Genomic mapping of DNA-repair reaction intermediates in living cells with engineered DNA structure-trap proteins

Abstract

Diverse DNA structures occur as reaction intermediates in various DNA-damage and -repair mechanisms, most of which results from replication stress. We harness the power of proteins evolutionarily optimized to bind and “trap” specific DNA reaction-intermediate structures, to quantify the structures, and discern the mechanisms of their occurrence in cells. The engineered proteins also allow genomic mapping of sites at which specific DNA structures occur preferentially, using a structure-trapping protein and ChIP-seq- or Cut-and-Tag-like methods. Genome-wide identification of sites with recurrent DNA-damage intermediates has illuminated mechanisms implicated in genome instability, replication stress, and chromosome fragility. Here, we describe X-seq, for identifying sites of recurrent four-way DNA junctions or Holliday-junctions (HJs). X-seq uses an engineered, catalysis-defective mutant of Escherichia coli RuvC HJ-specific endonuclease, RuvCDefGFP. X-seq signal indicates sites of recombinational DNA repair or replication-fork stalling and reversal. We also describe methods for genomic mapping of 3ʹ-single-stranded DNA ends with SsEND-seq, in E. coli. Both methods allow genomic profiling of DNA-damage and -repair intermediates, which can precede genome instability, and are expected to have many additional applications including in other cells and organisms.

1. Introduction

Non-B-form DNA structures can arise as reaction intermediates in DNA replication mishaps, and DNA damage and its repair. The reaction intermediates are repaired or removed by DNA damage-processing proteins to promote faithful genetic transmission. Engineered proteins that bind specific DNA structures that are intermediates in DNA-damage and -repair reactions have been used to visualize, quantify and map the intermediates in spontaneous and induced DNA damage processing in living cells (Henrikus et al., 2019; Renzette et al., 2005; Rogakou, Boon, Redon, & Bonner, 1999; Scully et al., 1997). The advantage of protein-based detectors is high substrate specificity, a result of evolutionary optimization, which allows the discrimination of different DNA intermediates. Engineering of catalytically dead “trap”-protein detectors takes a further step by blocking processing once the DNA intermediate is bound (Chen et al., 2017; Kotlajich et al., 2018; Mei et al., 2021; Schmutz, Timashev, Xie, Patel, & de Lange, 2017; Shee et al., 2013; Xia et al., 2016). Transient DNA intermediates thus become persistent and so are more likely to be captured. Chromatin immunoprecipitation followed by sequencing (ChIP-seq) is a well-established method for detecting DNA-protein interactions (Park, 2009), as are Cut-and-Run (Skene & Henikoff, 2017) and Cut-and-Tag (Kaya-Okur et al., 2019) technologies. Here, we describe two modified ChIP-seq approaches for mapping DNA intermediate structures bound by engineered proteins.

1.1. Overview of the RuvCDefGFP Holliday-junction trap and X-seq

Holliday junctions (HJs) are four-way DNA junctions (Fig. 1A). HJs are intermediates in homologous recombination, which is required for repair of double-strand breaks (DSBs) (Fig. 1Bi), and single-strand gaps (ssDNA gaps) (Fig. 1Bii), and also occur during replication-fork stalling and reversal (Dunderdale et al., 1991; Seigneur et al., 1998) by the pairing of newly synthesized DNA strands with each other, which creates a HJ (Fig. 1Biii).

Fig. 1.

Origins of recurrent HJs in E. coli can be distinguished by the proteins that promote or oppose their accumulation. The various origins of Holliday junctions (HJs) in E. coli cells can be distinguished genetically, by their requirements for or reduction by specific proteins shown. Orange boxes, HJs; lines, single DNA strands; dotted lines, newly synthesized DNA. (A) HJs can be formed by (Bi) homology-directed repair of DNA double-strand breaks (DSBs), which requires RecA recombinase, and has a unique requirement for RecBCD DNA exonuclease, which resects DSB ends and loads RecA only at DSB ends (Churchill, Anderson, & Kowalczykowski, 1999). (Bii) Formation of HJs during single-strand (ss)DNA gap repair requires RecA recombinase, RecQ helicase, RecJ 5ʹ-ssDNA-dependent exonuclease, and RecF protein, which loads RecA at ssDNA gaps (Morimatsu & Kowalczykowski, 2003; Xia et al., 2016). (Biii) HJs that occur when stalled replication forks are reversed are prevented by RecQ and RecJ, and destroyed by RecBCD, and so are increased in mutants that lack any of these proteins (Xia et al., 2016). RecA is not required for fork reversal (Seigneur, Bidnenko, Ehrlich, & Michel, 1998), but its overproduction induces fork reversal (Xia et al., 2016). (C) Human counterparts of the reversed fork-promoting RAD51 (RecA ortholog), and reducing BLM, EME1 and RECQL4 proteins, inferred from cancer transcriptome data (Xia et al., 2016). Adapted from Fitzgerald, D. M., & Rosenberg, S. M. (2021). Biology before the SOS response—DNA damage mechanisms at chromosome fragile sites. Cells, 10.

We developed a HJ-trap protein, RuvCDefGFP, based on a catalytically inactive mutant of E. coli RuvC HJ-specific endonuclease (Xia et al., 2016). RuvCDefGFP possesses mutations that encode D7N and E66D amino acid substitutions, which inactivate the catalytic activity but not RuvC binding affinity for HJs. Based on a previous ChIP-seq method (Bonocora & Wade, 2015), we developed HJ mapping by X-seq (Mei et al., 2021; Xia et al., 2016), in which HJs are bound by RuvCDefGFP in cells that produce the RuvCDefGFP protein (Fig. 2A). The DNA is prepared, sheared by sonication, immunoprecipitated using anti-RuvC antibody, and then converted into sequencing libraries using ligation-based library preparation as shown in Fig. 2A. The DNA ends at HJs are blunted, A-tailed and eventually ligated to sequencing adapters while the HJ-RuvCDefGFP complex remains bound to complexes of antibody-IgG-Dynabeads with Protein A (Fig. 2A). This simplifies DNA purification between each step.

