DNA methylation enables recurrent endogenization of giant viruses in an animal relative

Abstract

5-methylcytosine (5mC) is a widespread silencing mechanism that controls genomic parasites. However, in many eukaryotes 5mC has gained complex roles in gene regulation beyond parasite control. Animals are a quintessential case for 5mC evolution, as they show widespread variability across lineages, ranging from gene regulation and transposable element control to loss of this base modification. Here we show that the protist closely related to animals Amoebidium appalachense features both transposon and gene body methylation, a pattern reminiscent of invertebrates and plants. Unexpectedly, large hypermethylated regions of the Amoebidium genome derive from viral insertions, including hundreds of endogenized giant viruses contributing 14% of the encoded genes, to an extent never reported before in any eukaryotic genome. Using a combination of inhibitors and functional genomic assays, we demonstrate that 5mC silences these giant virus insertions. Moreover, alternative Amoebidium isolates show polymorphic giant virus insertions, highlighting a dynamic process of infection, endogenization and purging. Our results indicate that 5mC is critical for the controlled co-existence of newly acquired viral DNA into eukaryotic genomes, making Amoebidium a unique model to understand the hybrid origins of eukaryotic genomes.