We have isolated a new nucleocytoplasmic large DNA virus (NCLDV) from hot spring water in Japan, named medusavirus. This new NCLDV is phylogenetically placed at the root of the eukaryotic clades based on the phylogenies of several key genes, including that encoding DNA polymerase, and its genome surprisingly encodes the full set of histone homologs. Furthermore, its laboratory host, Acanthamoeba castellanii , encodes many medusavirus homologs in its genome, including the major capsid protein, suggesting that the amoeba is the genuine natural host from ancient times of this newly described virus and that lateral gene transfers have repeatedly occurred between the virus and amoeba. These results suggest that medusavirus is a unique NCLDV preserving ancient footprints of evolutionary interactions with its hosts, thus providing clues to elucidate the evolution of NCLDVs, eukaryotes, and virus-host interaction. Based on the dissimilarities with other known NCLDVs, we propose that medusavirus represents a new viral family, Medusaviridae .

Members of the Marseilleviridae family are large DNA viruses with icosahedral particle structures that infect Acanthamoeba cells. The first Marseillevirus to be discovered was isolated in 2009. Since then, several other members of the Marseilleviridae family have been reported, including Lausannevirus, Senegalvirus, Cannes 8 virus, Insectomime virus, Tunisvirus, Melbournevirus, Port-Miou virus, and Brazilian Marseillevirus, which have been isolated from Europe, Africa, Australia, and South America. The morphological and genomic properties of a new Marseilleviridae family member, Tokyovirus, discovered in a water/soil sample from a Japanese river in Tokyo, were described in the present study. Tokyovirus possesses icosahedral particles of up to 200 nm in diameter, as revealed by a transmission electron microscopy (TEM) analysis, which form a giant virion factory in Acanthamoeba cells. A preliminary genome analysis predicted 487 coding sequences. A dot plot analysis and phylogenetic analysis using family B DNA polymerase, proliferating cell nuclear antigen (PCNA), and DNA-directed RNA polymerase alpha subunit genes revealed that Tokyovirus shares similarities with Marseillevirus, Melbournevirus, and Cannes 8 virus (Marseilleviridae subclade A), but not with Lausannevirus and Port-Miou virus (subclade B), Tunisvirus and Insectomime virus (subclade C), or Brazilian Marseillevirus (subclade D), suggesting that Tokyovirus has evolved separately from the previously described Marseilleviridae members.

The mechanistic evolutionary origin of the eukaryotic cell nucleus remains unknown. Among several plausible hypotheses, the most controversial is that large DNA viruses, such as poxviruses, led to the emergence of the eukaryotic cell nucleus. Several recent findings, including the discovery of a nucleus-like structure in prokaryotic viruses and prokaryotes possessing nucleus-like inner membranes, suggest genomic DNA compartmentalization not only in eukaryotes but also in prokaryotes. The sophisticated viral machinery of mimiviruses is thought to resemble the eukaryotic nucleus: DNA replicates both inside the viral factory and nucleus, which is at least partially surrounded by membranes and is devoid of ribosomes. Furthermore, several features of the recently identified Acanthamoeba castellanii medusavirus suggest that the evolutionary relationship between ancestral viral factory and eukaryotic nucleus. Notably, Ran, DNA polymerase, and histones show molecular fossils of lateral transfer of nuclear genes between the virus and host. These results suggest viral innovation in the emergence of the eukaryotic nucleus. According to these results, a new scenario explaining the origin of the eukaryotic nucleus from the perspective of viral participation is proposed. This new scenario could substantially impact the study of eukaryogenesis and stimulate further discussion about viral contributions to the evolution of the eukaryotic nucleus.