Fig. 3: NPAS4–NuA4-bound sites undergo recurrent DNA breaks in vivo.
From: A NPAS4–NuA4 complex couples synaptic activity to DNA repair
a, A model for the dual function of NPAS4–NuA4 in stimulating transcription and DNA repair in active neurons. b, IGV tracks displaying aggregate signals for ATAC-seq (n = 3), NPAS4 CUT&RUN (n = 5), ARNT2 CUT&RUN (n = 2) and γH2AX ChIP–seq (n = 3) at the activity-inducible gene Bdnf. n = 3–5 mice pooled per replicate. c, Boxplots of average γH2AX ChIP–seq normalized counts (n = 3 pools of 3–5 mice) at all regulatory elements (left) and activity-inducible regulatory elements (right), subset by quartiles of NPAS4 CUT&RUN signal (Methods). All regulatory elements: Q1 = 44,864 sites, Q4 = 7,378 sites. Activity-inducible elements: Q1 = 1,017 sites, Q4 = 764 sites. Boxplots show the median (line), IQR (box) and 1.5× IQR (whiskers), and notches indicate the median ± 1.58× IQR/sqrt(n). ***P < 2.2 × 10−16. P values were calculated using unpaired, two-tailed Wilcoxon rank-sum tests. d, Schematic of sBLISS-seq to map DSBs in brain nuclei. A tissue aliquot from each sBLISS-seq sample was kept for paired RNA-seq analysis. e, Volcano plot depicting the DeSeq2 log2-transformed fold change versus –log10(Benjamini–Hochberg adjusted P value) of sBLISS-seq signals between 0 and 2 h KA stimulation. Dotted line indicates adjusted P < 0.1. Elements with log2-transformed fold change > 0, adjusted P < 0.1 and that are within Q4 of NPAS4 IgG-normalized CUT&RUN signals are indicated in blue. Right inset, IGV tracks of aggregate sBLISS-seq (n = 8 individual mice) and NPAS4 CUT&RUN (n = 5 pools of 3–5 mice) at Cgref1 and Rgs7bp promoters. Coloured bars represent statistically defined peaks. f, Aggregate plots showing sBLISS-seq coverage (fragment depth per bp per peak) at activity-inducible regulatory elements, subset by quartiles of NPAS4 binding. Signals are mean ± s.e.m. (n = 8 individual mice). ***P < 2.2 × 10−16. P values were calculated using the average signal extracted in a 500 bp window around the element centre using unpaired, two-tailed Wilcoxon rank-sum tests.
