Recurrent Domestication by Lepidoptera of Genes from Their Parasites Mediated by Bracoviruses

Abstract

Bracoviruses are symbiotic viruses associated with tens of thousands of species of parasitic wasps that develop within the body of lepidopteran hosts and that collectively parasitize caterpillars of virtually every lepidopteran species. Viral particles are produced in the wasp ovaries and injected into host larvae with the wasp eggs. Once in the host body, the viral DNA circles enclosed in the particles integrate into lepidopteran host cell DNA. Here we show that bracovirus DNA sequences have been inserted repeatedly into lepidopteran genomes, indicating this viral DNA can also enter germline cells. The original mode of Horizontal Gene Transfer (HGT) unveiled here is based on the integrative properties of an endogenous virus that has evolved as a gene transfer agent within parasitic wasp genomes for ≈100 million years. Among the bracovirus genes thus transferred, a phylogenetic analysis indicated that those encoding C-type-lectins most likely originated from the wasp gene set, showing that a bracovirus-mediated gene flux exists between the 2 insect orders Hymenoptera and Lepidoptera. Furthermore, the acquisition of bracovirus sequences that can be expressed by Lepidoptera has resulted in the domestication of several genes that could result in adaptive advantages for the host. Indeed, functional analyses suggest that two of the acquired genes could have a protective role against a common pathogen in the field, baculovirus. From these results, we hypothesize that bracovirus-mediated HGT has played an important role in the evolutionary arms race between Lepidoptera and their pathogens.

Fig 1. Map of bracovirus sequences inserted into lepidopteran genomes.

The seven examples of high homology regions between lepidopteran sequences and bracovirus circles (CcBV) described in this paper are shown (a to g correspond to the different insertions of bracovirus sequences related to CcBV found in Lepidoptera genomes). The level of similarity is indicated by grey colour intensity. Sequences of Lepidoptera contigs flanking the homology regions correspond to lepidopteran genomic DNA identified as such by specific genes and/or repetitive sequences of lepidopteran genomes. CcBV sequences are shown as in their integrated proviral form in the wasp genome in direct orientation or reverse complement (indicated by a c after contig length). Position of primers used to extend sequences or to verify insertions in different species or to check for splicing are shown. Gene annotations (reported from Genbank) and detected transcripts (TSA) are indicated.

Fig 2. Analysis of BEN 9 encoding insertions in the Danaina subtribe.

A) Analysis of BEN9 encoding insertions in genomic DNA of individuals from different species of the Danaina subtribe by ben9 gene PCR amplification from Lepidoptera of the species Danaus chrysippus chrysippus (Oman), Danaus genutia (Thailand), Danaus plexippus (Q, caterpillar sampled in Québec, A, adults from Australia), Tirumala septentrionis septentrionis (Malaysia). C1, C2, C3: control PCR (without DNA) performed with primer pairs used respectively for D. plexippus, D. chrysippus/D. genutia and T. septentrionis PCRs B) RT-PCR analysis of Ben9 expression in D. plexippus caterpillars from Québec. Ben9 expression was detected in three individuals. No PCR amplification of Ben9 was observed on RNA samples that were not subjected to RT (No RT). C) PCR fragments obtained from D. plexippus genomic DNA and cDNA and schematic represention of Ben9 gene and D. plexippus Ben9 cDNA organization. The black bar indicates that exon 3 is not to scale. Note that in the amplified fragment corresponding to D. plexippus cDNA, the two Ben9 intron sequences have been excised as observed in Ben9 cDNA obtained from Manduca sexta parasitized by Cotesia congregata [33]. The phylogenetic tree is adapted from [34]. Dating of the common ancestor is reported from [35].

Fig 3. Measure of selection operating on Ben4 and Ben9 genes in D. plexippus and 4 related species.

a) Phylogenetic tree based on the nucleotide sequence alignment of the region shared between Ben4 and Ben9 in Danaus species samples and CcBV. The values in brackets indicate the number of lepidopteran individuals used in the analysis. b) Plot of the dN/dS value of each codon along the Ben genes based on the alignment of the butterfly sequences. The red bars represent values that are significantly under positive or negative selection (HyPhy, p-value ≤ 0.1). The asterisks identify the sites also under positive selection with the PAML approach. The yellow blocks under the dN/dS graphs represent the Ben gene structure composed of three exons. The first exon corresponds to a PHA02737 domain and the BEN domain (represented in purple) is encoded by the end of the third exon. Note that the truncated Ben4 gene conserved in D. plexippus corresponds to the third exon of CcBV Ben4 gene, which contains the BEN domain.

Fig 4. Phylogenetic tree of bracovirus-lectin like proteins from different Spodoptera species and their homologs from bracovirus, hymenopteran, lepidopteran and dipteran species.

Evolutionary distance was calculated for aligned sequences by Maximum likelihood analysis. The tree was performed using the alignment of the C-lectin domains of the proteins analysed (S3 Fig).

Fig 5. Regulatory sequence involved in bracovirus circle production retained in a bracovirus insertion in Spodoptera exigua genome.

