Inseparable RNA binding and chromatin modification activities of a nucleosome-interacting surface in EZH2

Polycomb repressive complex 2 (PRC2) interacts with RNA in cells, but there is no consensus on how RNA regulates PRC2 canonical functions, including chromatin modification and the maintenance of transcription programs in lineage-committed cells. We assayed two separation-of-function mutants of the PRC2 catalytic subunit EZH2, defective in RNA binding but functional in methyltransferase activity. We find that part of the RNA-binding surface of EZH2 is required for chromatin modification, yet this activity is independent of RNA. Mechanistically, the RNA-binding surface within EZH2 is required for chromatin modification in vitro and in cells, through interactions with nucleosomal DNA. Contrarily, an RNA-binding-defective mutant exhibited normal chromatin modification activity in vitro and in lineage-committed cells, accompanied by normal gene repression activity. Collectively, we show that part of the RNA-binding surface of EZH2, rather than the RNA-binding activity per se, is required for the histone methylation in vitro and in cells, through interactions with the substrate nucleosome.

For citations made in this section, see reference list in the main text.
For the immunoblotting of mESC, cells were harvested and washed three times by 1x PBS before lysed using high salt buffer (50 mM Tris-HCl pH7.2, 300 mM NaCl, 1 mM EDTA pH7.4,0.5% NP40, 1x protease inhibitors).Cells were then sonicated three times (10 sec on and 1 min off per cycle) with MSE Soniprep 150 Plus sonicator equipped with a 3mm Micro Tip.The lysates were then clarified at 20,000 RCF for 20 mins at 4 °C.Protein concentrations were determined using Bradford protein assay (Bio-rad # 5000006).

CUT&Tag
CUT&Tag experiments were performed according to the protocol that was previously described 31 .In short, nuclei were isolated from harvested cells and lightly crosslinked using formaldehyde to a final concentration of 0.1% prior to slowly being frozen using Mr. Frosty container (Thermo Fisher Scientific #5100-0001).An aliquot of 200,000 crosslinked nuclei was conjugated to 10 µL of activated Concanavalin A coated beads (Bangs laboratories #BP531) per sample.Subsequently, samples were incubated with 1 µL primary antibody (H3K27me3: Cell Signaling Technology #9733; H3K27ac: Abcam #4729; Rabbit IgG control: Cell Signaling Technology #2729) overnight at 4 °C in 50 µL antibody buffer (20 mM HEPES pH 7.5, 150 mM NaCl, 0.5 mM spermidine, 2 mM EDTA, 0.1% BSA, 1x protease inhibitors).Incubation of the secondary antibody (Guinea pig anti-rabbit antibody: Antibodies online #ABIN101961) was performed at RT for 1h in 100 µL Wash buffer (20 mM HEPES pH 7.5, 150 mM NaCl, 0.5 mM spermidine, 1x protease inhibitors).After three washes using Wash buffer, samples were incubated with pA-Tn5 adapter complex at RT for 1h in 100 µL 300-Wash buffer (20 mM HEPES pH 7.5, 300 mM NaCl, 0.5 mM spermidine, 1x protease inhibitors).Tagmentation was done at 37 °C for 1h in 300 µL Tagmentation buffer (300-Wash buffer supplemented with the final concentration of 10 mM MgCl2).Reversing the crosslinks was done by adding 10 µL of 0.5 M EDTA, 3 µL of 10% SDS and 2.5 µL of 20 mg/mL Proteinase K and incubating 1h at 55 °C.DNA was then extracted using Phenol/Chloroform and precipitated using 100% ethanol.PCR was performed using NEBNext HiFi 2x PCR Master Mix (NEB #M0541), with indexed primers that were described previously 49 and a total of 13 cycles.PCR products were purified using 1.3x volume of Ampure XP bead slurry and eluted in 25 µL of 10 mM Tris-HCl pH 8.0.DNA was sequenced using Illumina Hiseq PE 150 or Novaseq PE150 Sequencing.
The analysis of CUT&Tag data followed the code outlined in a published online protocol 50,51 with some modifications.First, the --local flag was used for Bowtie2 alignment in order to remove adapter sequences.Second, potential duplicates were removed from the BAM files using the MarkDuplicates function from the Picard toolkit (https://broadinstitute.github.io/picard/),with these BAM files were then used to calculate scaling factors for H3K27me3 and H3K27ac datasets using the ChIPseqSpikeInFree R package 52 .These scaling factors were used to generate bigwig files for heatmap, enrichment profiles, and genome browser tracks.Heatmaps and enrichment profiles were generated using the command-line tool deepTools 53 and genome tracks were created using pyGenomeTracks 54,55 .Correlation matrices were generated following the cited protocol and the plots were generated using the R package pheatmap (version 1: 726) 56 .
CUT&Tag peaks of H3K27me3 were calculated using SEACR 57 as previously described 50,51 .The consensus peaks of all 3 replicates were found using the findOverlapsOfPeaks() function within the ChIPpeakAnno R package 58 .This function was also used to find overlaps between mutants and the wild type.Peaks were associated to genes by annotating them to windows of 1000 bp downstream to 1000 bp upstream of transcription start sites (TSSs).Peaks were associated with differentially expressed genes using gene names.

