All experimental procedures and protocols performed were reviewed and approved by the University of East Anglia Animal Welfare and Ethical Review Body (UEA AWERB) and were conducted in accordance with the specification of the United Kingdom Animal Scientific Procedures Act, 1986 (Amendment Regulations 2012) under the Home Office project licence PP9417531. Reporting of the study outcomes comply with the ARRIVE (Animal Research: Reporting of In Vivo Experiments) guidelines61.
Generation of C3GnT−/− and C3GnT+/+ littermates
We first bred C57BL/6 mice with the C3GnT−/− mice to generate mice heterozygous for the gene (C3GnT+/−). Once the pups (F1) reached sexual maturity (~ 6 to 8 weeks old), males and females C3GnT+/− were bred to generate the F2 mouse colony. The F2 litters were composed of 25% homozygous wild-type (C3GnT+/+), 50% heterozygous C3GnT+/− and 25% homozygous knock-out (C3GnT−/−) mice. After weaning (at ~ 3 weeks old), the mice litters were genotyped by analysis of genomic DNA (gDNA) from ear biopsies. Briefly, gDNA was extracted from ear biopsies using Quick Extract DNA extraction solution (Lucigen) following supplier’s advice. PCR was carried out using the GO TAQ (R) hot start polymerase (Promega, UK), with primers used at 1 µM final concentration (Table S2). The conditions consisted of an initial 2 min 95 °C denaturation step followed by 30 cycles of 94 °C for 45 s, 58 °C for 45 s, 72 °C for 45 s, and a final extension step of 5 min at 72 °C. The PCR products were analysed by electrophoresis on 1.5% agarose gels in 1× Tris–Acetate EDTA (TAE) for 60 min at 80 V in the presence of 100 bp DNA markers (New England Biolabs, USA). DNA was stained by adding 1 μL of Midori green direct DNA stain (Geneflow, UK) to 9 μL of PCR product prior to loading on the gel and imaged under UV light using an Alphaimager.
The C3GnT+/+ and C3GnT−/− mice were then individually caged for 4 weeks (isolation period), in order to allow the development of the gut microbiota according to the genotype of mice coming from the same breeding pairs (littermate-control). Faecal samples were collected before the isolation time (~ 21 days after birth referred to as ‘Time 1’) and after the isolation period (~ 75 days after birth referred to as ‘Time 2’). When the mice reached ~ 9 to 10 weeks old, they were culled by raising concentration of CO2 and, after confirmation of death by dislocation of the neck, the brain, blood, caecal content, scraped mucus from colon were collected and stored for downstream analyses. C3GnT+/+ (n = 8) and C3GnT−/− (n = 8) mice were also used for the battery of behavioural tests (Fig. S9).
DNA was extracted from faecal samples and scraped colonic mucus using the Fast DNA™ SPIN kit for Soil DNA extraction (MP Biomedicals, USA) with the following modifications. The weight of faecal material was measured in tared tubes. The samples were resuspended in 978 μL of sodium phosphate buffer (provided) before being incubated at + 4 °C for 1 h following addition of 122 μL of lysis solution MT Buffer. The samples were then transferred into the lysing tubes and homogenised in a FastPrep® Instrument (MP Biomedicals) 3 times for 40 s at a 6.0 m/s speed with 5 min interval on ice between each bead-beating step. The protocol was then followed as recommended by the supplier.
16S rDNA sequencing
Following extraction of DNA from faecal pellet and mucus, the concentration and quality of DNA was assessed by Qubit and Nanodrop. DNA was normalised to 5 ng/μL and sequenced in-house. The V3 and V4 16S rDNA gene regions were amplified using universal primers. For the first PCR, each well contained 4 μL kapa2G buffer, 0.4 μL dNTPs, 0.08 μL kapa2G polymerase, 0.4 μL 10 μM forward tailed specific primer, 0.4 μL 10 μM reverse tailed specific primer, 13.72 μL PCR grade water and 1 μL normalised DNA. The PCR conditions were 95 °C for 5 min followed by 30 cycles of 95 °C for 30 s, 55 °C for 30 s and 72 °C for 30 s with a final step at 72 °C for 5 min. A 0.7× SPRI using KAPA Pure Beads (Roche, UK) was then performed and the DNA was eluted in 20 μL of EB (10 mM Tris–HCl).
