Cell culture

ES-E14TG2a (E14) mES cells from male Mus musculus origin (RRID: CVCL9108361 [35]) were cultured as previously described [36] at 37 °C and 5% CO2 in DMEM base medium (Sigma Aldrich) supplemented with 10% FBS (Sigma Aldrich), 1X nonessential amino acids (Corning), 2mM L-glutamine (Corning), β-mercaptoethanol (Acros Organics) and LIF on 10 cm plates precoated with 0.2% gelatin. For LIF + 2i media, LIF media was further supplemented with 3 µM CHIR99021 GSK inhibitor (p212121), and 1 µM PD0325091 MEK inhibitor (p212121). Cells were passaged every ~48 h using trypsin and split at a ratio of ~1:8 with fresh medium. Routine anti-mycoplasma TC hood cleaning was conducted (LookOut DNA Erase spray, Sigma Aldrich) and cell lines were screened to confirm no mycoplasma presence. While generating clones, cells were grown in media with the addition of 1X penicillin/streptomycin (Corning) to prevent bacterial contamination.

Guide RNA cloning

Homology constructs for H3f3a and H3f3b (Supplemental Table 3) were designed to mutate residue 122 from lysine to alanine and mutate the PAM sequence (both highlighted in red). Oligos generated for the H3f3a or H3f3b gRNA were phosphorylated and annealed, then ligated to the px330-puro plasmid backbone using a simultaneous Fast Digest Bbs1 (ThermoFisher) digestion and Quick Ligase ligation (NEB). Plasmids were purified using the GenElute HP Plasmid Maxiprep Kit (Sigma Aldrich) following the manufacturer’s instructions and confirmed using Sanger sequencing (Genewiz).

Homology constructs were synthesized using IDT gBlocks Gene Fragments (IDT). The lyophilized construct was resuspended in 5 µL of nuclease-free water and 1 µL was used to clone the construct into the TOPO vector using the Zero Blunt TOPO PCR Cloning Kit for Sequencing (ThermoFisher) according to the manufacturer’s directions. The resulting plasmids were purified using the GenElute HP Plasmid Maxiprep Kit (Sigma Aldrich) and confirmed using Sanger sequencing (Genewiz).

H3.3K122A cell line generation

The targeting strategy used to generate cell lines where all four H3.3 alleles were mutated from lysine 122 to alanine is depicted in Fig. 1A. To generate cell lines, 2 × 105 cells were plated in a single well in a 6 well dish 24 h prior to transfection. One hour prior to transfection, medium was replaced with fresh, antibiotic-free medium. Three µg of pX330-puro Cas9/guide RNA vector and 3 µg of recombination template vector were combined in 100 µL of pre-warmed OptiMEM (Gibco). 24 µL of FuGENE HD (Promega) was added and the mixture was incubated at room temperature for 10 min. The mixture was added in drops evenly around the well of the dish and then gently swirled to ensure even distribution. 14–16 h post-transfection, cells were split using 0.5% trypsin (Gibco) and plated at varying densities on gelatinized 10 cm plates in antibiotic-free media. 48 h post-transfection, the media was replaced with media containing 2 µg/mL puromycin. 72 h post-transfection, the media was replaced with antibiotic-free media. Individual clones were picked 8 days post-transfection. Two days after picking clones, media was replaced with fresh penicillin/streptomycin media. Five days after picking clones, cells were split to 3 identical plates to allow for gDNA extraction and genetic screening and parallel storage for post-screening use. Targeting was performed in LIF and LIF + 2i containing media.

