High-Resolution Mapping of Protein-DNA Interactions in Mouse Stem Cell-Derived Neurons using Chromatin Immunoprecipitation-Exonuclease (ChIP-Exo)

Summary

Precise determination of protein-binding locations across the genome is important for understanding gene regulation. Here we describe a genomic mapping method that treats chromatin-immunoprecipitated DNA with exonuclease digestion (ChIP-exo) followed by high-throughput sequencing. This method detects protein-DNA interactions with near base-pair mapping resolution and high signal-to-noise ratio in mammalian neurons.

Abstract

Identification of specific protein-DNA interactions on the genome is important for understanding gene regulation. Chromatin immunoprecipitation coupled with high-throughput sequencing (ChIP-seq) is widely used to identify genome-wide binding locations of DNA-binding proteins. However, the ChIP-seq method is limited by its heterogeneity in length of sonicated DNA fragments and non-specific background DNA, resulting in low mapping resolution and uncertainty in DNA-binding sites. To overcome these limitations, the combination of ChIP with exonuclease digestion (ChIP-exo) utilizes 5’ to 3’ exonuclease digestion to trim the heterogeneously sized immunoprecipitated DNA to the protein-DNA crosslinking site. Exonuclease treatment also eliminates non-specific background DNA. The library-prepared and exonuclease-digested DNA can be sent for high-throughput sequencing. The ChIP-exo method allows for near base-pair mapping resolution with greater detection sensitivity and reduced background signal. An optimized ChIP-exo protocol for mammalian cells and next-generation sequencing is described below.

Introduction

The locations of protein-DNA interactions provide insight into the mechanisms of gene regulation. Chromatin immunoprecipitation coupled with high-throughput sequencing (ChIP-seq) has been used for a decade to examine genome-wide protein-DNA interactions in living cells^1,2. However, the ChIP-seq method is limited by heterogeneity in DNA fragmentation and unbound DNA contamination that lead to low mapping resolution, false positives, missed calls, and non-specific background signal. The combination of ChIP with exonuclease digestion (ChIP-exo) improves upon the ChIP-seq method by trimming ChIP DNA to the protein-DNA crosslinking points, providing near base-pair resolution and a low background signal^3,4,5. The greatly improved mapping resolution and low background provided by ChIP-exo allow accurate and comprehensive protein-DNA binding locations to be determined across the genome. ChIP-exo is able to reveal functionally distinct DNA-binding motifs, cooperative interactions between transcription factors (TF), and multiple protein-DNA crosslinking sites in a certain genomic binding location, not detectable by other genomic mapping methods^3,4,6,7.

ChIP-exo was initially used in budding yeast to examine the sequence-specific DNA binding of TFs, to study the precise organization of the transcription pre-initiation complex, and sub-nucleosomal structure of individual histones across the genome^4,8,9. Since its introduction in 2011⁴, ChIP-exo has been successfully utilized in many other organisms including bacteria, mice, and human cells^{7,10,11,12,13,14,15,16,17}. In 2016, Rhee et al.¹⁴ used ChIP-exo in mammalian neurons for the first time to understand how neuronal gene expression was maintained after the downregulation of programming TF Lhx3, which forms a heterodimer complex with another programming TF Isl1. This study showed that in the absence of Lhx3, Isl1 is recruited to new neuronal enhancers bound by Onecut1 TF to maintain gene expression of neuronal effector genes. In this study, ChIP-exo revealed how multiple TFs dynamically recognize cell type and cell stage-specific DNA regulatory elements in a combinatorial manner at near-nucleotide mapping resolution. Other studies also used the ChIP-exo method to understand the interplay between proteins and DNA in other mammalian cell lines. Han et al.⁷ used ChIP-exo to examine genome-wide organization of GATA1 and TAL1 TFs in mouse erythroid cells using ChIP-exo. This study found that TAL1 is directly recruited to DNA rather than indirectly through protein-protein interactions with GATA1 throughout erythroid differentiation. Recent studies also used ChIP-exo to profile the genome-wide binding locations of CTCF, RNA Polymerase II, and histone marks to study epigenomic and transcriptional mechanisms in human cell lines^18,19.

There are several versions of the ChIP-exo protocol available^3,5,20. However, these ChIP-exo protocols are difficult to follow for researches who are not familiar with next-generation sequencing library preparation. An excellent version of the ChIP-exo protocol was published with easy-to-follow instructions and a video²¹, but contained many enzymatic steps that require a significant amount of time to complete. Here we report a new version of the ChIP-exo protocol containing reduced enzymatic steps and incubation times, and explanations for each enzymatic step²¹. End repair and dA-tailing reactions are combined in a single step using end prep enzyme. The incubation times for index and universal adapter ligation steps are reduced from 2 h to 15 min using a ligation enhancer. The kinase reaction after the index adapter ligation step described in the previous ChIP-exo protocol is removed. Instead, a phosphate group is added during oligo DNA synthesis to one of the 5’ ends of the index adapter (Table 1), which will be used for the lambda exonuclease digestion step. While the previous ChIP-exo protocol used RecJ_f exonuclease digestion to eliminate single-stranded DNA contaminants, this digestion step is removed here because it is not critical for the quality of the ChIP-exo library. In addition, to purify reverse-crosslinked DNA after ChIP elution, a magnetic beads purification method is used instead of the phenol:chloroform:isoamyl alcohol (PCIA) extraction method. This reduces the incubation time of DNA extraction. Importantly, it removes the majority of adapter dimers formed during index adapter ligation, which may impact the efficiency of ligation-mediated PCR.

