Dietary fibre-adapted gut microbiome clears dietary fructose and reverses hepatic steatosis

Mouse studies

Animal studies followed protocols approved by the Institutional Animal Care and Use Committee of the University of California, Irvine (AUP-22-121). Male C57BL/6 mice (8 weeks old) were purchased from Jackson Laboratory. The duration of each experiment (for example, diet feeding) is indicated in the figure legends. In this study, only males were used because MASLD is more prevalent in men than women. Generation of Khk-C transgenic mice was previously reported in ref. ¹⁶. Mice were group-housed on a normal light–dark cycle (07:00–19:00 h) with free access to chow and water. We modified the open-source diet (Research Diets, D11112201) to reduce dextrose in the diet and replace corn starch with inulin. Mice were fed either a control (Research Diets, D21050401) or 10% (gm%) inulin diet (Research Diets, D21050402). The composition of both diets is shown in Supplementary Table 1. For HFCS provision, mice were provided either normal drinking water or 15% (weight/weight) fructose and 15% glucose mixture in drinking water. Animals were randomly assigned to experimental groups, and no specific method of randomization was used. No statistical methods were used to pre-determine sample sizes, but our sample sizes are similar to those reported in previous publications^15,16. Unless otherwise indicated, experiments were replicated independently at least twice.

Daily water and food intake were determined by measuring the total consumption in a cage divided by the number of mice in the cage. For antibiotic treatment, a cocktail of antibiotics, including 0.5 g l⁻¹ ampicillin, neomycin, metronidazole and 0.25 g l⁻¹ vancomycin, was dissolved in HFCS-water. For gut bacteria transplantation, faecal or small intestinal contents (200 mg) were freshly collected from donor mice at 09:00–10:00 h using sterilized utensils and microcentrifuge tubes. The samples were immediately dissolved in 2 ml of sterilized anaerobic PBS containing 0.1 g l⁻¹ of l-cysteine. The materials were then homogenized using sterilized pellet pestles and centrifuged at 500g for 3 min at 20 °C to remove particulate matter, based on studies by others that performed gut bacteria transplantation^89,90,91. The resulting bacterial supernatant was supplemented with 10% sterilized glycerol and then transferred to antibiotic-treated recipient mice by oral gavage (200 µl per mouse) with a plastic feeding tube (Instech Laboratories) on the same day. The remaining solution was stored at −80 °C until use for oral gavage. Gut bacteria transplantation was performed weekly for 5 weeks for faeces (five times) and every 4 days for 13 days for jejunal contents (four times). For measurement of circulating levels of ²H-labelled fatty acid using ²H₂O, mice received ²H₂O dissolved in 0.9% NaCl by intraperitoneal injection (30 µl g⁻¹) at 10:00 h. Mice were transferred to new cages without food. At 16:00 h, serum was collected by tail snip. For measurement of DNL flux, mice received ²H₂O dissolved in 0.9% NaCl by intraperitoneal injection (30 µl g⁻¹) at 18:00 h and serum was collected via tail snip at 09:00 h the following morning. For ¹³C-fructose tracing, mice received a solution of unlabelled glucose and ¹³C-fructose (2 g kg⁻¹ body weight each) by oral gavage (10 µl g⁻¹ body weight) at 09:00 h. For ¹³C-cysteine tracing, mice received a solution of 0.2 M ¹³C-cysteine by oral gavage (5 µl g⁻¹ body weight) at 09:00 h. At 10:00 h, tail blood and tissues were collected. Intestine-specific Khk-C transgenic mice received a solution of unlabelled glucose and ¹³C-fructose (2 g kg⁻¹ body weight each) by oral gavage (10 µl g⁻¹ body weight) at 09:00 h, and liver tissues were collected at 10:00 h. Tail blood was collected by tail snip to measure circulating metabolites after different durations. Tissues were quickly dissected and snap-frozen in liquid nitrogen with a pre-cooled Wollenberger clamp. Multiple cohorts were used, and data were combined if there was no statistical difference between cohorts. For fasting glucose and insulin measurements, serum was collected after 10 h of fasting (08:00–18:00 h) by tail snip, glucose was measured using liquid chromatography–mass spectrometry (LC–MS) and insulin was measured using an Ultra Sensitive Mouse Insulin ELISA Kit (cat. no. 90080; Crystal Chem).

