A hybrid mechanism of action for BCL6 in B cells defined by formation of functionally distinct complexes at enhancers and promoters

Abstract

The BCL6 transcriptional repressor is required for the development of germinal center (GC) B cells and diffuse large B cell lymphomas (DLBCLs). Although BCL6 can recruit multiple corepressors, its transcriptional repression mechanism of action in normal and malignant B cells is unknown. We find that in B cells, BCL6 mostly functions through two independent mechanisms that are collectively essential to GC formation and DLBCL, both mediated through its N-terminal BTB domain. These are (1) the formation of a unique ternary BCOR-SMRT complex at promoters, with each corepressor binding to symmetrical sites on BCL6 homodimers linked to specific epigenetic chromatin features, and (2) the "toggling" of active enhancers to a poised but not erased conformation through SMRT-dependent H3K27 deacetylation, which is mediated by HDAC3 and opposed by p300 histone acetyltransferase. Dynamic toggling of enhancers provides a basis for B cells to undergo rapid transcriptional and phenotypic changes in response to signaling or environmental cues.

Figure 1

Genome wide distribution of BCL6 and corepressors BCOR, SMRT and NCOR in DLBCL cells. (A) Tumor growth plot in DLBCL xenografted mice (Farage cell line) after treatment with RI-BPI vs control peptide (50 mg/kg/day for 5 consecutive days) (B) Genomic distribution of BCL6 peaks and BCL6 peaks coinciding with BCOR (BCL6-BCOR), SMRT (BCL6-SMRT) and NCOR (BCL6-NCOR) peaks based on their location relative to RefSeq transcripts (hg18) in OCI-Ly1 cells. (C) Venn diagrams representing the overlap of BCL6, BCOR and SMRT ChIP-seq peaks in DLBCL cells. (D) GSEA analysis integrating ChIP-seq and mRNA-seq results after BCL6 knockdown. The enrichment of promoter target genes bound by BCL6 ternary complexes (BCL6-BCOR-SMRT), BCL6-BCOR only, BCL6-SMRT only, or BCL6 only was tested based on decreasing gene all expression log ratios (48 h; siBCL6/siNT). Weighed statistic and 5000 sample permutations were used. NES: Normalized enrichment score, FDR: False Discovery Rate. (E) Pathway analysis using BCL6-ternary target genes (n=126) upregulated more than 1.5 fold after BCL6 knockdown. See also Figure S1 and Table S1.

Figure 2

BCL6 can recruit both BCOR and SMRT corepressor complexes simultaneously through homodimerization of its BTB domain in promoters. (A–B) BCL6 binding (green) at the promoters of CD69 and BANK1 coincides with binding of corepressors BCOR (blue) and SMRT (red) respectively. y-axis values represent read densities normalized to total number of reads. (C) ChIP-re-ChIP assay in BCL6-BCOR-SMRT promoters using BCOR and SMRT antibodies. IgG was used as a negative control. A BCL6 intron 9 locus is shown as a negative control. Experiment was performed in duplicate using triplicate wells. The y-axis depicts enrichment as % input +/−SEM. (D) FRET assay for A488-BCOR and BODIPY-SMRT peptides in solution with BCL6 BTB. Fluorescence (AU) is shown as a function of increasing BCL6 BTB concentration. FRET emission is generated when both peptides bind to the BCL6 BTB dimer. Higher concentrations of BCL6 BTB dimers increase single peptide binding evens decreasing FRET emission. (E) Hybrid model of the BCL6 BTB dimer (each monomer in violet and pink) simulated in complex with two peptides corresponding to BCOR 498–514 (green) and SMRT 1414–1430 (cyan) used in FRET assay. See also Figure S2.

Figure 3

Potent BCL6 repression occurs within repressed chromatin states and is linked to RNA Pol II pausing. (A) Graphical representation of weighted PCA analysis integrating histone mark and DNA methylation enrichment levels surrounding all TSS. The top two PCs are shown. Color key indicates the weights of the original variables assayed by ChIP-seq and ERRBS. (B) Correlation of genes corresponding in each PC with basal levels of gene expression and gene expression changes after BCL6 knockdown (48h) is indicated. (C) Cumulative distribution of RNA pol II Pausing Ratios (calculated as the fraction of normalized read density ratio of Ser5-P Pol II (paused) around the TSS (−100 to +200bp) to the Ser2-P Pol II (elongating) density at the TES (TES+2kb)) comparing BCL6 target genes upregulated after BCL6 siRNA versus the rest of BCL6 target genes. p-value is indicated. See also Figure S3.