Following the first PCR, a second PCR was performed using 5 μL of the clean PCR product, 4 μL kapa2G buffer, 0.4 μL dNTPs, 0.08 μL kapa2G polymerase, 2 μL of each P7 and P5 primers of Nextera XT Index Kit v2 index primers (Illumina Catalogue No. FC-131-2001 to 2004) were added to each well. Finally, the 5 µL of the clean specific PCR mix was added and mixed. The PCR was run using 95 °C for 5 min, 10 cycles of 95 °C for 30 s, 55 °C for 30 s and 72 °C for 30 s followed by a final step at 72 °C for 5 min. The obtained libraries were then quantified using the Quant-iT dsDNA Assay Kit, high sensitivity kit (Fisher Scientific, UK) on a FLUOstar Optima plate reader. Libraries were pooled following quantification in equal quantities. The final pool was cleaned with 0.7× SPRI using KAPA Pure Beads. The final pool was quantified on a Qubit 3.0 instrument and run on a High Sensitivity D1000 ScreenTape (Agilent) using the Agilent Tapestation 4200 to calculate the final library pool molarity.
The pool was run at a final concentration of 8 pM on an Illumina MiSeq instrument using MiSeq® Reagent Kit v3 (600 cycles, Illumina) following the Illumina denaturation and loading recommendations which included a 20% PhiX spike in (PhiX Control v3 Illumina). The raw data were analysed locally on the MiSeq using MiSeq reporter.
Raw reads produced by Illumina MiSeq sequencing of the 16S amplicons have been processed with the Dadaist2 0.8 workflow CIT62, using SeqFu 1.9 CIT63 to remove the universal primers used to amplify the target region, and then processed via DADA2 1.1664 to identify the Amplicon Sequence Variants (ASVs) and to generate a contingency table (with the raw abundance of each ASV in each sample). The DECIPHER package65 has been used to assign taxonomy to each ASV identified using the SILVA 138 database66, and the results have been exported to a PhyloSeq CIT67 object for downstream analysis.
RNA extraction and RT-qPCR
Mouse colonic tissue and brain were harvested and stored in RNAlater at + 4 °C until RNA extraction. Samples (10–100 mg) were then transferred in a tube containing 1 mL of QIAzol Lysis Reagent (Qiagen, UK) and stainless-steel beads of 5 mm (Qiagen, UK). Homogenisation was achieved using the FastPrep®-24 by 2 intermittent runs of 60 s at 4.0 m/s speed with 5 min interval at room temperature. RNA extraction was performed using the RNeasy Lipid Tissue Mini Kit (Qiagen) following the manufacturer’s instructions for purification of total RNA from animal tissues. Elution was performed as recommended with 50 μL RNAse-free water. The quality and concentration of the RNA samples was assessed using the NanoDrop 2000 spectrophotometer, the Qubit RNA HS assay on Qubit® 2.0 fluorometer (Life Technologies, UK) or Agilent RNA 600 Nano kit on Agilent 2100 Bioanalyzer (Agilent Technologies, Stockport, UK). Following RNA extraction, cDNA synthesis was carried out using QuantiTect Reverse Transcriptase (Qiagen) according to the manufacturer’s instructions including controls lacking the transcriptase to test for DNA contamination (-RT control).
For quantitative reverse transcription PCR (RT-qPCR) reactions were performed using SYBR green detection technology on the Roche light cycler 480 (Roche Life Science, UK). The samples were loaded into 384 microtiter plates in a randomised way and the primers were also randomised across triplicates. Primer sequences are given in Table S3.
