RNA m5C oxidation by TET2 regulates chromatin state and leukaemogenesis

Animals and tissues

Tet2^{−/−} mice were generated as described⁴¹. These mice used in this study were backcrossed for more than six generations with C57BL/6 mice. WT C57BL/6 and Tet2^{−/−} mice (aged 6–8 weeks), including both male and female, were used throughout this study and maintained under standard laboratory housing conditions with food and water ad libitum. All the mice were randomly assigned to experimental groups and data analyses were blindly performed by two lab members independently. All animal studies were performed with the approval from the Institutional Animal Care and Use Committee (IACUC), protocol number 30979/20190086AR at The University of Texas Health Science Center at San Antonio (UTHSCSA) and conducted in accordance with the institutional and national guidelines and regulations.

Xenotransplantation of human leukaemia cells

For in vivo xenotransplantation study procedures, 1 × 10⁶ K-562 cells were injected intravenously via the tail vein into adult NSG mice (aged 6–8 weeks) pretreated with 250 cGy whole body irradiation. At 28–39 days after transplantation, PB was collected from the submandibular vein, and the BM was isolated from the tibias and femurs. Human CD33⁺ chimerism in BM and PB cells were analysed by BD FACSCelesta flow cytometer (BD Biosciences).

2 × 10⁴ THP-1 cells were injected intravenously through the tail vein into adult NSG mice (6–8 weeks old) pretreated with 250 cGy whole-body irradiation. At 20–22 days after transplantation, human CD33⁺CD45⁺ chimerism in BM and PB cells were analysed using the BD FACSCelesta flow cytometer.

A cohort of mice from each transplantation group was monitored until they became moribund or died.

Competitive repopulation assay

The competitive repopulation assay was performed to assess the effect of TET2 and/or MBD6 KD on the repopulating potential of HSPCs in vivo. In total, 2 × 10⁴ Lin^−KIT⁺ cells isolated from the BM cells of 8-week-old WT or Tet2-KO mice (CD45.2⁺) were lentivirally transduced with short hairpin RNA (shRNA) plasmid pLKO.1-shC002 (MilliporeSigma, SHC002: shNC) or pLKO.1-shMbd6 (Millipore-Sigma, TRCN0000178563) and incubated in suspension culture containing 20% FBS in complete RPMI-1640 medium supplemented with 100 ng ml^−1 mSCF, 10 ng ml^−1 mIL-3, 10 ng ml^−1 IL-6 and 20 ng ml^−1 mFlt3. Then, 48 h after transduction, Lin^−KIT⁺ cells from each transduction were transplanted along with 1 × 10⁶ 8-week-old BoyJ (CD45.1⁺) BM competitor cells into lethally irradiated (800 cGy) BoyJ recipients through the tail-vein injection. The CD45.2/CD45.1 chimeras in the PB were monitored monthly for 6 months. Recipients were euthanized 6 months after transplantation to analyse the CD45.2/CD45.1 chimeras in the BM and spleen.

Haematopoietic stem and progenitor cell sorting, colony assay and in vitro differentiation assay

For haematopoietic stem and progenitor Lin^−KIT⁺ cell selection, magnetic-activated cell sorting was applied with autoMACS Pro Separator (Miltenyi Biotec). In brief, the lineage-positive cells (Lin⁺) were depleted from total BM cells of 6–8-week-old mice using the Direct Lineage Cell Depletion Kit (Miltenyi Biotec, 130-110-470), and the Lin^− cells were then sorted with KIT (CD117) MicroBeads (Miltenyi Biotec, 130-091-224). The purity of selected cells was analysed by flow cytometry.

For colony assay, HSPCs were plated in triplicate in methylcellulose medium (MethoCult, M3134) supplemented with mouse stem cell factor (mSCF; 100 ng ml^−1), interleukin-3 (mIL-3; 10 ng ml^−1), thrombopoietin (mTPO; 50 ng ml^−1), granulocyte-macrophage colony-stimulating factor (mGM-CSF; 10 ng ml^−1), human erythropoietin (hEPO; 4 U ml^−1) and interleukin-6 (hIL-6; 50 ng ml^−1, PeproTech). The colonies were imaged using STEMvision (StemCell Technologies) and scored on day 7, and these colonies were then sequentially replated every 7 days for replating assay. Colony cells were also collected and analysed for expression of stem and progenitor markers and myeloid linage markers by flow cytometry.

The HSPCs were also incubated in suspension culture containing 30% FBS and 2% BSA in complete RPMI-1640 medium supplemented with 100 ng ml^−1 mSCF, 10 ng ml^−1 mIL-3, 50 ng ml^−1 mTPO and 10 ng ml^−1 mGM-CSF. Cells were collected and analysed for expression of stem/progenitor markers at day 7 and myeloid lineage markers at day 14 by flow cytometry.

Flow cytometry analysis

Cells were stained with PerCP-Cy5.5 mouse lineage antibody cocktail (BD Biosciences, 561317) and PE rat anti-mouse CD117 (BD Biosciences, 553869) antibody for haematopoietic stem and progenitor cells analysis. Brilliant Violet 421 (BV421) anti-mouse/human CD11b (Mac-1) (BioLegend, 101236) was used to analyse myeloid lineage. PerCP-Cy5.5 mouse anti-mouse CD45.2 (BD Biosciences, 552950) and FITC mouse anti-mouse CD45.1 (BD Biosciences, 553775) antibodies were used for analysing CD45.2/CD45.1 chimeras in a competitive repopulation assay.

Human CD33 chimerism was analysed with PE mouse anti-human CD33 (BD Biosciences, 561816) and PE-Cy7 rat anti-mouse CD45 (BD Biosciences, 552848) in PB and BM cells from NSG mice that were xenotransplanted with K-562 cells. Human CD33/CD45 chimerism was analysed with PE mouse anti-human CD33 (BD Biosciences, 561816) and APC mouse anti-human CD45 (BD Biosciences, 555485) in PB and BM cells from NSG mice that were xenotransplanted with THP-1 cells. All flow cytometry data were analysed using FlowJo-V10 software (TreeStar). Examples of the gating strategies are provided in Supplementary Figs. 2 and 3.

Cell culture

WT and Tet2^{−/−} mES cells were gifts from the B. Ren laboratory^26,48. The control and KO mES cells have been shown to be pluripotent by chimera formation assay. All mES cells were kept in DMEM (Gibco, 11995065) supplemented with 15% heat-inactivated stem-cell-qualified fetal bovine serum (Gemini Bio Products, 100-525), 1× l-glutamine (Gibco, 25030081), NEAA (Gibco, 25030081), LIF (Millipore-Sigma, ESG1107), 1× β-mercaptoethanol (Gibco, 21985023), 3 μM CHIR99021 (StemCell Technologies, 72052) and 1 μM PD0325901 (StemCell Technologies, 72182) at 37 °C and 5% CO₂. For stable TET2 overexpression mES cells, empty vector, WT Tet2 or Tet2 HxD mutant bearing piggyBac plasmids were constructed and transfected into Tet2-KO or Pspc1-KO mES cells using Lipofectamine 3000 Transfection Reagent (Invitrogen, L3000001) according to the standard protocol. Stable expression clone selection was performed using 0.1 mg ml^−1 hygromycin B (Gibco, 10687-010) for 2 weeks. The medium was replaced every 24 h. ES cells were passaged on gelatin-coated plates twice to clear feeder cells before experiments.

