Fig 2: Schematic models of BRDT and BRD4 bromodomains interacting with acetylated nucleosomes.(a) BRDT interacts with acetylated nucleosomes via its BD1 domain. Binding may be initiated through non-specific interactions with DNA, which allow BRDT to localize to chromatin. Specificity is generated through recognition of tandem acetylated lysine residues (K5ac/K8ac) on the histone H4 tail, while the affinity of the interaction is enhanced by bivalent interaction with both the histone tail and DNA. In contrast, BRDT-BD2 does not interact with acetylated nucleosomes and therefore is flexibly tethered to nucleosomes via BRDT-BD1. BRDT-BD2 may function to recruit as-yet-unknown acetylated non-histone proteins to the chromatin. (b) BRD4 interacts with acetylated nucleosomes via both its BD1 and BD2 domains. Our results indicate that BRD4-BD1 binds to nucleosomes through the acetylated histone H4 tail and does not additionally interact with DNA. Bivalent binding of BRD4 through both bromodomains has previously been shown to enhance BRD4 binding affinity for nucleosomes by 2.6-fold37.

Fig 3: Other BET bromodomains interact with DNA.(a) Structures of the human BET bromodomains (PDB codes: BRDT-BD1 (2RFJ); BRD3-BD1 (2NXB); BRD3-BD2 (2OO1); BRD4-BD1 (2OSS, N-terminally truncated to R58 for comparison with other BD1 structures); BRD4-BD2 (2OUO)14; BRD2-BD1 (1X0J)26; BRDT-BD2 structure is a homology model generated using the Phyre2 web server28; BRD2-BD2 (2DVV)). Structures are shown as cartoons with side chain sticks and partially transparent electrostatic surfaces generated using PBD2PQR and APBS565758. (b) EMSA of BET bromodomains (as indicated) interacting with 167 bp double stranded DNA. DNA (0.5 μM) was mixed with BET bromodomains (100 μM) in a final volume of 4 μl and incubated for 30 min before native-PAGE electrophoresis at 4 °C and visualization with ethidium bromide staining. (c) ITC profiles for BRD4(1) interactions with H4K5acK8ac peptides and nucleosomes, as indicated.

Fig 4: The nucleosome structure augments acetylated histone tail binding by BRDT-BD1 but prevents interaction with BRDT-BD2.(a) Domain layout of human BRDT. (b) ITC profiles for N-BRDT(1) interactions with either acetylated histone H4 tail peptides or acetylated nucleosomes (both H4K5acK8ac), as indicated. (c) ITC profiles for BRDT(2) interactions with either acetylated histone H3 tail peptides or acetylated nucleosomes (both H3K18acK23ac), as indicated.

Fig 5: DNA binding by BD1 is important for BRDT localization to chromatin in cells.(a) EMSA of N-BRDT(1) lysine mutants (concentrations: 3, 12.5, 50 and 200 μM) binding to 167 bp double-stranded DNA. DNA (0.4 μM) was mixed with N-BRDT(1) 3KE in a final volume of 5 μl and incubated for 30 min before native-PAGE electrophoresis at 4 °C and visualization with ethidium bromide staining. (b) ITC profiles for N-BRDT(1) 3KE interactions with either acetylated histone H4 tail peptides or acetylated nucleosomes (both H4K5acK8ac), as indicated. (c) FRAP analysis of human and murine ΔC-sBRDT constructs in the presence of TSA-induced histone hyperacetylation. Cos7 cells were transfected by vectors expressing GFP-tagged WT and mutant ΔC-sBRDT constructs (as indicated) and cells were treated with the histone deacetylase inhibitor TSA (100 ng ml−1) to induce histone hyperacetylation. A decrease in fluorescence recovery half-life (t1/2) indicates an increase in protein mobility, and reduced chromatin association. The indicated fluorescence recovery half-lives are mean values obtained from 10 independent cells (n=9 for WT murine ΔC-Brdt). Error bars show the s.e.m. (d) Confocal microscopy images of representative transfected cells following TSA treatment (Scale bars, 10 μm).