Genome-wide profiling of histone modifications in Plasmodium falciparum using CUT&RUN

We present a detailed step-wise protocol for reliable genome-wide profiling of histone modifications using CUT&RUN that requires only low numbers of Plasmodium falciparum blood stages.


Introduction
Post-translational histone modifications play a central role in regulating chromatin accessibility and partitioning in nearly all eukaryotes. Numerous modifications have been described with a wide variety of functions and dynamics reviewed in Rothbart and Strahl [2014]. The ability to profile these modifications genomewide is a key tool for assigning their role in regulating gene expression and, more generally, for investigating nuclear functions. Traditionally, this has been accomplished by chromatin immunoprecipitation (ChIP) (Solomon et al, 1988), which relies on affinitybased pull-down from randomly fragmented chromatin of protein-DNA complexes with antibodies specific to one of the complex components followed by the extraction of the genomic fragments. High-throughput sequencing (HTS) methods are then used to determine the relative population abundance of the antibody target at each genomic position based on its relative enrichment in the precipitated chromatin versus its genomic frequency.
ChIP-seq has several drawbacks that derive from the random fragmentation of chromatin and the requirement for stringency during immunoprecipitation. The former results in relatively high background, although both make profiling low-affinity interactions difficult. This can be overcome by protein-DNA cross-linking before fragmentation. However, cross-linking presents another challenge as it can lead to non-specific recovery of untargeted proteins correlated with duration of fixation treatment (Baranello et al, 2016) and interfere with antibody recognition by epitope disruption (O'Neill & Turner, 2003).
The recent development of Cleavage Under Targets & Release Using Nuclease (CUT&RUN) addresses many of these challenges (Fig 1) (Skene et al, 2018;Meers et al, 2019). Instead of isolating and randomly fragmenting chromatin, cells are immobilized on beads and permeabilized to allow antibodies to diffuse into the nucleus to bind their targets in situ. After removal of excess antibodies, a fusion of protein A/G, which binds the Fc portion of antibodies, and micrococcal nuclease (MNase) is added under low calcium conditions that keep the MNase inactive. After the removal of excess protein A/G-MNase, calcium is added to activate MNase cleavage, thereby releasing the antibody-bound nucleosomes but leaving the unbound fraction of the genome intact. Because of their small size, liberated nucleosomes can then diffuse out of the cell, whereas the rest of the genome remains cell associated. DNA can then be extracted from the supernatant for standard HTS library generation, high-throughput sequencing, and analysis. Because of this, CUT&RUN has exceptionally low background and has been successfully used to profile chromatin interactions from very small number of cells (Liu et al, 2018;Skene et al, 2018). CUT&RUN has been successfully used for profiling a wide array of chromatin proteins in many experimental systems. In this study, we specifically focus on its use for profiling post-translational histone modifications on nucleosome.
We recently adapted a CUT&RUN protocol (Meers et al, 2019;Janssens & Henikoff, 2019 Preprint) for genome-wide profiling of chromatin modifications in the human malaria parasite Plasmodium falciparum (Harris et al, 2022 Preprint). Using the step-by-step protocol described below, we were able to generate high-quality profiles of multiple histone modifications using only a small fraction of the cells required for ChIP-seq. Using antibodies against two commonly profiled histone modifications, H3K4me3 and H3K9me3, we show here that CUT&RUN profiling is highly reproducible (Figs 2 and S1) and closely recapitulates previously published ChIP-seq-based abundance profiles of histone marks (Figs 3, S1, and S2) (Bártfai et al, 2010;Bunnik et al, 2018;Fraschka et al, 2018). Finally, we show that CUT&RUN requires substantially lower sequencing coverage for accurate profiling compared with ChIPseq (Fig 4).