WT THP-1, K-562 and TF-1 cells were obtained from the American Type Culture Collection (ATCC). The SKM-1 cell line was obtained from DSMZ (German Collection of Microorganisms and Cell Cultures). WT OCI-AML3 cell was a gift from L. Godley. WT and TET2^{−/−} K-562 and THP-1 cells were gifts from B. K. Jha as previously generated⁴⁹. THP-1, K-562, SKM-1 and OCI-AML3 cells were kept in RPMI-1640 (Gibco, 61870036) with 10% fetal bovine serum (FBS, Gibco 26140079) at 37 °C under 5% CO₂. TF-1 was kept in RPMI-1640 (Gibco, 61870036) with 10% FBS (Gibco 26140079) and 2 ng ml^−1 recombinant GM-CSF (Peprotech, 300-03) at 37 °C under 5% CO₂. U-87 MG (HTB-14), LN-229 (CRL-2611), Hep G2 (HB-8065), HeLa (CCL-2), HCT 116 (CCL-247), A549 (CCL-185) and A-375 (CRL-1619) cells were obtained from the American Type Culture Collection (ATCC). U-87 MG and LN-229 were kept in ATCC-formulated Eagle’s minimum essential medium (ATCC, 30-2003) supplemented with 10% FBS (Gibco, 26140079) and 5% FBS (Gibco, 26140079), respectively. Hep G2, HeLa, HCT 116, A549 and A-375 cells were kept in DMEM (Gibco, 11995065) supplemented with 10% FBS (Gibco, 26140079). All cell types were kept at 37 °C and 5% CO₂.

shNC and shMBD6 THP-1 and K-562 cell lines were constructed by lentivirus transduction with TransDux MAX Lentivirus Transduction Reagent (System Biosciences, LV860A-1). Lentiviral particles were prepared by using HEK293T cells and lentiviral packaging plasmids pCMV-VSV-G and pCMV-dR8.2 (pCMV-VSV-G and pCMV-dR8.2 were gifts from B. Weinberg (Addgene plasmid, 8454; and Addgene plasmid, 8455)) and shRNA plasmid pLKO.1-shC002 (Millipore-Sigma, SHC002) or pLKO.1-shMBD6 (Millipore-Sigma, TRCN000038787). Then, 48 h after transfection, lentiviral particles were precipitated using the PEG-it Virus Precipitation Solution (System Biosciences, LV810-1). shNC and shMBD6 THP-1 and K-562 cell lines were kept in RPMI-1640 (Gibco, 61870036) with 10% fetal bovine serum (FBS, Gibco) and 1 μg ml^−1 puromycin (Gibco, A1113803) at 37 °C under 5% CO₂. Small interfering RNA (siRNA) or gene overexpression plasmids transfection in K-562 and THP-1 cells were performed according to the manufacturer’s instructions for SF Cell Line 4D-Nucleofector X Kit (Lonza Biosciences, V4XC-2032, FF-120 for K-562) or SG Cell Line 4D-Nucleofector X Kit (Lonza Biosciences, V4XC-3024, FF-100 for THP-1)

TET2-KO THP-1 cell line for PDX model was generated using CRISPR–Cas9 system. Single-guide RNAs were designed using the CRISPick tool (https://portals.broadinstitute.org/gppx/crispick/public) and then cloned into LentiCRISPR V2-GFP vector by Synbio Technologies. THP-1 cells were infected by lentiviral particles for 72 h and followed by GFP-positive cell selection using the BD FACSMelody Cell Sorter (BD Biosciences). KO efficiency was verified by western blotting.

shNC, shMBD6 (Millipore-Sigma, TRCN0000178563), shNsun2 (Millipore-Sigma, TRCN0000325347), shNsun5 (Millipore-Sigma, TRCN0000097512) or shTrdmt1 (Millipore-Sigma, TRCN0000328293) Lin^−KIT⁺ HSPCs were constructed by electroporation with the P3 Primary Cell 4D-Nucleofector X Kit S (Lonza Bioscience, V4XP-3032) by program CV-137.

siRNA and plasmid transfection

Two or three individual siRNAs, or a pool of four siRNAs targeting different regions of the same transcript (Dharmacon siRNA) were used for KD of human or mouse transcripts. siRNA transfections in mES cells and other adherent cell lines were performed using Lipofectamine RNAiMAX Transfection Reagent (Invitrogen, 13778075) according to the manufacturer’s instructions. Transfections in human leukaemia cells (THP-1, TF-1, OCI-AML3, SKM-1) were performed by electroporation using the SG Cell Line 4D-Nucleofector X Kit L (Lonza Bioscience, V4XC-3024) with program FF-100. Transfections in K-562 cells were performed with the SF Cell Line 4D-Nucleofector X Kit L (Lonza Bioscience, V4XC-2012) with program FF-120.

Plasmid transfections in mES cells or HEK293T cells were performed using the Lipofectamine 3000 Transfection Reagent (Invitrogen, L3000015) according to the manufacturer’s instructions.

Cell proliferation assay

The cell proliferation assays for adherent and suspension cells were performed similarly. Cells were seeded into 96-well plates before assaying in 100 μl settings with CellTiter 96 Aqueous One Solution Cell Proliferation Assay (Promega, G3582) according to the manufacturer’s instructions. Then, 2,000–10,000 cells were seeded per well at day 0 and the cell proliferation was monitored every 24 h by incubating the cell suspension with MTS reagent at 37 °C for 1 h.

DNase I–TUNEL assay

For cell line samples, mES cells were reseeded to 10 cm cell culture dishes 12 h before siRNA transfection. The DNase I–TUNEL assay was performed using DeadEnd Fluorometric TUNEL System (Promega, G3250) according to the manufacturer’s instructions after cell fixation with paraformaldehyde and permeabilization with Triton X-100. Two independent experiments were performed. Cells were treated with 1 U ml^−1 of DNase I (Thermo Fisher Scientific, EN0521) for 5 min at 37 °C before rTdT labelling. Flow cytometry was performed on a BD Fortessa (BD Biosciences), and data were analysed using Flowjo (TreeStar).

Nascent RNA imaging assay

mES cells were reseeded in Nunc Lab-Tek II Chambered Coverglass (Thermo Fisher Scientific, 155409) 12 h before treatment. The nascent RNA synthesis assay was performed using Click-iT RNA Alexa Fluor 488 Imaging Kit (Invitrogen, C10329) according to the manufacturer’s instructions. 5-Ethynyl uridine incubation was performed for 1 h before washing away by cell medium. Cell nucleus was counterstained with Hoechst 33342 (Abcam, ab228551). The samples were imaged on a Leica SP8 laser scanning confocal microscope at University of Chicago. The fluorescence intensity across different samples were quantified with Cellprofiler v.3.0 with a custom workflow. The total RNA synthesis rate was obtained by multiplying the average intensity in each cell by the area of each cell.

