Read on Twitter
Our latest study on dynamics of chromatin remodeler activity is out now!
Thread
: https://www.nature.com/articles/s41588-020-00768-w
Thread
: https://www.nature.com/articles/s41588-020-00768-w
Regulatory regions are usually characterised by features of "open chromatin", as defined by increased accessibility to nucleases or transposases. These are regulated by chromatin remodelers, which orchestrate nucleosomes and allow temporal access to DNA.
But how dynamic is this process? You could imagine a scenario where remodelers are only required to establish a certain "open chromatin" site, which is then self-mantained possibly until the next replication cycle. Alternatively, you would need constant activity to keep it open.
We used chemical inhibition of the SWI/SNF remodeler complexes in mouse embryonic stem cells to try to address this question. The BRM014 compound we used targets both SWI/SNF ATPases Brm and Brg1 and was initially described here: https://pubs.acs.org/doi/10.1021/acs.jmedchem.8b01318 https://www.biorxiv.org/content/10.1101/812628v1
We first validated the specificity of BRM014 in mouse. We saw that treatment with BRM014 caused changes in chromatin accessibility that were overlapping those caused by genetic deletion of Brg1 (Brm is not expressed in mESCs).
Importantly, changes induced by treatment with SWI/SNF inhibitor did not resemble changes caused by deletion of Snf2h (ISWI) or Ino80 , both at the level of chromatin accessibility or transcriptome, showing the compound does not target ATPases unspecifically.
So the inhibitor works... let's use it! We then performed time-course of inhibition followed by ATAC-seq. We saw very fast response (already after 10 minutes), with regions that immediately show some loss of accessibility which gradually increases over time.
This fast response argues for continuous requirement of remodeler activity independently of cell cycle. To further test this, we measured the effect of SWI/SNF inhibition in cells sorted on DNA content and found that BRM014 affected accessibility in very similar way at all stages
What about transcription factors you ask? TFs whose binding is regulated by SWI/SNF show reduced accessibility at their binding motifs, while independent TFs are unaffected. This is also evident at short time frames. Binding of these TFs is also reduced as measured by ChIPseq.
We then asked whether this process was reversible, and with what dynamics. To answer this, we induced SWI/SNF inhibition, then washed out the compound, allowed cells to recover and measured chromatin accessibility again.
We saw that, with similarly fast dynamics, chromatin accessibility profile was reestablished in ESCs upon reintroduction of remodeler activity, arguing for a cell-intrinsic competence for chromatin accessibility establishment.
All features were quickly recovered, including accessibility profiles, transcriptome changes and nucleosome accumulation over respondent TFs as measured by MNase-seq.
Together, these data argue that the generation and maintenance of open chromatin sites are highly dynamic processes that require constant remodeler activity. Rather than “open chromatin”, it is more like a revolving door.
It would be interesting to see whether this is something we can generalise about chromatin remodelers, or whether SWI/SNF and SWI/SNF-regulated regions are special in that regard.
I really enjoyed working on this with Michael, Francesca, Zainab, Giorgio and Dirk. Trying to get this out during 2020 has been an interesting journey. I can also say without a doubt that the paper is much better now that it was before peer-review, so thank you reviewers! :)
Please also check out @_SandraSchick et al. from the Kubicek Lab, also out today. Our groups independently reached similar conclusions and upon discovery decided to coordinate submission. I think these studies nicely complement each other. https://www.nature.com/articles/s41588-021-00777-3
