Cell culture and transfections

For details of cell lines used, see Supplementary Table 5. CSB−/− HAP1 cells complemented with doxycycline-inducible expression constructs were established by first transducing CSB−/− HAP1 cells with pT2A-hygro-containing lentivirus, which bears the TET3G-T2A element (VectorBuilder). After hygromycin-B selection (200 µg ml−1; Thermo Scientific), single cell-derived clones were established and validated. One clone was transduced with lentiviral particles containing pTREG3-HA-EV-puro, pTREG3-HA-CSB-puro (all synthesized by VectorBuilder) or pTREG3-HA-CSBK538R-puro. Cells were selected with 1 µg ml−1 puromycin and expression of proteins was induced using 1 µg ml−1 doxycycline for 48 h. Generation of CSB−/− RPE1 cells expressing GFP–EV, GFP–CSBWT and GFP–CSBK538R was performed using lentiviral delivery and 3 µg ml−1 puromycin selection, using pGFP-lentiPuro constructs generated by VectorBuilder.

RPE1 SPRTN-ΔC and RPE1 CSB−/−/SPRTN-ΔC double mutant genome-edited cell lines were generated from RPE1-TetOn-Cas9-PuroS-TP53−/− and -TP53−/−/CSB−/− (referred to as RPE1 WT and CSB−/− throughout this manuscript, respectively), by co-transfection with lenti multi-guide plasmid containing gRNA_SPRTN-ΔC#1 and gRNA_SPRTN-ΔC#2 and px330 (Addgene, 82580) containing gRNA_SPRTN-ΔC#2 (Supplementary Table 6) (1 µg + 1 µg DNA). At 96 h after transfection, puromycin (1 µg ml−1) was added and increased to 2 µg ml−1 the following day. After selection, single cell-derived clones were selected and expanded. Editing efficiency was assessed by western blotting.

To generate XPG−/− RPE1 cells, RPE1-TetOn-Cas9-PuroS-TP53−/− cells were transfected with Cas9-2A-EGFP (pX458; Addgene #48138) containing sgXPG-4. Single clones were grown and verified by western blot and Sanger sequencing.

GFP–/GFP–DNMT1-expressing U2OS cells were established by transfecting U2OS cells with pEGFP–C1 or pEGFP–C1–DNMT1 (synthesized by VectorBuilder) using Lipofectamine 2000. At 36 h after transfection, cells were selected with 1 mg ml−1 G418. Single cell-derived clones were selected, expanded and GFP/GFP–DNMT1 expression was validated by live cell imaging and western blotting.

SiRNA transfections were performed using Lipofectamine RNAiMAX (Thermo Fisher Scientific) according to the manufacturer’s instructions. The details of siRNAs used in this study can be found in Supplementary Table 6.

CRISPRi screens

A pooled single guide RNA library was generated by combining the human protein-coding genome-wide CRISPRi-v2 library (containing 5 gRNA per gene; Addgene #83969) with multiple CRISPRi non-coding libraries (10 gRNA per gene; Addgene #86538, #86539, #86544, #86548 and #86549), covering the all long non-coding RNA genes expressed in K562 cells. Sublibraries were amplified in E. cloni cells (10G Elite duo, Lucigen) and mixed according to the relative amount of gRNA per individual sublibrary. Library transduction was performed as described73. Briefly, cells were transduced by spinoculation and transduced cells were selected with 0.75 µg ml−1 puromycin. A reference sample was collected 48 h post-transduction, the multiplicity of infection was calculated at day 6 and drug treatment was started at day 12. Every other day, cells were counted and replated at 500× representation with or without formaldehyde or 5-aza-dC at half-maximum inhibitory concentration (155 µM and 55 nM, respectively). Cell pellets were collected for genomic DNA (gDNA) extractions and library preparation/NGS at day 2 (initialome), day 10 (essentialome) and day 24 (endpoint). gDNA was extracted using the QIAamp DNA Blood Maxi Kit (Qiagen). Library preparation was then performed by subjecting 561 µg gDNA from the reference sample and 5.84 µg from experimental samples to 28 PCR cycles with NEBNext Ultra II Q5 Master Mix (New England Biolabs), using custom indexed primers (Supplementary Table 6) for simultaneous amplification of the backbone and addition of Illumina adaptors. PCR products were purified by excision and extraction from 1% agarose using the QIAquick Gel Extraction Kit (Qiagen). Sample concentrations were checked by a NanoDrop and diluted to a final concentration of 20 nM. Samples were multiplexed at 20 nM and sequenced on an Illumina HiSeq1500 using the CRISPRi_F1 primer. Enriched/depleted gRNAs in the endpoint samples compared to the 48 h reference sample were determined using DrugZ74.

