Mice

Two previously reported germline CD45.2 Sirt7−/− mouse models that contained exon 4 to 10 (ref. 24; Sirt7Δ4–10) or exon 4 to 9 deletions25 (Sirt7Δ4–9) and that were maintained on the 129Sv or C57BL/6 genetic background, respectively, were used. 129Sv Sirt7−/− mice were used for most experiments. IgHEL28 mice (CD45.2) were on the 129Sv background, whereas Pax5−/−5 and CD45.1 mice were on the C57BL/6 background. Heterozygous CD45.1/CD45.2 mice were generated by crossing CD45.1 C57BL/6 and wild-type CD45.2 129Sv mice for one generation to avoid rejection following transplantation of 129Sv cells. All mice were bred at the Comparative Medicine and Bioimage Centre of Catalonia animal facility of the Germans Trias i Pujol Research Institute. Animal studies were conducted at Josep Carreras Leukemia Research Institute (IJC) (Spain) according to national authorities and institutional ethics committees (Germans Trias i Pujol Reserach Institute Ethics Committee). The collection of BM samples from C57BL/6 wild-type and Sirt7−/− mice and the generation of Pax5−/− mouse B cell progenitors were conducted according to national authorities and institutional ethics committees at Max Plank Institute for Heart and Lung Research (MPI-HLR) (Germany) and Lund University (Sweden), respectively. Mouse housing conditions included a temperature of 21–25 °C, 40–70% humidity and a light cycle from 0800 to 2000 h with a 15-min intensity ramp simulating sunrise and sunset.

Determination of anti-HEL isotypes

To analyze antigen-specific responses, mice were intraperitoneally injected with 100 μl of an emulsion containing 50 μg of NP–HEL along with complete Freund adjuvant. After 14 days, mice were bled, and serum samples were collected. For antibody isotype detection, enzyme-linked immunosorbent assay plates were coated overnight at 4 °C with HEL peptide (6 μg ml–1). Following removal of excess peptide, serum dilutions from control or immunized mice were added and incubated for 1 h at room temperature. After three washes with a 0.05% Tween–PBS solution, rat anti-mouse immunoglobulin subclasses (IgM, IgG1 and IgG3; 1:250) were added for 1 h at room temperature. After 3 washes, an anti-rat immunoglobulin (1:5,000) was added for 1 h at room temperature, followed by three additional washes. Finally, TMB developing solution was added for 15 min and stopped with 1 N H2SO4. Color production was measured at 450 nm and 570 nm (for background subtraction) using a Multiskan Sky (Thermo Fisher) plate reader.

Transplantation experiments

Two million short-term expanded pro-B cells or 1.5 million splenic Lin–IgM+IgD+ B cells were washed and resuspended in 200 µl of PBS supplemented with 1% heat-inactivated fetal calf serum. Cells were injected via the tail vein into sublethally irradiated (5 Gy) CD45.1/CD45.2 or CD45.2 6- to 10-week-old randomized recipient mice. Four weeks after transplantation, the spleen (CD45.1/CD45.2 mice) or BM and thymus (CD45.2 mice) were collected and analyzed by fluorescence-activated cell sorting (FACS).

