Therefore, the Last Eukaryotic Common Ancestor had a dual methylation system: 6mA for transcriptionally permissive chromatin and 5mC for silencing. This dual pattern is maintained in some extant eukaryotes but has diverged in many others, including ALL multicellular groups that lost AMT1. 8/9

So why this wavy pattern? In ciliates 6mA is associated with inter-nucleosome linker regions, and that's the case for many eukaryotes as well. We further checked if there's any correlation with H3K4me3, and they are tightly linked across many eukaryotes, implying some level of co-deposition. 7/9

How does 6mA relate to 5mC evolution? They are totally independent, one can be lost without the other being affected and vice versa. But when they are together, there's clear compartmentalisation, with 6mA staying post TSS and 5mC usually restricted to silencing or downstream of the gene body. 6/9

But is 6mA directing transcription? We checked in 5 different species across stages and conditions, and we could detect plenty of transcriptional changes not followed by 6mA reprogramming. So correlation does not mean causation (as most times with chromatin features). 5/9

6mA is found consistently after the TSS, biased towards the 5' end of gene bodies. This is a remarkably consistent pattern across species that have diverged for 1 billion years, and in stark contrast to 5mC evolutionary variation. Also, 6mA consistently correlates with transcription. 4/9

However there's been rampant simplifications across the tree of life. We then took advantage of 6mA basecalling with @nanoporetech.com to profile several eukaryotes branching in very distant groups, some encoding AMT1 or not. AMT1 encoding eukaryotes presented symmetric ApT methylation. 3/9

First we traced the evolutionary history of AMT1 (aka MTA1), the main 6mA enzyme in eukaryotes. We find that a complex repertoire including AMT1 and its heterodimeric partner AMT6/7 can be traced back to the Last Eukaryotic Common Ancestor. Same goes for the RNA methyltransferases METTL3/14. 2/9

Then we surveyed species that kept 5mC for giant virus endogenization events, following on our previous work. We find that this is quite common across very distantly related species, rendering support that 5mC can help the host tolerate large influx of potentially deadly DNA. So no more n = 1! 5/7

Then we focus on 3 species that kept DNMTs from previously unstudied eukaryotic supergroups : Acanthamoeba, Naegleria & Cyanophora. All show 5mC as a repressive mark, marking transposons and silenced genes together with heterochromatin (H3K9me3 & K27me3), with no hint of gene body methylation. 3/7

We first characterise the evolution of DNMTs in eukaryotes. Unlike other gene families that expanded from the eukaryotic ancestor, DNMTs come from various prokaryotic sources, integrating throughout eukaryotic history. However, DNMT presence is very patchy, implying pressure for simplification. 2/7

So happy to be in 🇭🇰 "de-virtualising" collaborators, talking about TFs and epigenetics, such a welcoming scientific community at HKU. I came for the science, but there's always time for dim sum, dai pai dongs, cha chaan teng and the charm of the old Hong Kong...

Not only the 5 nucleobases... but the meteorite contains 5mC. So epigenetics existed even before genetics. Something for you @maxvcg.bsky.social

How can these Sox make iPSC? Through partnering with Oct4, and probably also other POU factors. Intriguingly, the same choanoflagellates with SOX genes also posses POU orthologues! However, these are divergent and can't bind DNA as animal POUs do (and can't make iPSC either).5/6

Unicellular Sox genes are sister group to the rest of animal Soxes. Yet they already bind the same DNA motifs. So we tested their capacity to substitute an animal Sox, in particular, Sox2 in the classic Yamanaka cocktail. We could generate iPSC, and these could make chimeric "choano"-mice! 3/6

This has been a fully collaborative project with Ralf Jauch group in Hong Kong University, none of this would be possible without their enthusiasm, expertise and effort. It all started with an email about the origins of these genes and a BLAST search. The distribution of hits, was intriguing... 2/6