The ChIP-exo protocol presented here is optimized for the detection of precise protein-DNA interactions in mammalian neurons differentiated from mouse embryonic stem (ES) cells. Briefly, harvested and crosslinked neuronal cells are lysed, to allow chromatin to be exposed to sonication, then sonicated so that appropriately sized DNA fragments are obtained (Figure 1). Antibody-coated beads are then used to selectively immunoprecipitate fragmented, soluble chromatin to the protein of interest. While the immunoprecipitated DNA is still on the beads, end-repair, ligation of sequencing adapters, fill-in reaction and 5’ to 3’ lambda exonuclease digestion steps are performed. The exonuclease digestion step is what gives ChIP-exo its ultra-high resolution and high signal-to-noise ratio. Lambda exonuclease trims the immunoprecipitated DNA a few base-pairs (bp) from the crosslinking site, thus causing contaminating DNA to be degraded. The exonuclease-treated ChIP DNA is eluted from the antibody-coated beads, protein-DNA crosslinks are reversed, and proteins are degraded. DNA is extracted and denatured to single-stranded ChIP DNA, followed by primer annealing and extension to make double-stranded DNA (dsDNA). Next, ligation of a universal adapter to the exonuclease-treated ends is performed. The resulting DNA is purified, then PCR amplified, gel purified, and subjected to next-generation sequencing.

The ChIP-exo protocol is longer than the ChIP-seq protocol, but is not very technically challenging. Any successfully immunoprecipitated ChIP DNA can be subjected to ChIP-exo, with several additional enzymatic steps. The notable advantages of ChIP-exo, such as ultra-high mapping resolution, a reduced background signal, and decreased false positive and negatives, regarding genomic binding sites, outweigh the time cost.

Protocol

NOTE: Autoclaved distilled and deionized water (ddH₂O) is recommended for making buffers and reaction master mixes. Sections 1−4 describe cell lysis and sonication, sections 5−7 describe chromatin immunoprecipitation (ChIP), sections 8−11 describe enzymatic reactions on beads, sections 12 and 13 describe ChIP elution and DNA purification, and sections 14−19 describe library preparation.

1. Harvesting and crosslinking cells

1. After differentiating mouse ES cells into postmitotic neurons, add 11% formaldehyde to the harvested cells to a final concentration of 1% (v/v). Rock cells on a rocking platform (a rocker, shaker, or rotator) for 15 min, at room temperature (RT, 25 °C).
   NOTE: Depending on the target protein to be crosslinked, less formaldehyde crosslinking (for example, a final concentration of 0.5%) or double crosslinking with disuccinimidyl glutarate (DSG) can be used.
2. Add 2.5 M glycine to a final concentration of 150 mM to stop the crosslinking reaction. Rock cells on a rocking platform at RT, for 5 min.
3. Centrifuge the crosslinked cells in 15 mL conical tubes at RT, for 6 min, at 1,350 x g. Aspirate the solution, then resuspend cells in 5 mL of 1x phosphate-buffered saline (PBS).
   NOTE: The crosslinked cell pellets can be stored at -80 °C after flash freezing with liquid nitrogen.

2. Cell lysis

NOTE: The following steps in this protocol are for approximately 2 x 10⁷ neuronal cells differentiated from mouse ES cells. To break open cells, lysis buffers containing various detergents will be used. Add 50 µL of 1000x complete protease inhibitor (CPI) stock to 50 mL of buffer just prior to use.

1. Prepare lysis buffers 1−3 as described in Table 2.
2. Thaw crosslinked cell pellets on ice, then thoroughly resuspend cell pellets in 3 mL of lysis buffer 1 in 15 mL conical tubes. Rock at 4 °C for 15 min on a rocking platform. Centrifuge at 1,350 x g for 5 min at 4 °C and aspirate supernatant.
3. Thoroughly resuspend pelleted cells in 3 mL of lysis buffer 2. Rock at 4 °C for 10 min on a rocking platform. Centrifuge at 1,350 x g for 5 min at 4 °C and aspirate supernatant.
4. Add 1 mL of lysis buffer 3 to each pellet, then keep on ice. Immediately proceed to sonication.

3. Sonicating chromatin

NOTE: Keep samples on ice or at 4 °C during this sonication procedure to reduce crosslink reversal.

1. Using a sterile spatula, add sonication beads (Table of Materials) up to the 0.2 mL graduation mark on 15 mL polystyrene tubes. Wash sonication beads by vortexing in 600 µL of 1x PBS until no dry spots are visible. Centrifuge the tubes for 5 s at 30 x g and aspirate 1x PBS.
   NOTE: Sonication in polystyrene tubes is more efficient than sonication in polypropylene tubes. If a smaller number of cells (for example, <10⁶ cells) is sonicated, use 1.5 mL polystyrene tubes without adding sonication beads.
2. Thoroughly resuspend the nuclear lysates in 1 mL of lysis buffer 3 (from step 2.4), then transfer to the 15 mL polystyrene tubes containing sonication beads. Briefly vortex the polystyrene tubes.
3. To fragmentize the crosslinked chromatin DNA, sonicate nuclear lysates at 4 °C for an optimized number of cycles, with power amplitude at 30 W, and sonication cycles set to 30 s on/30 s off.
   NOTE: Optimization of sonication for each cell type and batch will result in the best ChIP-exo yield. For 2 x 10⁷ mouse neuronal cells, 20−30 sonication cycles at the mentioned settings will fragment the chromatin in the range of 100−500 bp.
4. After sonication is complete, transfer all the supernatant (sonicated lysates), from each sample, to 1.5 mL tubes. Add 10% 2-[4-(2,4,4-trimethylpentan-2-yl)phenoxy]ethanol) to each sample at a final concentration of 1%. Mix by pipetting.
5. To pellet cell debris, centrifuge samples at 13,500 x g for 10 min at 4 °C. Transfer all the sonicated lysates (supernatant) from each sample to new 1.5 mL tubes.
6. Take 30 µL of sonicated lysate from each sample (~3% of sonicated lysate) and add to new 1.5 mL tubes to check sonication. Store remaining sonicated lysates at 4 °C until ready for incubation with antibody-coated beads.