ATAC–see analysis

Assay of transposase-accessible chromatin with visualization (ATAC–see) of mES cells was performed as described in the original report⁵⁰. ATTO-590-labelled imaging oligos were purchased from Integrated DNA Technologies (IDT) and the oligonucleotide sequences are as follows: Tn5MErev, 5′-[phos]CTGTCTCTTATACACATCT-3′; Tn5ME-A-ATTO590, 5′-/5ATTO590/TCGTCGGCAGCGTCAGATGTGTATAAGAGACAG-3′; Tn5ME-B-ATTO590: 5′-/ATTO590/GTCTCGTGGGCTCGGAGATGTGTATAAGAGACAG-3′. The oligos were assembled with recombinant Tn5 transposase (Active motif, 81286) to produce the Tn5 transposome. Cell fixation, permeabilization and labelling were performed as described in the original report⁵⁰.

Recombinant protein purification

Standard molecular cloning strategies were used to generate C-terminally MBP–6×His-tagged MBD domain of MBD6 (residues 1–100). The human MBD6 coding sequence was obtained from Origene (Origene, SC324058). The full-length coding sequence was cloned using PrimeSTAR GXL DNA Polymerase (TaKaRa Bio, R050B). Recombinant proteins were expressed in E. coli BL21 (DE3) grown to an optical density at 600 nm of 0.6 in LB medium. The expression was induced with 0.6 mM IPTG at 16 °C for 20 h and cells were collected by centrifugation.

For purification of MBP tagged MBD domain of MBD6, bacterial pellet was resuspended in a lysis buffer containing 25 mM Tris-HCl (pH 7.5), 500 mM NaCl, 20 mM imidazole, 10 mM β-mercaptoethanol (β-ME) and protease inhibitors (ethylenediaminetetraacetic-acid-free protease inhibitor cocktail tablet, Millipore-Sigma 4693132001) and disrupted by sonication for 3 min. The cell lysates were clarified by centrifugation at 26,000g for 30 min and the supernatant was applied to Ni²⁺-NTA resin (Thermo Fisher Scientific, 88221) and washed with lysis buffer, and the bound proteins were eluted with lysis buffer supplemented with 250 mM imidazole. The eluted protein was bound back to amylose resin (NEB, E8021S) before washing with lysis buffer. The bound protein was eluted with 1% maltose in lysis buffer. The eluted protein was analysed by SDS–PAGE and concentrated by centrifugal filtration (Amicon Ultra-15). Final concentrated protein was aliquoted, flash-frozen and stored at −80 °C for future use.

RT–qPCR

To quantify expression levels of transcripts, total RNA was reverse transcribed using the PrimeScript RT Master Mix (TaKaRa Bio, RR0361) with oligo dT primer and random hexamers as primers. The cDNA was then subjected to quantitative PCR (qPCR; LightCycler 96 system, Roche) using FastStart Essential DNA Green Master (Roche, 06402712001) with gene-specific primers. The relative changes in expression were calculated using the ΔΔC_t method.

Western blot analysis

Protein samples were prepared from respective cells by lysis in RIPA buffer (Thermo Fisher Scientific, 89900) containing 1× Halt protease and phosphatase inhibitor cocktail (Thermo Fisher Scientific 78441). Protein concentration was measured by NanoDrop 8000 Spectrophotometer (Thermo Fisher Scientific). Lysates of equal total protein concentration were heated at 90 °C in 1× loading buffer (Bio-Rad, 1610747) for 10 min. Denatured protein was loaded into 4–12% NuPAGE Bis-Tris gels (Invitrogen, NP0335BOX) and transferred to PVDF membranes (Thermo Fisher Scientific, 88585). Membranes were blocked in Tris-buffered saline, 0.1% Tween-20 (TBST) with 3% BSA (Millipore-Sigma, A7030) for 30 min at room temperature, incubated in a diluted primary antibody solution at 4 °C overnight, then washed and incubated in a dilution of secondary antibody conjugated to HRP for 1 h at room temperature. Protein bands were detected using SuperSignal West Dura Extended Duration Substrate kit (Thermo Fisher Scientific, 34075) with a FluroChem R (Proteinsimple). Blot intensities were quantified with Fiji (ImageJ) Analyse-Gel module. Uncropped gels with size marker indications are provided in Supplementary Fig. 1.

Dot blot

Oligonucleotide probes end-labelled with Alexa Fluor 488 dye was spotted on a positively charged Nylon membrane (Roche, 11209299001). The membrane was dried at room temperature for 5 min before UV cross-linking at 254 nm with a Stratalinker (Stratagene) for two times to achieve a 4,500 J m^−2 UV flux. The membrane was then blocked in Tris-buffered saline, 0.1% Tween-20 (TBST) with 3% BSA (Millipore-Sigma, A7030) for 30 min at room temperature. Primary antibodies were diluted according to the manufacturer’s instructions and incubated with the membrane for 60 min at room temperature. The membrane was washed and incubated in a dilution of secondary antibody conjugated to HRP for 60 min at room temperature. The final membrane was detected using SuperSignal West Dura Extended Duration Substrate kit (Thermo Fisher Scientific, 34075) with the iBright 1500 system (Invitrogen, A44241).

Cell fractionation

Fractionation of mES cells, K-562 or THP-1 cells was performed according to the published protocol⁵¹ with the optimized concentration of NP-40 (MilliporeSigma, 492018) for each cell line. In brief, 5 × 10⁶ to 1 × 10⁷ cells were collected and washed with 1 ml cold PBS/1 mM EDTA buffer, then centrifuged at 4 °C and 500g to collect the cell pellet. Then, 200 μl ice-cold lysis buffer (10 mM Tris-HCl, pH 7.4, 0.05% NP-40, 150 mM NaCl) were added to the cell pellet and incubated on ice for 5 min, then gently pipetted up the cell lysate over 2.5 volumes of chilled sucrose cushion (24% RNase-free sucrose in lysis buffer) and centrifuged at 4 °C and 15,000g for 10 min. All the supernatant was collected as cytoplasmic fraction and the nuclei pellet was washed once by gently adding 200 μl ice-cold PBS/1 mM EDTA to the nuclei pellet without dislodging the pellet. The nuclei pellet was resuspended in 200 μl prechilled glycerol buffer (20 mM Tris-HCl, pH 7.4, 75 mM NaCl, 0.5 mM EDTA, 0.85 mM DTT, 0.125 mM PMSF, 50% glycerol) with gentle flicking of the tube. An equal volume of cold nucleus lysis buffer (10 mM HEPES, pH 7.6, 1 mM DTT, 7.5 mM MgCl₂, 0.2 mM EDTA, 0.3 M NaCl, 1 M urea, 1% NP-40) was then added, followed by vigorous vertexing for 5 s twice. The nuclei pellet mixtures were incubated for 2 min on ice, then centrifuged at 4 °C and 15,000g for 2 min. The supernatant was collected as the soluble nuclear fraction (nucleoplasm). The pellet was gently rinsed with cold PBS/1 mM EDTA without dislodging and was then collected as the chromosome-associated fraction.