Soluble mucin extraction and purification
The mucus was scraped from the colon and resuspended in cold PBS and kept in ice. The solution was then centrifuged at 1000×g for 30 min to separate bacteria (pellet) and soluble proteins (supernatant). The bacteria pellet was stored at − 80 °C until DNA extraction. The presence of Muc2 in the soluble fraction was confirmed by slot blot, using an anti-MUC2 antibody, after blocking the PVDF membrane with Pierce™ Protein-Free (PBS) Blocking Buffer (ThermoFisher). The supernatant from samples from 15 C3GnT+/+ and 12 C3GnT−/− were pooled together according to the genotype and processed for mucin extraction and purification. Briefly, caesium chloride was added to the samples at a final density of 1.4 g/mL and the samples were ultracentrifuged at 42,000 rpm on a Beckman Coulter 70 Ti rotor. Fractions (1 ml each) were collected and those with a density higher than 1.4 g/mL were pooled together, dialysed against water and freeze-dried. The purified mucins were subjected to β-elimination under reductive conditions (0.1 M NaOH, 1 M sodium borohydride, NaBH4) for 20 h at 45 °C. The reaction was stopped by slowly adding 5 drops of glacial acetic acid (Sigma). A desalting column was assembled by packing a Paster pipette, with Glass Wool to control the flow (Sigma) and Dowex 50 W × 8 hydrogen form beads (Sigma). Following elution of the desalting column with 15 mL of glacial acetic acid, the samples, were loaded onto the column. The collected fractions were dried under a stream of nitrogen to evaporation the excess of borates.
Permethylation was performed on released O-glycans from mucin samples. Samples were solubilised in 200 μL dimethyl sulfoxide (DMSO). Then NaOH (1 pellet) and 300 μL iodomethane were added to the suspension in anhydrous conditions and the samples vigorously shaken at room temperature for 90 min. The permethylation reaction was stopped by addition of 1 mL acetic acid (5% vol/vol). Permethylated O-glycans were extracted in 1–2 mL of chloroform and 5 ml of ultrapure water, mixed thoroughly by vortexing and centrifuged at 14,000×g for 2 min to allow the mixture to set into two layers. The upper aqueous part (containing DMSO) was removed and discarded. The lower chloroform layer was washed five times with MilliQ water (1 mL) and then dried under nitrogen using an evaporator unit. The dried samples were dissolved in 30% acetonitrile in 0.1% aqueous trifluoroacetic acid (TFA, Sigma) and mixed 1:1 with 20 mg/mL 2,5-dihydroxy-benzoic acid (DHBA) in the same solvent. 2 μL were spotted on a matrix assisted laser desorption ionization (MALDI) target plate for analysis.
MALDI-ToF MS mucin glycosylation analysis
MALDI-TOF (Tandem time of flight) and TOF/TOF–MS data were acquired using the Bruker Autoflex analyzer mass spectrometer (Applied Biosystems, Foster City, CA, US) in the positive-ion and reflectron mode. The relative quantification of sialylation on mucins was calculated based on the sum of all areas of mass peaks corresponding to sialylated structures divided by the sum of all areas of mass peaks corresponding to defined O-glycans. Similar calculation was done to determine the relative quantification of fucosylation or sulfation on mucins. The identification of glycan structures was conducted using Glycoworkbench software68.
FITC-dextran in vivo permeability assay
To assess intestinal permeability, C3GnT+/+ and C3GnT−/− mice were fasted for 4 h before administration of 150 μL of fluorescein isothiocyanate dextran (FITC) (80 mg/mL 4 kD; Sigma-Aldrich) in sterile 1× PBS by oral gavage. Mice were culled following schedule-1 by raising concentration of CO2 4 h following the FITC treatment, and blood was collected by intracardial puncture. The blood samples were diluted 1:4 in PBS and the concentration of the FITC-dextran was determined using a fluorimeter (FLUOstar OPTIMA, BMG LABTECH) with an excitation wavelength at 490 nm and an emission wavelength of 520 nm. Serial-diluted FITC-dextran was used to generate a standard curve (from 8000 to 0.125 ng/mL) (Fig. S3).
The caecal content (100–200 mg) from C3GnT+/+ (n = 6) and C3GnT−/− (n = 6) mice, the brain (100–200 mg) from C3GnT+/+ (n = 9) and C3GnT−/− (n = 9) mice and the serum (50–60 µL) pooled from C3GnT+/+ (n = 5) and C3GnT−/− (n = 5) mice were processed and analysed by Metabolon, Inc, USA. A total of 705 metabolites were detected in the brain, 911 in the caecal contents and 945 detected the serum. Metabolites were identified by comparison to library entries of purified standards or recurrent unknown entities. Metabolon maintains a library based on authenticated standards that contains the retention time/index (RI), mass to charge ratio (m/z), and chromatographic data (including MS/MS spectral data) on all molecules present in the library. Furthermore, biochemical identifications are based on three criteria: retention index within a narrow RI window of the proposed identification, accurate mass match to the library ± 10 ppm, and the MS/MS forward and reverse scores between the experimental data and authentic standards. The MS/MS scores are based on a comparison of the ions present in the experimental spectrum to the ions present in the library spectrum.
