Fig 2: CHD1 and CHD2 control H3 and H3.3 occupancy at active regulatory regions. (A) Chromatin immunoprecipitation (ChIP-qPCR) analysis of control CHD1, CHD2 and double knockdown experiments at promoters associated with DNase hypersensitivity sites (DHS) show increased H3 occupancy with CHD1, CHD2 and double knockdowns in K562 cells. The strongest effect on H3 occupancy is observed with CHD2 knockdown. CHD2 knockdown also reduced H3.3 enrichment. (B) At intergenic and intragenic DNase sensitivity sites CHD1 and CHD2 knockdown increase H3 occupancy with the strongest trend observed in CHD2 and double knockdown cells. As for promoter sites, CHD2 knockdown reduced H3.3 enrichment. (C) Active tRNA loci-associated DNase sensitive sites show similar trends for both H3 and H3.3 occupancy. (D) Control loci from nonaccessible chromatin demonstrate the effects of CHD2 knockdown on H3 and H3.3 occupancy are specific to active regions. All data points represent mean values plotted from three biological replicates per loci tested. Differences compared to control were tested by one-way ANOVA, *P <0.05, **P <0.01, ***P <0.001.

Fig 3: CHD1 and CHD2 are enriched at an enhancer-like chromatin state with RNA polymerase II (Pol II). (A)The occupancy of CHD1 and CHD2 occurs primarily at enhancer-like chromatin state 4, which also showed the strongest binding by Pol II (Kruskall-Wallis test with Dunn’s post test correction, k = 3, P <0.0001, comparing enhancer state 4 to all other enhancer-like chromatin states). (B) In contrast H3K4 methylation was observed to different extents at all enhancer-like chromatin states.

Fig 4: CHD1 and CHD2 co-occupy active promoters but show different binding profiles across the transcription start site (TSS). (A) Statistically significant high confidence peaks from CHD1 and CHD2 chromatin immunoprecipitation sequencing (ChIP-seq) data were intersected with chromatin state annotations from K562 cells using model-based analysis for ChIP-seq (MACS) with a stringent cut-off p value of 1 x 10-15. The majority of CHD1 (94%) and CHD2 (78%) binding sites fall within an active promoter or enhancer chromatin state or both as defined by Ernst et al. (B) The majority of CHD1 and CHD2 bound sites are overlapping; 60% of the 7,736 CHD2 binding sites overlap with at least one of 9,827 CHD1 peaks. (C) At highly expressed genes defined in the top 20% expression quintile, CHD1 binding profile follows the enrichment of H3k4me2/3-marked nucleosomes. In contrast, CHD2 binds strongest across the micrococcal nuclease (MNase) sensitive nucleosome depleted region (NDR) at the TSS.

Fig 5: Recruitment of CHD1 and CHD2 at CpG island (CGI) promoters correlates with RNA polymerase II (Pol II) occupancy independently of micrococcal nuclease (MNase) sensitivity. Genes with CGI overlapping the transcription start sites (TSS) were organized into quintiles of ascending Pol II occupancy in K562 cells. The average MNase-seq signal, CHD1 and CHD2 occupancy was then plotted across the region from -2500 to +2500 bases from the annotated TSS. Light gray lines represent the bottom quintile, medium gray the second lowest quintile, dark gray the third quintile, black the fourth and red the fifth quintile of Pol II occupancy.