Bacterial culture

B. acidifaciens (DSM 15896) was purchased from a German collection of microorganisms and cell cultures. B. pseudolongum (ATCC 25526) was obtained from American Type Culture Collection. B. acidifaciens was cultured on Columbia CNA agar medium containing sheep blood (Thermo Scientific) and chopped meat medium (CMM; Fisher Scientific) at 37 °C in the EZ Anaerobe Container System (BD). B. pseudolongum was cultured on modified BHI agar and broth (BD) at 37 °C in the EZ Anaerobe Container System (BD). To measure the growth of B. acidifaciens in response to inulin, PBS, 0.4 g l⁻¹ glucose (Sigma-Aldrich) or 0.4 g l⁻¹ inulin (Sigma-Aldrich) was added to 60% v/v CMM. B. acidifaciens was inoculated, and optical density at 600 nm was measured using a microplate reader (Victor). To measure the fructose catabolic activity of B. acidifaciens, the bacterium was cultured in CMM + glucose or CMM + inulin until the early stationary phase, washed with PBS and then the bacterial pellets were resuspended to the same optical density in CMM + glucose or CMM + inulin supplemented with 0.1 g l^{−1 13}C-fructose for a 3 day culture. To prepare a bacterial solution for oral delivery to mice, bacterial cultures were centrifuged and washed with PBS after 72 h of incubation. The bacterial pellets were resuspended in 25% (v/v) glycerol and stored at −80 °C before oral gavage to mice; 25% glycerol without any bacterium was used for the control. After HFCS provision and antibiotics treatment as described above, 200 µl of bacteria solution and glycerol were transferred to recipient mice by oral gavage (200 µl per mouse) with a plastic feeding tube (Instech Laboratories) every day for 2 weeks.

Histology

Freshly collected liver tissues were fixed in 4% paraformaldehyde overnight, embedded in paraffin, sectioned and stained with haematoxylin and eosin (H&E). Tissues were submitted to the Experimental Tissue Shared Resource Facility at the University of California Irvine. For trichrome staining, Gomori’s Trichrome Stain Kit (Polysciences) was used. Images were captured with a high-resolution image scanner (Ventana DP200, Roche). Hepatic lipid accumulation was quantified by analysing the digital slides using QuPath (v.0.4.4)⁹². A pixel classifier was trained from representative images among the groups. This pixel classifier was then applied to annotations of the same size within each slide. These regions were measured, and the area values (in μm²) for the regions classified as lipids were used.

Indirect calorimetry, ¹³C FAO analysis and echo magnetic resonance imaging

O₂ consumption, CO₂ release, respiratory exchange ratio, locomotor activity and heat production were monitored for individually housed mice using Phenomaster metabolic cages (TSE Systems). The climate chamber was set to 21 °C and 50% humidity, with a 12–12 h light–dark cycle (07:00–19:00 h) as the home-cage environment. Animals were entrained for 24 h in the metabolic cages before the start of each experiment to allow for environmental acclimation. Data were collected at 40 min intervals, and each cage was recorded for 3.25 min before time point collection. For measuring ¹³CO₂, environmental levels of ¹³CO₂ and total CO₂ in the sealed cages were calibrated to ±1.1% ¹³CO₂, as the natural abundance of ¹³C. Body composition was measured using an EchoMRI Whole Body Composition Analyzer, which provides whole-body fat and lean mass measurements. To account for potential changes in ¹³CO₂ loss caused by CO₂ fixation reactions (for example, those catalysed by pyruvate carboxylase or urea carboxylase)^56,57, ¹³CO₂ production was measured using indirect calorimetry after ¹³C-acetate administration. In brief, mice were fasted for 6 h and administered ¹³C-acetate by oral gavage (0.3 mg g⁻¹ body weight), and ¹³CO₂ recovery was measured for individually housed mice using Phenomaster metabolic cages (TSE Systems). Under the same conditions, we performed oral ¹³C-palmitate administration and ¹³CO₂ measurements. To estimate the oxidation of the total circulating palmitate pool from circulating triglycerides and exogenous ¹³C-palmitate, we measured circulating un-esterified ¹³C-palmitate enrichment (%) in serial time points and used these values to normalize ¹³CO₂ (ref. ⁵⁸).