Fig. 2.

HJ mapping by X-seq in E. coli. (A) Diagram of X-seq (Mei et al., 2021; Xia et al., 2016). RuvCDefGFP-bound HJs are immunoprecipitated with anti-RuvC antibody. While RuvCDefGFP remains bound to immobilized antibody-protein A complex, DNA ends are blunted, A-tailed, and ligated with NEBNext adapters prior to elution. Adapter-ligated DNA fragments are converted into a sequencing library by PCR amplification, then sequenced. (B) X-seq detects three significant spontaneous HJ sites within the terminus region of the E. coli genome, one at the TerA unidirectional replication-barrier site, and two that flank the dif site of chromosome decatenation. All three form dependently on DSB repair-proteins RecA and RecBCD indicating their origins in DSB repair (see Fig. 1B). Replication terminus region at top, and origin, oriC, at the bottom. Orange, X-seq with RuvC antibody; sky blue, chromatin immunoprecipitation and sequencing (ChIP-seq) with nonspecific immunoglobulin G2a (IgG2a). Italics capital letters A–H, locations and orientations of TerA-TerH sites. Panel (B): Data from Mei, Q., Fitzgerald, D. M., Liu, J., Xia, J., Pribis, J. P., Zhai, Y., et al. (2021). Two mechanisms of chromosome fragility at replication-termination sites in bacteria. Science Advances, 7(25), eabe2846.

We have used X-seqto map spontaneous recurrent HJs in the E. coli genome and identified three major chromosome fragile sites (Mei et al., 2021). These peaks (Fig. 2B), which represent recurrent HJs, occur primarily in the replication terminus region, and are formed by homology-directed repair of DNA DSBs, which, we found, occur via two separate mechanisms. One-ended DNA-break ends form near the TerA site, dependently on termination of runaway replication forks that cross the terminus region and would otherwise continue “backward” in the genome towards the origin (Fig. 2B). These DNA-break ends occur via replication-blocking Tus protein binding at the TerA unidirectional replication-terminus site, and so are Tus-dependent. Two other major HJ sites flank the site at which chromosomes are decatenated following replication, the dif site, apparently caused by DNA breakage and repair during chromosome decatenation (Fig. 2B). Both of these mechanisms may be useful models of human chromosome fragility (Fitzgerald & Rosenberg, 2021; Mei et al., 2021).

1.2. Overview of the ExID single-stranded DNA 3ʹ-end trap and SsEND-seq

We created a trap protein for DNA with a 3ʹ-single-stranded (ss)DNA end or overhang, ExID, for Exonuclease I Defective (Liu et al., n.d.). ExID is based on the protein encoded by the sbcB15- mutant form of E. coli Exonuclease I (Exo I), a 3ʹ-specific ssDNA-dependent exonuclease that degrades ssDNA from a 3ʹ end (Lehman & Nussbaum, 1964). We have shown that ExID binds to recurrent induced and spontaneous 3ʹ-ssDNA-ends, which, we found, are created in a replication-dependent, transcription-influenced manner (Liu et al., n.d.). The approach for genomic mapping of 3ʹ-ssDNA ends with ExID is called Single-strand END-seq or SsEND-seq.

Fig. 3A and B shows schematic diagrams of SsEND-seq. In cells that carry a doxycycline-inducible gene encoding ExID-FLAG, 3ʹ-ssDNA ends are bound by ExID-FLAG. The DNA is prepared, sheared by sonication, then immunoprecipitated using anti-FLAG antibody, and purified using the TaKaRa ChIP Elute kit, after which it is converted into a strand-specific sequencing library using the TaKaRa DNA SMART ChIP-Seq Kit. Sequencing then identifies sites of ExID enrichment: recurrent 3ʹ-ssDNA ends (Fig. 3C) (Liu et al., n.d.).

Fig. 3.

SsEND-seq mapping of 3ʹ-ssDNA-ends in E. coli. (A) Diagram of SsEND-seq. ExID-FLAG-bound 3ʹ-ssDNA ends are immunoprecipitated with anti-FLAG antibody, followed by strand-specific library preparation using the Takara DNA SMART ChIP-Seq Kit, in which 5ʹ and 3ʹ ends of ssDNA fragments are tagged with known adapter sequences, then sequenced and a genomic map of recurrent 3ʹ-ssDNA ends is made. (B) SsEND-seq detects asymmetrical peaks. SsEND-seq reads are mapped to the top (TOP, orange) or bottom (BTM, purple) strand, and form different asymmetrical peaks in which summits indicate recurrent DNA 3ʹ ssDNA termini that ExID traps. (C) SsEND-seq detects peaks of predicted asymmetry and strandedness (Liu et al. (n.d.)). Heatmaps of SsEND-seq signals around peak summits of ~220 recurrent 3ʹ-ssDNA end sites. The signals were normalized by Reads Per Kilobase per Million mapped reads (RPKM), and sorted according to signal intensity. Signals are displayed along ±300-nt regions around and centered along the peak summits.

We have used SsEND-seq to map 3ʹ-ssDNA end sites that arise spontaneously in replicating E. coli (Liu et al., n.d.). Recurrent 3ʹ-ssDNA end sites detected by SsEND-seq give rise to asymmetrical peaks whose steep sides correspond to 3ʹ-ssDNA ends trapped by ExID, and the gentle slopes correspond to the random distribution of 5ʹ ends created by DNA sonication (Fig. 3B).

Detailed protocols for both X-seq and SsEND-seq in E. coli are presented below. In principle, these approaches can be applied to any organism or cell, as long as the trap protein can be produced in the organism and remain functional.

