Extended Data Fig. 10: Chromatin density analysis by ChaiN, RNA-seq data, and a schematic model for topological surveillance of DSB loci.

a, Schematic depiction of ChaiN analysis to quantify chromatin density in 3D-SIM images based on histone H2B–GFP distribution. Reconstructed and aligned 3D-SIM images were used to segment volumes occupied by 53BP1-MDs and subjected to an HMM process to derive seven discrete GFP–H2B chromatin density classes within the segmented region. Class 1 represents chromatin-free interchromatin space, while class 2–7 feature increasing chromatin densities. An equivalent analysis of the whole nucleus serves as a control for global chromatin distributions outside 53BP1-MDs. b, ChaiN analysis in undamaged nuclei in wild-type or RIF1-depleted cells (n = 12 per condition). Median ± 95% CI. *P = 0.0348, 0.0226 (class 2 and 4), NS P = 0.2525, 0.7373, 0.0990, 0.4874, 0.9496 (classes 1, 3, 5–7); two-tailed Student’s t-test. c, A hypothetical model. A DSB triggers accumulation of 53BP1 in the damaged and several neighbouring chromatin nanodomains. Saturation of 53BP1 at chromatin nanodomains prompts recruitment of RIF1 to the boundaries between them. Through functional crosstalk with cohesin, RIF1 locally stabilizes the nanodomain topology into an ordered and circular microdomain, which confines repair factors such as BRCA1 to DSBs and locally concentrates shieldin–CST–POLα to restrain DNA-end resection. Absence of RIF1 leads to topological disorder that results in excessive spreading of BRCA1, inability to concentrate DNA-end protection factors and DSB hyper-resection. d, RNA-seq data for TP53BP1, RIF1 and SHLD1 transcripts per million kilobases in cancerous cells (U2OS, HeLa) and normal cells (IMR90, HBL100). Data were derived from publicly available RNA-seq data in the EMBL-EBI expression atlas (see Methods). Scale bars (a), 5 μm in the whole nucleus and 200 nm in the magnified 53BP1-MD image (right). For detailed image information see Supplementary Table 1.

Source data