Table of Contents
Cell lines (MDA-MB-231, 4T1 and RPE-1) were purchased from the American Type Culture Collection (ATCC). TP53-knockout MCF10A, TP53-knockout RPE-1 and Trex1 knockout 4T1 cells were gifts from the Maciejowski laboratory at the Memorial Sloan Kettering Cancer Center (MSKCC). OVCAR-3 cells were a gift from J. D. Gonzales. All gifted cells had been obtained from the ATCC. Catalogue numbers from the ATCC are as follows: MDA-MB-231, HTB-26; 4T1, CRL-2539; RPE-1, CRL-4000; MCF10A, CRL-103172; and OVCAR-3, HTB-161. MDA-MB-231 and RPE-1 cells were cultured in DMEM (Corning), and 4T1 and OVCAR-3 cells were cultured in RPMI (Corning); both media were supplemented with 10% FBS and 50 U ml−1 penicillin–streptomycin. All cells were periodically tested for mycoplasma contamination.
Immunofluorescence microscopy and histone PTM quantification
Cells were grown on coverslips until 80–90% confluent. Cells were fixed using ice-cold (−20 °C) methanol for 15 min in the freezer. Cells were then permeabilized with 0.5% Triton-X-100 in bovine serum albumin (TBS-BSA) 1% solution for 5 min at room temperature. Subsequently, cells were washed with TBS-BSA (1%) and incubated for 5 min at room temperature. For samples from patients with HGSOC, specimens were collected as previously reported40. All protocols for collection were approved by the institutional review board of MSKCC. Patients gave consent according to the institutional review board-approved standard operating procedures for informed consent. Written informed consent was obtained from all patients before conducting any study-related procedures. The study was conducted in accordance with the Declaration of Helsinki and the good clinical practice guidelines. Antibodies (see Supplementary Table 1 for a complete list and dilutions) were diluted with 1% BSA-TBS solution and incubated at 4 °C overnight. The stain DAPI (0.5 µg ml−1) was then added along with secondary antibodies diluted in TBS-BSA (1%). Cells were mounted with Fluoro-Gel with Tris buffer (Electron Microscopy Sciences). Images were collected as described below (super-resolution immunofluorescence microscopy section). Histone PTM positive/negative quantification was done in triplicates for each histone PTM, with 100 micronuclei counted per replicate. Histone PTM intensity quantification was done using ZEN 2 (blue edition) software, for which the fluorescence intensity of the corresponding PTM is normalized to the fluorescence intensity of the DAPI channel.
Cells were lysed with 1× RIPA buffer (EMD Millipore), kept on ice and vortexed 3 times in 10-min intervals. The lysates were then centrifuged at 10,000 rpm for 10 min and debris was discarded. Protein was quantified using the DC protein assay (Bio-Rad) and equal amounts of protein were loaded in each SDS–PAGE gel well and run at 120 V for 1 h. Proteins were then transferred to a nitrocellulose membrane using the standard protocol in the Bio-Rad Trans-Blot Turbo transfer system. The membrane was blocked using TBS blocking buffer (Li-Cor) incubated with primary antibodies (Supplementary Table 2) overnight at 4 °C. Secondary antibodies were incubated for 1 h at room temperature. Western blot membrane scanning was done using an Odyssey CLx system (Li-Cor). For dot blots, isolated nuclei were lysed with 1× RIPA buffer (EMD Millipore), kept on ice and vortexed 3 times in 10-min intervals. The lysates were then centrifuged at 10,000 rpm for 10 min and debris was discarded. No quantification was done for dot blots; an approximation of a 1:20 ratio between micronuclei and primary nuclei was used instead. Then 2 µl of the RIPA extract of the fraction of either the micronucleus or the primary nucleus was blotted on top of a nitrocellulose membrane and air-dried for at least 1 h. The membrane was then blocked using TBS blocking buffer (Li-Cor) incubated with primary antibodies (Supplementary Table 2) overnight at 4 °C. Secondary antibodies were incubated for 1 h at room temperature. Scanning was done using the Odyssey CLx system (Li-Cor) and quantification was done using ImageStudio software (Li-Cor). Antibodies used for immunoblots are listed in Supplementary Table 2. Full uncropped membranes are shown in Supplementary Fig. 1.
Cells were treated with reversine (Cayman Chemical Company) at a concentration of 0.5 µM for either 24 h (for MDA-MB-231) or 48 h (4T1). The HDAC inhibitor vorinostat (Sigma-Aldrich; also known as SAHA) was used at a concentration of 0.5 µM for 24 h, and the methyltransferase inhibitor GSK-126 (XcessBio; EZH2) was used at a concentration of 5 µM for 5 days. Long-term reversine treatment was done 24 h after cells were passaged, and the medium was replaced with normal cell medium after 48 h. GSK923295 (Selleck Chemicals) was used at the concentration of 50 nM for 24 h. DMSO was used as the vehicle control.