Figure 4

BCL6-SMRT complexes mediate enhancer silencing (A) Overlap of distal/intronic BCL6-SMRT peaks with H3K4me3 and H3K4me1 peaks in DLBCL cells. (B) BCL6, SMRT and H3K4me1,3 ChIP-seq read density profiles in BCL6-SMRT enhancers centered at the BCL6 peak summit. y-axis represents average read densities normalized to total number of reads. (C) GSEA analysis of genes proximal to BCL6-SMRT bound enhancers or BCL6 non-SMRT enhancers. Genes were ranked based on decreasing log₂ RPKM (siBCL6/siNT) at 48 h BCL6 knockdown (weighted p2 statistic, 5000 permutations). (D) Comparison of fold expression induction (siBCL6/siNT RPKM, 48 h) of genes proximal to BCL6-SMRT enhancers or BCL6 non-SMRT enhancers versus other genes. (E) A BCL6-SMRT enhancer located −13kb upstream of CDKN1A is illustrated. UCSC tracks of BCL6, SMRT, H3K4me1,3 and total RNA polymerase II density normalized to the total number of reads are represented. Location of BCL6 sequence motifs is indicated (F) Reporter assays performed in DLBCL cells with constructs containing CDKN1A promoter alone (Prom), promoter and wild type −13kb enhancer (Prom + wt Enh) or promoter and mutant enhancer (Prom+ mut Enh). Cells were treated with siBCL6 and siNT as indicated in quadruplicates. Y-axis represents fold repression of reporter based on relative luciferase (vs. TK-Renlla) compared to the basal activity of the promoter construct. Data are represented as mean +/− SEM. See also Figure S4.

Figure 5

BCL6 recruits SMRT to deacetylate H3K27 leading to enhancer inactivation. (A) Selected BCL6-SMRT bound enhancers or BCL6 only enhancers were tested for enrichment of H3K27ac by Q-ChIP in OCI-Ly1 cells nucleofected with siBCL6 or siNT. Relative enrichment is normalized to siNT and shown as mean +/− SEM (B) H3K27ac immunoblot of in vitro histone deacetylation reactions using immunoprecipitated SMRT and HDAC3 in the presence or absence of TSA (Trichostatin A). IgG was used as a negative IP control and H3 as a loading control. Coomassie stain indicates that equal amount of antibody was added in each IP. (C) H3K27ac immunoblot using whole cell extracts of B220+ cells isolated from two Hdac3^+/+/ROSA-GFP/CD19-Cre and two Hdac3^FL/−/ROSA-GFP/CD19-Cre mice. Hdac3 depletion in null cells was confirmed. H3 and β-globin were used as loading controls. (D) Biological replicates of H3K27ac QChIP performed in triplicates in OCI-Ly1 cells exposed to 50 µM of 79-6¹⁰⁸⁵ or vehicle (DMSO) for 30 min. Fold H3K27ac enrichment vs. vehicle is shown (y-axis). Error bars represent the mean +/− SEM. See also Figure S5.

Figure 6

BCL6-SMRT complexes antagonize p300 acetyltransferase activity to mediate enhancer toggling. (A) Comparison of the average normalized H3K27ac read density levels in BCL6 enhancers bound by SMRT but not p300, bound by both SMRT and p300 or bound by p300 but not SMRT (B) Log₂ change of H3K27ac levels upon BCL6 knockdown in BCL6 enhancers bound by SMRT but not p300, bound by both SMRT and p300 or bound by p300 but not SMRT. p-values are indicated (C) Graphical representation of weighted PCA analysis using distal BCL6 binding sites. Correlation with gene expression changes after BCL6 depletion (24 h) links PC3 (715 distal enhancers, p < 1e-8) to gene derepression. Color key indicates the weights of the original variables assayed by ChIP-seq. See also Figure S6.

Figure 7

Model of BCL6 repression mechanism. BCL6 dimers can simultaneously recruit PRC1-like BCOR complexes and HDAC3-containing SMRT complexes in B cell promoters to most effectively repress transcription in a repressed chromatin environment. Alternatively BCL6 selectively recruits SMRT to functionally inactivate a network of B cell enhancers through H3K27 deacetylation opposing the effect of p300 complexes that mediate H3K27 acetylation. BCL6 BTB inhibitors dismiss BCOR and SMRT complexes and reactivate the BCL6 targeted gene programs thereby killing lymphoma cells while BCL6-BTB point mutations abrogate GC formation.