Animal experimentation at the IRB Barcelona was performed according to protocols approved by the Science Park of Barcelona (PCB) Ethics Committee for Research and Animal Welfare. Mice were housed in a specific pathogen-free facility on a 12-hour light–dark cycle at an ambient temperature of 20–24 °C and humidity of 30–70%. Adult mice were fed ad libitum with SAFE R40 pellet diet (https://safe-lab.com/safe_en/) containing 0.02 mg per kg body weight vitamin B12. In general, mice of 8–16 weeks of age of both sexes were treated with 1 mg ml−1 doxycycline hyclate BioChemica (PanReac, A2951) in the drinking water (supplemented with 7.5% sucrose) for 7 d. Antibiotic treatment was conducted using a broad-spectrum cocktail (1 mg l−1 each of ampicillin (BioChemica, A0839), neomycin sulfate and metronidazole (Sigma, M1547); 0.5 mg l−1 vancomycin (Cayman Chemical, CAY-15327) all dissolved in water supplemented with 7.5% sucrose) for 3 weeks before doxycycline initiation and was maintained during doxycycline treatment. Vitamin B12 (Sigma, V2876) supplementation was provided at 1.25 mg l−1 and folate supplementation was provided as folic acid (Sigma, F7876) at 40 mg l−1 in the drinking water, both for 7 d concomitant with doxycycline treatment. For the B12 bolus experiment, mice were administered 5 µg vitamin B12 (Sigma, V2876) dissolved in water by oral gavage on day 6 after the start of doxycycline treatment, and blood samples were taken by submandibular collection just before and 24 h after the bolus. OSKM transgenic mice are the i4F-B strain (derived on a C57/BL6J background and bred in house) described in ref. 3 and are available upon request. WT mice were i4F-B WT littermate controls where specified, or WT C57/BL6J (Charles River France).
Mice were treated with 2.5% (wt/vol) DSS, colitis grade (36,000–50,000; MP Biomedicals, MFCD00081551) in drinking water for 5 consecutive days. On day 5, the DSS was removed and drinking water was supplemented with doxycycline hyclate BioChemica (1 mg ml−1; PanReac, A2951; with 7.5% sucrose) for 48 h, after which regular water was returned. Mice in the B12 experimental group also received supplementation of vitamin B12 (1.25 mg l−1; Sigma, V2876) from the point of DSS removal (that is, day 5) until experimental endpoint. The MAT2Ai group received FIDAS-5 (MedChemExpres, HY-136144) and were dosed with 20 mg per kg body weight per day dissolved in PEG400 by oral gavage as previously described79.
FITC–dextran intestinal permeability assay
On day 9 (relative to the start of DSS administration), food was withdrawn from mice for 4 h, after which mice were gavaged with FITC–dextran (MW 4,000; Sigma-Aldrich, FD4) at a dose of 44 mg per 100 g of body weight dissolved in PBS. Food restriction was maintained for 3 additional hours, at which point blood was sampled by submandibular vein bleeding. Whole blood was diluted at a ratio of 1:4 in PBS, and 100 µl of blood/PBS mixture from each mouse was loaded into a 96-well plate. Fluorescence intensity was measured on a BioTek Synergy H1 Microplate Reader (excitation 490 nm; emission 520 nm).
Microbial analysis from murine stool samples
Fresh stool samples were collected directly from mice and snap frozen. gDNA was isolated using a QIAamp DNA Stool Mini Kit (QIAGEN, 51504) according to the manufacturer’s protocols.
Library preparation and sequencing
Libraries were prepared using the NEBNext Ultra DNA Library Prep Kit for Illumina (E7370L) according to the manufacturer’s protocol. Briefly, 50 ng of DNA was fragmented to approximately 400 bp and subjected to end repair plus ‘A’-tailing, ligation of NEB adaptor and Uracil excision by USER enzyme. Then, adaptor-ligated DNA was amplified for eight cycles by PCR using indexed primers. All purification steps were performed using AMPure XP Beads (A63881). Final libraries were analysed using an Agilent DNA 1000 chip to estimate the quantity and check size distribution, and were then quantified by qPCR using the KAPA Library Quantification Kit (KK4835, KapaBiosystems) before amplification with Illumina’s cBot. Libraries were sequenced (2 × 125 bp) on Illumina’s HiSeq 2500.
Reads were aligned to the mm10 genome using STAR 2.7.0a with default parameters80. DNA contaminated reads were filtered out from the analysis. The first and final ten bases of the non-contaminated reads were trimmed using DADA2 1.10.1 (ref. 81). Taxonomic assignments were carried out through Kaiju 1.7.0 (ref. 82) using the microbial subset of the NCBI BLAST non-redundant protein database (nr). Resulting sequencing counts were aggregated at genus level. Reads that could not be assigned to any specific genus were classified to the nearest known taxonomic rank (marked by the term _un). The gut microbial compositional plot displays the relative abundances (percentage) at genus level. Only the 17 most abundant taxa are shown, while the rest were moved to the ‘others’ category. For all genera, the treatment effect (finish versus start) was compared between OSKM and control (WT) mice. This was accounted in a model with an interaction term (drug:treatment) using DESeq2 with default options83. The paired nature of the experimental design was taken into account in the model as an adjusting factor.
