Fundamental Difficulties Prevent the Reconstruction of the Deep Phylogeny of Viruses

Abstract

The extension of virology beyond its traditional medical, veterinary, or agricultural applications, now called environmental virology, has shown that viruses are both the most numerous and diverse biological entities on Earth. In particular, virus isolations from unicellular eukaryotic hosts (heterotrophic and photosynthetic protozoans) revealed numerous viral types previously unexpected in terms of virion structure, gene content, or mode of replication. Complemented by large-scale metagenomic analyses, these discoveries have rekindled interest in the enigma of the origin of viruses, for which a description encompassing all their diversity remains not available. Several laboratories have repeatedly tackled the deep reconstruction of the evolutionary history of viruses, using various methods of molecular phylogeny applied to the few shared “core” genes detected in certain virus groups (e.g., the Nucleocytoviricota). Beyond the practical difficulties of establishing reliable homology relationships from extremely divergent sequences, I present here conceptual arguments highlighting several fundamental limitations plaguing the reconstruction of the deep evolutionary history of viruses, and even more the identification of their unique or multiple origin(s). These arguments also underline the risk of establishing premature high level viral taxonomic classifications. Those limitations are direct consequences of the random mechanisms governing the reductive/retrogressive evolution of all obligate intracellular parasites.

Host Range and Coding Potential of Eukaryotic Giant Viruses

“…Giant viruses with larger genome size can potentially better manipulate specific hosts in a variety of pathways and cellular processes, but there are clearly factors that limit their genome size or cause reductive genome evolution as in other parasitic entities. Random gene losses intrinsically lead to genomic reduction during viral evolution [ 139 ]. There is almost no limit to the genes that can be lost, even genes central to information processing.…”

“…There is almost no limit to the genes that can be lost, even genes central to information processing. Largely speaking, the repertoires of translation-related genes are the most variable [ 66 ], followed by transcription-related genes and then by genes for DNA replication [ 13 , 55 , 139 ]. A 16% reduction in genome size was observed in mimivirus subcultured 150 times in axenic Acanthamoeba cultures, which was accompanied by marked changes in virion morphology [ 140 ].…”

“…Answers to this question would depend on our understanding of the genomic compositions of these ancient viruses. Although it can be still disputable [ 139 ], NCLDVs are generally believed to have evolved from a common ancestor. Phylogenomic and comparative genomic analyses suggest that the NCLDV common ancestor had a small viral genome [ 13 ] rather than that of a cellular organism [ 143 ], but what this common ancestor was like and its relationship to cellular eukaryotes are much more debated.…”

“…However, no NCLDV-like infection in prokaryotic cells has been reported so far. As single-gene trees can be misleading in inferring the ancient past of eukaryotes [ 147 ], it is also disputable whether we could take the virus-cell DDRP tree at face value [ 148 ] and assume they have evolved in viral lineages without being lost, regained, or replaced, which are especially problematic when inferring deep viral phylogenies [ 139 ]. Besides, if the NCLDV families Iridoviridae , Marseilleviridae , and Pithoviridae had originated before LECA [ 88 ], it is difficult to imagine that today they have such confined host range across the eukaryotic tree of life ( Figure 1 ).…”

Phylogeny of the Varidnaviria Morphogenesis Module: Congruence and Incongruence With the Tree of Life and Viral Taxonomy

“…For some authors, the determination of viral phylogeny based on protein sequences comparison is a futile or at least risky exercise because they originated from different cell lines and that the core proteins characteristic of a modern viral lineage could be an artifact due to the random losses of proteins initially present their cellular ancestors ( Claverie, 2020 ; Nasir et al, 2020 ). This is probably not the case for the core proteins of Nucleocytoviricota since we have shown that the concatenation and single phylogenies of their eight-core genes were mostly congruent ( Guglielmini et al, 2019 ).…”