Macromolecular interactions dictate Polycomb-mediated epigenetic repression [ChIP-seq]

Dynamic regulation of epigenetic states relies on complex macromolecular interactions such as at the protein-protein and protein-DNA level. It remains an unsolved question how H3K27me3, the hallmark histone modification for facultative heterochromatin, and its writer Polycomb repressive complex 2 (PRC2) are spatiotemporally regulated during development. Here we engineered separation-of-function mutants to surgically dissect the roles of individual macromolecular interactions required for PRC2 function. We show that Polycomb-mediated silencing is precisely regulated by the dynamic interactions among the PRC2 core complex, its accessory proteins, DNA sequences, and other histone modifications. Combining CRISPR-mediated engineering of separation-of-function mutants, human stem cell differentiation models, next-generation sequencing approaches, and reconstituted biochemical assays, we identified distinct regulatory functions of these macromolecular interactions on epigenetic repression in human pluripotent stem cells and cardiac differentiation. Disruption of key interactions led to distinct and opposing effects on cardiomyocyte differentiation, suggesting their highly specified roles in cell fate determination. Together, these results reveal the importance of individual macromolecular interactions at the center of PRC2 controlling epigenetic repression. Two independent CRSIPR clones of each of the two edited cell lines (WT or mutant) were analyzed. For ChIP-seq, hiPSCs at 70–80% confluency were collected by Accutase treatment and crosslinked with 1% formaldehyde in PBS at room temperature for 10 min. The crosslinking reaction was quenched by adding one volume of 1.25 M glycine to 10 volumes of PBS for 2 min. Cells were washed in ice-cold PBS and cell pellets stored at -80°C. Aliquots of 10 million cells were lysed in 1 mL ChIP lysis buffer (800 mM NaCl, 25 mM Tris pH 7.5, 5 mM EDTA, 1% Triton X-100, 0.1% SDS, 0.5% sodium deoxycholate, 1X protease inhibitor cocktail) on ice for 30 min. Lysates were sonicated in MilliTubes (Covaris, 520135) with the Covaris S220 focused-ultrasonicator at 4°C for 10 min at a PIP of 140 W, a duty factor of 5%, and 200 cycles/burst. Lysate was cleared by centrifugation at 16,000xg for 10 min at 4°C, and supernatant was collected. For each IP, 1 mL chromatin dilution buffer (25 mM Tris pH 7.5, 5 mM EDTA, 1% Triton X-100, 0.1% SDS, 1X protease inhibitor cocktail) together with 20 ng spike-in chromatin from Drosophila melanogaster S2 cells (Active Motif, 53083) was added to the cleared lysate from 2.5-5.0 million cells. Chromatin samples were incubated with 2-4 μg of anti-CBX7 (4 μL), anti-FLAG (2 μL), anti-H2AK119ub (5 μL), anti-H3K27me3 (6 μL), anti-H3K36me3 (3 μL), anti-H3K4me3 (4 μL), anti-JARID2 (4 μL), anti-MTF2 (5 μL), anti-RING1B (7 μL), or anti-SUZ12 (4 μL) antibody at 4°C overnight. A spike-in antibody against the Drosophila-specific histone variant H2Av was added to all chromatin samples. Then, 25 μl of washed Pierce Protein A/G Magnetic Beads (Thermo Fisher, 88803) were added to the IP solution to incubate at room temperature for 2 h. Beads were washed twice with 1 mL of LSB buffer (20 mM Tris pH 8.0, 2 mM EDTA, 150 mM NaCl, 0.1% SDS, 1% Triton X-100), HSB buffer (20 mM Tris pH 8.0, 2 mM EDTA, 500 mM NaCl, 0.1% SDS, 1% Triton X-100), LiCl buffer (10 mM Tris pH 8.0, 1 mM EDTA, 250 mM LiCl, 1% sodium deoxycholate, 1% Igepal), and TE buffer (10 mM Tris pH 8.0, 1 mM EDTA). Chromatin was then eluted by incubating the beads with 120 μl of elution buffer (100 mM sodium bicarbonate, 1% SDS) for 20 min at room temperature. Formaldehyde crosslinking was reversed by adding 5 μL of 5 M NaCl and incubating at 65°C for 3 h. Protein and RNA were removed by adding 14 μL of 1M Tris pH 8, 3 μL of 500 mM EDTA, 3 μL of proteinase K (Thermo Fisher, 25530049), and 1 μL of 1 mg/mL RNase A (Thermo Fisher, EN0531) and incubating at 37°C for 1 h. DNA was extracted by phenol–chloroform–isoamyl extraction followed by ethanol precipitation. DNA samples were resuspended in 30 μl of 10 mM Tris pH 8.0 and quantified using a Qubit fluorometer (Invitrogen). Between 2 ng and 100 ng of each sample was used for library preparation using the KAPA HyperPlus Kit according to the manufacturer’s instructions, except for skipping the fragmentation step. DNA libraries were pooled and sequenced on a NovaSeq 6000 or NovaSeq X Plus using a 150-cycle kit (paired end, 2 x 150 bp). ChIP-seq analysis was performed similar to what was described previously.14 Adapter sequences were trimmed from the read pairs using Cutadapt (-a AGATCGGAAGAGCACACGTCTGAACTCCAGTCA -A AGATCGGAAGAGCGTCGTGTAGGGAAAGAGTGT). Trimmed reads were aligned to the human reference genome (hg38) using BWA (MEM). Properly paired read alignments with mapping scores greater than 30 were used for downstream analysis. PCR duplicates were removed using Picard MarkDuplicates (https://broadinstitute.github.io/picard/). Peak calling was performed with MACS215 with the options of ‘macs2 callpeak -g hs -f BAMPE --keep-dup all --bdg’. Bigwig files were generated using bedGraphToBigWig tool. Peak metaplots and heatmaps were generated on pooled replicates using deepTools. Gene-specific ChIP enrichment analysis was performed by comparing the number of reads mapped to gene bodies (defined by the featureCounts tool using the GENCODE v42 annotation) in mutant versus WT cell lines. For each biological replicate (individual clone), the read counts of input and pulldown were normalized by total number of aligned reads for the genes whose counts were great than one. Similar to what was previously described,14 a generalized linear model was constructed to estimate log2Foldchange of the normalized read counts in pulldown relative to input while adjusting for cell line variability. Shrunken fold change was estimated using the R package DESeq2. The statistical significance of the log2Foldchange was calculated empirically by the R package fdrtool, with the input of Wald statistics calculated using DESeq2. The differential ChIP enrichment between the mutant and WT lines was calculated with an interaction term of cell line (mutant versus WT) and library (pulldown versus input) in a regression model. Plots were generated using ggplot2 and Tableau software. Gene ontology analysis was performed using ShinyGO.