Affinity map of bromodomain protein 4 (BRD4) interactions with the histone H4 tail and the small molecule inhibitor JQ1

Abstract

Bromodomain protein 4 (BRD4) is a member of the bromodomain and extra-terminal domain (BET) protein family. It binds to acetylated histone tails via its tandem bromodomains BD1 and BD2 and forms a complex with the positive transcription elongation factor b, which controls phosphorylation of RNA polymerase II, ultimately leading to stimulation of transcription elongation. An essential role of BRD4 in cell proliferation and cancer growth has been reported in several recent studies. We analyzed the binding of BRD4 BD1 and BD2 to different partners and showed that the strongest interactions took place with di- and tetra-acetylated peptides derived from the histone 4 N-terminal tail. We also found that several histone 4 residues neighboring the acetylated lysines significantly influenced binding. We generated 10 different BRD4 BD1 mutants and analyzed their affinities to acetylated histone tails and to the BET inhibitor JQ1 using several complementary biochemical and biophysical methods. The impact of these mutations was confirmed in a cellular environment. Altogether, the results show that Trp-81, Tyr-97, Asn-140, and Met-149 play similarly important roles in the recognition of acetylated histones and JQ1. Pro-82, Leu-94, Asp-145, and Ile-146 have a more differentiated role, suggesting that different kinds of interactions take place and that resistance mutations compatible with BRD4 function are possible. Our study extends the knowledge on the contribution of individual BRD4 amino acids to histone and JQ1 binding and may help in the design of new BET antagonists with improved pharmacological properties.

Keywords: Acetylation; Anticancer Drug; BET Family; BRD4; Bromodomain; Cancer Therapy; Chromatin; Histones; Mutant.

FIGURE 1.

Binding of BRD4 BD1 and BD2 to acetylated peptides. A, TR-FRET quantification of BRD4 BD1 and BD2 (100 nm) binding to acetylated peptides derived from H3, H4 or RelA (200 nm each). B, determination of BRD4 BD1 and BD2 affinities to H4 K5(ac)K8(ac)K12(ac)K16(ac) using a TR-FRET homogeneous competition assay. 200 nm biotinylated H4 peptide and 50 nm BRD4 BD1 or 500 nm BRD4 BD2 were titrated with unlabeled H4 peptide at the concentrations indicated. Delta F values were plotted against the concentrations of unlabeled peptide (competitor). The fitting of data to the four-parameter equation described under “Experimental Procedures” (line) served to calculate the K_D values indicated in Table 1. C, SPR sensorgrams of the interaction between BRD4 BD1 (immobilized on a 1000 m polycarboxylate chip) and histone H4 K5(ac)K8(ac)K12(ac)K16(ac). D, SPR sensorgrams of the interaction between BRD4 BD2 (immobilized on a 1000M polycarboxylate chip) and histone H4 K5(ac)K8(ac)K12(ac)K16(ac). The insets in C and D show SPR equilibrium resonance values Req versus [BRD4 BD1] and Req versus [BRD4 BD2] plots from which the K_D values indicated in the text were calculated. Red lines represent the fit of the data to the Langmuir 1:1 and single-site equilibrium binding equations. All data are the mean values of at least two experiments with multiple replicates each.

FIGURE 2.

Alanine scanning of the tetra-acetylated H4 tail. The bar chart shows the average blank-subtracted TR-FRET Delta F (%) values corresponding to the interaction of BRD4 BD1 (100 nm) at three NaCl concentrations with tetra-acetylated histone H4 peptides (200 nm each) modified as indicated. The data represent the mean values with multiple replicates each.

FIGURE 3.

Specific contacts of BRD4 with substrates and the small molecule JQ1 and conservation of key interacting residues within the bromodomain family. A, two-dimensional representation of BRD4 BD1 contacts with K5(ac)K8(ac) H4 region as seen in the three-dimensional structure of the complex (PDB code 3UVW). B, two-dimensional representation of BRD4 BD1 interactions with JQ1 as seen in the three-dimensional structure of the complex (PDB code 3MFX). For A and B, properties of the interacting residues and nature of the interactions are indicated in the legend. C, alignment of selected bromodomain protein sequences. Bold characters indicate the residues that were mutated in the BRD4 BD1 sequence. Equivalent positions in other bromodomains are also in bold and highlighted in red when they are not conserved. Residues analyzed in previous works and mentioned in the text are underlined. Roman figures denote the different bromodomain subfamilies. No member of family VI is shown due to the low sequence conservation in the compared region. Secondary structure elements (α-helices) of human BRD4 are displayed above the sequence alignment. h, human; m, mouse; r, rat; d, Drosophila.