Free floating immunolabelling of brain sections
Following transcardial flushing with 4% paraformaldehyde (PFA) in PBS, mouse brains were extracted and kept overnight in PFA at + 4 °C. Next, the brains were washed in PBS and dehydrated by incubation in an ascending alcohol series, starting from 30–50–70–90 to 100% of ethanol for 1 h each, followed by hydration in decreasing concentrations of ethanol, from 100 to the 30% and using PBS in the final step. The brains were then embedded in 3% agar with the bulbs facing the top and 60 µm slices were cut using a vibratome (Leica, VT1200S). The floating brain sections were then transferred into a multi-well plate (Corning, UK) with a paintbrush. For immunohistochemistry, free floating sections were selected according to the region of interest, following a mouse brain atlas (Paxinos and Franklin’s the Mouse Brain in Stereotaxic Coordinates, Compact 5th Edition).
Brain sections (Bregma coordinates from − 1.22 to − 2.46 mm, including hippocampus and hypothalamus) were first incubated with antigen unmasking buffer (10 mM citrate buffer, 0.05% Tween 20, pH 6.0) pre-warmed at 70 °C for 15 min at 70 °C in a water-bath. The sections were then blocked for 2 h with a PBS solution containing 20% normal goat serum (NGS, Gibco) and 1% Triton X100 and then incubated overnight at + 4 °C with primary antibodies (Table S4) in a solution containing 0.2% NGS and 0.1% Triton X100. The sections were then washed five times by incubation of 1 h each at room temperature in 0.2% NGS and 0.1% Triton X100 and then incubated overnight at + 4 °C with the relevant secondary antibodies diluted in the buffer used for the primary antibodies (Table S4). Following washing in PBS (6 times, 30 min incubation each), the sections were stained with 4ʹ,6-diamidin-2-fenilindolo DAPI (1 µg/mL, Thermo Fisher, UK) for 5 min, washed, mounted and cover-slipped with mounting medium (VECTASHIELD Antifade Mounting Medium-H100).
Western blot analysis
Fresh extracted brains from C3GnT+/+ and C3GnT−/− littermates were separated in the two hemispheres to get sagittal sections and stored at − 80 °C until use; for the tissue lysis then the exposed hippocampal area was excised with a blade from the right and left hemisphere, resulting in the hippocampus and surrounding regions. The tissue extracts (15–20 mg) were homogenized in 500 µL of lysis buffer (20 mM Tris–HCl, pH 8.0, 150 mM NaCl, 5 mM EDTA, 1 mM phenylmethylsulphonyl fluoride, 70 µg/mL aprotinin, 10 µg/mL leupeptin, 2% Triton X-100) as described69. Lysates were treated with 25 ng/µL endosialidase for 45 min on ice. The proteins were then separated by 4–12% SDS-PAGE (50 µg of total protein per lane), followed by Western blotting for 16 h at 4 °C. Membranes were incubated with 0.4 µg/mL of NCAM-specific rat monoclonal antibody (mAb) H28 and 1 µg/mL of polySia-specific mouse mAb 735 (IgG2a) (kindly provided by Prof Herbert Hildebrandt). To identify tight junction proteins, ZO-1-specific rabbit polyclonal antibody (PABs) 1:1000 (Thermo Fischer Scientific) and Occludin-specific mouse mAb 1:1000 (Thermo Fischer Scientific) were used. GAPDH was used as loading control using a GAPDH- specific PABs (Abcam). Bound antibodies were detected with peroxidase-coupled anti-mouse or anti-rabbit IgG (Vector Laboratories) and developed by enhanced chemiluminescence using Clarity Western ECL Substrate (BioRad). Quantification was carried out using ImageLab software (Bio-Rad).
