明石 基洋

Abstract While mutations are the engine of evolution, how mutation rates themselves evolve remains poorly understood. Previous simulation and experimental studies have often treated the mutation rate as a simple, cell-wide constant, an assumption that may not fully capture the asymmetric and dynamic nature of replication fidelity in living systems. The disparity mutagenesis model explains the conservation of replicated genetic information via unequal mutation rates and theoretically avoiding extinction due to intracellular damage. This model has been extensively compared with the parity model, in which the two mutation rates are equal under a constant mutation rate. In this study, we extended our investigation and analyzed the evolution of two freely mutating mutation rates that could vary between 0 and 1, and were imported at the time of replication using a simple simulation model. Our simulation imposed no constraints on the mutation rates and highlighted the a priori functionality of the mutation rate parameter. In all trials, the difference between two mutation rates, the fidelity difference (FD), oscillated significantly above 0 during evolution as gene scores, which function as fitness score, increased. After the total gene score reached a plateau, the FD remained constant and stabilized at two unequal mutation rates. The high average mutation rate indicated the generation of the ‘genome guarantee effect’ as one of the two mutation rates always remained high and the other automatically decreased. Our results reveal a crucial new insight: if a life form replicates with a doubling of genetic information, the FD never converges to zero and automatically remains in the disparity state throughout the evolutionary process. This offers a new perspective on the long-standing issue of mutator hitchhiking in evolutionary biology.

Giant viruses are distinguished not only by their large particle size, but also by their extensive genomes, often reaching megabase levels. Many sequences within these genomes are considered to have been introduced by hosts, surrounding organisms, or other viruses. Since the natural hosts of many giant viruses remain unidentified, analyzing sequences potentially derived from other organisms may aid in clarifying their hosts. In the present study, we identified eukaryote-homologous sequences by isolating those not shared among viruses, an aspect previously overlooked. Our primary focus was on pandoravirus, which, with a genome size of ~2 Mb, is the largest among giant viruses. We obtained 375 BLAST hits with an average sequence identity of ~90%. Among the 102 detected species, those with higher hits included Mus musculus, Lampetra planeri, Melanogrammus aeglefinus, Lampetra fluviatilis, Scylla paramamosain, Cardiocondyla obscurior, Monodelphis domestica, Vespula pensylvanica, Micromonas pusilla, Physcomitrium patens, and Peromyscus californicus. Similar anal-yses of Cedratvirus and Pithovirus, which share an amphora-shaped particle structure with pandoraviruses, yielded fewer data (48 and 5 hits, respectively), with no common taxa at the order level. Thirteen BLAST hits exceeded 100 bp, including conserved non-coding elements (CNEs) in fish and other taxa, along with sequences of unknown functions. These results indicate the presence of short regions with sequence similarity in non-shared sequences, although direct host identification proved difficult.

Giant viruses, categorized under Nucleocytoviricota, are believed to exist ubiquitously in natural environments. However, comprehensive reports on isolated giant viruses remain scarce, with limited information available on unrecoverable strains, viral proliferation sites, and natural hosts. Previously, the author highlighted Pandoravirus hades, Pandoravirus persephone, and Mimivirus sp. styx, isolated from brackish water soil, as potential hotspots for giant virus multiplication. This study presents findings from nearly a year of monthly sampling within the same brackish water region after isolating the three aforementioned strains. This report details the recurrent isolation of a wide range of giant viruses. Each month, four soil samples were randomly collected from an approximately 5 × 10 m plot, comprising three soil samples and one water sample containing sediment from the riverbed. Acanthamoeba castellanii was used as a host for virus isolation. These efforts consistently yielded at least one viral species per month, culminating in a total of 55 giant virus isolates. The most frequently isolated species was Mimiviridae (24 isolates), followed by Marseilleviridae (23 isolates), Pandoravirus (6 isolates), and singular isolates of Pithovirus and Cedratvirus. Notably, viruses were not consistently isolated from any of the four samples every month, with certain sites yielding no viruses. Cluster analysis based on isolate numbers revealed that soil samples from May and water and sediment samples from January produced the highest number of viral strains. These findings underscore brackish coastal soil as a significant site for isolating numerous giant viruses, highlighting the non-uniform distribution along coastlines.