Adaptative evolution of the Vkorc1 gene in Mus musculus domesticus is influenced by the selective pressure of anticoagulant rodenticides

Abstract Anticoagulant rodenticides are commonly used to control rodent pests worldwide. They specifically inhibit the vitamin K epoxide reductase (VKORC1), which is an enzyme encoded by the Vkorc1 gene, involved in the recycling of vitamin K. Therefore, they prevent blood clotting. Numerous mutations of Vkorc1 gene were reported in rodents, and some are involved in the resistant to rodenticides phenotype. Two hundred and sixty‐six mice tails were received from 65 different locations in France. Coding sequences of Vkorc1 gene were sequenced in order to detect mutations. Consequences of the observed mutations were evaluated by the use of recombinant VKORC1. More than 70% of mice presented Vkorc1 mutations. Among these mice, 80% were homozygous. Contrary to brown rats for which only one predominant Vkorc1 genotype was found in France, nine missense single mutations and four double mutations were observed in house mice. The single mutations lead to resistance to first‐generation antivitamin K (AVKs) only and are certainly associated with the use of these first‐generation molecules by nonprofessionals for the control of mice populations. The double mutations, probably obtained by genetic recombination, lead to in vitro resistance to all AVKs. They must be regarded as an adaptive evolution to the current use of second‐generation AVKs. The intensive use of first‐generation anticoagulants probably allowed the selection of a high diversity of mutations, which makes possible the genetic recombination and consequently provokes the emergence of the more resistant mutated Vkorc1 described to date.


| INTRODUCTION
House mice (Figure 1) are the most widespread mammals on the earth and are present in every continents and all environments (rural, urban, and insular;Angel, Wanless, & Cooper, 2009). Mus musculus domesticus belongs to the list of the UICN as one of the 100 most invasive species in the world. To manage rodent populations, chemical controls have been organized since 1950 using antivitamin K (AVK) anticoagulant rodenticides. The intensive use of such molecules for pest control has selected many resistant strains of rodents. Resistance was first detected in brown rats in 1958 (Boyle, 1960) and in house mice in the early 1960s (Dodsworth, 1961) in the UK. Since this initial observation, resistance has been reported worldwide, in many European countries, in the US (Jackson & Kaukeinen, 1972), in Canada (Siddiq & Blaine, 1982), in Japan (Tanaka et al., 2012), and in Australia (Saunders, 1978). The emergence of such resistance to anticoagulants belonging to the first generation (i.e., warfarin, diphacinone, coumatetralyl, chlorophacinone) led to the development of new AVK belonging to the second-generation molecules (i.e., bromadiolone, difenacoum, flocoumafen, brodifacoum, and difethialone) in the 1970s and 1980s.
The resistance to AVKs was proposed to be supported by two major mechanisms in rodents 1/a metabolic resistance due to an accelerated detoxification system involving cytochrome P-450 (Ishizuka et al., 2007;Sugano et al., 2001) and 2/a target resistance due to the inefficiency of AVKs to specifically inhibit the vitamin K epoxide reductase (VKOR) activity. This VKOR activity is involved in the recycling of vitamin K by allowing the reduction in vitamin K epoxide in vitamin K quinone. Vitamin K is necessary for the activation of clotting factors II, VII, IX, and X. Inhibition of VKORC1 enzyme by AVK molecules results in the absence of gamma-carboxylated clotting factors II, VII, IX, and X and thus compromises the coagulation process.
While the VKOR activity was described in the 1970s, the VKORC1 enzyme catalyzing this activity was identified in 2004 only by two different teams (Li et al., 2004;Rost et al., 2004) Rost et al., 2004). This enzyme of 163 amino acids is coded by the vkorc1 gene. This gene is located on the chromosome 7 in mice and on the chromosome 1 in rats. Single nucleotide polymorphisms of this gene were immediately proposed to be responsible for resistance to AVK (Grandemange et al., 2009; Hodroge, Longin-Sauvageon, Fourel, Benoit, & Lattard, 2011; Pelz et al., 2005;Rost et al., 2004) and appeared to support the resistance process in western Europe, even if cohabitation of target resistance and metabolic resistance had been demonstrated in Denmark (Markussen, Heiberg, Fredholm, & Kristensen, 2007.
In this paper, we report the different mutations of M. musculus domesticus Vkorc1 gene observed in different parts of France, six of these mutations being described for the first time. By the use of recombinant VKORC1, we thus analyzed the catalytic consequences of all the different mutations described to date in house mice in order to evaluate the resistant phenotype associated with these mutations.
This characterization allowed us to better understand the origin of the different evolution of the Vkorc1 gene between rats and mice. The diversity of Vkorc1 mutations observed exclusively in M. musculus domesticus led to the emergence of double mutants described for the first time in this study. These double mutations of the Vkorc1 gene are associated with severe resistant to all AVK.

