Finally, our model predicts that only the endogenous loci should be affected by loss of H3K27Ac as the ectopic landing pad is located within a chromatin neutral environment. And indeed, CREs like the Xpa enhancer lose accessibility endogenously but not when inserted at the ectopic site. 11/11
We tested this idea by perturbing TF binding. At the motif level, we leveraged the natural genetic variation between mice species crossed into F1 lines. Here, SNPs reduced TF motif affinity across alleles. CREs losing single TF binding showed <10% reductions in chromatin accessibility frequency.6/11
We developed FootprintCharter to classify molecules into accessible or nucleosome-occupied at TF motifs by unsupervisedly detecting TF and nucleosome footprints. FootprintCharter is distributed on Bioconductor at bit.ly/3XLe8RC. For more details, read our annex preprint at bit.ly/3EkRqJh. 3/11
Can transcription factors still find and bind their motif at this ectopic site?
Yes! But while they do bind, they typically open chromatin to a lower extent than at their endogenous counterparts—with the notable exception of the insulator CTCF. 8/11
To dissect the contribution of TF binding vs. chromatin context in driving chromatin opening, we developed a Multiplexed Chromatin-Integrated Reporter Assay. It measures the ability of hundreds of CREs (<300 bp) to recruit TFs and establish accessibility at a defined ectopic locus. 7/11
To test this, we chemically inhibited p300, a histone acetyltransferase that deposits H3K27Ac at a subset of enhancers in mESCs. As predicted, chromatin accessibility dropped, and the amplitude of the drop scaled with the degree of H3K27Ac reduction. 10/11
We found that the frequency of chromatin accessibility at CREs scales with their increasing number of bound TF motifs. Overall, any additional TF binding event increases chromatin accessibility frequency and width rather moderately, i.e., by ~10%. 5/11
When we stratified endogenous CREs by their H3K27Ac ChIP-seq signal, we found that the higher the acetylation, the more frequently the chromatin is accessible.
But at the ectopic site, this scaling disappears—suggesting that H3K27Ac enhances accessibility frequency endogenously. 9/11
We quantified the heterogeneity of chromatin accessibility at mouse ESC CREs. We observed that, within the same cell population, accessible molecules systematically co-exist with nucleosome-occupied ones. At enhancers this is much more prominent than at promoters. 4/11
Bulk assays such as ATAC-seq selectively sequence accessible loci over nucleosome-occupied ones, losing information on accessibility frequency at CREs. Instead, SMF marks accessibility by methylation footprinting and sequences all molecules, revealing the frequency of accessibility across cells.2/11
