indicated that mutations in this gene might have a causal role in the pathogenesis of VKCFD2 and warfarin resistance. On the basis of the mutations and the functional data reported below, we named the gene vitamin K epoxide reductase complex subunit 1 (VKORC1).

An orthologue of the VKORC1 gene was present in mice (NM_178600), and the orthologues in rat and in Fugu rubripes (Fig. 2) were established by homology searches and polymerase chain reaction with reverse transcription (RT–PCR). The corresponding proteins share 79–84% identity with the human protein. Database searches did not detect homology to any known genes or to characterized protein domains. Topology prediction programs anticipated three transmembrane domains. The PSORT II (http://www.psort.org/) program predicted a carboxy-terminal endoplasmic reticulum (ER) membrane retention signal (Lys-Lys-X-X or Lys-X-Lys-X-X) in all VKORC1 proteins15, in accordance with location of the VKOR activity in the ER membrane fraction (ref. 5 and Fig. 3).

 Figure 2 Amino acid sequence alignment of VKORC1 and VKORC1L1.   Full legend
 
High resolution image and legend (195k)
 

 Figure 3 Subcellular location of VKORC1.   Full legend
 
High resolution image and legend (48k)
 

Tblastn searches with VKORC1 detected paralogous human (BC027734) and mouse (AK009497) genes, each with 50% protein identity. We established the complete complementary DNA sequences in mouse, rat and Fugu rubripes. We named these genes VKORC1-like 1 (VKORC1L1). Human, mouse and rat VKORC1L1 proteins were highly conserved, sharing roughly 97% identity. We also detected a homologous protein in Xenopus laevis (AAH43742) and, with weaker homology, in Anopheles gambiae (EAA06271) (Fig. 2). Pseudogenes of VKORC1 and VKORC1L1 were present in the human, mouse and rat genomes.

VKORC1 seems to be conserved in vertebrates, as it is present in human, rodents and fish. Notably, a close homologue is present in Anopheles gambiae but not in the Drosophila genome, although a gene encoding -carboxylase has been described in the fruitfly16. -Carboxyglutamate residues have been also found in the peptide venom of the marine snail genus Conus17. Like in vertebrates, -carboxylation in the mollusc and Drosophila requires reduced vitamin K as a cofactor. Thus, protein modification by -carboxylation seems to antedate the evolutionary emergence of VKORC1.

The tissue distribution of more than 100 human expressed sequence tags available at UniGene suggested that VKORC1 is widely expressed. On analysis of fetal and adult human tissues by northern blotting, we detected a single transcript of 1.0 kilobases, providing no evidence for alternative splicing. We found the highest expression in fetal and adult liver, followed by fetal heart, kidney and lung, adult heart and pancreas (data not shown). Thus, like the GGCX gene, VKORC1 is highly expressed in the adult liver; however, its expression seems to be broader. For example, it is clearly transcribed in fetal liver and both fetal and adult heart, where only low expression of GGCX has been described18.

From biochemical fractionation experiments it is known that the VKOR activity purifies with the microsomal membrane fraction5, 19. Furthermore, -glutamyl carboxylase has been localized to the membranes of the ER by immunocytochemistry20. To study the subcellular localization of human VKORC1, we generated constructs expressing green fluorescent protein (GFP)- and Myc-tagged VKORC1 fusion proteins for transient transfection of COS-7 cells. Primary antibodies against the epitope tags and fluorochrome-labelled secondary antibodies were used to visualize the fusion proteins; an antibody against the ER-specific protein calnexin was used as a control. The green immunofluorescence of the VKORC1 fusion proteins decorated the mesh-like structures of the ER in the cytoplasm and perfectly colocalized with the red label of the ER marker calnexin (Fig. 3). Thus, VKORC1 seems to be located in the ER.

To study the effect of VKORC1 on the reduction of vitamin K 2,3-epoxide, we overexpressed wild-type and mutated VKORC1 in HEK293 cells. VKOR activity was measured by the dithiothreitol (DTT)-dependent conversion of vitamin K 2,3-epoxide to vitamin K quinone. Dose–response curves to warfarin inhibition were measured at final concentrations of 5–80 µM warfarin19. Untransfected and mock-transfected cells showed a low basal activity, which was sensitive to warfarin. Overexpression of wild-type VKORC1 resulted in a marked stimulation of VKOR activity: production of vitamin K quinone was increased 14–21-fold as compared with untreated and mock-transfected cells. The activity was inhibited by warfarin in a dose-dependent manner (Table 1).