Quantitative PCR with reverse transcription

RNA samples were prepared using TRIzol Reagent (Invitrogen) according to the manufacturer’s instructions. RNA was reverse-transcribed to cDNA using the iScript kit (Bio-Rad). The resulting cDNA was analysed by qPCR with reverse transcription using SYBR green master mix (Life Technologies) on a QuantStudio6 Real-Time PCR system (Life Technologies). Relative mRNA expression was calculated from the comparative threshold cycle values relative to housekeeping genes Actin, 36b4 and Tbp. Primer sequences were: Tgfb1 (forward, CTCCCGTGGCTTCTAGTGC; reverse, GCCTTAGTTTGGACAGGATCTG), Acta2 (forward, ATGCTCCCAGGGCTGTTTTCCCAT; reverse, GTGGTGCCAGATCTTTTCCATGTCG), Vim (forward, TTTCTCTGCCTCTGCCAAC; reverse, TCTCATTGATCACCTGTCCATC), Col1a1 (forward, GCTCCTCTTAGGGGCCACT; reverse, CCACGTCTCACCATTGGGG), Mmp13 (forward, CTTCTTCTTGTTGAGCTGGACTC; reverse, CTGTGGAGGTCACTGTAGACT), Glut5 (forward, TCTTTGTGGTAGAGCTTTGGG; reverse, GACAATGACACAGACAATGCTG), Khk-a (forward, TTGCCGATTTTGTCCTGGAT; reverse, CCTCGGTCTGAAGGACCACAT), Khk-c (forward, TGGCAGAGCCAGGGAGAT; reverse, ATCTGGCAGGTTCGTGTCGTA), Aldob (forward, CACCGATTTCCAGCCCTC; reverse, GTTCTCCACCTTTATCCTTTGC), Tkfc (forward, GCATCTCAGAGCAGAAGTGTG; reverse, CAAGTCAGGGTTAGAGGCTAC), Acly (forward, CAGCCAAGGCAATTTCAGAGC; reverse, CTCGACGTTTGATTAACTGGTCT), Acss2 (forward, ATGGGCGGAATGGTCTCTTTC; reverse, TGGGGACCTTGTCTTCATCAT), Fasn (forward, GGAGGTGGTGATAGCCGGTAT; reverse, TGGGTAATCCATAGAGCCCAG), Scd1 (forward, TTCAGAAACACATGCTGATCCTCATAATTCCC; reverse, ATTAAGCACCACAGCATATCGCAAGAAAGT), Tbp (forward, CCCTATCACTCCTGCCACACCAGC; reverse, GTGCAATGGTCTTTAGGTCAAGTTTAC), Actin (forward, CCCTGTATGCTCTGGTCGTACCAC; reverse, GCCAGCCAGGTCCAGACGCAGGATG) and 36b4 (forward, GGAGCCAGCGAGGCCACACTGCTG; reverse, CTGGCCACGTTGCGGACACCCTCC).