Decontamination from host and trimming was done following the same routines as for the taxonomic analysis. Cleaned sequences for all samples were assembled into contigs using megahit 1.2.4 (ref. 84), and prodigal 2.6.3 (ref. 85) was then used to predict the open reading frames inside the obtained contigs. Protein mapping and KEGG and COG annotations were obtained using the EggNOG mapper 2.0.0 (ref. 86). The abundance of the annotated genes was finally measured by counting aligned reads to them via Bowtie2, version 2.2.2, under default parameters87. Resulting counts data were aggregated at protein level. The treatment effect (finish versus start) was compared between OSKM and control (WT) mice. This was accounted in a model with an interaction term (drug:treatment) using DESeq2 with default options83. The paired nature of the experimental design was considered in the model as an adjusting factor. The top 500 protein hits from the fitted model (nondirectional set) as well as the top 200 positive hits and the top 200 negative hits (directional sets), in all cases ordered by statistical significance, were used to explore enrichment of functional annotations. In this regard, GO terms for bacteria and archaea were considered using the AmiGO 2 GO annotations database88, removing from the analysis gene sets with few genes (less than 8) and too many genes (more than 499). Statistically enriched GO terms were identified using the standard hypergeometric test. Significance was defined by the adjusted P value using the Benjamini and Hochberg multiple-testing correction. To take into consideration the compositional nature of the data, all DESeq2-based results were complemented with graphical representations of abundance log-ratio (between finish and start matched samples) rankings. This provides a scale invariant way (with regard to the total microbial load) to present the data89.
Blood was collected via submandibular vein bleed (D0, D2, D4) or intracardiac puncture following deep carbon dioxide anaesthetisation (D7) at approximately 12:00–14:00 h (4–6 h into the light cycle) of each day. Whole blood was spun down for 10 min at 3,381g at 4 °C and supernatant (serum) was separated and stored at −80 °C.
In vivo time-course metabolomics analysis
Standard and reagents
Acetonitrile (Sigma-Aldrich), isopropanol (Sigma-Aldrich), methanol (Sigma-Aldrich), chloroform (Sigma-Aldrich), acetic acid (Sigma-Aldrich), formic acid (Sigma-Aldrich), methoxyamine hydrochloride (Sigma-Aldrich), MSTFA (N-methyl-N-(trimethylsilyl) trifluoroacetamide; Sigma-Aldrich), pyridine (Sigma-Aldrich), 3-nitrophenylhydrazine (Sigma-Aldrich), N-(3-dimethylaminopropyl)-N′-ethylcarbodiimide hydrochloride (EDC; Sigma-Aldrich) and sulfosalicylic acid (Sigma-Aldrich) as previously described90.
Sample preparation serum (lithium heparin)
A volume of 25 µl of serum were mixed with 250 µl a cold solvent mixture with ISTD (methanol/water/chloroform, 9:1:1, −20 °C), into 1.5 ml microtube, vortexed and centrifugated (10 min at 15,000g, 4 °C). The upper phase of supernatant was split into three parts: 50 µl was used for gas chromatography coupled to mass spectrometry (GC–MS) experiments in the injection vial, 30 µl was used for the short-chain fatty acid ultra-high performance liquid chromatography (UHPLC)–MS method, and 50 µl was used for other UHPLC–MS experiments.
Widely targeted analysis of intracellular metabolites GC coupled to a triple-quadrupole mass spectrometer
The GC–MS/MS method was performed on a 7890B gas chromatography system (Agilent Technologies) coupled to a triple-quadrupole 7000C (Agilent Technologies) equipped with a high-sensitivity electronic impact source (EI) operating in positive mode.
Targeted analysis of bile acids by ion pairing UHPLC coupled to a triple-quadrupole mass spectrometer
Targeted analysis was performed on an RRLC 1260 system (Agilent Technologies) coupled to a triple-quadrupole 6410 (Agilent Technologies) equipped with an electrospray source operating in positive mode. Gas temperature was set to 325 °C with a gas flow of 12 l min−1. Capillary voltage was set to 4.5 kV.
Targeted analysis of polyamines by ion pairing UHPLC coupled to a triple-quadrupole mass spectrometer
Targeted analysis was performed on an RRLC 1260 system (Agilent Technologies) coupled to a triple-quadrupole 6410 (Agilent Technologies) equipped with an electrospray source operating in positive mode. The gas temperature was set to 350 °C with a gas flow of 12 l min−1. The capillary voltage was set to 3.5 kV.
Targeted analysis of short-chain fatty acid by ion pairing UHPLC coupled to a 6500 + QTRAP mass spectrometer
Targeted analysis was performed on an RRLC 1260 system (Agilent Technologies) coupled to a 6500 + QTRAP (Sciex) equipped with an electrospray ion source.
Untargeted analysis of intracellular metabolites by UHPLC coupled to a Q-Exactive mass spectrometer (reversed-phase acetonitrile method)
The profiling experiment was performed with a Dionex Ultimate 3000 UHPLC system (Thermo Scientific) coupled to a Q-Exactive (Thermo Scientific) equipped with an electrospray source operating in both positive and negative mode and full scan mode from 100 to 1,200 m/z. The Q-Exactive parameters were: sheath gas flow rate, 55 arbitrary units (a.u.); auxiliary gas flow rate, 15 a.u.; spray voltage, 3.3 kV; capillary temperature, 300 °C; S-Lens RF level, 55 V. The mass spectrometer was calibrated with sodium acetate solution dedicated to low mass calibration.
