Updated Virophage Taxonomy and Distinction from Polinton-like Viruses

Abstract

Virophages are small dsDNA viruses that hijack the machinery of giant viruses during the co-infection of a protist (i.e., microeukaryotic) host and represent an exceptional case of "hyperparasitism" in the viral world. While only a handful of virophages have been isolated, a vast diversity of virophage-like sequences have been uncovered from diverse metagenomes. Their wide ecological distribution, idiosyncratic infection and replication strategy, ability to integrate into protist and giant virus genomes and potential role in antiviral defense have made virophages a topic of broad interest. However, one limitation for further studies is the lack of clarity regarding the nomenclature and taxonomy of this group of viruses. Specifically, virophages have been linked in the literature to other "virophage-like" mobile genetic elements and viruses, including polinton-like viruses (PLVs), but there are no formal demarcation criteria and proper nomenclature for either group, i.e., virophage or PLVs. Here, as part of the ICTV Virophage Study Group, we leverage a large set of genomes gathered from published datasets as well as newly generated protist genomes to propose delineation criteria and classification methods at multiple taxonomic ranks for virophages 'sensu stricto', i.e., genomes related to the prototype isolates Sputnik and mavirus. Based on a combination of comparative genomics and phylogenetic analyses, we show that this group of virophages forms a cohesive taxon that we propose to establish at the class level and suggest a subdivision into four orders and seven families with distinctive ecogenomic features. Finally, to illustrate how the proposed delineation criteria and classification method would be used, we apply these to two recently published datasets, which we show include both virophages and other virophage-related elements. Overall, we see this proposed classification as a necessary first step to provide a robust taxonomic framework in this area of the virosphere, which will need to be expanded in the future to cover other virophage-related viruses such as PLVs.

Keywords: giant virus; polinton; polintovirus; virophage; virus taxonomy.

Figure 1

Detection of virophage marker genes across sequences currently classified in the Lavidaviridae family on GenBank. Marker genes were detected based on an hmmsearch performed between predicted proteins from the Lavidaviridae genomes and HMM profiles previously built for the four conserved genes in the virophage morphogenesis gene module [12]. Panel (A) displays the pattern of conserved gene detection for virophage isolates, virophage genomes assembled from metagenomes, and virophage-like elements currently classified as Lavidaviridae on GenBank but not encoding any of the virophage conserved genes. Panel (B) displays the distribution of the hmmsearch score for each marker for both the virophages (isolates or assembled from metagenomes) and the virophage-like elements. Accession numbers for the different genomes used here are indicated in Table S1.

Figure 2

Genetic diversity in virophage genomes. (A) Size distribution and functional annotation of the 30 largest protein clusters obtained from the virophage sequences. The four conserved genes are colored in orange, while other protein clusters functionally annotated are colored in blue. (B) Distribution of pairwise amino-acid identity (AAI) across all virophage sequences for the four conserved morphogenesis genes. Two standard AAI cutoffs (30% and 40% AAI) are highlighted in red and blue, respectively. MCP: Major Capsid Protein (hexon). Penton: minor capsid protein.

Figure 3

Definition of new virophage clades based on MCP phylogeny and genome features. For both panels, an MCP phylogeny is presented based on complete and near-complete virophage genomes. For panel (A), genomes features are displayed including from the inner to the outer ring: (1) groups based on genome-wide amino-acid identity (gwAAI) considering the seven largest groups, (2) genome length, (3) Major Capsid Protein (MCP) length, (4) average GC content, (5) detection of the typical virophage penton protein based on HMM profiles, and (6) ecosystem from which the genome was obtained. In panel (B), the same tree is decorated with the classification of virophage genomes in curated groups. The “core” sequences (solid colors, inner ring) of each group represent the members of the monophyletic clade assigned to this group based on the gwAAI clustering (see panel (A)). The “affiliated” ring displays the results of a “best BLAST hit” affiliation of all sequences in the tree (ignoring self-hits) in light colors (outer ring).

Figure 4

Detection and classification of virophages from published datasets. (A) Schematic overview of the hmmsearch- and BLAST-based classification approach used to identify and assign virophages to the proposed families. For panels (B–D), three datasets were processed with the approach described in panel (A). The first includes non-redundant previously published virophage sequences gathered as part of this study and not included in the complete and near-complete genome set (n = 591; “Incomplete virophages”). The second includes contigs identified as virophages and Polinton-like viruses assembled from Gossenköllesee metagenomes and published in [19] (n = 114; “Gossenköllesee”). Finally, the third dataset included adintoviruses published in several previous studies [23,32,33] (n = 64; “Adintovirus”). In panel (B), the percentage of sequences assigned as Maveriviricetes (based on MCP HMM) or with a hit to the PLV HMM profile is indicated. The score distribution of score for the hits to the Maveriviricetes HMM or PLV HMM is displayed in panels (C,D), respectively, for each dataset and each group of sequence.