Fractionation of HSPCs was performed similar to ES cells with minor modifications. In brief, HSPCs were cultured in vitro for 2 h after sorting on the autoMACS Pro Separator, and then ice-cold lysis buffer (10 mM Tris-HCl, pH 7.4, 0.15% IGEPAL CA-630, 75 mM NaCl) was used to separate the cytoplasmic fraction. The procedures for isolating the nuclear fraction and chromosome-associated fraction were the same as that of ES cells.

Quantitative analysis of modified base levels using UHPLC–MS/MS

The nucleic acid digestion step for RNA was as follows: 75 ng ribo-depleted RNA was digested by nuclease P1 (MilliporeSigma, N8630) in 20 μl buffer containing 20 mM ammonium acetate at pH 5.3 for 2 h at 42 °C. Then, 1 U of FastAP thermosensitive alkaline phosphatase (Thermo Fisher Scientific, EF0651) was added to the reaction and FastAP buffer was added to a 1× final concentration before incubation for 2 h at 37 °C. For DNA, genomic DNA was purified from cells according to the standard protocol of the Monarch Genomic DNA Purification Kit (NEB, T3010S). An additional RNase A (Thermo Fisher Scientific, EN0531) digestion step was performed on the purified DNA and the reaction was recovered with DNA Clean & Concentrator-5 (Zymo Research, D4014). Then, 200 ng DNA was digested with Nucleoside Digestion Mix (NEB, M0649S) at 37 °C for 2 h.

The samples were diluted and filtered (0.22 μm, Millipore) and injected into a C18 reversed-phase column coupled online to the Agilent 6460 LC–MS/MS spectrometer in positive electrospray ionization mode. The nucleosides were quantified using retention time and the nucleoside to base ion mass transitions (for RNA: 268 to 136 for A; 284 to 152 for G; 258 to 126 for m⁵C and 274 to 142 for hm⁵C; for DNA: 228 to 112 for dC, 242 to 126 for 5mdC, 258 to 142 for 5hmdC). Quantification was performed by comparing with the standard curve obtained from pure nucleoside standards running with the same batch of samples.

Chromatin-associated RNA-seq

Chromatin-associated RNA-seq analyses of mES cells, K-562 and HSPCs were performed similarly. After caRNA isolation, ERCC RNA spike-in mix (Invitrogen, 4456740) was added to purified total caRNA according to the ratio recommended by the standard protocol. Ribosomal RNA was depleted from isolated chromatin-associated RNA with RiboMinus Eukaryote System v2 (Invitrogen, A15026) followed by size-selection using the standard protocol of RNA Clean & Concentrator-5 (RCC-5, Zymo Research, R1013). RNA libraries were constructed with SMARTer Stranded Total RNA-Seq Kit v2 – Pico Input Mammalian (TaKaRa Bio, 634411) according to the manufacturer’s instructions. Three replicates were performed for each condition. Libraries were sequenced on the NovaSeq 6000 sequencer.

ATAC–seq analysis

ATAC–seq was performed using the ATAC–seq kit (Active Motif, 53150) according to the manufacturer’s instructions. In brief, 50,000 to 100,000 cells were aliquoted for each replicate and mixed with equal amounts of Drosophila spike-in (Active Motif, 53154). Cells were then permeabilized with buffer containing 0.1% Tween-20 and 0.01% Digitonin, both supplied by the original kit. Accessible chromatin regions were tagged with pre-assembled Tn5 transposome. Tagged genomic DNA was extracted from cells and DNA libraries were obtained by PCR amplification. Pooled libraries were sequenced on the NovaSeq 6000 sequencer. For ATAC–qPCR, tagged genomic DNA was extracted and amplified by PCR for 8 cycles using the indexing primers from the original kit. Amplified DNAs were subjected to qPCR analysis using individual primer sets.

m⁵C methylated RNA immunoprecipitation with spike-in

m⁵C modified or unmodified mRNA spike-ins were in vitro transcribed from firefly luciferase or Renilla luciferase coding sequences with mMESSAGE mMACHINE T7 Transcription Kit (Invitrogen, AM1344) and manually reconstituted dNTP mixes with 20% m⁵CTP/CTP ratio. 5-methylcytidine-5-triphosphate was obtained from TriLink Biotechnologies (N-101405). Yielded RNA was purified by using the standard protocol of RNA Clean & Concentrator-5 (Zymo Research, R1013). The spike-in RNA mixes were then applied to RNA before fragmentation.

Total RNAs from whole cell or the chromatin-associated fractions were randomly fragmented by incubation at 94 °C for 4 min using 1× fragmentation buffer (NEB, E6186A). Fragmentation was stopped by adding 1× stop solution. Spike-in RNAs were added to each sample. Then, 4 μg anti-m⁵C antibody (Diagenode, MAb-081-100) was conjugated with 30 μl of protein G beads (Invitrogen, 1003D) in 300 μl IP buffer (10 mM Tris-HCl pH 7.5, 150 mM NaCl, 0.05% Triton X-100 (v/v), 1 mM spermidine) for 2 h at 4 °C on a rotating wheel. The same procedure was performed for a control reaction using mouse IgG isotype control (Abcam, ab37355). Bead–antibody complexes were washed three times with IP buffer and finally brought to 250 μl with IP buffer. After heat denaturation and quick chill on ice, 10 μg samples of RNA were added to the bead–antibody complexes and incubated with 1 μl SUPERase•In RNase Inhibitor (Invitrogen, AM2694) overnight at 4 °C on a rotating wheel. After several washes with IP buffer, RNA was incubated in 100 μl elution buffer (5 mM Tris-HCl pH 7.5, 1 mM EDTA, 0.05% SDS, and 200 μg proteinase K (Invitrogen, 25530049)) for 1 h at 50 °C. Beads were removed by centrifugation in a microcentrifuge, and the supernatant was purified with RCC-5 without size selection. Immunoprecipitated RNAs were eluted in water and then analysed using RT–qPCR. For next-generation sequencing, the immunoprecipitated RNAs were used as inputs for library constructions with the SMARTer Stranded Total RNA-Seq Kit v2—Pico Input Mammalian (TaKaRa Bio, 634411) according to the manufacturer’s instructions. Libraries were sequenced on the NovaSeq 6000 sequencer.

