Analysis of Spounaviruses as a Case Study for the Overdue Reclassification of Tailed Phages

Abstract

Tailed bacteriophages are the most abundant and diverse viruses in the world, with genome sizes ranging from 10 kbp to over 500 kbp. Yet, due to historical reasons, all this diversity is confined to a single virus order-Caudovirales, composed of just four families: Myoviridae, Siphoviridae, Podoviridae, and the newly created Ackermannviridae family. In recent years, this morphology-based classification scheme has started to crumble under the constant flood of phage sequences, revealing that tailed phages are even more genetically diverse than once thought. This prompted us, the Bacterial and Archaeal Viruses Subcommittee of the International Committee on Taxonomy of Viruses (ICTV), to consider overall reorganization of phage taxonomy. In this study, we used a wide range of complementary methods-including comparative genomics, core genome analysis, and marker gene phylogenetics-to show that the group of Bacillus phage SPO1-related viruses previously classified into the Spounavirinae subfamily, is clearly distinct from other members of the family Myoviridae and its diversity deserves the rank of an autonomous family. Thus, we removed this group from the Myoviridae family and created the family Herelleviridae-a new taxon of the same rank. In the process of the taxon evaluation, we explored the feasibility of different demarcation criteria and critically evaluated the usefulness of our methods for phage classification. The convergence of results, drawing a consistent and comprehensive picture of a new family with associated subfamilies, regardless of method, demonstrates that the tools applied here are particularly useful in phage taxonomy. We are convinced that creation of this novel family is a crucial milestone toward much-needed reclassification in the Caudovirales order.

Keywords: Caudovirales; Herelleviridae; phylogenetics; phylogenomics; spounavirus; virus classification; virus taxonomy.

Figure 1.

a) Network representation of predicted protein content similarity of dsDNA viruses generated with vConTACT v2.0. Viruses are represented as circles (nodes) connected with each other (edges) based on a significant number of shared protein clusters, with more similar genomes displayed closer together on the network. The genomes belonging to the new family Herelleviridae are indicated with a circle. Genomes previously assigned to the subfamily Spounavirinae are indicated in pink. b) Clustering of dsDNA bacteriophages that possess subfamily assignments in the order Caudovirales generated with GRAViTy, darker colors in the heatmap represent higher degrees of similarity between genomes. The phages are clustered using UPGMA into a dendrogram, showing bootstrap values (100 pseudoreplicates) on each branch.

Figure 2.

a) VICTOR and b) DICE score trees. The trees were rooted at Brochothrix phage A9. The scale bars represent the calculated distance metric, branch support values at the VICTOR trees were calculated from 100 pseudobootstrap replicates. Genera and subfamilies are delineated with colored squares and colored circles, respectively.

Figure 3.

a) Virus Proteomic Tree (VIPTree) and b) GOAT tree. The trees were rooted at Brochothrix phage A9. The scale bar represents the distance metric. Genera and subfamilies are delineated with colored squares and colored circles, respectively.

Figure 4.

Maximum-likelihood tree based on concatenated alignment of 10 marker proteins generated using IQ-tree. The scale bar represents the number of substitutions per site, branch support values were calculated from 1000 ultrafast bootstrap (UFBOOT) replicates. The trees were rooted at Brochothrix phage A9 to facilitate comparison. Branches corresponding to genera and subfamilies are delineated with colored squares and circles, respectively.