Thanks KJ. You’re exactly right. Since AEBP2-L is the broadly expressed isoform, it needs to get more attention. Chen’s lab did a great job of figuring out how its disordered N-terminal region works. What’s also cool is that it binds to PRC2 without JARID2 in many somatic cell types

16/ 📖 Again, please check out the full paper here: surl.li/cgwqcq

Huge thanks again to all collaborators and co-authors who made this project possible! 🙏

15/ This work highlights how evolution created a regulatory switch within a single protein isoform to restrain PRC2 activity in somatic tissues — a major step toward understanding PRC2 regulation during development, but also has relevance in terms of PRC2 dysregulation in cancer surl.li/jkrdbt

14/ 💡 Key insight: We found that AEBP2-L ⏸️ acts as a brake on PRC2 in somatic cells, whereas AEBP2-S ⚡ boosts PRC2, but is only present in very early development.

Since most studies focus on AEBP2-S, we feel it's time the broadly expressed AEBP2-L 👀 gets the attention it deserves!

13/ ⚙️ Mechanistically, AEBP2-L contains a mammalian-specific, disordered N-terminal region with acidic tracts that we find inhibit PRC2 binding to chromatin. Mutations that neutralize this region relieve inhibition and boost H3K27 methyltransferase activity:

12/ Interestingly, across mouse & human tissues, AEBP2-L is the predominant isoform from early embryogenesis, and broadly expressed.

Strikingly, the AEBP2-S isoform, which most labs study, is barely expressed beyond very early development!! 🤯

11/ We next wished to explore PRC2 integrity in the absence of AEBP2 isoforms with @michielvermeulen.bsky.social. We found that while loss of either isoform reduced JARID2 incorporation, JARID2 can associate with PRC2.2 with either isoform or even without both

10/ Indeed, loss of AEBP2-S KO impaired repression of 398 genes normally silenced by PRC2 during differentiation, whereas loss of AEBP2-L did not cause any defect

9/ If AEBP2-L acts like a Trithorax group protein, could AEBP2-S act like a Polycomb group protein? 🤔 We examined the roles of both isoforms during naive-to-primed pluripotency differentiation, which we previously used to study JARID2 during differentiation pubmed.ncbi.nlm.nih.gov/37030288/

8/ Notably, complete loss of AEBP2 (both isoforms) was shown by Brockdorff & Cooper (2016) to produce a Trithorax phenotype in mice, accompanied by increased PRC2 and H3K27me3 at target genes ⬆️

pubmed.ncbi.nlm.nih.gov/27317809/

7/ AEBP2-L acts like a Trithorax group protein! 😲 We found that loss of AEBP2-L, but not AEBP2-S, increases ⬆️ PRC2 binding and H3K27me3 deposition on PcG target genes, mirroring the effect of losing both isoforms

6/ Next, to avoid any potential overexpression issues, we used CRISPR 🧬✂️ to KO either the short, the long or both isoforms in mouse ESCs:

5/ We then confirmed in cells that ectopically expressed AEBP2-L does not binds PcG target genes as well as AEBP2-S. Curiously, ectopic overexpression of AEBP2-L, but not AEBP2-S, reduced overall SUZ12 (PRC2) binding ⬇️⬆️

4/ 🔍We focused on AEBP2, a PRC2 accessory subunit with two isoforms: AEBP2-S (short) and AEBP2-L (long).

Strikingly, we found AEBP2-L inhibits DNA binding by PRC2, whereas AEBP2-S promotes it:

3/ PRC2 is a chromatin repressor complex essential for development and deregulated in disease. While several accessory proteins enhance PRC2, no endogenous inhibitor in somatic cells had been identified — until now! ⚡️

2/ 🧵🌍 Firstly, a huge thank you to first authors Marlena Mucha, Zhihao Lai, Nicholas McKenzie, Francesca Matra. This was a joint effort with @davidovichlab.bsky.social with wonderful collaborations with labs at Radboud, NKI, USC & Garvan Institute 🙏🧵

2/ Firstly, this was a fantastic collaborative effort that would not have been possible without super-talented co-first author @dr-chromatin.bsky.social, and co-senior authors @conwayer1.bsky.social and @adrianbracken.bsky.social.