4. Checking sonication

1. Make 20 mL of 2x proteinase K buffer by adding 2 mL of 0.5 M Tris-HCl (pH 8.0), 2 mL of 0.5 M EDTA, 10 mL of 10% sodium dodecyl sulfate (SDS) and 6 mL of autoclaved ddH₂O. Do not add CPI to this buffer. Store in 50 mL tube at RT.
2. To reverse the protein-DNA crosslink, by removing the proteins, take the 30 µL of sonicated chromatin from each sample (from step 3.6), add 166 µL of autoclaved ddH₂O, 200 µL of 2x proteinase K buffer and 4 µL of 20 mg/mL proteinase K. Briefly vortex the samples, and then incubate at 65 °C for 1−3 h at 24 x g.
3. NOTE: The sonicated lysates can be reverse crosslinked at 65 °C overnight.
4. Extract DNA using PCIA (25:24:1) and ethanol precipitation method as follows.
   CAUTION: PCIA is toxic. Use under a fume hood with standard personal protective equipment (PPE).
   1. Add 400 µL of PCIA to each sample in 1.5 mL tubes, set vortex to maximum speed and vortex samples for 20 s. Centrifuge at 18,400 x g for 6 min at RT. Two phases will be observed. Carefully transfer the upper aqueous layer (clear phase) of each sample to new 1.5 mL tubes.
   2. Add 1 µL of 10 mg/mL RNase A. Incubate samples at 37 °C for 30 min.
   3. Add 1 µL of 20 mg/mL glycogen to each sample, then precipitate with 1 mL of ice-cold 100% ethanol (stored in a -20 °C freezer). Mix briefly, then incubate samples in -80 °C freezer for 30 min to 1 h. Centrifuge samples at 18,400 x g for 10 min at 4 °C and carefully pour out 100% ethanol.
   4. Wash pellets with 500 µL of ice-cold 70% ethanol (stored in a -20 °C freezer). Centrifuge at 18,400 x g for 5 min at 4 °C. Carefully pour out 70% ethanol.
   5. Incubate samples in 1.5 mL tubes at 50 °C until remaining ethanol is evaporated. Resuspend DNA pellets in 15 µL of autoclaved ddH₂O or nuclease-free ddH₂O.
   6. Run the extracted DNA samples, along with a DNA ladder, in a 1.5% agarose gel at 120−180 V and check the size of the sonicated DNA.

5. Antibody incubation with beads

NOTE: The following steps in this protocol are for approximately 2 x 10⁷ neuronal cells differentiated from mouse ES cells. Do not freeze and thaw magnetic beads at any point during the ChIP-exo protocol as the beads may crack causing contamination of the sample or the antibody’s performance may be compromised.

1. After cell lysis and sonication, prepare Protein G magnetic beads (Table of Materials) for ChIP, mix magnetic beads until homogenous, then add 25 µL of magnetic beads to 2 mL protein low bind tubes.
   NOTE: The type of magnetic beads used will depend on the species of the antibodies.
2. Wash beads with 1 mL of blocking solution (Table 2), mix well, then place on a magnetic rack for 1 min. While still on a magnetic rack, remove supernatant once it is clear.
3. Add 1 mL of blocking solution to the magnetic beads. Rock tubes for 10 min at 4 °C on a rocking platform. Briefly spin, then place tubes on a magnetic rack and remove supernatant.
4. Repeat step 5.3 two more times.
5. Add 500 µL of blocking solution to magnetic beads. Briefly spin the antibody against Isl1 (0.04 µg/µL, Table of Materials), then add 4 µg of antibody to corresponding 2 mL protein low bind tubes containing the magnetic beads in blocking solution.
6. Repeat step 5.5 without antibody (i.e., the no antibody control for ChIP).
   NOTE: The amount of antibody to add can be determined empirically by considering the quality of the antibody and the number of cells used for ChIP.
7. Rock samples at 4 °C for 6−24 h on a rocking platform.

6. Chromatin immunoprecipitation (ChIP)

1. Wash antibody-coated beads from step 5.8 with 1 mL of blocking solution. Place samples on a rocking platform at 4 °C for 5 min. Remove supernatant.
2. Repeat step 6.1 two more times.
3. Resuspend antibody-coated beads in 50 µL of blocking solution, then transfer to new 2 mL protein low bind tubes. For each ChIP sample, add sonicated lysates (~1.0 mL from step 3.6) to antibody-coated beads in 2 mL protein low bind tubes.
4. Incubate each sample on a rocking platform overnight at 4 °C.

7. ChIP washes

NOTE: Keep samples on ice or at 4 °C to maintain protein-DNA crosslinking during ChIP washes.

1. Make high salt wash buffer, LiCl wash buffer and 10 mM Tris-HCl buffer (pH 7.4) as described in Table 3. Store in 50 mL tubes at 4 °C. Add 50 µL of 1000x CPI stock to all buffers just prior to use.
2. Briefly spin samples in 2 mL protein low bind tubes to collect liquid from the caps, then place on a magnetic rack for 1 min and remove supernatant carefully with a pipette.
3. For ChIP washes, add 1 mL of lysis buffer 3 (at 4 °C) to each tube. Mix on a rocking platform at 4 °C for 5 min. Briefly spin samples, then place on a magnetic rack for 1 min and remove the supernatant with a pipette.
4. Add 1 mL of cold high salt wash buffer to each tube. Mix on a rocking platform at 4 °C for 5 min. Briefly spin samples, then place on a magnetic rack for 1 min and remove the supernatant with a pipette.
5. Add 1 mL of cold LiCl wash buffer to each tube. Mix on a rocking platform at 4 °C for 5 min. Briefly spin samples, then place on a magnetic rack for 1 min and remove the supernatant with a pipette.
6. Add 1 mL of cold 10 mM Tris-HCl buffer (pH 7.4) to each tube. Mix on a rocking platform at 4 °C for 5 min. Briefly spin samples, then place on a magnetic rack for 1 min and remove the supernatant with a pipette.
   NOTE: Tris-EDTA buffer (pH 8.0) can be used instead of Tris-HCl buffer (pH 7.4).
7. Repeat steps 7.4−7.6 three times.
8. Transfer the sample beads with 500 µL of Tris-HCl buffer (pH 7.4) to fresh 1.5 mL tubes. Briefly spin samples, then place on a magnetic rack for 1 min and remove the supernatant with a pipette.