Plasmids and molecular biology
Expression of H2B-mCherry and GFP–cGAS was done using plasmids obtained from the Maciejowski laboratory as previously described20. Puromycin at the concentration of 2 µg ml−1 was used to select for cells expressing both proteins, and subsequently fluorescence-activated cell sorting (FACS) was used to enrich for cells expressing the respective proteins. Lami-B2-mCherry overexpression was achieved using a plasmid obtained from the Hetzer laboratory, as previously described7. Blasticidin was used at the concentration of 10 µg ml−1 to select for cells overexpressing lamin B2. Cells were transduced using a standard lentiviral protocol (H2B-mCherry and GFP–cGAS) or retroviral protocol (lamin B2-mCherry) using HEK-293T cells to produce the lentivirus and retrovirus vectors. After viral transduction, cells were cultured for 24 h before selection was performed.
We plated 500,000 cells for each condition in a 6-well plate containing a coverslip for each well and allowed to adhere overnight. The next day, cells were transfected with 30 pmol of small interfering RNA (siRNA) anti-lamin A (pre-designed siRNA, ThermoFisher) or its selected negative control (ThermoFisher) following the standard protocol for lipofectamine RNAiMAX (ThermoFisher). After 48 h the cells were lysed with RIPA buffer, and the lysate subjected to western blot was analysed using anti-lamin A antibody (Thermo Scientific) to test for efficient knockdown. Coverslips were fixed and stained as previously described in the ‘Immunofluorescence microscopy and histone PTM quantification’ section.
Fluorescence lifetime imaging microscopy
Fluorescence lifetime images were collected with a Zeiss LSM 880 (Carl Zeiss Microscopy) microscope equipped with a Spectra-Physics MaiTai HP laser (Spectra-Physics) for two-photon excitation, tuned at 800 nm, and collected using a Zeiss 63×/1.41 NA oil objective. The fluorescence signal was collected by a photomultiplier tube (H7422P-40, Hamamatsu) and recorded using a FLIMbox (model A320 ISS) to obtain the lifetime information. The pixel dwell time for the acquisitions was 16 μs and the images were taken with sizes of 256 × 256 pixels, accumulating 15 frames. The data from each pixel were recorded and analysed using the SimFCS software (Laboratory for Fluorescence Dynamics, University of California, Irvine, CA) to perform the phasor transformation and with custom MATLAB code to quantify the lifetime. The phasor position was corrected for the instrument response function by acquiring the fluorescence of a solution with a known lifetime (0.4 mM Coumarin 6 dissolved in ethanol, τ = 2.5 ns). The lifetime was determined as τ = s/ωg, where τ is the fluorescence lifetime, g and s are the phasor coordinates and ω is the laser frequency (80 MHz).
Cell labelling and image processing
The samples were labelled with one drop per well of Hoechst 33342 (NucBlue, Thermofisher Scientific) and imaged after 20 min of incubation. For every field of view for fluorescence lifetime imaging microscopy (FLIM), we also recorded a 2-channel fluorescence-intensity image corresponding to the Hoechst 33342 and enhanced GFP emissions to identify the ruptured micronuclei. The four subregions considered (heterochromatin, euchromatin and intact and ruptured micronuclei) were determined manually according to the following criteria. For heterochromatin, in the Hoechst 33342 channel, bright spots in the nucleus correspond to heterochromatin-enriched regions in 4T1 or MDA-MB-231 cells. For euchromatin, in the Hoechst 33342 channel, regions in the nucleus with low Hoechst 33342 signal were classified as euchromatin-enriched regions. For micronuclei, in the Hoechst 33342 channel, micronuclei were identified as nucleic-acid-positive compartments that were smaller than the main nucleus. They were further classified as intact, if there was no signal in the eGFP channel, or ruptured, if there was a positive signal in the eGFP channel. The subregions selected above were used to segment the lifetime image, and a single lifetime value was obtained as the median value of the entire subregion. For every cell, one lifetime value for heterochromatin, euchromatin and micronuclei (either ruptured or intact) was obtained, for a total of 73 cells (31 intact, 42 ruptured micronuclei) over five biological replicates. Outliers were removed by the MATLAB function ‘isoutlier’. Given that FLIM was performed on live cells we postulate that, as well as chromatin compaction states, it would enable the observation of digested DNA that would otherwise be lost during permeabilization used in the other methods (ATAC-see and ATAC-seq). To address this, we repeated all experiments in TREX1-knockout cells as shown in Fig. 3b.
ATAC-see with immunofluorescence microscopy
Tn5 production, transposome assembly and ATAC-see were performed as described28 with some adjustments for subsequent immunofluorescence experiments. In brief, cells were grown on coverslip until 80–90% confluent, and were fixed using 1% paraformaldehyde in PBS at room temperature for 10 min. Permeabilization was done in lysis buffer (10 mM Tris-Cl, pH 7.4, 10 mM NaCl, 3 mM MgCl2, 0.01% Igepal CA-630) for 10 min at room temperature. Coverslips with cells were washed twice with PBS for 5 min each at room temperature. The transposase mixture solution (25 μl of 2× TD buffer, final concentration of 100 nM Tn5-ATTO-59ON and dH2O up to 50 μl) was added on the coverslip and incubated for 30 min at 37 °C in a humid chamber box. Coverslips were then washed three times with PBS containing 0.01% SDS and 50 mM EDTA for 15 min at 55 °C each time. After washing, coverslips were subjected to the immunofluorescence procedure described above, starting from washing with TBS-BSA (1%).
