Fig. 3

H3K27me3 micro-domains may be favored depending on long-range interactions with distant insulators depending on co-factors. a Genomic view of the Mio locus showing the ChIP-seq reads (y-axis) of Beaf32, H3K27me3, and CP190 in Beaf-KD-depleted and control cells. See panel c for a zoom of the micro-domain. The single asterisk indicates the downregulated genes (in Beaf-KD compared to WT control; e.g., Tsp39D) and double asterisks the upregulated genes (Mio (also called or “Mondo”) and crc). b Upper panel: Genome-wide analysis representing the relative enrichment (log odds ratio) of genes with decreasing (down-) or increasing (up-) H3K27me3 levels upon Beaf-KD compared to control cells, depending on binding of the indicated insulator protein co-factor CP190 alone or in combination with GAF (or dCTCF; not shown). Lower panel: same analysis for the insulator co-factor cohesin alone or in combination with GAF. The indicated p values (asterisks) were calculated using Fisher’s exact test in presence compared to absence of binding (see Additional file 3: Fig. S3). c Zoom of the micro-domain associated with the Mio locus of chromosome 2L. The arrows represent the sites where H3K27me3 levels decrease significantly upon Beaf-KD compared to control cells. d Chromosome conformation capture (3C) analysis of the long-range interactions with the micro-domain containing Mio promoter and a distant Beaf32 peak. The graph represents the relative frequency chimera products as measured by qPCR from Cohesin-depleted (blue), CP190-depleted (orange), or control siRNA-depleted (blue) cells. Proximal ligation products were estimated after HindIII restriction at the indicated Beaf32 (orange triangle) as anchor site using reverse primer and TaqMan-MGB probe (see “Methods”) with systematic measurements using primers spanning the whole Mio locus. Variations were tested by Student’s t test. (Additional file 3: Fig. S3C; reciprocal 3C)