Genetic stability of Aedes aegypti populations following invasion by wMel Wolbachia

Abstract

Background: Wolbachia wMel is the most commonly used strain in rear and release strategies for Aedes aegypti mosquitoes that aim to inhibit the transmission of arboviruses such as dengue, Zika, Chikungunya and yellow fever. However, the long-term establishment of wMel in natural Ae. aegypti populations raises concerns that interactions between Wolbachia wMel and Ae. aegypti may lead to changes in the host genome, which could affect useful attributes of Wolbachia that allow it to invade and suppress disease transmission.

Results: We applied an evolve-and-resequence approach to study genome-wide genetic changes in Ae. aegypti from the Cairns region, Australia, where Wolbachia wMel was first introduced more than 10 years ago. Mosquito samples were collected at three different time points in Gordonvale, Australia, covering the phase before (2010) and after (2013 and 2018) Wolbachia releases. An additional three locations where Wolbachia replacement happened at different times across the last decade were also sampled in 2018. We found that the genomes of mosquito populations mostly remained stable after Wolbachia release, with population differences tending to reflect the geographic location of the populations rather than Wolbachia infection status. However, outlier analysis suggests that Wolbachia may have had an influence on some genes related to immune response, development, recognition and behavior.

Conclusions: Ae. aegypti populations remained geographically distinct after Wolbachia wMel releases in North Australia despite their Wolbachia infection status. At some specific genomic loci, we found signs of selection associated with Wolbachia, suggesting potential evolutionary impacts can happen in the future and further monitoring is warranted.

Keywords: Aedes aegypti; Evolve-and-resequence; Genome-wide association study; Population replacement; Wolbachia.

Fig. 1

Locations of sampled Aedes aegypti populations. Samples from Gordonvale were collected in 2010, 2013 and 2018, while samples from other locations were collected in 2018. Axes show longitude (lon) and latitude (lat). The map is derived from ggmap package in R, interface to Google Maps application program interface (API) (https://developers.google.com/maps/)

Fig. 2

LOESS-smoothed curves of genome-wide nucleotide diversity (π). Six populations of Ae. aegypti measured in 10 kbp non-overlapping windows. GV10 and GV13 represent samples collected in 2010 and 2013 from Gordonvale; GV18, EH, YK and RL represent samples collected in 2018 from Gordonvale, Yorkeys Knob, Edge Hill and Redlynch respectively

Fig. 3

Density distributions of Tajima’s D values. Tajima’s D measured at (a) the genome level measured in 10 kbp non-overlapping windows and at (b) the gene level. GV10 and GV13 represent samples collected in 2010 and 2013 from Gordonvale; GV18, EH, YK and RL represent samples collected in 2018 from Gordonvale, Yorkeys Knob, Edge Hill and Redlynch respectively

Fig. 4

Principal components analysis based on MAF or pairwise Fst. PCA plots of (a) allele frequency of Aedes aegypti samples (MAF > 0.1, minimal coverage > 50); (b, d) pairwise Fst throughout the genome with 100 kbp non-overlapping windows, and (c, e) pairwise Fst within genes. Colors in (b, c) represent geographical comparisons. The blue color in (d, e) represents comparisons between Wolbachia-infected and uninfected samples, while red represents comparisons within Wolbachia-infected samples and green represents the comparison within uninfected samples. The 95% confidence ellipses show data clustering

Fig. 5

Manhattan plot of SNPs in covariate model representing Wolbachia infection status. Points represent SNPs with a positive Bayes Factor (BF) in covariate model. Orange labels represent genes and their positions associated with “strongly” outliers that identified from the combination of two Bayesian models