iiinova — where your digital self keeps evolving

Together these data reveal a temporal cascade in which FACT depletion causes 5’ loss of nucleosomes and their associated histone modifications. This chromatin disruption occurs prior to transcriptional inhibition More in the paper so take a look!

100

Ultimately, we observe transcriptional collapse and more specifically downregulation of genes. But this happens well after the initial reduction in nucleosome phasing observed.

100

So what impact does this have? Well, loss of histone recycling leads to a build up of RNAPII in the 5' end of genes and reduced transcription elongation, as indicated by changes to H3K36me3.

100

I told you that FACT's impact is greatest at highly transcribed locations, but does it require transcription? Again inspired by the Robert lab, we show that yes, FACT's function is dependent on transcription

100

Can we rescue this defect by bringing FACT back? Work led by our amazing undergraduate research Bryona Jackson found that there is only a partial restoration. Note for others: washing out dTAG doesn't always work, so we resorted to crafty ways of rescuing. But not full restoration.

100

Is the 5' nucleosome loss and increased TF invasion occurring at the same place? Yes! Over highly transcribed locations especially, which rely more on FACT activity

100

We corroborated our short read sequencing results with single molecule long read Fiber-seq and found reduce 5' and increased 3' nucleosomes in FACT depletion

How is this happening? Beautiful work from F. Robert group in budding yeast found FACT is important for recycling of histone PTMs, and inspired by this we hypothesized that chromatin opening would permit TF invasion. We found that 5' nucleosome phasing is lost over time in FACT depletion.

However, using rapid dTAG depletion and high resolution time course analysis, we find that FACT depletion results in TF impingement into gene bodies, in a sequence independent manner, but occurring more readily over highly transcribed genes.

100

For both these remodelers, these tend to be OSN bound and also Suz12 bound, and enriched as bivalent promoters. Together, these data suggest that the enhancer-promoter loops esBAF promotes and INO80C restricts are pluripotency and bivalency related.

Enhancer-promoter loops impacted by BRG1 or INO80 loss are enriched for OCT4, SOX2, NANOG and SUZ12, and are described as bivalent locations.

100

Therefore, to investigate the impact of these remodelers more directly on enhancer-promoter loops, we performed promoter capture microC (PCMC) and found that KD of BAF ATPase BRG1 results in decreases in some E-P loops and KD of INO80C ATPase Ino80 results in increases in some E-P loops.

Promoter capture microC (PCMC) enriches for TSS-based loops and shows decreased loops in Brg1 KD and increased loops in Ino80 KD

100

Although we sequenced our Hi-C datasets deeply, we had a tough time calling non-architectural loops (most loops were CTCF-based, and not as many were TSS-based)

Our Hi-C data detects limited TSS-based loops, but is enriched for CTCF-based (architectural) loops

Similar to work from the Schubeler lab for BAF, we find that loss of neither BAF nor INO80C impact TAD structures significantly.

100

However, there is a reproducible effect on subcompartment structures, where upon KD of either ATPase, we observed a modest increase in the total amount of the genome assigned to inactive subcompartments, especially at locations that these remodelers bind and regulate transcription.

subcompartment analysis of Hi-C data showing impact of remodeler depletion on subcompartments

100

48hr KD of the ATPase subunit for either complex has almost not impact on compartments, assessed by Hi-C:

heatmaps from Hi-C data showing compartment structures

100

the Rtf1 mutation not being sufficient to break the interaction, and/or represent the complexity and redundancy present in this more complicated system.

So in a murine cell line system (ES cells) what happens? We made the CHCT deletions and an Rtf1 mutation ES cell line, but did not observe the same nucleosome shifts. This may be due to reduced protein levels observed in the Chd1 and Chd2 CHCT deletion cell lines...

100

Chd1 and Rtf1 are conserved proteins, so we wanted to know if this interaction is conserved in mammalian systems. Y2H using murine constructs demonstrate an interaction between Chd1 and Rtf1 as well as the related remodeler, Chd2, and Rtf1. CHCT deletions and Rtf1 mutations reduce this interaction

100

Histone PTMs, K4me3 and K36me3 are also shifted 5'

100

As Chd1 is an important nucleosome remodeler, what does this loss of appropriate localization mean to genic nucleosomes? Well, nucleosomes (and overlapping dinucleosomes aka hexasome-nucleosomes) are also shifted 5', and this is exacerbated in Isw1 delete.

100

By mutating either side of this interaction (deleting the CHCT domain or precise point mutations in Rtf1), we find that while Rtf1 occupancy is unaltered, Chd1 occupancy is shifted 5'

100

Sarah Hainer @hainerlab.bsky.social · 01/01/2025

Building off the finding from the Arndt lab in 2003 (10.1093/emboj/cdg179), we narrowed down an interaction between the understudied Chd1 CHCT domain and a N terminal LALA box in Rtf1, a member of the Paf1 elongation complex, which form a direct interaction in budding yeast

100

Happy New Year from us to you!

080

Excited by these findings we made some mutants in mES cells to ask if disrupting this interaction resulted in similar changes to CHD distribution and nucleosome positions. Perhaps due to complexity or resolution, we don't observe these changes.

Is this mechanism for Chd1 distribution conserved in mammals? To test these we asked if murine CHD1 and CHD2 (close homolog) interact with murine RTF1 using yeast two hybrid, and yes, they do! And in fact, this is the conserved LALA box (LLSLA in mammals) and CHD1/2 CHCT

This also results in a concordant 5' shift in resident histone posttranlational modifications

This disrupted interaction and 5' Chd1 shift then leads to a 5' shift of nucleosomes, subnucleosomes, dinucleosomes, and hexasome-nucleosome (OLDN) complexes.

Using precise mutants that disrupt this interaction, Sarah found 5' shift in Chd1 localization across genes, with modest effects on RNAPII and Rtf1 occupancy