Super-resolution immunofluorescence microscopy
Images were acquired with an inverted Zeiss LSM880 (Carl Zeiss Microscopy) equipped with an Airy Scan detector (gain 850, digital gain 1) in super-resolution mode. The 848 × 848 pixels images were acquired with a Plan-Apochromat 63×/1.4 NA oil objective using a pixel size of 40 nm and a dwell time of 4.96 μs. The optical configuration used split the laser lines (405 nm at 1% power, 488 nm at 3% and 561 nm at 7%) with a 488/561/633 MBS and an invisible light beam splitter at 405. For the detection we applied a bandpass filter at 495–550 nm and a longpass filter at 570 nm. Each image is acquired as a z stack of 5 stacks with an interval of 500 nm; a maximum-intensity projection and Airy Scan processing with default parameters were performed for all images before quantification using the Zeiss ZEN black software. For the ATAC-see images, only two laser lines were used, 405 nm and 594 nm, at 1% and 30% power, respectively, using the optical configuration described above. The number of micronuclei number was scored in 10 different sections per coverslip (3 coverslips accounting for 3 biological replicates) at a magnification of 63×. Missegregation scoring was done to at least 50 cells undergoing anaphase per biological replicate (n = 3).
Purification of micronuclei
Micronucleus purification was performed as described previously41, except that the primary nucleus fraction was also collected. In brief, cells were expanded in 245 × 245 × 25 mm dishes and treated with reversine as specified above. Cells were then collected and washed in DMEM. Washed cells were resuspended in prewarmed (37 °C) DMEM supplemented with 10 μg ml−1 cytochalasin B (Sigma-Aldrich) at a concentration of around 5 × 106 cells per ml DMEM and incubated at 37 °C for at least 30 min. Subsequently, cells were centrifuged at 300g for 5 min and resuspended in cold lysis buffer (10 mM Tris-HCl, 2 mM Mg-acetate, 3 mM CaCl2, 0.32 M sucrose, 0.1 mM EDTA, 0.1% (v/v) NP-40, pH 8.5) at a concentration of 107 cells per ml lysis buffer, which was freshly complemented with 1 mM dithiothreitol, 0.15 mM spermine, 0.75 mM spermidine, 10 μg ml−1 cytochalasin B and protease inhibitors. Resuspended cells were then dounce-homogenized 10 times using a loose-fitting pestle. Then, lysates were mixed with an equal volume of cold 1.8 M sucrose buffer (10 mM Tris-HCl, 1.8 M sucrose, 5 mM Mg-acetate, 0.1 mM EDTA, pH 8.0) freshly complemented with 1 mM dithiothreitol, 0.3% BSA, 0.15 mM spermine and 0.75 mM spermidine before use. In a 50-ml conical tube, 10 ml of the mixture was then layered on top of a two-layer sucrose gradient (20 ml of 1.8 M sucrose buffer on top of 15 ml 1.6 M sucrose buffer). This mixture was then centrifuged in a swing bucket centrifuge at 950g for 20 min at 4 °C. The first resulting 2-ml top fraction is discarded; the next 5–6 ml, mostly containing micronuclei, was collected, and the subsequent 3 ml, mostly containing primary nuclei, was collected in a separate container. Collected fractions were diluted 1:5 with FACS buffer (ice-cold PBS freshly supplemented with 0.3% BSA, 0.1% NP-40 and protease inhibitors). Diluted micronuclei were then centrifuged at 950g in a JS-5.2 swing-bucket centrifuge for 20 min at 4 °C. The resulting supernatant was removed by aspiration and either micronuclei or primary nuclei were resuspended in 2–4 ml FACS buffer supplemented with 2 μg ml−1 DAPI (but no DAPI was used for micronuclei purification for ATAC-seq and CUT&RUN experiments). Resuspended samples were filtered through a 40-μm ministrainer (PluriSelect) into FACS tubes. Micronuclei were then sorted by FACSAria (BD Biosciences) into FACS buffer at the MSKCC Flow Cytometry Core Facility. Default FSC and DAPI thresholds were lowered, and a log scale was used to visualize the micronuclei. For ATAC-seq micronuclei purification, FSC and mCherry were used instead of FSC and DAPI. Sorted micronuclei were centrifuged at 4,000g in a swing-bucket rotor for 20 min at 4 °C and the pellets were stored at −80 °C for later use or processed immediately or ATAC-seq or CUT&RUN experiments.
Viably frozen cells were thawed, pelleted and incubated for at least 30 min in 360 μl buffer ATL + 40 μl proteinase K at 55 °C. DNA isolation proceeded with the DNeasy Blood & Tissue Kit (Qiagen) according to the manufacturer’s protocol modified by replacing AW2 buffer with 80% ethanol. DNA was eluted in 60 µl 0.5× buffer AE heated to 55 °C. For nuclei, DNA from nuclei and micronuclei was isolated with the DNeasy Blood & Tissue Kit (Qiagen) according to the manufacturer’s protocol, modified by replacing AW2 buffer with 80% ethanol. DNA was eluted in 60 µl 0.5 × buffer AE heated to 55 °C.