For analysing the effects of GC ratio and m⁵C modification levels, we designed three different in vitro transcription templates to get 70%, 50% or 30% GC ratio RNA products based on firefly luciferase mRNA (Supplementary Table 2). DNA oligos were purchased from Integrative DNA Technologies and annealed with a complementary DNA oligo (T7; Supplementary Table 2) to enable T7 DNA polymerase binding. In vitro transcription was performed using the mMESSAGE mMACHINE T7 Transcription Kit (Invitrogen, AM1344) and manually reconstituted dNTP mixes with a 0%, 0.2%, 2% or 20% m⁵CTP/CTP ratio. 5-methylcytidine-5-triphosphate was obtained from TriLink Biotechnologies (N-101405). Yielded RNA was purified using the standard protocol of the RNA Clean & Concentrator-5 (Zymo Research, R1013) kit. meRIP–qPCR experiments were performed according to the protocol mentioned above, and yeast tRNA (Invitrogen, AM7119) was mixed with RNA probes as a carrier.

RNA amplicon bisulfite sequencing

caRNAs were isolated from Tet2 WT or Tet2-KO mES cells as aforementioned. Ultrafast bisulfite (UBS) conversion was performed according to the published protocol²⁸. Reverse transcription was then performed with SuperScript III Reverse Transcriptase (Invitrogen, 18080093) using individual RT primers (Supplementary Table 2). The resulting cDNA was amplified for 10 cycles using NEBNext Ultra II Q5 Master Mix (NEB, M0544S) according to the standard protocol except that the T_m was set to 50 °C. Amplified DNA was quantified using the universal p5 primer (Supplementary Table 2) and p7 primer from NEBNext Multiplex Oligos for Illumina (NEB, E7500S). cDNAs amplified from different amplicons were then pooled together based on qPCR quantifications to achieve equal sequencing depth in the final DNA library. A final amplification was performed using the two primers (universal p5 primer and p7 primer from NEBNext Multiplex Oligos for Illumina) for 15 cycles using NEBNext Ultra II Q5 Master Mix (NEB, M0544S). PCR products were recovered using 1.0 volume of AMPure XP beads (Beckman Coulter, A63882) and subjected to sequencing on a NovaSEQ-X sequencer.

meDIP analysis

For methyl-DNA immunoprecipitation (meDIP) analysis, genomic DNA was extracted from cultured cells using the Monarch Genomic DNA Purification Kit (New England Biolabs, T3010S). Unmethylated lambda DNA (Promega, D1521) was spiked at a 0.5% ratio for quality control of the immunoprecipitation. DNAs were then fragmented to 200–1,000 bp by incubation for 22 min with NEBNext dsDNA Fragmentase (New England Biolabs, M0348S). The fragmented DNA was then denatured at 95 °C for 5 min and immediately cooled on ice for another 5 min. The input samples were removed and saved on ice for later use. The reaction was conducted in IP buffer (150 mM NaCl, 10 mM Tris-HCl, pH 7.5, 0.1% NP-40) at 4 °C overnight. The beads were then washed three times with IP buffer, followed by three washes by high-salt wash buffer (500 mM NaCl, 10 mM Tris-HCl, pH 7.5, 0.1% NP-40). Immunoprecipitated DNA was extracted by proteinase K digestion (Invitrogen, 25530049) before qPCR analysis. High-throughput sequencing libraries were constructed using xGen Methyl-Seq Lib Prep kits (IDT, 10009860) and sequenced on the NovaSEQ-X sequencer.

RNA synthesis rate assay

The RNA synthesis rate was measured with a procedure modified from the protocol Click-iT Nascent RNA Capture Kit, for gene expression analysis (Invitrogen, C10365). mES cells were seeded to 6 cm dishes at the same density in three replicates. After 42 h, cells were treated with 1 mM 5-ethynyl uridine for 10 min, 20 min and 40 min before RNA collection using TRIzol Reagent (Invitrogen, 15596026). Ribosomal RNA was depleted from total RNA preps before the click reaction with biotin azide (PEG4 carboxamide-6-azidohexanyl biotin). Biotinylated RNA was enriched using Dynabeads MyOne Streptavidin T1 (Invitrogen, 65601). ERCC RNA spike-in mix (Invitrogen, 4456740) was added to the eluted RNA with the amount proportional to the total RNA of each sample before rRNA depletion. Spiked RNAs were used as an input for RNA-seq library construction using the SMARTer Stranded Total RNA-Seq Kit v2—Pico Input Mammalian (TaKaRa Bio, 634411) according to the manufacturer’s instructions. Libraries were sequenced on the NovaSeq 6000 sequencer.

CUT&Tag analysis

Cleavage under targets and tagmentation (CUT&Tag) analysis was performed using the CUT&Tag-IT Assay Kit (Active motif, 53160) according to the manufacturer’s instructions. In brief, 0.2 million cells were used as an input for one replicate and washed with 1× wash buffer. Washed cells were conjugated to concanavalin A beads and permeabilized with Digitonin-containing buffer before incubation with primary antibodies (anti-H3K27me3, anti-H2AK119ub or normal rabbit IgG). Preassembled protein A-Tn5 transposome-enabled DNA tagmentation was performed after secondary antibody conjugation. Equal amounts of Drosophila spike-in chromatin preps (Active Motif, 53083) were added to each samples and subjected to the Tn5 tagmentation reaction. Tagged DNA was extracted by proteinase K digestion and amplified by PCR with indexed primers to yield DNA libraries. DNA libraries were subjected to qPCR analysis with gene-specific primers or high-throughput sequencing on the NovaSeq 6000 sequencer.

Construction of induced tethering mES cell lines

Cell lines stably expressing dCas13 protein fusion with catalytic domain of mouse TET2 (TET2-CD) or catalytic dead mutants were constructed first from WT mES cells. The coding sequence of dCas13 was cloned from plasmid pCMV-dCas13-M3nls, which was a gift from D. Liu (Addgene plasmid, 155366). The coding sequence of TET2-CD was cloned from the plasmid pcDNA3-FLAG-mTET2 (CD), which was a gift from Y. Xiong (Addgene plasmid, 89736), and the catalytic-dead mutant was cloned from the plasmid pcDNA3-Flag-Tet2 CD Mut, which was a gift from Y. Zhang (Addgene plasmid, 72220). pLR5-CBh-dCas9-hEzh2-IRES-Hyg was a gift from H. Ochiai (Addgene plasmid, 122375). The coding sequences of TET2-CD (or mutant) and dCas13 or dCas9 were fused. The fusion protein was delivered to mES cells with the piggyBac transposon system using the pLR5 vector and selected with hygromycin B (Gibco, 10687010). Sequences expressing guide RNA for dCas13 were cloned into a plasmid expressing a Tet operator controlled H1 operator (H1-2O2)⁵². This tet-pLKO-sgRNA-puro plasmid was a gift from N. Gray (Addgene plasmid, 104321). The guide-RNA expression plasmid was delivered into the TET2-CD-fusion protein-expressing mES cells by lentivirus. The resulting cell lines were selected with puromycin (Gibco, A1113803).

