RDL mutations in Guangxi Anopheles sinensis populations along the China–Vietnam border: distribution frequency and evolutionary origin of A296S resistance allele

Anopheles sinensis adults were caught in seven different sites from April 2015 to August 2017. The seven sample-collecting sites were located in different villages along the China–Vietnam border. Rice is the main crop planted in these villages. The commonly used insecticides for rice pest control in these areas are diamides (e.g. chlorantraniliprole), neonicotinoids (e.g. imidacloprid) and organophosphorates (e.g. acephate and dimethoate). Methods about sample collection and species identification were described by Yang et al. [6].

A fragment of the AsRDL gene that includes codons 296 and 327 (named as RDL7) was amplified using primers (Primers ARE-7F and 7I-R, Table 1) and individual genomic DNA as template. The amplification reaction consisted of 25 μl of 2× Taq PCR Master Mix, 0.2 μM each primer, and 50–150 ng of DNA template in a total volume of 50 μl. The amplification was programmed as 95 °C for 5 min; 42 cycles of 95 °C for 30 s, 62 °C for 30 s, 72 °C for 70 s; and 72 °C for 10 min. PCR products were directly sequenced (TSINGKE Biotech, Beijing, China). For clone sequencing, three to ten clones were sequenced. The sequences from clone sequencing were cross-checked with those from direct PCR product sequencing to clarify the two haplotypes in each heterozygote.

Primers 8F-1 and 8R-1 (Table 1) were used to amplify a fragment of 167 bp covering codon 345 of the AsRDL gene (named as RDL8). The reaction mixture (30 μl) included 2× Taq PCR MasterMix, 0.2 μM each primer and 30–100 ng of genomic DNA. The reaction procedure was set as 95 °C for 5 min; 38 cycles of 95 °C for 30 s, 56 °C for 30 s, 72 °C for 10 s; and 72 °C for 10 min. PCR products were digested with the restricted DNA digestion enzyme HpyCH4 III (New England Biolabs) (Fig. 1a) for 2 h in a total volume of 20 μl according to the manufacturer’s instruction. The digestion products were detected by 3% agarose gel electrophoresis. The genotype of each sample was identified based on the resulting gel profiles: one band of 167 bp for homozygous TCG (SS), one band of combined 83 bp and 84 bp for homozygous ACG (TT) (complete digestion of the 167 bp DNA by HpyCH4 III resulted in two fragments with a length of 83 bp and 84 bp respectively, which could not be distinguished on a routine agarose gel); two bands (167 bp + 83 bp) for T/ACG (S/T) heterozygote (Fig. 1b).

Each sequencing result of RDL7 was manually checked and the terminal ambiguous regions were trimmed. All the confirmed sequences (360 bp) were aligned using Muscle 3.8 [14], and the polymorphic sites in the sequence were recorded. Hardy–Weinberg equilibrium (HWE) of genotypes in each population was tested by probability test using the online software GENEPOP v.4.2 [15]. The haplotypes of heterozygotes were clarified by clone sequencing. Haplotype network analysis was conducted using Network 5.0 [16].

Ten polymorphic sites (PS) were observed in RDL7 in a total length of 360 bp, three PSs in exon 7 and seven in intron 7 (Fig. 2). The 2nd and 3rd PSs represented synonymous mutations, causing amino acid substitutions at sites 296 (A296S) and 327 (V327I), respectively.

The distribution and frequency of the three insecticide resistance-related mutations are shown in Table 2 and Fig. 3. Overall, the frequency of A296S was very high (79–100%) in the seven examined populations. Notably, this allele was fixed in four of the seven populations. By contrast, the frequency of the susceptible allele was rare: A296 was only detected in three populations (PXSS, JXTD, JXXW) at frequencies ranging from 1.5 to 21%.

The three possible genotypes for site 327 were detected in all the seven populations, and all genotypes agreed with Hardy–Weinberg equilibrium (Table 2). The frequency of the mutant 327I ranged from 26.9 to 53.2% (Fig. 3). The mutant allele (345S) was detected in five of the seven populations, but absent in DXEC and CZSK. Overall, the frequency of 345S was low (0–28.8%), with the highest frequency being observed in FCGDX (Fig. 3).

When considering the three resistance-related sites together, 11 types of triple-locus genotypes (C1 to C11) were observed (Table 3). The number of triple-locus genotype combinations observed in each sampling site ranged from 3 (CZSK) to 9 (JXTD). Three combinations (C1, C4, C8) were distributed in all the seven populations, while C6 and C7 were uniquely found in PXSS and FCGDX, respectively. The triple-locus wild genotype (C11) was distributed in JXTD and JXXW. The alleles carrying 327I were found only in individuals that harbor 296S, and 345S existed in individuals that were homozygous for 296S. Notably, the triple-locus mutant allele was detected (C6 and C7), although at low frequencies.

To track the possible origin of mutant 296S and 327I, the DNA sequences of 360 bp-in-length (Fig. 2) were used for haplotype identification and network analysis. A total of 21 haplotypes (H1 to H21) were identified (Table 4). Among them, 4 (H1 to H4) were wild haplotypes without 296S and 327I mutations, 11 (H5–H15) were haplotypes with only the 296S mutation, and 6 (H16–H21) carried both 296S and 327I mutations. The 327I mutation was found always together with 296S. Network analysis indicated that the 296S allele might have evolved independently from more than one origin (Fig. 4). The single mutation haplotypes H13 and H10 may be evolved from H2 and H3, respectively, while the five double-mutation haplotypes (H17–H21) could be derived from some 296S haplotypes via only one mutational step (Fig. 4).