After PicoGreen quantification and quality control by Agilent TapeStation, 97–165 ng genomic or micronuclear DNA was sheared using a LE220-plus focused ultrasonicator (Covaris) and sequencing libraries were prepared using the KAPA Hyper Prep Kit (Kapa Biosystems) with modifications. In brief, libraries were subjected to a 0.5 × size selection using aMPure XP beads (Beckman Coulter) after post-ligation clean-up. Libraries were amplified with 5 cycles of PCR and pooled at equimolar ratios. Samples were run on a NovaSeq 6000 in a PE150 run, using the NovaSeq 6000 S4 Reagent Kit (300 cycles) (Illumina). The average coverage per sample was 18×. For micronuclei samples, copy-number analysis was done using HMMcopy42 v.1.38 with the MoCaSeq pipeline43, setting C57BL-6NJ 6NJ44 as the normal pair for all three. Plots for tumour/normal relative copy number ratios (Extended Data Fig. 6g) were also made using MoCaSeq, using its CNV_PlotHMMcopy R script. For the copy-number comparisons with ATAC-seq counts of micronucleus and primary nucleus samples (Extended Data Fig. 6i), reads were mapped to the mm10 genome using BWA-MEM (default settings), and duplicate marking and sorting were done using NovoSort MarkDuplicates (v. 3.08.02), a multi-threaded bam sort/merge tool by Novocraft Technologies. Counts were obtained for 5-kb bins using bedtools (v. 2.25.0) for ATAC-seq and WGS mapped reads. For Fig. 3f, hierarchical clustering was done using the pheatmap package in R. Copy-number profiles were projected onto the clusters by intersecting ATAC-seq peaks with the surrounding 1-kb window of the WGS data. For Extended Data Fig. 6j, copy-number box plots were created by cutting the row dendrogram of the ATAC-seq heat map to produce micronuclei up and down groups and then plotting the corresponding 1-kb windows from the WGS copy-number ratios. For RPE-1 samples, the copy-number variation (CNV; GATK4) pipeline from Basepair was used to determine the copy number.
CUT&RUN of cell lines
CUTANA CUT&RUN was performed with RPE-1 and DLD-1 samples on an automated protocol (autoCUT&RUN) derived from those previously described45,46. In brief, for each CUT&RUN reaction, 500,000 cells (5 million cells per ml prepared in 20 mM HEPES, pH 7.5, 150 mM NaCl, 0.5 mM spermidine, 1 mM trichostatin A, 1× EDTA/EGTA-free complete protease inhibitor) were dispensed to individual wells of a 96-well plate, immobilized onto concanavalin-A beads (EpiCypher) and incubated overnight (4 °C) with 0.5 µg antibody (IgG, H3K4me3, H3K27me3, H3K27ac) (all antibodies validated to histone PTM-defined SNAP-ChIP nucleosome standards). pAG-micronuclease (EpiCypher) was added and activated (2 h at 4 °C), and CUT&RUN-enriched DNA was purified using Serapure beads after mixing at a 2:1 (bead:DNA) ratio. Recovered DNA was quantified using PicoGreen and reactions were normalized to 5 ng DNA (or the entirety of the reaction if less than 5 ng DNA was recovered) before preparing sequencing libraries (CUTANA CUT&RUN Library Prep kit; EpiCypher). All CUT&RUN steps were optimized and performed on Tecan Freedom EVO robotics platforms with gentle rocking for incubation steps and magnetic capture for medium exchange and washing steps as described previously45,46.
CUT&RUN of micronuclei and primary nuclei
CUTANA CUT&RUN was performed with isolated micronuclei and primary nuclei samples using a CUTANA ChIC/CUT&RUN kit (EpiCypher, 14-1048), following the manufacturer’s manual for nuclei samples. Around 50,000 primary nuclei, 1,000,000 intact micronuclei and 100,000 ruptured micronuclei were subjected to CUT&RUN for each replicate. In brief, nuclei were immobilized onto concanavalin-A beads from the kit and incubated overnight (4 °C) with 0.5 µg of antibody (IgG, H3K4me3). pAG-micronuclease (EpiCypher) was added and activated (2 h at 4 °C), and CUT&RUN-enriched DNA was purified using the kit’s purification protocol and reagents. The reaction was used to prepare sequencing libraries using the protocol from the CUTANA CUT&RUN Library Prep kit (EpiCypher).