8. End repair and dA-tailing reaction on beads

NOTE: Sonication often generates non-blunt ended dsDNA. An end-repair reaction is required to make blunt-ended DNA prior to the dA-tailing reaction followed by the index adapter ligation step. For sticky end DNA ligation with index adapter DNA, dATP is added to the 3’ end of a blunt, dsDNA fragment by the dA-tailing reaction. End prep reaction mix contains dATP.

1. After ChIP washes, add 38 µL of autoclaved ddH₂O to the sample beads in 1.5 mL tubes for end repair and dA-tailing reaction.
2. Add 5.6 µL of end prep reaction mix and 2.4 µL of end prep enzyme mix (Table of Materials) to each sample (total reaction volume: 46 µL). Incubate samples at 20 °C for 30 min.
3. Wash beads as described in previous ChIP wash steps 7.4−7.6.

9. Index adapter ligation on beads

NOTE: The index adapter has 6−10 bases of barcoded index sequences, which are specific to a given sample used for multiplexing multiple samples in high-throughput sequencing. Index adapter DNA sequences are described in Table 1.

1. For index adapter ligation, add 27 µL of cold 10 mM Tris-HCl buffer (pH 7.4) to the sample beads. Add 2 µL of 15 µM index adapter, 0.5 µL of ligation enhancer, and 15 µL of ligase master mix (Table of Materials) to each sample (total reaction volume: 44.5 µL).
2. Incubate samples at 20 °C for 15 min. Wash beads as described in steps 7.4−7.6.

10. Fill-in reaction on beads

NOTE: After adapter ligation, there is no phosphodiester bond between the 5’ end of the adapter and the 3’ end of the ChIP DNA. The nick can be repaired by a fill-in reaction.

1. Add 47 µL of cold 10 mM Tris-HCl buffer (pH 7.4) to the sample beads.
2. Make the fill-in reaction mix (Table 4 and Table of Materials) on ice.
3. Add 11.1 µL of fill-in mix to each sample (total reaction volume: 58.1 µL). Incubate samples at 30 °C for 20 min. Wash beads as described in steps 7.4−7.6.

11. Lambda exonuclease digestion on beads

1. After fill-in reaction on beads, add 50 µL of cold autoclaved ddH₂O to the sample beads.
2. To digest the ChIP DNA in the 5’ to 3’ direction, add 6 µL of 10x lambda exonuclease buffer and 2 µL of 5 U/µL lambda exonuclease (Table of Materials, total reaction volume: 58 µL). Incubate samples at 37 °C for 30 min. Wash beads as described in steps 7.4−7.6.

12. Elution and reverse crosslinking

1. Make ChIP elution buffer as described in Table 5.
   NOTE: Do not add CPI to ChIP elution buffer.
2. To elute ChIP samples from the beads, resuspend samples in 75 µL of ChIP elution buffer and incubate at 65 °C for 15 min at 130 x g.
3. Add 2.5 µL of 20 mg/mL proteinase K (Table of Materials) to the samples. Briefly vortex, then incubate the samples overnight at 65 °C.

13. DNA extraction

1. After samples have incubated overnight at 65 °C, briefly spin samples, place on a magnetic rack for 1 min, then transfer the supernatant from each sample to new 1.5 mL tubes.
2. Add 1 µL of 10 mg/mL RNase A to the samples. Briefly vortex, then incubate the samples at 37 °C for 30 min. Add ~25 µL of autoclaved ddH₂O to volume up to 100 µL.
3. Purify DNA using magnetic beads (Table of Materials). Elute DNA with 16 µL of autoclaved ddH₂O or nuclease-free ddH₂O.

14. Denaturing, primer annealing and primer extension

1. After ChIP elution and reverse crosslinking, transfer 16 µL of the extracted DNA samples from step 13.3 to PCR tubes.
2. Make denaturing and primer annealing mix (Table 6 and Table of Materials) on ice. Add 1.2 µL of denaturing and primer annealing mix to each sample (total reaction volume: 17.2 µL). Run samples using the program described in Table 6 to denature and anneal primers to the template DNA.
3. Make primer extension mix (Table 7 and Table of Materials) on ice.
4. Once the program to denature and anneal primers is complete, add 3 µL of primer extension mix to the samples (total reaction volume: 20.2 µL). Run samples using the program for primer extension described in Table 7.
5. Immediately proceed to dA-tailing reaction.

15. dA-tailing reaction

NOTE: For sticky end DNA ligation with universal adapter DNA, dATP is added to the 3’ end of blunt, dsDNA by the dA-tailing reaction.

1. Make dA-tailing mix (Table 8 and Table of Materials) on ice.
2. Add 4.1 µL of dA-tailing mix to each sample (total reaction volume: 24.3 µL). Run samples using the program for dA-tailing (Table 8).
3. Immediately proceed to universal adapter ligation.

16. Universal adapter ligation

NOTE: The universal adapter includes high-throughput sequencing-specific sequences for DNA sample recognition for sequencing chemistry. The universal adapter DNA sequences are described in Table 1.

1. After dA-tailing reaction, make universal adapter ligation mix (Table 9 and Table of Materials) for universal adapter ligation.
2. Add 21.5 µL to each sample (total reaction volume: 45.8 µL) and incubate samples for 15 min at 20 °C.

17. DNA cleanup

1. Purify DNA using magnetic beads (Table of Materials). Elute DNA with 21 µL of autoclaved ddH₂O or nuclease-free ddH₂O.
2. Store samples at -20 °C until ready to perform ligation-mediated PCR (LM-PCR) and PCR purification.

18. Ligation-mediated PCR

NOTE: LM-PCR primer sequences are described in Table 1.

1. After DNA cleanup, make LM-PCR mix (Table 10 and Table of Materials) on ice.
2. Add 29 µL of LM-PCR mix to each sample (total reaction volume: 50 µL) and run samples using the program for LM-PCR (Table 10).
3. Immediately proceed to gel purification of PCR amplified DNA, followed by DNA purification for high-throughput sequencing.