2. Considerations for experimental design with X-seq in E. coli

2.1. Interpretation of X-seq data

Before the experimental set-up, it is useful to consider in which cells to perform X-seq. Mapping HJs in genomes with X-seq indicates the presence and positions of four-way-junction structures but does not reveal how recurrent HJs formed, unless genetic tests are carried out (Fitzgerald & Rosenberg, 2021; Mei et al., 2021; Xia et al., 2016). In any organism or cell, HJs can be formed during homology-directed repair of DNA DSBs (Fig. 1Bi) and ssDNA gaps (Fig. 1Bii), or by replication-fork stalling and reversal (Fig. 1Biii and C). These sources can be distinguished definitively in E. coli by use of cells with null mutations in genes that encode the proteins that promote or reduce HJ formation via any of these mechanisms (Fitzgerald & Rosenberg, 2021; Mei et al., 2021; Xia et al., 2016) (Fig. 1B). Particularly, cells that lack the RecBCD double-strand end-dependent DNA exonuclease activity are defective for DSB repair and cannot form the DSB-repair HJs (Fig. 1Bi); so, recB or recC null mutants lack HJs (and X-seq signal) caused by DSB repair (Mei et al., 2021; Xia et al., 2016). By contrast, because RecBCD nuclease destroys reversed-fork HJs (Fig. 1Biii) (Michel, Ehrlich, & Uzest, 1997; Seigneur et al., 1998), fork-reversal HJs are more abundant in recB or recC deficient mutant cells and, with previous assays, were undetectable in wild-type cells, visible only in recB-null mutants (Michel et al., 1997; Seigneur et al., 1998). These opposite roles of RecBCD and these mutants can be used to distinguish HJs formed in DSB repair from those formed by fork reversal (Fig. 1Bi and iii) (Fitzgerald & Rosenberg, 2021; Mei et al., 2021; Xia et al., 2016). In addition, loss of RecQ or RecJ increased HJs, detected as RuvCDefGFP foci, formed by fork reversal (Xia et al., 2016). Similarly, HJ signal that is unaffected by RecBCD presence or absence, and instead requires RecF, which loads RecA recombinase onto ssDNA at ssDNA gaps, is likely to represent HJs formed during ssDNA gap repair (Fitzgerald & Rosenberg, 2021; Magner et al., 2007; Mei et al., 2021; Xia et al., 2016). Thus, analysis of the source of recurrent HJs detected by X-seq can be accomplished with an isogenic set of RuvCDefGFP-expressing cells that carry no rec mutations (“wild-type”) and each single null mutation of recA, recB, and recF, with recQ and recJ null mutant cells to shore up the conclusions. One such set is given in Section 4.1. A summary of these relationships between cell genotype and the appearance of detectable HJ structures that occur during repair of DSBs or ssDNA gaps, or those indicating reversed replication forks is given in Table 1.

Table 1.

Phenotypes of RuvCDefGFP-trapped HJ signals formed by various mechanisms in E. coli.

Processes that form HJs	E. coli strain genotype
	WT	ΔrecA	ΔrecB	ΔrecF	ΔrecJ	ΔrecQ	References
DSB repair	+	−	−	+	+	+	Mei et al. (2021) and Xia et al. (2016)
ssDNA gap repair	+	−	+	−	−	−	Xia et al. (2016)
Reversed forks (RecA overproduced)a	+	not tested	+	+	+++	+++	Xia et al. (2016)

^aObserved in cells with overproduced RecA (Xia et al., 2016), except the ΔrecA data for reversed forks, which are from pulsed-field gel electrophoresis, and were observed only in recB null-mutant cells without use of a HJ-trap protein (Michel et al., 1997; Seigneur et al., 1998).

2.2. Bioinformatic analysis of X-seq data

Illumina paired-end sequencing reads are aligned by BWA-MEM to the E. coli genome sequence of strain W3110 [National Center for Biotechnology Information Reference Sequence database accession: NC_007779.1] for strains given in Section 4.1, or to other reference sequences for other strain backgrounds used. Reads that have multiple primary hits or low-mapping quality are discarded using samtools. PCR duplicates are removed by Picard Tools MarkDuplicates. X-seq peaks are broad, and are difficult to detect with conventional peak-calling algorithms that are optimized for narrow peaks. To estimate the boundaries of X-seq peaks, we first divide the genome into 1-kb bins and count the reads in each bin. Next, numbers of reads are normalized to the median number of total reads. We then use change-point analyses to detect peak boundaries by detecting the change point in the mean for a minimum of two independent experiments and datasets. We use the cpt.mean function in the “changepoint” R package (Mei et al., 2021), with the PELT (Pruned Exact Linear Time) algorithm and Akaike’s information criterion as penalty. The input is normalized reads in 1-kb bins.

3. Considerations for SsEND-seq in E. coli

3.1. Bioinformatic analysis of SsEND-seq data

We process SsEND-seq data as follows (Liu et al., n.d.). Paired-end sequencing reads are mapped to the E. coli K12 MG1655 genome (NCBI: NC_000913.3) using BWA-MEM (v0.7.12). PCR duplicates are removed by Picard Tool MarkDuplicates. Other reference sequences would be used for other strain backgrounds. Mapped reads are then divided into two classes, top-strand (reference sequence in the 5ʹ–3ʹ direction) and bottom-strand (reference sequence in the 3ʹ–5ʹ direction) reads. The TaKaRa DNA SMART ChIP-Seq Kit ligates the read-2 adapter (Fig. 3A) to 3ʹ ends of the original ssDNA fragments and ligates the read-1 adapter to the 5ʹ ends (Fig. 3A). It then converts ssDNA to duplex DNA. Therefore, the reads 2, aligned to the top strand, and their readmate, reads 1, aligned to the bottom strand, are classified as reads mapped to top strand, shown as orange signal in Fig. 3B and C. Similarly, the reads 2 aligned to the bottom strand, and reads 1 aligned to the top strand, are reads mapped to the bottom strand, shown as purple signal in Fig. 3B and C. Significant peaks (adjusted P-value ≤0.00001) are called by MACS2 (v1.4.2) on the top and bottom strands individually.

