Comparative Genomics Reveals Factors Associated with Phenotypic Expression of Wolbachia

Abstract

Wolbachia is a widespread, vertically transmitted bacterial endosymbiont known for manipulating arthropod reproduction. Its most common form of reproductive manipulation is cytoplasmic incompatibility (CI), observed when a modification in the male sperm leads to embryonic lethality unless a compatible rescue factor is present in the female egg. CI attracts scientific attention due to its implications for host speciation and in the use of Wolbachia for controlling vector-borne diseases. However, our understanding of CI is complicated by the complexity of the phenotype, whose expression depends on both symbiont and host factors. In the present study, we perform a comparative analysis of nine complete Wolbachia genomes with known CI properties in the same genetic host background, Drosophila simulans STC. We describe genetic differences between closely related strains and uncover evidence that phages and other mobile elements contribute to the rapid evolution of both genomes and phenotypes of Wolbachia. Additionally, we identify both known and novel genes associated with the modification and rescue functions of CI. We combine our observations with published phenotypic information and discuss how variability in cif genes, novel CI-associated genes, and Wolbachia titer might contribute to poorly understood aspects of CI such as strength and bidirectional incompatibility. We speculate that high titer CI strains could be better at invading new hosts already infected with a CI Wolbachia, due to a higher rescue potential, and suggest that titer might thus be a relevant parameter to consider for future strategies using CI Wolbachia in biological control.

Keywords: Drosophila; Wolbachia; comparative genomics; cytoplasmic incompatibility; genomics; symbiosis.

Fig. 1.

Phylogenetic relationship between the nine Wolbachia genomes in this study. Maximum likelihood tree based on the concatenated alignment of 714 orthologous single copy genes. All nodes have bootstrap values of 100. Branches of the SYTMA clade (in orange) were lengthened to more clearly show their internal relationships; they follow the orange scale bar. The branch connecting supergroups A and B was halved. The numbers shown on each node indicate the number of protein clusters that are unique to each subclade or node (white boxes), unique between the close relatives wSY and wTei or wMa and wNo in comparison to each other (light gray boxes), or duplicated in wSY or wTei in comparison to each other (dark gray boxes). The “Mod” and “Resc” columns show if a strain is capable (+) or uncapable (−) of mod and resc in D. simulans. CI strength is indicated as low (+), medium (++) or strong (+++). Colored arrows connecting the “Resc” and “Mod” columns show the ability of a strain to fully (full lines) or partially (dashed lines) rescue the modification induced by wTei (blue), wMel (green), wRi (orange), and wNo (gray). Data for CI strength, inter-strain compatibility, resc and mod phenotypes are summarized from Martinez et al. (2015) and Zabalou et al. (2008).

Fig. 2.

The Octomom and neighboring phage regions. (A) Comparison of the Octomom (blue) and neighboring phage regions (orange) in the SYTM genomes. The wMel WD0513 gene and its homologs in wSYT are shown in pink. Other genes are represented in light gray and mobile elements in dark gray. Pseudogenes are marked by diagonal lines. Similarity between sequences is indicated by gray lines, where darker is more similar. Blastn was used for comparisons between wSYT genomes and tblastx was used for the comparison between wTei and wMel. (B) Maximum likelihood tree of the WD0513 protein of wMel and homologs from the wSYT genomes as well as other species identified through blast searches in the nr database. Bootstrap values are shown on nodes. The tree was midpoint-rooted in Figtree. Accession numbers for the proteins featured in the tree are available in supplementary table S5.

Fig. 3.

The Dozen Island. The wSYT Dozen Island genes and their homologs in wCon and wLug are shown in blue. Other genes are shown in light gray and mobile elements in dark gray. Pseudogenes are marked by diagonal lines. Similarity between sequences is indicated by gray lines, where darker is more similar.

Fig. 4.

The CI-associated cifA and cifB genes. (A) Maximum likelihood trees of cifB and cifA homologs showing five clades that correspond to Types I–V, as indicated by labels and colors. Homologs found in our nine genomes are shown in bold. Trees were midpoint-rooted in Figtree. Bootstrap values below 70 are not shown. (B) Comparison of cifAB homologs in our nine Wolbachia genomes. Distinct colors identify CifAB homologs of different Types, with Type I in blue, Type II in orange, Type III in yellow, and Type IV in green. For each cifAB pair, cifA is shown in a lighter tone than cifB. Other genes are represented in light gray and mobile elements in dark gray. Pseudogenes are marked by diagonal lines. Similarity between sequences is indicated by gray lines, where darker is more similar.

Fig. 5.

Comparison of the genomic region containing the CI-associated gene WD0462. Homologs of WD0462 are shown in yellow, with the predicted HAUS Augmin3 domain (PF14932) indicated in dark yellow. The neighboring gene WD0463 is shown in green, with the predicted AAA-ATPase domain (PF00004) indicated in dark green. The flanking genes pyrF and valS are shown in orange and blue, respectively. Other genes are represented in light gray and mobile elements in dark gray. Pseudogenes are marked by diagonal lines. Domain predictions with scores below the significance threshold are marked with horizontal lines. Similarity between sequences is indicated by gray lines, where darker is more similar.