Planning and preparation
Prepare stock solutions Timing: 4 h 1 M HEPES-NaOH, pH 7.5: dissolve 238.3 g HEPES in 800 ml of distilled water. Adjust pH to 7.5 using 1 M NaOH, then adjust the final volume with distilled water to 1 liter. Store at RT. 1 M HEPES-KOH: dissolve 238.3 g HEPES in 800 ml of distilled water. Adjust pH to 7.5 using 1 M KOH, then adjust the final volume with distilled water to 1 liter. Store at RT. 1 M KCl: dissolve 74.55 g KCl in 800 ml of distilled water and adjust the volume to 1 liter. Store at RT. 5% sorbitol: dissolve 50 g sorbitol in 800 ml of distilled water and adjust the volume to 1 liter. Filter, sterilize, and store at RT. 90% percoll with 6% sorbitol: mix 20 ml of minimal RPMI with 180 ml percoll and dissolve 12 g D-sorbitol. Filter, sterilize, and store at 4°C. 100 mM CaCl 2 : dissolve 1.1 g CaCl 2 in 100 ml of distilled water. 100 mM MnCl 2 : dissolve 1.25 g MnCl 2 in 100 ml of distilled water. 100 mM EGTA, pH 8.0: dissolve 3.8 g EGTA in 80 ml of distilled water while maintaining the pH at 8.0 using 10 N NaOH. Adjust volume to 100 ml and store at RT. 500 mM EDTA, pH 8.0: dissolve 186.1 g EDTA in 800 ml of distilled water while maintaining the pH at 8.0 using 10 N NaOH. Adjust volume to 1 liter and store at RT. 1 M Tris-HCL, pH 8.0: dissolve 121.14 g Tris in 800 ml of distilled water. Adjust pH to 8.0 using concentrated HCl/10 N Figure 1. Overview of genome-wide profiling of histone modifications using CUT&RUN. (A) Infected RBCs are immobilized on concanavalin-coated beads. (B) Immobilized cells are permeabilized, allowing histone modification-specific antibody to diffuse into the nucleus to bind to its targets. (C) After removal of excess antibodies, MNase fused to protein A/G is added under low calcium conditions and binds to chromatin-associated antibodies. (D) After removal of excess proteinAG-MNase, cleavage adjacent to bound nucleosomes is activated by addition of calcium-cations. (E, F) Liberated nucleosomes are allowed to diffuse out of the permeabilized cells for subsequent (F) DNA extraction, library generation, high-throughput sequencing, and analysis.
NaOH. Adjust the final volume with distilled water to 1 liter and store at RT. 5 M NaCl: dissolve 292.2 g NaCl in 800 ml distilled water. Adjust the volume to 1 liter and store at RT. 5% digitonin: dissolve 0.05 g of digitonin in 1 ml of DMSO. Prepare 150 µl aliquots and store at −20°C for up to 6 mo.
25× Protease Inhibitor Cocktail: dissolve one EDTA-free cOmplete protease inhibitor tablet in 1 ml of distilled water. Split into 200 µl aliquots and store at −20°C for up to 3 mo.
2 M spermidine: prepare a fresh work solution each time by diluting 31.25 µl of stock spermidine (6.4 M) in 100 µl of distilled water.
P. falciparum culturing and synchronization of erythrocytic stages Timing:~1 wk We have found that CUT&RUN using parasites containing 10-50 million nuclei yielded robust signal for the antibodies tested. At a parasitemia and hematocrit of 5% each, this is the equivalent of 0.4-2 ml of culture for 1 N parasites (ring, early trophozoite, and gametocyte stages) or as little as 20-100 µl of culture containing segmenting schizonts.
(1) Maintain parasite cultures in T25 flasks following established culturing techniques (Moll et al, 2013) using red blood cells at 3-5% hematocrit resuspended in standard malaria complete media.

Solutions and buffers
Store at 4°C for up to 6 mo. Note that the binding buffer uses HEPES-KOH! Store at 4°C for up to 1 wk. Critical: Protease Inhibitor Cocktail (EDTA-free) should be added fresh on the day of use.
Add 20 µl of Protease Inhibitor Cocktail stock solution per 1 ml of wash buffer.
Store on ice or at 4°C for up to 1 d. Critical: mix well before each use. Note: digitonin has batch variability, and the minimal concentration that yields full of permeabilization needs to be determined empirically for your digitonin stock. We have obtained good results with as low as 0.025% digitonin. To test for permeabilization, pellet 150 µl of culture and resuspend in 150 µl Dig-wash buffer containing 8 µM Hoechst 33342 and 10 µg/ml propidium iodide. Incubate cell suspension for 10 min at RT and wash once. Pellet and resuspend cells in 20 µl of media and observe under fluorescence microscope or flow cytometer. Determine the lowest digitonin concentration at which all nuclei are positive for both dyes (Fig 5).
Always prepare fresh before use. Critical: add the amount of antibody indicated right before incubation and keep on ice at all times.
Store at 4°C for up to 1 wk. Critical: Protease Inhibitor Cocktail (EDTA-free) should be added fresh on the day of use.
Add 20 µl of 25× stock solution per 1 ml of buffer.
Store at 4°C for up to 1 wk. Critical: Protease Inhibitor Cocktail (EDTA-free) should be added fresh on the day of use.
Add 20 µl of 25× stock solution per 1 ml of buffer.
Store at 4°C for up to 1 wk. Step-by-step instructions Activation of concanavalin A-coated beads Timing: 5 min Concanavalin A lectins on magnetic beads are activated and stabilized by Ca 2+ and Mn 2+ cations.
(1) Gently resuspend and withdraw 28 µl of the ConA bead slurry per sample/antibody combination.
(2) Transfer beads to a 1.5-ml microcentrifuge tube and add 1 ml of binding buffer. (3) Place the tube on a magnetic bead separator stand and allow to clear (30-120 s) before removing and discarding the supernatant. (4) Remove tube from stand and add 1 ml of binding buffer. Mix gently and return to the magnetic stand. (5) Repeat steps 3-4 one additional time.
(6) Resuspend beads in the same volume of binding buffer that was added in step 1 and keep at RT. (7) Place wash buffer at RT to warm for later use. TE buffer.