4. Materials for X-seq in E. coli

4.1. Bacterial growth, induction of RuvCDefGFP (RDG) and HJ-RDG crosslinking

• Orbital shaker

• LB medium: 0.5% NaCl, 0.5% Bacto-Yeast extract (Gibco), 1% Bacto-Tryptone (Gibco)

• Doxycycline

• Formaldehyde Solution, 37%

• Phosphate-buffered saline (PBS), pH 7.0–7.2

• 2.5 M glycine

• Bacterial strains (Table 2): Note that the 594 strain background is derived from the W3110 “wild-type” E. coli K12 strain and carries the following mutations: galK2(Oc) galT22 rpsL179 IN(rrnD-rrnE)1 lac3350 (Bachmann, 1996). It lacks the F conjugative plasmid and contains no lambda prophage.

Table 2.

E. coli strains that can diagnose origins of HJ signal observed with X-seq.

Name	Relevant genotype	References
SMR19425 (“WT”)	594 hsdrk− mK+ Δattλ::PN25tetR FRT zfe2512.7::PN25tetOruvCDefgfp FRT	Xia et al. (2016)
SMR19406	594 hsdrk− mK+ Δattλ::PN25tetR FRT zfe2512.7::PN25tetOruvCDefgfp FRTcatFRT ΔrecA::FRTKanFRT	Xia et al. (2016)
SMR19407	594 hsdrk− mK+ Δattλ::PN25tetR FRT zfe2512.7::PN25tetOruvCDefgfp FRTcatFRT ΔrecB::FRTKanFRT	Xia et al. (2016)
SMR19427	594 hsdrk− mK+ Δattλ::PN25tetR FRT zfe2512.7::PN25tetOruvCDefgfp FRT ΔrecF::FRT	Xia et al. (2016)
SMR19392	594 hsdrk− mK+ Δattλ::PN25tetR FRT zfe2512.7::PN25tetOruvCDefgfp FRTcatFRT ΔrecQ1906::FRT	Xia et al. (2016)
SMR19398	594 hsdrk− mK+ Δattλ::PN25tetR FRT zfe2512.7::PN25tetOruvCDefgfp FRTcatFRT ΔrecJ::FRT	Xia et al. (2016)

4.2. DNA sonication

• Labquake rotator (Fisher Scientific)

• Bioruptor Pico Sonicator (Diagenode)

• 15 mL Bioruptor Pico Tube (Diagenode)

• Nanodrop (Thermo Scientific)

• FA lysis buffer:

  0.05 M HEPES

  150 mM NaCl

  1% Triton

  0.1% sodium deoxycholate

  0.1% SDS

  1 mM ethylenediaminetetraacetic acid (EDTA)

• Lysozyme (Fisher Scientific)

4.3. Immunoprecipitation of RDG

• Labquake rotator (Fisher Scientific)

• Magnetic rack for 1.5 mL microtube

• Anti-RuvC antibody 200 μg/mL (Santa Cruz sc-53,437)

• Dynabeads Protein A magnetic beads (Invitrogen 10002D)

• FA lysis buffer

4.4. Library construction

• Magnetic rack for 0.2 mL PCR strip

• Quick Blunting and Quick Ligation kits (NEB E0542L)

• Klenow Fragment (3ʹ→5ʹ exo-) (NEB M0212L)

• NEBNext Multiplex Oligos for Illumina (NEB E7335L, E7500L, E7700L, E7730L)

• FA lysis buffer

• FA lysis buffer 500:

  0.05 M HEPES

  500 mM NaCl

  1% Triton

  0.1% sodium deoxycholate

  0.1% SDS

  1 mM EDTA

• Wash buffer:

  50 mM HEPES

  0.5 M NaCl

  1% Triton

  0.1% sodium deoxycholate

  0.1% SDS

  1 mM EDTA

• Deoxycholate buffer:

  10 mM TE, pH 8.0

  250 mM LiCl

  0.5% sodium deoxycholate

  0.5% Nonidet-P40

• ChIP elution buffer:

  50 mM Tris–HCl, pH 7.5

  1% SDS

  10 mM EDTA

• 10 mM Tris–HCl, pH 7.5

• 10 mM Tris–HCl, pH 8.0

• TE buffer, pH 7.5

• 10 μM ChIP_p1 Primer: 5ʹ-AATGATACGGCGACCACCGAGATC TACAC

• 10 μM ChIP_p2 Primer: 5ʹ-CAAGCAGAAGACGGCATACGAGAT

4.5. Sequencing-library amplification and bead-based size selection

• Thermocycler

• Agilent TapeStation (or equivalent)

• Nanodrop (Thermo Scientific)

• NEBNext Multiplex Oligos for Illumina (NEB E7335L, E7500L, E7700L, E7730L)

5. Methods for X-seq in E. coli

5.1. Overview

The protocol can be completed in 5 days as follows:

Day 0 and 1: Bacterial growth, induction of RDG and HJ-RDG crosslinking.

Day 2: DNA sonication and immunoprecipitation of RDG.

Day 3: Library construction.

Day 4: Sequencing-library amplification and bead-based size selection.

5.2. Bacterial growth, induction of RDG and HJ-RDG crosslinking

1. Start a 3 mL LB shaking overnight culture for cells that carry the doxycycline-inducible ruvCDefgfp expression cassette.

2. Dilute saturated overnight cultures 1:500 (400 μL) into 20 mL LB medium with 100 ng/mL doxycycline to induce the production of RDG, then grow for 3 h at 37 °C shaking until the OD₆₀₀ is 0.4 ~ 0.8.

3. Add 550 μL 37% formaldehyde (~1% final concentration) to cross-link HJs with RDG. Incubate at room temperature for 20 min.

