Mucin-derived O-glycans supplemented to diet mitigate diverse microbiota perturbations

Abstract

Microbiota-accessible carbohydrates (MACs) are powerful modulators of microbiota composition and function. These substrates are often derived from diet, such as complex polysaccharides from plants or human milk oligosaccharides (HMOs) during breastfeeding. Host-derived mucus glycans on gut-secreted mucin proteins serve as a continuous endogenous source of MACs for resident microbes; here we investigate the potential role of purified, orally administered mucus glycans in maintaining a healthy microbial community. In this study, we liberated and purified O-linked glycans from porcine gastric mucin and assessed their efficacy in shaping the recovery of a perturbed microbiota in a mouse model. We found that porcine mucin glycans (PMGs) and HMOs enrich for taxonomically similar resident microbes. We demonstrate that PMGs aid recovery of the microbiota after antibiotic treatment, suppress Clostridium difficile abundance, delay the onset of diet-induced obesity, and increase the relative abundance of resident Akkermansia muciniphila. In silico analysis revealed that genes associated with mucus utilization are abundant and diverse in prevalent gut commensals and rare in enteric pathogens, consistent with these glycan-degrading capabilities being selected for during host development and throughout the evolution of the host-microbe relationship. Importantly, we identify mucus glycans as a novel class of prebiotic compounds that can be used to mitigate perturbations to the microbiota and provide benefits to host physiology.

Fig. 1. Human milk oligosaccharides (HMOs) are consumed by and shape the commensal microbiota.

a Germ-free and mice bi-colonized with Bt and B. infantis were fed MD diet supplemented with 1% HMOs (w/v in water). HPLC-FL-based chromatograms of glycan content in stool samples at day 7 reveal degradation of HMOs in vivo. b HMO supplementation provides a competitive advantage to B. infantis over Bt in bi-colonized mice (n = 4 mice/group, mean ± SEM, ***P < 0.001, unpaired t-test). Abundance of Bt and B. infantis from feces was determined with CFU dilution plating. c Bray–Curtis dissimilarity metric indicates significant differences to community composition between mice on MD diet and MD diet supplemented with 1% PMGs, or HMOs. MD diet alone is not significantly different from MD diet supplemented with 1% GOS. (Data were combined across sampling timepoints: n = 12 MD, n = 19 MD + GOS, n = 21 MD + HMOs, n = 11 MD + PMG, n = 61 MAC⁺, n = 10 MAC⁺ + PMG. a. *MAC⁺ vs. MD; ****MAC⁺ vs. MD + GOS, MD + HMOs, and MD + PMG. b. ***MAC⁺ + PMG vs. MD; ****MAC⁺ + PMG vs. MD + GOS, MD + HMOs, and MD + PMG. ***P < 0.001, ****P < 0.0001, pairwise t-tests with Bonferroni multiple hypothesis correction). d Individual ASVs with significantly different abundance due to HMO supplementation (positive log₂-fold change) versus MD diet alone (negative log₂-fold change, mean ± SEM, adjusted P value < 0.01, parametric Wald test). Highest resolution taxonomic assignment indicated to the left.

Fig. 2. Structural analysis of porcine mucin glycans (PMGs).

a Six detectable monosaccharides were identified after total acid hydrolysis of PMGs, with four monosaccharides associated with mucin glycans dominating. Amino sugars are the N-acetyl forms. b 13 abundant O-linked glycans were detected in purified PMGs using MALDI-TOF mass spectrometry. c, d Structures of the O-glycans quantified in (b) were predicted with GlycoWork Bench. Three of the identified PMGs (c) were validated with MS/MS.

Fig. 3. Treatment with PMGs leads to accelerated postantibiotic recovery.

a MD diet supplemented with 1% PMGs (+PMG) leads to accelerated recovery of alpha diversity compared to MD diet alone (Control, mean ± SEM shown, **P < 0.01, ****P < 0.0001, pairwise t-tests with Bonferroni correction for postantibiotic treatment timepoints). Baseline alpha diversity measurements are during MAC⁺ diet prior to clindamycin treatment. b Unweighted UniFrac distance (compared to preantibiotic MAC⁺ baseline) reveals that PMG supplementation to MD diet (+PMG) leads to a faster trajectory back to baseline community than MD diet alone (Control, ****P < 0.0001, pairwise t-tests with Bonferroni correction). c Mean relative abundance (%) of the phylum Proteobacteria is suppressed in mice treated with 1% PMGs compared to the MD control. Bars are colored by the genus. Total Proteobacteria abundance is significantly higher days 7 and 14 postantibiotic treatment in the control group. (*P < 0.05, pairwise t-tests with Bonferroni correction). d Relative abundance of Cd is suppressed with PMG supplementation compared to MD diet alone (*P < 0.05, **P < 0.01, pairwise t-tests with Bonferroni correction).

Fig. 4. PMGs mitigate the effect of high-fat diet on host physiology and the gut microbiota.

a Weight gain in individual mice fed HFD (n = 5 mice, blue), HFD supplemented continuously with 1% PMGs (n = 5, red), or maintained on MAC⁺ diet (n = 4, red). Weight was compared to baseline day 0 prior to diet switch (mean ± SEM, repeated measures two-way ANOVA with Dunnett’s multiple comparison’s test, asterisk indicates significance versus MAC⁺). b Fat pads as percent total body mass at day 21 (mean ± SEM, ANOVA). c Unweighted UniFrac reveals separation of the three diet groups; PMG supplementation to HFD leads to a unique microbial community from HFD alone. d Changes in the top 100 most abundant taxa over time for mice maintained on MAC⁺, HFD, or HFD supplemented with 1% PMGs. f__ indicates that a strain is not assigned at the family level in the Greengenes database; NA indicates lack of taxonomic assignment at the family level. e Diet-dependent decreases in cecal short-chain fatty acids and organic acids were determined by GC-MS in cecal contents of mice fed MAC⁺, HFD, or HFD + 1% PMG. f PMGs administered twice for 7-day durations in water (1% w/v) are sufficient to reduce host fat accumulation due to HFD. Fat pads as percent total body mass at day 35 (mean ± SEM, n = 5 mice/group). For a, b, e, f: *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

Fig. 5. Prevalent gut commensals harbor high numbers of mucin-degrading carbohydrate gene clusters.

a Numbers of putative mucin-degrading CGCs per genome in 839 taxa that met criteria for at least 1 mucin-degrading CGC and 10% prevalence in healthy humans. A healthy cohort from the Human Microbiome Project (HMP) dataset is used to define prevalence. Stars indicate the pathogenic strains queried that contained putative mucin CGCs. b The top 50 taxa within the HGM database harboring the highest number of mucin-degrading CGCs within their genomes. Total putative mucin-targeting glycoside hydrolases identified within these candidate mucin CGCs are indicated (GH sum).