| Mice tissue sampling
Mus musculus domesticus samples were collected from the national network of pest control operators (PCOs) in 27 of 95 departments (French administrative areas) covering all the country. The tails of dead mice were cut, and the samples were sent to the laboratory by mail in individual tubes in 70% alcohol. They were frozen at −20°C until analysis. For each tail, PCO filled a questionnaire indicating the site where the mouse was collected. Sampling was performed by trapping or by collecting mice found dead after chemical control.

| Heterologous expression of wild-type and mutated M. musculus domesticus VKORC1
The coding sequence corresponding to the M. musculus domesticus VKORC1 fused with a c-myc tag via a flexible (GGS) 3 in its 3′-extremity was optimized for heterologous expression in yeast and synthetized by GenScript (Piscataway, NJ, USA). The synthetized nucleotide sequence included EcoRI and XbaI restriction sites at its 5′-and 3′-extremities, respectively. This nucleotide sequence was subcloned into pPICZ-B (Invitrogen) and sequenced on both strands.
Construction of mVKORC1 mutant was carried out using pPICZ-mVKORC1 as template with the Quick change site-directed mutagenesis kit (Stratagene) according to the manufacturer's recommendations.
Each mutant was checked by sequencing, and thus expressed in Pichia pastoris as described by Hodroge et al. (2011Hodroge et al. ( , 2012.

| Subcellular fractionation of yeast cells
Microsomes were prepared from yeast cells by differential centrifugation. Briefly, yeast cells were resuspended in 50 mmol/L Phosphate Buffer (pH 7.4) containing 1.15% (w/v) of KCl. They were broken with Zircon beads using Dispermat ® LC30 (VMA-GETZMANN, Germany; 30 min-3,500 rpm) continuously at 4°C and further submitted to differential centrifugation a continuously at 4°C. The 100,000 g pellet corresponding to the membrane fraction was suspended by Potter homogenization in HEPES-glycerol buffer (50 mmol/L Hepes, 20% glycerol, pH 7.4). Protein concentrations were evaluated by the method of Bradford (1976) using bovine serum albumin as a standard. Microsomal fractions were frozen at −80°C and used for kinetic analysis.

