DNA methyltransferases, or DNMTs, are enzymes that catalyze the deposition of methyl groups onto the DNA. This epigenetic phenomenon is an evolutionary conserved process that maintains genome stability by providing an extra layer of regulation beyond the information coded within the DNA. In mammals, cytosine DNA methylation plays key roles during embryonic development. So it is no surprise that genetic loss of any of the three catalytically active enzymes Dnmt1, Dnmt3A or Dnmt3B is lethal in mice. In most mammalian species, DNMT3A and/or DNMT3B are chiefly responsible for adding novel DNA methylation throughout the genome. However, rodents use a third enzyme, DNMT3C, to specifically methylate transposon sequences. Understanding which evolutionary transitions drove the specialization of these enzymes may shed a new light onto the function of all DNMTs.
Central to the function of these DNMTs is their role in suppressing transposons, selfish genetic elements that propagate within host genomes thereby compromising reproductive fitness. These opportunist genetic elements are particularly troublesome in reproductive cells where their activity lead to the vertical transmission of novel copies potentially disrupting genome stability. Retrotransposons are reminiscent of retroviruses and some share a common ancestry. Much is known about the evolutionary arms race between viruses and immune genes in the host but does such a race exist between transposons and DNMTs?
The Malik lab (Basic Sciences Division) studies the evolutionary forces driving pathogen-host interactions. The lab uses Drosophila, yeast and mammalian models to study the mechanisms driving genetic conflict. In recent work, published in Molecular Biology and Evolution, they studied the dynamics of potential genetic conflicts between transposons and DNMTs and uncovered the evolutionary forces driving DNMT evolution in mammals. The study was led by Antoine Molaro, a postdoc in the lab in collaboration with Deborah Bourc’his, a world expert in DNMTs, at the Curie Institute, France.
Chromatin modifying enzymes, such as DNMTs, are thought to be highly conserved between species as they perform universal housekeeping functions, such as the addition of methyl groups to DNA in the case of DNMTs. Many of these enzymes have been around for over a billion years. It was surprising to see that some de novo DNMTs are subject to rapid evolution, i.e. evolutionary diversification, said Molaro. “Usually, these evolutionary signatures are seen in immunity genes that are under pressure to adapt to pathogens. Thus, one of the broad implication to the field is that evolutionary conservation is not the rule,” he explained.
The team used a phylogenomic approach to study the DNMT3C in rodents in order to understand the evolutionary forces that shape its unique function. By analyzing rodent genomes in detail, they found that the enzyme arose through a single duplication of Dnmt3B that occurred around 60Mya in the last common ancestor of muroid rodents. Their analysis revealed that the N- and C-terminal domains of DNMT3C evolved independently, with the gene sequence coding for the C-terminus being re-written by recurrent gene conversion with Dnmt3B, whereas the N-terminus has separately evolved under strong diversifying selection. But what was the driving force behind this strong positive selection?
While DNMT3C is absent in primates, the N-terminus of DNMT3A has also evolved under similar diversifying selection, the team observed. “In the two cases discussed in our manuscript (DNMT3A in primates and DNMT3C in rodents), we hypothesize that pressure to control the activity of retrotransposons and ancient viruses by DNA methylation is driving DNMTs to diversify,” Molaro said.
This drive down history lane of DNMT evolution had some broader implications. Catalytically active DNMT3s have three well-defined domains that stretch from the center to the C-terminus with the most C-terminus region encoding the catalytic domain. In contrast, the N-terminal portion of DNMTs remained largely uncharacterized. “In our study, we found that residues located within an uncharacterized domain - the N-terminal tail of the enzyme, making up to a 4th of the protein - are subject to intense natural selection. This implies that the N-term is a novel functional domain that had been completely overlooked by prior studies. This illustrate nicely how phylogenomics can teach us invaluable lessons about gene function,” said Molaro, further highlighting the significance of their work.
Other questions remain outstanding. Are DNMTs chasing or being chased by these selfish genetic elements? Molaro wonders. “We are now investigating this question in mouse models and in cell culture,” he added.
Molaro A, Malik H, Bourc’his D. 2020. Dynamic evolution of de novo DNA methyltransferases in rodent and primate genomes. Molecular Biology and Evolution. DOI:10.1093/molbev/msaa044.
This work was supported by the Damon Runyon Cancer Research Foundation, the National Institutes of Health, the Howard Hughes Medical Institute and the Laboratoire d’Excellence (LABEX).