Transposon-mediated insertional mutagenesis unmasks recessive insecticide resistance in the aphid Myzus persicae

Abstract

The evolution of resistance to insecticides threatens the sustainable control of many of the world's most damaging insect crop pests and disease vectors. To effectively combat resistance, it is important to understand its underlying genetic architecture, including the type and number of genetic variants affecting resistance and their interactions with each other and the environment. While significant progress has been made in characterizing the individual genes or mutations leading to resistance, our understanding of how genetic variants interact to influence its phenotypic expression remains poor. Here, we uncover a mechanism of insecticide resistance resulting from transposon-mediated insertional mutagenesis of a genetically dominant but insecticide-susceptible allele that enables the adaptive potential of a previously unavailable recessive resistance allele to be unlocked. Specifically, we identify clones of the aphid pest Myzus persicae that carry a resistant allele of the essential voltage-gated sodium channel (VGSC) gene with the recessive M918T and L1014F resistance mutations, in combination with an allele lacking these mutations but carrying a Mutator-like element transposon insertion that disrupts the coding sequence of the VGSC. This results in the down-regulation of the dominant susceptible allele and monoallelic expression of the recessive resistant allele, rendering the clones resistant to the insecticide bifenthrin. These findings are a powerful example of how transposable elements can provide a source of evolutionary potential that can be revealed by environmental and genetic perturbation, with applied implications for the control of highly damaging insect pests.

Keywords: resistance; transposon; voltage-gated sodium channel.

Fig. 1.

Clones of M. persicae that are heterozygous for the kdr + skdr allele but differ in sensitivity to bifenthrin show equivalent expression of genes encoding key detoxification enzymes but marked differences in the expression of the kdr + skdr and wild-type alleles. (A) Number of genes significantly differentially expressed in comparisons of RNA-seq data obtained from different M. persicae clones (n = 4). (B) Venn diagram displaying differentially expressed genes shared in different comparisons of bifenthrin-susceptible M. persicae clones (1X and 4H) with resistant (88H2 and 62H2) clones. (C) Fold change in expression of E4/FE4-like esterase genes (MYZPE13164_G006_v1.0_000082180 and MYZPE13164_G006_v1.0_000124330) in M. persicae clone comparisons (n = 4). Significant differences (P < 0.05) in expression for each gene between clones is denoted using letters above bars as determined by EdgeR (55) (Dataset S1), letters without asterisks designate significance of expression levels of MYZPE13164_G006_v1.0_000082180, and letters with asterisks designate significance of expression levels of MYZPE13164_G006_v1.0_000124330. (D) Expression of alleles of the VGSC with or without the kdr and skdr mutations as assessed by the percentage of RNA-seq reads with or without these mutations mapping at the kdr and skdr loci.

Fig. 2.

A heterozygous MULE insertion is observed in the coding sequence of the VGSC subunit 1 gene in clones 88H2 and 62H2 that introduces a premature termination codon. (A) DNA-seq reads mapped to the region of the VGSC gene encoding domains IIS1 to IIS6. (B) Close up of soft-clipped reads that diverge in sequence at the start of the region of VGSC encoding the IIS1 transmembrane helix domain that identify a heterozygous insertion that results in the introduction of several premature termination codons. (C) Alignment of the insertion against four sequences (with 96.6 to 99.9% sequence identity to the query sequence) obtained by BLAST searches against the M. persicae G006 reference genome. The gray regions indicate similarity between sequences, and black regions indicate sequence differences. Indels are indicated by gaps in the sequences. The 109-bp imperfect terminal inverted repeats that flank both ends of the returned sequences are indicated using red boxes on the consensus sequence. (D) Features of the truncated and full-length versions of the MULE transposon.

Fig. 3.

The MULE insertion down-regulates the expression of the bifenthrin-susceptible allele of the VGSC. (A) Results of PCR amplification and sequencing of alleles of the VGSC with or without the MULE insertion. Gel shows the results of PCR amplification from DNA of different aphid clones using primers specific for alleles of the VGSC with or without the MULE insertion, as illustrated in accompanying schematic. NTC: no template control. DNA size marker is the GeneRuler 1 kb plus DNA ladder (Thermo Fisher). Chromatograms show the genotype at the skdr and kdr loci obtained for each PCR product following sequencing. (B) Schematic of the impact of the MULE insertion on expression of the allele of the VGSC without the kdr + skdr mutations in clones 62H2 and 88H2 compared to clone 4H, which is heterozygous for kdr + skdr but lacks the MULE insertion. Expression level was inferred from RNA-seq data (SI Appendix, Table S2).

Fig. 4.

Monoallelic expression of the VGSC subunit 1 reduces total expression levels and results in a temperature-dependent reproductive fitness cost. (A) qPCR analysis of the expression of the VGSC subunit 1 and 2 genes in M. persicae clones with (62H2 and 88H2) and without (1X and 4H) the MULE insertion. Expression is shown as fold change relative to clone 4H. The error bars show 95% CI (n = 4). Significant differences (P < 0.01) in expression between clones is denoted using letters above bars as determined by one-way ANOVA with post hoc testing (Tukey Honestly Significant Difference [HSD]). (B) Measurement of reproductive fitness and lifespan of M. persicae clones that are all heterozygous for kdr + skdr but carry this resistance allele in combination with an allele of the VGSC with (62H2 and 88H2) or without (C118 and 4H) the MULE insertion at 24 and 27.5 °C. The error bars display SD (n = 18). Significant differences (P < 0.01) in fitness measurement between clones at 27.5 °C is denoted using letters above bars as determined by one-way ANOVA with post hoc testing (Tukey HSD). No significant differences between clones for any fitness measurement were observed at 24 °C.