Genome Sequencing of the Behavior Manipulating Virus LbFV Reveals a Possible New Virus Family

Abstract

Parasites are sometimes able to manipulate the behavior of their hosts. However, the molecular cues underlying this phenomenon are poorly documented. We previously reported that the parasitoid wasp Leptopilina boulardi which develops from Drosophila larvae is often infected by an inherited DNA virus. In addition to being maternally transmitted, the virus benefits from horizontal transmission in superparasitized larvae (Drosophila that have been parasitized several times). Interestingly, the virus forces infected females to lay eggs in already parasitized larvae, thus increasing the chance of being horizontally transmitted. In a first step towards the identification of virus genes responsible for the behavioral manipulation, we present here the genome sequence of the virus, called LbFV. The sequencing revealed that its genome contains an homologous repeat sequence (hrs) found in eight regions in the genome. The presence of this hrs may explain the genomic plasticity that we observed for this genome. The genome of LbFV encodes 108 ORFs, most of them having no homologs in public databases. The virus is however related to Hytrosaviridae, although distantly. LbFV may thus represent a member of a new virus family. Several genes of LbFV were captured from eukaryotes, including two anti-apoptotic genes. More surprisingly, we found that LbFV captured from an ancestral wasp a protein with a Jumonji domain. This gene was afterwards duplicated in the virus genome. We hypothesized that this gene may be involved in manipulating the expression of wasp genes, and possibly in manipulating its behavior.

Fig. 1.—

Superparasitism phenotype associated with LbFV infection: distribution of wasp eggs deposited by (A) nonsuperparasitizing uninfected females and by (B) LbFV-infected superparasitizing females within Drosophila larvae. Wasp were tested under controlled laboratory conditions (see “Methods” section for details). The number of female tested was 6 and 20, respectively, for NS and S lines.

Fig. 2.—

Characterization of the viral nucleic acids. (A) Electrophoresis of viral nucleic acids extracted from Leptopilina boulardi wasps and (B) viral nucleic acids were treated with DNAse I or RNAse A. ND, not digested. The ladders are expressed in kilobases.

Fig. 3.—

Simplified De Bruijn graph showing the connexions among viral contigs after assembly of Illumina reads. Each circle, so called node in the Velvet terminology, represents a contig. Its diameter is proportional to its length. Each edge connection between two contigs indicates contiguity among contigs. Edge width is proportional to the number of connexions found among nodes. The dark grey circles represent the eight large contigs identified by Velvet assembler (13, 101, 65, 42, 99, 70, 67, 12).

Fig. 4.—

LbFV contigs identified after assembling Illumina and 454 reads are flanked by homologous regions (hrs). Alignment of homologous regions found in the LbFV genome. The contig name, start and end position of each sequence is indicated on the left. The last line of the alignment corresponds to the final consensus sequence of all contigs.

Fig. 5.—

Connexion among contigs identified by (A) mapping paired-end reads on the LbFV contigs masked for repeated sequence. The eight LbFV contigs are arbitrarily ranked according to their length and represented within a circular chromosome. Grey areas indicate genomic regions homologous to the hrs motif. The blue lines represent the mapping of paired-end reads on different contigs extremities. In total 96 blue lines are represented corresponding to 96 paired Illumina reads. (B) PCR amplification. The 16 PCR primers used in this experiment are symbolically represented by the arrows (not scale). 51 out of 120 primer combinations gave a positive PCR and are represented by a blue line connecting contig extremities. The red lines indicate two cases where the PCR was positive with a single primer. (C) The PCR amplifications often gave several bands. As an example, the PCR product obtained with the 7f and 2r primers (identified in fig. 5B by stars) is shown. The full set of gel images is provided in supplementary figure S2, Supplementary Material online. λ, molecular weight marker; 7f2r, PCR product.

Fig. 6.—

Representation of the gene and repeat content of the circular dsDNA genome of LbFV. The black line represents the coverage for Illumina reads and the grey one represents the coverage for the MinION 2D reads (left y axis). The green line indicates the GC in a sliding window of 2000 bp (right y axis). Unresolved sequences correspond to genomic regions where only MinION reads are available and where polymorphism is very likely.

Fig. 7.—

The phylogeny of two putative anti-apoptotic genes found in LbFV genome suggest that they have been acquired from eukaryotes. (A) ORF27 and (B) ORF66. The branch supports were estimated by approximate Likelihood ratio tests (aLRT). Only aLRT > 0.70 are represented. Eukaryotic lineages are represented in blue and virus lineages are in red.

Fig. 8.—

The phylogeny of LbFV ORFs 11 and 13 suggests that they have been horizontally acquired from eukaryotes and evolved subsequently through a duplication in the genome of LbFV. The branch supports were estimated by approximate Likelihood ratio tests (aLRT). Only aLRT > 0.70 are represented. Eukaryotic lineages are represented in blue and virus lineages are in red.

Fig. 9.—

Phylogenetic reconstruction based on the concatenation of 6 genes (ORF37, ORF52, ORF58, ORF85 and ORF106). The conserved blocks were concatenated and the phylogeny was built using PhyML. The phylogenetic signal was congruent among the six genes. The branch supports were estimated by approximate Likelihood ratio tests (aLRT). Only aLRT > 0.70 are represented.