The main character of many movies has a memorable nemesis who can’t help but try to bring them down and get in their way. Now, it turns out that the nemesis of the inactive Dntm3b isoform may just be development’s best friend.
Before differentiation, Dnmt3b transcript levels are kept low, but get a boost at differentiation. Suddenly, the transcripts increase, and there’s an elevation of the active Dnmt3b1 form made compared to the Dnmt3b6 form. After this process is finished, the hubbub dies down and Dnmt3b gene expression returns to a low basal level. Humaira Gowher’s lab (Purdue University) wondered how these changes happened, so they analyzed mouse embryonic stem cells. In these cells, inactive Dnmt3b6 is the major isoform that’s expressed at a basal level pre-differentiation.
Dnmt3bas is located 1 kb down from the Dnmt3b promoter. Curiously, the group noticed that Dnmt3bas levels decrease 3 days after differentiation, whereas Dnmt3b peaks at that time. This reverses at day 6. This pattern was their first clue that the lncRNA could have a dastardly plan.
With knock-down (KD) and overexpression (OE) experiments of Dnmt3bas, the team made these observations around day 3:
- KD of Dnmt3bas doubled Dnmt3b transcript levels versus control, whereas OX of Dnmt3bas meant less Dnmt3b
- Alternative splicing produced more active Dnmt3b1 than inactive Dnmt3b6 when the lncRNA was knocked down—there was still a higher Dnmt3b1:Dnmt3b6, but it was less pronounced and there was still quite a bit of Dnmt3b6 around
- OE of Dnmt3bas really boosted the Dnmt3b1:Dnmt3b6 ratio so that a lot of protein was made and hardly any of it was in the inactive form
- At the proximal enhancer and the CpG island (CGI) region near the promoter, there was less H3K27me3 in KD cells, and more of it in OE cells; there was an opposite trend for H3K27Ac at CGI
So, the team suspected that at differentiation, Dnmt3bas recruits its partner-in-crime PRC2, which adds H3K27me3, as some lncRNAs are known to do. ChIP and RNA pull-downs showed this was indeed the case, and this changed as expected when Dnmt3bas was knocked down or overexpressed. They also thought it might bind to its henchman hnRNPL to regulate splicing, and they were right again. They surmised that hnRNPL then brings over RNA Pol II to splice sites.
This all led the group to come up with a full story in which Dnmt3bas doesn’t seem to like inactive Dnmt3b much, keeping it at very low basal levels before differentiation via PRC2 repression. Then, as Dnmt3b transcription picked up by day 3 after the differentiation signal, Dnmt3bas kept it mostly in the active form via alternative splicing. Finally, the lncRNA helped revert to low basal Dnmt3b levels again by day 6 by bringing PRC2 back to put H3K27me3 at the promoter CGI.
Read more about the backstory of this caper at Cell Reports, June 2023.