Determinants of raffinose family oligosaccharide use in Bacteroides species

Abstract

Bacteroides species are successful colonizers of the human colon and can utilize a wide variety of complex polysaccharides and oligosaccharides that are indigestible by the host. To do this, they use enzymes encoded in polysaccharide utilization loci (PULs). While recent work has uncovered the PULs required for the use of some polysaccharides, how Bacteroides utilize smaller oligosaccharides is less well studied. Raffinose family oligosaccharides (RFOs) are abundant in plants, especially legumes, and consist of variable units of galactose linked by α-1,6 bonds to a sucrose (glucose α-1-β-2 fructose) moiety. Previous work showed that an α-galactosidase, BT1871, is required for RFO utilization in Bacteroides thetaiotaomicron. Here, we identify two different types of mutations that increase BT1871 mRNA levels and improve B. thetaiotaomicron growth on RFOs. First, a novel spontaneous duplication of BT1872 and BT1871 places these genes under the control of a ribosomal promoter, driving high BT1871 transcription. Second, nonsense mutations in a gene encoding the PUL24 anti-sigma factor likewise increase BT1871 transcription. We then show that hydrolases from PUL22 work together with BT1871 to break down the sucrose moiety of RFOs and determine that the master regulator of carbohydrate utilization (BT4338) plays a role in RFO utilization in B. thetaiotaomicron. Examining the genomes of other Bacteroides species, we found homologs of BT1871 in a subset and showed that representative strains of species with a BT1871 homolog grew better on melibiose than species that lack a BT1871 homolog. Altogether, our findings shed light on how an important gut commensal utilizes an abundant dietary oligosaccharide.

Importance: The gut microbiome is important in health and disease. The diverse and densely populated environment of the gut makes competition for resources fierce. Hence, it is important to study the strategies employed by microbes for resource usage. Raffinose family oligosaccharides are abundant in plants and are a major source of nutrition for the microbiota in the colon since they remain undigested by the host. Here, we study how the model commensal organism, Bacteroides thetaiotaomicron utilizes raffinose family oligosaccharides. This work highlights how an important member of the microbiota uses an abundant dietary resource.

Keywords: BT1871; Bacteroides; melibiose; polysaccharide utilization loci; raffinose family oligosaccharide; α-galactosidase.

Fig 1

Duplication of the BT1871-BT1872 locus provides B. thetaiotaomicron a growth advantage on melibiose. (A) Structure of the BT1871-BT1872 (BT1871—PUL24 α-galactosidase, BT1872—PUL24 β-glucosidase) duplication in a dupl+ strain (bottom) compared to a strain without the duplication (dupl−, top). Red gene models represent genes in an IS3 element transposon while green gene models represent genes in a ribosomal RNA operon. (B) qRT-PCR to measure BT1871 mRNA levels in the indicated strains grown to mid-log phase in rich (TYG) media. The bars represent the mean and SD of n = 3 biological replicates. All values are normalized to the level in WT dupl−. Statistical significance was determined by one-way ANOVA. **P < 0.01; ****P < 0.0001. (C) Growth curves of the indicated strains on minimal media with melibiose as the sole carbon source. The points and error bars represent the mean and SD of n = 3 biological replicates.

Fig 2

B. thetaiotaomicron mutants derived from WT dupl− cells grow on AMG. (A) Growth curves of WT dupl+ and WT dupl− strains on glucose, melibiose, and AMG. On glucose, the WT dupl− strain had a mean growth rate of 0.50 h⁻¹ compared to 0.54 h⁻¹ for the isolates (P = 0.49 by two-tailed t-test) (B) Growth curves of WT dupl− cells inoculated as described in the text in 96-well plates containing minimal media with AMG as the sole carbon source (left panel). Subculture of strains to fresh media shows that isolates that began growing between 40 and 60 hours show no lag (right panel). New isolates arise again in some wells. (C) Growth curves of WT dupl+, WT dupl−, and six independent mutants of WT dupl− that can grow on AMG as the sole carbon source. In (A) and (C), points and error bars represent the mean and SD of n = 3 biological replicates. For each growth curve, the sugar used as the sole carbon source is indicated at the top along with the concentration used.