The peak areas (corrected to quality control) corresponding to each annotated metabolite identified in the serum of reprogrammable mice (n = 6 per group) at day 5 and day 7 after doxycycline treatment were converted to log2 values. Data were represented as log2 fold change (log2 FC) values to each mouse at day 0 (before doxycycline administration). Metabolic pathway impact was calculated by Global ANOVA pathway enrichment and Out-degree Centrality Topology analysis through the MetaboAnalyst 4.0 software91, using KEGG library (2019) as a reference. The colour gradient from white to red indicates the P value, where red is most significant. Bubble size indicates the relative contribution of the detected metabolites in their respective KEGG pathway. Pathway impact scores the centrality of the detected metabolites in the pathway.
Doxycycline serum analysis
A total of 30 µl of mouse plasma was acidified with 3 µl solution of 15% phosphoric acid (vol/vol). Afterwards, 42 µl of methyl tert-butyl ether was added and vigorously mixed using a vortex. After 20 min of reequilibration, samples were centrifuged for 10 min at 21,130g at 4 °C. Next, 90 µl of acetonitrile were added to 10 µl of the aqueous phase to facilitate protein precipitation. After another cycle of centrifugation, the supernatant was transferred into a vial before LC–MS analysis.
The extracts were analysed by a UHPLC system coupled to a 6490 triple-quadrupole mass spectrometer (QqQ, Agilent Technologies) with electrospray ion source (LC–ESI–QqQ) working in positive mode. The injection volume was 3 µl. An ACQUITY UPLC BEH HILIC column (1.7 µm, 2.1 × 150 mm, Waters) and a gradient mobile phase consisting of water with 50 mM ammonium acetate (phase A) and acetonitrile (phase B) were used for chromatographic separation. The gradient was as follows: isocratic for 2 min at 98% B, from 2 to 9 min decreased to 50% B, for 30 s raised to 98%, and finally column equilibrated at 98% B until 13 min. The flow rate was 0.4 ml min−1. The mass spectrometer parameters were as follows: drying and sheath gas temperatures, 270 °C and 400 °C, respectively; source and sheath gas flow rates, 15 and 11 l min−1, respectively; nebulizer flow, 35 psi; capillary voltage, 3,000 V; nozzle voltage, 1,000 V; and iFunnel HRF and LRF, 130 and 100 V, respectively. The QqQ worked in MRM mode using defined transitions. The transitions for doxycycline and the collision energy (CE(V)) were 445 → 428(17), 445 → 98(60).
Determination of methionine, SAM, SAH and homocysteine in serum
In total, 25 μl of serum was mixed with 25 μl of TCEP and 70 μl of 1% formic acid in methanol. Samples were vortexed and left at −20 °C for 1 h, centrifuged for 10 min at 21,130g and 4 °C and transferred to glass vials for their analysis by LC–MS.
LC–MS was performed with a Thermo Scientific Vanquish Horizon UHPLC system interfaced with a Thermo Scientific Orbitrap ID-X Tribrid Mass Spectrometer.
Metabolites were separated by HILIC chromatography with an InfinityLab Poroshell 120 HILIC-Z 2.7 μm, 2.1 mm × 100 mm column (Agilent Technologies). The mobile phase A was 50 mM ammonium acetate in water, and mobile phase B was acetonitrile. Separation was conducted under the following gradient: 0–2 min, isocratic 90% B; 2–6 min raised to 50% B; 6–7 min, isocratic 50% B; 7–7.2 min, increased to 90% B; 7.2–10.5 min, reequilibration column 90% B. The flow rate was 0.4 ml min−1. The injection volume was 5 μl.
Samples were analysed in positive mode in targeted SIM mode and the following setting: isolation window (m/z), 4; spray voltage, 3,500 V; sheath gas, 50 a.u.; auxiliary gas, 10 a.u.; ion transfer tube temperature, 300 °C; vaporizer temperature, 300 °C; Orbitrap resolution, 120,000; RF lens, 60%; AGC target, 2e5; maximum injection time, 200 ms.
SAM (m/z 399.145) was monitored from 5–7 min; Met (m/z 150.0583) from 3.2–5.2 min; SAH (m/z 385.1289) from 4–6 min; Hcy (m/z 136.0428) from 3.4–5.5 min, as previously optimized using pure standards.
Determination of cyanocobalamin in stool
Approximately, 20 mg of dry and pulverized stool samples were mixed with with 75 μl of TCEP and 210 μl of 1% formic acid in methanol. Samples were vortexed and subjected to three freeze–thaw cycles using liquid nitrogen. Subsequently, samples were left in ice for 1 h, centrifuged for 10 min at 21,130g and 4 °C and transferred to glass vials for their analysis by LC–MS.
LC–MS was performed with a Thermo Scientific Vanquish Horizon UHPLC system interfaced with a Thermo Scientific Orbitrap ID-X Tribrid Mass Spectrometer.
Metabolites were separated by HILIC chromatography with an InfinityLab Poroshell 120 HILIC-Z 2.7 μm, 2.1 mm × 100 mm column (Agilent Technologies). The mobile phase A was 50 mM ammonium acetate in water, and mobile phase B was acetonitrile. Separation was conducted under the following gradient: 0–2 min, isocratic 90% B; 2–6 min raised to 50% B; 6–7 min, isocratic 50% B; 7–7.2 min, increased to 90% B; 7.2–10.5 min, reequilibration column 90% B. The flow was 0.4 ml min−1. The injection volume was 5 μl.