Mitochondria and cytosolic fractionation

To isolate liver mitochondria⁹³, freshly isolated mouse liver was weighed and rapidly extracted and homogenized in ice-cold liver homogenization buffer (1 ml per 200 mg of liver; 200 mM sucrose, 5 mM Tris, 1 mM EGTA, 100 μg ml⁻¹ digitonin pH 7.4), followed by centrifugation at 1000g for 1 min at 4 °C. The resulting 200 μl supernatants were combined with 500 μl spin buffer (150 mM sucrose, 5 mM Tris, 1 mM EGTA, 25 mM ammonium bicarbonate pH 7.4) to reduce suspension density. Each 700 μl mixture was carefully layered above 300 μl liver oil mix (60:40 silicone oil to dioctyl phthalate; density, 1.066 g ml⁻¹ at 20–23.5 °C) and 100 μl of 23% glycerol in pre-mixed tubes, then centrifuged at 9,727g for 1 min at 4 °C. The cytosolic layer and most of the oil were aspirated, and the remaining oil was removed after freezing the lower layer in a dry ice–ethanol bath and washing with dry ice–cold hexane. The mitochondrial pellets were pooled to obtain the final sample.

Mitochondrial DNA analysis

For mtDNA copy number analysis, 50 ng of DNA was extracted with the Purelink Genomic DNA Mini (Invitrogen), and qPCR was performed with a pair of primers for mtDNA (MT-16S rRNA-ND1) with 18S for normalization. The following primers were used: MT-16S-ND1 (forward, CACCCAAGAACAGGGTTTGT; reverse, TGGCCATGGGTATGTTGTTAA) and 18S (forward, TAGAGGGACAAGTGGCGTTC; reverse, CGCTGAGCCAGTCAGTGT).

RNA-seq analysis

RNA samples were prepared using TRIzol Reagent (Invitrogen) according to the manufacturer’s instructions. RNA concentration was quantified using fluorimetry (Qubit 2.0 fluorometer; Life Technologies), and quality was assessed using an Agilent BioAnalyzer 2100 (Agilent Technologies). Ribosomal RNA depletion and sample library preparation were performed using the Illumina TruSeq Stranded Total RNA with RiboZero. The libraries were then sequenced on a NovaSeq 6000 (Illumina) using paired-end sequencing. The quality of the raw sequencing data was assessed using FastQC, and all samples passed the quality control analysis. Raw sequencing reads were aligned to the mouse reference genome mm10 (GRCm38) using STAR, and RNA-seq counts were obtained using featureCounts. These raw counts were further analysed using DESeq2 for differential gene expression analysis. To elucidate the biological pathways involved in our study, we used Kyoto Encyclopedia of Genes and Genomes pathway enrichment analysis using Fisher’s exact test. For gene-set enrichment analysis, gene sets from the Molecular Signatures Database (v.2023.2) were used.