| Vitamin K epoxide reductase activity assays and kinetics
Microsomal VKOR activity was assayed as described previously (Bodin et al., 2005;Haniza et al., 2015). Briefly, standard reactions were performed in 200 mmol/L Hepes buffer (pH 7.4) containing 150 mmol/L KCl, 1 mmol/L dithiothreitol, and 1 g/L of total proteins. The reaction was started by the addition of vit K 1 >O solution in 1% Triton X-100 and incubated at 37°C for 30 min. In these conditions, the reaction was linear according to the time of incubation and the quantity of incubated proteins. After incubation at 37°C for 30 min, the reaction was stopped by adding of 2 ml of isopropanol. After centrifugation at 3,000 g for 10 min in order to precipitate proteins, 2 ml of hexane was added. After centrifugation at 3,000 g for 10 min, the hexane layer was removed and dried under nitrogen. The dry residue was immediately dissolved in 0.2 ml of methanol, and reaction product was analyzed by liquid chromatography-mass spectrometry.
The LC-APCI/MS/MS used was a 6120 Quadrupole LC/MS with an atmospheric pressure chemical ionization (APCI) interface and a LCMS Chemstation software from Agilent Technologies (Palo Alto, CA, USA). Chromatographic separation was performed using a XTerra MS C18 column (2.1 mm × 50 mm, 2.5 μm; Waters, Milford, MA, USA) with a mobile phase of methanol, 0.1% acetic acid (96:4) in isocratic condition. The column temperature was 48°C. The flow rate in the LC column was 0.4 ml/min. The injection volume was 20 μl. The temperature of the auto sampler tray was set to 5°C, and the samples were protected from the daylight. Detection was by MS with APCI source in positive mode. Nebulizer pressure was set to 60 psi, dry gas temperature to 350°C, dry gas flow to 5 L/min, and vaporizer temperature to 400°C. Capillary voltage was set to 4,000 V, corona needle to 10 μA.
Identification criteria for vit K 1 are the retention time (tr = 234 s) and the selected ion 451.4. Identification criteria for vit K 1 >O are the retention time (tr = 174 s) and the selected ion 467.0. Linearity and accuracy were tested from 25 to 2,000 ng/ml (n = 20). The response was linear throughout the concentration range tested with a coefficient of correlation (r 2 ) above .99. Accuracy was between 80% and 120% of the theoretical concentrations. to 20 × K i . Data were fitted by nonlinear regression to the noncompetitive inhibition model v = (V max /(1 + (I/K i ))) × (S/(K m + S)) using the R-fit program.
Twelve Vkorc1 genotype were observed in the French mice. Six Vkorc1 genotypes detected in the French mice led to a single mutation in the corresponding VKORC1 protein (i.e., A26T, A26S, W59G, L124M, L128S, or Y139C). The locations of these genotypes are presented in Figure 3. The observed allelic frequencies of A26T, A26S, W59G, L124M, L128S, and Y139C in our sampling, were, respectively, 1.3%, 0.4%, 1.1%, 3.4%, 14.9%, and 12.3%. The allelic frequencies of these mutations were different between geographical areas. The results are summarized in Table 2 with the number of samples and the number of mutated samples. The departments were gathered according to five geographical areas (i.e., Brittany, North, Marne, South West and South East) presented in Figure 3. Mice carriers for two of these mutations, the A26T and Y139C mutations (one mouse), but also the L128S and Y139C mutations (four mice) at the heterozygous state, were detected. Cloning and sequencing of the Vkorc1 gene from these mice revealed that these mutations were present on different alleles. T A B L E 1 Detail of SNPs and mutations of Vkorc1 found in French mice 0.4% in France. The association of the g.969T>G mutation with the g.2177C>A mutation was only found in Brittany (Table 2 and Figure 3), and the association of the g.969T>G mutation with the g.2190T>C mutation was only found in Marne (Table 2 and Figure 3). The g.76G>T mutation was found to be associated with the g.976G>T mutation or the g.2190T>C mutation at the homozygous state leading to proteins with two combined mutations, the A26S and R61L or the A26S and L128S. The observed allelic frequencies of these two genotypes in our sampling were, respectively, 0.4% and 0.8% in France. Both genotypes were found in the North of France (Table 2 and Figure 3). The g.76G>A mutation was found to be associated with the g.2190T>C mutation at the homozygous state leading to a protein with two combined mutations, the A26T and L128S. The allelic frequency of this genotype in our sampling was 5.3%, and this genotype was detected only in the North of France.

| Functional consequences of VKORC1 mutations
To assess the consequences of VKORC1 mutations on the functional properties of VKORC1, wild-type VKORC1 and its mutants were overexpressed as c-myc-fused proteins in P. pastoris. All the French single and double mutants, but also mutants detected in other countries in previous studies, were characterized in this study. All proteins were efficiently expressed in P. pastoris with the same expected molecular mass of approximately 20 kDa.
The ability of each membrane protein to catalyze the reduction of K>O to K was determined. Five single or double mutants (i.e., W59G, W59L, W59S, W59G/L124L, and W59G/L128S) presented less than 2% of the VKOR activity determined for wild-type VKORC1 preventing additional studies (Table 3). The other mutants were all able to reduce the vitamin K epoxide with K m similar to wild-type VKORC1 (Table 3).
In order to compare the susceptibilities to AVKs of wild-type VKORC1 and its mutants, their respective inhibition constants K i toward various AVKs of the first generation (i.e., coumatetralyl, chlorophacinone) or the second generation (i.e., bromadiolone, difenacoum, difethialone, or brodifacoum) were determined. All AVKs were able to inhibit the VKOR activity in a noncompetitive manner for all the mutants. Nevertheless, the concentrations of AVKs necessary to inhibit the VKOR activity were different between mutants and between inhibitors. Results are presented in Figure 4 as a ratio between the K i obtained for the mutated protein and the K i obtained for the wild-type protein, this ratio representing the resistance factor of the mutated protein.