Samples were analysed in positive mode in targeted SIM mode and the following setting: isolation window (m/z), 4; spray voltage, 3,500 V; sheath gas, 50 a.u.; auxiliary gas, 10 a.u.; ion transfer tube temperature, 300 °C; vaporizer temperature, 300 °C; Orbitrap resolution, 120,000; RF lens, 60%; AGC target, 2e5; maximum injection time, 200 ms. Cyanocobalamin was monitored from (m/z 1355.5747 and m/z 678.291) from 5–5.5 min, as previously optimized using a pure standard.
Vitamin B12 serum analysis
Mouse serum was diluted at a 1:20 ratio in PBS and holotranscobalamin (holoTC) was measured using an ADVIA Centuar Immunoassay System (SIEMENS) with ADVIA Centuar Vitamin B12 Test Packs (07847260) according to the manufacturer’s instructions.
In vitro SIL experiment
Cell pellets were mixed with 50 μl of TCEP and 140 μl of 1% formic acid in methanol (containing 150 μg l−1 of Tryptophan-d5 as internal standard). Samples were vortexed and subjected to three freeze–thaw cycles using liquid nitrogen. Subsequently, samples were left at −20 °C for 1 h, centrifuged for 10 min at 21,130g and 4 °C and transferred to glass vials for their analysis by LC–MS/MS.
Samples were analysed with an UHPLC 1290 Infinity II Series coupled to a QqQ/MS 6490 Series from Agilent Technologies (Agilent Technologies). The source parameters applied operating in positive electrospray ionization (ESI) were gas temperature: 270 °C; gas flow: 15 l min−1; nebulizer: 35 psi; sheath gas heater, 400 a.u.; sheath gas flow, 11 a.u.; capillary, 3,000 V; nozzle voltage: 1,000 V.
The chromatographic separation was performed with an InfinityLab Poroshell 120 HILIC-Z 2.7 μm, 2.1 mm × 100 mm column (Agilent Technologies), starting with 90% B for 2 min, 50% B from minute 2 to 6, and 90% B from minute 7 to 7.2. Mobile phase A was 50 mM ammonium acetate in water, and mobile phase B was acetonitrile. The column temperature was set at 25 °C and the injection volume was 2 μl.
MRM transitions for SAM (RT: 6.1 min) were 399→298 (4 V), 399→250 (12 V), 399→97 (32 V) and 399→136 (24 V) for M + 0, and 400→299 (4 V), 400→251 (12 V), 400→97 (32 V), 400→137 (24 V), 400→250 (12 V) and 400→136 (24 V) for M + 1.
Samples were fixed overnight at 4 °C with neutral buffered formalin (HT501128-4L, Sigma-Aldrich). Paraffin-embedded tissue sections (2–3 μm in thickness) were air-dried and further dried at 60 °C overnight for immunohistochemical staining.
Sections were stained with haematoxylin and eosin (H&E) for histological evaluation by a board-certified pathologist who was blinded to the experimental groups. Additionally, periodic acid–Schiff staining (AR16592-2, Artisan, Dako, Agilent) was used to visualize mucus-producing cells on 3–4-µm sections of colon that were counterstained with haematoxylin.
In the reprogramming model, the findings were evaluated by focusing mainly on the appearance of hyperplastic and dysplastic changes of the epithelial cells of the digestive mucosa and pancreatic acini. Inflammation and loss of the intestinal goblet cells were also reported. To document the severity and extension, a semi-quantitative grading system was used based on previously used histological criteria:
Gastric and colon mucosa inflammatory cell infiltrate and multifocal areas of crypt (large intestine) or glandular (stomach) epithelial cell dysplasia were scored from 0 to 5, where 0 indicates absence of lesion and 5 indicates very intense lesions.
Intestinal crypt hyperplasia: 1, slight; 2, twofold to threefold increase of the crypt length; 3, >threefold increase of the crypt length.
Goblet cell loss of the mucosa of the large intestine: 1, <10% loss; 2, 10–50% loss; 3, >50% loss.
Histological total score was presented as a sum of all parameters scored for a given tissue.
In the colitis model, the following parameters were semi-quantitatively evaluated as previously described92 as follows:
Inflammation of the colon mucosa: 0, none; 1, slight, 2, moderate; 3, severe.
Depth of the injury: 0, none; 1, mucosa; 2, mucosa and submucosa; 3, transmural.
Crypt damage: 0, none; 1, basal and 1/3 damaged; 2, basal and 2/3 damaged; 3, only the surface epithelium intact; 4, entire crypt and epithelium lost.
Tissue involvement: 0, none; 1, 0–25%; 2, 26–50%; 3, 51–75%; 4, 76–100%.
The score of each parameter was multiplied by the factor of tissue involvement and summed to obtain the total histological score.
