CUT&RUN Profiling of H3K27me3 in 100,000 Mouse Embryonic Stem Cells Using Targeted MNase Cleavage and Spike-In Normalization

Two states (naive 2i/LIF, primed EpiLC), biological triplicate (n=3), plus per-sample IgG controls = 9 CUT&RUN reactions minimum. 100,000 cells per reaction. Drosophila S2 chromatin (or heterologous spike-in DNA via the pAG-MNase reaction) added at fixed ratio for calibration. H3K4me3 included as a positive antibody control on a separate aliquot. Samples randomized across the bead-binding plate; library indices assigned by a randomization script. Two technical sequencing replicates pooled per library.

BSL-1 for mESC and S2/E. coli spike-in handling. Digitonin is toxic if ingested/inhaled — weigh in a fume hood, wear gloves and mask. Concanavalin A is a lectin (irritant). Phenol-chloroform (if used) is corrosive and toxic — handle in a fume hood, collect as halogenated organic waste; SPRI is the preferred lower-hazard alternative. Standard biohazard waste handling for cells and antibody reagents.

Negative control: normal rabbit IgG in place of primary antibody — should yield <10% of target read count and no domain enrichment. Positive antibody control: H3K4me3 — should produce sharp peaks at active TSS. Spike-in control: fixed E. coli (or Drosophila) DNA per reaction for cross-sample quantitative scaling. No-cell ConA-bead control to detect background DNA. A no-CaCl2 reaction confirms cleavage is calcium-dependent (should give minimal yield).

Fragment-size distribution showing a sub-nucleosomal peak (~120 bp) plus mono-nucleosomal (~150-170 bp) ladder. H3K27me3 broad domains (kilobases to megabases) over Hox clusters and developmental genes. FRiP (fraction of reads in peaks) >20% for a successful H3K27me3 experiment; IgG FRiP <2%. Spike-in normalization revealing a global H3K27me3 increase in primed cells. Library yield 5-50 ng from 100k cells.

• E14 mouse ESCs cultured in 2i/LIF (CHIR99021 3 µM, PD0325901 1 µM, LIF 1000 U/mL) on gelatin
• Concanavalin A-coated magnetic beads (Bangs Labs BP531 or EpiCypher 21-1401)
• Anti-H3K27me3 rabbit mAb (Cell Signaling C36B11 #9733), 1:50
• Anti-H3K4me3 (positive control) and normal rabbit IgG (negative control)
• pAG-MNase fusion (EpiCypher 15-1016), 2.5 µL per reaction
• Wash buffer: 20 mM HEPES pH 7.5, 150 mM NaCl, 0.5 mM spermidine, protease inhibitors
• Digitonin (0.01-0.05%, titrated for >95% permeabilization by trypan blue)
• Calcium chloride 100 mM; EGTA/EDTA STOP buffer with E. coli spike-in DNA (0.5 ng) and RNase A
• Proteinase K (20 mg/mL), phenol-chloroform or SPRI for DNA cleanup
• NEBNext Ultra II DNA Library Prep Kit (NEB E7645)

To generate high-signal-to-noise genome-wide maps of the repressive histone mark H3K27me3 from only 100,000 mESCs by tethering a protein A/G-MNase fusion to chromatin-bound antibody and releasing antibody-proximal fragments through targeted, calcium-activated cleavage. Because cleaved fragments diffuse out of permeabilized nuclei, background is dramatically lower than ChIP, enabling sequencing at 3-8 million reads per sample.

1. Harvest 100,000 cells per reaction, wash in wash buffer, bind to 10 µL activated ConA beads, 10 min RT.
2. Resuspend beads in 50 µL antibody buffer (wash + 0.01% digitonin + 2 mM EDTA) containing primary antibody (1:50); rotate 4°C overnight.
3. Wash twice in Dig-wash; add 50 µL Dig-wash + pAG-MNase; incubate 4°C 1 h.
4. Wash twice; place on ice, add 100 µL Dig-wash, then 2 µL 100 mM CaCl2 to activate cleavage; incubate 0°C (ice-water) 30 min.
5. Add 100 µL STOP buffer (with E. coli spike-in DNA); incubate 37°C 10 min to release fragments.
6. Place on magnet, collect supernatant; add Proteinase K + SDS to 0.1%, 50°C 30 min.
7. Purify released DNA by SPRI (1.8x) or phenol-chloroform; elute in 25 µL.
8. Prepare libraries (NEBNext Ultra II) with no size selection, using a low-temperature/short extension to preserve sub-nucleosomal fragments; 12-14 PCR cycles.
9. Bead clean-up (1.1x), Qubit + TapeStation QC; expect a sub-nucleosomal and mono-/di-nucleosomal ladder.

Independent variable: pluripotency state (naive 2i/LIF vs primed EpiLC). Dependent variables: spike-in-normalized H3K27me3 read density per genomic bin, Polycomb domain width, and peak number. Controlled variables: cell number (100,000), antibody dilution (1:50), pAG-MNase amount, cleavage time/temperature (30 min on ice-water), and spike-in mass per reaction.

We hypothesize that the transition from naive (2i/LIF) to primed (EpiLC) pluripotency redistributes H3K27me3, with broadening of Polycomb domains over developmental regulator promoters (e.g., Hox clusters, Gata6, Sox17) and net global gain in H3K27me3, detectable only with spike-in normalization because conventional read-depth scaling would mask a genome-wide level change.

Trim with Trim Galore; align to mm10 with Bowtie2 (--local --no-mixed --no-discordant --dovetail -I 10 -X 700). Align spike-in reads to the spike genome to compute a per-sample scale factor (reads-per-spike). Generate spike-normalized bigWigs (deepTools bamCoverage --scaleFactor). Call broad domains with SEACR (relaxed/stringent) using IgG as the background track, or MACS2 --broad. Quantify domain coverage and differential signal in DiffBind/csaw.

Low yield/no fragments: verify digitonin permeabilization (>95% trypan-positive nuclei) and confirm CaCl2 was added — re-titrate digitonin per cell line. High IgG background: increase wash stringency (NaCl to 150-300 mM) and reduce pAG-MNase incubation. Over-digestion (excess sub-nucleosomal smear, no nucleosome ladder): shorten cleavage to 20 min or keep strictly at 0°C. Weak H3K27me3 domains: antibody lot variability — validate with a SNAP-CUTANA spike-in nucleosome panel. Adaptor dimers: add a 0.9x SPRI re-clean and lower adaptor concentration.

Differential occupancy across consensus domains tested with csaw/edgeR using the spike-in-derived normalization factors as library sizes; quasi-likelihood F-test with n=3 per state. FDR controlled by Benjamini-Hochberg at q<0.05; require |log2FC|>1. Power: triplicate CUT&RUN at >20% FRiP reliably detects 2-fold domain changes. Report spike-normalized fold changes with 95% CIs and a sign test on global domain-level direction of change.