Clone screening

To extract gDNA, cells were washed once in 1xDPBS, lysed with ES cell lysis buffer (10 mM Tris pH 7.5, 10 mM EDTA, 10 mM NaCl, and 0.5% sarkosyl in water) including 1 mg/mL Proteinase K, and incubated overnight at 55 °C. The following day, plates were centrifuged for 2 min at 4 °C and 1000 rpm. DNA was precipitated using ice cold 100% ethanol with 75 mM NaCl and incubated for about 30 min until the solution cleared before centrifuging for 5 min at 4 °C and 3000 rpm. The supernatant was removed and the DNA was washed twice with 70% ethanol, then dried for an hour and resuspended in 50 µL 1X TE. PCR was performed on the extracted DNA in 96-well plates using primers that amplified the targeted H3f3a or H3f3b region (see Supplemental Table 2).

esiRNA generation and knockdown

Endoribonuclease-digested short interfering RNAs (esiRNAs) were generated as previously described [36]. An ideal target DNA sequence was identified within H3f3b and screened for unique nucleotide sequence via DEQOR. Amplified cDNA from wildtype mES cells was in vitro transcribed using T7 RNA polymerase. To generate esiRNAs, IVT products were digested with ShortCut RNase III (NEB) and purified using a PureLink RNA Mini Kit (Invitrogen). Transient transfections to deplete H3.3b were performed on wildtype cells (control for transfection efficiency) or H3f3a K122A clones on 6 wells using 5 µL of Lipofectamine 3000 (ThermoFisher) and 900 ng of esiRNAs for 48 h. Cells were harvested and counted with trypan blue on a BioRad Cell Counter. Following RNA isolation, depletion was quantified using RT-qPCR.

CRISPRi

A guide RNA for H3f3a was cloned into dCas9-KRAB-MeCP2 plasmid (Addgene 110821), as previously described [37]. dCas9-KRAB-MeCP2 was a gift from Alejandro Chavez & George Church (Addgene plasmid #110821; RRID: Addgene 110821). Plasmids were transfected into wildtype mES cells (control for transfection efficiency and CRISPRi activity) or H3f3b K122A clones on 6 wells using 5 µL of Lipofectamine 3000 (ThermoFisher), 5 µL of p3000 reagent, and 5 µg of plasmid for 48 h. Cells were harvested and counted with trypan blue on a BioRad Cell Counter. Following RNA isolation, transcript abundance was quantified using RT-qPCR.

Protein extraction and western blot analysis

For each protein, the following primary/secondary antibodies and dilutions were used: H3 (abcam: ab1791, lot GR300978-2, 1:1000), H3.3 (Millipore: 09-838, lot 3987735, 1:1000), K122ac (Invitrogen: PA5-112508, lot YJ4090633A, 1:1000), and ACTIN (Sigma: A1978, lot 037M4782V, 1:5000), IRDye 800CW Goat anti-rabbit IgG Secondary (LI-COR: 926-32211, 1:10000), and IRDye 800CW Goat anti-mouse IgG Secondary (LI-COR: 926-32210, 1:10000).

Cells were trypsinized and collected in a 15 mL conical tube, then centrifuged at 600 rcf for 5 min at 4 °C. Cell pellets were washed twice with cold 1XPBS and protein was extracted using 200 µL Triton Extraction Buffer (PBS containing 0.5% Triton X 100 (v/v), 2 mM phenylmethylsulfonyl fluoride (PMSF), 0.02% (w/v) NaN3). Concentration was determined using the Pierce BCA protein assay kit (ThermoFisher), and 7.5–10 µg protein per sample was loaded onto 15% Tris-acrylamide gels with Precision Plus ladder (Biorad). Gels were run for 40 V for 1 h, 120 V for 2.5 h, and then transferred at 20 V overnight to a nitrocellulose membrane (Biotrace). Successful transfer assessed using REVERT 700 total protein stain (LI-COR) and membranes were blocked using Intercept blocking buffer (LI-COR). Primary antibody incubations were conducted overnight at 4 °C, and secondary antibody incubations were conducted for 2 h at room temperature. All imaging was performed using a LI-COR Odyssey DLx Imager according to the manufacturers specifications. Protein quantification was conducted using ImageJ [38]. For total protein quantification, identical areas for each lane were selected and the mean pixel intensity was measured and subtracted from the mean pixel background intensity. For target protein quantification, identical size areas immediately surrounding the target band were selected and mean pixel intensity was measured and subtracted from mean background pixel intensity. Target protein quantification was then made relative to total protein quantification, and then made relative to wildtype. Three biological replicates were conducted per target for each cell line.

