Fig. 4

Defining targets of a ZEB1/CHD4/NuRD complex. a Flow chart depicts the criteria for choosing 37 candidate ZEB1/HDAC co-repressor complex target genes from ENCODE ChIP-sequencing data. Briefly, the ENCODE ChIP-seq data for ZEB1 and CHD4 was mined for potential co-binding. The list was refined by various mRNA datasets to identify targets which inversely correlate with ZEB1 expression. Lastly, ZEB1 binding (−/+250 bp) was applied as ZEB1 preferentially binds to this region. Validation of ZEB1/CHD4/NuRD targets was initially determined by comparison of mRNA in b 393P constitutively expressing ZEB1 and c 344SQ inducibly expressing miR200a-b-429; all asterisks indicate statistical significance by t-test (n ≥ 3, *p ≤ 0.005); error bars represent standard error mean. d ChIP-qPCR in the human NSCLC cell line H1299 confirm that ZEB1 and CHD4 co-occupy the promoter of KCNK1, TBC1D2, TBC1D2b and EPS8L2, as compared to non-specific IgG control. e Transient depletion of CHD4 in H1299 decreases CHD4 binding, but increases H3K27 acetylation (f) at the promoters of TBC1D2a, TBC1D2b, and EPS8L2. g CHD4 and (H) H3K27ac ChIP was performed in the cell line H1299 after 24 h treatment with mocetinostat (1 μM). Relative CHD4 binding decreased upon treatment, while H3K27 acetylation increased. j ZEB1 overexpression in H358 significantly enhances both ZEB1 and i CHD4 binding to TBC1D2b and EPS8L2. Enrichment in ChIP experiments are relative to non-specific IgG control at each locus and is presented as the mean +/− standard error mean from three independent experiments; all asterisks indicate statistical significance by t-test (n ≥ 3, *p ≤ 0.05); error bars represent standard error mean