Giant viruses infect protists. To be able to calculate codon usage preferences, high quality genomes are essential. The availability of these in the public databases is limited to very few protist species. We therefore sequenced six selected protists to subsequently generate high quality genome assemblies, that were contained in a small number of fragments with high completeness scores. The codon usage patterns of these protists show to be highly distinct at the genus level. When comparing giant virus genomes, we found that they also have markedly different codon usage patterns at both the family and genus levels. We then performed direct comparisons of the codon usage preferences of giant viruses and hosts. This analysis gave surprising results as not all giant viruses have close codon usage preferences, and some even opposite, to their best-known hosts. Interestingly, this analysis also revealed a potential novel host in which certain giant viruses might perform better. Our experimental results suggest that the cultivation conditions for obtaining high viral fitness in alternative hosts are closer to those of a natural environment. Nonetheless, the results of this study also suggest that tRNA content and codon usage preferences alone are not an accurate predictor of giant virus host range. The experimental data that we already have and that we are currently generating will aid us in defining the additional factors to be considered for making more accurate predictions at the genomic level.
To better understand the factors that drive the evolution of large viral genomes with a wide range of codon usage preferences, we used experimental evolution. For this we used three different giant viruses and three different hosts. We performed weekly passages of these viruses in their hosts for a total duration of half a year. For Tupanvirus, that has a wide host range, we included additional setups with (i) one host-switch at half of the experiment, (ii) a monthly host switch, (iii) and a weekly host switch. For the three viruses, we observed three different outcomes: (i) a decrease in cytopathic effects (CPE) and viral copy number, (ii) no change in CPE and viral copy number, and (iii) an increase in CPE and viral copy number. Although we have not completely finalized analyzing the genome sequences of the ancestral and evolved viruses, the timing of the increase in the number of mutations correlates well with the observed increase in CPE as well as viral copy number. In line with our initial hypothesis, frequent host switches generate viral variants that are able to perform well in different host species. However, lower viral replication success appears to be an important trade-off for maintaining a mixed viral population that works well in multiple hosts.
To be able to answer our original question on whether codon usage preferences define the rate of genome adaptation of giant viruses, we need to finalize the analysis of the viral genome sequencing along the evolution experiment, and potentially perform additional experiments. Like many other projects, this project was affected by the COVID-19 pandemic. Therefore, we were unfortunately not able disseminate the results at international conferences or at events targeting the general public within the project period. We recently presented our results for the first time at two international conferences: Viruses of Microbes and the IX European Congress of Protistology & Annual Congress of the International Society of Protistologists joint meeting.