FIGURE 4.

Binding of BRD4 BD1 mutants to acetylated peptides. A, TR-FRET delta F (%) quantification (±S.D.) of BRD4 BD1 (wild-type and mutants, 50 nm each) binding to different acetylated H4 histone tails (200 nm each). B, determination of BRD4 BD1 affinity to H4 K5(ac)K8(ac)K12(ac)K16(ac) with a TR-FRET homogeneous competition assay. 200 nm biotinylated H4 peptide and 50 nm protein were titrated with unlabeled H4 peptide at the concentrations indicated. Delta F values were plotted against the concentrations of unlabeled peptide (competitor). The fitting of the normalized data (not shown) to the four-parameter equation described under “Experimental Procedures” (colored lines) served to calculate the K_D values indicated in Table 1. All data represent the mean values of at least two experiments with multiple replicates each. C, SPR sensorgrams of the interaction of BRD4 BD1 (wild-type and mutants, 10 μm each) with a biotinylated H4 K5(ac)K8(ac)K12(ac)K16(ac) peptide captured on a Biacore SA chip.

FIGURE 5.

Effect of bromodomain mutation on BRD4 dissociation from chromatin by FRAP evaluation. A, nuclei of U2OS cells transfected with GFP-tagged wild-type BRD4, bromodomain mutants or with wild-type BRD4 and treated with JQ1. The bleached area is indicated by a red circle. B, time dependence of fluorescent recovery in the bleached area for wild-type, mutant, or treated cells. Curves represent the means at each time point of at least 12 cells in each group.

FIGURE 6.

Determination of BRD4 BD1 wild-type and mutant affinities for the small molecule inhibitor JQ1 in vitro and in cells. A, FP analysis of BRD4 BD1 wild-type and mutant binding to TAMRA-labeled JQ1. The TAMRA-labeled JQ1 tracer (10 nm) was titrated with increasing concentrations of protein as indicated, and blank-subtracted fluorescence polarization values were plotted versus protein concentration. The lines show the fit of the data to a single-site equilibrium binding equation described under “Experimental Procedures.” The K_D values obtained from these fits are indicated in Table 2. B, TSA analysis of BRD4 BD1 wild-type and mutant binding to JQ1. The compound-induced changes in free energy of unfolding (ΔΔG°_u) were plotted versus JQ1 concentration, and the data were fitted to the equation described under “Experimental Procedures” (colored lines). These fits served to calculate the K_D° values shown in Table 2. C, TR-FRET competition assay. 200 nm biotinylated H4 peptide and 50 nm wild-type or mutant proteins were titrated with JQ1 at the concentrations indicated in the graph. Plots of the Delta F data versus JQ1 concentration were fitted to the four-parameter equation described under “Experimental Procedures” (colored lines) to calculate the K_D values indicated in Table 2. D, cellular stabilization of BRD4 wild-type and mutants by JQ1. HEK293 cells expressing wild-type and mutant variants of BRD4 BD1 were incubated with JQ1 and tested following the protocol given under “Experimental Procedures.” Fitting the plots of TR-FRET ratio at increasing JQ1 concentrations was used to calculate the EC₅₀ values for protein stabilization.

FIGURE 7.

Three-dimensional affinity map of BRD4 interactions with H4 tail and JQ1. Residues investigated in this study are highlighted in the crystal structures of BRD4 complexed with H4K5(ac)K8(ac) (A) and JQ1 (B). The colors represent the loss of absolute free energy of binding (dΔG) associated with the individual mutations. dΔG values were calculated by subtracting free energy of binding of the wild-type from the mutants. Non-mutated BRD4 regions are shown in gray, the H4K5(ac)K8(ac) peptide is shown in green, and JQ1 is in magenta. The dΔG values were calculated from the TR-FRET experiment (panel A) and the dΔG° values from the TSA experiment (panel B). n.i., not interpreted.