Alamar blue assays

RPE1 cells (1,000 per well) were seeded in triplicate in 24-well plates and treated the next day with compounds (methanol-free formaldehyde, Fisher Scientific; 5-aza-dC, Sigma-Aldrich and illudin S, Santa Cruz). After 5 days, growth medium was replaced with 0.5 ml Alamar blue cell viability reagent (36 μg ml−1 resazurin in phosphate-buffered saline (PBS)) and plates were incubated for 1 h at 37 °C. Viability was assessed by using fluorescence (560 nm excitation/590 nm emission).

For assays with siRNA-transfected cells, cells were transfected 48 h before seeding and treated with the appropriate compounds the following day (72 h after transfection).

Colony formation assays

Cells were seeded at 400–1,000 per well in 6-well plates in technical triplicate. After 24 h, drugs (5-aza-dC, formaldehyde or illudin S) were added in fresh media. Formaldehyde-containing media were replaced with fresh media after 24 h. After 6–8 days, cells were fixed in 0.25% (wt/vol) crystal violet with 25% ethanol and washed with distilled water. Plates were scanned and colonies counted using FIJI. Technical triplicates were averaged and treated as one biological replicate. To assess UVC sensitivity, 4,000 cells per dish were seeded in triplicate in 6 cm dishes. The following day, cells were irradiated in PBS with UVC and PBS was replaced with fresh media. After 7 days, cells were stained with crystal violet. For assays with siRNA transfections, cells were first transfected in 6-well plates and clonogenic survival assays were seeded after 48 h. Excess cells were re-seeded for collection of western blot lysates 24 h later.

RNAPII co-immunoprecipitation

Elongating RNAPII immunoprecipitation was performed as previously described52. Cells were treated with formaldehyde (1 mM or 1.75 mM) or UVC (20 J m−2) and collected after 1 h. Chromatin enrichment was performed on ice in IP-130 buffer (30 mM Tris pH 7.5, 130 mM NaCl, 2 mM MgCl2, 0.5% Triton X-100 (Sigma-Aldrich) and cOmplete ethylenediaminetetraacetic acid (EDTA)-free protease inhibitor cocktail (Roche)) for 30 min, followed by centrifugation at 10,000g for 10 min and removal of the supernatant. Cell pellets (chromatin fractions) were lysed in IP-130 with 500 U ml−1 Benzonase (Merck), and 2 μg RNAPII-S2 antibody (Abcam; Supplementary Table 7) for 3 h at 4 °C. Protein complexes were immunoprecipitated by 1.5 h incubation with Protein A Agarose Beads (Sigma-Aldrich). Beads were washed 6× with IP-130 and immunoprecipitates were eluted by boiling in 2× NuPAGE lithium dodecyl sulfate (LDS) sample buffer.

Dsk2 pulldown

RPE1 cells were synchronized in 1% fetal bovine serum (FBS)-containing media for 24 h before formaldehyde (250 μM) or UVC (20 J m−2) treatment and collection 1 h later or as indicated. For pulldowns with dC/5-aza-dC, cells were grown to 60% confluency and synchronized at the G1/S boundary with 2 mM thymidine (Sigma-Aldrich) for 20–24 h. Cells were released into S-phase or 30 min and treated with 10 μM dC/5-aza-dC. Cells were collected by scraping in PBS and centrifugation at 500g for 5 min. Pellets were snap frozen on dry ice and stored at −80 °C.