RNA was extracted using the manufacturer’s protocols in the Qiagen RNeasy mini plus kit (Qiagen). Library preparations are TruSeq stranded and sequenced using the poly(A) RNA-sequencing workflow, with a read depth of 80–100 million, 100-base-pair (bp) paired-end reads. Library preparation and sequencing were performed in MSKCC Integrated Genomics Core. RNA-seq data were analysed with a Basepair toolkit using the expression count (STAR) workflow to obtain values of fragments per kilobase of transcript per million mapped reads. Principal component analysis was done using randomized principal component analysis applied to the normalized count matrix of all genes (Extended Data Fig. 8g) or a list of oncogenes and tumour-suppressor genes (Extended Data Fig. 9c) using Partek Flow software, version 10.0 (Partek). Principal component analysis of TP53-knockout hTERT RPE-1 cells (Extended Data Fig. 9c) was done using the prcomp function in the stats package (v. 4.2.1) in R (v.4.2.1) and RStudio (version 2022.07.0+548). A GSEA was performed on normalized gene counts from control and lamin-B2-overexpressing TP53-knockout hTERT RPE-1 cells treated with reversine or DMSO (Extended Data Fig. 9a,b) using the GSEA software version 4.2.3 (The Broad Institute), with the following parameters: metric for ranking genes, Signal2Noise; weighted enrichment statistics; permutation type, gene set. The gene sets used for GSEA were derived from the ATAC-seq analysis of chromatin accessibility in intact or ruptured micronuclei in 4T1 (Fig. 3f). Mouse gene symbols were converted to human gene symbols using the Ensembl biomaRt package (version 2.48.3). Expression datasets from the TCGA breast-cancer cohort were accessed through cbioportal47 (https://www.cbioportal.org/study/summary?id=brca_tcga_pan_can_atlas_2018). GSEA comparing tumours belonging to the top or bottom tertile of the fraction of genome altered, as reported by the TCGA, was performed as described above, and using the same parameters as above (Fig. 3f). The log2-transformed fold change was computed using the DESeq2 package (v. 1.36.0), and KEGG pathways (Extended Data Fig. 9d) were generated using the pathview package (version 1.36.0) in Bioconductor (v. 3.15).
The ATAC-seq procedure on cells was done as described previously48. The ATAC-seq procedure on nuclei samples was done immediately after isolation using the same protocol but without the lysis step. The ratio of primary nuclei to micronuclei was 1:20. Libraries were sequenced with 50-bp paired-end reads to approximately 50 million reads per sample for cells, whereas for nuclei samples, sequencing was done with 50-bp paired-end reads to 350 million total reads for a triplicate of each sample at the Memorial Sloan Kettering Cancer Center Integrated Genomics Core. For chromosome-4-missegregation ATAC-seq experiments in DLD-1 and RPE-1 cells, the raw fastq data were analysed using the Basepair toolkit. The raw reads were trimmed using fastp to remove low-quality bases from reads (quality <20) and adapter sequences. The trimmed reads were aligned using Bowtie2 as described49 to the UCSC genome assembly (hg38). For long-term dnMCAK and micronuclei ATAC-seq experiments in RPE-1 and 4T1 cells, reads were mapped to the hg38 (RPE-1) or mm10 (4T1) genome assembly using Bowtie2 with the following parameters: -X2000 –no-mixed –no-discordant and reads were filtered using samtools for the 1804 FLAG and mapq score of 30. PCR duplicates were identified using the MarkDuplicates function from Picard tools. Peaks were called per sample using macs2 with the following parameters: callpeak -f BAM -g hs –nomodel –shift 37 –extsize 73 –keep-dup all -B –SPMR –call-summits -q 1e-2. Final peaks across each dataset were determined based on an extension of 250 bp from the summit of each peak and merged across all samples using bedtools. Normalized bigwig files were generated using the bamCoverage function from deeptools using reads per kilobase of transcript per million mapped reads normalization. Heat maps for genomic regions were generated using the deeptools computeMatrix scale-region and plotHeatmap functions. Differential peak calling was performed using DESeq2 on raw counts for the merged peaks (an adjusted P-value cut-off of 0.01 and log2-transformed fold change of 1.5 was applied to the mouse 4T1 cell lines and an adjusted P-value of 0.05 and log2-transformed fold change of 0.75 was applied to the RPE-1 cell line). The normalized counts for peaks were used to plot the heat maps of differential peaks. Pie charts were generated using the plotAnnoPie function in the ChIPseeker R package. Pathway enrichment was performed on genes associated with differential peaks using the enrichGO function in the Bioconductor package clusterProfiler and similar pathways were merged in R using the ‘simplify’ function with a similarity cut-off of 0.7.
Micronuclei ATAC-seq and H3K4me3 CUT&RUN analyses
The k-means clustering of ATAC and H3K4me3 CUT&RUN data was performed on normalized count matrices with increasing k until the clusters became redundant. Hierarchical clustering was done using the pheatmap package in R. Peak annotation pie charts were created in R with the annotatePeak and plotAnnoPie functions from the package ChIPseeker. High-, medium- and low-expression groups were based on normalized expression from RNA-seq results in 4T1 dnMCAK cells and divided using the top, middle and bottom third of TPM (transcript per million) values. Peaks were annotated by gene and assigned to the expression group of the gene. ATAC and H3K4me3 CUT&RUN pie charts were based on differential peak overlaps obtained using bedtools intersect. Scatterplots were created using the normalized ATAC or H4K4me3 signal at ±500 bp surrounding a TSS and plotting the log2-transformed fold change of the averaged micronucleus/primary nucleus signal against the normalized gene expression of the matching gene.
