Oyster RNA-seq Data Support the Development of Malacoherpesviridae Genomics

Abstract

The family of double-stranded DNA (dsDNA) Malacoherpesviridae includes viruses able to infect marine mollusks and detrimental for worldwide aquaculture production. Due to fast-occurring mortality and a lack of permissive cell lines, the available data on the few known Malacoherpesviridae provide only partial support for the study of molecular virus features, life cycle, and evolutionary history. Following thorough data mining of bivalve and gastropod RNA-seq experiments, we used more than five million Malacoherpesviridae reads to improve the annotation of viral genomes and to characterize viral InDels, nucleotide stretches, and SNPs. Both genome and protein domain analyses confirmed the evolutionary diversification and gene uniqueness of known Malacoherpesviridae. However, the presence of Malacoherpesviridae-like sequences integrated within genomes of phylogenetically distant invertebrates indicates broad diffusion of these viruses and indicates the need for confirmatory investigations. The manifest co-occurrence of OsHV-1 genotype variants in single RNA-seq samples of Crassostrea gigas provide further support for the Malacoherpesviridae diversification. In addition to simple sequence motifs inter-punctuating viral ORFs, recombination-inducing sequences were found to be enriched in the OsHV-1 and AbHV1-AUS genomes. Finally, the highly correlated expression of most viral ORFs in multiple oyster samples is consistent with the burst of viral proteins during the lytic phase.

Keywords: Herpesvirales; Malacoherpesviridae; OsHV-1; RNA-seq; bivalve; viromes.

Figure 1

Number of ORFs identified in Herpesvirales genomes. The black part of each bar indicates the ORFs fraction with a predicted signal peptide. Malacoherpesviridae genomes are highlighted by the red bars.

Figure 2

Phylogenetic tree of catalytic subunit of DNA polymerase. The blue lines represent Malacoherpesviridae hits. Background colors highlight alpha-Herpesvirales in green, beta-Herpesvirales in blue, and gamma-Herpesvirales in yellow. The circular cladogram was computed with the Neighbor Joining algorithm and Jukes-Cantor distance estimation. Bootstrap values are reported for each node as percentages calculated over 1,000 performed replicates, with a significance cutoff set at 500.

Figure 3

Venn diagram of the protein domains detected in bivalve Malacoherpesviridae, gastropod Malacoherpesviridae, or other Herpesvirales. Organization of protein domains (with PFAM IDs) of bivalve-exclusive (left) and gastropod-exclusive (right) viral proteins.

Figure 4

Coverage graph of Malacoherpesviridae reads (A). Coverage graph of the whole OsHV-1 genome, as obtained by mapping 4.9 M viral reads with length and similarity mapping parameters set to 0.9. Yellow arrowheads indicate annotated ORFs along the virus genome. Maximum coverage was set to 10,000x. (B) Details of the ORF104 coverage graph, with the alignment including the 9-nt deletion causing coverage drop-off. (C) Details of ORF107 coverage graph, with two nested ORFs (arrows) that could explain the local coverage bias. One ORF encoded three transmembrane regions (highlighted in blue).

Figure 5

Radar graph depicting the viral TPM values in 14 RNA-seq samples selected as “high”. Expression values are reported with a logarithmic scale.

Figure 6

(A) SNPs occurrence in OsHV-1 ORFs. The graph depicts the ratio of nsSNP vs. total SNP for the ORFs presenting at least one variation. Globe size is proportional to the number of SNP for each ORF. (B) Common and exclusive SNPs indicated in the Venn diagram resulted from the comparison of oyster RNA-seq samples grouped by origin. (C) “Genotyping” graph based on 20 polymorphic common OsHV-1 SNPs in the eight samples richest in viral reads (indicated by IDs).