| DISCUSSION
All the 15 missense mutations in the mice VKorc1 observed all over Europe have been characterized using the same method which was also previously used to characterize the catalytic consequences of the human (Hodroge et al., 2012)  Resistance factors corresponding to the ratio between K i obtained for the mutated VKORC1 and for the wild-type VKORC1 were determined for each VKORC1 protein toward each anticoagulant rodenticide. These in vitro resistance factors used for the first time to study the consequences of the Vkorc1 mutations in brown rats (Hodroge et al., 2011) were shown to be totally coherent with resistance factors obtained using liver microsomes (Lasseur et al., 2005) and with resistance factors obtained by BCR tests after peros administration of AVK to strains of rats homozygous for wild-type Vkorc1 or Vkorc1-Y139F (Grandemange et al., 2009)  To determine the VKOR activity, standard reactions were performed in 200 mmol/L Hepes buffer (pH 7.4) containing 150 mmol/L KCl and 0.25-2 g/L of microsomal proteins expressing membrane wild-type or mutant VKORC1. Each data point represents the mean ± SD of three individual determinations. Comparison between two groups was made using Mann-Whitney test. *p < .05 compared to wild-type VKORC1. and L124Q) lead to a moderate resistance (Figure 4) to first-generation AVK such as chlorophacinone and coumatetralyl with resistance factors of about 5; second-generation molecules remaining still efficient.
Such resistance factors allow certainly mice homozygous for one of these mutations to survive when pest control is performed using firstgeneration AVK. On the contrary, the use of second-generation AVK will eliminate such mice. On the other hand, Vkorc1 mutations observed in mice such as L128S, Y139C, or the Vkorc1 spr genotype lead to a severe resistance to first-generation AVK, but also confer a limited resistance to some second-generation AVK.
The number of mutations reported in this gene is thus comparable between rats, mice, and humans. Therefore, the existence of such a diversity of mutations in the Vkorc1 gene cannot be associated with the management of rodent populations by the intensive use of AVK molecules while on the contrary the frequency of these mutations is clearly linked to this massive AVK use for pest control.
In humans, the allelic frequencies of Vkorc1 mutations are well below 1% and the resistance levels to AVK conferred by all the mutations are limited (Hodroge et al., 2012) compared to that conferred by some Vkorc1 mutations in rats (Hodroge et al., 2011). Indeed, the use of warfarin or other AVK in humans cannot be considered as a selection pressure as the average age of patients treated with AVK is around 70 years.
In rats, if 15 Vkorc1 mutations have been reported in Europe, only three mutations are frequently and mostly observed in Europe. In France and Belgium, the major mutation observed in the Vkorc1 gene of brown rats is the Y139F mutation (Baert, Stuyck, Breyne, Maes, & Casaer, 2012;Grandemange et al., 2010), the observed allelic frequency of the Y139F mutation was reported in 2010 to be 21% from a sampling of 268 rats trapped from 91 sites well distributed in France (Grandemange et al., 2010), while the other mutations found in France (i.e., Y139C, L128Q, L120Q, E155K, S103Y) accounted for only 6.5% of the Vkorc1 alleles analyzed. The repetition of this resistance monitoring conducted between 2014 and 2015 by our laboratory suggests an increase in the allelic frequency of the Y139F in France since 2010, with an observed allelic frequency that has now reached over 60% in a population of 180 rats trapped in 51 sites from 18 French administrative departments (data not shown). Moreover, this allelic frequency can locally reach levels close to 80% in rural areas (Berny, Fourel, & Lattard, 2014). In Germany, the Netherlands, and England, the predominance of one mutation in brown rats populations seems also to be the case with the Y139C mutation distributed in foci of resistance at frequencies similar to that observed in France for the Y139F mutation (Haniza et al., 2015;Meerburg, van Gent-Pelzer, Schoelitsz, & van der Lee, 2014;Pelz, 2007   , and L120Q mutations lead to drastic resistance to AVK of the first generation such as chlorophacinone and coumatetralyl, but also to some AVK of the second generation such as bromadiolone and to a lesser extent difenacoum (Hodroge et al., 2011). During pest control management with one of these molecules, only the brown rats carrier for one of these mutations, in the homozygous state and also, but to a lesser extent, in the heterozygous state, survive. Nevertheless, brown rats carriers of one of these mutations are still susceptible to AVK such as difethialone, brodifacoum, and flocoumafen and control with one of these molecules is still possible (Hodroge et al., 2011;Lasseur et al., 2005).
In mice, 15 mutations have been detected in Europe in this study and in a previous study performed in 2012 (Pelz et al., 2012).