4. Add 5 mL 2.5 M glycine to quench the formaldehyde.

5. Transfer cells to centrifuge bottles and centrifuge at 7000 rpm for 7 min at 4 °C using a refrigerated centrifuge.

6. Resuspend the cell pellet in 1 mL PBS and transfer cells to a 1.5 mL microtube.

7. Centrifuge at 7000 rpm for 2 min at 4 °C.

   Discard the supernatant.

   Store cell pellets overnight at −20 °C.

5.3. DNA sonication

1. Resuspend cell pellet in 0.5 mL FA lysis buffer containing 4 mg/mL lysozyme.

2. Incubate for 30 min at 37 °C with rotation on a Labquake rotator.

3. Chill cell lysate on ice for 5 min.

4. Add ~750 g sonication beads to a 15 mL Bioruptor Pico tube.

   Wash beads once with 0.9 mL PBS.

5. Transfer lysate to the prepared Bioruptor Pico Tube.

   Sonicate in a 30 s on/off continuous pulse for 30 min.

6. Transfer sonicated lysate to a 15 mL tube.

   Centrifuge at 11,000 rpm for 20 min at 4 °C.

   Transfer the supernatant to a new 15 mL tube and discard cell debris (pellet).

7. Measure the DNA concentration of the lysate with these steps:

   • take a 15 μL sample of lysate and boil at >95 °C for 10 min to decrosslink
   • spin and take the supernatant
   • purify the supernatant with a PCR purification kit
   • use Nanodrop to measure the concentration of purified DNA

8. Adjust the lysate to ~150ng/μL.

9. Split diluted lysate into two 1.5 mL microtubes with 1 mL in each tube.

   Store lysates at −80 °C or proceed to the next step.

5.4. Immunoprecipitation of RDG

1. Transfer 50 μL Dynabeads to a 1.5 mL microtube.

   Place tubes on magnetic rack for 1 min and discard supernatant.

2. Add 1 mL FA lysis buffer to wash Dynabeads.

   Pipette up and down 10 times.

   Place tubes on magnetic rack for 1 min and discard supernatant.

3. Add 50 μL anti-RuvC antibody to Dynabeads.

   Incubate at room temperature with rotation on a Labquake rotator for 30 min.

4. Centrifuge briefly to bring all the liquid to the bottom of the tube.

   Place tubes on magnetic rack for 1 min and discard supernatant.

5. Add 1 mL FA lysis buffer to wash antibody-coated Dynabeads.

   Pipette up and down 10 times.

   Place tubes on magnetic rack for 1 min and discard supernatant.

6. Add 1 mL normalized cell lysates.

   Incubate overnight at 4 °C with rotation.

5.5. Library construction

1. Centrifuge briefly to bring all the liquid to the bottom of the tube.

   Place tubes on magnetic rack for 1 min and discard supernatant.

2. Resuspend Dynabeads in 200 μL FA lysis buffer and transfer to a 0.2 mL PCR strip.

   Place PCR strip on magnetic rack for 1 min and discard supernatant.

3. Add 200 μL FA lysis buffer.

   Pipette up and down 10 times.

   Place PCR strip on magnetic rack for 1 min and discard supernatant.

4. Add 200 μL 10 mM Tris–HCl, pH 7.5.

   Pipette up and down 10 times.

   Place PCR strip on magnetic rack for 1 min and discard supernatant.

5. Repeat step 4 once.

6. Prepare blunting solution as follows (per sample):

   10 μL 10×quick blunting buffer

   10 μL 1 mM dNTP Mix

   80 μL dH2O

   2 μL blunt enzyme mix

   Make a master mix for the number of samples planned.

7. Add ~100 μL blunting solution to each tube and incubate at room temperature for 30 min. Resuspend with pipette every 10 min.

8. Place PCR strip on magnetic rack for 1 min and discard supernatant.

9. Add 200 μL FA lysis buffer.

   Pipette up and down 10 times.

   Place PCR strip on magnetic rack for 1 min and discard supernatant.

   Repeat this step once.

10. Add 200 μL 10 mM Tris–HCl, pH 8.0.

    Pipette up and down 10 times.

    Place PCR strip on magnetic rack for 1 min and discard supernatant.

    Repeat this step once.

11. Prepare A-tailing solution master mix as follows (per sample):

    10 μL 10×NEB buffer 2

    2 μL 100 mM dATP

    88 μL dH2O

    2 μL Klenow fragment (3ʹ→5ʹexo–)

    Make a master mix for the number of samples planned.

12. Add ~100 μL A-tailing solution and incubate at 37 °C for 30 min.

    Resuspend with pipette every 10 min.

13. Place tubes on magnetic rack for 1 min and discard supernatant.

14. Add 200 μL FA lysis buffer.

    Pipette up and down 10 times.

    Place PCR strip on magnetic rack for 1 min and discard supernatant.

    Repeat this step once.

15. Add 200 μL 10 mM Tris–HCl, pH 7.5.

    Pipette up and down 10 times.

    Place PCR strip on magnetic rack for 1 min and discard supernatant.

    Repeat this step once.

16. Dilute the NEBNext Adapter for Illumina (15 μM) 10-fold in 10 mM Tris–HCl, pH 7.5 to a final concentration of 1.5 μM.

17. Prepare adapter ligation solution master mix as follows (per sample):

    50 μL 10×ligase buffer

    50 μL H2O

    2 μL quick ligase

    1 μL diluted NEBNext Adapter

    Make a master mix for the number of samples planned.

18. Add ~100 μL adapter ligation solution and incubate at room temperature for 15 min.

    Resuspend with pipette every 10 min.

19. Add 2 μL USER enzyme and incubate at 37 °C for 15 min.

    Resuspend with pipette every 10 min.

20. Place PCR strip on magnetic rack for 1 min and discard supernatant.

21. Add 200 μL FA lysis buffer.

    Pipette up and down 10 times.

    Place PCR strip on magnetic rack for 1 min and discard supernatant.

    Repeat this step once.

