A key genetic factor for fucosyllactose utilization affects infant gut microbiota development

Abstract

Recent studies have demonstrated that gut microbiota development influences infants' health and subsequent host physiology. However, the factors shaping the development of the microbiota remain poorly understood, and the mechanisms through which these factors affect gut metabolite profiles have not been extensively investigated. Here we analyse gut microbiota development of 27 infants during the first month of life. We find three distinct clusters that transition towards Bifidobacteriaceae-dominant microbiota. We observe considerable differences in human milk oligosaccharide utilization among infant bifidobacteria. Colonization of fucosyllactose (FL)-utilizing bifidobacteria is associated with altered metabolite profiles and microbiota compositions, which have been previously shown to affect infant health. Genome analysis of infants' bifidobacteria reveals an ABC transporter as a key genetic factor for FL utilization. Thus, the ability of bifidobacteria to utilize FL and the presence of FL in breast milk may affect the development of the gut microbiota in infants, and might ultimately have therapeutic implications.

Figure 1. Infant gut microbiota community profiles during the first month of life.

(a) Microbiota profiles in stool samples from 12 subjects (n=202; ∼17 sampling days per subject), temporally ordered from left to right. Each row represents taxonomic groups at the family level. The top 15 families are displayed and sorted according to relative abundance. Abundances are represented using the colour scale. (b) Characteristics of infant gut microbiota, illustrated by PCoA and PAM clustering analyses. Data from individuals (points) were clustered, and the centres of gravity (rectangles) were computed for each class. The coloured ellipses encompass 67% of the samples in each cluster. (c) Box plots showing the relative abundances of the main contributors to each cluster. Different letters (a–c) above the boxes indicate significant differences between clusters (P<0.05, Mann–Whitney U-test with Bonferroni's correction). (d) Temporal shift from Staphylococcaceae- or Enterobacteriaceae-dominant microbiota to Bifidobacteriaceae-dominant microbiota. S, Staphylococcaceae-dominated (yellow); E, Enterobacteriaceae-dominated (blue); B, Bifidobacteriaceae-dominated (red); NT, not tested; —, sample not provided.

Figure 2. Gut microbiota community profiles of 27 1-month-old infants and 22 adults.

(a) Bacterial families representing more than 1% (on average) of the microbiota in infants or adults are shown on the colour scale. Samples were hierarchically clustered by measuring Euclidean distances with complete-linkage clustering, as shown in the upper tree. (b) Characteristics of infant and adult gut microbiota, as illustrated by PCoA and PAM clustering analyses. Cluster B, Bifidobacteriaceae-predominant; cluster E, Enterobacteriaceae-predominant; cluster AD, adult-type microbiota. (c) Abundances of the main contributors to each cluster. Different letters (a–c) indicate significant differences between clusters (P<0.05, Mann–Whitney U-test with Bonferroni's correction). Differences in other bacterial families are shown in Supplementary Fig. 7). (d) Network diagram showing co-occurrence relationships among the main contributors in infants and adults. Node sizes indicate the abundances of each bacterial family, and the widths of the edges reflect the calculated Spearman's rank correlation coefficient.

Figure 3. Relationships between bacterial family abundances and gut environments.

(a) Spearman's rank correlation coefficients between bacterial family abundances and gut environmental factors such as pH, organic acid concentrations and faecal oligosaccharide concentrations are shown in numerical and colour-scale formats. (b) Spearman's rank correlations of oligosaccharide concentrations with pH values and acetate concentrations. (c) Relationships between bacterial abundances, faecal oligosaccharide concentrations and the relative abundances of bifidobacterial species. The upper tree shows hierarchical clustering on the basis of the bacterial family compositions. (d) Growth curves of 29 bifidobacterial strains in medium containing HMOs (see Supplementary Fig. 14 for more details). (e) Glycoprofiles of bacterial supernatants after 40 h of cultivation. Samples are ordered based on their OD₆₀₀ values after 40 h of cultivation.

Figure 4. Identification of bifidobacterial genes responsible for HMO utilization.

(a) Syntenic relationships of putative FL-utilization gene clusters identified among the bifidobacteria. The genes and their orientations are depicted with arrows. Strains lacking fucosidase genes (10 strains among 29) are not represented. (b) Growth curves of the B. breve BR-A29 strain and the corresponding FL-SBP gene knockout strain in medium containing HMOs. (c) Glycoprofiles of BR-A29 and FL-SBP gene-knockout strains after 40 h of cultivation.

Figure 5. Impact of FL-utilizing bifidobacteria on gut microbial ecosystems.

Box plots showing differences among Bifidobacteria-dominant microbiota, with or without FL-utilizing bifidobacteria (clusters B1 and B2, respectively) and Enterobacteriaceae-dominant microbiota (cluster E). Different letters (a–c) above the boxes indicate significant differences between clusters (P<0.05, Mann–Whitney U-test with Bonferroni's correction).

Figure 6. Infant microbiota development and molecular mechanisms of FL utilization by bifidobacteria.

Infant gut microbiota showed 3 distinct clusters, which underwent directional transition to Bifidobacteriaceae-dominant microbiota, and displayed individual variations in the pace of progression. Isolated bifidobacterial strain showed differences in FL utilization. Colonization of FL-utilizing bifidobacteria are associated with altered gut acetate concentrations, pH, and Enterobacteriaceae and Bifidobacteriaceae abundances in the cohort, which have been previously shown to affect infant health. We subsequently identified a SBP of the multiple-sugar ABC transporter system is a key genetic factor of FL utilization.