Immunohistochemistry was performed using a Ventana discovery XT for NANOG and Sca1/Ly6A/E, the Leica BOND RX Research Advanced Staining System for H3K36me3, keratin 14 and vitamin B12, and manually for Ki67. Antigen retrieval for NANOG was performed with Cell Conditioning 1 buffer (950-124, Roche) and for Sca1/Ly6A/E with Protease 1 (5266688001, Roche) for 8 min followed with the OmniMap anti-Rat HRP (760-4457, Roche) or OmniMap anti-Rb HRP (760-4311, Roche). Blocking was done with casein (760-219, Roche). Antigen–antibody complexes were revealed with ChromoMap DAB Kit (760-159, Roche). For H3K36me3 and keratin 14, antigen retrieval was performed with BOND Epitope Retrieval 1 (AR9961, Leica) and for vit B12 with BOND Epitope Retrieval Solution 2 (Leica Biosystems, AR9640) for 20 min, whereas for Ki67, sections were dewaxed as part of the antigen retrieval process using the low pH EnVision FLEX Target Retrieval Solutions (Dako) for 20 min at 97 °C using a PT Link (Dako-Agilent). Blocking was performed with Peroxidase-Blocking Solution at room temperature (RT; S2023, Dako-Agilent) and 5% goat normal serum (16210064, Life technology) mixed with 2.5% BSA diluted in wash buffer for 10 and 60 min at RT. Vitamin B12 also was blocked with Vector M.O.M. Blocking Reagent (MK-2213, Vector) following the manufacturer’s procedures for 60 min. Primary antibodies were incubated for 30, 60 or 120 min. The secondary antibody used was the BrightVision poly HRP-Anti-Rabbit IgG, incubated for 45 min (DPVR-110HRP, ImmunoLogic) or the polyclonal goat Anti-Mouse at a dilution of 1:100 for 30 min (Dako-Agilent, P0447). Antigen–antibody complexes were revealed with 3-3′-diaminobenzidine (K346811, Agilent or RE7230-CE, Leica). Sections were counterstained with haematoxylin (CS700, Dako-Agilent or RE7107-CE, Leica) and mounted with Mounting Medium, Toluene-Free (CS705, Dako-Agilent) using a Dako CoverStainer. Specificity of staining was confirmed by staining with a rat IgG (6-001-F, R&D Systems, Bio-Techne), a Rabbit IgG (ab27478, Abcam) or a mouse IgG1, kappa (Abcam, ab18443) isotype controls. See Supplementary Table 5 for primary antibody details.
In situ hybridization—RNAscope
Ready-to-use reagents from RNAscope 2.5 LS Reagent Kit-RED (322150, RNAScope, ACD Bio-Techne) were loaded onto the Leica Biosystems BOND RX Research Advanced Staining System according to the user manual (322100-USM). FFPE tissue sections were baked and deparaffinized on the instrument, followed by epitope retrieval (using Leica Epitope Retrieval Buffer 2 at 95 °C for 15 min) and protease treatment (15 min at 40 °C). Probe hybridization, signal amplification, colorimetric detection and counterstaining were subsequently performed following the manufacturer’s recommendations.
Hybridization was performed with the RNAscope LS 2.5 Probe – Mm-Lgr5 – Mus musculus leucine rich repeat containing G-protein-coupled receptor 5 (312178, RNAScope, ACD Bio-Techne). Control probe used was the RNAscope 2.5 LS Probe – Mm-UBC – Mus musculus ubiquitin C (Ubc), as a housekeeping gene (310778, RNAScope – ACD Bio-Techne). The bacterial probe RNAscope 2.5 LS Negative Control Probe_dapB was used as a negative control (312038, RNAScope – ACD Bio-Techne).
Brightfield images were acquired with a NanoZoomer-2.0 HT C9600 digital scanner (Hamamatsu) equipped with a ×20 objective. All images were visualized with a gamma correction set at 1.8 in the image control panel of the NDP.view 2 U12388-01 software (Hamamatsu, Photonics).
Brightfield images of immunohistochemistry were quantified using QuPath software93 with standard detection methods. Where the percentage of tissue staining is calculated, pixels were classified as positive and negative using the Thresholder function. Where the percentage of cells is quantified, the Positive Cell Detection function was used.
Cellular and molecular methods
MEFs were cultured in standard DMEM medium with 10% FBS (Gibco, LifeTechnologies, 10270106) with antibiotics (100 U ml−1 penicillin–streptomycin; Life Technologies, 11528876). Reprogramming of the doxycycline-inducible 4-Factor (i4F) MEFs with inducible expression of the four Yamanaka factors Oct4, Sox2, Klf4 and cMyc (OSKM) was performed as previously described3. Briefly, i4F MEFs were seeded at a density of 3 × 105 cells per well in six-well tissue culture plates coated with gelatin and treated with doxycycline (PanReac, A2951) 1 mg ml−1 continuously to induce expression of the OSKM transcription factors in the presence of ‘complete KSR media’ (15% (vol/vol) Knockout Serum Replacement (KSR, Invitrogen, 10828028) in DMEM with GlutaMax (Life Technologies, 31966047) basal media, with 1,000 U ml−1 LIF (Merck, 31966047), non-essential amino acids (Life Technologies, 11140035) and 100 μM beta-mercaptoethanol (Life Technologies, 31350010) plus antibiotics (penicillin–streptomycin, Gibco, 11528876)), which was replaced every 48–72 h. After 10 d, iPS cell colonies were scored by alkaline phosphatase staining according to the manufacturer’s protocol (AP blue membrane substrate detection kit, Sigma, AB0300). Vitamin B12 (Sigma, V2876; 2 μM final), MAT2Ai PF-9366 (MedChemExpress, HY-107778; 2 µM final), SAM (S-(5′-adenosyl)-l-methionine iodide, Merck, A4377; 100 µM final) and NSC636819 (Sigma-Aldrich, 5.31996; 10 µM final) were added continuously to the culture media and replaced every 48–72 h.
