Fig. 5: DOT1L inhibition decreases the accessibility of SOX2 target enhancers and SOX2 binding in NPC48h. | Nature Communications

Fig. 5: DOT1L inhibition decreases the accessibility of SOX2 target enhancers and SOX2 binding in NPC48h.

From: DOT1L-mediated murine neuronal differentiation associates with H3K79me2 accumulation and preserves SOX2-enhancer accessibility

Fig. 5

a Characterization of dynamic ATAC peaks. Left panel: stacked barplot summarizing the genomic distribution of the 2000 most dynamic ATAC peaks (blue: TTS, green: exon, yellow: intron, black: promoter-TSS). Central panel: scatterplot showing the association between the log2 fold-change of dynamic ATAC peaks (x-axis) and the log2 fold-change of genes overlapping or proximal to at least one dynamic ATAC peak (y-axis) upon EPZ treatment in NPC48h. ATAC peaks overlapping enhancer regions are shown in purple. Gene symbols are shown for differentially expressed genes upon DOT1L inhibition. Genes associated with dynamic ATAC peaks that are downregulated in both mESC and NPC48h are highlighted in boxes. Top panel: density plot of the accessibility log2 fold-change of ATAC peaks overlapping enhancer upon EPZ treatment, stratified according to genomic annotation. b Differential motif analysis on differentially accessible ATAC peaks upon EPZ treatment in mESC (left) and NPC48h (right). The size and color of each dot is proportional to −log10(p-value) associated with the enriched motif. c SOX2 preferentially binds in vivo to regions with reduced accessibility upon EPZ treatment. Metaprofile and heatmap of SOX2 binding profile in brain-derived NPC38 over open regions losing accessibility (ATAC-Down), gaining accessibility (ATAC-Up) and over unaffected regions (background-ATAC) upon EPZ treatment in NPC48h. Metaprofile and heatmap of the corresponding log2 ratio of EPZ vs DMSO ATAC-seq signal, H3K27ac, H3K4me1 and H3K4me3 input subtracted coverage of DMSO treated sample. d SOX2 binding decreases on enhancers with decreased accessibility upon DOT1L inhibition compared to control. For each gene, % of immunoprecipitated input is shown (y-axis) for SOX2 ChIP on DMSO and EPZ-treated NPC48h. Enhancers overlapping with or proximal to Jarid1, Fgfr2, Msi2, and Pantr1 were tested. One enhancer overlapping Schip1 showing no dynamic accessibility upon DOT1L inhibition was chosen as negative control. Four independent biological replicates were compared for each locus (shown as dots). Error bars represent standard deviation. Significance for differential SOX2 binding was computed with a one-sided paired student t-tests. *p-value < 0.5; **p-value < 0.05; ***p-value < 0.005; n.s: not significant. Exact p-values: Jarid2_1 = 0.0010, Jarid2_2 = 0.0283, Fgfr2 = 0.0134, Msi2 = 0.0009, Pantr1 = 0.0025, Schip1 = 0.2664.

Back to article page