Fig 3

A BT1876 mutant strain grows better on RFOs than the WT strain due to increased BT1871 expression. (A) Genomic organization of B. thetaiotaomicron PUL24. The location of known transcription start sites [derived from reference (63) in the text] is indicated by bent arrows. (B) Growth curves of WT, ΔBT1876 anti-sigma mutant, and a complemented strain. On glucose, the WT dupl− strain had a mean growth rate of 0.51 h⁻¹ compared to 0.52 h⁻¹ for the ΔBT1876 strain (P = 0.75 by two-tailed t-test). On melibiose, the WT dupl− strain had a mean growth rate of 0.29 h⁻¹ compared to 0.29 h⁻¹ for the complemented strain (P = 0.60 by two-tailed t-test). On raffinose, the WT dupl− strain had a mean growth rate of 0.17 h⁻¹ compared to 0.17 h⁻¹ for the complemented strain (P = 0.84 by two-tailed t-test), whereas on stachyose the growth rates and maximal OD were similar (mean growth rate of 0.11 h⁻¹ for the WT dupl− strain compared to 0.11 h⁻¹ for the complemented strain, P = 0.18 by two-tailed t-test; mean maximal OD₆₃₀ of 0.53 for the WT dupl− strain compared to 0.53 for the complemented strain). Points and error bars represent the mean and SD of n = 3 biological replicates. For each growth curve, the sugar used as the sole carbon source is indicated at the top along with the concentration used. (C) qRT-PCR showing relative BT1871 mRNA levels in WT, BT1876 anti-sigma mutant, and the complemented mutant strain. The bars depict the mean and SD of n = 3 biological replicates. All values are normalized to the WT grown on melibiose. Differences between WT and the BT1876 mutant are significant based on two-way ANOVA. *P < 0.05.

Fig 4

PUL22 is important for RFO utilization in B. thetaiotaomicron and its GH32 family sucrases act redundantly to promote RFO utilization. (A) Genomic organization of B. thetaiotaomicron PUL22. (B) Growth of WT, ΔBT1754 (sensor kinase), and ΔBT1758 (putative fructose transporter) on glucose and raffinose. (C) Growth curves of individual GH32 sucrase mutant strains on glucose, sucrose, and raffinose. BT1759, BT1760, and BT1765 are part of PUL22. BT3082 is not part of PUL22 but is coregulated with PUL22 genes in response to fructans. On sucrose, the mean growth rate of the WT strain was 0.55 h⁻¹ compared to 0.54 h⁻¹ for the ΔBT1759 strain, 0.57 h⁻¹ for the ΔBT1760 strain, 0.56 h⁻¹ for the ΔBT1765 strain, and 0.52 h⁻¹ for the ΔBT3082 strain (P > 0.05 for all individual comparisons to the WT strain by two-tailed t-test). On raffinose, the mean growth rate of the WT strain was 0.20 h⁻¹ compared to 0.17 h⁻¹ for the ΔBT1759 strain, 0.16 h⁻¹ for the ΔBT1760 strain, 0.20 h⁻¹ for the ΔBT1765 strain, and 0.18 h⁻¹ for the ΔBT3082 strain (P > 0.05 for all individual comparisons to the WT strain by two-tailed t-test). (D) Growth curves of WT, BT1871 mutant, a quadruple GH32 sucrase mutant (Δsucrase) and the Δsucrase mutant in a BT1871 mutant background (ΔsucraseΔBT1871) on RFOs. On sucrose, the WT strain had a mean growth rate of 0.44 h⁻¹ compared to 0.39 h⁻¹ for the Δsucrase strain (P = 0.14 by two-tailed t-test). On stachyose, the Δsucrase strain had a mean growth rate of 0.14 h⁻¹ compared to 0.25 h⁻¹ for the Δsucrase ΔBT1871 strain (P = 0.14 by two-tailed t-test). Points and error bars represent the mean and SD of n = 3 biological replicates. For each growth curve, the sugar used as the sole carbon source is indicated at the top along with the concentration.