Reprogramming of WT MEFs was performed as previously described94. Briefly, HEK-293T (American Type Culture Collection, ATCC-CRL-3216) cells were cultured in DMEM supplemented with 10% FBS and antibiotics (penicillin–streptomycin, Gibco, 11528876). Around 5 × 106 cells per 100-mm-diameter dish were transfected with the ecotropic packaging plasmid pCL-Eco (4 μg) together with one of the following retroviral constructs (4 μg): pMXs-Klf4, pMXs-Sox2, pMXs-Oct4 or pMXs-cMyc (obtained from Addgene) using Fugene-6 transfection reagent (Roche) according to the manufacturer’s protocol. The following day, media were changed and recipient WT MEFs to be reprogrammed were seeded (1.5 × 105 cells per well of a six-well plate). Retroviral supernatants (10 ml per plate/factor) were collected serially during the subsequent 48 h, at 12-h intervals, each time adding fresh media to the 293T cells cells (10 ml). After each collection, supernatant was filtered through a 0.45-µm filter, and each well of MEFs received 0.5 ml of each of the corresponding retroviral supernatants (amounting to 2 ml total). Vitamin B12 supplementation (Sigma, V2876; 2 µM final concertation) began on the same day as viral transduction. This procedure was repeated every 12 h for 2 d (a total of four additions). After infection was completed, media were replaced by ‘complete KSR media’ (see above). Cell pellets were harvested on day 5 (relative to the first infection) and histone extracts were processed for immunoblot as described below. On day 14 (relative to the first infection), iPS cell colonies were scored by alkaline phosphatase staining according to the manufacturer’s protocol (AP blue membrane substrate detection kit; Sigma, AB0300).
Doxycycline-inducible i4F MEFs were cultured as described in ‘Cell culture’ above, with 1 mg ml−1 doxycycline, with without continuous vitamin B12 supplementation. At 72 h after the addition of doxycycline, cells were transferred to complete KSR media containing a final concentration of 0.5 mM l-Serine-13C3 (Sigma-Aldrich, 604887). This is the same concentration of unlabelled l-serine normally found in the complete KSR media, and was generated by ordering custom, serine-free DMEM (Life Technologies, ME22803L1) and custom, serine-free non-essential amino acid mixture (Life Technologies, ME22804L1). Six hours after the addition of labelled media, a subset of wells was harvested by scraping in PBS and centrifugation (300g for 5 min); supernatant was removed and pellets were snap frozen. At 72 h after the addition of the labelled media (that is, 6 days into reprogramming), cells still in culture were transferred back to unlabelled complete KSR media, which was changed every 48–72 h. iPS cell colonies were analysed by alkaline phosphatase staining according to the manufacturer’s protocol (AP blue membrane substrate detection kit; Sigma, AB0300) on day 10. Doxycycline and vitamin B12 supplementation were continuous throughout the entire reprogramming protocol, and replenished with every media change (that is, every 48–72 h).
i4F MEFs were cultured in the presence doxycycline ±2 µM of vitamin B12 over 3 or 10 days (culture conditions as described above) and histone extracts were prepared using EpiQuik Total Histone Extraction Kit (EpiGentek, OP-0006-100) according to the manufacturer’s instructions. Around 200 ng of total histone extract was used per well in the EpiQuik Histone H3 Modification Multiplex Assay Kit (Colorimetric; EpiGentek, P-3100) according to the manufacturer’s instructions.
Cell lysis and immunoblot
Histone extracts were prepared using an EpiQuik Total Histone Extraction Kit (EpiGentek, OP-0006-100) according to the manufacturer’s instructions and quantified using DC Protein Assay Kit (Bio-Rad, 5000111). Whole-cell extracts were prepared in RIPA buffer (10 mM Tris-HCl, pH 8.0; 1 mM EDTA; 0.5 mM EGTA; 1% Triton X-100; 0.1% sodium deoxycholate; 0.1% SDS; 140 mM NaCl). A total of 10 μg of lysate was loaded per lane and hybridized using antibodies against H3K36me3, MS, vinculin, total histone H3 and LI-COR fluorescent secondary reagents (IRDye 800 CW anti-mouse, 926-32210; IRDye 680 CW anti-mouse, 926-68070; IRDye 800 CW anti-rabbit, 926-32211; IRDye 680 CW anti-mouse, 926-68071) all at a dilution of 1:10,000 according to manufacturer’s instructions. Immunoblots were visualized on an Odyssey FC Imaging System (LI-COR Biosciences). See Supplementary Table 5 for primary antibody details.
GSEAPreranked was used to perform a GSEA of annotations from MsigDB M13537, with standard GSEA and leading edge analysis settings. We used the RNA-seq gene list ranked by log2 fold change, selecting ‘gene set’ as the permutation method with 1,000 permutations for Kolmogorov–Smirnoff correction for multiple testing95.
Selection of genes to measure by qPCR from methionine deprivation signature
Genes belonging to the leading edge of the GSEA using the Met derivation signature (MsigDB, M13537) in the pancreas of reprogramming mice were selected. These genes were then compared to genes belonging to the leading edge of the same gene signature from i4F MEFs treated with doxycycline in vitro for 72 h, as compared to OSKM MEFs treated with vitamin B12 (that is, genes in MsigDB M13537 whose upregulation was relieved by B12 supplementation in vitro). We selected 11 of these genes for which we had qPCR primers available.
Analysis of mRNA levels by qPCR
Total RNA was extracted from MEFs with TRIzol (Invitrogen) according to the manufacturer’s instructions. Up to 5 µg of total RNA was reverse transcribed into cDNA using the iScript Advanced cDNA Synthesis Kit (Bio-Rad, 172-5038; pancreas) or iScript cDNA Synthesis Kit (Bio-Rad, 1708890; all other organs) for RT–qPCR. Real-time qPCR was performed using GoTaq qPCR Master Mix (Promega, A6002) in a QuantStudio 6 Flex thermocycler (Applied Biosystem) or 7900HT Fast Real-Time PCR System (Thermo Fisher). See Supplementary Table 6 for primer sequences.