Preparation of GST–Dsk2 affinity resin and subsequent Dsk2 pulldowns were performed essentially as previously described53. Cells were lysed in TENT buffer (50 mM Tris/HCl ph 7.5, 2 mM EDTA, 150 mM NaCl and 1% Triton X-100) supplemented with cOmplete Protease Inhibitor Cocktail tablets and 2mM N-ethylmaleimide (Sigma-Aldrich). Samples were incubated on ice for 10 min and sonicated extensively before centrifugation at full speed for 5 min at 4 °C. Supernatants were transferred to fresh tubes, protein concentrations were determined by the Bradford assay and standardized in TENT buffer. Ten per cent was removed as an input and prepared for western blotting.

Next, 0.5 ml GST–Dsk2 beads per 1 mg protein were equilibrated in ice-cold TENT buffer before distribution across samples. Pulldowns were performed overnight at 4 °C with rotation. Beads were washed three times by centrifugation (500g for 5 min at 4 °C) and resuspension in TENT buffer, then washed once with cold PBS. Elution was performed by boiling in 2× SDS loading dye containing 1 mM dithiothreitol (DTT) at 95 °C for 5 min.

RPB1 degradation and ATF3 induction and degradation assays

For RPB1 degradation experiments, cells were pre-treated with 100 µg ml−1 cycloheximide for 1 h, then treated with 1.75 mM formaldehyde for 1 h or 20 J m−2 UVC and allowed to recover for the indicated time or, for chronic treatments, treated with 250 µM formaldehyde for the indicated times. Where indicated, cells were pre-treated with or without 2 µM NEDDi (MLN4924) for 1 h before UVC/formaldehyde treatment. To assess disappearance/recovery of hypophosphorylated RPB1, cells were synchronized in 1% FBS-containing medium and treated with 250 µM formaldehyde for the indicated times. For ATF3 induction and degradation experiments, cells were synchronized in medium containing 1% FBS and treated with 250 µM formaldehyde for the indicated times, or treated with 9 J m−2 UVC and allowed to recover for 24 h. Cells were lysed in 1× NuPAGE LDS sample buffer, followed by SDS–polyacrylamide gel electrophoresis and western blotting with the indicated antibodies.

RRS assays

For recovery of RNA synthesis (RRS) assays, adherent cells on coverslips or in 24-well imaging plates were synchronized with 1% FBS-containing media for 24 h. Cells were then treated with 1.75 mM (RPE1) or 0.75 mM (HAP1) formaldehyde, or irradiated with UVC 20 J m−2. After 1 h, formaldehyde-treated cells were washed twice with PBS and allowed to recover for the indicated times before pulse labelling with 400 μM EU (Jena Bioscience) for 1 h, followed by a 30 min chase with Dulbecco’s modified Eagle medium. Cells were fixed with 4% formaldehyde in PBS for 15 min. For EU incorporation upon 5-aza-dC treatment, cells were synchronized using a double thymidine block, based on two cycles of overnight incubation with 2 mM thymidine (Sigma-Aldrich) and a 9 h release in between. Cells were released into S-phase and treated with 5-aza-dC for the indicated time periods, or 30 min in the case of HAP1 cells. In the final 30 min of treatment, EU was added followed by medium chase and fixation as above. Click-iT labelling and analysis was performed as described below.

EU and EdU detection by Click-iT

For Click-iT labelling, cells were permeabilized with 0.5% Triton X-100 in PBS for 15 min at room temperature (RT) and blocked in 1.5% bovine serum albumin (Thermo Scientific) in PBS for 15 min. Click reactions were performed for 1 h with 60 μM AF488-Azide (Jena Bioscience), 4 mM CuSO4 (Sigma-Aldrich), 10 mM (+)-sodium l-ascorbate (Sigma-Aldrich) and 0.5 μg ml−1 4,6-diamidino-2-phenylindole (Thermo Scientific) in a 50 mM Tris buffer. Coverslips/imaging plates were washed three times with PBS and mounted in Prolong Gold Antifade Mountant (Thermo Scientific; coverslips) or stored in PBS (imaging plates). Images were acquired from coverslips using a Zeiss LSM710 and software ZEN 2009 (Carl Zeiss) version 5.5.0.443 and analysed using CellProfiler. Images from 24-well imaging plates were acquired using a PerkinElmer Opera Phenix and analysed using Harmony software. For representative images, brightness was adjusted using FIJI (ImageJ) software. For comparisons between cell lines, per-nucleus EU intensity values were normalized to the mean EU intensity of untreated cells. For visualization purposes, graphs were plotted to exclude outlier cells exceeding 200% normalized EU intensity.

