Genomic evolution of the pathogenic Wolbachia strain, wMelPop

Abstract

Most strains of the widespread endosymbiotic bacterium Wolbachia pipientis are benign or behave as reproductive parasites. The pathogenic strain wMelPop is a striking exception, however: it overreplicates in its insect hosts and causes severe life shortening. The mechanism of this pathogenesis is currently unknown. We have sequenced the genomes of three variants of wMelPop and of the closely related nonpathogenic strain wMelCS. We show that the genomes of wMelCS and wMelPop appear to be identical in the nonrepeat regions of the genome and differ detectably only by the triplication of a 19-kb region that is unlikely to be associated with life shortening, demonstrating that dramatic differences in the host phenotype caused by this endosymbiont may be the result of only minor genetic changes. We also compare the genomes of the original wMelPop strain from Drosophila melanogaster and two sequential derivatives, wMelPop-CLA and wMelPop-PGYP. To develop wMelPop as a novel biocontrol agent, it was first transinfected into and passaged in mosquito cell lines for approximately 3.5 years, generating wMelPop-CLA. This cell line-passaged strain was then transinfected into Aedes aegypti mosquitoes, creating wMelPop-PGYP, which was sequenced after 4 years in the insect host. We observe a rapid burst of genomic changes during cell line passaging, but no further mutations were detected after transinfection into mosquitoes, indicating either that host preadaptation had occurred in cell lines, that cell lines are a more selectively permissive environment than animal hosts, or both. Our results provide valuable data on the rates of genomic and phenotypic change in Wolbachia associated with host shifts over short time scales.

Keywords: Wolbachia; endosymbiont; evolution; genomics.

Fig. 1.—

Timeline of the history of the wMelPop strains described in this article. The Wolbachia strain wMelPop was purified from Drosophila melanogaster w¹¹¹⁸ and transinfected into the Aedes albopictus-derived cell line Aa23. After approximately 27 months of serial passaging, the Wolbachia infection was transferred to the RML12 cell line and passaged for a further 17 months, then transinfected into A. aegypti mosquitoes. This strain was also transinfected back into D. melanogaster w¹¹¹⁸ after approximately 35 months of cell-line passage; this strain, wMelPop-CLA, showed reduced pathogenesis in flies compared with the original wMelPop strain. We sequenced the genomes of three variants of popcorn: wMelPop from D. melanogaster w¹¹¹⁸, wMelPop-CLA after approximately 44 months of cell-line passage, and wMelPop-PGYP from A. aegypti approximately 48 months after transinfection into the mosquito.

Fig. 2.—

Maximum likelihood phylogeny based on an alignment of the genome sequences of wMel, wMelCS, and wMelPop strains. The two sequences labeled wMelPop and wMelPopCS/Canton-S were produced in this article. The other sequences used in the phylogeny are the published wMel genome (Wu et al. 2004), labeled AE017196, and ten consensus genome sequences generated by Richardson et al. (2012); roman numerals indicate the Wolbachia clades identified in that paper. wMelPop branches within the wMelCS clade (VI), separately from the wMel clades (I–IV).

Fig. 3.—

Sequencing coverage for wMelCS, wMelPop, wMelPop-CLA, and wMelPop-PGYP. Reads were mapped against the wMel genome using BWA or Newbler with default settings. For depth calculation, reads mapping to repeat regions were assigned to a randomly chosen instance of the repeat, and mean per-site coverage was calculated for nonoverlapping 50-nt windows along the genome. The region corresponding to the genes WD0507–WD0514 in the wMel genome is single copy in wMelCS, triplicated in wMelPop, and deleted in wMelPop-CLA and wMelPop-PGYP. The narrow peak of increased coverage visible in wMelPop-CLA and wMelPop-PGYP slightly downstream of this region represents the duplication of two ankyrin repeats in the orthologs of WD0550 in these strains. This repeat expansion is also present in wMelPop (confirmed by PCR), but is not apparent in the sequence coverage plot. Coverage along the genome is clearly more variable for wMelCS (100-nt Illumina reads) and wMel-CLA (shotgun 454 reads) than for the two strains sequenced with paired-end 454 reads.

Fig. 4.—

The qPCR analyses showing relative amplification of genes in the putative region of copy number variation (WD0512, WD0513, and WD0514, top two panels) and a control gene outside this region (WD1213, third panel), normalized against the single copy gene wsp. For genes WD0512, WD0513, and WD0514, normalized amplification is three times higher in wMelPop than in wMelCS, whereas there is no amplification in wMelPop-PGYP. There are no significant differences in normalized amplification between strains for the control gene WD1213. Note that amplification relative to wsp is dependent on primer efficiency, so values on the y axis do not represent copy number, and should only be compared across strains, not across genes.

Fig. 5.—

The genomic differences detected between wMelPop and the strain wMelPop-CLA derived from it through serial passaging in cell lines. (A) Insertion of an additional IS5 element between the orthologs of wMel genes WD0765 and WD0766. (B) Deletion of a 57-kb region corresponding to the triplicated orthologs of genes WD0507 to WD0514. (C) Two point mutations and one 10-nt deletion.

Fig. 6.—

(A) Timing of genomic changes during cell line passaging. Asterisks indicate the time points at which PCR assays were performed to test for the presence of each of the three structural genomic changes that occurred. Circular genome icons correspond to the symbols used in figure 5, with arrows labeled P indicating primers used for PCRs. The horizontal black lines above these icons show the period during which each mutation segregated in the population, from first detection to the time at which it was fixed. Small squares labeled a–d indicate time points shown in gel below. (B) Ethidium bromide gel showing amplification patterns of these three markers at time points a–d.