Flow cytometry and cell sorting

BM, spleen and thymus samples were collected from wild-type, Sirt7−/−, IgHEL+/–, Sirt7−/−IgHEL+/–, CD45.1/CD45.2 and CD45.2 mice. BM samples were crushed in staining buffer (3% fetal bovine serum (FBS) and 2 mM EDTA in PBS) to obtain single-cell suspensions. Spleen and thymus samples were similarly processed. Red blood cells were lysed in ACK buffer (Gibco), and the reaction was stopped by adding 5 volumes of staining buffer. Cells were filtered through 40-μm sterile strainers and incubated with Fc-block (eBioscience) before staining for 30 min (4 °C) with anti-B220 (RA3-6B2, eBioscience, 1:400), anti-CD19 (HIB19, eBioscience, 1:400), anti-CD43 (eBioR2/60 or 1G10, BD Biosciences, 1:800), anti-IgM (II/41, eBioscience, 1:800), anti-IgD (11-26c, eBioscience, 1:400), anti-CD21 (7G6, BD Biosciences, 1:800), anti-CD23 (B3B4, BD Biosciences, 1:200), anti-CD93 (AA4.1, eBioscience, 1:400), anti-Gl7 (Gl7, eBioscience, 1:200), anti-CD38 (90, eBioscience, 1:400), anti-CD138 (300506, Invitrogen, 1:400), anti-Fas (SA367H8, Biolegend, 1:600), anti-IgG1 (A85-1, BD Pharmigen, 1:600), anti-CD127 (eBioSD/199, eBioscience, 1:200), anti-CD45.1 (A20, eBioscience, 1:400), anti-CD45.2 (104, eBioscience, 1:200), anti-TCRβ (H57-597, BD Biosciences, 1:200), anti-NKP46 (29A1.4, Biolegend, 1:100), anti-CD4 (GSK1.5, eBioscience, 1:800), anti-CD8 (53-6.7, eBioscience, 1:800), anti-hCD4 (RPA-T4, Biolegend, 1:400), anti-CD3e (145-2C11, BD Biosciences, 1:400), anti-Ly76 (TER-119, BD Biosciences, 1:400), anti-CD11b (M1/70, BD Biosciences, 1:400) or anti-GR1 (RB6-8C5, BD Biosciences, 1:400). After staining, cells were washed twice and resuspended in staining buffer before analysis with a FACS Canto II (BD Biosciences) or sorting with a FACSAria II (BD Biosciences). B cell subsets are defined as pre-pro-B (B220+CD19−), pro-B (B220+CD19+IgM−CD43+), pre-B (B220+CD19+IgM−CD43−), large pre-B (B220+CD19+IgM−CD43−FSChi), small pre-B (B220+CD19+IgM−CD43−FSClo), immature B (B220+CD19+IgM+), BM mature B (B220hiCD19+), marginal zone B (B220+CD19+CD21hiCD23−), transitional B (B220+CD19+CD21+CD23+CD93+), follicular B (B220+CD19+CD21+CD23+CD93−), germinal center B (B220+CD19+IgM+Gl7+Fas+), memory B (CD19+CD38+CD138−Gl7−), class-switched IgG1+ B cells (B220+IgG1+) and plasma cells (B220loCD138+).

Apoptosis was measured in BM pro-B and pre-B cells stained with 7AAD (BD Biosciences) and Annexin V-FITC (Abcam), following the manufacturer’s instructions. To measure cell cycle distribution, BM pre-B cells were stained with surface markers and fixed and incubated with a solution containing 1 µg ml–1 DAPI in permeabilization buffer for at least 1 h before FACS analysis. All flow cytometry experiments were analyzed with FlowJo.

For SIRT7 and Pax5 intracellular staining, BM single-cell suspensions were stained and washed before fixation for 30 min and permeabilization with Foxp3/Transcription Factor Fixation/Permeabilization buffers (eBioscience) according to manufacturer’s instructions. Permeabilized cells were blocked with 2% FBS for 10 min, stained with anti-SIRT7 (D3K5A, Cell Signaling, 0.25 µg per sample) or anti-Pax5 (1H9, eBioscience, 0.2 µg per sample) for 1 h at room temperature and washed twice with permeabilization buffer containing 2% FBS. For SIRT7 staining, cells were subsequently incubated with a polyclonal anti-IgG (H + L) secondary antibody (Invitrogen, 0.25 µg per sample) for 1 h at room temperature before FACS analysis.

For phospho-STAT5-Y694 staining (47/Stat5(pY694), BD Biosciences, 20 µl per sample) in pre-B cells, BM cells were incubated in RPMI (Gibco) for 30 min at 37 °C and stimulated with IL-7 (5 ng ml–1) in RPMI for an additional 30 min at 37 °C. Cells were fixed with Foxp3/Transcription Factor Fixation buffer, followed by washing and incubation with ice-cold 100% methanol for 1 h. After Fc blocking, cells were stained and analyzed with a FACS Canto II.

