Butyrate Levels in the Transition from an Infant- to an Adult-Like Gut Microbiota Correlate with Bacterial Networks Associated with Eubacterium Rectale and Ruminococcus Gnavus

Abstract

Relatively little is known about the ecological forces shaping the gut microbiota composition during infancy. Therefore, the objective of the present study was to identify the nutrient utilization- and short-chain fatty acid (SCFA) production potential of gut microbes in infants during the first year of life. Stool samples were obtained from mothers at 18 weeks of pregnancy and from infants at birth (first stool) at 3, 6, and 12-months of age from the general population-based PreventADALL cohort. We identified the taxonomic and SCFA composition in 100 mother-child pairs. The SCFA production and substrate utilization potential of gut microbes were observed by multiomics (shotgun sequencing and proteomics) on six infants. We found a four-fold increase in relative butyrate levels from 6 to 12 months of infant age. The increase was correlated to Eubacterium rectale and its bacterial network, and Faecalibacterium prausnitzii relative abundance, while low butyrate at 12 months was correlated to Ruminococcus gnavus and its associated network of bacteria. Both E. rectale and F. prausnitzii expressed enzymes needed for butyrate production and enzymes related to dietary fiber degradation, while R. gnavus expressed mucus-, fucose, and human milk oligosaccharides (HMO)-related degradation enzymes. Therefore, we believe that the presence of E. rectale, its network, and F. prausnitzii are key bacteria in the transition from an infant- to an adult-like gut microbiota with respect to butyrate production. Our results indicate that the transition from an infant- to an adult-like gut microbiota with respect to butyrate producing bacteria, occurs between 6 and 12 months of infant age. The bacteria associated with the increased butyrate ratio/levels were E. rectale and F. prausnitzii, which potentially utilize a variety of dietary fibers based on the glycoside hydrolases (GHs) expressed. R. gnavus with a negative association to butyrate potentially utilizes mucin, fucose, and HMO components. This knowledge could have future importance in understanding how microbial metabolites can impact infant health and development.

Keywords: gut microbiota, infant, short-chain fatty acids, metaproteomics.

Figure 1

Flowchart of the study. The figure represents the workflow of the project. Sampling was performed by the Oslo University Hospital/University of Oslo, Østfold Hospital Trust and Karolinska Institutet, Stockholm [16]. In this study, we analyzed fecal samples from children in four different age groups (newborn, 3 months, 6 months, and 12 months) and their respective mothers, using a multiomics approach, which included 16S rRNA sequencing, short-chain fatty acids, shotgun sequencing, and metaproteomics. The stippled line represents the number (n) of fecal samples from infants analyzed at the different age groups at gas chromatography for short-chain fatty acid composition, and the full line represents the number of fecal samples from each age group that were analyzed by 16S rRNA after rarefaction and filtering for poor quality sequences.

Figure 2

Taxonomic and short-chain fatty acid (SCFA) composition. The bar chart shows the relative abundance (%) of bacterial orders acquired from sequencing processed by the QIIME pipeline (A) and SCFAs composition (B) for the respective age groups. The dominant orders of bacteria and SCFAs in their respective colors are displayed on the top right. The asterisks (*) represent a p-value < 0.05, determined by Kruskal–Wallis–Dunn’s test, FDR correct by the Benjamini–Hochberg method. The exact p-values are shown Supplementary Tables S1 and S2. SCFAs illustrated represent percent based on average total SCFAs detected. SCFAs included in the “other” group are isobutyrate, isovalerate, and valerate.

Figure 3

Relative abundance of E. rectale, R. gnavus, and Faecalibacterium, at 12 months. The dot plot shows the relative abundance (Log10 of percentage abundance) and median of the E. rectale group, R. gnavus group, and Faecalibacterium, at 12 months of age for all infants. The asterisk (*) represents a significant difference in relative abundance (p < 0.05, Wilcoxon rank sum test, FDR correct by Benjamini–Hochberg method) between the bacteria.

Figure 4

Metadata association with microbiota. ANOVA-simultaneous component analysis (ASCA-ANOVA) was used to determine the association of microbiota to known factors of the children. The Principal Component Analysis (PCA) plot shows the effect (Y-axis) of delivery method: vaginal and C-section (A), breastfeeding between 3 and 6 months of age (B), and age (C) based on the taxonomic groups acquired from 16S rRNA sequencing (X-axis). (C) The Y-axis is a gradient of age, where the effect points towards a young age or old age, showing the outer points of the scale: newborns and mother.

Figure 5

Bacterial and SCFA correlations at 12 months. The illustration shows all OTUs from 16S rRNA represented as nodes, with color indicating their correlation to SCFAs; blue = no correlation, red = negative correlation to butyrate, green = positive correlation to butyrate, and black = positive correlation to propionate. The three different node sizes represent the general abundance of the respective bacteria. The thickness of the lines between nodes represents a correlation between the bacteria, of which a thick line is a strong correlation. Blue lines indicate a positive correlation between the bacteria, while brown lines indicate a negative correlation. Prominent nodes in the networks are highlighted with their respective OTU taxonomy. The highlighted green circle with positive association to butyrate but without correlations to other bacteria was assigned to Faecalibacterium.

Figure 6

Protein expression. (A) shows a Volcano-plot highlighting (green) proteins differentially expressed between children with E. rectale and the R. gnavus community. The volcano plot was created using the Perseus software. Proteins marked as black circles represent significant differential expression in proteins related to the butyrate pathway with their respective Enzyme Commission (E.C.) numbers. (B) shows the Log2 Label-Free Quantification (LFQ) intensity of the proteins, with missing values imputed with the constant 10. The intensity is represented as a gradient from light yellow (not detected) to red (highly abundant). Clustering is shown on the left, including four clusters, expressed in children with the R. gnavus network, E. rectale network, scattered, or expressed in all. A dendrogram is included at the top, to show the similarities and dissimilarities between the technical replicates and the children with E. rectale- or R. gnavus-dominated communities, with their respective technical duplicate (e.g., “R. gnavus 1.2” annotates child number one with the R. gnavus network and technical duplicate two). E. rectale infant 1 did not have any technical duplicates.

Figure 7

Protein presence related to butyrate production. The figure gives an overview of bacterial proteins (E.C. number) in relation to the butyrate pathway butyryl-CoA:Acetate CoA-transferase. The figure shows proteins detected (green box) in infants with the E. rectale network (A), or R. gnavus network (B). Bacterial taxonomy is shown next to each E.C. number, representing the bacterial source of the given protein. The bacterial sources are divided by two different colors, where orange represents detection in three or fewer samples in (A) or two or fewer in (B), and green represents detection in four or more in (A) or three or more in (B).

Figure 8

Expressed glycoside hydrolases/carbohydrate esterases. The figure shows proteins expressed within the glycoside hydrolase and carbohydrate esterase groups expressed based on the contiguous sequences assembled from shotgun sequencing and protein expression derived from nanoLC-Orbitrap tandem mass spectrometry (MS/MS). The abundance represents the number of unique proteins expressed from a given taxa within the glycoside hydrolases (GH) or carbohydrate esterase (CE) group. The y-axis shows the relevant taxonomies, and the dendrogram represents the Euclidian distance between the taxonomic groups based on GH and CE expression. The plot was created using ggplot2 [18]. GH numbers related to mucus and fucose degradation are marked in green, and degradation of starch, glycogen, and hemicellulose are marked in blue.