RNA-seq
RNA was extracted from 1 million cells using RNeasy kit (QIAGEN #74004) as per manufacturer's instructions.cDNA libraries were constructed using NEBNext® Ultra TM RNA Library Prep Kit for Illumina®.cDNA libraries were pooled and sequenced using Illumina Novaseq instrument with a 2x150bp paired-end configuration.
RNA-seq reads were obtained in FASTQ format and were quantified against the hg38 transcriptome index using Salmon (version 0.14.1) with the --gcBias flag 59 .These transcriptlevel quantifications were then imported into R using the R packages DESeq2 60 and tximport 61 .Batch effect removal was done with limma, using the function removeBatchEffect() 62 .Next, this normalized data was visualized using a principal component analysis (PCA) plot using the DESeq2 function plotPCA() 60 .
Differential expression analysis was done using DESeq2 60 .Genes were considered differentially expressed if identified under a Benjamini-Hochberg adjusted p-value of 0.01.Gene ontology (GO) enrichment analysis and visualisation was done using the R package clusterProfiler 63 .GO terms were considered if obtained under a Benjamini-Hochberg adjusted p-value of 0.01.

Quantitative ChIP-Seq (ChIP-Rx)
K562 were collected, counted and washed once with PBS before crosslinking for 10 minutes with PBS containing 1% formaldehyde (Sigma).Cells were crosslinked in 15 mL falcon tubes at a density of ~5x10 6 cells/mL.Crosslinking was quenched with 0.125 M Glycine prior to two PBS washes.Crosslinked cells were then snap-frozen on liquid nitrogen and stored at -80 ºC before proceeding with the rest of the ChIP assay.Crosslinked K562 cells were thawed on ice and lysed in 4 mL of SDS-Lysis buffer (100 mM NaCl, 50 mM Tris-HCl pH 8.1, 5 mM EDTA pH 8.0, 0.5% SDS and 1X protease inhibitor cocktail (Sigma #P8340)).Chromatin/nuclei were pelleted by centrifugation at 1200 RPM for 6 mins at room temperature.The supernatant was then discarded, and the nuclear/chromatin pellet was resuspended in 2 mL of ChIP buffer, obtained by combining 1.33 ml of SDS-Lysis buffer with 0.67 ml of Triton Dilution buffer (100 mM Tris-HCl pH 8.6, 100mM NaCl, 5 mM EDTA pH 8.0, 5% Triton X-100, 1X protease inhibitor cocktail).
To allow for a quantitative ChIP-seq (ChIP-Rx) 64 , with respect to a reference genome, a total of 10% mESC (WT BRUC4 mESC) was added to each K562 chromatin lysate.Spike in percentage was calculated based on cell number and was added prior to sonication.Chromatin was sheared to approximately 200 bp to 500 bp fragments by sonication using a Bioruptor Plus (Diagenode) at high power.Total "on" sonication time was 15 min with 30 sec of "on" and 30 sec of "off" pulses.Sonicated chromatin was incubated overnight with the antibodies while rotating at 4 ºC.Prior to overnight incubation with antibodies, an input sample was taken (1%).10 million cells were used per ChIP.In the following morning, samples were clarified by centrifugation at 20,000 g for 20 min at 4 ºC.Following clarification, the chromatin was incubated for 3 hours with Protein G Dynabeads (ThermoFisher #10004D), where 50 µL beads were used per ChIP.After incubation, the beads were washed three times in Mixed Micelle Buffer (150 mM NaCl, 20 mM Tris-HCl pH 8.1, 5 mM EDTA pH 8.0, 5.2% Sucrose,1% Triton X-100, 0.2% SDS), twice with Buffer 500 (0.1% Sodium Deoxycholate, 1 mM EDTA pH 8.0, 50 mM HEPES pH 7.5, 1% TritonX-100), twice with LiCl detergent wash (0.5% Sodium Deoxycholate, 1 mM EDTA pH 8.0, 250 mM LiCl, 0.5% NP-40,10 mM Tris-HCl pH 8.0) and finally, one wash with TE.All washes were performed for 5 min with rotating at 4 ºC.Immunoprecipitated material was eluted from the beads with 100 µL elution buffer (0.1 M NaHCO3, 1% SDS) while shaking for 1 hour at 65 ºC.The supernatant was retained and incubated overnight at 65 ºC while shaking to reverse the crosslinks.The eluted material was then subjected to RNase A (DNase and protease-free Thermo Fisher Scientific #EN0531) and Proteinase K (ThermoFisher #EO0491) treatments prior to DNA clean up (QIAGEN MinElute PCR Purification Kit #28004).
Following the ChIP experiment, the precipitated DNA is already including the spike in reference (see above) and is quantified using the Qubit dsDNA High Sensitivity Assay Kit (ThermoFisher #Q32854).A total of 1-5 ng of DNA from each ChIP experiment was used for library preparation using the NEBNext Ultra II DNA Library Kit for Illumina (NEB #E7645) and NEBNext Multiplex Oligos for Illumina (NEB Set 1 #E7335 & Set 2 #E7500).Following adaptor ligation, DNA was PCR amplified for 5-10 cycles, depending on the amount of the input DNA.The DNA was purified using NEBNext Sample Purification Beads (NEB #E7767S).The quality and size distributions of DNA libraries were verified on a High Sensitivity D1000 Screen Tape (Agilent).The resulting libraries were then used for cluster generation and sequencing using an Illumina based sequencer with 150 bp paired-end read length (Azenta).
For the bioinformatic analysis of ChIP-Rx datasets, chromosome names for the mouse genome were modified with the prefix 'mm10_' and a metagenome was created by concatenating the human and mouse reference genomes (hg38 and mm10, respectively) before indexing with Bowtie2 65 as described previously 64 .Reads were aligned to the metagenome using Bowtie2 with default parameters 65 .Non-unique read alignments were filtered out using SAMtools 66 to exclude those with an alignment quality of <2, and the mm10_ prefix appended to chromosome names was used to separate reads as aligned to the reference human or spike-in mouse genomes.SAMtools was used to convert SAM files to BAM files and to remove duplicate aligned reads.Spike-in normalisation factors were calculated for each ChIP using the formula for normalised reference-adjusted reads per million (RRPM) as described 64 (1 per million spike-in reads).BigWig files were generated using the bamCoverage tool from the deepTools suite (version 3.3.0) 67with a bin size of 10 and data were subsequently visualised as ChIP-Rx normalised tracks using the IGV genome browser.Transcription start sites (TSS) for the hg38 human genome build were defined and annotated, to include TSS of all unique protein coding genes, including 5 kb upstream and downstream to them (n=17,139; downloaded from ENSEMBL biomart).Average enrichment profiles were generated with this TSS annotation file using the computeMatrix and plotProfile tools from the deepTools suite 67 .