ASO and plasmid transfection in HSPCs

The steric-blocking antisense oligonucleotides (ASOs) (Integrated DNA Technologies) targeted to the hypermethylated motifs were fully modified with 2′-O-methoxyethyl (2′MOE) bases and phosphorothioate bonds, which were also incorporated with a fluorescent dye Cy5 at the 3′ end to monitor transfection efficiency. The NC5 ASO was used as a negative control that was not targeted to the human or mouse genome.

IAPEz-int 2′MOE: AGTTGAATCCTTCTTAACAGTCTGCTTTACGGGAAC

Sequence: /52MOErA/*/i2MOErG/*/i2MOErT/*/i2MOErT/*/i2MOErG/*/i2MOErA/*/i2MOErA/*/i2MOErT/*/i2MOErC/*/i2MOErC/*/i2MOErT/*/i2MOErT/*/i2MOErC/*/i2MOErT/*/i2MOErT/*/i2MOErA/*/i2MOErA/*/i2MOErC/*/i2MOErA/*/i2MOErG/*/i2MOErT/*/i2MOErC/*/i2MOErT/*/i2MOErG/*/i2MOErC/*/i2MOErT/*/i2MOErT/*/i2MOErT/*/i2MOErA/*/i2MOErC/*/i2MOErG/*/i2MOErG/*/i2MOErG/*/i2MOErA/*/i2MOErA/*/i2MOErC//3Cy5Sp/

MERVL 2′MOE: ACCATTACTGGGTATGTTAT

Sequence: /52MOErA/*/i2MOErC/*/i2MOErC/*/i2MOErA/*/i2MOErT/*/i2MOErT/*/i2MOErA/*/i2MOErC/*/i2MOErT/*/i2MOErG/*/i2MOErG/*/i2MOErG/*/i2MOErT/*/i2MOErA/*/i2MOErT/*/i2MOErG/*/i2MOErT/*/i2MOErT/* /i2MOErA/*/i2MOErT//3Cy5Sp/

NC5 2′MOE: GCGACTATACGCGCAATATG

Sequence: /52MOErG/*/i2MOErC/*/i2MOErG/*/i2MOErA/*/i2MOErC/*/i2MOErT/*/i2MOErA/*/i2MOErT/*/i2MOErA/*/i2MOErC/*/i2MOErG/*/i2MOErC/*/i2MOErG/*/i2MOErC/*/i2MOErA/*/i2MOErA/*/i2MOErT/*/i2MOErA/* /i2MOErT/*/i2MOErG//3Cy5Sp/

The crRNA targeting the primary m⁵C sites on IAPEz sequence based on our RNA bisulfite sequencing results was custom-synthesized and cloned into the pLentiRNAGuide_002-hU6-RfxCas13d-DR-BsmBI-EFS-EGFP:P2A:Puro-WPRE vector. The catalytic domain of mouse TET2 (mTET2-CD) or a catalytically dead mutant TET2(H1304Y/D1306A) (mTET2CDHxDCD) was cloned into the pLV[Exp]-[EF-1sc>[NLS-RfxCas13d]:[Linker]:P2A:mCherry(ns):T2A:Bsd vector. All of these plasmids were synthesized, constructed and confirmed by VectorBuilder.

All of the ASOs and plasmids were transfected into HSPCs using electroporation with the P3 Primary Cell 4D-Nucleofector X Kit S (Lonza Bioscience, V4XP-3032) with the program CV-137.

ASO transfections

We designed ASOs targeting the primary m⁵C sites on IAPEz or MERVL sequences based on our RNA m⁵C sequencing results. ASO transfections in mES cells were performed using the Lipofectamine RNAiMAX Transfection Reagent (Invitrogen, 13778075) according to the manufacturer’s instructions.

Cross-linking and immunoprecipitation and PAR-CLIP

Cultured mES cells or human leukaemia cells (SKM-1, WT and TET2^{−/−} THP-1 and K-562) were UV cross-linked at 254 nm with a Stratalinker (Stratagene) twice to achieve a 4,500 J m^−2 UV flux and then flash-frozen in liquid nitrogen. For photoactivatable ribonucleoside-enhanced crosslinking and immunoprecipitation (PAR-CLIP), 4-thiouridine was added to the cell culture medium 14 h before UVA irradiation (365 nm) three times, 1,500 J m^−2 each. The pellets were thawed on ice and resuspended in 3 volumes of ice-cold CLIP lysis buffer (50 mM HEPES pH 7.5, 150 mM KCl, 2 mM EDTA, 0.5% (v/v) NP-40, 0.5 mM DTT, 1 × Halt protease and phosphatase inhibitor cocktail (Thermo Fisher Scientific, 78442), 1 × RNaseOUT recombinant ribonuclease inhibitor (Invitrogen, 10777019)). The pellets were lysed by rotating at 4 °C for 15 min after passing through a 26 G needle (BD Biosciences). Embryo suspensions were sonicated on the Bioruptor system (Diagenode) with 30 s on/30 s off for 5 cycles. Lysates were cleared by centrifugation at 21,000g for 15 min at 4 °C on a benchtop centrifuge. The supernatants were applied to Flag-antibody-conjugated (Abcam, ab205606) protein A beads (Invitrogen, 1001D) and left overnight at 4 °C on an end-to-end rotor. The beads were washed extensively with 1 ml wash buffer (50 mM HEPES pH 7.5, 300 mM KCl, 0.05% (v/v) NP-40, 1 × Halt protease and phosphatase inhibitor cocktail, 1 × RNaseOUT recombinant ribonuclease inhibitor) at 4 °C five times. Protein–RNA complex conjugated to the beads was treated with 8 U μl^−1 RNase T1 (Thermo Fisher Scientific, EN0541) at 22 °C for 10 min with shaking. The input samples were digested in parallel. Then, input and IP samples were separated on an SDS–PAGE gel and gel slices at corresponding size ranges were treated by proteinase K (Invitrogen, 25530049) elution. RNA was recovered with TRIZol reagent (Invitrogen, 15596026). T4 PNK (Thermo Fisher Scientific, EK0031) end repair was then performed with purified RNA before library construction with the NEBNext Small RNA Library Prep Set for Illumina (NEB, E7330S). Libraries were pooled and sequenced on the NovaSeq 6000 sequencer.

Electrophoretic mobility shift assay

Recombinant MBD6-MBD–MBP–His protein was purified from Escherichia coli BL21 (DE3). Different concentrations of proteins were mixed with 100 nM FAM-labelled oligo probes in 1 × binding buffer (20 mM HEPES pH 7.5, 40 mM KCl, 10 mM MgCl₂, 0.1% Triton X-100, 10% glycerol and 1 × RNaseOUT Recombinant Ribonuclease Inhibitor (Invitrogen, 10777019)). The probe–protein mixture was incubated on ice for 30 min. The mixtures were loaded to a 10% Novex TBE Gel (Invitrogen, EC62755BOX). After gel running at 4 °C in 0.5× TBE for 2 h, the gel was washed twice in 0.5× TBE for 5 min. Washed gel was imaged with the GelDoc imaging system (Bio-Rad) with channel ‘FAM’. Individual K_D values were determined from a regression equation Y = [P]/(K_D + [P]), where Y is the fraction of probe bound at each protein concentration. The fraction bound is determined from the background-subtracted signal intensities using the expression: bound/(bound + unbound). [P] is protein concentration in each sample.

