Fig. 2: ZEB1 enhances ERα recruitment at common binding sites.

a ChIP-qPCR of ERα on the GREB1 (+5 kb) and TFF1 (+0.5 kb) binding sites in MCF7-V-ZEB1 cells. ERα ChIP values were normalized to a non-binding region and the input. Recruitment was compared to −DOX with the graph showing the means ± SEM of n = 6 biologically independent experiments. b, c Volcano plots of ERα ChIP-seq with wild-type MCF7-V (data from our previously published data set³⁷) and MCF7-V-ZEB1 cells showing the FDR values as a function of the fold-changes of the normalized ERα values of MCF7-V-ZEB1 cells (n = 4 biologically independent experiments) compared to the ERα peaks of wild-type MCF7-V cells (n = 2 biologically independent experiments) treated with E2 (b) or FI (c). d, e Functional annotations for the biological functions of E2– or FI-only ERα binding sites, respectively, using GREAT. f Table summarizing the number of ERα (ESR1), ZEB1, and ESR1-ZEB1 shared motifs in the ERα ChIP-seq, as found with FIMO (FDR < 0.05; for p values, see Supplementary Data 2). g, h ChIP-qPCR of candidate ERα binding sites from top hits of the ChIP-seq data for E2 (g) and FI (h). g n = 4 biologically independent experiments each including n = 2 technical replicates for TBX2, and n = 3 biologically independent experiments for ANXA3. In panel h, n = 3 biologically independent experiments. i ZEB1 ChIP-qPCR with MCF7-V-ZEB1 cells with (+DOX) or without (−DOX) ZEB1 (n = 3 biologically independent experiments). j Venn diagram shows the intersections between ZEB1-binding sites and E2– or FI-induced ERα binding sites from the ChIP-seq data of MCF7-V-ZEB1 (+DOX) and MCF7-V cells, respectively. k Genome browser views of ZEB1 and ERα binding sites adjacent to the ERα target genes GREB1, TFF1, XBP1, and CDH1. Highlighted sites were analyzed by re-ChIP-qPCR (see next panel). l Re-ChIP experiment showing that ZEB1 and ERα co-occupy the indicated shared binding sites (n = 3 biologically independent experiments). The GREB1 (–20 kb) site is a negative control site as highlighted in k. veh, E2, and FI stand for vehicle, 17β-estradiol, and forskolin + IBMX, respectively. Error bars represent the standard errors of the means; p values are indicated above the bars. Statistical significance was determined with a two-way ANOVA. Source data are provided as a Source Data file and Supplementary Data 1 and 2.