22. Add 200 μL FA lysis buffer 500.

    Pipette up and down 10 times.

    Place PCR strip on magnetic rack for 1 min and discard supernatant.

23. Add 200 μL wash buffer.

    Pipette up and down 10 times.

    Place PCR strip on magnetic rack for 1 min and discard supernatant.

24. Add 200 μL deoxycholate buffer.

    Pipette up and down 10 times.

    Place PCR strip on magnetic rack for 1 min and discard supernatant.

25. Add 200 μL TE buffer.

    Pipette up and down 10 times.

    Place PCR strip on magnetic rack for 1 min and discard supernatant.

26. Add 50 μL ChIP elution buffer and incubate at 70 °C for 15 min.

    Place PCR strip on magnetic rack for 1 min.

27. Transfer the supernatant to a 1.5 mL microtube and boil at >95 °C for 10 min.

28. Use 1.1 × Ampure XP beads to clean the DNA according to the manufacturer’s protocol.

29. Elute DNA with 24 μL water.

    Store at 4 °C or proceed to the next step.

5.6. Sequencing-library amplification and bead-based size selection

1. Prepare 1st round of library amplification solution as follows (per sample):

   20 μL adapter-ligated DNA

   25 μL NEBNext Q5 Hot Start HiFi PCR Master Mix

   2.5 μL 10 μM Universal PCR Primers/i5 Primer

   2.5 μL 10 μM Index Primers/i7 Primer

2. Run the following PCR protocol:

   98 °C	30 s
   98 °C	10 s	8 cycles
   65 °C	75 s
   65 °C	5 min
   4 °C	∞

3. Add 25 μL (0.5 ×) resuspended AMPure XP beads.

   Mix well by pipetting up and down 10 times.

   Incubate at room temperature for 5 min.

4. Place PCR strip on the magnetic rack until the supernatant is clear.

   Carefully transfer the supernatant to a new well.

   Discard beads that contain the large fragments.

5. Add 20 μL (0.4 ×) AMPure XP beads to the supernatant.

   Mix well by pipetting up and down 10 times.

   Incubate at room temperature for 5 min.

6. Put PCR strip on the magnetic rack until the supernatant is clear.

   Carefully remove and discard the supernatant.

7. Add 200 μL of freshly prepared 80% ethanol while the plate is still on the magnetic rack.

   Incubate at room temperature for 30 s and carefully remove and discard the supernatant.

8. Repeat step 7 once.

9. Air dry beads for 5 min while the plate is on the magnetic rack.

10. Remove PCR strip from the magnetic rack.

    Add 23 μL water.

    Mix well by pipetting up and down.

11. Incubate at room temperature for 2 min.

    Put PCR strip on the magnetic rack until the solution is clear.

    Transfer 20 μL of the supernatant to a new PCR tube.

12. Prepare 2nd round of library amplification solution as follows (per sample):

    20 μL size-selected DNA

    25 μL NEBNext Q5 Hot Start HiFi PCR Master Mix

    2.5 μL 10 μM ChIP_p1 Primer

    2.5 μL 10 μM ChIP_p2 Primer

13. Run the following PCR protocol:

    98 °C	30 s
    98 °C	10 s	15 cycles
    65 °C	75 s
    65 °C	5 min
    4 °C	∞

14. Clean the DNA with 45 μL (0.9 ×) Ampure XP beads according to the manufacturer’s protocol.

15. Quantify the library with Nanodrop.

16. Proceed to next-generation sequencing.

6. Materials for SsEND-Seq in E. coli

6.1. Bacterial growth, ExID-FLAG induction and crosslinking to 3ʹ-single-stranded DNA ends

• Orbital shaker

• LB medium: 0.5% NaCl, 0.5% Bacto-Yeast extract (Gibco), 1% Bacto-Tryptone (Gibco)

• Doxycycline

• Formaldehyde Solution, 37%

• Phosphate-buffered saline (PBS), pH 7.0–7.2

• 4.5 M Tris, pH 8.0

• Bacterial strain (Table 3): Note that the MG1655 strain background is the first sequenced E. coli K12 wild-type strain (Blattner et al., 1997). It lacks the F conjugative plasmid and contains no lambda prophage.

Table 3.

E coli strain that produces ExID-3xFLAG protein.

Name	Relevant genotype	Reference
SMR24776	MG1655 ΔaraBAD567 Δattλ::PBAD ΔypeC::PN25tetR FRT (mntH-nupC)::PN25tetOExID-3xFLAG FRT	Liu et al. (n.d.)

6.2. DNA sonication

• Labquake rotator (Fisher Scientific)

• Bioruptor Pico Sonicator (Diagenode)

• 1.5 mL Bioruptor Pico microtube (Diagenode)

• FA lysis buffer:

  0.05 M HEPES

  150 mM NaCl

  1% Triton

  0.1% sodium deoxycholate

  0.1% SDS

  1 mM EDTA

• Lysozyme (Fisher Scientific)

6.3. Immunoprecipitation of ExID-FLAG

• 1.5 mL snaplock microfuge tube (Axygen MCT-150-C-S)

• Labquake rotator (Fisher Scientific)

• Magnetic rack for 1.5 mL microtube

• ANTI-FLAG M2 antibody (Sigma-Aldrich F1804)

• Dynabeads Protein G magnetic beads (Invitrogen 10003D)

• Wash buffer:

  50 mM HEPES

  0.5 M NaCl

  1% Triton

  0.1% sodium deoxycholate

  0.1% SDS

  1 mM EDTA

• Deoxycholate buffer:

  10 mM TE, pH 8.0

  250 mM LiCl

  0.5% sodium deoxycholate

  0.5% Nonidet-P40

• PBS with 0.02% Tween 20

• TaKaRa ChIP Elute Kit

• QuantStudio^™ 3 Real-Time PCR (or equivalent)

• DEPC (Diethyl pyrocarbonate)-treated water (Fisher Bioreagents)