Quantitative analysis of RNA modification levels of CLIP RNA

Cultured mES cells were washed twice with DPBS before UV cross-linking at 254 nm with a Stratalinker (Stratagene) and flash-frozen in liquid nitrogen. The pellets were thawed on ice and resuspended in 3 volumes of ice-cold CLIP lysis buffer (50 mM HEPES pH 7.5, 150 mM KCl, 2 mM EDTA, 0.5% (v/v) NP-40, 0.5 mM DTT, 1 × Halt protease and phosphatase inhibitor cocktail (Thermo Fisher Scientific, 78442), 1 × RNaseOUT recombinant ribonuclease inhibitor (Invitrogen, 10777019)). The pellets were lysed by rotating at 4 °C for 15 min after passing through a 26 G needle (BD Biosciences). The cell suspensions were sonicated on the Bioruptor system (Diagenode) with 30 s on/30 s off for 5 cycles. Lysates were cleared by centrifugation at 21,000g for 15 min at 4 °C on a benchtop centrifuge. The supernatants were applied to Flag-antibody (Abcam, ab205606) conjugated protein A beads (Invitrogen, 1001D) and left overnight at 4 °C on an end-to-end rotor. Beads were washed extensively with 1 ml wash buffer (50 mM HEPES pH 7.5, 300 mM KCl, 0.05% (v/v) NP-40, 1 × Halt Protease and Phosphatase Inhibitor Cocktail, 1 × RNaseOUT Recombinant Ribonuclease Inhibitor) at 4 °C five times. Then, the input and IP samples were treated by proteinase K (Invitrogen, 25530049) to release cross-linked RNA. RNA was recovered with TRIZol reagent (Invitrogen, 15596026). Ribosomal RNA was then removed using the RiboMinus Eukaryote System v2 (Invitrogen, A15026) with purification and size-selection using the RNA Clean & Concentrator-5 (Zymo Research, R1013) kit. Recovered RNAs were subjected to digestion and MS/MS analysis.

Biotinylation of immunoprecipitated RNAs

Biotin labelling of immunoprecipitated RNA was performed according to a published protocol⁵³.

Fluorescence microscopy

For immunolabelling, cells were fixed with 4% PFA in DPBS at 37 °C for 5 min, permeabilized with methanol at −20 °C for 8 min, dried at room temperature for 10 min and then washed three times with DPBS at room temperature. The chambers were blocked in blocking buffer (DPBS, 0.5% BSA, 0.05% Triton X-100, 1:100 SUPERase·In (Invitrogen, AM2694)) for 1 h at room temperature and primary antibodies were diluted in blocking solution according to the suggested fold from the manufacturer’s and incubate at room temperature for 1 h. Chambers were washed three times with 0.05% Triton X-100 in DPBS, then 1:1,000 diluted goat anti rabbit IgG-AF568 conjugate (Invitrogen, A-11011) in blocking solution was added to each well and the chambers were incubated at room temperature for 1 h. The chambers were then washed three times with 0.05% Triton X-100 in DPBS and fixed with 4% PFA in DPBS for 30 min at room temperature and washed three times with DPBS. Nuclei were counterstained with 2 µg ml^−1 Hoechst 33342 (Abcam, ab145597) in DPBS at room temperature for 20 min, wash with DPBS three times. The chambers were stored at 4 °C before proceeding to imaging on a Leica SP8 laser-scanning confocal microscope at University of Chicago.

Lifetime profiling

Transcription inhibitor actinomycin D (Act D, Abcam ab141058) was applied to a final concentration of 2.5 μM in mES cell medium to cultured mES cells or cultured Lin^−KIT⁺ mouse HSPCs. Actinomycin D treatment started at 48 h after siRNA transfection (if any). RNAs were extracted from cells at different timepoints after actinomycin D treatment (10 min, 3 h and 6 h). Custom spike-in RNA (in vitro transcribed from firefly luciferase coding sequence) was added proportional to the yield of total RNA for different samples for RNA quantifications. RNA abundance was normalized to the value at 10 min for each condition.

DNA-seq data analysis

Raw reads were trimmed with Trimmomatic (v.0.39)⁵⁴ and then mapped to mouse genome (mm10) or human genome (hg38), together with Drosophila melanogaster chromatin (spike-in chromatin), using bowtie2 (v.2.4.1)⁵⁵ using the default mode, where multiple alignments are searched and the best one is reported. Mapped reads were deduplicated using the Picard tool MarkDuplicates (v.2.26.2; http://broadinstitute.github.io/picard/).

For ATAC–seq, reads that mapped to the mitochondrial genome were discarded before deduplication. Peaks were identified using MACS2⁵⁶ with the default mode, except for the parameters ‘–shift −75 –extsize 150 –nomodel –call-summits’. For CUT&Tag–seq, peaks were called using MACS2 with the default mode, except for the parameters ‘–broad –broad-cutoff 0.01’. For both ATAC–seq and CUT&Tag-seq, peaks that appeared in at least two biological replicates were retained for subsequent downstream analyses. The chromatin accessibility (ATAC) and H2AK119ub levels (CUT&Tag) were normalized by considering both sequencing depth and spike-in Drosophila melanogaster chromatin.

For meDIP–seq, differentially methylated regions were identified using MEDIPS⁵⁷ with the following settings: diff.method = ‘edgeR’, p.adj = ‘bonferroni’, MeDIP = True, CNV = False, minRowSum = 10. Regions with an adjusted P value of less than 0.1 were defined as significantly differentially methylated regions.

Nascent RNA-seq data analysis

Raw reads were trimmed with Trimmomatic (v.0.39)⁵⁴, and then aligned to mouse genome and transcriptome (mm10, version M19) as well as external RNA Control Consortium (ERCC) RNA spike-in control (Thermo Fisher Scientific) using HISAT2 (v.2.2.1)⁵⁸. Annotation files (version M19 for mouse) were obtained from GENCODE database (https://www.gencodegenes.org/)⁵⁹. Reads on each GENCODE annotated gene were counted using HTSeq (v.0.12.4)⁶⁰ and then normalized to counts per million (CPM) using edgeR packages in R⁶¹. CPM was converted to attomole by linear fitting of the RNA ERCC spike-in. The RNA level and EU adding time were fitted using a linear mathematical model, and the slope was estimated as transcription rate of RNA.

