Thank you for your lovely preview of our work! 😄

Congratulations to everyone involved in this collaborative effort.
Read our web story here: edin.ac/4jeAkLw
@cmvm-edinburghuni.bsky.social

Thank you Elena! Congrats on your recent Mol Cell paper too 🥳 Hope you’re well!

In summary: H3K27M rewires cPRC1 in DMG, creating a unique dependency on a minority CBX4-PCGF4-cPRC1 complex. CBX4's essential disordered region specifies formation of cPRC1 with PCGF4/BMI1, which drives oncogenic gene repression. (11/11)
Read more: www.cell.com/molecular-ce...

IP-mass spec revealed CBX4 preferentially forms a distinct complex with essential PCGF4. CBX2 lacks this specificity. Deleting the key disordered region shifts complex composition – the PCGF4 preference is lost, mimicking CBX2. This region defines CBX4’s unique function. (10/11)

But what makes CBX4 so unique? We used a CRISPR tiling screen to pinpoint key domains. Result: the chromodomain (H3K27me3 binding) and a section of the central disordered region are essential for function. The chromodomain makes sense - what about that disordered part? (9/11)

CUT&RUN-Rx says no - all forms of cPRC1 are displaced equally by PRC2i. Still, part of our hypothesis held: CBX4-cPRC1 drives repression of key genes, like p16-INK4A. KO of CBX4 derepressed genes, while KO of abundant CBX2/CBX8 had no significant impact on gene expression! (8/11)

H3K27me3 retention and CBX4-cPRC1 accumulation at Polycomb target genes correlates with stronger gene repression in DMG. PRC2 inhibition derepresses target genes and DMG is sensitive to this. We thought: could PRC2 inhibition work by selectively displacing CBX4-cPRC1? (7/11)

To dig deeper, we used CUT&RUN-Rx to map cPRC1 complexes in isogenic DMG cells with or without H3K27M. Strikingly, H3K27M rewires cPRC1 binding, driving CBX2/CBX4 accumulation and displacing CBX8 from CGIs at Polycomb targets genes. (6/11)

This was surprising because paralogs like PCGF2, CBX2 and CBX8 are also expressed!
Mass spec revealed CBX4 is <5% of the CBX pool in DMG cells, while non-essential CBX2 and CBX8 are far more abundant. Yet, it’s this tiny CBX4-cPRC1 minority that’s functionally essential! (5/11)

What about canonical PRC1? cPRC1 complexes bind H3K27me3 via the CBX subunit to repress gene expression. Our CRISPR-Cas9 screens in patient-derived DMG lines showed that only 2 specific cPRC1 components - CBX4 and PCGF4 (aka BMI1) - are functionally essential in DMG. (4/11)

It was known that while H3K27M inhibits PRC2, residual H3K27me3 at CGIs is crucial for repressing oncogenic programs in DMG. We built on this by showing that while all PRC2 component combinations exist in DMG cells, only the catalytic core is essential for disease biology. (3/11)

First, this was a collaborative effort that would not have been possible without super talented co-first author @dairegannon.bsky.social, co-senior authors Dr Richard Phillips, @adrianbracken.bsky.social & @gerrybrien.bsky.social. Thanks to many other international collaborators too! (2/11)