19. DNA purification of LM-PCR amplified DNA

1. Run the LM-PCR amplified DNA samples, along with a DNA ladder, in a 1.5% agarose gel at 120−180 V and excise 200–400 bp bands.
2. Purify DNA from the excised gel using a gel extraction kit (Table of Materials).
3. After DNA purification, check the concentration (Table of Materials) of each ChIP-exo library sample. Submit samples for high-throughput sequencing for single-read sequencing.
   NOTE: The preferred read length for single-read sequencing is at least 35 bp, which is sufficient to uniquely align the sequencing reads to the reference genome in mouse or human cells. For most transcription factors in mouse or human cells, 10−20 million reads per ChIP-exo sample are sufficient to identify their genomic binding locations. At least 30 million reads per sample are required to map genomic regions enriched in histone marks in mammalian cells.

Representative Results

Figure 2A illustrates sonication results after cell lysis and sonication, with various cycles, of motor neuron cells differentiated from mouse ES cells. The optimal number of sonication cycles (for example, 12 cycles in Figure 2A) generated strong DNA intensity in 100−400 bp DNA fragments. High-quality ChIP-exo libraries are based on the size and quantity of fragmented chromatin DNA. Thus, optimization of sonication is recommended for each cell type and batch of cells before starting ChIP-exo.

Figure 2B shows ligation-mediated PCR (LM-PCR) products of the ChIP-exo DNA samples amplified by 18−21 PCR cycles. LM-PCR amplifies ChIP-exo DNA fragments that are ligated with DNA adapters for next-generation sequencing. PCR primer is a part of the DNA adapter sequences. The size of sonicated DNA fragments is around 100−400 bp (Figure 2A). After lambda exonuclease digestion, the size of DNA fragments would become the half size of the starting material (50−200 bp). Approximately 125 bp of DNA adapters were ligated to the ChIP-exo DNA fragments thus, the expected size of LM-PCR products is 175−325 bp. The minimal number of PCR cyclesc while still being sufficient to amplify the ChIP-exo library, are performed to avoid over-amplification of ChIP-exo DNA.

Here, ChIP-exo for Isl1 was performed in mouse ES cell-derived motor neurons. Isl1 is a motor neuron programming transcription factor, which is specifically expressed in postmitotic motor neurons. The results for Isl1 ChIP-exo show significant amounts of 200−400 bp LM-PCR amplified ChIP-exo DNA (Figure 2B). The band around 100 bp indicates PCR artifacts from adapters and PCR primers. Running a no antibody control alongside an experimental sample is also important to ensure non-specific, background DNA was digested by the lambda exonuclease treatment. As an example, we identified Isl1 bound locations in mouse ES cell-derived motor neurons using ChIP-exo and ChIP-seq (Figure 3). The ChIP-exo signal was highly focused at Isl1-binding sites, detecting multiple clustered Isl1 transcription factor binding patterns. The ChIP-seq signal displayed broader signals, indicating that ChIP-exo of Isl1 has higher mapping resolution than ChIP-seq of Isl1.

Figure 1: Schematic of the ChIP-exo protocol. After ChIP (steps 1−7), the end-repair and dA-tailing reactions make ChIP DNA blunt-ended by adding a phosphate group to the 5’ end of the DNA and adding dATP to the 3’ end of the DNA (step 8). Sonicated ends of ChIP DNA, on the beads are ligated with the index adapter (step 9). After the fill-in reaction, the adapter DNA with the 5’ overhang becomes blunt-ended (step 10). Lambda exonuclease digests the sonicated DNA 5’ to 3’ up to the protein (TF)-DNA crosslinking points (step 11). After elution, reverse-crosslinking and DNA extraction (steps 12 and 13), denatured single-stranded DNA is made double-stranded by index PCR primer extension (step 14), followed by dA-tailing (step 15). The universal adapter is then ligated to the exonuclease-treated end (step 16). The resulting library is cleaned up, PCR-amplified, and subjected to next-generation sequencing (steps 17−19). Mapping the 5’ ends of the resulting sequencing tags to the reference genome demarcates the exonuclease barrier and thus the precise site of protein-DNA crosslinking.

Figure 2: Sonication and LM-PCR amplification of a ChIP-exo library. (A) 1.5% agarose gel of the electrophoresed sonicated DNA. 12, 18, and 24 cycles of sonication were conducted to find optimal sonication cycles for 2 x 10⁷ cells of mouse ES cell-derived motor neurons. Following sonication, the DNA sample was reverse crosslinked, and extracted using PCIA and ethanol precipitation. Each sample contained 4 x 10⁵ cell equivalents, which is 2% of sonicated cells. 12 cycles of sonication produced a greater yield of sonicated DNA with the desired size (100−500 bp). (B) 1.5% agarose gel of electrophoresed ChIP-exo libraries following 18−21 cycles of LM-PCR for no antibody control (No Ab) and Isl1 in mouse ES cell-derived motor neurons. Each sample contained 10 x 10⁶ cell equivalent of mouse motor neurons. The no antibody ChIP-exo result (lane 2) demonstrates that non-specific, contaminating DNA is eliminated by lambda exonuclease. The ChIP-exo libraries for Isl1 (lane 3) show amplified DNA libraries around 200−400 bp, indicating the ChIP-exo libraries were successfully amplified by adapter ligation-mediated PCR. Please click here to view a larger version of this figure.

Figure 3: Comparison of ChIP-exo to ChIP-seq for Isl1 at specific loci. The blue and red filled plots show the distribution of sequencing tags for ChIP-seq and ChIP-exo Isl1-bound locations, respectively, proximal to Slit3 and Fgfr1 gene in nascent spinal motor neurons differentiated from mouse ES cells.

Table 1: Oligonucleotides used in this protocol.