Isolation of primary B cell progenitors

BM wild-type and Sirt7−/− pro-B cells were isolated by magnetic-activated cell separation enrichment of CD19+ cells, followed by cell sorting. Briefly, BM single-cell suspensions were prepared in staining buffer. Cells were stained with Fc-block for 20 min and an additional 30 min with a biotinylated antibody to CD19 (1D3, BD Biosciences, 0.1 µg per 10 million cells). After washing, stained cells were incubated with Streptavidin MicroBeads (Miltenyi) and separated magnetically. CD19+ cells were cultured overnight in Opti-MEM supplemented with 10% heat-inactivated FBS, 25 mM HEPES, 50 µg ml–1 gentamicin and 50 µM β-mercaptoethanol in the presence of 10 ng ml–1 IL-7, 10 ng ml–1 SCF and 10 ng ml–1 FTL3-L before Lin−CD19+B220+IgM− cell sorting. Fetal liver Pax5−/− B cell progenitors were obtained as described previously13.

Cells and reagents

HAFTL, TANOUE, NALM-20, REH, KOPN-8, SD-1, SEM and TOM-1 cells were cultured in RPMI supplemented with 10% heat-inactivated FBS and 100 U ml–1 penicillin/streptomycin (Gibco), whereas HEK293F, SIRT7−/− HEK293F (described in Simonet et al. 42) and Platinum E cells were grown in DMEM (Gibco) supplemented with 10% FBS and 100 U ml–1 penicillin/streptomycin. OP9 cells were maintained in MEM-α (Gibco) supplemented with 20% FBS and 100 U ml–1 penicillin/streptomycin. KOPN-8, NALM-20, REH, TANOUE, SD-1, TOM-1 and SEM cells were kindly provided by M. Parra (IJC; purchased from the DSMZ German Collection of Microorganisms and Cell Cultures). OP9 and HEK293F cells were purchased from ATCC, and Platinum E cells were purchased from Cell Biolabs. None of the cell lines used were found in the Commonly misidentified lines database. Primary pro-B cells from wild-type, Sirt7−/− or Pax5−/− mice were plated onto a layer of mitomycin C-inactivated OP9 feeder cells and grown on Opti-MEM supplemented with 10% heat-inactivated FBS, 25 mM HEPES, 50 µg ml–1 gentamicin and 50 µM β-mercaptoethanol in the presence of 10 ng ml–1 IL-7, 10 ng ml–1 SCF and 10 ng ml–1 FTL3-L. All cells were cultured at 37 °C in a humidified atmosphere containing 5% CO2. For transient transfections, HEK293F or SIRT7−/− HEK293F cells were transfected using polyethylenimine and the corresponding plasmids. For retroviral transduction of pro-B cells and B-ALL cell lines, Platinum E cells were transiently transfected with polyethylenimine, a pVSV-G vector encoding the viral envelop and pMIG bicistronic vectors encoding either the selection marker (hCD4 or GFP) alone or together with the SIRT7WT, SIRT7H187Y, Pax5WT, Pax5K198Q or Pax5K198R coding sequences. Pro-B cells were resuspended in retroviral supernatants, centrifuged for 1.5 h at 1,000g (32 °C) and selected by hCD4+ cell sorting 96 h after infection. Treatments were performed with 5 mM nicotinamide (Sigma) for 48 h, 100 µg ml–1 cycloheximide (Sigma-Aldrich) for the indicated times, 2 µM lactacystin (Santa Cruz Biotechnology) for 8 h and 1 µM trichostatin A (Sigma-Aldrich) for 3 h.

Histology

Spleens from wild-type and Sirt7−/− mice were collected, fixed in 10% formalin for 24 h, embedded in paraffin and sectioned at 4 µm before staining with hematoxylin and eosin. Histological sections were visualized with an Olympus BX53 microscope.