ChIP sample preparation
i4F MEFs were cultured in the presence or absence of doxycycline ±2 µM of vitamin B12 (Merck, V2876) over 3 days in six-well plates (culture conditions as described above). Cells were fixed with 1% (vol/vol) PFA (Fisher Scientific, 50980487) for 2 min and then quenched with 750 mM Tris (PanReac AppliChem, A2264) for 5 min. Cells were washed twice with PBS, scraped, and spun down at 1,200g for 5 min. Pellets were lysed with 100 µl (per well) lysis buffer (50 mM HEPES-KOH pH 7.5, 140 mM HCl, 1 mM EDTA pH 8, 1% Triton X-100, 0.1% sodium deoxycholate, 0.1% SDS, protease inhibitor cocktail; Sigma, 4693159001) on ice for 10 min, then sonicated using a Diagenode BioRuptor Pico (Diagenode, B01060010) for ten cycles (30 s on, 30 s off) at 4 °C. Lysates were clarified for 10 min at 8,000g, 1% input samples were reserved, and supernatant was used for immunoprecipitation with Diagenode Protein A-coated Magnetic beads ChIP–seq grade (Diagenode, C03010020-660) and H3K3me3 monoclonal antibody (Cell Signaling Technologies, 4909) with 0.1% BSA (Sigma, 10735094001). The following day, cells were washed once with each buffer: low salt (0.1% SDS, 1% Triton X-100, 2 mM EDTA, 20 mM Tris-HCl pH 8.0, 150 mM NaCl), high salt (0.1% SDS, 1% Triton X-100, 2 mM EDTA, 20 mM Tris-HCl pH 8.0, 5,000 mM NaCl), LiCl (0.25 M LiCl, 1% NP-40, 1% sodium deoxycholate, 1 mM EDTA, 10 mM Tris-HCl pH 8.0) and eluted in 1% SDS, 100 mM NaHCO3 buffer. Cross-links were reversed with RNase A (Thermo Fisher, EN0531), proteinase K (Merck, 3115879001) and sodium chloride (Sigma, 71376), and chromatin fragments were purified using QIAquick PCR purification kit (Qiagen, 28104).
RNA-seq RNA extraction
i4F MEFs were cultured in the presence or absence of doxycycline and the indicated compounds over 3 days in six-well plates (culture conditions as described above). After 72 h, RNA was extracted using an RNeasy Kit (Qiagen, QIA74106) according to the manufacturer’s instructions.
The concentration of the DNA samples (inputs and immunoprecipitations) was quantified with a Qubit dsDNA HS kit, and fragment size distribution was assessed with the Bioanalyzer 2100 DNA HS assay (Agilent). Libraries for ChIP–seq were prepared at the IRB Barcelona Functional Genomics Core Facility. Briefly, single-indexed DNA libraries were generated from 0.5–1.5 ng of DNA samples using the NEBNext Ultra II DNA Library Prep kit for Illumina (New England Biolabs). Eleven cycles of PCR amplification were applied to all libraries.
The final libraries were quantified using the Qubit dsDNA HS assay (Invitrogen) and quality controlled with the Bioanalyzer 2100 DNA HS assay (Agilent). An equimolar pool was prepared with the 24 libraries and sequenced on a NextSeq 550 (Illumina). 78.9 Gb of SE75 reads were produced from two high-output runs. A minimum of 23.97 million reads were obtained for all samples.
For MEFs 1–3
The concentration of total RNA extractions was quantified with the Nanodrop One (Thermo Fisher), and RNA integrity was assessed with the Bioanalyzer 2100 RNA Nano assay (Agilent). Libraries for RNA-seq were prepared at the IRB Barcelona Functional Genomics Core Facility. Briefly, mRNA was isolated from 1.5 μg of total RNA using the kit NEBNext Poly(A) mRNA Magnetic Isolation Module (New England Biolabs). The isolated mRNA was used to generate dual-indexed cDNA libraries using the NEBNext Ultra II Directional RNA Library Prep Kit for Illumina (New England Biolabs). Ten cycles of PCR amplification were applied to all libraries.
The final libraries were quantified using the Qubit dsDNA HS assay (Invitrogen) and quality controlled with the Bioanalyzer 2100 DNA HS assay (Agilent). An equimolar pool was prepared with the 12 libraries and submitted for sequencing at the Centre Nacional d’Anàlisi Genòmica (CRG-CNAG). A final quality control by qPCR was performed by the sequencing provider before paired-end 50-nucleotide sequencing on a NovaSeq 6000 S2 (Illumina). Around 77.7 Gb of PE50 reads were produced from three NovaSeq 6000 flow cells. A minimum of 55.7 million reads were obtained for all samples (Extended Data Fig. 7).
For MEFs 4–5
Total RNA extractions were quantified with a Nanodrop One (Thermo Fisher), and RNA integrity was assessed with the Bioanalyzer 2100 RNA Nano assay (Agilent). Libraries for RNA-seq were prepared at the IRB Barcelona Functional Genomics Core Facility. Briefly, mRNA was isolated from 1.2 μg of total RNA and used to generate dual-indexed cDNA libraries with the Illumina Stranded mRNA ligation kit (Illumina) and UD Indexes Set A (Illumina). Ten cycles of PCR amplification were applied to all libraries.