Oligo name	Length (nt)	Sequence						Note
Index adapter-forward*	66	5'/Phos/CAAGCAGAAGACGGCATACGAGATXXXXXXXXGTGACTGGAGTTCAGACGTGTGCTCTTCCGATCT 3'						5' phosphate, index (XXXXXXXX), 3' T-overhang
Index adapter-reverse*	33	5'GATCGGAAGAGCACACGTCTGAACTCCAGTCAC 3'
Ligation adapter-forward*	58	5'AATGATACGGCGACCACCGAGATCTACACTCTTTCCCTACACGACGCTCTTCCGATCT 3'
Ligation adapter-reverse*	34	5'GATCGGAAGAGCGTCGTGTAGGGAAAGAGTGTAG 3'
Index PCR primer*	22	5'CAAGCAGAAGACGGCATACGAG 3'
Universal PCR primer*	19	5'AATGATACGGCGACCACCG 3'
*The above sequencing adapters and PCR primers are designed for Illumina HiSeq and NextSeq Sequencing Platforms.

Table 2: Recipes for lysis buffers 1−3 and blocking buffer. Store in 50 mL tubes at 4 °C. Add 50 µL of 1000x CPI (complete protease inhibitor) stock to all buffers just prior to use.

Lysis buffer 1
Reagent	Volume (mL)	[Final]
1 M HEPES (pH 7.3)	2.5	50 mM
5 M NaCl	1.4	140 mM
0.5 M EDTA (pH 8.0)	0.1	1 mM
50% Glycerol	10	10%
10% Octylphenol ethoxylate	2.5	0.50%
10% 2-[4-(2,4,4-trimethylpentan-2-yl)phenoxy]ethanol	1.25	0.25%
Autoclaved ddH2O	Fill to 50
Lysis buffer 2
Reagent	Volume (mL)	[Final]
0.5 M Tris-HCl (pH 8.0)	1	10 mM
5 M NaCl	2	200 mM
0.5 M EDTA (pH 8.0)	0.1	1 mM
Autoclaved ddH2O	Fill to 50
Lysis buffer 3
Reagent	Volume (mL)	[Final]
1 M Tris-HCl (pH 8.0)	0.5	10 mM
5 M NaCl	1	100 mM
0.5 M EDTA (pH 8.0)	0.1	1 mM
10% Deoxycholic Acid	0.5	0.10%
30% N-Lauroylsarcosine sodium salt solution	0.83	0.50%
Autoclaved ddH2O	Fill to 50
Blocking solution
Reagent	Amount	[Final]
Bovine serum albumin (BSA)	250 mg	0.50%
Complete Protease Inhibitor (CPI, 1000x)	50 µL	1x
Phosphate buffered saline (PBS)	Fill to 50 mL

Table 3: Recipes for ChIP washes: High Salt Wash buffer, LiCl Wash buffer and 10 mM Tris-HCl buffer (pH 7.4). Store in 50 mL tubes at 4 °C. Add 50 µL of 1000x CPI stock to all buffers just prior to use.

High salt wash buffer
Reagent	Volume (mL)	[Final]
1 M HEPES (pH 7.3)	2.5	50 mM
5 M NaCl	5	500 mM
0.5 M EDTA (pH 8.0)	0.1	1 mM
10% 2-[4-(2,4,4-trimethylpentan-2-yl)phenoxy]ethanol	5	1%
5% Deoxycholic acid	1	0.10%
5% SDS	1	0.10%
Autoclaved ddH2O	Fill to 50
LiCl wash buffer
Reagent	Volume (mL)	[Final]
0.5 M Tris-HCl (pH 8.0)	2	20 mM
0.5 M EDTA (pH 8.0)	0.1	1 mM
1 M LiCl	12.5	250 mM
10% Octylphenol ethoxylate	2.5	0.50%
5% Deoxycholic acid	5	0.50%
Autoclaved ddH2O	Fill to 50
10 mM Tris-HCl buffer
Reagent	Volume (mL)	[Final]
1 M Tris-HCl (pH 7.4)	0.5	10 mM
Autoclaved ddH2O	Fill to 50

Table 4: Fill-in reaction master mix.

Reagent	1x (µL)	[Final]
20 mg/mL BSA	0.6	0.2 mg/mL
10x phi29 DNA polymerase buffer	6	1x
3 mM dNTPs	2.5	150 µM
10 U/µL phi29 DNA polymerase	2	10 U
Reaction mix volume	11.1

Table 5: Recipe for ChIP Elution buffer. Store in 50 mL tubes at RT.

Reagent	Volume (mL)	[Final]
0.5 M Tris-HCl (pH 8.0)	5	50 mM
0.5 M EDTA (pH 8.0)	1	10 mM
10% SDS	5	1%
Autoclaved ddH2O	up to 50 mL

Table 6: Denaturing and Primer Annealing reaction master mix and program.

Denaturing and primer annealing mix
Reagent	1x (µL)	[Final]
20 mg/mL BSA	0.2	0.2 mg/mL
3 mM dNTPs	0.5	75 µM
10 µM Index PCR primer	0.5	0.25 µM
Reaction mix volume	1.2
Denaturing and primer annealing program
Temperature (°C)	Time
95	5 min
65	3 min
60	3 min
57	3 min
54	3 min
51	3 min
30	Forever

Table 7: Primer Extension reaction master mix and program.

Primer extension mix
Reagent	1x (µL)	[Final]
10x phi29 DNA polymerase buffer	2	1x
10 U/µL phi29 DNA polymerase	1	75 µM
Reaction mix volume	3
Primer extension program
Temperature (°C)	Time
30	20 min
65	10 min
4	Forever

Table 8: dA-Tailing reaction master mix and program.

dA-Tailing mix
Reagent	1x (µL)	[Final]
3 mM dATP	0.7	120 µM
Klenow fragment buffer	2.4	1x
5 U/µL Klenow fragment	1	5 U
Reaction mix volume	4.1
dA-Tailing program
Temperature (°C)	Time
37	30 min
75	20 min
4	Forever

Table 9: Universal Adapter Ligation reaction master mix.

Reagent	1x (µL)	[Final]
Autoclaved water	5
15 µM Universal adapter*	1	320 nM
Ligation enhancer	0.5	1x
Ligase master mix	15	1x
Reaction mix volume	21.5
*Universal adapter DNA sequence is described in Table 1.

Table 10: Ligation-Mediated PCR master mix and program.

