Fig. 2

Functional characterization of the RUNX1 intronic silencer. a Top, DHS fragments (~ 1 kb) were cloned upstream of P2 into pNL1.1. The resultant constructs were co-transfected with pGL4.54 into K562 cells. Results were compared to the pNL1.1-P2 control. Bottom, ENCODE ChIP-seq data at the repressive DHS in K562 cells, obtained from the UCSC genome browser. b Various deletion/mutant constructs were co-transfected with pGL4.54 into the cells. Blue and red lines represent the GFI1/GFI1B and SNAI1 motifs, respectively. Numbers indicate the genomic positions. Data are presented as in panel a. c The 392-bp silencer was cloned into pNL3.1 and the construct was co-transfected with pGL4.54 and pCI vectors expressing different transcription factors (TF) into HeLa cells. Parallel experiments using empty pNL3.1 were performed for each TF group. Results were compared to the empty pCI group (EV). d ChIP-qPCR analysis of LSD1 binding to the silencer. Results were compared to flanking intron 1 regions (20 kb from the silencer) as well as exon 5 (Body) of RUNX1. In panels a, c and d, data were analyzed by one-way ANOVA followed by Dunn’s test. *, ** and *** indicate p < 0.05, P < 0.01 and P < 0.001, respectively. e 3C analysis of chromatin interactions between P2 and the silencer. A physical map of the 10 analyzed EcoRI sites is also shown. The open box represents the silencer. In panels a-e, data are expressed as mean ± SE from three independent experiments. f CRISPR/Cas9 disruption of the silencer. Top, a representative cell population (DEL) showing biallelic deletion of the silencer. WT represents the wild-type genotype. The deletion was verified by Sanger sequencing. The two guide RNA (blue) flanking the target region is shown and the PAM sites are underlined. Bottom, RUNX1 P2 (RUNX1b/RUNX1a) and P1 (RUNX1c) transcript expression in cell populations with (DEL) (n = 6) or without (WT) (n = 6) deletion of the silencer. Data were analyzed by the Mann-Whitney test