PLA

U2OS GFP–DNMT1 cells in 96-well CellCarrier Ultra (PerkinElmer) imaging plates were synchronized at the G1/S boundary and released into S-phase as described for RRS assays with 5-aza-dC. At 30 min after release into S-phase, cells were treated with 10 μM dC or 5-aza-dC. Then, 1 h later, cells were pre-extracted in ice-cold PBS with 0.2% Triton X-100 on ice for 2 min and fixed with 4% formaldehyde for 15 min. PLA was performed with anti-RPB1 CTD-pS2 and anti-GFP, using the Duolink In Situ PLA Probe Anti-Mouse Minus and Anti-Rabbit Plus probes and the DuoLink In Situ Detection Reagents FarRed Kit (Sigma-Aldrich) as per the manufacturer’s instructions. Images were acquired on a Zeiss 880 confocal microscope analysed using CellProfiler.

RT–qPCR with exogenous RNA spike-in for absolute quantification

Following treatment, cells were collected, counted and lysed in TRIzol reagent (Invitrogen). Exogenous ‘spike-in’ RNA from the DT40 chicken cell line was added to the lysed cells in proportion to the cell count. Total RNA extraction was then performed with the GENEzol TriRNA Pure Kit (Geneaid). RNA concentration and purity were measured by a NanoDrop 2000c (Thermo Scientific). The complementary DNA was prepared using qScript cDNA Synthesis Kit (QuantaBio, 95047). Expression level of genes was determined by real-time PCR conducted on a Bio-Rad CFX384 system using iTaq Universal SYBR Green Supermix (Bio-Rad). Gene expression was normalized to the chicken transcript cRPL4 for absolute quantification. All primers were validated for specificity and linearity.

TTchem-seq

Nascent RNA-seq (TTchem-seq) was performed as previously described75. Cells were mock treated or exposed to 1 mM formaldehyde for 1 h. Following drug removal, cells were allowed to recover as indicated and labelled with 1 mM 4SU (Glentham Life Sciences) during the last 30 min. Immediately after 4SU labelling, cells were washed with PBS and lysed in TRIzol reagent. RNA was isolated by TRIzol–chloroform isolation and ethanol precipitation. RNA was rDNAse-treated according to the manufacturer’s protocol (Macherey-Nagel) followed by phenol/chloroform isolation and ethanol precipitation. Then, 200 μg RNA (in a total volume of 116 μl) was fragmented by adding 4 μl 5 M NaOH and incubating on ice for 30 min, then stopped by addition of 80 μl of 1 M Tris pH 7 and cleaned up twice using the Micro Bio-Spin P-30 Gel Columns (Bio-Rad, 7326223) according to the manufacturer’s instructions. Biotinylation of 4SU residues was performed in a total volume of 250 μl, containing 10 mM Tris–HCl pH 7.4, 1 mM EDTA and 5 mg of MTSEA biotin-XX linker (Biotium, BT90066) for 30 min at RT in the dark. RNA was purified by phenol–chloroform extraction, denatured by 10 min incubation at 65 °C and added to 150 μl μMACS Streptavidin MicroBeads (Milentyl, 130-074-101). RNA was incubated with beads for 15 min at RT and beads were applied to a μColumn in the magnetic field of a μMACS magnetic separator. The beads were washed twice with 55 °C pulldown wash buffer (100 mM Tris–HCl pH 7.4, 10 mM EDTA, 1 M NaCl and 0.1% Tween-20). Biotinylated RNA was eluted twice by addition of 100 mM DTT and cleaned up using the RNeasy MinElute kit (Qiagen, 74204) using 1,050 μl ethanol (≥99%) per 200 μl reaction after addition of 700 μl of RLT buffer to precipitate RNA of less than 200 nucleotides. At least 30 ng of the purified 4SU-labelled RNA was then used as input for the TruSeq Stranded Total RNA kit (Illumina, 20020596) for library preparation. The libraries were amplified according to the manufacturer’s instructions. The library was amplified with ten PCR cycles and quality-control checked on the TapeStation (Agilent) using the High Sensitivity DNA Kit before pooling and paired-end sequencing on the NextSeq 550 (Illumina) system.

