Active regulatory elements recruit cohesin to establish cell-specific chromatin domains [ChIP-seq]

The genome adopts varied and specific 3D conformations in different cell types. The loop extruding factor, cohesin, is thought to play a key role in the formation of this 3D structure, however it is unknown how cell-type specificity is achieved. One proposed mechanism is that cis-regulatory elements, which are active in unique combinations in different cell types, recruit cohesin and therefore direct the formation of specific 3D conformation. We have identified genome-wide natural variation in enhancer sites in healthy human donors, by performing cohesin ChIP-seq in these cells we show that cohesin occupancy is directly correlated with enhancer activity. Identification of the genetic variants associated with the gain or loss of enhancer-like elements and associating this with the recruitment of cohesin provides evidence of a direct link between widespread genetic variation in the non-coding genome and cell-specific changes in 3D genome organization. Therefore, enhancer elements have evolved not only the ability to upregulate transcription at promoters, but also the ability to stimulate the mechanisms that insure robust interactions between them. Next, using enhancer KO models in the well characterised alpha globin locus, we show how cohesin recruitment and 3D genome organization correlates with changes in the local regulatory landscape. We then selected the strongest of the alpha globin enhancers and inserted it into a neutral region of the genome, to test if a single active enhancer was capable of recruiting cohesin and initiating the formation of a 3D chromatin domain when isolated from its native context. Our results show that an enhancer is able to direct the recruitment of cohesin when in an active state and that this is correlated with changes in the local chromatin environment, giving rise to open chromatin sites and de novo chromatin interactions in an otherwise “neutral” region of the genome. These results support the hypothesis that tissue-specific recruitment of cohesin is facilitated by enhancers and that this plays a role in establishing cell-specific changes in 3D genome conformation. Overall design: ChIP-seq was performed on aliquots of 3-10x10^6CD71+ cells. For calibrated ChIP-seq, a 1% or 4% spike-in of WT HEK293 cells were added prior to sonication. Cells were double cross-linked using disuccinimidyl glutarate (DSG, Sigma) and 1% formaldehyde (Sigma) for a total fixation time of 1 hour. Fixed chromatin samples were fragmented using the Covaris sonicator (ME220) for 10 minutes (75 power, 1000 cycles per burst, 25% duty factor) at 4°C. 50µL per sample was removed as an input control. Sonicated samples were pre-cleared to remove background signal through incubation with a 1:1 mix of Protein A/G Dynabeads (InVitrogen). Antibody was added at a concentration of 1 µg/µL per sample and immunoprecipitated at 4°C overnight. Immunoprecipitated samples were incubated with a 1:1 mix of Protein A/G Dynabeads for 5 hours at 4°C. Beads were then washed 4x in RIPA buffer on a magnetic stand, followed by 1x wash with TE (Sigma Aldridge) + 50mM NaCl (ThermoFisher). Chromatin was eluted from the beads using elution buffer and incubated for 30 minutes at 65°C with shaking. Input samples were diluted 1:1 with elution buffer. Samples and input controls were incubated at 65°C overnight for de-crosslinking, before RNase (Roche) and proteinase K (BioLabs) treatment. DNA fragments were purified using the Zymo ChIP DNA Clean & Concentrator kit (Zymo Research) and eluted in water. DNA concentration was quantified using the Qubit dsDNA HS assay (Invitrogen) as per manufacturers protocol. To assess sonication efficiency the D1000 Tapestation (Agilent) assay was performed on input sample. An approximately equal mass of input and IP DNA was used for indexing (0.5 ng – 1 µg). NEBNext Ultra II DNA Library Prep Kit (New England Biolabs) was used to prepare indexed sequencing libraries following the manufacturer supplied protocol. PCR amplification was performed for 7-11 cycles (depending on input DNA concentration) using NEBNext Mulitplex Oligos (New England Biolabs). Indexed sample concentration was quantified using the KAPA Library Quantification Complete Kit (Universal)(Roche). Samples were pooled as a 4 nM library and sequenced with a high-Output v2 75 cycle kits on the Illumina NextSeq platform.