LM-PCR mix
Reagent	1x (µL)	[Final]
10 µM Index PCR primer*	2	0.2 µM
10 µM Universal PCR primer*	2	0.2 µM
2x Taq DNA polymerase PCR master mix	25	1x
Reaction mix volume	29
*PCR primer sequences are described in Table 1.
LM-PCR program
Temperature (°C)	Time	Cycles
98	30 s	1
98	10 s	15−25
65	75 s
65	5 min	1
4	Hold

Discussion

In this protocol, ChIP followed by exonuclease digestion is used to obtain DNA libraries for the identification of protein-DNA interactions in mammalian cells at ultra-high mapping resolution. Many variables contribute to the quality of the ChIP-exo experiment. Critical experimental parameters include the quality of antibodies, optimization of sonication, and the number of LM-PCR cycles. These critical experimental parameters are also what can limit ChIP-exo experiments and will be discussed below.

In any ChIP protocol, antibody quality is one of the most important considerations. The use of ChIP-grade antibodies is recommended for ChIP-exo. The antibody quality can be checked prior to conducting the ChIP-exo protocol. ChIP-seq or ChIP-qPCR can be used to validate ChIP-grade antibodies by confirming specific DNA binding locations of the protein of interest. Subsequently, optimizing the concentration of antibodies added to the beads is ideal to ensure that protein-DNA complexes are immunoprecipitated^22,23.

The sonication of chromatin is another critical step in the ChIP-exo protocol. Sonication is critical for the non-specific shearing of chromatin, necessary for optimal immunoprecipitation to the protein of interest²⁴. The size and amount of sonicated DNA are dependent on the number of sonication cycles. A high concentration of fragmented DNA is important to ensure that enough DNA is immunoprecipitated to the protein of interest for a ChIP-exo DNA library to be created. Obtaining 100−500 bp fragmented DNA is ideal because the resolution of ChIP-exo is better if the sonicated chromatin fragments have smaller starting sizes. Therefore, the optimal  chromatin sonication should be determined for each type and batch of cells by varying the number of sonication cycles and sonication buffers.

The final critical parameter is obtaining the amplification of ChIP-exo library DNA with the minimal number of LM-PCR cycles to avoid over-amplification of PCR artifacts. To determine the minimal number of PCR cycles to amplify ChIP-exo DNA, a small portion of ChIP-exo DNA can be used to run multiple PCR cycles (for example, 10, 15, and 20 cycles) and compare the PCR products.

ChIP-exo has many more steps than a traditional ChIP-seq and, as a result, each step should be performed carefully and precisely for the ChIP-exo protocol to work. Importantly, the correct volume of every reagent in enzymatic reactions must be added to each reaction master mix²¹. Thus, creating a spreadsheet to calculate the volume of each reagent required for each master mix, then printing the tables and checking each reagent after addition to the master mixes is recommended. Careful excision of the amplified DNA is also important; fragments below 200 bp often contain adapter dimers, which need to be removed before next-generation sequencing.

Notably, most of the critical parameters and limitations of ChIP-exo are identical to those of ChIP-seq. However, unlike ChIP-seq, ChIP-exo delivers high-resolution genome mapping with low background. Thus, ChIP-exo is the ideal method for identifying precise protein-DNA interactions in a variety of systems and organisms, helping to reveal the significant roles of DNA-binding proteins in the cell. The ChIP-exo method described above can be used to examine transcription factors, histone modification marks and chromatin regulatory proteins in living cells at higher mapping resolution than ChIP-seq^25,26,27. In addition, ChIP-exo can detect the individual binding locations of DNA-binding proteins within a cluster, while ChIP-seq cannot due to its mapping resolution. Importantly, ChIP-exo displays a higher signal-to-noise ratio compared to ChIP-seq. These advantages of ChIP-exo allow us to identify a comprehensive set of bound locations across the genome. This protocol will provide a foundation for researchers interested in examining DNA-binding locations of various proteins across the genome at near base-pair resolution.

