Cressdnaviricota: a Virus Phylum Unifying Seven Families of Rep-Encoding Viruses with Single-Stranded, Circular DNA Genomes

LETTER

Eukaryotic single-stranded DNA (ssDNA) viruses are classified into 10 families (Table 1), but many remain unclassified (1, 2). Viruses from seven families (Bacilladnaviridae, Circoviridae, Geminiviridae, Genomoviridae, Nanoviridae, Redondoviridae, and Smacoviridae) have small circular genomes, with some encompassing only two genes, encoding the replication initiation protein (Rep) and the capsid protein (3–5). The genomes are replicated by the rolling-circle mechanism initiated by the virus-encoded Rep proteins of the HUH endonuclease superfamily (6–10). In eukaryotic ssDNA viruses, the HUH endonuclease domain is followed by a superfamily 3 helicase (S3H) domain, which is not found in prokaryotic ssDNA virus Reps, although homologous two-domain Reps are encoded by several groups of bacterial plasmids (5). Informally, this assemblage of ssDNA viruses has been referred to as “circular, Rep-encoding single-stranded (CRESS) DNA viruses” (4, 8). Viruses from the seven families infect hosts across the eukaryotic domain, including animals (Circoviridae), diatoms (Bacilladnaviridae), fungi (Genomoviridae), and plants (Nanoviridae and Geminiviridae), whereas members of the Smacoviridae and the newly created Redondoviridae have been discovered by metagenomics and are suspected to infect animals (3, 11), although an association of smacovirids with methanogenic archaea has also been suggested (12).

Phylogenetic analysis of classified and unclassified CRESS DNA viruses revealed 13 well-supported groups (Fig. 1). Six groups correspond to unclassified viruses, denoted CRESSV1 through CRESSV6 (13). The phylogeny splits into two large clades: (i) geminivirids, genomovirids, and CRESSV6 viruses and (ii) all other CRESS DNA viruses (Fig. 1). These clades are supported by detailed analyses of the conserved catalytic motifs of Rep proteins (5, 14), with Reps of viruses in clade 1 sharing the GRS motif (15, 16), which is not found in Reps of viruses from clade 2. In contrast, Reps of all members of clade 2 contain a conserved arginine finger in the helicase domain, absent in geminivirids and genomovirids, supporting the monophyly of the Reps in clade 2 (5).

Unlike the Reps, the capsid proteins of CRESS DNA viruses are not orthologous and apparently were acquired from RNA viruses on multiple, independent occasions (5, 14, 17–19). Thus, the gene encoding the characteristic two-domain Rep is the only universally conserved gene in this virus supergroup and thus can serve as a scaffold for mapping evolutionary events. Accordingly, it has been proposed that Reps be used as a basis for megataxonomic classification of CRESS DNA viruses (20).

Recently, the International Committee on Taxonomy of Viruses (ICTV) extended its taxonomic rank structure to facilitate the classification of the entire spectrum of virus diversity (21, 22). This new taxonomic framework was adopted to formalize the evolutionary relationships among CRESS DNA viruses. Below, we outline these taxonomic changes, which were ratified by the ICTV in March 2020 and are now part of the official virus taxonomy:

• •

  Clade 1. The families Geminiviridae and Genomoviridae were included in the new order Geplafuvirales (gemini/genomo, plants, fungi, and the suffix -virales for order taxa). In anticipation of the future classification of the CRESSV6 virus group, the order Geplafuvirales has been assigned into a new class, Repensiviricetes (a portmanteau of Rep-encoding single strand and the suffix -viricetes for class taxa).

• •
  Clade 2. Viruses of the families Bacilladnaviridae, Circoviridae, Nanoviridae, Redondoviridae, and Smacoviridae and unclassified groups CRESSV1 to -5 are sufficiently distinct from each other, precluding their unification into a single order. The differences extend beyond the separation of the Reps in the phylogenetic tree and include nonhomologous capsid proteins and different genome lengths (Table 1) and genome architectures (e.g., nanovirids have multipartite genomes, whereas other viruses have monopartite genomes). Notably, circovirids cluster with CRESSV1 to -3 viruses, whereas nanovirids cluster with CRESSV4 and -5 viruses, justifying order-rank taxa for these assemblages. Thus, five orders have been created for the classification of clade 2 viruses:

  • - Baphyvirales (to include Bacilladnaviridae): after bacillariophytes (diatoms, hosts of viruses in this taxon) and the suffix -virales
  • - Cirlivirales (to include Circoviridae and, in due course, CRESSV1 to -3 viruses): circo-like and the suffix -virales
  • - Cremevirales (to include Smacoviridae): CRESS, metagenomics, and the suffix -virales
  • - Mulpavirales (to include Nanoviridae and, in due course, CRESSV4 and CRESSV5 viruses): after multipartite genomes of viruses in this taxon and the suffix -virales
  • - Recrevirales (to include the newly approved Redondoviridae): redondoviruses, CRESS, and the suffix -virales

Viruses from the established families within the five orders encode order-specific capsid proteins, which are not recognizably similar at the sequence level across the orders. This feature and the Rep phylogeny should help in assigning newly identified viruses to a particular taxon. Notably, Reps of satellite nucleic acids from the family Alphasatellitidae (23) and nanovirids form a clade (Fig. 1); thus, a taxon unifying Alphasatellitidae and Nanoviridae might be considered in the future. The five orders were further assigned to the class Arfiviricetes (arginine finger and the suffix -viricetes for class taxa). Finally, the two classes, Repensiviricetes and Arfiviricetes, were grouped into the phylum Cressdnaviricota (a contraction of CRESS DNA and the suffix -viricota for phylum taxa), thereby unifying all officially recognized families of CRESS DNA viruses.