Binding cells to concanavalin A-coated beads
Timing: 1 h Infected red blood cells (iRBCs) bind to beads coated with concanavalin A, a lectin that binds specifically to extracellular glycoproteins.
(8) Harvest fresh parasite cultures at the desired stage and synchrony. If feasible, enrich for iRBCs by magnetic purification (Mata-Cantero et al, 2014), centrifugation in a continuous percoll gradient (Rivadeneira et al, 1983), or lysis of uninfected RBCs (Brown et al, 2020). The stage, synchrony, and growth conditions are determined by the scientific question the experimenter is asking. In this study, we used percoll/sorbitol isolated asexual blood stages synchronized to a window of 36 ± 4 h post-invasion. (9) Determine iRBC density and purity using a hemocytometer, coulter counter, or volumetric flow cytometry. (10) Pellet cells by centrifugation (800g for 3 min at RT) and resuspend cells containing 1-5 × 10 7 nuclei to a cell density to 1 × 10 7 cells/ml in wash buffer by gentle pipetting.
Note: the average nuclear content of parasites can be determined by flow cytometry after staining cultures with 8 µM Hoechst 33342 (DNA specific) and 0.1 µg/ml thiazole orange (RNA specific with Hoechst 33342) for 30 min at 37°C, followed by two brief washes. A high ring-stage culture can be used to determine the mean fluorescence value equivalent to a 1 N nuclear content (Fig 6). However, we did not attempt to optimize the minimal number of nuclei required. Based on CUT&RUN in other systems and our down-sampling analysis (Fig 4), substantially fewer nuclei are likely sufficient.
Optional: at this point, it is possible to cryopreserve infected RBCs for later processing. Harvest cells by centrifugation (800g for 3 min at RT) and resuspend in 1 ml of 10% DMSO prepared in 1× PBS and freeze slowly to −80 o C using an isopropanol freezing chamber to minimize cellular damage. Store frozen cells at −80°C until you are ready to use the cells, then thaw them quickly by placing cryovials in a 37°C water bath for 3-5 min. Then transfer the cell suspension to a 1.5-ml microcentrifuge tube and spin at 800g for 3 min at RT. Wash twice with 1× PBS before proceeding to step 13.
Critical: from here on, mix gently and avoid vortexing.
(11) Split each sample into 1 ml aliquots (one per antibody including isotype controls) in 1.5-ml LoBind DNA microcentrifuge tubes and wash twice with 1 ml of wash buffer (800g for 3 min at RT).  (14) Incubate at RT on rotator for 10 min at an orbital speed of 10 rpm.
Cell permeabilization and antibody binding Timing: 2.5 h Cells attached to conA beads are permeabilized to allow antibodies to diffuse into the nucleus and bind their targets.
(15) For each antibody, make an antibody master mix by combining antibody buffer with 0.8 µg of antibody to a final volume of 165 µl per sample. For antibodies other than those used in this protocol, final concentrations may have to be determined empirically based on affinity and target abundance. (16) Ensure bead-bound cells are in homogenous suspension by inverting the tube several times, then divide bead-bound cells into aliquots of 250 µl in 1.5-ml microcentrifuge LoBbind DNA tubes.
Note: if antibodies from multiple species are used, include the appropriate isotype control for each species. The amount of antibody needed for optimal signal depends on multiple factors, including its affinity for the target, the abundance of the target, and the number of cells used. In our experience, high background is more often caused by using too much antibody rather than too little.
(17) Place tubes on a magnetic stand to clear, then carefully remove and discard supernatant. (18) Remove the tube from the stand and, while holding the tube at an angle, add 150 µl of antibody master mix to the side of the tube opposite from the beads. Then tap gently to dislodge the beads. (19) Incubate on a rotator for 2 h at an orbital speed of 10 rpm at RT.
Pause point: antibody incubation may proceed overnight at 4°C.