Sequencing-ready libraries were quantified using the Qubit dsDNA HS assay (Invitrogen) and quality controlled with the Tapestation HS D5000 assay (Agilent). An equimolar pool was prepared with the 15 libraries for SE75 sequencing on a NextSeq 550 (Illumina). Sequencing output was above 539 million 75-nucleotide single-end reads and a minimum of 28 million reads was obtained for all samples (Extended Data Fig. 7).
RNA-seq data processing
All analyses were performed in the R programming language (version 4.0.5)96 unless otherwise stated. Stranded paired-end reads were aligned to the Mus musculus reference genome version mm10 using STAR80 with default parameters. STAR indexes were built using the ENSEMBL annotation version GRC138.97. SAM files were converted to BAM and sorted using sambamba (version 0.6.7)97. Gene counts were obtained with the featureCounts function from the Rsubread package98 with the gtf file corresponding to ENSEMBL version GRC138.97 and parameters set to: isPairedEnd = TRUE and strandSpecific = 2. Technical replicates were collapsed by adding the corresponding columns in the count matrix.
We obtained a reprogramming gene signature from published data48 and selected genes with false discovery rate (FDR) lower than 0.05 and fold change between MEF and d3-EFF larger than 2. The reprogramming score was defined as the average of all genes in the signature after scaling the rlog transformed matrix.
Computation of cryptic transcript ratios between first and intermediate exons
Exon counts were generated using the featureCounts function with parameters: isPairedEnd = TRUE, strandSpecific = 2, GTF.featureType = exon, GTF.attrType = transcript_id, GTF.attrType.extra = gene_id, allowMultiOverlap = TRUE and useMetaFeatures = FALSE and the same GTF as for gene counts. Technical replicates were collapsed by adding the corresponding counts. For each gene, the longest annotated transcript was selected. Genes with less than four exons of RPKMs lower than exp(−2) were discarded from the analysis. Intermediate exons were defined as those from the fourth to the penultimate. A total of 9,365 genes were used to compute the ratio between the intermediate and first exons. Fold changes between untreated and B12-treated samples were computed as the ratio between the exon ratios.
Comparison of cryptic transcript ratios between conditions
Genes were separated by their expression after transcript length and library size normalization (RPKM). For each sample, we computed the median ratios for genes in each decile.
Analysis of CT in DSS time course
Data were accessed from GSE131032. Reads were processed and ratios computed as previously described. log2 ratios for all transcripts were summarized through the median by sample. Comparisons between days were performed fitting a linear model to the medians using ‘cage’ as a covariable. The function glht from the multcomp R package was used to find coefficients and P values.
Functional enrichment in genes with exon ratios affected by vitamin B12 treatment
To select genes most affected by the B12 treatment after reprogramming, we compared ratios between the doxy and MEF conditions and between the doxy and doxy + B12 conditions. Genes that increased the ratios in the first comparison (upper 25th percentile) and decreased the ratio in the second comparison (bottom 25%) were selected for functional enrichment analysis. A hypergeometric test was performed to find significant overlap between the defined gene set and the Biological Processes GO collection99.
ChIP–seq data processing
Reads were aligned to the mm10 reference genome with bowtie100 version 0.12.9 with parameters –n 2 and –m 1 to keep reads with multiple alignments in one position. SAM files were converted to BAM and sorted using sambamba version 0.6.7.
Heat maps of average coverage in gene bodies
For each sample, aligned reads were imported into R using the function scanBam from the Rsamtools package101. Whole-genome coverage was computed using the coverage function from the IRanges package102 and binned into 50-bp windows. Gene annotations were imported from Ensembl version GRCm38. The average coverage over gene bodies was computed using the normalizeToMatrix function from the EnrichedHeatmap package103 with parameters extend = 1,000, mean_mode = w0 and w = 50. Genes were filtered to coincide with those used in the exon ratio calculation from the RNA-seq data. Rows in the heat map were split by the average RNA-seq RPKM values in all samples.
Visualization of ChIP tracks
BAM files were transformed to TDF files using the count function from IGVtools (version 2.12.2)104 with parameters –z 7, –w 25 and –e 250. Visualization of TDF files was generated using IGV (version 2.9.4)105.
Analysis of human RNA-seq data
Data were accessed from GSE109142. Reads were processed and ratios computed as previously described except using the ENSEMBL GRCm38.101 human gene annotation and the hg38 genome assembly version. The log2 ratios for all transcripts were summarized through the median by sample. Comparison between diagnosis status was performed fitting a linear model to the medians with sex and the expression quantiles as covariables. The model was fitted using the lm R function and coefficients and P values with the coeff function.
Statistics and data availability
Statistical analysis and figure preparation
Unless otherwise specified, data are presented as the mean ± s.d. Statistical analysis was performed by Student’s t-test or one-way analysis of variance (ANOVA) as indicated, using GraphPad Prism v9.0.0, and specific statistical tests as indicated for each experiment for bioinformatic analyses. P values of less than 0.05 were considered as statistically significant. No statistical methods were used to predetermine sample size in the mouse studies, but our sample sizes are similar to those reported in previous publications3,9,16,17,19. Animals and data points were not excluded from analysis with the exception of the MEFs that failed to reprogram in the ChIP experiment, which is clearly detailed in the text. Mice were allocated at random to treatment groups, with attempts to balance initial body weight and sex as possible. The investigators were blinded during histological assessment of the mice; other data collection and analysis was not performed blind to the conditions of the experiments. Data distribution was assumed to be normal, but this was not formally tested. Figures were prepared using Illustrator CC 2019 (Adobe).
Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.