Fig. 5.

Ars2 broadly associates with enhancers selectively in NSCs. (A) Ars2 ChIP-seq data were collected from self-renewing neurospheres (nsph), a model for NSCs, and following their differentiation (diff. nsph). Libraries were made ±RNase treatment. Representative data are shown at Sox2, a previously characterized Ars2 target. Its 5′ UTR/3′ UTR peaks are RNA dependent, whereas its enhancer peaks are RNase independent. Furthermore, limited binding is observed in ChIP-seq from differentiated neurospheres. Arrows indicate Ars2 peaks located on Sox2 5′UTR and 3′UTR, which disappear after RNase treatment. (B) Venn diagram summarizes Ars2 peak overlaps in NSCs without RNase treatment, NSCs with RNase treatment and differentiated neurospheres. (C) Classification of overlaps between Ars2 ChIP ±RNase treatment by ChIP peak strength. The overlaps are greater amongst highest-bound peaks. (D) Ars2 peaks near transcription start sites (TSS) were reduced after RNase treatment. (E) ChIP-qPCR validation across a range of Ars2 associations in NSCs. HDAC6 and Bmi1 had minimal binding in ChIP-seq data, serving as negative controls. We observed general concordance between ChIP-seq and ChIP-qPCR data. Data are mean±s.e.m. (F) Representative NSC genes (e.g. Sox2, Sox21 and Egfr) were validated by ChIP-qPCR using mouse tissue. Compared with ChIP from adult cortex, strong enrichment was observed in NSC-containing tissues. Data are mean±s.e.m. (G) The vast majority of Ars2 common peaks locate in intronic (47.6%) and intergenic (49.1%) regions. (H) Ars2 peaks are locally conserved. The y-axis shows PhastCons values aggregated for 4428 common NSC Ars2 peaks. (I) Venn diagram shows the genome-wide overlaps between 4428 common Ars2 and active enhancer marker H3K27Ac in NSCs.