Secondary antibody binding (optional)
Timing: 1.5 h Note: the affinity of protein A/G to IgG antibodies varies based on host species and IgG subtype (www.bio-rad.com/ en-us/applications-technologies/protein-g-affinity). If mouse primary antibodies are being used, it is recommended to use a secondary rabbit-anti-mouse antibody to increase binding to protein A. Although the version of pA/G-MNase sold by Epicypher has an improved antibody compatibility relative to protein A alone, Jenssen and Henikoff suggest that rabbit anti-mouse secondary antibody should still be used in some cases (Janssens & Henikoff, 2019 Preprint). Otherwise, proceed to step 26.
(20) For each mouse antibody, make a secondary antibody master mix by combining antibody buffer with 0.8 µg of rabbit antimouse IgG to a final volume of 165 µl per sample with mouse primary antibody. Remove liquid from cap and side with a quick pulse on a micro-centrifuge, then place on the magnetic stand to clear then remove and discard the supernatant.  DNA purification by phenol/chloroform extraction Timing: 2 h After the release and diffusion of the nucleosome fragments into the supernatant, DNA is purified and subjected to quality control. Note: the pellet may not be visible at this point, so be careful to not discard the pellet.

Chromatin digestion and release
(46) Carefully pour off the liquid and dry on a paper towel. Let tube air dry on its side for 5 min. (47) When the pellet is dry, dissolve in 30 µl TE buffer. (48) Take 2 µl of each sample to quantify DNA concentration by Qubit Fluorometer using the Qubit dsDNA HS (High Sensitivity) assay kit. In our experience, concentrations at this step are nearly always below 1 ng/µl and have little predictive value of downstream success. (49) Keep the samples on ice at all times.
Note: although samples can be stored at −20°C, it is strongly advisable to continue with the protocol until libraries have been PCR amplified.

End repair and A-tailing
Timing: 2 h The ends of the extracted DNA are repaired and A-tailed to allow subsequent HTS adapter ligation.
(50) Place AMPure XP Beads at RT for later use.
(51) In a 1.5-ml microcentrifuge tube, prepare the End repair master mix.
*Master mix recipes include 10% extra volume for pipetting error.
(52) Combine 5 µl of end repair master mix to PCR tubes with 4 ng of DNA template and add ddH 2 O to a total volume of 30 µl. Mix by vortexing and quick spin to collect at the bottom of the tube. (53) Place the tube in a PCR machine with the heated lid set to ≥75°C and run the following program.
Note: compared with manufacturer instructions, the A-tailing temperature was decreased to 50°C to avoid DNA melting of shorter DNA fragments and the reaction time was increased to 60 min to compensate for lower enzymatic efficiency.

Ligation of barcoded HTS adapters
Timing: 25 min Dual barcode HTS adapters are added to the endrepaired and A-tailed fragments.
(54) Remove tube from PCR machine and place on ice. (55) In a 1.5-ml microcentrifuge tube, combine the following reagents.
Note: we use adapters from NEXTFLEX Unique Dual Index Barcodes (25 µM concentration) diluted 1:500 in ddH 2 O (0.05 µM final stock concentration). Because adapters are supplied in duplex form, avoid heating above RT. Dual index barcodes should be used when sequencing on Illumina systems that use patterned flow cells. If in doubt, use dual index barcodes.
(56) Add 17 µl of Adapter ligation master mix (final volume of 47 µl) and mix by vortexing. Quick spin to collect liquid at the bottom of the tube.

Post-ligation clean-up
Timing: 20 min Excess HTS adapters are removed.
Note: before use, equilibrate AMPure XP beads to RT for at least 30 min. Library PCR amplification Timing: 35 min Fragments are PCR amplified using KAPA polymerase, which is more efficient at amplifying AT-rich fragments.
Note: we use NEXTFLEX Primer Mix (12.5 µM stock concentration) that comes with NEXTFLEX Unique Dual Index barcodes sets, at a final concentration of 1.2 µM.

Expected outcomes
When starting with 1 × 10 7 schizont stages with an average nuclear content of 3-5 N, post-clean-up library yield was in the range of 10-40 ng after 15 cycles of PCR. Fig 2 shows examples of CUT&RUN library size profiles obtained with anti-H3K9me3 and anti-H3K4me3 antibodies. The ratio of primer dimers and adapters in our libraries was around 4%, which is 5× lower than normally reported for these assays. Library size profile obtained with rabbit IgG antibody represents unspecific bound regions that have been PCR amplified. Also, a PCR negative control with no template DNA was shown to dismiss this sort of contamination during library preparation.