Metabolite measurements using LC–MS

For aqueous metabolites extraction, serum (5 µl) was mixed with 150 µl of extraction solvent (40:40:20 methanol:acetonitrile:water, v:v:v) at −20 °C, vortexed and immediately centrifuged at 16,000g for 10 min at 4 °C. The supernatant (70 µl) was collected for LC–MS analysis. Frozen tissue samples were ground at liquid nitrogen temperature with a CryoMill (Retsch). The resulting tissue powder (approximately 20 mg) was weighed and then mixed with −20 °C extraction solvent containing 0.5% formic acid (40 µl per mg tissue), vortexed and neutralized with 15% NH₄HCO₃ (3.5 µl per mg tissue). Following vortexing and centrifugation at 16,000g for 10 min at 4 °C, the supernatant (70 µl) was loaded into LC–MS vials. Metabolites were analysed by a quadrupole–orbitrap mass spectrometer (Q-Exactive Plus Hybrid Quadrupole–Orbitrap, Thermo Fisher) coupled to hydrophilic interaction chromatography by heated electrospray ionization. LC separation was performed on an Xbridge BEH amide column (2.1 mm × 150 mm, 2.5 µm particle size, 130 Å pore size; Waters) at 25 °C using a gradient of solvent A (5% acetonitrile in water with 20 mM ammonium acetate and 20 mM ammonium hydroxide) and solvent B (100% acetonitrile). The flow rate was 150 µl min⁻¹. The LC gradient was: 0 min, 90% B; 2 min, 90% B; 3 min, 75% B; 7 min, 75% B; 8 min, 70% B; 9 min, 70% B; 10 min, 50% B; 12 min, 50% B; 13 min, 25% B; 14 min, 20% B; 15 min, 20% B; 16 min, 0% B; 20.5 min, 0% B; 21 min, 90% B; and 25 min, 90% B. Autosampler temperature was set at 4 °C and the injection volume of the sample was 3 μl. MS analysis was acquired in negative and positive ion modes with Full MS scan mode from m/z 70 to 830 and 140,000 resolution with the following operational parameters: AGC target, 3 × 10⁶; maximum IT, 500 ms; sheath gas flow rate, 40; aux gas flow rate, 10; sweep gas flow rate, 2; spray voltage, +3.8 kV and −3.5 kV; spray current, 33 μA; capillary temperature, 300 °C; s-lens RF level, 50; aux gas heater temperature, 360 °C. MS2 analysis was acquired in negative and positive ion modes with Full MS/dd-MS2 from m/z 70 to 830. For full MS, the parameters were: resolution, 70,000; AGC target, 1 × 10⁶; maximum IT, 200 ms. For MS/dd-MS2, the parameters were: resolution, 17,500; AGC target, 1 × 10⁵; maximum IT, 50 ms; loop count, 15; isolation window, 1.2 m/z; stepped CE, ±20 and 50 eV with the following operational parameters: sheath gas flow rate, 40; aux gas flow rate, 10; sweep gas flow rate, 2; spray voltage, +3.8 kV and −3.5 kV; spray current, 33 μA; capillary temperature, 300 °C; s-lens RF level, 50; aux gas heater temperature, 360 °C. Data were analysed using the EI-MAVEN software and Compound Discoverer software (Thermofisher Scientific). The identity of metabolites was confirmed based on the retention time and accurate m/z of authentic synthesized chemical standards from Sigma-Aldrich, as well as MS2 fragmentation patterns available in HMDB (https://hmdb.ca) and mzCloud database (https://www.mzcloud.org). Natural isotope correction was performed with AccuCor2 R code (https://github.com/wangyujue23/AccuCor2). Labelled ion counts refer to the sum of all labelled forms, in which each form is weighted by the fraction of carbon atoms labelled. It is used to calculate fractional carbon labelling (%) by normalizing against the ion count of the total pool. The concentrations of selected metabolites were determined by calibration curves using authentic synthesized standards. The metabolite concentrations in tissues or intestinal contents were calculated using the following equation: concentration (μmol g⁻¹) = concentration of extracted sample (μM) × volume of extraction solution (μl) / tissue weight (mg).

Lipid measurements using LC–MS

For lipid extraction, samples were mixed with −20 °C isopropanol (150 µl per 5 µl serum and 40 µl per mg tissue), vortexed and immediately centrifuged at 16,000g for 10 min at 4 °C¹⁵. The supernatant (70 µl) was loaded into LC–MS vials. Lipids were analysed by a quadrupole–orbitrap mass spectrometer (Q-Exactive Plus Hybrid Quadrupole–Orbitrap) coupled to reverse-phase chromatography with electrospray ionization. LC separation was on an Atlantis T3 Column (2.1 mm × 150 mm, 3 µm particle size, 100 Å pore size; Waters) at 45 °C using a gradient of solvent A (1 mM ammonium acetate, 35 mM acetic acid in 90:10 water:methanol) and solvent B (1 mM ammonium acetate, 35 mM acetic acid in 98:2 isopropanol:methanol). The flow rate was 150 µl min⁻¹. The LC gradient was: 0 min, 25% B; 2 min, 25% B; 5.5 min, 65% B; 12.5 min, 100% B; 16.5 min, 100% B; 17 min, 25% B; and 30 min, 25% B. MS analysis was acquired in positive ion mode with Full MS and MS/dd-MS2 scan mode from m/z 290 to 1,200. The MS operational parameters and data analysis are the same as for the metabolite analysis.