CLIP–seq data analysis

Low-quality reads were filtered using ‘fastq_quality_filter’, and adapters were clipped using ‘fastx_clipper’, then adapter-free reads were collapsed to remove PCR duplicates using ‘fastx_collapser’ and, finally, reads longer than 15 nucleotides were retained for further analysis (http://hannonlab.cshl.edu/fastx_toolkit/). Reads from rRNA were removed. The preprocessed reads were mapped to mouse genome (mm10) using bowtie (v.1.0.0)⁶² with ‘-v 3 -m 10 -k 1 –best –strata’ parameters. Mapped reads were separated by strands with samtools (v.1.16.1)⁶³ and peaks on each strand were called using MACS2 (v.2)⁵⁶ with parameter ‘-nomodel, –keep-dup 5, -g 1.3e8, -extsize 150’ separately. Significant peaks with q < 0.01 identified by MACS2 were considered. Peaks identified in at least two biological replicates were merged using bedtools (v.2.31.0)⁶³ and were used in the following analyses.

RNA-seq data analysis

Raw reads were trimmed with Trimmomatic (v.0.39)⁵⁴, then aligned to mouse (mm10) or human (hg38) genome and their corresponding transcriptome, together with external RNA Control Consortium (ERCC) RNA spike-in control (Thermo Fisher Scientific) when applicable, using HISAT2 (v.2.2.1)⁵⁸. Annotation files (version M19 for mouse, and version v29 for human in gtf format) were obtained from GENCODE database (https://www.gencodegenes.org/)⁵⁹. Reads were counted for each GENCODE annotated gene using HTSeq (v.0.12.4)⁶⁰ and for caRNAs using featureCounts⁶⁴, and then differentially expressed genes were called using DESeq2 package in R⁶⁵ with P < 0.05. In this step, the spike-in normalization factor was calculated by dividing the number of reads mapped to ERCC spike-ins by the total number of mapped transcriptomic reads. This factor was then included in the size factor calculation for DESeq2.

m⁵C meRIP–seq data analysis

Raw reads were trimmed with Trimmomatic (v.0.39)⁵⁴, then aligned to mouse (mm10) or human (hg38) genome and transcriptome, together with m⁵C modified or unmodified mRNA spike-ins (see the ‘m⁵C methylated RNA immunoprecipitation with spike-in’ section for details), using HISAT2 (v.2.1.0)⁵⁸. Annotation files (version M19 for mouse, and version v29 for human in gtf format) were downloaded from the GENCODE database (https://www.gencodegenes.org/)⁵⁹. Mapped reads were deduplicated using a Picard tool ‘MarkDuplicates’ (v.2.26.2) (http://broadinstitute.github.io/picard/). The remaining reads were separated by strands with samtools (v.1.16.1)⁶³ and peaks on each strand were called using MACS2 (v.2)⁵⁶ with the parameters ‘–nomodel –extsize 150’. Genome-specific parameters ‘-g hs’ for human and ‘-g mm’ for mouse were separately applied. We required significant peaks (q < 0.01) to appear in all biological samples to be considered validated. Peaks within the same conditions were then merged using bedtools (v.2.31.0)⁶³ for subsequent analysis. To quantify m⁵C methylation levels, we initially compared reads mapped to m⁵C-methylated spike-ins with those mapped to their unmethylated counterparts to confirm satisfactory pull-down efficiency. The m⁵C methylation levels were determined by calculating the log₂-transformed fold changes between immunoprecipitation) and input samples. The normalization factor was calculated by dividing the number of reads mapped to the m⁵C-methylated spike-in by the total number of transcriptomic reads. This approach enabled us to quantify the global changes in m⁵C levels under different conditions.

Chromatin-associated RNA UBS amplicon-seq analysis

Adapter sequences and low-quality reads were trimmed using cutadapt (v.4.0). Only properly paired reads with a length less than 20 nucleotides were retained. The 7 nucleotides of the UMI at the 5′ end of the insert fragments (R2) were extracted. Clean reads were then mapped to the mouse genome sequence (mm10) using the HISAT-3n tool⁶⁶ with the ‘–base-change C,T’ argument. To leverage the strand-specific property of the library, the ‘–directional-mapping’ parameter was applied. To increase the accuracy of site identification, only properly paired reads without soft clipping were retained. To eliminate unconverted clusters, reads containing more than three unconverted C sites, or where more than one-third of the total C sites were unconverted, were discarded. A binomial model was used to calculate a P value for each site, and sites with a P value less than 0.01 were classified as m⁵C sites.

Antibodies

The antibodies used in this study are summarized below: rabbit monoclonal anti-H2AK119ub antibody (Cell Signaling Technology, 8240S, 1:1,000 for western blot, 1:50 for CUT&Tag); rabbit monoclonal anti-H3 antibody (Cell Signaling Technology, 4499S, 1:1,000); mouse monoclonal anti-TET2 antibody (MilliporeSigma, MABE462, 1:500); rabbit monoclonal anti-GAPDH antibody, HRP conjugate (Cell Signaling Technology, 8884S, 1:1,000); rabbit monoclonal anti-DDDDK tag antibody (Abcam, ab205606, 1:1,000 for western blot, 1:50 for immunoprecipitation); rabbit polyclonal anti-SNRP70/U1-70K antibody (Abcam, ab83306, 1:1,000); mouse monoclonal anti-5-methylcytosine antibody (Diagenode, C15200081-100, 1:1,000 for dot blot, 1:50 for meRIP); mouse monoclonal anti-hm⁵C antibody (Diagenode, C15200200-100, clone Mab-31HMC, 1:1,000 for dot blot); rabbit monoclonal anti-H3K27me3 antibody (Cell Signaling Technology, 9733S, only for CUT&Tag experiments, 1:50); mouse monoclonal anti-BAP1 antibody (Santa Cruz, sc-28383, 1:50 for CUT&Tag). Goat anti-rabbit IgG, HRP conjugated antibody (Cell Signaling Technology, 7074S, 1:2,000) and horse anti-mouse IgG, HRP conjugated antibody (Cell Signaling Technology, 7076S, 1:2,000) were used as secondary antibodies. Mouse IgG-isotype control (Abcam, ab37355, 1:50 for immunoprecipitation) and rabbit IgG-isotype control (Abcam, ab37415, 1:50 for immunoprecipitation) were used as normal IgG controls. PerCP-Cy5.5 mouse lineage antibody cocktail (BD Biosciences, 561317, 1:100); PE rat anti-mouse CD117 (BD Biosciences, 553869, 1:100); Brilliant Violet 421 (BV421, 1:100) anti-mouse/human CD11b (Mac-1) (BioLegend, 101236, 1:100); APC mouse anti-human CD45 (BD Biosciences, 555485, 1:100); PE mouse anti-human CD33 (BD Biosciences, 561816, 1:100); PE-Cy7 rat anti-mouse CD45 (BD Biosciences, 552848, 1:100); PerCP-Cy5.5 mouse anti-mouse CD45.2 (BD Biosciences, 552950, 1:100) and FITC mouse anti-mouse CD45.1 (BD Biosciences, 553775, 1:100). All antibodies were applied at a dilution fold according to the manufacturer’s suggestions for specific use unless specified elsewhere in the Methods.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.