Quantification and statistical analysis
The libraries shown in these examples were pooled with other CUT&RUN libraries at equimolar ratios and sequenced using an Illumina NextSeq2000 sequencer in paired-end 50 bp P2 mode. We initially aimed at a target of 10 million PE50 read pairs per library to ensure good coverage. Subsequent down-sampling analysis showed that as low as 1 million read pairs provide excellent signalto-noise ratios for these antibodies (Fig 4).
For the comparison of the CUT&RUN profiles to published profiles (Fig 3), we obtained raw read files from public repositories using SRAtools (https://github.com/ncbi/sra-tools). H3K9me3 and input raw ChIP-seq reads from Bunnik et al (2018) were downloaded from NCBI Sequence Read Archive (accession number SRP091939). PfHP1 and input raw ChIP-seq reads from Fraschka et al (2018) were downloaded from NCBI's Gene Expression Omnibus (GEO) (accession number GSE102695). H3K4me3 and input raw ChIP-seq reads from Bártfai et al (2010) were downloaded from NCBI's GEO (accession number GSE210062).
Raw reads were quality trimmed using Trimmomatic v0.38 (Bolger et al, 2014) to remove residual adapter sequences and lowquality leading and trailing bases to retain properly paired reads of length ≥ 30 bases after trimming, as previously published (Campelo Morillo et al, 2022). Trimmed reads were aligned to PlasmoDB version 51 P. falciparum 3D7 reference genome (Warrenfeltz et al, 2018) using BWA v.0.7.1 ). SAMtools v.1.10 ) was used to sort and index BAM files. Normalized fold enrichment tracks were generated by using the MACS2 v.2.2.7.1 (Zhang et al, 2008) callpeak function with settings: -f BAMPE -B -g 2.3 × 10 7 -q 0.05 -nomodel-broad-keep-dup auto-max gap 500. BedGraph outputs were then passed into the bdgcmp function with the setting -m FE (fold enrichment) to generate signal tracks to profile histone modification enrichment levels compared with either input (for ChIP-seq) or isotype IgG control (CUT&RUN). Using the bed-GraphToBigWig tool (Kent et al, 2010), we converted the genomewide coverage and enrichment tracks from bedGraph into BigWig format. These files are available for download at NCBI GEO under accession number GSE210062 and can be directly loaded into the PlasmoDB genome browser (Warrenfeltz et al, 2018) for detailed inspection.
Coverage and enrichment tracks were visualized using Integrative Genome Viewer IGV (Thorvaldsdóttir et al, 2013) or the GenomicRanges v.1.44.0 and GViz Viz v1.38.1 (Hahne & Ivanek, 2016) R packages within the Bioconductor project (release 3.13) (Huber et al, 2015). The full analysis pipeline can be found at https:// github.com/KafsackLab/PfCUTandRUN. Representative tracks of H3K9me3, H3K4me3, and HP1 produced using CUT&RUN or ChIP-seq at chromosome 8 of P. falciparum are shown in Fig 4. Troubleshooting Problem 1: low DNA concentration after library amplification If DNA is low or undetectable in step 84, check the following.
-Beads were not activated properly.
-Beads got stuck on tube lid or walls during incubation steps.
Potential solution Remember to activate beads before initiating the protocol. Give a quick spin after each incubation to recover beads from tube lid and walls. Try not to pipette too harshly because cells can be very fragile after permeabilization. Always perform the digestion at 0°C, and do not agitate the tube once high Ca 2+ incubation buffer is added to the isolated DNA from supernatants saved in step 36. No DNA should be detected in this fraction; if so, this is indicative of premature diffusion of nucleosomes during digestion. For low-abundance targets, increasing the cycle number during PCR library amplification may help.

Limitations
This protocol relies on the availability of a specific antibody against histones PTM that would be able to recognize correspondent modification in P. falciparum. This should be tested empirically by Western blot or immunoprecipitation assays. We have only used rabbit antibodies. Antibodies generated in other species should be tested according to experimental set up and availability, and antibodies raised in other species may benefit from the inclusion of rabbit secondary antibodies. This protocol was optimized for use with abundant chromatin marks on nucleosomes. For rare modifications, or non-nucleosomal targets, like transcription factors and histone modifiers, further optimization may be required.

Data Availability
The full analysis pipeline can be found at https://github.com/ KafsackLab/PfCUTandRUN. Raw (fastq) and processed (BigWig) high-throughput sequencing data have been deposited in the NCBI GEO under accession number GSE210062.