Saponified fatty acid measurement using LC–MS

Serum (5 µl) or liver powder (20 mg) was incubated with 0.5 ml of 0.3 M KOH in 90% methanol at 80 °C for 1 h in a 2 ml glass vial. Then, formic acid (50 µl) was added for neutralization. The saponified fatty acids were extracted by adding 500 µl of hexane and vortexing. After 5 min for separation of the layers, 250 µl of the top hexane layer was transferred to a new glass vial. Samples were then dried under a nitrogen gas stream and redissolved in 100 µl (for serum) or 500 µl (for liver) of 1:1 isopropanol:methanol for LC–MS analysis¹⁵. Fatty acids were analysed by a quadrupole–orbitrap mass spectrometer (Q-Exactive Plus Hybrid Quadrupole–Orbitrap) coupled with reverse-phase chromatography with electrospray ionization. LC separation was performed on an Atlantis T3 Column (2.1 mm × 150 mm, 3 µm particle size, 100 Å pore size; Waters) at 45 °C using a gradient of solvent A (1 mM ammonium acetate, 35 mM acetic acid in 90:10 water:methanol) and solvent B (1 mM ammonium acetate, 35 mM acetic acid in 98:2 isopropanol:methanol). The flow rate was 150 µl min⁻¹. The LC gradient was: 0 min, 25% B; 2 min, 65% B; 5.5 min, 100% B; 16.5 min, 100% B; 16.5 min, 25% B with a flow rate of 200 µl min⁻¹; 19 min, 25% B with a flow rate of 200 µl min⁻¹; 19.1 min, 25% B with a flow rate of 150 µl min⁻¹; and 20 min, 25% B. MS analysis was acquired in negative ion mode with Full MS scan mode from m/z 200 to 530. The MS operational parameters and data analysis are the same as for the metabolite analysis

Body water enrichment and DNL calculation

To quantify body water enrichment, 5 μl of serum, 5 μl of water, 4 μl of 1 M sodium hydroxide and 10 μl of acetone were mixed in a glass vial and incubated overnight at room temperature to promote ²H exchange between ²H₂O in serum and acetone⁹⁴. The resulting ²H-acetone was derivatized to 2,4-dinitrophenylhydrazine (2,4-DNPH)⁹⁵. The 2,4-DNPH solution was prepared by dissolving 20 mg of 2,4-DNPH in 10 ml of ethanol with 100 μl of H₂SO₄ and 150 μl of water. The precipitate formed upon mixing was then isolated by filtration (cat. no. 09-790-D; Fisher Scientific). The filtrate was treated with 1 ml of H₂SO₄, and 5 μl of this solution was mixed with 20 μl of the sample solution. After a 2 h incubation at room temperature, 200 μl of ethanol was added, and the samples were transferred to 1.5 ml tubes for centrifugation at 4 °C for 20 min. The clear supernatant was transferred into a glass vial for LC–MS analysis. ²H₁ acetone and unlabelled acetone were measured by LC–MS analysis, using the same method as for SCFA analysis (see below), and were used to calculate the fraction of ²H₁ acetone. Calibration standards of known ²H fraction water were prepared by mixing naturally labelled water and 99.9% ²H₂O. The ²H₁ acetone fraction of ²H₂O serial dilution in naturally labelled water was used to generate a standard curve. Then, the ²H₁ acetone fraction in the serum samples was substituted into the standard curve equation to calculate body water enrichment, as previously described^53,96. The contribution of fatty acid synthesis was determined using equation (1).

²H-labelled palmitate enrichment was calculated using equation (2), where ²H₁, ²H₂ and ²H₃ indicate the ²H-labelled fraction of each isotopologue.

The number of exchangeable hydrogens (n) was calculated using the fractions of ²H₁ and ²H₂ palmitate as in equation (3). The rate of DNL per hour was determined by dividing the time elapsed since ²H₂O administration.