We also determined VKOR activity after transfection with constructs expressing mutated VKORC1 (Table 1). Recombinant expression of the Arg98Trp mutation observed in the two VKCFD2 families only slightly increased VKOR activity in HEK293 cells. Spontaneous bleeding episodes and high concentrations of serum vitamin K epoxide in these patients suggest that the efficiency of vitamin K recycling is also markedly decreased in vivo21. The five WR variants showed a reduced VKOR activity ranging from 5% for the Leu128Arg mutation to 96% for the Val29Leu mutation. The Val45Ala, Arg58Gly and Tyr139Cys variants showed, respectively, about 23%, 21% and 48% activity (Table 1).

Reduced VKOR activity associated with higher vitamin K demand and death from spontaneous bleeding has been observed in heterozygous and homozygous Rw rats13, 14, 22. Similarly in our expression system, which mimics homozygous conditions, WR mutations led to a lower functional efficiency of the VKOR complex. Whereas at the phenotypic level all WR variants showed at least partial resistance towards the anticoagulation effect of warfarin, both wild-type and mutant proteins were sensitive to warfarin in vitro. At concentrations above 20 µM, the Val29Leu and Tyr139Cys variants retained higher VKOR activities than did the wild type, whereas in the Val45Ala, Arg58Gly and Leu128Arg variants VKOR activity fell below the detection limit (Table 1). Inconsistent response to warfarin has been reported previously in warfarin-resistant rats14, 22. Further studies are required to resolve in more detail the function of VKORC1 in warfarin resistance and VKCFD2.

In summary, by using a positional cloning approach that integrated mapping information from three species, we have identified a gene that is mutated in VKCFD2 and WR. From the segregation data, the absence of mutations in controls and the activity measurements of recombinant proteins, we conclude that mutations in this gene are causative for both VKCFD2 and WR. The molecular characterization of this first component of the VKOR complex will be instrumental in isolating the other subunits of this complex and in resolving the binding mechanisms of substrate and inhibitors. This should lead to a better understanding of the pharmacological action of coumarin drugs and may provide a basis for the rational design of anticoagulants targeting VKOR.

Methods
Subjects We carried out mutation screening of candidate genes on DNA samples from two VKCFD2 index patients and three unrelated WR patients. Clinical data of the VKCFD2 families have been described21. Patients C, D and F are sporadic cases. Patient E has two brothers suffering from warfarin resistance. Whereas the normal range of weekly warfarin doses is 10–60 mg, patient C required 100 mg, and patient E and his two brothers needed 220–250 mg of warfarin per week to reach therapeutic anticoagulation. Patients D and F did not respond to warfarin at all doses tested. All patients gave informed consent to the study.

Rw rats An Rw rat colony was established as described10. A male brown Norway rat caught in the wild from the Münsterland area, Germany, that was homozygous resistant to warfarin and bromadiolone was crossed with a Wistar albino female. One F1 heterozygous resistant male was used to establish a segregating backcross population. At the time of the study, the colony was abandoned. We analysed DNA samples from seven susceptible and eight resistant offspring of four matings between two sixth and seventh generation heterozygous resistant males and Wistar albino rats, and three susceptible and two resistant offspring of a backcross between a seventh generation heterozygous resistant male and his susceptible daughter from an earlier mating. Rats were tested for warfarin and bromadiolone resistance with the blood-clotting response method. Wild-caught rats were characterized for resistance to coumarin-derivatives and VKOR activity as described14.

Sequence analysis Genome sequences and annotation were obtained from the National Center for Biotechnology Information (NCBI) (http://www.ncbi.nlm.nih.gov/), University of California Santa Cruz (UCSC) (http://genome.ucsc.edu/) and Ensembl (http://www.ensembl.org/). Primers for mutation screening were designed using Primer3 software integrated into a script to allow for automatic primer design (http://ihg.gsf.de/ihg/ExonPrimer.html). We amplified exon sequences and exon–intron boundaries with intronic primers and sequenced them directly by the BigDye Terminator Cycle sequencing kit (ABI). Primer sequences are available from the authors on request. Topology predictions were done using PHDhtm (http://cubic.bioc.columbia.edu/predictprotein/).