Immunoprecipitation, gel filtration high-performance liquid chromatography and immunoblotting

For immunoprecipitation, cell pellets were lysed in RIPA buffer (50 mM Tris-HCl (pH 8.0), 150 mM NaCl, 0.5% sodium deoxycholate, 0.1% SDS, 1% NP-40 and 2 mM MgCl2) containing cOmplete Protease Inhibitor (Roche) and incubated for 8 h with benzonase nuclease (Millipore) at 4 °C. Cell lysates were clarified by centrifugation (17,000g for 10 min at 4 °C) and incubated overnight with anti-Flag beads (Millipore) at 4 °C with gentle rotation. The immunoprecipitated protein complexes were washed five times with lysis buffer (20 mM Tris-HCl (pH 8.0), 500 mM NaCl, 10% glycerol and 1 mM EDTA) and eluted with Laemmli buffer supplemented with 10% β-mercaptoethanol. Samples were then boiled at 95 °C for 5 min and analyzed by immunoblotting. Densitometric quantification of immunoblotting experiments was performed with ImageJ software.

Cellular fractionation experiments were performed using the Dignam method, as described in Simonet et al.42. For size-exclusion chromatography analysis, nuclei from HAFTL cells were purified and lysed under native conditions according to the Dignam method, as previously described42. Nuclear lysates were incubated overnight with benzonase nuclease before clarification and concentration with Amicon Ultra centrifugal filters (Millipore). Concentrated nuclear lysates were then fractionated by molecular weight on the gel filtration column Superose 6 (Cytiva) with a fractionation range of 5 × 103–5 × 106 Da. The eluted fractions containing size-excluded proteins and protein complexes were denatured in Laemmli buffer supplemented with 10% β-mercaptoethanol and analyzed by immunoblotting. The following antibodies were used for immunoblotting: anti-SIRT7 (D3K5A, Cell Signaling, 1:1,000), anti-Pax5 (D19F8, Cell Signaling, 1:1,000), anti-H3 (ab1791, Abcam, 1:1,000), anti-fibrillarin (B1, Santa Cruz Biotechnology, 1:1,000), anti-acetyl-lysine (9814, Cell Signaling, 1:200), anti-Flag (M2, Sigma-Aldrich, 1:10,000), anti-HA (6908, Sigma), anti-Myc (9B11, Cell Signaling, 1:1,000), anti-V5 (ab9116, Abcam, 1:1,000) and anti-actin (A1978, Sigma, 1:5,000).

SIRT7 and Pax5 purification and in vitro deacetylation assay

SIRT7−/− HEK293F cells were transiently transfected with vectors encoding Pax5-Myc–Flag or SIRT7–Flag for 48 h. Before collection, cells expressing these constructs were treated overnight with 5 mM nicotinamide and for 3 h with 1 µM trichostatin A to hyperacetylate Pax5. Cell pellets were lysed, incubated with benzonase nuclease for 8 h, clarified and incubated overnight with anti-Flag beads (Millipore). The immunoprecipitated protein complexes were washed five times with BC500 buffer (20 mM Tris-HCl (pH 8.0), 500 mM NaCl, 10% glycerol and 1 mM EDTA), eluted with synthetic Flag peptide (0.6 µg ml–1; GenScript) and dialyzed in BC100 buffer (20 mM Tris-HCl (pH 8.0), 100 mM NaCl, 10% glycerol and 1 mM EDTA). Purified PAX5 and SIRT7 proteins were incubated for 1 h at 37 °C, with or without 1.25 mM NAD+, in deacetylation buffer (10 mM Tris-HCl (pH 8.0), 150 mM NaCl, 1 mM DTT and 10% glycerol), and the reaction was stopped with 5× Laemmli buffer containing 10% β-mercaptoethanol. PAX5 acetylation was determined by immunoblotting using anti-pan-acetyl-lysine.