The novel object recognition (NOR), a measure of recognition memory, was performed as described previously70,71 with slight modifications. Briefly, on day 1, mice (n = 8/group) were habituated in a grey 50 × 50 × 50 cm apparatus illuminated with low lux (100 lx) lighting, by being placed into the empty maze and allowed to move freely for 10 min then returned to the cage. On day 2, mice were placed back in the maze and conditioned to a single object for a 10 min period. On day 3, mice were placed into the same experimental area in the presence of two identical objects for 15 min, after which they were returned to their respective cages and an inter-trial interval of 1 h was observed. One of the familiar objects was replaced with a novel object. Mice were placed back within the testing area for a final 10 min. Videos were analysed for a 5 min period, after which, if an accumulative total of 8 s in the presence of both objects (familiar and novel) failed to be reached, the analysis continued for the full 10 min or until the mouse would complete 8 s of time spent close to the object. The mice not reaching the required time were excluded from the analysis72. A discrimination index (DI) was calculated as follows: DI = (TN − TF)/(TN + TF), where TN is the time spent exploring the novel object and TF is the time spent exploring the familiar object.
The Y-maze spontaneous alternation test, a measure of spatial working memory, was performed on the final day of behavioural testing, as previously described73. Briefly, the Y-maze apparatus made of white Plexiglas of three distinct arms (dimensions 38.5 × 8 × 13 cm, each arm at a 120° angle from each other) was illuminated with low lux (100 lx) lighting. The mouse to test was placed in the maze and allowed to explore freely for 7 min whilst tracking software recorded zone transitioning and locomotor activity (Ethovision XT, Noldus, UK). It is expected that the mouse following the natural exploratory instinct, goes around the maze and visits each arm equally.
The open field test (OFT) was conducted as previously described74. Briefly, the mouse to be tested was placed in the centre of the OFT and a video tracking system (Ethovision XT, Noldus, UK) recorded the total distance the mice travelled, as well as with the time they spent in the centre of the field within the first 5 min. The open field maze was cleaned between each mouse with 20% ethanol to eliminate odours. It is expected that the mouse goes around exploring the box, staying close to walls (thigmotaxis)42.
All statistical analyses were conducted using the R version 4.1.0. In the glycan analysis from intestinal mucins, at least three technical replicates were performed for each fraction and t-tests were used to compare carried out for each fraction between C3GnT+/+ and C3GnT−/− littermates.
In the analysis of the gut microbiota, 16S data were filtered such that only ASVs with count greater than 3 in at least 20% of samples were included. ASVs were then aggregated into families for primary analysis. The relative abundances and centred-log ratios of each family were plotted by genotype with means and confidence intervals.
Bray–Curtis distance matrix was calculated based on total sum scaled data, and multidimensional scaling (MDS) ordination was plotted in strata corresponding to sample point (Time 1 vs Time 2). PERMANOVA was conducted using adonis2 from the vegan R package. For differential abundance analysis, two approaches were used. First, DESeq2 was applied on the count data (aggregated at family level) separately for faecal samples (Time 2) and scraped mucus independently to compare family counts across genotypes. All DESeq2 defaults were accepted, including Benjamini–Hochberg correction to p-values.
Second, standard linear regression models were applied to centred-log ratios calculated using the microbiome package in R. Analyses were also repeated on three other datasets: (1) all ASVs, (2) top 20 ASVs only aggregated into families, (3) data aggregated to Phylum level.
Differences in gene expression between groups were calculated by estimating the effect of genotype on Ct values, adjusting for the Ct values of both housekeeping genes (Gapdh and Tbp), with individual mouse included as a random effect and replicate as a fixed effect. The data from three samples were removed prior to analysis, as outliers based on visual inspection of the Ct values. Data are presented as estimated ratio of gene expression (C3GnT−/− vs C3GnT+/+) with 95% confidence intervals for each gene. P-values were adjusted using the Benjamini–Hochberg method taking into account fourteen genes being simultaneously tested.
An estimate of the false discovery rate (q-value) was calculated to take into account the multiple comparisons that normally occur in metabolomic-based studies. The q-value describes the false discovery rate; a low q-value (q < 0.10) is an indication of high confidence in a result.
In the behavioural study, data were analysed using the “Origin 9” software (OriginLab, Northampton, USA). Unpaired t-test with normality checks of QQ-plots were used for the assessment of behavioural tests in mice. Outliers were detected by ROUT or Grubbs test.