Cell growth assay

Cells were plated in 6 wells at a density of 50,000 cells per well. At days 1, 2, 3, 4, and 6 post-plating, cells were trypsinized using 0.5 mL of trypsin and stopped with 1.5 mL of ES cell medium. After achieving single cell suspension, cells were counted on a TC20 cell counter (BioRad) using trypan blue stain to distinguish live cells. Three biological replicates were conducted for each cell line.

Alkaline phosphatase staining

Cells were washed twice in 1X Dulbecco’s Phosphate-Buffered Saline (DPBS, Gibco) and crosslinked for five minutes at room temperature using 1% formaldehyde (Fisher) in DPBS. Crosslinking was quenched using 500 mM glycine and cells were washed using 1X DPBS twice. Cells were stained using Vector Red Alkaline Phosphatase Staining Kit (Vector Labs) per manufacturer’s instructions in a 200 mM Tris-HCl buffer, pH 8.4. 2 mL working solution was added to each 6 well and incubated in the dark for 30 min before being washed with 1XPBS and imaged. Three biological replicates were conducted for each cell line.

4sU-labelling and RNA isolation

4sU-labelled RNA was generated and isolated from cells using TT-seq as previously described [28, 39]. Media was aspirated from cell plates and replaced with 10 mL of 500 nM 4-thiouridine (4sU, Carbosynth T4509) containing ES cell media and the plates incubated at 37 °C with 5% CO2 for 5 min. After 5 min, the 4sU-containing media was aspirated and the cells were washed with 1xPBS and 2 mL TRIzol (Invitrogen) was added directly to the plate. Cell lysate were collected, and RNA was extracted according to ThermoFisher’s recommendations. RNA concentration was determined by Qubit with the Qubit RNA broad range quantification kit (ThermoFisher). 100 µg of total RNA was diluted to a concentration of 240 ng/µL at a volume of 416.67 µL in 1XTE and then fragmented with a Diagenode Bioruptor Pico on high power for one 30 s cycle. The fragmented RNA was then combined with 283.33 µL 1XTE, 100 µL 10X Biotinylation buffer (100 mM Tris pH 7.4 and 10 mM EDTA), and 200 µL of 1 mg/mL biotin-HPDP (ThermoFisher) in dimethylformamide (DMF; freshly prepared). Samples were vortexed, then incubated in a thermomixer at 37 °C shaking at 1000 RPM in the dark for 2 h. Samples were chloroform extracted, isopropanol/salt precipitated, and resuspended in 22 µL of nuclease-free water. Sixty uL of Streptavidin C1 beads (Invitrogen) were prepared for RNA separation by washes with 1 mL of 1 M NaOH and 50 mM NaCl and resuspended in 60 µL of TT-seq Binding buffer (10 nM Tris pH 7.4, 300 mM NaCl, 0.1% Triton). Then, 60 µL of prepared streptavidin C1 beads were added to each sample and rotated at room temperature for 20 min. Following incubation, the samples were magnetized for 1 min and the supernatant (containing the unlabeled RNA) was placed in a separate 1.5 mL tube and put on ice. The unlabeled RNA from the supernatant was Phenol: Chroloform: Isopropanol(PCI)/chloroform extracted, isopropanol/salt precipitated, and resuspended in 100 µL of nuclease-free water. The bead-bound labeled nascent RNA was washed twice with 500 µL of High Salt buffer (50 mM Tris pH 7.4, 2 M NaCl, 0.5% Triton), twice with 500 µL of TT-seq Binding buffer, and once with 500 µL of Low Salt buffer (5 mM Tris pH 7.4, 0.1% Triton), rotating for 1 min at room temperature, re-magnetizing and resuspending the beads during each wash. The nascent RNA was recovered from the beads through two rounds of 100 µL of freshly prepared 100 mM DTT and incubating in a thermomixer at 65 °C and 1000 RPM shaking for 5 min. Eluted nascent RNA was recovered with a PCI extraction and an isopropanol/salt/glycogen precipitation. RNA pellets were resuspended in 25 µL of nuclease-free water. The total RNA and nascent RNA from each sample were used for RT-qPCR and TT-seq libraries as described below.

