Key bacterial taxa and metabolic pathways affecting gut short-chain fatty acid profiles in early life

Abstract

Infant gut microbiota development affects the host physiology throughout life, and short-chain fatty acids (SCFAs) are promising key metabolites mediating microbiota-host relationships. Here, we investigated dense longitudinally collected faecal samples from 12 subjects during the first 2 years (n = 1048) to identify early life gut SCFA patterns and their relationships with the microbiota. Our results revealed three distinct phases of progression in the SCFA profiles: early phase characterised by low acetate and high succinate, middle-phase characterised by high lactate and formate and late-phase characterised by high propionate and butyrate. Assessment of the SCFA-microbiota relationships revealed that faecal butyrate is associated with increased Clostridiales and breastfeeding cessation, and that diverse and personalised assemblage of Clostridiales species possessing the acetyl-CoA pathway play major roles in gut butyrate production. We also found an association between gut formate and some infant-type bifidobacterial species, and that human milk oligosaccharides (HMO)-derived fucose is the substrate for formate production during breastfeeding. We identified genes upregulated in fucose and fucosylated HMO utilisation in infant-type bifidobacteria. Notably, bifidobacteria showed interspecific and intraspecific variation in the gene repertoires, and cross-feeding of fucose contributed to gut formate production. This study provides an insight into early life SCFA-microbiota relationships, which is an important step for developing strategies for modulating lifelong health.

Fig. 1. Infant gut microbiota community profiles during the first 2 years of life.

Microbiota profiles of faecal samples from 12 infants (n = 1048; up to 92 samples per subject) were investigated. A Order-level dynamics of three infants (see Fig. S2A for all subjects). Vertical bars along the x-axis indicate every 2 months; dots along the x-axis indicate every week until 1 month. B Temporal shift of bacterial abundance in 12 infants at the order level. C Temporal shift in α-diversity. D Characteristics of infant gut microbiota, illustrated using JSD with PCoA and PAM clustering analyses. E Box plots showing relative abundances of the main contributors to each cluster. F Temporal shift of microbiota clusters. Light blue, Enterobacterales dominant; red, Bifidobacteriales dominant; yellow, Clostridiales dominant; grey, not tested. White and black arrowheads indicate the initiation of solid food and cessation of breastfeeding, respectively. Subjects were ordered with a stable colonisation of Bifidobacteriales.

Fig. 2. Infant gut SCFA profiles during the first 2 years.

A Faecal SCFA and pH dynamics of three infants (see Fig. S3A for all subjects). B Temporal shift of faecal SCFA concentrations and pH. C Characteristics of infant gut SCFA profiles, illustrated using JSD with PCoA and PAM clustering analyses. D Box plots showing faecal SCFA concentration of the main contributors to each cluster. E Temporal shift of the SCFA clusters. Blue, acetate declined and succinate elevated (type 1); orange, lactate and formate elevated (type 2); green, propionate and butyrate elevated (type 3); grey, not tested. White and black arrowheads indicate the initiation of solid food and cessation of breastfeeding, respectively. The subjects were ordered as in Fig. 1F.

Fig. 3. The relationship between infant microbiota composition and gut SCFA profiles.

A Relationship between three microbiota clusters and three SCFA patterns. B Microbiota composition (ordered from cluster En, Bi to Cl) and SCFA profiles (ordered from types 1 to 3) are shown. C Personalised association between gut SCFA profile and microbiota at the order- level (see Fig. S4 for the other SCFAs). Association with the top 6 bacterial lineages (average abundance > 2%) is presented. Numbers represent r values (Spearman’s correlation). Underlined numbers indicate FDR-corrected p values < 0.01.

Fig. 4. Clostridiales members contribute to gut butyrate production.

A Plot of Clostridiales abundance and butyrate concentrations with respect to age. Six representative subjects are presented (see Fig. S5A for all subjects). Blue and orange dotted vertical lines represent the initiation of solid food and cessation of breastfeeding, respectively. B Pathway for butyrate production [35]. C Age-dependent heatmap of Clostridiales phylotypes and their potent butyrate production. The top 40 Clostridiales phylotypes are shown. Left tree represents the taxonomic relationship based on 16S rRNA sequences. Right figure is a summary of the presence of genes for butyrate production (see Fig. S6B). Phylotypes mentioned in the text are highlighted in bold.

Fig. 5. Infant bifidobacteria contribute gut lactate and formate production.

A–C Bifidobacterium abundance and concentration of lactate and formate with respect to age. Six infants whose lactate and formate production were driven by B. infantis or B. bifidum are presented (see Fig. S7 for all subjects). D Volcano plots of transcriptional data during utilisation of fucose and FL compared to lactose. E List of upregulated genes and their locus tags (INF29_xxxxx are abbreviated), KEGG orthology (KO), annotation and transcripts per million (TPM). F Predicted metabolic pathway from fucosyllactose and fucose to formate. G Organisation of fucose utilisation genes and their phenotypes (see Fig. S10A,B for the other strains). Numbers below the arrow represent the last digits of the locus tags.

Fig. 6. Fucosylated HMO utilisation and production of formate by the bifidobacterial community.

A Proposed cross-feeding model for FL metabolism by B. bifidum and B. breve to produce SCFA. B Formate production by B. infantis, B. breve, B. bifidum and a combination of these species. C Relationship between key marker genes for fucosylated HMO utilisation and concentration of lactate and formate (see Fig. S11 for all subjects). D Samples were divided into five subgroups based on feeding, Bifidobacteriales-colonisation and presence of key genes for fucosylated HMO utilisation. E Difference in faecal SCFA concentration among the subgroups. Digits in parenthesis represent the number of samples assigned to the subgroup.