A sequencing quality profile was generated using FastQC (version 0.11.9). If needed, sequences were trimmed using TrimGalore (version 0.6.5). Reads were aligned to the human genome 38 using STAR (version 2.7.7a/gcc-8.3.1) with genome GRCh38_no_alt_analysis_set and sjdbGTFfile Homo_sapiens.GRCh38.94.correctedContigs.gtf (https://ftp.ncbi.nlm.nih.gov/genomes/all/GCA/000/001/405/GCA_000001405.15_GRCh38/seqs_for_alignment_pipelines.ucsc_ids/GCA_000001405.15_GRCh38_no_alt_analysis_set.fna.gz). Bam files were converted into stranded TagDirectories and University of California Santa Cruz (UCSC) genome tracks using Hypergeometric Optimization of Motif EnRichment tools (version 4.8.2)76. Example genome tracks were generated in Integrative Genomics Viewer (version 2.4.3). A list of 49,948 genes was obtained from the UCSC genome database (https://genome.ucsc.edu/cgi-bin/hgTables) selecting the known canonical table containing the canonical TSSs per gene. To prevent contamination of binding profiles, only genes of at least 3 kb in size and non-overlapping with at least 2 kb between genes (9,944 genes) were included.

TTchem-seq aggregated profiles were defined using the AnnotatePeaks.pl tool of Hypergeometric Optimization of Motif EnRichment without normalization of read counts. Genes were selected on size and strongest TTchem signal in the first 3 kb of genes in untreated WT cells. To compare different aggregate profiles, total reads per individual profile were normalized to nascent transcript levels, quantified by EU labelling performed in parallel to the TTchem-seq experiment.

PxP

First, 1 × 106 cells were seeded in 10-cm dishes and 24 h later synchronized with 1% FBS-containing medium. After 24 h, cells were treated with formaldehyde (1.75 mM, Fisher Scientific). Formaldehyde-treated cells were collected or washed twice with PBS, and recovered in fresh medium for 3 h or 6 h. Cell pellets were frozen at −80 °C and processed the next day. For PxP, cells were washed and resuspended in PBS at 2.5 × 104 cells µl−1. Then, 10 µl of cell suspension was lysed in 1× NuPAGE LDS sample buffer (Thermo Scientific) as input sample for western blot. The remaining cell suspension was pre-warmed for 45 s at 45 °C and mixed with an equal volume of low-melt agarose (2% in PBS, Bio-Rad) and immediately cast into plug molds (1703713, Bio-Rad). Plugs were placed at 4 °C for 5 min, then transferred into 1 ml ice-cold lysis buffer (1× PBS, 0.5 mM EDTA, 2% sarkosyl, cOmplete EDTA-free protease inhibitor cocktail (Merck) and 0.04 mg ml−1 Pefabloc SC (Merck)). Cells were lysed on a rotating wheel at 4 °C for 4 h. For electro-elution, plugs were transferred to wells of 10-well SDS–polyacrylamide gel electrophoresis gels (12%, 1.5 mm Novex WedgeWell or BOLT gels, Thermo Fisher). Electrophoresis was performed in 300 ml 3-(N-morpholino)propanesulfonic acid buffer at 20 mA per gel for 60 min in a Mini Gel Tank (Thermo Fisher). Following electro-elution, plugs were retrieved and transferred to tubes containing 1 ml 1×TBS (for PxP–MS) on a rotating wheel at 4 °C for 10 min. This wash was repeated once.

For western blotting, NuPAGE LDS sample buffer was added to plugs after electro-elution, before melting at 99 °C for 25 min.