L128S/Y139C
Ki d e t a t u m is predominant in the brown rat. This difference could be due to a difference in the selection pressure exerted by AVK. Management of mice is essentially realized by nonprofessionals, and we can assume that first-generation AVK can be used more extensively by nonprofessionals not informed about resistance problems while management of rats is realized more frequently by better trained and better informed PCO. Moreover, the feeding behavior of mice promotes diversity food sources and thereby induces a limitation of the amount of ingested bait. To that extent, Vkorc1 mutations conferring even limited in vitro resistance factors to first-generation AVK have been selected (i.e., A26T, E37G, R58G, L124M and L124Q). Therefore, the preservation of such a diversity of mutations observed in the three exons, (while the three main mutations in the rats concern two codons only and are in the same exon) and the high prevalence of each of these mutations in the homozygous state are responsible for the emergence of double mutants (i.e., A26T/L128S, A26S/L128S, W59G/L124M, W59G/ L128S) due certainly, to genetic recombination between mutated alleles. Indeed, in the geographical areas where double mutations on the same Vkorc1 allele were observed, mice carrying the corresponding single mutations in the homozygous state were systematically present.
For example in the Brittany region, mice homozygous for the W59G mutation, mice homozygous for the L124M mutation, and mice homozygous for the W59G/L124M double mutations are found and within the same site of capture, the three types of genome were also encountered.
The emergence of the double mutations (A26T/L128S, A26S/ L128S) confer an evident benefit for mice carrier for these double mutations compared to mice carrier for the corresponding single mutations ( Figure 5). While the single A26T, A26S, or L128S mutations lead to moderate resistance to first-generation AVK, the double mutations lead to resistance to all AVK currently available with resistance factors reaching levels higher than 10 toward difenacoum and brodifacoum and close to 10 toward difethialone for A26T/L128S ( Figure 5). Such levels of resistance toward such molecules has never been found for any of the isolated mutations observed in mice, but also in brown rats. The emergence of these double mutations must therefore be regarded as an adaptive response to the use of AVKs. Interestingly, our observation of mice heterozygous for A26T and Y139C or L128S and Y139C on different alleles can be considered as a necessary step prior to the recombination and makes likely the future recombination.
Nevertheless, such a recombination has not yet been demonstrated in living mice but their consequences, as shown in Figure 5 would be certainly important.
In this adaptive context, the probable use of the first-generation AVK for the management of the mice populations and the feeding behavior of this species lead to a high diversity of the mutations in the The recombination of mutations in rat seems unlikely due to the population structure and the small number of mutations encountered and the predominance of one mutation per geographical area.

ACKNOWLEDGMENTS
This work was supported by grant ISI no. I1301001W "NEORAMUS" from BPI France. We want to thank all the PCO who sent us tails of mice.

CONFLICT OF INTEREST
None declared.

DATA ACCESSIBILITY
Species identification and Vkorc1 genotype of mice trapped in this study has been deposited on Dryad (doi: 10.5061/dryad.bs888).