Research ArticleMICROBIOLOGY

Evolutionary adaptation in fucosyllactose uptake systems supports bifidobacteria-infant symbiosis

See allHide authors and affiliations

Science Advances  28 Aug 2019:
Vol. 5, no. 8, eaaw7696
DOI: 10.1126/sciadv.aaw7696
  • Fig. 1 A platform for evaluating in vivo substrate specificity of HMO transporters.

    Specificity of the two candidate HMO transporters from B. infantis was analyzed by their heterologous expression in the pMSK65-harboring B. longum 105-A ΔlnbX ΔgltA, the strain expressing intracellular HMO-degrading enzymes but deficient in growth on purified HMO mixture. (A) E. coliBifidobacterium shuttle vector pMSK65 carrying the genes encoding intracellular HMO-degrading exo-glycosidases from B. infantis. (B) B. longum 105-A ΔlnbX ΔgltA strain with pMSK65 as a host for evaluating in vivo specificity of HMO transporters. The candidate gene cluster (shown in red) is introduced into the host using a compatible plasmid. (C) Growth of the recombinant B. longum strain with an empty vector (open circles) or with the plasmid expressing FL transporter-1 (locus tags Blon_0341–0343, blue squares) or FL transporter-2 (locus tags Blon_2202–2204, green squares) in medium supplemented with 0.5% (w/v) purified HMO mixture or Lac. Experiments were technical duplicates, and data are means ± SD. OD600, optical density at 600 nm. (D) Gene organization of candidate FL transporter paralogs from B. infantis JCM 1222T analyzed in this study. TMDs, transmembrane domains; aa, amino acid; bp, base pair. (E) Changes in the concentration of each HMO in culture supernatant during growth on 0.5% (w/v) purified HMO mixture. Experiments were technical triplicates, and data are means ± SD. The symbols are the same as in (C). Sugars are depicted according to the nomenclature committee of the Consortium for Functional Glycomics (www.functionalglycomics.org/static/consortium/consortium.shtml).

  • Fig. 2 Two FL transporters differently contribute to in vitro growth of B. infantis JCM 1222T on major fucosylated HMOs.

    (A) Schematic representation of the gene organization at the genomic loci for the two FL transporters in B. infantis JCM 1222T variants. (B) Growth of WT and mutant B. infantis strains on 0.5% (w/v) Gal, 2′-FL, 3-FL, LDFT, LNFP I, or purified HMO mixture. The symbols used are as follows: open circles (WT), blue triangles [FL transporter-1 disruptant (ΔFL1)], green triangles (ΔFL2), orange diamonds, [FL2-BP insertional mutant (FL2-BPINS)], and gray triangles [double mutant (ΔFL1 FL2-BPINS)]. Experiments were biological duplicates, and the data are means ± SD. Sugars are depicted as in Fig. 1. (C) In vitro growth competition assay between WT and double-mutant (ΔFL1 FL2-BPINS) strains on 0.5% (w/v) GOSs, 2′-FL, and purified HMO mixture. The population of each strain was determined by quantitative polymerase chain reaction (qPCR), and the relative abundance (%) of WT (white bar) and double mutant (gray bar) is shown. Experiments were biological triplicated, and the data are means ± SD. Student’s two-tailed t test was used for evaluating statistical significance.

  • Fig. 3 Structure of 2’-FL- or 3-FL-complexed SBP of FL transporter-2 (FL2-BP) and distribution of FL-BP homologs among gut microbes.

    (A) Cartoon model of the overall structure of FL2-BP complexed with 2′-FL. Domain 1 and domain 2 are in brown and green, respectively. The sugars are depicted in stick models with carbon atoms of Fuc in white, Gal in yellow, Glc in blue, and oxygen atoms in red. (B and C) Close-up view of the ligand binding site of FL2-BP in complex with 2′-FL (B) and 3-FL (C). The same colors are used as in (A). (D) Distribution of FL1-BP and FL2-BP homologs (>61% identity) among Bifidobacterium species/subspecies. FL1-BP and FL2-BP from B. infantis JCM 1222T were used as queries for BlastP analysis. The cluster numbers (I to IV) are indicated based on the phylogenetic analysis shown in fig. S6C. The sequences of B. longum without subspecies identification in the National Center for Biotechnology Information (NCBI) database were excluded from the analysis. No close homologs (≥60% identity) were found in the genomes of other bacterial taxa in the NCBI database.