Northern blots Human multiple tissue northern blots, Fetal Blot 1 (Stratagene) and Human 8-Lane (BD Biosciences Clontech), were hybridized to radiolabelled full-length human VKORC1 cDNA in MiracleHyb high-performance hybridization solution (Stratagene). A -actin probe supplied with the multiple tissue northern blot was used as a control.

Cloning and construction of expression vectors The complete coding sequence of VKORC1 was amplified from human liver and kidney cDNA by using Marathon-Ready cDNA (BD Biosciences Clontech) and primers with added cleavage sites for HindIII and EcoRI (VKORC1-HindIII-F, 5'-ATTAAGCTTCACCATGGGCAGCACCTGGGGGAGCCCT-3'; VKORC1-EcoRI-R, 5'-ATTGAATTCCGTGCCTCTTAGCCTTGCCCTG-3'), and cloned into pBluescript II (Stratagene). For immunocytochemistry, we re-cloned the insert into the mammalian expression vectors pEGFP-N1 (BD Biosciences Clontech) and pcDNA3.1/myc-His (Invitrogen).

For expression studies, the VKORC1 cDNA was cloned into pcDNA3 (Invitrogen) after amplification with the primers VKORC1-pcdna3-F (5'-GGGCGGAAGCTTGAGATAATGGGCA-3') and VKORC1-pcdna3-R (5'-GCTTGAATTCAGGGCTCAGTGC-3'). Mutagenesis was done with the QuikChange mutagenesis kit (Stratagene). Wild-type and mutated cDNAs were recloned for expression in pCEP4 (Invitrogen) using the HindIII and XhoI sites. We verified all constructs by sequencing.

Cell culture, transient transfection and immunocytochemistry COS-7 cells (DSMZ) were grown in DMEM medium with 10% fetal calf serum (FCS) on glass coverslips for 18–24 h and transfected with the expression vectors by using Effectene (Qiagen) in accordance with the manufacturer's specifications. After 48–60 h of further culturing, immunostaining was done as described23. Primary antibodies, Living Colors A.v. JL-8 (BD Biosciences Clontech), anti-Myc antibody (Invitrogen) and anti-calnexin (Sigma), were diluted 1:100 in the blocking solution. The same incubation and washing procedures were used for the secondary antibodies fluorescein isothiocyanate (FITC)-conjugated anti-mouse IgG (Sigma) and Cy3-conjugated anti-rabbit IgG F(ab')2 fragment (Sigma). Coverslips were counterstained with 4',6-diamidino-2-phenylindole dihydrochloride (DAPI; 1:500), washed and visualized under a fluorescence microscope (Leica).

Expression studies and VKOR activity measurements We grew HEK293-EBNA cells (Invitrogen) in MEM medium plus 10% FCS. For each experiment, 6  105 cells were plated onto 94-mm Petri dishes. After 30 h at 37 °C and 5% CO2, transfection (20 µg of DNA construct per dish) was done by the calcium phosphate method. After 40 h at 35 °C and 3% CO2, transfected cells were washed in PBS, collected and lysed in 0.25 mM imidazole (pH 7.6) plus 0.5% CHAPS. Transfection efficiency was checked by sequencing of RT–PCR products.

VKOR enzymatic activity was measured in whole-cell extracts as described19 with 5 µM vitamin K 2,3-epoxide as a substrate in the presence of 5 mM DTT. After extraction by isopropanol:hexane (3:2), vitamin K quinone was separated from the epoxide by HPLC on a reversed-phase C18 column. During the extraction procedure, vitamin K hydroquinone was quantitatively oxidized to the quinone form. Measurements were run in duplicate and the activity is given as the percentage of substrate converted into quinone. Vitamin K 2,3-epoxide was prepared by the oxidation of vitamin K quinone (Sigma-Aldrich) with H2O2. We added warfarin (Sigma-Aldrich) in dimethyl sulphoxide (<1 vol%).

Received 5 September 2003;accepted 31 October 2003