Nested patterns of commensals and endosymbionts in microbial communities of mosquito vectors

Abstract

Background: Mosquitoes serve as vectors for numerous pathogens, posing significant health risks to humans and animals. Understanding the complex interactions within mosquito microbiota is crucial for deciphering vector-pathogen dynamics and developing effective disease management strategies. Here, we investigated the nested patterns of Wolbachia endosymbionts and Escherichia-Shigella within the microbiota of laboratory-reared Culex pipiens f. molestus and Culex quinquefasciatus mosquitoes. We hypothesized that Wolbachia would exhibit a structured pattern reflective of its co-evolved relationship with both mosquito species, while Escherichia-Shigella would display a more dynamic pattern influenced by environmental factors.

Results: Our analysis revealed different microbial compositions between the two mosquito species, although some microorganisms were common to both. Network analysis revealed distinct community structures and interaction patterns for these bacteria in the microbiota of each mosquito species. Escherichia-Shigella appeared prominently within major network modules in both mosquito species, particularly in module P4 of Cx. pipiens f. molestus, interacting with 93 nodes, and in module Q3 of Cx. quinquefasciatus, interacting with 161 nodes, sharing 55 nodes across both species. On the other hand, Wolbachia appeared in disparate modules: module P3 in Cx. pipiens f. molestus and a distinct module with a single additional taxon in Cx. quinquefasciatus, showing species-specific interactions and no shared taxa. Through computer simulations, we evaluated how the removal of Wolbachia or Escherichia-Shigella affects network robustness. In Cx. pipiens f. molestus, removal of Wolbachia led to a decrease in network connectivity, while Escherichia-Shigella removal had a minimal impact. Conversely, in Cx. quinquefasciatus, removal of Escherichia-Shigella resulted in decreased network stability, whereas Wolbachia removal had minimal effect.

Conclusions: Contrary to our hypothesis, the findings indicate that Wolbachia displays a more dynamic pattern of associations within the microbiota of Culex pipiens f. molestus and Culex quinquefasciatus mosquitoes, than Escherichia-Shigella. The differential effects on network robustness upon Wolbachia or Escherichia-Shigella removal suggest that these bacteria play distinct roles in maintaining community stability within the microbiota of the two mosquito species.

Keywords: Escherichia-Shigella; Wolbachia; Community assembly; Mosquitoes; Nestedness theory.

Fig. 1

Differences in mosquito microbiota diversity and community assembly between Culex pipiens f. molestus and Culex quinquefasciatus. (A) Shannon diversity and (B) Pielou’s evenness indices showed significant differences between microbiota of Cx. pipiens f. molestus and Cx. quinquefasciatus. (C) Venn diagram showing the number of bacterial taxa that are shared or unique among the networks of two mosquito species. *p < 0.05 (D) Beta diversity of mosquito microbiota of two species represented in PCoA plot obtained by Betadisper function. There are significant differences in dispersions (variances) (ANOVA, p < 0.01). (E) Heatmap representing the abundance (expressed as *Centered Log-Ratio) of the 10 taxa whose abundance was higher in Cx. pipiens f. molestus group and 16 taxa whose abundance was higher in Cx. quinquefasciatus. (F, G) Bacterial co-occurrence networks were inferred from 16SrRNA sequences obtained from laboratory reared mosquitoes of two species (F) Cx. pipiens f. molestus and (G) Cx. quinquefasciatus (SparCC > 0.5 or < -0.5). Nodes correspond to taxa (family or genus level). The colours of nodes specify modules in which taxa occur. The size of nodes is related to their eigenvector centrality, the bigger the node, the higher eigenvector centrality value it has. Positive (purple) and negative (coral) correlations are shown by the colour of the edges. (H) Core Association Network (CAN) (SparCC > 0.75 or < -0.75). Positive correlations are shown by purple edges. Nodes correspond to taxa. (I) Differential network between Cx. pipiens f. molestus and Cx. quinquefasciatus natural networks illustrating the correlations that vary between identical taxa in two bacterial networks. Grey nodes represent taxa, and edges represent differential associations between taxa. P^* - Cx. pipiens f. molestus; Q^* - Cx. quinquefasciatus; CLR* - Centered log ratio

Fig. 2

Differences in mosquito bacterial modules between Culex pipiens f. molestus and Culex quinquefasciatus. (A, B) Bacterial co-occurrence networks of (A) Cx. pipiens f. molestus and (B) Cx. quinquefasciatus divided by modules. Node colours are based on modularity class metric, each module is represented by a different colour. Grey colored nodes represents single node modules. The size of nodes is related to their eigenvector centrality, the bigger the node, the higher eigenvector centrality value it has. Positive (purple) or negative (coral) correlations are shown by the colour of the edges. Bacterial taxa (family or genus level) with at least one connection are symbolized by nodes, whilst connected edges represent correlations between them (SparCC ≥ 0.5 or ≤ -0.5). (C, D) Co-occurrence networks in the same layout were extrapolated from the microbiota of (C) Cx. pipiens f. molestus and (D) Cx. quinquefasciatus. Bacterial taxa (family or genus level) with at least one connection are symbolized by nodes, whilst connected edges represent a significant correlation between them (SparCC ≥ 0.75 or ≤ -0.75). Node colours are based on determined clusters and sized according to the node’s eigenvector centrality. Positive (purple) or negative (red) correlations are shown by the colour of the edges

Fig. 3

Differences in the local connectivity of Escherichia-Shigella and Wolbachia between Culex pipiens f. molestus and Culex quinquefasciatus microbiota. (A, B) Sub-networks of the local connectivity of Escherichia-Shigella and (C, D) Wolbachia were extracted from (A, C) Cx. pipiens f. molestus and (B, D) Cx. quinquefasciatus natural networks. The size of nodes is related to their eigenvector centrality, the bigger the node, the higher eigenvector centrality value it has. Positive (purple) or negative (coral) correlations are shown by the colour of the edges. Bacterial taxa (family or genus level) with at least one connection are symbolized by nodes, whilst connected edges represent correlations between them (SparCC ≥ 0.5 or ≤ -0.5)

Fig. 4

Differences in the microbial network response to node removal and addition between Culex pipiens f. molestus and Culex quinquefasciatus. (A, D, G) The resistance to cascading attack was measured and compared between Cx. pipiens f. molestus and Cx. quinquefasciatus (A) natural networks, (D) networks without Escherichia-Shigella or (G) Wolbachia. The robustness to nodes addition (from 0 to 1000) based on the size of the largest connected component (LCC) (B, E, H) and average path length (avg. path length) (C, F, I) was measured and compared between Cx. pipiens f. molestus and Cx. quinquefasciatus natural networks (B, C), networks without Escherichia-Shigella (E, F) and Wolbachia (H, I). Different curve colours represent different groups