ATAC-seq and CUT&RUN normalization for DLD-1 cells
CNV calling from WGS data
WGS CRAM files were obtained from EGA (dataset ID EGAD00001004163) and unmapped from GRCh37 using samtools (version 1.3.1), and then aligned to GRCh38 using BWA-MEM (version 0.7.17-r1188). Copy-number profiles were called using Python software using the CNVKit library (version 0.9.7). Integer copy numbers were obtained using the CNVKit ‘call’ command, and non-integer copy numbers were obtained from the log2-transformed copy-number ratios reported in the CNVKit .cns files j.
ATAC-seq and CUT&RUN alignments were filtered from the BAM files to remove PCR duplicates, reads aligning with mitochondrial DNA, and alignments with low mapping quality (MAPQ score <15)
We used the ATAC-seq signal not to identify peaks corresponding to active transcriptional sites but to measure the total accessibility of DNA within a region. We generated ATAC-seq read counts for 10,000-bp genomic windows using samtools (version 1.13). We then calculated the signal values for each window by normalizing the read counts by copy number, chromosome length and total ATAC-seq reads across all chromosomes (Supplementary Fig. 5). Windows overlapping with unmappable regions (using the AmpliconSuite-pipeline genome annotation files), or with a copy number close to 0 (<0.3), were excluded from further analyses. The signal calculation for CUT&RUN data was done in the same way. The normalized signal of a clone c in a given window i with total (ATAC-seq or CUT&RUN) reads T(i) and copy number C(i) was calculated as
where R is the total number of ATAC-seq reads mapped across all chromosomes in clone c; L denotes the sum of ‘effective lengths’ of all the chromosomes. The effective length l of a chromosome is calculated as follows. Let P denote the expected ploidy (P = 1 for the Y chromosome; P = 2 otherwise) of the chromosome, and l(i),C(i) denote the length (in bp) and average copy number, respectively, of window i in the chromosome. Then,
In the case of the Y chromosome, we used only the region up to 26.68 Mbp for the above calculations because the Y chromosome did not have any confidently aligned ATAC reads beyond 26.68 Mbp until the telomeric region in any of the clones (including the parental clone).
Differential analysis of Y-chromosome accessibility
The normalized ATAC signals of a clone can be compared to the parental signals by taking the log-transformed signal fold changes with respect to the parental signal in the same window.
The signal fold change (F) of a clone c in a given window I is given by:
where p is the parental clone.
Intra-clone variability and change in accessibility
We computed the mean and variance of log-transformed signal fold changes for each chromosome in a single clone to understand the global changes in accessibility (Supplementary Fig. 5). To control for high fold changes caused by error in windows with low read counts, windows in which neither the clone nor the parent has more than 200 reads were excluded from this calculation.
To understand whether the Y-chromosome accessibility changes were similar for different clones, we calculated the variance of log-transformed signal fold changes of different clones in the same genomic window. Windows where neither the clone nor the parent has more than 200 reads were excluded. Moreover, only clones with a copy number >0.3 in the window were included for the variance calculation.
Structural variations and change in accessibility
We used manta (version 1.6.0) to call somatic variants in all the clones using a tumour/normal approach with the parental clone as a normal sample. We then checked for any correlation between the density of variant breakends and the variance of log-transformed signal fold changes. We also compared the signal fold change variance in just the regions containing breakends with the variance across all regions to determine whether the change in accessibility is limited to regions with structural variations.
Immunofluorescence staining of human samples
The immunofluorescence detection of cGAS, H3K27me3 and H3K27ac were performed at the Molecular Cytology Core Facility of MSKCC using the Discovery Ultra processor (Ventana Medical Systems, Roche). After 32 min of heat (95 °C) and treatment with Cell Conditioning 1 solution (Ventana), the tissue sections were blocked first for 30 min in Background Blocking reagent (Innovex). A mouse monoclonal cGAS antibody (LSBio) was used at a dilution of 1:200. The primary antibody was incubated for 5 h, after which the biotinylated anti-mouse secondary (Vector Labs) was incubated at a concentration of 5.75 µg ml−1. Blocker D, streptavidin–HRP and TSA Alexa 488 (Life Tech) were diluted according to the manufacturer’s instructions at 1:150 and incubated for 16 min. A mouse monoclonal antibody for H3K27me3 (Active Motif) was used at a dilution of 1:500. The primary antibody was incubated for 6 h, after which the biotinylated anti-mouse secondary (Vector Labs) was incubated at a concentration of 5.75 µg ml−1. Blocker D, streptavidin–HRP and Tyramide-CF594 (Biotium) were diluted according to the manufacturer’s instructions at 1:2,000 for 16 min. A rabbit polyclonal H3K27ac antibody (Abcam) was used at a concentration of 0.05 µg ml−1. The primary antibody was incubated for 6 h, after which the biotinylated goat anti-rabbit IgG (Vector labs) was incubated for 60 min at a concentration of 5.75 µg ml−1. Blocker D, streptavidin–HRP and TSA Alexa 647 (Life Tech) were diluted according to the manufacturer’s instructions at 1:150 and incubated for 16 min. All slides were counterstained in 5 µg ml−1 DAPI (Sigma D9542) for 5 min at room temperature, mounted with anti-fade mounting medium Mowiol [Mowiol 4–88 (CALBIOCHEM)] and a coverslip was added. The 956 × 956 pixels z stack images were acquired with an inverted Zeiss LSM880, using the optical configuration described above. The acquisition was performed with a pixel size and dwell time of 40 nm and 4.39 μs, respectively, and using the following lasers and relative powers: 405 nm at 0.4%, 488 nm at 3%, 561 nm at 0.5% and 647 nm at 0.5%. The images were processed as described above, using the Zeiss ZEN black software.
