In this study, we sought to determine the prevalence of anticoagulant rodenticide resistance in Singapore to inform rodent management strategies. We did not find any evidence of anticoagulant rodenticide resistance in the R. norvegicus populations. Moreover, SNPS were absent for the whole of exon 3 for all Norway rat samples, identical to that of the wild-type. This finding was surprising, given that R. norvegicus was the most common species of rodents found in urban environments of Singapore [38], and that anticoagulant rodenticide has been widely used by pest control operators (PCOs) to treat heavily rodent infested places in the last 20 years. The findings in our study for R. norvegicus differed from others undertaken in Netherlands, France, England, Germany, Hungary, Azores, USA and Argentina [12, 14, 15, 28, 30]. The majority of the missense SNPs proven to cause resistance in Norway rats that were detected in the European regions were Tyr139Cys, Tyr139Phe, Leu128Gln, Leu128Ser and Leu120Gln. Interestingly, R. norvegicus samples in Indonesia and Thailand (which are located in Southeast Asia) were reported to contain both missense and silent mutations [12]. A possible explanation for the absence of the above-mentioned SNPs in Singapore is that the urban conditions could be relatively harsh for Norway rats to thrive. In Singapore, there are well established refuse management and public cleaning programmes that reduces access to food waste and keeps the environment clean. Strict laws and regulations are in place to minimize littering, poor waste management and food hygiene practices in food retail establishments that can generate food sources to facilitate rodent population growth. This reduces the accessibility to food sources, leading to tougher competition amongst rat colonies. Even if spontaneous resistance mutations occurred within a few individuals of the rat population, they might not be fit enough to survive until they produce the next generation of pups. Similarly, a good and comprehensive rodent control programme that comprises of pulsed-baiting techniques [39] and alternating between types of rodenticide usage and physical trapping strategy may also result in low resistance frequency.
In contrast, the Roof rats contained multiple SNPs within exon 3 itself comprising of a variety of SNPs that are either silent or missense mutations. Although SNPs were not detected in the widely studied codons of 139, 128 or 120 that have been proven to confer resistance [20, 25, 26], it was interesting to find other mutations and this suggests at the possibility of local rat populations in acquiring rodenticide resistance phenotypes. In our study, roof rats carried multiple silent mutations within exon 3 of Vkorc1, namely at codons Ser103Ser, Ile107Ile, Leu108Leu and Thr137Thr. This essentially means the nucleotide variants do not result in a substitution of amino acid, allowing the VKORC1 enzyme to retain its protein structure and function. The two most common silent mutations observed were Ile107Ile and Thr137Thr and this was found in all of the Roof rats carrying both SNPs in homozygous form. In fact, all of the R. rattus samples tested in New Zealand and India demonstrated the exact same pattern of silent mutation at both of these codons [36, 40]. One reason for this finding could be the outcome of genetic bottleneck events or founder effect that occurred in the past. As SNPs in the Vkorc1 gene is heritable from parents [15, 16], all that is required is for a single founder individual having these two SNPs to pass down the mutations to modern day rat populations. Due to Singapore’s strategic location as a trading hub since the 19th century, ship vessels frequent this island and it is likely that stowaway rats from neighbouring countries are imported many years ago. Nevertheless, it is interesting to note that R. norvegicus from Indonesia and Thailand have also been reported to exhibit similar SNP mutation patterns at codons 107 and 137 [12]. Hence, the silent mutations in these two codons are not only exclusive to R. rattus, but also able to occur in R. norvegicus as well.
A new silent SNP mutation that has not been described before was identified in this study. It was located in codon 108 with the substitution of a single cytosine nucleotide to tyrosine. About a quarter of the Roof rat population possess the Leu108Leu variant, which is quite substantial given that there is no substitution of amino acid involved. The final silent mutation detected in this study was the Ser103Ser variant. What makes this SNP interesting is that all the Roof rats with it simultaneously also carried the Ala143Val SNP, the only amino acid substitution observed in this study. Rats that are heterozygous for the SNP at codon 103 were observed to be heterozygous for the Ala143Val variant and likewise if it was homozygous. It is quite likely that the Ala143Val genotype has a strong association with the 103Ser103 silent mutation in exon 3. Although there were only eight samples exhibiting this genotype amongst the roof rats, they were obtained from the four corner regions of Singapore with distances of 6km or more apart. Therefore, it is quite unlikely that these groups of roof rats were related in any manner. As this genotype is known to be seen in the wild type VKORC1 protein of other species such as humans and mice, it is considered to be a neutral mutation [12]. However, a slightly more recent study in Indonesia in 2013 tested the Asian House Rats, R. tanezumi, sampled from areas with intensive coumatetralyl usage had comparable genotypic patterns. It was reported that the Ala143Val mutation was found in 9 rats and 7 of them displayed resistant phenotypes [31]. Further research is required to elucidate the mechanism by which this amino acid substitution truly confers resistance.
Molecular analysis of mitochondrial DNA from fossils found in present day traced that Norway rats historically originated from Southwestern China about 1.3 million years ago [41]. Also, for the Roof rats, a strong phylogeographic pattern indicates that they were native to South Asia and Indochina regions [35]. Despite studies depicting these two species having Asian origins, rodenticide resistance was first described in Europe and then United States about one to two decades after the introduction of rodenticide in the early 1950s [16, 42, 43]. This strongly suggests that the evolution of anticoagulant resistance could be the result of massive or improper use of warfarin in the western regions of the world where it was developed and used. As such, the exon 3 SNP profile of rats in Singapore was unexpected, since rodenticides have been used for the past 20 years. Likewise for China, anticoagulant rodenticides have been widely used for over 30 years and yet the frequency of resistance has been reported to be low [44].
Vkorc1 has been extensively studied and alterations to the protein structure is known to affect the blood clotting mechanism. Future studies can include investigating SNP mutation profiles of exon 1 and 2 for other non-synonymous mutations, coupled along with blood clotting response (BCR) test. It will also be interesting to investigate rodents’ ability in metabolising the ingested rodenticide by measuring gene expression of cytochrome P450 gene. The P450 enzyme breaks down a wide range of drugs and an enhanced expression of certain P450 genes can increase drug metabolism rate [45]. This could result in anticoagulants being cleared from their cardiovascular system quickly, causing rodenticides to be ineffective.
Our results indicate that the prevalence of SNP mutations in the rat population in Singapore was low and it is reassuring that there is no widespread resistance. Nonetheless, further studies to provide a better geographical representation and spatial resolution of assessed rodenticide resistance would benefit the review of site-specific intervention strategies. Regular monitoring of the extent and evolution of mutations in rodent populations would facilitate the review of resistance management strategies.
Given that this was the first study carried out in Singapore since the registration and use of rodenticides, we did not have historical information on the status of anticoagulant resistance or types of SNP mutations for comparison over time. We obtained our samples through convenience sampling and our results many not be generalizable to the entire population of rodents in Singapore. However, we have no reason to believe that rodenticide resistance in the areas from which the rodent samples were obtained had lower resistance compared to other areas that had no samples since population activity was higher and control activities more extensive in the former. While screening for SNPs in Vkorc1 does provide some information on the mutations associated with resistance, it is a proxy of the actual level of resistance towards rodenticides.