  • Fig. 4 The abundance of Bifidobacterium in breast-fed infant guts is associated with FL transporter genes.

    The results presented here were obtained by analyzing Japanese subjects recruited in this study (A to D) and by data mining of a deposited metagenome dataset (MG-RAST, accession no. qiime:621) (3) (E to G). (A) Relative abundance (%) of genus Bifidobacterium and two FL-BP homologs in stool DNA of breast-fed infants (n = 36) and adults (n = 31). The copy number of the genes attributable to Bifidobacterium 16S ribosomal RNA (rRNA), FL1-BP, and FL2-BP, which was determined by qPCR analysis, was divided by the copies of the 16S rRNA gene attributable to total bacteria. Mann-Whitney U test was used for evaluating the statistical significance. The data are shown by box plot, in which the middle bar indicates the median, while the top and bottom of the box indicate the third and first quartiles, respectively. Whiskers represent the lowest and highest values within 1.5 times the interquartile range from the first and third quartiles, respectively. (B) Spearman’s rank correlation coefficient analysis between the relative abundances of genus Bifidobacterium and either (left and middle panels) or both (right panel) of FL-BP genes in the breast-fed infant (orange circles) and adult (purple squares) groups. The data obtained in (A) were used for the analysis. (C) The concentrations of 2′-FL, 3-FL, LDFT, and LNFP I (substrates for FL transporter-1 and -2) in breast milk and infant stools (32 mother-infant pairs) were compared between FL-BP gene–detected (positive, n = 25) and undetected (negative, n = 7) groups. Mann-Whitney U test was used for evaluating the statistical significance. (D) Spearman’s rank correlation coefficient analysis between the relative abundance of the FL1-BP or FL2-BP gene and the fecal concentration of each substrate HMO. The data obtained in (C) were used for the analysis, and the results are shown as a heat map. *P < 0.05 and **P < 0.01. Sugars are depicted as in Fig. 1. (E and F) The deposited metagenome data of 83 individuals from the United States (n = 50), Malawi (n = 18), and Venezuela (n = 15) were used for the analysis (132,556 ± 76,952 reads per sample) (3). The abundances (%) of genus Bifidobacterium and two FL-BP homologs in the fecal samples obtained from breast-fed (n = 34) and formula-fed (n = 27) infants (≤1 year old) and adults (n = 22, ≥18 years old) are shown in (E), and those in the fecal samples of breast-fed infants living in the United States (n = 10), Malawi (n = 14), and Venezuela (n = 10) are shown in (F). See Materials and Methods for read count and abundance calculation for FL-BP genes. Different letters (a, b, and c) indicate statistically significant differences among the three groups (P < 0.05, Mann-Whitney U test with Bonferroni correction). (G) Spearman’s rank correlation coefficient analysis between the relative abundances of genus Bifidobacterium and either or both of two FL-BP genes in the breast-fed infant group. The data obtained in (F) were used for the analysis, and the results are shown as a heat map. *P < 0.05, **P < 0.01, and ***P < 0.001.

  • Fig. 5 Schematic model showing selective growth advantage of bifidobacteria with FL transporter-1 and -2 in the breast-fed infant guts.

    FL transporters and HMO-degrading enzymes are shown with their locus tag numbers of B. infantis ATCC 15697T (GenBank accession no. CP001095.1). Sugars are depicted as in Fig. 1.

  • Table 1 Binding parameters of FL1-BP and FL2-BP to mono- and difucosyllactose determined by ITC and SPR analysis.