Cellular fractionations
K562 cells were lysed in 400 μL pre-extraction buffer (20 mM HEPES pH 7.2, 0.5% Triton X-100, 50 mM NaCl, 3 mM MgCl2, 300 mM Sucrose, 1X protease inhibitor cocktail) and incubated on ice for 30 minutes.200 μL of suspension was removed and labelled "Total extract" and 200 μL of 2X SDS-PAGE sample buffer was added.The remaining lysate was clarified at 20,817 g in a 4°C centrifuge for 10 minutes.Supernatant was kept and labelled "Soluble" and 200 μL of 2X SDS-PAGE sample buffer was added.The insoluble pellet was washed once in 1mL of pre-extraction buffer before resuspension in 200 μL of pre-extraction buffer and 200 μL of 2X SDS-PAGE sample buffer.All samples were boiled at 99°C for 5 minutes and sonicated 3 times for 15 s at 10% amplitude in a MSE Soniprep 150 Plus sonicator prior to separation by SDS-PAGE and western blotting.
1.5 μg of anti-SUZ12 antibody (Cell Signalling #3737S) or 1.5 μg control IgG (Cell Signalling #3900S) was added to the lysate and incubated on a rotator for 2 h at 4 °C.50 μl of Dynabeads Protein G beads (Thermo Fisher Scientific #10003D) prewashed with PBS was added and rotated at 4 °C for an additional hour.Beads were washed with 1 ml lysis buffer (1:1 buffers C:D) once, then once with 1 ml of each of the Wash Buffers 2, 3 and 4 (Wash Buffer 2: 50 mM Tris-HCl pH 8.0, 150 mM NaCl; Wash Buffer 3: 50 mM Tris-HCl pH 8.0, 450 mM NaCl; Wash Buffer 4: 50 mM Tris-HCl pH 8.0).Beads were resuspended in a 100 μL of 1x LDS buffer and heated to 95 °C for 10 minutes.Lysate was loaded on a 10% SDS-PAGE gel and then western blot was carried out as described above, with the exception that a titer of 1:250 of the SUZ12 antibody was used.