Transient transcriptome sequencing (TT-seq)

For TT-seq, three replicates for wildtype mES cells, and one replicate for each mutant cell line were performed. 4sU labelling for TT-seq was performed as described in 4sU labelling section. RNA pellets were resuspended in 1X TE buffer, at approximately 100 µL per 100 µg of RNA. Strand-specific nascent RNA library building was performed as previously described [39] using the NEBNext Ultra II Directional Library kit with the following changes: The rRNA depleted RNA was fragmented at 94 °C for fifteen minutes following the manufacturer’s instructions for intact RNA, the fragmentation, first strand cDNA synthesis, and second strand cDNA synthesis were performed at double the reaction volume in the manufacturer’s instructions, the adaptors were diluted 1:5 in Adaptor Dilution Buffer, the primers were diluted 1:5 in nuclease-free water, and 7 cycles of PCR were used to amplify the libraries. Finished libraries were quantified by Quibit with the dsDNA High Sensitivity kit and run on a Fragment Analyzer to confirm high quality of each library prior to sequencing. Libraries were sequenced paired-end on an Illumina NextSeq500 to ~ 40 million mapped reads.

TT-seq analysis

Raw paired end fastq files were analyzed with fastQC [40] using default parameters to ensure high data quality and adapter sequences were trimmed using cutadapt [41]. Reads were aligned with the mm10 Gencode annotation (GRCm38.p6, vM23) using STAR [42] with the following parameters: outFilterMismatchNoverReadLmax 0.02, outFilterMultimapNmax (1) samtools [43] was used to filter reads and generate indexed bams with the following parameters: -q 7 -f (2) Feature counts were generated using featurecounts [44] options -B -t “exon” -g “gene_name” -F GTF -p -s 2 with the mm10 (GRCm38.p6, vM23) genome annotation (GTF format). Differential gene expression analysis was conducted using DEseq2 [45] with cutoff as |log2(FC)|≥0.75 and FDR ≤ 0.05. Gene Ontology analysis on gene sets of interest was performed the R package “clusterProfiler” [46] using all Gencode features as background. Upset plots were generated using the R package “upsetR” (https://github.com/hms-dbmi/UpSetR), and volcano plots were generated using R package “EnhancedVolcano” (https://bioconductor.org/packages/release/bioc/html/EnhancedVolcano.html).

Reverse transcription and quantitative PCR (RT-qPCR)

cDNA was generated from 1 µg of total RNA or 100 ng of 4sU labelled RNA with random hexamers (Promega) with homemade reverse transcriptase (RT). Generated cDNA was used as a template in qPCR reactions using 2X SYBR FAST mix (KAPA) and 5 µM target specific primers (see Supplemental Table 2) on a Lightcycler 96 (Roche). Target transcript abundance in samples was determined using the ΔΔCT normalization method relative to wildtype samples, using Gapdh transcript abundance for internal normalization as previously described [47].

CUT&RUN

The following antibodies and dilutions were used for CUT&RUN: H3.3 (Active Motif 91191, lot # 01022005; 1:50), K122ac (Invitrogen PA5-105108, lot # WJ3411065A; 1:50), and K27ac (abcam ab4729, lot # GR3416784-1; 1:100).

