Novel high-quality amoeba genomes reveal widespread codon usage mismatch between giant viruses and their hosts

Abstract

The need for high-quality protist genomes has prevented in-depth computational and experimental studies of giant virus-host interactions. In addition, our current knowledge of host range is highly biased due to the few hosts used to isolate novel giant viruses. This study presents six high-quality amoeba genomes from known and potential giant virus hosts belonging to two distinct eukaryotic clades: Amoebozoa and Discoba. We employ their genomic data to investigate the predictability of giant virus host range. Using a combination of long and short-read sequencing, we obtained highly contiguous and complete genomes of Acanthamoeba castellanii, Acanthamoeba griffini, Acanthamoeba terricola, Naegleria clarki, Vermamoeba vermiformis, and Willaertia magna, contributing to the collection of sequences for the eukaryotic tree of life. We found that the six amoebae have distinct codon usage patterns and that, contrary to other virus groups, giant viruses often have different and even opposite codon usage with their known hosts. Conversely, giant viruses with matching codon usage are frequently not known to infect or replicate in these hosts. Interestingly, analyses of integrated viral sequences in the amoeba host genomes reveal potential novel virus-host associations. Matching of codon usage preferences is often used to predict virus-host pairs. However, with the broad-scale analyses performed in this study, we demonstrate that codon usage alone appears to be a poor predictor of giant virus host range. We discuss the potential strategies that giant viruses employ to ensure high viral fitness in non-matching hosts. Moreover, this study emphasises the need for more high-quality protist genomes. Finally, the amoeba genomes presented in this study set the stage for future experimental studies to better understand how giant viruses interact with different host species.