PxP sample preparation for MS

For analysis by MS, 100 µl denaturation solution (4 M urea, 100 mM Tris–HCl (pH 7.5), 2 mM DTT and 10 mM chloroacetamide) were added per 100 mg plug in a 1.5 ml tube. After vortexing, samples were incubated at 37 °C for 30 min with agitation (1,500 rpm). Then, 20 µl trypsin solution (2 M urea, 50 mM Tris–HCl (pH 7.5), 1 mM DTT, 5 mM chloroacetamide and 20 µg ml−1 trypsin (Sigma-Aldrich)) was added per sample. After overnight incubation at 25 °C with rapid agitation (1,500 rpm), samples were centrifuged at high speed for 10 min. Supernatants were transferred to fresh tubes. Trifluoroacetic acid (TFA) was added at 1% (final concentration).

Purification and desalting of peptides on three layers of styrenedivinylbenzene reverse-phase sulfonate StageTips

StageTips were equilibrated by adding 100 µl of 100% acetonitrile (ACN), 100 µl of Solution 2 (30% methanol, 0.2% TFA and 0.2% TFA solution respectively). After the addition of each solution, tips were centrifuged at 600g at RT until no liquid was visible at the tip. Thereafter, samples were loaded into StageTips. Following centrifugation at 600g, samples were sequentially washed with 100 µl isopropanol (twice) and 100 µl 0.2% TFA (twice). After centrifugation at 600g in each washing step, samples were eluted with 60 µl elution buffer (1.25% ammonium hydroxide (NH4OH) and 80% ACN) into PCR tubes. Eluted peptides were dried using a SpeedVac centrifuge (Eppendorf, Concentrator plus) at 45 °C for 25 min and then samples were resuspended in A* Buffer (2% ACN/0.1% TFA).

LC–MS/MS measurements

LC–MS/MS analysis was performed on the Orbitrap Exploris 480 mass spectrometer coupled to a Thermo Scientific Vanquish Neo ultra high performance liquid chromatography system. First, 200 ng peptide per sample was loaded into a house-packed −50 cm reversed-phase column (75 μm diameter; packed with ReproSil-Pur C18-AQ 1.9 μm resin), and eluted with a gradient starting at 5% buffer B (80% ACN) and increased to 30% in 75 min with a flow rate of 0.300 µl min−1, 60% in 5 min and 95% in 10 min. For DIA (data-independent acquisition) measurement, the full scan range was set to 350–1,000 m/z at a resolution of 120,000. The full MS automatic gain control was set to 300% at a maximum injection time of 45 ms. Higher energy collisional dissociation (30%) was used for precursor fragmentation and the resulting fragment ions were analysed in 84 DIA windows at a resolution of 15,000 and an automatic gain control of 1,000. The windows were fixed size with 7.7 m/z width and 1 m/z overlap.

Raw MS data analysis

Raw files from MS were processed in DIA-NN 1.8.2 beta 22 using an in silico library (N-terminal methionine excision, cysteine carbamidomethylation and one missed cleavage are enabled). Peptide length range was 7–35, and precursor m/z range was 350–1,000. match-between-runs was unchecked. Besides these, default parameters were used.

Statistical analysis of MS data

Statistical analysis of MS data was performed in R (version 4.2.2). Due to an overall lower number of identified proteins, one replicate (no. 3) out of four was excluded from downstream analysis. Label-free quantitation intensities were log2 transformed and filtered for proteins identified in the three remaining replicates at the 0 h timepoint in both WT and CSB−/− cells. Values were quantile normalized between replicates using the R package preprocessCore (version 1.60.0) and missing values were imputed using the MinDet method from the R package MSnbase (2.24.0)77. Differentially abundant proteins were identified using a moderated t-test using the R package limma78 with Benjamini–Hochberg FDR correction79. Proteins with a log2-fold change >1 and a FDR ≤0.01 were considered significant.