    Binding parameters of FL1-BP and FL2-BP to mono- and difucosyllactose determined by ITC and SPR analysis..

    ProteinLigandITC*SPR
    Ka
    (×105 M−1)
    Kd
    (μM)
    ΔG0
    (kcal mol−1)
    ΔH
    (kcal mol−1)
    TΔS0
    (kcal mol−1)
    ΔS0
    (cal K−1 mol−1)
    n
    (site)
    Kd
    (μM)†
    Rmaxχ 2 §
    FL1-BP2′-FL1.05 ± 0.239.75−6.8−23.6 ± 0.516.8−56.30.85 ± 0.149.0 ± 0.714.70.09
    FL2-BP2′-FL1.86 ± 0.125.40−7.2−14.2 ± 0.77.0−23.60.90 ± 0.049.9 ± 1.69.30.13
    3-FL1.69 ± 0.115.95−7.1−14.0 ± 0.46.9−23.21.12 ± 0.145.9 ± 0.510.20.06
    LDFT0.05 ± 0.00191.11−5.1−6.8 ± 0.51.7−5.71.02 ± 0.02163.1 ± 25.012.00.24

    *Values of association constants (Ka), enthalpy of binding (ΔH), and binding stoichiometry (n) obtained in ITC analysis are expressed as means ± SD of duplicate experiments. Dissociation constants (Kd) were calculated from the reciprocal of Ka. The Gibbs free energy change (ΔG0) and the entropy change (ΔS0) were calculated from the equations ΔG0 = −RTlnKa and TΔS0 = ΔH − ΔG0, respectively (R, gas constant; T, absolute temperature).

    †The Kd values obtained in the SPR analysis are means ± SD of duplicate experiments.

    ‡The maximum binding level from the fits to a one-site binding model.

    §Statistical goodness of the fit to a one-site binding model.

    Supplementary Materials

    • Supplementary material for this article is available at http://advances.sciencemag.org/cgi/content/full/5/8/eaaw7696/DC1

      Fig. S1. Representative HMO molecules analyzed in this study.

      Fig. S2. Schematic representation of two different gene disruption methods used in this study.

      Fig. S3. Evaluation of the HMO transporter–heterologous expression system.

      Fig. S4. ITC and SPR analyses of SBPs of FL transporter-1 and -2.

      Fig. S5. Structural basis of the dual recognition of 2′-FL and 3-FL by FL2-BP from B. infantis JCM 1222T.

      Fig. S6. Phylogenetic and structural analyses revealed the signature sequences that dictate specificity difference between FL1-BP and FL2-BP homologs.

      Fig. S7. HMO consumption analysis using samples collected from 32 mother-infant pairs.

      Table S1. Data collection and refinement statistics for the SBP of FL transporter-2 from B. infantis JCM 1222T (FL2-BP).

      Table S2. Polar contacts between FL2-BP and FL molecules.

      Table S3. Primers and probes used in this study.

    • Supplementary Materials

      This PDF file includes:

      • Fig. S1. Representative HMO molecules analyzed in this study.
      • Fig. S2. Schematic representation of two different gene disruption methods used in this study.
      • Fig. S3. Evaluation of the HMO transporter–heterologous expression system.
      • Fig. S4. ITC and SPR analyses of SBPs of FL transporter-1 and -2.
      • Fig. S5. Structural basis of the dual recognition of 2′-FL and 3-FL by FL2-BP from B. infantis JCM 1222T.
      • Fig. S6. Phylogenetic and structural analyses revealed the signature sequences that dictate specificity difference between FL1-BP and FL2-BP homologs.
      • Fig. S7. HMO consumption analysis using samples collected from 32 mother-infant pairs.
      • Table S1. Data collection and refinement statistics for the SBP of FL transporter-2 from B. infantis JCM 1222T (FL2-BP).
      • Table S2. Polar contacts between FL2-BP and FL molecules.
      • Table S3. Primers and probes used in this study.

      Download PDF

      Files in this Data Supplement:

    Stay Connected to Science Advances

    Navigate This Article