Fig 5

The global regulator BT4338 is important for RFO utilization in B. thetaiotaomicron. (A) qRT-PCR was used to measure relative levels of BT1875 (PUL24 SusC-like transporter) and BT1871 (PUL24 α-galactosidase) mRNAs. The media and strains used are indicated. The bars depict the mean and SD of n = 3 biological replicates. All values are normalized to the level in the WT strain grown on glucose. (B) Growth curves of WT, BT4338 (global regulator of carbon utilization) mutant, BT1876 (PUL24 anti-sigma) mutant, and their double mutant strains on RFOs. Points and error bars represent the mean and SD of n = 3 biological replicates. For each growth curve, the sugar used as the sole carbon source is indicated at the top along with the concentration used. (C) 5′ RACE to identify 5′ ends for BT1875 and BT1871 mRNAs. The bands that were sequenced for identifying novel 5' ends are highlighted in red boxes and the middle lane corresponds to a DNA ladder. Fig. S5 depicts the location of sequenced 5′ ends with respect to the flanking genes.

Fig 6

The alpha-galactosidase BT1871 and its homologs are important for melibiose utilization in Bacteroides species. (A) Growth curves of Bacteroides species containing a BT1871 homolog on RFOs. (B) Growth curves of Bacteroides species lacking a BT1871 homolog on RFOs. (C) Growth curves of wild-type B. ovatus ATCC 8483 (BO) and a mutant strain lacking two BT1871 homologs on RFOs. On melibiose, the WT strain had a mean growth rate of 0.32 h⁻¹ compared to 0.37 h⁻¹ for the double mutant strain, P = 0.06 by two-tailed t-test. On raffinose, the WT strain had a mean growth rate of 0.32 h⁻¹ compared to 0.37 h⁻¹ for the double mutant strain, P = 0.12 by two-tailed t-test. On stachyose, the WT strain had a mean growth rate of 0.40 h⁻¹ compared to 0.51 h⁻¹ for the double mutant strain, P = 0.14 by two-tailed t-test). (D) Growth curves of B. fragilis 638R and B. eggerthii DSM 20697 alongside their derived strains expressing BT1871 from a constitutive promoter. On raffinose, the B. fragilis WT strain had a mean growth rate of 0.33 h⁻¹ compared to 0.32 h⁻¹ for the BT1871 expressing strain (P = 0.69 by two-tailed t-test) while both the strains of B. eggerthii failed to grow. On stachyose, the B. fragilis WT strain had a mean growth rate of 0.70 h⁻¹ compared to 0.57 h⁻¹ for the BT1871 expressing strain (P = 0.21 by two-tailed t-test) while both the strains of B. eggerthii failed to grow. In all panels, points and error bars represent the mean and SD of n = 3 biological replicates. For each growth curve, the sugar used as the sole carbon source is indicated at the top along with the concentration used.

Fig 7

Model for raffinose family oligosaccharide utilization by B. thetaiotaomicron. After crossing the outer membrane (OM), RFOs are acted upon by BT1871 which cleaves the α-1,6 galactoside bond between glucose and galactose. Next, GH32 family enzymes act redundantly on the liberated sucrose moiety in the periplasm. The monosaccharide subunits are transported across the inner membrane (IM) by dedicated transporters. Expression of PUL22 genes depends on the HTCS BT1754 which binds fructose and activates PUL22 genes, including the sucrases that cleave the bond between glucose and fructose. The natural signal that induces PUL24 genes, including BT1871, is unknown, but basal levels of BT1871 transcription depend on BT4338. When the PUL24 anti-sigma factor is deleted, there is an RFO and BT4338-dependent upregulation from novel transcription start sites in PUL24 whose approximate locations are indicated by bent arrows. A dashed arrow indicates that BT4338 may also control the expression of PUL22 genes.