SCFA measurement using LC–MS

Serum (5 µl) or intestinal content (1 mg) was mixed with derivatizing reagent (100 µl) and incubated for 1 h at 4 °C. The derivatizing reagent was prepared by mixing 12 mM N-(3-dimethylaminopropyl)-N′-ethylcarbodiimide, 25 mM 3-nitrophenylhydrazine and pyridine (4% v/v) in 40:40:20 methanol:acetonitrile:water (v/v/v). Samples were centrifuged at 16,000g for 10 min at 4 °C, and 10 µl of supernatant was mixed with 90 µl of the quenching reagent (0.5 mM β-mercaptoethanol in water). After centrifugation at 16,000g for 10 min at 4 °C, the supernatants were collected for LC–MS analysis. SCFAs were analysed using a quadrupole–orbitrap mass spectrometer (Q-Exactive Plus Hybrid Quadrupole–Orbitrap) coupled with reverse-phase chromatography with electrospray ionization. LC separation was performed on an Atlantis T3 Column (2.1 mm × 50 mm, 3 µm particle size, 100 Å pore size; Waters) using a gradient of solvent A (water) and solvent B (methanol) at 60 °C. The flow rate was 300 µl min⁻¹. The LC gradient was: 0 min, 10% B; 2.3 min, 80% B; 3.6 min, 80% B; 3.7 min, 10% B; and 5 min, 10% B. MS analysis was acquired in negative ion mode with Full MS scan mode from m/z 100 to 300. The MS operational parameters and data analysis are the same as for the metabolite analysis.

Malondialdehyde measurement using LC–MS

Malondialdehyde in liver tissue (10 mg) was mixed with 20 µl of butylated hydroxytoluene solution (1 g l⁻¹ in ethanol) and 220 µl of 50% ethanol solution⁹⁷. Samples were centrifuged at 16,000g for 10 min at 4 °C, and 200 µl of supernatant was mixed with 200 µl of 2,4-dinitrophenyl-hydrazine solution (0.05 M in acetonitrile:acetic acid 9:1 (v/v)). After incubating samples for 2 h at 60 °C, the samples were mixed with 530 µl of water and 1 ml of hexane. Samples were vortexed and incubated for 10 min at 25 °C to separate the layers. Then, 500 µl of the top hexane layer was transferred to a new glass vial. Samples were then dried under a nitrogen gas stream and redissolved in 200 µl of acetic acid solution (0.03% in acetonitrile:water 4:6 (v/v)). After centrifugation at 16,000g for 10 min at 4 °C, 150 µl of the supernatant was collected for LC–MS analysis. Malondialdehyde was analysed using a quadrupole–orbitrap mass spectrometer (Q-Exactive Plus Hybrid Quadrupole–Orbitrap) coupled to reverse-phase chromatography with electrospray ionization. LC separation was performed on an Atlantis T3 Column (2.1 mm × 150 mm, 3 µm particle size, 100 Å pore size; Waters) at 45 °C using a gradient of solvent A (1 mM ammonium acetate, 35 mM acetic acid in 90:10 water:methanol) and solvent B (1 mM ammonium acetate, 35 mM acetic acid in 98:2 isopropanol:methanol). The flow rate was 150 µl min⁻¹. The LC gradient was: 0 min, 25% B; 2 min, 100% B; 5 min, 100% B; 11.5 min, 100% B; 11.6 min, 25% B; and 15 min, 25% B. MS analysis was acquired in negative ion mode with Full MS scan mode from m/z 200 to 400. The MS operational parameters and data analysis are the same as for the metabolite analysis. Malondialdehyde was also quantified using a thiobarbituric acid reactive substances (TBARS) assay kit (10009055, Cayman Chemical).

Immunofluorescence imaging

Tissue samples were fixed in 4% buffered paraformaldehyde for 1 h at room temperature. After fixation, tissues were dehydrated with 30% sucrose in PBS overnight, frozen and embedded in Frozen Section Media (Leica) and cut into 20 μm-thick sections using a Cryocut Microtome (Leica). The samples were then blocked in protein block serum (Agilent) with 0.3% Triton X-100 (Thermo Fisher) for 1 h. For 4-HNE staining, the sections were incubated overnight at 4 °C with anti-4-HNE antibody (clone 12F7; Invitrogen, 1:200) in antibody diluent (Agilent), then with anti-mouse secondary antibody conjugated to Alexa Fluor 488 (1:1,000; Jackson ImmunoResearch) for 2 h at room temperature. For reactive oxygen species staining, the sections were incubated with 5 µM dihydroethidium (309800, Sigma-Aldrich) for 30 min at 37 °C. After staining, the samples were washed three times for 10 m each in PBS with 0.3% Triton X-100. Finally, tissue sections were mounted with Vectashield plus DAPI (4′,6-diamidino-2-phenylindole) (Vector Labs) and imaged with an LSM 980 confocal microscope (Zeiss).