• BIO-RAD iTaq Universal SYBR Green Supermix

• 100 μM qPCR Primer 1: 5ʹ-ACGGCATGATTGAAATATCGA GTTC

• 100 μM qPCR Primer 2: 5ʹ-GGTACTGGATTTTCGTGAGACAG

6.4. Strand-specific library construction

• Vacufuge (Eppendorf )

• TaKaRa DNA SMART ChIP-Seq Kit

• Thermocycler

• Agilent TapeStation (or equivalent)

6.5. DNA polyacrylamide gel size selection and DNA extraction

• Vertical electrophoresis cells

• 30% Acrylamide/Bis Solution, 29:1 (Bio-Rad 161,015)

• 10 × TBE:

  1 M Tris-base, pH 7.5

  1 M boric acid

  20 mM EDTA

• 10% Ammonium persulfate (APS, store aliquotes at −20 °C)

• N,N,Nʹ,Nʹ-Tetramethylethylenediamine (TEMED)

• DNA extraction buffer:

  10 mM Tris–HCl, pH 8.0

  0.3 M NaCl

  1 mM EDTA

• Ethidium bromide

• 3 M sodium acetate

• Glycogen (20 mg/mL)

• Isopropyl alcohol

• KAPA Library Quantification kit (Roche)

7. Methods for SsEND-seq in E. coli

7.1. Overview

The protocol can be completed in 6 days as follows:

Day 0 and 1: Bacterial growth, ExID-FLAG induction and crosslinking to 3ʹ-single-stranded DNA ends.

Day 2: DNA sonication, and immunoprecipitation of ExID-FLAG steps 1 through 8.

Day 3: The rest of the steps of immunoprecipitation of ExID-FLAG.

Day 4: Strand-specific library construction.

Day 5: DNA polyacrylamide gel size selection and DNA extraction.

7.2. Bacterial growth, ExID-FLAG induction and crosslinking to 3ʹ-single-stranded DNA ends

1. Start 3 mL LB overnight cultures for cells that carry the doxycycline-inducible ExID-FLAG expression cassette.

2. Dilute saturated overnight cultures 1:100 putting 400 μL in 40 mL LB medium with 1 ng/mL doxycycline to induce the production of ExID-FLAG, then grow for 2 h shaking at 37 °C to OD600 0.5 ~ 1.0.

3. Take 1 mL of the sample into cuvette to measure the OD600.

4. Add 1.1 mL 37% formaldehyde (~1% final concentration) to cross-link chromatin and proteins.

   Incubate at room temperature for 20 min.

5. Take 12.5 mL of cells at OD600=1, or normalize from lower density from 12.5 mL to as much as 25 mL cells.

6. Add 2.5 mL~ 5 mL 4.5 M Tris, pH 8.0 (750 mM final concentration) to quench the formaldehyde.

7. Transfer cells to centrifuge bottles and centrifuge at 7000 rpm for 7 min at 4 °C using a refrigerated centrifuge.

8. Resuspend the cell pellet in 1 mL PBS and transfer cells to a 1.5 mL microtube.

9. Centrifuge at 7000 rpm for 2 min at 4 °C.

   Discard the supernatant and store cell pellet overnight at −80 °C.

Notes:

• Because 3ʹ-ssDNA ends are mostly DNA-replication intermediates, we use rich (LB) medium, which allows multifork replication, and a low concentration of doxycycline that induces low levels of ExID-FLAG, to minimize the amount of non-DNA-bound ExID-FLAG.

7.3. DNA sonication

1. Resuspend cell pellet in 1 mL FA lysis buffer containing 4 mg/mL lysozyme.

2. Incubate for 1 h at 37 °C with rotation on a Labquake rotator. Chill cell lysate on ice for 5 min.

3. Transfer 125 μL lysate to a 1.5 mL Bioruptor Pico microtube.

   Sonicate in 30 s on/off continuous pulse for 15 min; wait for 5 min to cool the lysate, then sonicate again for 15 min.

4. Transfer sonicated lysate to a 1.5 mL microtube.

5. Repeat step 3 and 4.

6. Centrifuge the sonicated lysate at 11,000 rpm for 10 min at 4 °C.

   Transfer supernatant to a new tube and discard the cell debris pellet.

7. Take 15 μL for agarose gel electrophoresis analysis.

   Visually inspect aiming for a DNA smear between 100 and 400 bp.

   Store the rest at −80 °C or proceed to the next step.

Notes:

• Sonication optimization might be necessary if a different cell density or a different lysate volume is desired.

• After sonication, agarose gel electrophoresis analysis can be performed to verify that DNA fragments range from 100 to 400 bp. Boil 15 μL lysate at 96 °C for 10 min to decrosslink lysate. Centrifuge at 11,000 rpm for 10 min at 4 °C. Transfer supernatant to a tube and add 1 μL RNase A solution. RNase A treatment reduces the background signal, improving visual assessment of shearing.

7.4. Immunoprecipitation of ExID-FLAG

1. Dilute 16 μL ANTI-FLAG M2 antibody into 400 μL PBS with 0.02% Tween 20.

   Prepare a master mix for the number of immunoprecipitation reactions planned.

2. Vortex the vial of Dynabeads for >30 s to resuspend.

   Transfer 100 μL Dynabeads to a 1.5 mL Snaplock microfuge tube.

3. Place tubes on magnetic rack for 1 min to separate beads from solution and discard supernatant.

4. Remove tubes from magnetic rack.

   Add ~416 μL of the prepared antibody master mix to the Dynabeads.

   Incubate at room temperature with rotation on a Labquake rotator for 1 h.

5. Centrifuge briefly to bring the liquid down to the bottom of the tube.

   Place tubes on magnetic rack for 1 min and discard supernatant.

6. Remove tubes from the magnetic rack.

   Wash the Dynabeads once with 400 μL PBS with 0.02% Tween 20.

   Rotate the tubes at room temperature for 2 min.