DPC-seq

Cells were seeded in 6-well plates at 160,000 cells per well in 10% FBS-containing media in technical triplicate and 24 h synchronized with 1% FBS-containing media. After 24 h, cells were treated with 1.75 mM formaldehyde for 1 h. Next, cells were washed twice with PBS and either collected immediately or released into drug-free media for 6 h. Cells were scraped in 150 µl 2% SDS, 20 mM Tris–HCl pH 7.5, transferred to 1.5 ml tubes snap frozen in liquid nitrogen and stored at −80 °C. For KCl–SDS precipitation, samples were thawed at 55 °C for 5 min (1,200 rpm shaking) and sonicated using Covaris Focused ultrasonicator E220evo in 130 µl tubes (microTUBE AFA Fiber Pre-Slit Snap-Cap 6 × 16 mm; 1× cycle and 120 s). Samples were transferred to 1.5 ml tubes and 270 µl 2% SDS, 20 mM Tris–HCl pH 7.5 was added. DNA extraction was performed on 10% of the total lysate using the GeneJET Genomic DNA Purification Kit (Thermo Scientific) and considered ‘Input’. For KCl–SDS precipitation, 400 µl of KCl buffer (200 mM KCl and 20 mM Tris–HCl pH 7.5) was added, incubated on ice for 5 min and centrifugated full speed at 4 °C (5 min). Supernatants (soluble DNA) were transferred to fresh tubes for quantification. Pellets (protein and cross-linked protein–DNA complexes) were washed three times according to the following protocol: addition of 400 µl KCl buffer, incubation at 55 °C for 5 min (1,200 rpm shaking), incubation on ice (5 min) and full-speed centrifugation 4 °C (5 min). Next, pellets were resuspended in 400 µl KCl buffer + Proteinase K (0.2 mg ml−1) and incubated at 55 °C for 45 min (800 rpm shaking). Then, 10 µl UltraPure bovine serum albumin (Thermo Scientific) was added followed by cooling on ice for 5 min and centrifugation at maximum speed at 4 °C for 5 min. Next, supernatants containing cross-linked DNA were collected. Soluble and cross-linked samples were treated with 0.2 mg ml−1 DNAse-free RNAse A (Sigma-Aldrich) for 30 min at 37 °C. DNA concentrations were determined using the Qubit dsDNA HS Assay Kit (Thermo Scientific). Relative DPC amounts between samples were calculated as a ratio of cross-linked DNA to total DNA (cross-linked plus soluble DNA). Then, 50 ng DNA was concentrated via ethanol precipitation with 300 mM sodium acetate, 1 µl glycogen and 2.5× ethanol and resuspended in 20 µl nuclease-free water. DNA was run on 1% agarose gels, stained with SYBR-Gold and the DNA smear of 400–1,000 bp was excised. Gel slices were immersed in 500 µl gel-extraction buffer (10 mM Tris pH 8.0, 1 mM EDTA and 0.02% SDS) and rotated at 4 °C overnight. Gel slices and buffer were loaded into Spin-X columns (Corning Costar, CLS8160) and centrifuged at 14,000g for 10 min at 4 °C. Eluted DNA was ethanol precipitated as before and resuspended in 50 µl nuclease-free water. DNA was then subjected to library preparation via the NEBNext Ultra II DNA library prep kit (NEB, E7645L) using a 1/10 adaptor dilution and seven PCR cycles. Libraries were analysed via a Qubit and Tapestation, pooled at equimolar concentrations and sequenced on an Illumina NovaSeq with PE50 cycles. For all DPC-seq experiments, three biological replicates were performed, each consisting of pooled technical triplicates. A step-by-step protocol for DPC-seq can be found at Nature Protocol Exchange80.

DPC-seq analysis

Fastq files were generated using bcl2fastq2 (v2.20), low-quality reads were filtered out using fastp81 (v0.23.2) and aligned to the hg38 human genome via Bowtie2 (ref. 82) (v2.4.5). Alignments were sorted and indexed using Samtools83 (v1.16.1). Read coverages were calculated from alignments using Deeptools84 (v3.5.0) bamCoverage or bedtools85 (v2.30.0) coverage with GRCh38 as a reference. Per-gene coverage was normalized to reads per sample and gene length. Further analysis was conducted in R (v4.1.2) using custom scripts. Read coverages were compared via log2-fold change and t-tests with Bonferroni correction. All box plots, dot plots and genome track plots were generated using ggplot2 (v3.4.0) whereas metagene line plots and heat maps were generated using Deeptools. Comparisons with other datasets from ATAC-seq (GSE209659) and RNAPII ChIP-seq (GSE141798) was completed by analysing read coverage of these datasets in the same way as with DPC-seq data. All code for upstream processing, downstream analysis and plot generation are available at ref. 86.

