Extensive duplication of the Wolbachia DNA in chromosome four of Drosophila ananassae

Abstract

Background: Lateral gene transfer (LGT) from bacterial Wolbachia endosymbionts has been detected in ~20% of arthropod and nematode genome sequencing projects. Many of these transfers are large and contain a substantial part of the Wolbachia genome.

Results: Here, we re-sequenced three D. ananassae genomes from Asia and the Pacific that contain large LGTs from Wolbachia. We find that multiple copies of the Wolbachia genome are transferred to the Drosophila nuclear genome in all three lines. In the D. ananassae line from Indonesia, the copies of Wolbachia DNA in the nuclear genome are nearly identical in size and sequence yielding an even coverage of mapped reads over the Wolbachia genome. In contrast, the D. ananassae lines from Hawaii and India show an uneven coverage of mapped reads over the Wolbachia genome suggesting that different parts of these LGTs are present in different copy numbers. In the Hawaii line, we find that this LGT is underrepresented in third instar larvae indicative of being heterochromatic. Fluorescence in situ hybridization of mitotic chromosomes confirms that the LGT in the Hawaii line is heterochromatic and represents ~20% of the sequence on chromosome 4 (dot chromosome, Muller element F).

Conclusions: This collection of related lines contain large lateral gene transfers composed of multiple Wolbachia genomes that constitute >2% of the D. ananassae genome (~5 Mbp) and partially explain the abnormally large size of chromosome 4 in D. ananassae.

Figure 1

Coverage of Hawaii sequencing data. The copy number for the 3 kbp mate pair library from the Hawaii genomic DNA was calculated over a 1 kbp window every 500 bp and plotted against the reference wRi genome (left) and the first 1.5 Mbp of the largest scaffold (gi|109914400|gb|CH902617.1|) of the Drosophila ananassae caf1 assembly. Approximately 5× coverage was expected to represent single copy Drosophila genes and single copy regions are apparent. However, most of the nuwt is present with at least two copies.

Figure 2

Confirmation of coverage metrics using a qPCR based assay. The copy number was confirmed and found to be accurate up to 8 copies per haploid genome using a qPCR assay measuring the ratio of the LGT abundance to that of six single copy Drosophila genes as shown for the 3 kbp mate pair library for the cured Hawaii (Panel A), India (Panel B), and Indonesia (Panel C). A significant positive correlation (p < 0.001, linear regression) is found for the cured Hawaii and the India line The results in Panels A and C represent the accumulation of multiple qPCR experiments while those in Panel B result from a single experiment.

Figure 3

Coverage of India sequencing data. The copy number for the 3 kbp mate pair library from the India (Panels A and B), and India x Florida (Panels C and D) genomic DNA was calculated over a 1 kbp window every 500 bp and plotted against the reference wRi genome (left) and the first 1.5 Mbp of the largest scaffold (gi|109914400|gb|CH902617.1|) of the Drosophila ananassae caf1 assembly. The India genome shows the same relative pattern of duplication but is less abundant than single copy nuclear genes, as is the genome from offspring of a backcross of India with Florida. The spikes in the coverage for the India × Florida line reflect the mapping of transposase sequences.

Figure 4

Comparison of the coverage patterns. The coverage of the Indonesia genome is shown relative to that of the Hawaii line by dividing the values for Hawaii line by those of the Indonesia line across the wRi genome (Panel A) and the Drosophila ananassae genome (Panel B). The coverage of the India line relative to the Hawaii line is illustrated in the same manner (Panels C and D). While the India line has the same pattern of duplication as Hawaii, that pattern is different in the Indonesia line.

Figure 5

Coverage of Indonesia sequencing data. The copy number for the 3 kbp mate pair library from the Indonesia genomic DNA was calculated over a 1 kbp window every 500 bp and plotted against the reference wRi genome (left) and the first 1.5 Mbp of the largest scaffold (gi|109914400|gb|CH902617.1|) of the Drosophila ananassae caf1 assembly. The Indonesia line has a nuwt with even coverage similar to the rest of the Drosophila genome. The spikes in the coverage for the Indonesia line reflect the mapping of conserved DNA sequences (e.g. those from rRNA or tRNA) from bacterial contaminants like those in the guts of adult flies.

Figure 6

Comparison of copy number of the nuwts in homozygous and heterozygous fly lines and cured v. uncured flies. Cured (Panel A) and uncured (Panel B) Hawaii flies show the same correlation of qPCR results and sequencing coverage, although the qPCR results are shifted by the amplification of genes in the bacterial genome as they are plotted against the sequencing coverage obtained from the cured insects. This indicates that the duplication of the nuwt does not occur during the tetracycline treatment of the flies to compensate for the loss of the infection. The results of the qPCR assay for coverage was compared for homozygous cured Hawaii (Panels A) and the offspring of a Hawaii/Mexico cross which is heterozygous for the nuwt (Panels C) for the 3 kbp mate pair library library. Should the Hawaii line be heterozygous, the cross should have resulted in 50% of the offspring being homozygous for absence of single copy regions of the nuwt. Instead, each of the 44 offspring (half with Hawaii mothers and Mexico fathers and half from the reciprocal cross) tested positive by PCR for single copy genes in the nuwt and the nuwt was at half of the abundance of the Hawaii line relative to internal single copy Drosophila nuclear gene controls in 4 of these as seen by a shifting of the y-intercept by ~1 ΔCt.

Figure 7

Comparison of copy number of the nuwts through the life stages. The copy number from the genome sequencing of adult Hawaii flies have a significant positive correlation to the qPCR results obtained for embryos (Panel A), 3^rd instar larvae (Panel B), pupae (Panel C), and adults (Panels D). While the slopes for embryos, pupae, and adults are approximately one, the slope for 3^rd instar larvae is almost four suggesting that the nuwts are 4-fold underreplicated in 3^rd instar larvae.

Figure 8

Fluorescence in situ hybridization of polytene chromosomes. Polytene chromosomes were hybridized with a Wolbachia-specific WRi_004290 probe amplified directly from genomic DNA from the Wb- Hawaii line and visualized at 40X magnification. A single band is observed and not the 8–10 bands expected (Panel A). Polytene chromosomes were again hybridized with the WRi_004290 probe amplified directly from genomic DNA from the Wb- Hawaii line and visualized at 20X (Panel B) and 40X (Panel C). A single band is observed and not the 8–10 bands expected. Polytene chromosomes were simultaneously hybridized with differentially-labelled probes for the Wolbachia-specific WRi_004290 (Panel D, pink) and a D. ananassae actin gene (Panels E &F, green) that were generated from sequence-verified plasmids and visualized at 40X. While hybridization by the plasmid-derived WRi_004290 probe is not detected, hybridization by the plasmid-derived actin probe is detected. This suggests that the previous detection with Wolbachia probes amplified directly from genomic DNA may be detecting hybridization of spurious amplification products and that the lateral gene transfer is heterochromatic.

Figure 9

Fluorescence in situ hybridization of mitotic chromosomes. Mitotic chromosomes were hybridized with a Wolbachia-specific WRi_004290 probe amplified from sequence-verified plasmids and visualized at 60X. While hybridization was not detected in polytene chromosomes with WRi_004290, it is detected in the mitotic chromosomes. The fluorescence appears to be localized to the centromere of the abnormally large fourth chromosomes of D. ananassae. The larger 2^nd and 3^rd chromosomes are visualized nearby as well as a likely X chromosome. Hybridization to the fourth chromosome is consistent with the LGT being largely heterochromatic.