In the insertion containing BV2-5, a sequence (BV2-5 DRJ) downstream of the gene strongly resembles C25 DRJ of CcBV (C25). A schematic representation of the C25 circle and the BV2-5 insertion in Spodoptera exigua genome (not to scale). An alignment of BV2-5 DRJ with DRJ sequences of 12 CcBV circles (including C25) is shown below. Note that the DRJ in the lepidoptera is in the same relative position as in C25 and that the similarity between C25 DRJ and the BV2-5 insertion extend beyond (residues in black) the most conserved region of the CcBV DRJs (residues in red). The presence of this DRJ sequence, which is important for bracovirus life cycle (production of DNA circles packaged in the particles), is a signature that the sequence originated from a bracovirus and shows that the direction of the transfer was from bracovirus to Lepidoptera.

Fig 6. Expression of BLL2 and BV2-5 in main larval tissues of S. exigua.

Values were normalized to the ATP synthase values and expressed relatively to the abundance in the midgut sample.

Fig 7. Cellular localization of BV2-5 and its effect on actin distribution.

Sf21 cells were infected with different recombinant viruses. The upper horizontal panel represents non-infected cells and the rest represent cells infected with AcMNPV-GFP, AcMNPV-BV2-5GFP and AcMNPV-GFP treated by latrunculin A, respectively. The fluorescence was visualized by confocal microscopy. Nuclei are visualized by DAPI and actin is visualized by phalloidin-TRITC staining.

Fig 8. Spodoptera exigua bracovirus-like genes affect baculovirus infection.

A) Effect of BV2-5 on baculovirus multiplication. One-step growth curve analysis of BV2-5 expressing virus (ph_BV2-5), virus expressing the truncated form (ph_BV2-5 t) and the control virus (ph). The results are the means ± standard deviations (error bars) for independent infection and titration experiments. BV accumulation is shown as the viral titer, calculated for each time point. Statistically different curves and P-Values (Dunnett’s test) are indicated by square brackets. B, C, D) Effect of Se-BLL2 on baculovirus infectivity. AcMNPV-GFP virions were preincubated with different concentrations of purified recombinant Se-BLL2 (50 μg/mL, 10 μg/mL, and 1μg/mL) and then used for the infection of Sf21 cells. B) Percentage of Sf21 cells infected with baculovirus (GFP positive) 36 hours after infection C) Representative images of the infected cells 36 hours after infection. D) One-step growth curve analysis of baculovirus in presence of BLL2. Statistically different curves (Dunnett’s test) are indicated by square brackets.

Fig 9. BLL expression and protection against baculovirus infection in S. exigua.

A) Changes in the expression of the BLL genes after baculovirus infection in the midguts of third-instar larvae of S. exigua (L3) B) Effect of Se-BLL2 on SeMNPV infection. The time to death was assessed by comparing the mortality curves using the Kaplan Meier method (GraphPad Prism 5). The statistical significance was determined using the log-rank analysis (Mantel-cox test), C refers to control (non-treated) larvae, Se-BLL2 refers to larvae treated with purified Se-BLL2, NPV refers to larvae treated with S. exigua baculovirus SeMNPV and NPV+Se-BLL2 refers to larvae treated simultaneously with SeMNPV and Se-BLL2 (0. 15mg/mL).

Fig 10. Production of bracovirus particles by the parasitoid wasp C. congregata and hypothesis on the process leading to transfer of bracovirus sequences to lepidopteran genomes.

The BV genome is integrated in the wasp genome (in grey). It is composed of proviral segments (in blue) used to produce dsDNA circles (blue circles) packaged in nucleocapsids (grey cylinders) that encode virulence genes introduced into the host (coloured rectangles) and of BV genes that are involved in particle production (grey rectangles). The latter originate from a nudivirus and encode structural proteins, they are expressed in wasp ovaries where production of bracovirus circles also occurs. Direct Repeat Junctions (DRJ, red triangles) are involved in site-specific recombination allowing circularisation of linear molecules from proviral segments. The circles thus produced are packaged in BV particles that also contain several integrase proteins. The particles are injected in the lepidopteran host during wasp oviposition. Once in the host BV particles infect many lepidopteran cell types but do not replicate. BV circles can integrate into lepidopteran host genomic DNA (in light blue) by a mechanism involving most likely an integrase and mediated by Host Integration Motifs (HIM) indicated by dark blue lines. When injected into a regular host (1) BV virulence gene (coloured squares) expression leads to modifications in lepidopteran host physiology, such as inhibition of wasp egg encapsulation and alteration of developmental programming allowing wasp larvae to complete their development safely in the host body. Hypothesis: when integration of viral circles occurs in the germline the integrated forms are not transmitted because the host dies. When bracoviruses are injected into a caterpillar, which is not a regular host species (2) or is a resistant host (interrupting oviposition, destroying wasp eggs, etc.) the integrated viral form in germline DNA can be transmitted vertically. As bracovirus genes are adapted for expression in lepidopteran cells they can be readily domesticated. Once integrated in lepidopteran genomes the bracovirus sequences undergo rearrangements. Ultimately, after several million years, only the domesticated genes remain from the original integrated circle. We propose that stinging of non-host species could be the main route for bracovirus sequence transfer to Lepidoptera. This is based on the fact that the genome of M. sexta which is the regular host of Cotesia congregata does not contain genes acquired from CcBV, conversely genes found in Spodoptera exigua, which is not a host of Cotesia congregata, are more closely related to CcBV. This figure is mostly based on the life cycle of CcBV associated with C. congregata parasitoid wasp of M. sexta, HIM motifs have been identified in the bracovirus of M. demolitor, the picture of S. exigua is shown as an example of C. congregata non-host species.