Pre-B cell proteome and Pax5 acetylation analysis

For determination of the pre-B cell proteome, pre-B cells were sorted from the BM of wild-type and Sirt7−/− mice. Proteins were extracted with 6 M urea, 100 mM Tris (pH 8.0) and the help of a bioruptor, quantified using a NanoDrop at 280 nm and precipitated with trichloroacetic acid/acetone. Samples were then reduced and alkylated with 10 mM DTT and 55 mM chloroacetamide, respectively. Proteins were then resuspended in 6 M urea and 100 mM Tris (pH 8.0) and digested with LysC/trypsin. LysC digestion was performed for 16 h, while trypsin was added for 8 h, and both reactions were performed at 30 °C. The reactions were stopped with 10% formic acid. The peptides were then desalted with a C18 reverse-phase ultramicrospin column and desiccated in a speedvac. Total proteome samples were separated using a C18 analytical column (nanoEaseTM M/Z HSS C18 T3; 75 µm × 25 cm, 100 Å; Waters) with a 180-min run comprising three consecutive steps with linear gradients from 3% to 35% B in 150 min, from 35% to 50% B in 5 min and from 50% to 85% B in 2 min, followed by isocratic elution at 85% B in 5 min and stabilization to initial. The mass spectrometer was operated in data-dependent acquisition mode, and the data were acquired with Xcalibur software 4.0.27.10 (Thermo Scientific).

For identification of acetylated Pax5 residues, PAX5 was purified from SIRT7−/− HEK293F cells transiently expressing Pax5-Myc–Flag together with an empty vector or a vector encoding SIRT7–Flag. PAX5-containing beads were washed three times with 100 mM Tris (pH 8.0) and then resuspended in 6 M urea and 100 mM Tris (pH 8.0). Reduction and alkylation were then performed by using 10 mM and 55 mM chloroacetic acid. The digestion was performed by adding 1 μg of trypsin for 16 h at 30 °C. Finally, the digestion was stopped with 10% formic acid, and the peptides were desalted with a polyLC C18 pipette tip and dried in a speedvac. The acetylomes were separated using an Evosep EV1000 column (150 μm × 150 mm, 1.9 μm; Evosep) with an 88-min run. The spectrometer was working in positive polarity mode, and single-charge state precursors were rejected for fragmentation. The data were acquired with Xcalibur software 4.2.28.14 (Thermo Scientific).

For both total proteome and acetylome analyses, the peptides were reconstituted with 3% acetonitrile and 0.1% formic acid aqueous solution at 100 ng µl–1, and 800 ng was injected into the mass spectrometer.

Semiquantitative PCR, RT–qPCR and RNA-seq

Semiquantitative PCR of Igh segments was performed as described in Ng et al.43 using genomic DNA extracted from sorted splenic IgM+ cells and degenerate primers. RNA for RNA-seq and RT–qPCR was extracted from frozen pellets using a Maxwell RSC simplyRNA Tissue kit (Promega). For RT–qPCR, cDNA was synthesized using a Transcriptor First Strand cDNA Synthesis kit (Roche) according to manufacturer’s instructions. RT–qPCR reactions were performed in a QuantStudio 5 Real-Time PCR System. Primer sequences (Integrated DNA Technologies) are shown in Extended Data Table 1.

RNA for RNA-seq was extracted from sorted wild-type and Sirt7−/− BM pro-B cells and pre-B cells or Pax5−/− ex vivo-expanded B cell progenitors expressing an empty vector, PAX5-WT, PAX5-K198Q or PAX5-K198R. After library construction, 150-bp paired-end sequencing was performed on a DNBSEQ-G400 (MGI Tech).

ChIP–seq

Chromatin was prepared using a truChIP Chromatin Shearing kit (Covaris) with modifications. Ten million pro-B cells were collected and washed once in PBS before fixation in 1 mg ml–1 DSG (Thermo Fisher) for 30 min, followed by an additional 10-min incubation with 1% formaldehyde. The cross-linking reaction was stopped by adding 1/20 quenching buffer, and fixed cells were washed twice with ice-cold 0.5% bovine serum albumin (wt/vol) in PBS. Nuclei were purified and sonicated following the manufacturer’s instructions. One volume of 2× dilution buffer supplemented with 0.1% SDS and protease inhibitors was added before clarifying the lysate (10,000g, 5 min, 4 °C). ChIP was performed overnight at 4 °C with 10 µg of anti-Pax5 (ab183575, Abcam) previously conjugated with 20 µl of Pierce ChIP-grade Protein A/G Magnetic Beads (Thermo Fisher) for 4 h at 4 °C. After immunoprecipitation, samples were serially washed once with low-salt wash buffer (0.1% SDS, 1% Triton, 2 mM EDTA, 20 mM Tris-HCl (pH 8.1) and 150 mM NaCl), high-salt wash buffer (0.1% SDS, 1% Triton, 2 mM EDTA, 20 mM Tris-HCl (pH 8.1) and 500 mM NaCl) and LiCl immune complex wash buffer (10 mM Tris-HCl (pH 8), 250 mM LiCl, 1% NP-40, 1% sodium deoxycholate and 1 mM EDTA) supplemented with protease inhibitors and twice with modified TE buffer (0.1 mM EDTA and 10 mM Tris-HCl (pH 8)). Cross-linking was reversed in elution buffer (0.1 M NaHCO3 and 1% SDS) with 10 µg of RNase A (Thermo Fisher) for 30 min at 37 °C, followed by an additional 6-h incubation at 65 °C with 50 µg of Proteinase K (Apollo Scientific). DNA was subsequently cleaned up using NucleoSpin Gel and PCR clean-up columns (Macherey-Nagel) before library construction and 100-bp paired-end sequencing on a DNBSEQ-G400.