Bacteria 16S rDNA quantification

Bacterial DNA was extracted from faecal pellets (10–20 mg) using Quick-DNA Faecal/Soil Microbe Kits (Zymo Research) according to the manufacturer’s instructions. Purified DNA was amplified by qPCR using SYBR green master mix (Life Technologies) on a QuantStudio6 Real-Time PCR system (Life Technologies). DNA from the E. coli DH5a strain was used as a standard for determining the copy number of the 16S rDNA gene of universal bacteria by qPCR. Primer pairs targeting the bacterial universal 16S rRNA gene, Bacteroides spp. and B. pseudolongum were selected from previous studies^98,99 as follows: bacterial universal 16S rRNA gene (forward, GTGGTGCACGGCTGTCGTCA; reverse, ACGTCATCCACACCTTCCTC), B. pseudolongum (forward, CRATYGTCAAGGAACTYGTGGCCT; reverse, GCTGCGAMGAKACCTTGCCGCT) and Bacteroides spp. (forward, CTGAACCAGCCAAGTAGCG; reverse, CCGCAAACTTTCACAACTGACTTA). Relative bacterial abundances of Bacteroides spp. and B. pseudolongum were calculated from threshold cycle values relative to universal bacterial abundance in each sample.

16S rRNA gene amplicon sequencing and analysis

The ZymoBIOMICS-96 MagBead DNA Kit (Zymo Research) was used to extract DNA from mouse jejunal or caecal contents. Bacterial 16S ribosomal RNA gene-targeted sequencing was performed using the Quick-16S NGS Library Prep Kit (Zymo Research). The bacterial 16S primers amplified the V3–V4 region of the 16S rRNA gene. The sequencing library was prepared using an innovative library preparation process in which PCR reactions were performed in real-time PCR machines to control cycles and, therefore, limit PCR chimera formation. The final PCR products were quantified with qPCR fluorescence readings and pooled together based on equal molarity. The final pooled library was cleaned with the Select-a-Size DNA Clean & Concentrator (Zymo Research), then quantified with TapeStation (Agilent Technologies) and Qubit (Thermo Fisher Scientific). The final library was sequenced on Illumina MiSeq with a v3 reagent kit (600 cycles). The sequencing was performed with 10% PhiX spike-in. Unique amplicon sequence variants were inferred from raw reads using the DADA2 pipeline. Taxonomy assignment was performed using Uclust from Qiime (v.1.9.1) with the Zymo Research 16S reference database. Composition visualization, alpha diversity and beta diversity analyses were performed with Qiime (v.1.9.1). Taxonomy that has significant abundance among different groups was identified by linear discriminant analysis effect size using default settings. To compare microbiome diversity among donor and recipient mice of the jejunal content transplant experiment, NMDS analysis based on Bray-Curtis dissimilarities was performed using Microbiome Analyst (v.2.0). Amplicon sequence variants were used to plot for NMDS analysis in Extended Data Fig. 3j.

Statistical analysis

Data collection and analysis were not performed blind to the conditions of the experiments. Data distribution was assumed to be normal, but this was not formally tested. Tukey’s HSD test and two-sided or one-sided Student’s t-test were used to calculate P values, with P < 0.05 considered significant. False discovery rate correction was performed for metabolomics with the Benjamini and Hochberg method. Outliers were defined as values more than 1.5 times the interquartile range below quartile 1 or above quartile 3 (ref. ¹⁰⁰).

Reporting summary

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