Materials

Name	Company	Catalog Number	Comments
Agarose, UltraPure	Invitrogen	16500	Checking Sonication (Section 4.3.6) and Gel Purification of LM-PCR Amplified DNA (Section 19.1)
Albumin, Bovine Serum (BSA), Protease Free, Heat Shock Isolation, Min. 98%	BioShop	ALB003	Blocking Solution
Antibody against Isl1	DSHB	39.3F7	Antibody incuation with beads (Section 5.6)
Bioruptor Pico	Diagenode	B01060010	Sonicating Chromatin (Section 3.3)
Bovine serum albumin (BSA), Molecular Biology Grade	New England BioLabs	B9000S	Fill-in Reaction on Beads (Section 10.2) and Denaturing, Primer Annealing and Primer Extension (Section 14.2)
Centrifuge 5424 R	Eppendorf	5404000138	Sonicating Chromatin (Section 3.5), Checking Sonication (Section 4.3.1, 4.3.3 and 4.3.4)
Centrifuge 5804 R	Eppendorf	22623508	Harvest, cross-linking and freezing cells (Section 1.3), Cell lysis (Section 2.2 and 2.3), Sonicating Chromatin (Section 3.1)
cOmplete, Mini, EDTA-free Protease Inhibitor Cocktail	Roche	4693159001	Added to all buffers, except Proteinase K buffer and ChIP Elution buffer
dATP Solution	New England BioLabs	N0440S	dA-Tailing Reaction (Section 15.1)
Deoxycholic Acid Sodium Salt	fisher scientific	BP349	Lysis Buffer 3, High Salt Wash Buffer and LiCl Wash Buffer
dNTP Mix, Molecular Biology Grade	Thermo Scientific	R0192	Fill-in Reaction on Beads (Section 10.2) and Denaturing, Primer Annealing and Primer Extension (Section 14.2)
DreamTaq Green PCR Master Mix, 2x	Thermo Scientific	K1081	Ligation-Mediated PCR (Section 18.1)
EDTA, 0.5 M, Sterile Solution, pH 8.0	BioShop	EDT111	Lysis Buffer 1-3, Checking Sonication (Section 4.1), High Salt Wash Buffer, LiCl Wash Buffer and ChIP Elution Buffer
Ethyl Alcohol Anhydrous, 100%	Commercial alcohols	P006EAAN	Checking Sonication (Section 4.3.3)
Formaldehyde, 36.5-38%, contains 10-15% methanol	Sigma	F8775	Harvest, cross-linking and freezing cells (Section 1.1)
Glycerol, Reagent Grade, min 99.5%	BioShop	GLY002	Lysis Buffer 1
Glycine, Biotechnology Grade, min. 99%	BioShop	GLN001	Harvest, cross-linking and freezing cells (Section 1.2)
Glycogen, RNA Grade	Thermo Scientific	R0551	Checking Sonication (Section 4.3.3)
HEPES, 1 M Sterile-filtered Solution, pH 7.3	BioShop	HEP003	Lysis Buffer 1, High Salt Wash Buffer
Klenow Fragment (3->5 exo-)	New England BioLabs	M0212S	dA-Tailing Reaction (Section 15.1)
Lambda exonuclease	New England BioLabs	M0262S	Lambda Exonuclease Digestion on Beads (Section 11.2)
Ligase Enhancer	New England BioLabs	E7645S	NEBNext Ultra II DNA Library Kit. Index Adapter Ligation on Beads (Section 9.1) and Universal Adapter Ligation (Section 16.1)
Ligase Master Mix	New England BioLabs	E7645S	NEBNext Ultra II DNA Library Kit. Index Adapter Ligation on Beads (Section 9.1) and Universal Adapter Ligation (Section 16.1)
Lithium Chloride (LiCl), Reagent grade	Bioshop	LIT704	LiCl Wash Buffer
Magnetic beads for ChIP (Dynabeads Protein G)	Dynabeads Protein G (magnetic beads for ChIP)	Dynabeads Protein G (magnetic beads for ChIP)	Antibody incubation with beads (Section 5)
Magnetic beads for DNA purification (AMPure XP Beads)	Beckman Coulter	A63880	DNA Extraction (Section 13.3) and DNA Clean-up (Section 17.1)
Magnetic rack (DynaMag-2 Magnet)	Invitrogen	12321D	Used in many steps in Sections: 5 - 11, 13
MinElute Gel Extraction Kit	Qiagen	28604	Gel Purification of PCR Amplified DNA (Section 19.2)
N-Lauroylsarcosine sodium salt solution, 30% aqueous solution, ≥97.0% (HPLC)	Sigma	61747	Lysis Buffer 3
Octylphenol Ethoxylate (IGEPAL CA630)	BioShop	NON999	Lysis Buffer 1 and LiCl Wash Buffer
Phenol:Chloroform:Isoamyl Alcohol, Biotechnology Grade (25:24:1)	BioShop	PHE512	Checking Sonication (Section 4.3.1)
phi29 DNA Polymerase	New England BioLabs	M0269L	Fill-in Reaction on Beads (Section 10.2) and Denaturing, Primer Annealing and Primer Extension (Section 14.3 and 14.4)
Phosphate-Buffered Saline (PBS), 1x	Corning	21040CV	Harvest, cross-linking and freezing cells (Section 1.3) and Sonicating Chromatin (Section 3.1), Antibody incuation with beads (Section 5.1)
PowerPac Basic Power Supply	BioRad	1645050	Checking Sonication (Section 4.3.6) and Gel Purification of LM-PCR Amplified DNA (Section 19.1)
ProFlex PCR System	Applied Biosystems	ProFlex PCR System	Used in Sections: 14.2, 14.4, 15.2, 16.2 and 18.2
Protein LoBind Tube, 2.0 mL	Eppendorf	22431102	Antibody Incubation with Beads (Section 5.2) and Chromatin Immunoprecipitation (Section 6.3)
Proteinase K Solution, RNA Grade	Invitrogen	25530049	Checking Sonication (Section 4.2) and Elution and Reverse Crosslinking (Section 12.3)
Qubit 4.0 Fluorometer	Invitrogen	Q33226	Gel Purification of PCR Amplified DNA (Section 19.3)
Quibit dsDNA BR assay kit	Invitrogen	Q32853	Gel Purification of PCR Amplified DNA (Section 19.3)
Rnase A, Dnase and Protease-free	Thermo Scientific	EN0531	Checking Sonication (Section 4.3.2) and DNA Extraction (Section 13.2)
Sodium chloride (NaCl), BioReagent	Sigma	S5886	Lysis Buffer 1-3, High Salt Wash Buffer
Sodium Dodecyl Sulfate (SDS), Electrophoresis Grade	BioShop	SDS001	Checking Sonication (Section 4.1), High Salt Wash Buffer and ChIP Elution Buffer
Sonication beads and 15 mL Bioruptor Tubes	Diagenode	C01020031	Sonicating Chromatin (Section 3.1 and 3.2)
ThermoMixer F1.5	Eppendorf	5384000020	Section 4.2, 4.3.2, 4.3.5, 8.2, 9.2, 10.3, 11.2, 12.2, 12.3, 13.2 and 16.2
Trizma hydrochloride solution (Tris-HCl), BioPerformance Certified, 1 M, pH 7.4	Sigma	T2194	10 mM Tris-HCl Buffer
Trizma hydrochloride solution (Tris-HCl), BioPerformance Certified, 1 M, pH 8.0	Sigma	T2694	Lysis Buffer 2, Lysis Buffer 3, Checking Sonication (Section 4.1), LiCl Wash Buffer and ChIP Elution Buffer
Ultra II End Repair/dA-Tailing Module (24 rxn -> 120 rxn)	New England BioLabs	E7546S	End Prep Reaction mix and End Prep Enzyme mix. End-repair and dA-Tailing Reaction on Beads (Section 8.2)
VWR Mini Tube Rocker, Variable Speed	VWR	10159-752	Used in many steps in sections: 1, 2, 5, 6 and 7
2-[4-(2,4,4-trimethylpentan-2-yl)phenoxy]ethanol, Triton X-100, Reagent Grade	BioShop	TRX506	Lysis Buffer 1, Sonicating Chromatin (Section 3.4) and High Salt Wash Buffer