Global evaluation of taxonomic relationships and admixture within the Culex pipiens complex of mosquitoes

Abstract

Background: Within the Culex pipiens mosquito complex, there are six contemporarily recognized taxa: Cx. quinquefasciatus, Cx. pipiens f. pipiens, Cx. pipiens f. molestus, Cx. pipiens pallens, Cx. australicus and Cx. globocoxitus. Many phylogenetic aspects within this complex have eluded resolution, such as the relationship of the two Australian endemic taxa to the other four members, as well as the evolutionary origins and taxonomic status of Cx. pipiens pallens and Cx. pipiens f. molestus. Ultimately, insights into lineage relationships within the complex will facilitate a better understanding of differential disease transmission by these mosquitoes. To this end, we have combined publicly available data with our own sequencing efforts to examine these questions.

Results: We found that the two Australian endemic complex members, Cx. australicus and Cx. globocoxitus, comprise a monophyletic group, are genetically distinct, and are most closely related to the cosmopolitan Cx. quinquefasciatus. Our results also show that Cx. pipiens pallens is genetically distinct, but may have arisen from past hybridization. Lastly, we observed complicated patterns of genetic differentiation within and between Cx. pipiens f. pipiens and Cx. pipiens f. molestus.

Conclusions: Two Australian endemic Culex taxa, Cx. australicus and Cx. globocoxitus, belong within the Cx. pipiens complex, but have a relatively older evolutionary origin. They likely diverged from Cx. quinquefasciatus after its colonization of Australia. The taxon Cx. pipiens pallens is a distinct evolutionary entity that likely arose from past hybridization between Cx. quinquefasciatus and Cx. pipiens f. pipiens/Cx. pipiens f. molestus. Our results do not suggest it derives from ongoing hybridization. Finally, genetic differentiation within the Cx. pipiens f. pipiens and Cx. pipiens f. molestus samples suggests that they collectively form two separate geographic clades, one in North America and one in Europe and the Mediterranean. This may indicate that the Cx. pipiens f. molestus form has two distinct origins, arising from Cx. pipiens f. pipiens in each region. However, ongoing genetic exchange within and between these taxa have obscured their evolutionary histories, and could also explain the absence of monophyly among our samples. Overall, this work suggests many avenues that warrant further investigation.

Keywords: Culicidae; Disease vector; Genetic exchange; Mosquito; Population structure; Species complex.

Fig. 1

Principal components analysis (PCA) using four-fold degenerate sites with reported samples from all six described members of the Culex pipiens complex (a) and with a four-taxon set that excluded the reported Australian endemic taxa, australicus and globocoxitus (b). These PCAs were implemented with PLINK and plotted in R. Shown are the first two PCs. Colors corresponding to the different reported taxa are consistent between the two PCAs

Fig. 2

World maps showing the described collection locations of samples (small circles inside gray boxes) and the relative proportions of three (a) or five (b) inferred populations as determined in our ADMIXTURE analysis (large circles), using four-fold degenerate sites. Each sample’s taxonomic designation was based on that reported in the literature (see Table 1, Additional file 1: Table S1). For the ADMIXTURE results the proportion of each color in the circle corresponds to the amount of cluster-associated ancestry. Note that for our sample designations, we defined five broad geographical regions, indicated on the map by the dashed gray boxes

Fig. 3

STRUCTURE bar plots for the samples in our subsampled dataset plotted for genetic clusters (K) from two through seven, using four-fold degenerate sites. Each horizontal bar represents one sample. The relative proportions of each color indicate the proportion of genetic diversity assigned to that cluster. Sample designations are reported along the left y-axis. Taxon groups are reported along the right y-axis. The two best-supported K values are given in black text at the bottom (K = 3 for Evanno’s DK; K = 5 for median posterior probability). For additional sample details, see Additional file 1: Table S1

Fig. 4

Maximum likelihood phylogeny using four-fold degenerate sites and a transversional mutation model with a proportion of invariable sites and a gamma distribution of rate heterogeneity (TVM + I + Γ; [79]). The colors for the branch tip labels correspond to the six different taxa in this study. The numbers at the major branch nodes indicate bootstrap support for each bifurcation in the tree (out of 100). The three-letter code in the middle of each sample name indicates its geographical region of origin (see Additional file 1: Table S1 for additional sample details). Samples under a broad dashed line were determined to be intra-taxonomically admixed (pipiens and molestus only). Samples under a fine dashed line were determined to be inter-taxonomically admixed. Within the pipiens and molestus samples, three broad geographical clusters are defined: North America, Mediterranean and northern Europe (including Russia)