Comparative genomics of mutualistic viruses of Glyptapanteles parasitic wasps

Abstract

Background: Polydnaviruses, double-stranded DNA viruses with segmented genomes, have evolved as obligate endosymbionts of parasitoid wasps. Virus particles are replication deficient and produced by female wasps from proviral sequences integrated into the wasp genome. These particles are co-injected with eggs into caterpillar hosts, where viral gene expression facilitates parasitoid survival and, thereby, survival of proviral DNA. Here we characterize and compare the encapsidated viral genome sequences of bracoviruses in the family Polydnaviridae associated with Glyptapanteles gypsy moth parasitoids, along with near complete proviral sequences from which both viral genomes are derived.

Results: The encapsidated Glyptapanteles indiensis and Glyptapanteles flavicoxis bracoviral genomes, each composed of 29 different size segments, total approximately 517 and 594 kbp, respectively. They are generated from a minimum of seven distinct loci in the wasp genome. Annotation of these sequences revealed numerous novel features for polydnaviruses, including insect-like sugar transporter genes and transposable elements. Evolutionary analyses suggest that positive selection is widespread among bracoviral genes.

Conclusions: The structure and organization of G. indiensis and G. flavicoxis bracovirus proviral segments as multiple loci containing one to many viral segments, flanked and separated by wasp gene-encoding DNA, is confirmed. Rapid evolution of bracovirus genes supports the hypothesis of bracovirus genes in an 'arms race' between bracovirus and caterpillar. Phylogenetic analyses of the bracoviral genes encoding sugar transporters provides the first robust evidence of a wasp origin for some polydnavirus genes. We hypothesize transposable elements, such as those described here, could facilitate transfer of genes between proviral segments and host DNA.

Figure 1

Summary of GiBV and GfBV sequence data. Homologous viral segments between GiBV and GfBV are depicted in the same row and have been assigned the same number; blank spaces represent the absence of a homologous segment in one of the genomes. To the right, columns indicate the locus containing the segment and whether the segment is complete (C), partial (P), or absent (-) in the viral and proviral sequences from each genome. Incomplete sequence indicators (<>) are shown only for segments for which both viral and proviral sequence are absent or incomplete.

Figure 2

Nucleotide conservation extends around the proviral segment excision site. Viral, 5' proviral, and 3' proviral motifs are shown for the GiBV and GfBV genomes, with the excision site highlighted. For each proviral motif, underlined sequence represents non-encapsidated sequence while unmarked sequence represents proviral segment sequence. Following excision, DNA is circularized at the excision site forming the viral motif.

Figure 3

Structural organization and synteny of proviral segments of GfBV and GiBV. Detailed diagrams of proviral loci for GfBV and GiBV are depicted. For each, the corresponding segment number of the encapsidated viral segment is given for each segment, with the > or < symbol depicting the directionality of segment excision. Black boxes represent long tandem repeats, and three repeat classes are listed as A, B, and C. Regions of synteny between proviral segment and flanking DNA are shaded in gray. Strand-specific protein coding genes within proviral segments are depicted by green boxes, while genes encoded in flanking DNA at each locus are depicted by purple boxes. The length for each locus is shown in kbp.

Figure 4

Clustering diagram of proviral segments, flanking sequences, and inter-segmental sequences by nucleotide composition. The clustering was generated using the neighbor-joining algorithm on relative trinucleotide frequences for each sequence region. Inter-segmental sequences are shown in black (unlabeled), proviral segment sequences in red, flanking sequences in green, and the GiBV TE in blue.

Figure 5

Histogram of dN/dS values for viral and flanking genes in GiBV and GfBV proviral sequences. dN/dS values were calculated for 72 and 41 homologous pairs of viral genes in flanking DNA, respectively. Genes in flanking DNA have dN/dS values centered near 0 while viral genes have dN/dS values centered near 1.

Figure 6

Bayesian phylogram of sugar transporter gene sequences from GiBV, GfBV and various insects. Mammal sugar transporters were used as an outgroup. Posterior probabilities >80% are listed above the corresponding branches, and higher taxonomic classification is shown to the right of the tree.

Figure 7

Detailed view of GfBV proviral locus 7. Genes are labeled by strand, and genes identical in nucleotide sequences are indicated by asterisks. TE classification and CcBV homologs are shown above the genes. The directionality of segment excision is indicated by an arrowhead next to the segment number.