Fig 2: BET family proteins shape ACD organization in the absence of Cohesin. Related to Fig.2.(a) Low-dose 1,6 Hexanediol (1,6-HD,2%) treatment for 1.5 hours with or without Cohesin depletion did not significantly affect chromatin accessibility in the linear genome. The heatmaps of ATAC-seq enrichment at CTCF sites, enhancers and promoters are shown for WT, Cohesin depletion, 1,6 HD treatment, or a combination of both.(b) Disrupting BET-chromatin interactions by JQ1 had no significant impact on accessible chromatin clustering after CTCF depletion. g(r) curves were plotted for indicated conditions. Specifically, 1µM JQ1 treatment for 12 hours was applied to both control and CTCF depleted cells (6-hour depletion). The two-sided Mann-Whitney U test was used for statistical testing.(c-d) Violin plot of 3D volumes (the number of voxels) of two ATAC-rich segments R2 (c) and R5 (d) before (grey) and after (red) auxin-mediated CTCF depletion with or without JQ1 or 1,6HD treatment. The number of alleles analyzed is indicated at the bottom. The black bar indicates the median value for each data set and Mann-Whitney U test was performed for the statistical analysis.(e) The disordered region predicted from PONDR program (Predictor of Natural Disordered Regions) for BET family proteins using the VSL2 algorithms. Except for the double bromodomain and the extra-terminal domain, other regions of BET proteins are predicted to be highly disordered with high disorder score.(f) RNA-seq analysis of the expression level of BET family genes in mouse ESCs and mouse embryonic fibroblasts (MEF). Except for the Brdt gene, Brd2/3/4 are all actively transcribed in ESCs. RNA-seq data was derived from GSM723776 and GSM723775.

Fig 3: BRD4 regulates BRD2 chromatin binding and function. Related to Fig.4.(a) Enrichment profile and heatmap of BRD2/3/4, CTCF, P300 ChIP-seq signal at their respective binding peaks before and after BRD2 depletion.(b) Enrichment profile and heatmap of ChIP-seq signal of BRD2, BRD3, BRD4 and CTCF at their cognate binding peaks after 6 hours of BRD3 depletion.(c) Heatmap of BRD2 binding signal specifically at BRD2/BRD4 co-bound regions before and after acute BRD4 depletion for 6 hours.(d) Representative genomic tracks (from integrated genomics viewer) of IgG, BRD3 and p300 after BRD4 depletion. BRD4 depletion markedly reduced P300 binding.(e) Validating the specificity of fluorescence signal or antibodies used for immunofluorescence after acute depletion of RAD21 or BRD2/BRD4. The RAD21-mAID-eGFP : BRD2-HA-dTAG : BRD4-Halo-dTAG triple degron mESC line was used in this study. RAD21 was detected by covalently tagged eGFP signal. BRD2 was detected by mouse monoclonal anti-HA antibody and goat-anti-mouse Alexa568 secondary antibody. BRD4 was detected by rabbit BRD4 primary antibody and goat-anti-rabbit Alexa647 secondary antibody. To validate the fluorescence signal specificity, RAD21 and BRD2/BRD4 were depleted by adding auxin (500µM) or dTAG13 (100nM) for 6 hours, respectively.(f) (Upper panel) Representative images of spatial distribution of RAD21, BRD2 and BRD4 signal. (Lower panel) Fluorescent Intensity profile of RAD21(green), BRD2 (cyan) and BRD4 (magenta) relative to DAPI (blue). RAD21 is enriched at both DAPI high and low regions whereas BRD2 and BRD4 are enriched at DAPI low regions. Scale bar, 3µm.(g) Representative single cell view of Cohesin (RAD21, green), BRD2 (cyan) and BRD4 (magenta) in their merged view. The while line indicates the cross-sectional line to analyze the fluorescent intensity profile. The cyan arrow indicates the BRD2 puncta showing little colocalization with BRD4. The magenta arrow indicates the BRD4 puncta poorly colocalized with BRD2. Scale bar, 1µm.(h) Fluorescence intensity profile of Cohesin (RAD21, green), BRD2(cyan) and BRD4 (magenta) along the while line in (g).(i) Whole nucleus 3D voxel-to-voxel correlation summary of RAD21 vs BRD2 and BRD4 vs BRD2 as shown in Fig.4c. The percentage of voxel colocalization of protein pair, the Pearson’s coefficient, and the Manders’ coefficient are computed and summarized.

Fig 4: Genome-wide BRD2 and BRD4 occupancy following stimulation of NK cells with IL-15 (10 ng/ml) and 24 h of JQ1(+) treatment of NK cells. (A) Coverage plot showing average read counts centered around the transcriptional start site (TSS) from BRD4 ChIP-seq following DMSO or JQ1(+) treatment of NK cells. (B) Coverage plot showing average read counts centered around the TSS from BRD2 ChIP-seq following DMSO or JQ1(+) treatment of NK cells. (C) Heatmap of BRD4 and BRD2 following treatment of NK cells with DMSO or JQ1(+) covering 1 kb upstream and downstream of the center of called peaks. (D) Representative ChIP-seq coverage profiles for BRD2 and BRD4 across the NCR3 locus. (E) Representative ChIP-seq coverage profiles for BRD2 and BRD4 across the IFNG locus.

Fig 5: Comparison of RVX-208 and JQ-1 interactions with the bromodomain ligand binding sites.A. Overlay of BRD4[BD1] (pale green, cyan and light gray) and BRD2[BD2] (green and wheat) when complexed with RVX-208. The three residue insertion difference in BRD2[BD2] is colored in yellow. In all three panels, the hydrogen bonds are illustrated with dashed lines and the waters are removed for clarity. B. A comparison of the BRD4[BD1] when complexed with RVX-208 (pale green, cyan and light gray) and with JQ-1(purple and light blue, 3MXF.pdb). C. A comparison of the BRD2[BD2] when complexed with RVX-208 (green and wheat) and with JQ-1 (burgundy and salmon, 3ONI.pdb).