Fig. 2

CTCF&BORIS bound regions enclose at least two closely located CTCF binding sites. a Upper panel: gene track represents CTCF (red) and BORIS (blue) binding at the TSP50 (PRSS50) promoter in K562 cells. Lower panel: the alignment of human and mouse sequences under the summit of CTCF (red nucleotide) and BORIS (blue nucleotide) peaks mapped by ChIP-seq at the TSP50 promoter. Two CTCF core motifs (shown on the top of the alignment) coincide with two CTCF binding sites (site1 and site2) previously identified by EMSA and methylation interference assay (MIA) [32] in both mouse and human TSP50 promoters (labeled and underlined at the bottom of the alignment). Asterisks denote the contact guanines mapped by MIA. The space between the two CTCF binding motifs in human and mouse TSP50 promoters varies from 33 bp to 58 bp (shown by brackets). The phastCons conservation track shows the high conservation of two CTCF sites in the TSP50 promoter among 100 vertebrates. The black arrow shows the beginning of TSP50 transcription. b The percentage of CTCF-only, CTCF&BORIS and BORIS-only peaks with two or more CTCF motifs. The top 1000 CTCF-only, CTCF&BORIS and BORIS-only binding regions (invariant in three cancer cell lines) were selected for analysis. The presence of a CTCF motif was calculated by FIMO (MEME suite) in the sequence extended 100 bp upstream and downstream of the summit of either CTCF (CTCF-only and CTCF&BORIS) or BORIS (BORIS-only) peaks. Each CTCF motif occurrence has a p value < 0.0001. c Upper panel: EMSA with five CTCF&BORIS (blue bracket) and four CTCF-only (red bracket) binding regions. The ~200-bp ³²P-labeled probes were incubated with either in vitro translated luciferase (−) or with the 11 ZF domain of CTCF (+). The slower (shown by arrow with two red dots) and faster (arrow with one red dot) migrating shifted bands correspond to CTCF binding to two CTCF sites at once or to one CTCF site, respectively (double and single occupancy). Lower panel: genome browser view of CTCF and BORIS occupancy in K562 and Delta47 cells at nine DNA sequences used in the EMSA. The brackets show the connection between upper and lower panels. ChIP-seq data are shown in combination with ChIP-Re-ChIP-seq data for K562 and Delta47 cells. ChIP-seq tracks are labeled with the molecule against which antibodies were directed and the cell line used. d First row: individual examples of 1xCTS and 2xCTS, differentially occupied by CTCF and BORIS in BORIS-positive (K562) and BORIS-negative (NHDF) cells. CTCF and BORIS ChIP-seq data are combined with digital genomic footprinting (DGF, ENCODE data) and phastCons46 conservation scores. The core 20-bp CTCF motif is marked by a grey box. Second row: heatmaps show DNaseI cleavage density at thousands of 1xCTSes (single CTCF motif on plus strand) and at hundreds of 2xCTSes (two CTCF motifs separated by a 30–40-bp linker, both on the minus strand). The tag density data were collected within a 300-bp window around the first (left) 20-bp CTCF core motifs (0) under a single CTCF ChIP-seq peak. Third row: average phastCons46 conservation score at 1xCTSes and 2xCTSes, the same genomic regions as in the second row. Fourth row: model of CTCF and BORIS differential occupancy in NHDF and K562 cells. 1xCTSes are occupied by CTCF monomer in both BORIS-positive and BORIS-negative cells, while 2xCTSes are co-occupied by CTCF and BORIS in BORIS-positive cells or by CTCF homodimer in BORIS-negative cells