A novel gene cluster allows preferential utilization of fucosylated milk oligosaccharides in Bifidobacterium longum subsp. longum SC596

Abstract

The infant intestinal microbiota is often colonized by two subspecies of Bifidobacterium longum: subsp. infantis (B. infantis) and subsp. longum (B. longum). Competitive growth of B. infantis in the neonate intestine has been linked to the utilization of human milk oligosaccharides (HMO). However, little is known how B. longum consumes HMO. In this study, infant-borne B. longum strains exhibited varying HMO growth phenotypes. While all strains efficiently utilized lacto-N-tetraose, certain strains additionally metabolized fucosylated HMO. B. longum SC596 grew vigorously on HMO, and glycoprofiling revealed a preference for consumption of fucosylated HMO. Transcriptomes of SC596 during early-stage growth on HMO were more similar to growth on fucosyllactose, transiting later to a pattern similar to growth on neutral HMO. B. longum SC596 contains a novel gene cluster devoted to the utilization of fucosylated HMO, including genes for import of fucosylated molecules, fucose metabolism and two α-fucosidases. This cluster showed a modular induction during early growth on HMO and fucosyllactose. This work clarifies the genomic and physiological variation of infant-borne B. longum to HMO consumption, which resembles B. infantis. The capability to preferentially consume fucosylated HMO suggests a competitive advantage for these unique B. longum strains in the breast-fed infant gut.

Figure 1

Growth of B. longum isolates on mMRS medium supplemented with 2% 2FL (A) or 3FL (B). Growth was measured as OD of the media at 600 nm. Fermentations were carried out in triplicate; controls consisted of inoculated medium lacking of substrates and un-inoculated medium containing a substrate which was also used as blank for OD measurements. Bi: B. infantis, Bal: B. animalis.

Figure 2. Glycoprofiling of the HMO consumption by select B. longum strains.

(A) Total utilization of HMO; (B) Total fucosylated HMO consumption. Bacteria were incubated with pooled HMO from breast milk and consumption was calculated as the percent difference in HMO between the start and the end of exponential phase. (C) Temporal glycoprofile of the consumption of select neutral and acidic HMO by B. longum SC596 at different phases during the exponential phase. Error bars represent experiments in triplicate.

Figure 3. Glycoprofiling HMO consumption on select B. longumw strains.

(A) Percentage of utilization of neutral HMO; (B) Percentage of utilization of fucosylated and sialylated HMO. Bacteria were incubated with pooled HMO from breast milk and consumption was calculated as the percent difference in HMO between the start and the end of exponential phase. Error bars represent experiments in triplicate.

Figure 4. Global transcriptome distances in substrate responses of B. longum SC596 to human milk oligosaccharides.

The heatmap shows the distances between the whole transcriptomes in response to LNT, LNnT, lactose (LAC), 2FL, 3FL, and four time points during HMO growth (early, mid1, mid2, late). Each experiment on each substrate was done in duplicate (i.e. samples ‘A’ and ‘B’). Branches represent the Euclidean distances of the whole transcriptomes after variance-stabilization of the count data. The intensity of the blue color in the heatmap indicates the degree of similarity, from white (dissimilar) to dark blue (most similar). Likewise, the dendograms also show the distances with larger branch lengths indicating greater distances.

Figure 5. Hierarchical clustering of features of carbohydrate metabolism genes induced inB.

longum SC596 during growth on HMO used in this study (See Table S7). Heat map shows intensity of these genes in response to LNT, LNnT, lactose (LAC), 2FL, 3FL, and four time points during HMO growth (early, mid1, mid2, late). Each experiment on each substrate was done in duplicate (i.e. samples ‘A’ and ‘B’). Clustering is calculated according to RPKM values for each gene.

Figure 6. Heat map of the digestion of individually-quantified oligosaccharides from pooled HMO, by β-galactosidases (BLNG00015 and BLNG01753) and α-fucosidases (BLNG01263 and BLNG01264) purified from B. longum SC596).

Enzymes were incubated with pooled HMO and products were quantified using LC-MS. The percentage digestion calculated for each of the specific oligosaccharide degraded by these enzymes is expressed via a heat map. HMO names and structures were reported in ref. . T: Terminal linkage; endo: internal linkage.

Figure 7. FHMO utilization cluster in B. longum SC596, and homologous genes in bifidobacteria.

Arrows represent genes, and numbers on top of each gene indicate the locus tag number in the respective genome. Text inside the arrows indicates predicted or validated annotations. Red numbers indicate percent identity between corresponding genes and homologs relative to strain SC596. Colors of each gene are indicative of the primary function of each respective gene: transcriptional regulators (black), oligosaccharide transport (blue), carbohydrate feeder pathways (green) and glycosyl hydrolases (red). SBP: Solute Binding Protein; L-Fuc DH: L-fuconate dehydrogenase; DHPS: Dihydropicolinate synthase; FucU: L-fucose mutarotase.