7. Centrifuge briefly to bring the liquid to the bottom of the tube.

   Place tubes on magnetic rack for 1 min and discard supernatant.

8. Remove tubes from the magnetic rack.

   Add 200 μL cell lysate and 200 μL PBS with 0.02% Tween 20.

   Incubate at 4 °C overnight with rotation.

9. Centrifuge briefly and place tubes on magnetic rack for 1 min.

   Discard supernatant.

10. Wash magnetic beads 3 times with 400 μL wash buffer.

    Between each wash, resuspend beads by rotating the tubes for 2 min at room temperature, then place the tubes on a magnetic rack for 1 min and discard supernatant.

11. Wash magnetic beads 3 times with 400 μL deoxycholate buffer.

    Between each wash, resuspend beads by rotating the tubes for 2 min at room temperature, then place the tubes on a magnetic rack for 1 min and discard supernatant.

12. Elute DNA using the TaKaRa ChIP Elute Kit according to its user manual.

13. Dilute 1 μL ChIPed ssDNA in 7.5 μL DEPC-treated water for qPCR quantification.

    Store the rest at −80 °C.

14. Prepare qPCR primer mix by diluting 10 μL of 100 μM qPCR primer 1 and 2 in 1.4 mL water.

15. Prepare qPCR reactions as follows (per sample):

    5 μL SYBR Green master mix

    2.5 μL diluted qPCR primer mix

    2.5 μL diluted ChIPed ssDNA

16. Run the following qPCR protocol:

    50 °C	2 min
    95 °C	3 min	40 cycles
    95 °C	15 s
    60 °C	30 s

17. Proceed to the next step with samples of Ct value less than 28.

Notes:

• Low-retention tubes are used throughout the DNA elution step to minimize DNA loss.

• Elute DNA first with 23 μL DEPC-treated water, then with 18 μL water to maximize DNA recovery.

• Quantitative PCR is performed to approximate the quantity of ChIPed ssDNA based on an empirical Ct value. Ct values <28 indicate adequate DNA for library construction.

7.5. Strand-specific library construction

1. Thaw ~40 μL ssDNA on ice.

   Vacufuge at 60 °C for 40 min to concentrate ssDNA.

2. Resuspend ssDNA in 10.5 μL DEPC-treated water and transfer to a PCR strip.

3. Prepare library with the Takara DNA SMART ChIP-Seq Kit according to its user manual.

4. Do 24-cycle PCR amplification for library construction.

5. Follow the single size selection protocol in the user manual.

   Elute library in 27 μL water.

6. Use TapeStation or equivalent for DNA analysis to ensure generation of a successful library with a visible DNA smear.

   Proceed to the next step with successful libraries or store the libraries at −20 °C.

Notes:

• 24 PCR cycles are used because the amount of ChIPed ssDNA is usually less than 100 pg, which is the lower limit of the kit.

• There might be non-specific PCR products between 170 and 200 bp if using >18 cycles of PCR. If seen, proceed to the next step for DNA separation and purification from polyacrylamide gel.

7.6. DNA polyacrylamide gel size selection and DNA extraction

1. Make an 8% nondenaturing DNA polyacrylamide gel (0.75 mm-thick, 5 mL):

   30% Acrylamide/Bis solution 29:1	1.35 mL
   10 × TBE	0.5 mL
   Water	3.15 mL
   10% APS	52 μL
   TEMED	3.125 μL

2. Thaw 25 μL library.

   Vacufuge at 60 °C for 15 min.

   Resuspend in 10 μL water.

3. Add 2 μL 6 × DNA loading dye.

   Load entire sample onto the 8% PAGE gel.

4. Separate by electrophoresis for ~1.5 h at 100 V until the xylene cyanol FF band reaches the bottom of gel.

5. Dilute 2 μL ethidium bromide in ~30 mL water.

   Submerge the gel in the staining solution for 5 min.

6. Cut out the smear 250–600 bp.

   Pierce the bottom of a 0.7 mL tube and place into a 1.5 mL tube and place gel slice into the pierced 0.7 mL tube.

7. Shred the gel slice by spinning at 13,000 rpm for 5 min.

   Discard the 0.7 mL tube.

8. Add 0.4 mL DNA extraction buffer.

   Incubate with rotation for >16 h at 4 °C.

9. Transfer gel and liquid into a Spin-X column and spin at 13,000 rpm for 5 min.

10. Add 40 μL 3 M sodium acetate, 2 μL glycogen (20 mg/mL), and 400 μL isopropyl alcohol to flow-through.

    Incubate at −80 °C for 30 min.

11. Centrifuge at 13,000 rpm for 10 min at 4 °C.

    Discard the supernatant without disturbing the pellet.

12. Rinse the pellet with 1 mL cold 70% ethanol.

    Centrifuge at 10,000 rpm for 5 min.

    Discard the supernatant carefully to avoid disturbing the pellet.

13. Air dry the pellet.

    Dissolve the nucleic acid pellet in 12 μL DEPC-treated water.

14. Use KAPA Library Quantification kit to measure library concentration.

15. Proceed to next-generation sequencing.

Notes:

• 6 × DNA loading dye contains two different dyes, bromophenol blue and xylene cyanol FF, for visual tracking of DNA migration during electrophoresis. In an 8% nondenaturing PAGE gel, bromophenol blue migrates to a position near 45 bp, and xylene cyanol FF 160 bp.

8. Conclusions

X-seq and SsEND-seq are both powerful approaches for identifying genomic sites with recurrent DNA-damage and -repair intermediate structures, which are otherwise transient. An informative landscape of replication stress-induced DNA damage intermediates can be obtained using these two approaches along with other DNA intermediate-mapping tools, such as END-seq (Canela et al., 2016), which we used with X-seq in a previous study (Mei et al., 2021). However, it should be noted that rare, or non-recurrent sites cannot be seen with bulk sequencing-based methods because their signals are indistinguishable from background noise.