CUT&RUN was performed as described [39, 48, 49] using recombinant ProteinA-MNase (pA/G-MN) or Protein A/Protein G-MNase (pA/G-MN), as specified. For each target, two replicates for wildtype and one replicates for each mutant cell line were performed. Briefly, 100,000 nuclei were isolated from cell populations using a hypotonic buffer (20 mM HEPES-KOH, pH 7.9, 10 mM KCl, 0.5mM spermidine, 0.1% Triton X-100, 20% glycerol, freshly added protease inhibitors) and flash frozen. Nuclei were thawed and bound to lectin-coated concanavalin A magnetic beads (50 µL bead slurry per 100,000 nuclei; Polysciences). Immobilized nuclei were pre-blocked with blocking buffer (20 mM HEPES, pH 7.5, 150 mM NaCl, 0.5mM spermidine, 0.1% BSA, 2 mM EDTA, fresh protease inhibitors) and washed in wash buffer (20 mM HEPES, pH 7.5, 150 mM NaCl, 0.5mM spermidine, 0.1% BSA, fresh protease inhibitors). Nuclei were incubated in wash buffer containing primary antibody (H3.3, K122ac, K27ac) overnight at 4 °C with rotation. Nuclei were incubated in wash buffer containing primary antibody for 1 h at room temperature with rotation. Next, samples were washed twice with wash buffer and incubated in wash buffer containing recombinant pA-MN (for K122ac, K27ac) or pA/G-MN (for H3.3) for 30 min at room temperature with rotation. Controls lacking a primary antibody were subjected to the same conditions but incubated in wash buffer without any primary antibody prior to incubation with pA-MN/pA/G-MN. After incubation, samples were washed twice with was buffer and equilibrated to 0 °C in an ice/water bath. MNase cleavage was activated upon addition of 3 mM CaCl2. Digestion was chelated for samples containing primary antibody K27ac, after 30 min on the ice water bath, using 20 mM EDTA, 4 mM EGTA, 200 mM NaCl, and 1.5 pg MNase-digested S. cerevisiae mononucleosomes were added as a spike-in control. Genomic fragments were released after an RNase A treatment. Released fragments were separated through centrifugation. Digestion was chelated for samples containing primary antibodies K122ac and H3.3 after 30 min on the ice water bath, using a low-salt treatment: 10 mM EDTA, 2 mM EGTA, 150 nM NaCl, 5 mM TritonX, and 1.5 pg MNase-digested S. cerevisiae mononucleosomes were added as a spike-in control. Genomic fragments were released after being incubated for 1 h at 4 °C and released fragments were separated from beads. The salt concentration of low salt treated samples were increased to 500 mM and additional RNase A.

Isolated fragments from all samples were used as input for a library build consisting of end repair and adenylation, adapter ligation, and subsequent purification with AMPure XP beads (Agencourt). Barcoded fragments were then amplified by 15 cycles of high-fidelity PCR and purified using agarose gel extraction. Libraries were pooled and sequenced on an Illumina NextSeq500 to a depth of ~ 10 million mapped reads.

CUT&RUN analysis

Paired-end fastq files were trimmed to 25 bp and mapped to the mm10 genome with bowtie2 [50] (options -q -N 1 -X 1000). Mapped reads were duplicate-filtered using reads with insufficient mapping quality (MAPQ ≥ 10) were removed using samtools [43]. Reads were converted to bigWig files using deepTools [51] with the spike in S. cerevisiae mapped reads used as a scaling factor. Heatmaps and metaplots were generated using deepTools. Peaks were called from CUT&RUN samples using MACS2 [52] with no antibody samples as input. For each target, peaks from all samples were combined to create a consensus peak set. After removing all peaks with less than 10 counts in 50% of samples, differential enrichment peak analysis was performed using edgeR [53] for all available replicates of each condition with RUVseq [54] correction. Differentially enriched peaks were defined as |log2(FC)|≥0.5 and FDR ≤ 0.05. For each target, samples peak annotations and motif enrichment analysis were performed using HOMER [55].