Public data analysis

Mouse and human BM 10x scRNA-seq data were derived from the Broad Institute Single Cell Portal (https://singlecell.broadinstitute.org; projects ‘Bone Marrow from B6 Mice, 10x’ and ‘A Census of Immune Cells’, respectively). For human scRNA-seq analysis, the associated Loom file was downloaded and reanalyzed using Scanpy44. Cell-type annotation was performed using DecoupleR (v1.34)45 and PanglaoDB46. Mouse scRNA-seq feature and t-distributed stochastic neighbor embedding plots were directly downloaded from the Single Cell Portal. Microarray normalized data from mouse B cell progenitors were from the Immgen Consortium datasets47 (accession number GSE15907). Pax5-regulated genes were retrieved from accession number GSE38046 (ref. 20). Pax5 peaks were obtained from a publicly available PAX5 ChIP–seq dataset17. The list of mammalian proteins with acetyltransferase activity was from a previous report31, whereas Pax5 protein-protein interactions (PPIs) combined the data from another study19 with the curated interactions compiled in the BioGRID repository. Normalized proteomics and RNA-seq data for individuals with B-ALL were from Yang et al.29. Proteomics and RNA-seq raw data from this project are available at the ProteomeXchange Consortium (identifier PXD010175) and at the European Genome–Phenome Archive (accession EGAS00001003079), respectively. Relative Pax5 and SIRT7 protein levels in B-CLL samples were obtained from Johnston et al.34, and raw data are deposited at the PRIDE archive (PXD002004). Children’s Oncology Group clinical trial P9906 RNA expression and outcome data are available at the Genomic Data Commons (https://portal.gdc.cancer.gov) and were generated by the Therapeutically Applicable Research to Generate Effective Treatments (https://www.cancer.gov/ccg/research/genome-sequencing/target; GSE11877) initiative, whereas PAX5 deletion data in these individuals were retrieved from Roberts et al.37 (EGAS00001000654).

Proteomics data analysis

The raw thermo files were processed with MaxQuant 1.6.7.0 using a mouse database downloaded from https://www.uniprot.org/ (December 2018) or with PEAKS X+ software using a mouse database downloaded from https://www.uniprot.org/ (November 2019) for the total proteome and acetylome, respectively. In both cases, only reviewed entries were included (16,997 or 17,024 entries, respectively). The search was performed using the following parameters: trypsin was selected as the enzyme, and a maximum of two or three missed cleavages was allowed for the total proteome and acetylome. For the total proteome, the modifications were carbamidomethylation as a fixed modification, whereas oxidation in methionines and acetylation of protein N termini were used as variable modifications. The iBAQ intensity was used to quantify the proteins. Alternatively, for the acetylome, carbamidomethylation was set as a fixed modification, whereas oxidation in methionines; acetylation at both lysines and the protein N terminus; deamidation at asparagine and glutamine; phosphorylation at serines, threonines and tyrosines and dehydration at aspartic acid, tyrosines, threonines, serines, glutamine and asparagine were used as variable modifications. The mass tolerance for the parental ion and MS/MS fragments were set to 10 ppm and 0.5 Da. In both analyses, the results were filtered at 1% FDR at the peptide and protein levels.