5-ethynyluridine-marked transcript detection and immunofluorescence
To detect nascent transcripts, cells were incubated for 30 min with 1 mM 5-ethynyluridine. Incorporation of 5-ethynyluridine was detected using the Click-iT RNA Alexa Fluor 594 Imaging Kit according to the manufacturer’s instructions (ThermoFisher). Samples were then incubated overnight in antibody. Sample staining and quantification was then performed using the immunofluorescence microscopy and histone PTM quantification methods described above with the following changes: paired measurements were performed for each micronucleus and its primary nucleus. For all measurements a z-score was calculated and experimental replicates were combined. Samples with a positive z-score were defined as having H3K27ac-positive staining.
For immunofluorescence combined with DNA FISH, the immunofluorescence procedure was performed as described above followed by fixation in 1% paraformaldehyde for 15 min and subsequently in Carnoy’s fixative at −20 °C for 15 min. Chromosome paint probes were purchased from MetaSystems. To perform FISH on interphase cells, unsynchronized cells cultured in chamber slides were washed with PBS and fixed with Carnoy’s fixative for 15 min at room temperature. The slides were then dehydrated in sequential washes in 70% ethanol, 85% ethanol and 100% ethanol and then air-dried for 2 min. FISH probes were applied to interphase cells on glass slides and sealed with a coverslip. Samples and probes were co-denatured at 75 °C for 2 min, followed by sealing with rubber cement and overnight hybridization at 37 °C in a humidified chamber. Slides were washed in 2× SSC three times for 30 s. Slides were rinsed in PBS, stained with DAPI, dehydrated in sequential washes in 70% ethanol, 85% ethanol and 100% ethanol and mounted in ProLong Gold anti-fade mounting solution.
Live cell imaging
On the day before imaging, 2,500 cells were plated in each well of an Ibidi 8-well plate chamber (µ-Slide 8 Well high Glass Bottom, Ibidi). After 24 h, cells were washed twice with PBS and stained with NucSpot Live 488 (Biotium) or with NucRed Live 647 (Invitrogen) dyes, following the manufacturer’s instructions. When using tNucSpot Live 488, the dye was incubated with the cells in the imaging solution (Live Cell Imaging Solution (Invitrogen) supplemented with 10% FBS). In the case of NucRed Live 647, cells were washed 4 times with Live Cell Imaging Solution and then imaged in Live Cell Imaging Solution supplemented with 10% FBS. Cells were imaged on an Axio Observer Zeiss wide-field microscope equipped with a Plan-Apochromat 20x/0.8 M27 air objective, incubation chamber with controlled temperature, CO2 and humidity, and a Hamamatsu camera. Images were acquired every 10 min in the 488 and 647 channels, with the light source X-Cyte lamp at 20% power for 48 h.
Chromosome 4 missegregation system
RPE-1 cells were transfected with GFP–Cas9 (IDT) using the target sequence TTTAGTGCCCGGCCGCAAGG using the IDT Alt-R CRISPR–Cas9 sgRNA, 2 nmol platform final sequence: mU*mU*mU* rArGrU rGrCrC rCrGrG rCrCrG rCrArA rGrGrG rUrUrU rUrArG rArGrC rUrArG rArArA rUrArG rCrArA rGrUrU rArArA rArUrA rArGrG rCrUrA rGrUrC rCrGrU rUrArU rCrArA rCrUrU rGrArA rArArA rGrUrG rGrCrA rCrCrG rArGrU rCrGrG rUrGrC mU*mU*mU* rU. We used Lipofectamine 3000 (ThermoFisher) and followed the manufacturer’s protocol. After 48 h, single cells were sorted into individual wells and clones were validated using chromosome 4 paint (Metasystems) according to the manufacturer’s protocol. Clones were screened by DAPI staining to evaluate for bridge formation and micronucleation rate. A subset of clones underwent secondary validation using chromosome 4 paint (Metasystems) according to the manufacturer’s protocol. Validated clones were then processed for DNA-seq, RNA-seq and ATAC-seq as mentioned above.