NER excision assay

Chemiluminescent excision assays were performed as previously described87. Following treatments, 4.5 × 106 cells were collected and low-molecular-weight DNA was extracted. Spike-in DNA was added as an internal control (50 fmol of a 50 nt oligomer). High-molecular-weight DNA was removed by reverse size selection with 1.2× PCR clean up beads (MagBio), and samples were subjected to biotinylation with biotin-16-ddUTP (Merck). Samples were separated on 12% urea–polyacrylamide gel, transferred to hybond N+ nylon membrane (Amersham) and biotin-labelled DNA was detected using streptavidin–horseradish peroxidase (Abcam, ab7403) and detected using enhanced chemiluminescence.

Stringent CUT&Tag

CUT&Tag was performed in biological triplicate as previously described88 with minor modifications. Per condition, 200,000 untreated, formaldehyde-treated or released cells were resuspended in high salt wash buffer (20 mM HEPES pH 7.5, 600 mM NaCl, 0.5 mM spermidine in nuclease-free water with a Roche Complete Protease Inhibitor EDTA-free tablet) and prepared with concanavalin A-coated magnetic beads (Bangs Laboratories, BP531). Primary antibodies (Supplementary Table 7) were incubated 1:50 overnight at 4 °C. Cells were washed with high salt Dig-wash buffer (0.05% digitonin, 20 mM HEPES pH 7.5, 600 mM NaCl and 0.5 mM spermidine in nuclease-free water with a Roche Complete Protease Inhibitor EDTA-free tablet). The guinea pig anti-rabbit secondary antibody (Supplementary Table 7) was added 1:50 and incubated for 2 h at 25 °C. Cells were washed three times with 300 mM NaCl Dig-wash buffer, and pA-Tn5 tagmentation and subsequent steps were performed in 300 mM NaCl as per the original protocol (Dig-300 buffer; 0.01% digitonin, 20 mM HEPES pH 7.5, 300 mM NaCl and 0.5 mM spermidine in nuclease-free water with a Roche Complete Protease Inhibitor EDTA-free tablet). Libraries were prepared as described previously88, quantified with the NEBNext Library Quant Kit for Illumina (NEB, 7630S) and sequenced on an Illumina NextSeq 2000 (P3, 100 cycles).

CUT&Tag data processing

Illumina sequencing paired-end output files were demultiplexed using demux Illumina version 3.0.9 using the flags; -c -d -i -e -t 1 -r 0.01 -R -l 9. Resultant fq.gz files underwent sequencing quality control using FastQC v0.11.8, and their summary was visualized by MultiQC v1.11. Bases with a quality score <20 were trimmed from both reads using cutadapt (cutadapt -q 20). Fastq files were aligned to the combined hg38 and Escherichia coli genomes using bwa 0.7.17-r1188 with only reads in the whitelist regions of hg38 continuing the process pipeline. Duplicates were removed using Picard version 2.20.3 (Picard MarkDuplicates). BigWig files were created using deepTools version 3.5.1 bamCoverage using the mapped BAM files and the flags --binSize 3 --normalizeUsing CPM --extendReads, for fragments of up to 600 nt (using –maxFragmentLength). Profile plots were generated using deeptools 3.5.1 on positive-sense non-overlapping genes.

Statistics and reproducibility

Genome-wide CRISPRi screens were performed with three independent cell populations per conditions. DPC-seq experiments were performed in biological triplicate, where each biological replicate consists of three pooled technical replicates. The TTchem-seq experiment was performed once. CUT&Tag experiments were performed in biological triplicate. PxP–MS was performed in four indepent replicates but one was excluded from analysis because fewer peptides overall were detected in the excluded sample. All other experiments were performed a minimum of twice but usually three to four times independently, as indicated in the figure legends. Statistical analysis was performed using two-sided tests as appropriate; details and P values can be found in the legends. For data visualization purposes, in quantification of RRS assays, outlier EU intensity values exceeding 200–210% of the mean were excluded, but raw data including these values can be found in Source data. No statistical method was used to predetermine sample sizes. The experiments were not randomized. The investigators were not blinded to allocation during experiments and outcome assessment.