Data processing and statistical analyses were performed using R (https://cran.r-project.org/) and RStudio (https://www.rstudio.com/) software. Total proteome statistical analysis was performed with the ‘limma’ package. For the PAX5 acetylome, the peptide spectrum matches were calculated by using a homemade algorithm and the PEAKS’ output files ‘peptide.csv’ and ‘DB search psm.csv’.

Bulk RNA-seq bioinformatic analysis

Bulk RNA-seq raw FastQ files were quality checked using FastQC, and raw counts were obtained by performing Salmon48 ‘quant’ pseudoalignment (mm10 reference mouse genome), with fragment-level GC bias correction (–gcBias), eight threads (-p 8) and selective alignment enabled (–validateMappings). The quant.sf files were used to import transcript-level quantification data into a Summarized Experiment using Tximeta49. The Summarized Experiment was imported into an R environment, and differential gene expression analysis was performed using DESeq2 (ref. 50). Counts were normalized by DESeq2’s median of ratios method51, and normalized counts were transformed to z score for data visualization (ggplot2). For generating BigWig files, raw FastQ files were mapped onto a reference mouse genome (mm10) using Bowtie2 (ref. 52) to generate SAM files. SAM files were converted to BAM files using SAMTools53, skipping alignments with MAPQ smaller than 37 (samtools view -bS -q 37). BAM files were sorted and filtered using Sambamba54, eliminating unmapped and duplicated reads (-F ‘[XS] == null and not unmapped and not duplicate’). Index BAI files were generated using SAMTools ‘index’, and BigWig files were created using deepTools55 ‘bamCoverage’ (–binSize 20–normalizeUsing BPM–ignoreForNormalization chrX–extendReads 150–centerReads–smoothLength 60). Visualization of BigWig files was performed using IGV56.

GSEA was performed using the software57 provided by the Broad Institute at https://www.gsea-msigdb.org/ with RNA-seq normalized count values and default parameters. Gene Ontology terms were obtained with the Enrichr tool58.

ChIP–seq bioinformatic analysis

ChIP–seq raw FastQ files were quality checked using FastQC and aligned onto a reference mouse genome (mm10) using Bowtie2 (ref. 52) to generate SAM files. SAM files were converted to BAM files using SAMTools53, skipping alignments with MAPQ smaller than 37 (samtools view -bS -q 37). BAM files were sorted and filtered using Sambamba54, eliminating unmapped and duplicated reads (-F ‘[XS] == null and not unmapped and not duplicate’). Index BAI files were generated using SAMTools ‘index’, and BigWig files were created using deepTools55 ‘bamCoverage’ (–binSize 20–normalizeUsing BPM–ignoreForNormalization chrX–extendReads 150–centerReads–smoothLength 60). Visualization of BigWig files was performed using IGV56. Peak calling was performed using MACS2 (ref. 59) ‘callpeak’ (-f BAM–nomodel–extsize 20 -g mm -B) taking into account both immunoprecipitation (-t) and input (-c) samples for each peak calling. Only significant peaks (q < 0.05) were included in downstream analyses. Differential peak analysis was performed with ChIPpeakAnno, and heat maps were generated with DiffBind or plotHeatmap.

Statistics and reproducibility

All statistical analyses were performed using GraphPad Prism 8.0.1. The statistical tests performed for data analysis are indicated in the corresponding figure legends. Individual P values of significant data are indicated in the figures. Results were similarly replicated in at least two independent experiments, and results pooled from independent experiments are indicated in the corresponding figure legends. For experiments involving mice, each replicate corresponds to an individual mouse, and no animals or data points were excluded. RNA-seq, ChIP–seq and proteomics replicates were generated on different days. No statistical methods were used to predetermine sample sizes. Data distribution was assumed to be normal, but this was not formally tested. Only recipient mice for transplantation experiments were randomized because the rest of the experiments required mouse genotyping and were performed on genetically identical mice. Data collection and analysis were not performed blind to the conditions of the experiments.

Reporting summary

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