Histone extraction and digestion for mass spectrometry
Histone proteins were extracted from the cell pellet as described50 to ensure good-quality identification and quantification of single histone marks. In brief, histones were acid-extracted with chilled 0.2 N sulfuric acid (5:1, sulfuric acid:pellet) and incubated with constant rotation for 4 h at 4 °C, followed by precipitation with 33% trichloroacetic acid overnight at 4 °C. The supernatant was then removed and tubes were rinsed with ice-cold acetone containing 0.1% HCl, centrifuged and rinsed again using 100% ice-cold acetone. After the final centrifugation, the supernatant was discarded and the pellet was dried using a vacuum centrifuge. The pellet was next dissolved in 50 mM ammonium bicarbonate, pH 8.0, and histones were subjected to derivatization as described previously50. In brief, 5 µl propionic anhydride and 14 µl ammonium hydroxide (all Sigma Aldrich) were added to samples to balance the pH at 8.0. The mixture was incubated for 15 min and the procedure was repeated. Histones were then digested with 1 µg of sequencing-grade trypsin (Promega) diluted in 50 mM ammonium bicarbonate (1:20, enzyme:sample) overnight at room temperature. The derivatization reaction was repeated to derivatize peptide N termini. The samples were finally dried in a vacuum centrifuge. Before mass-spectrometry analysis, samples were desalted using a 96-well plate filter (Orochem) packed with 1 mg Oasis HLB C-18 resin (Waters). The samples were next resuspended in 100 µl of 0.1% trifluoroacetic acid (TFA) and loaded onto the HLB resin, which was previously equilibrated using 100 µl of the same buffer. After washing with 100 µl of 0.1% TFA, the samples were eluted with 70 µl of buffer containing 60% acetonitrile and 0.1% TFA and then dried in a vacuum centrifuge.
Liquid chromatography tandem mass spectrometry acquisition
Samples were resuspended in 10 µl of 0.1% TFA and loaded onto a Dionex RSLC Ultimate 300 (Thermo Scientific), coupled online with an Orbitrap Fusion Lumos (Thermo Scientific). Chromatographic separation was performed with a two-column system, consisting of a C-18 trap cartridge (300 µm inner dimension, 5 mm length) and a picofrit analytical column (75 µm inner dimension, 25 cm length) packed in-house with reversed-phase Repro-Sil Pur C18-AQ 3 µm resin. Histone peptides were separated using a 30 min gradient from 1–30% buffer B (buffer A, 0.1% formic acid; buffer B, 80% acetonitrile + 0.1% formic acid) at a flow rate of 300 nl min−1. The mass spectrometer was set to acquire spectra in a data-independent acquisition mode. Briefly, the full MS scan was set to 300–1,100 m/z in the orbitrap with a resolution of 120,000 (at 200 m/z) and an AGC target of 5 × 10−5. Tandem mass spectrometry (MS/MS) was performed in the orbitrap with sequential isolation windows of 50 m/z with an AGC target of 2 × 10−5 and a higher-energy C-trap dissociation collision energy of 30. Histone peptide raw files were imported into EpiProfile 2.0 software51. From the extracted ion chromatogram, the area under the curve was obtained and used to estimate the abundance of each peptide. To achieve the relative abundance of PTMs, the sum of all different modified forms of a histone peptide was considered as 100% and the peak area of the particular peptide was divided by the total area for that histone peptide in all its modified forms. The relative ratio of two isobaric forms was estimated by averaging the ratio for each fragment ion with different mass between the two species. The resulting peptide lists generated by EpiProfile were exported to Microsoft Excel and further processed for a detailed analysis, as described below. Histone PTM quantification is done automatically using EpiProfile 2.0, a software specifically designed to extract histone peptide signals in mass-spectrometry raw files51. The abundance of (un)modified histone peptides is extracted by performing extracted ion chromatography, that is, an area under the curve of signals corresponding to the peptide calculated mass. To convert raw peak area into relative peptide abundance, the sum of the unmodified and all modified forms for a given histone peptide, that is, all forms with the same underlying sequence, is considered as the total. Next, the area of each peptide is divided by this total to obtain a normalized relative abundance in percentage. In the case of isobaric peptides, that is, peptides with the same intact mass (such as H3K18ac and H3K23ac), the MS/MS fragment ions are used first to calculate the ratio between the two species and split the peak area before the percentage calculation. Finally, the total abundance of a given PTM is obtained by summing the relative abundance of the peptides that contain that modification. For instance, the final relative abundance of H3K9me3 (occurring on the peptide amino acids 9–17) is obtained by summing the relative abundances of peptides modified as H3K9me3, K3K9me3K14ac, H3K9me3S10 and H3K9me3S10pK14ac. It is important to note that changes in histone PTMs in micronuclei could not be attributed to differences in the total levels of histone H3 between primary nuclei and micronuclei (LC-MS/MS AUC (liquid chromatography-mass spectrometry/mass spectrometry area under the curve) ± s.e.m. = 3.05 × 1010 ± 8.85 × 109 and 3.39 × 1010 ± 1.36 × 1010 for primary nuclei and micronuclei, respectively, P = 0.85, two-sided t-test) or specific degradation of the histone tail, as some of the PTMs were differentially enriched on the same lysine residue (for example, loss of acetylation and gain of trimethylation on H3K27).
Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.