Genomic Diversity of Enterotoxigenic Strains of Bacteroides fragilis

Abstract

Enterotoxigenic (ETBF) strains of Bacteroides fragilis are the subset of strains that secrete a toxin called fragilysin (Bft). Although ETBF strains are known to cause diarrheal disease and have recently been associated with colorectal cancer, they have not been well characterized. By sequencing the complete genome of four ETBF strains, we found that these strains exhibit considerable variation at the genomic level. Only a small number of genes that are located primarily in the Bft pathogenicity island (BFT PAI) and the flanking CTn86 conjugative transposon are conserved in all four strains and a fifth strain whose genome was previously sequenced. Interestingly, phylogenetic analysis strongly suggests that the BFT PAI was acquired by non-toxigenic (NTBF) strains multiple times during the course of evolution. At the phenotypic level, we found that the ETBF strains were less fit than the NTBF strain NCTC 9343 and were susceptible to a growth-inhibitory protein that it produces. The ETBF strains also showed a greater tendency to form biofilms, which may promote tumor formation, than NTBF strains. Although the genomic diversity of ETBF strains raises the possibility that they vary in their pathogenicity, our experimental results also suggest that they share common properties that are conferred by different combinations of non-universal genetic elements.

Fig 1. Phylogeny of B. fragilis ETBF and NTBF strains.

Evolutionary distance values are based on average nucleotide identity (ANI) as determined by CLGenomics software. The scale ranges from 99–100%.

Fig 2. Diversity of B. fragilis ETBF and NTBF strains.

(A) The genome sequences of ETBF strain were aligned to that of NTBF strain NCTC 9343 using Mauve. Blocks of homology between strains are color-coordinated. Non-homologous regions are white, and inverted regions are shown below the center line. Regions containing CTn9343 (which lacks the BFT PAI) and CTn86 are boxed. The diversity of the location of this mobile element is highlighted by interconnecting lines. (B) Intra-species variation is shown by pairwise comparison of non-orthologous coding sequences (<10% identity as determined by BLASTp using CLgenomics software) present in five ETBF and two NTBF strains to NTBF strain NCTC 9343. (C) Orthologous protein clusters observed in three ETBF and three NTBF strains. The Venn diagram indicates overlapping clusters between strains. The total number of clusters that are unique to a single strain and that are shared between 2–6 strains is also shown.

Fig 3. Genes conserved in ETBF strains are located predominantly in the BFT PAI and CTn86.

The three regions of the ETBF strain 20656-2-1 genome that contain genes present in all five ETBF strains but not present in the NTBF reference genomes NCTC 9343, 638R, and YCH46 are shown. The conserved genes are shown in red. Other genes that encode proteins with predicted functions are listed and color-coded based on the COG database classification [51] Hypothetical genes are shown in gray. The penultimate four digits of the Genbank accession numbers of genes of known or predicted function are indicated (the final “0” in the accession numbers have been omitted).

Fig 4. ETBF strains secrete variable amounts of mature Bft.

Culture supernatants were obtained from ETBF strains at the indicated growth phases. The volume of each supernatant was normalized based on cell number (OD₆₀₀) and proteins were concentrated 40-fold. Bft was then detected by Western blot using an anti-peptide antiserum.

Fig 5. NTBF strain NCTC 9343 outcompetes ETBF strains and secretes a growth-inhibitory protein.

(A) In vitro competition assay. Strain NCTC 9343 was co-cultivated with one ETBF strain and passaged after 24 h. The competitive index (CI, ratio of ETBF/NCTC 9343 DNA) was determined by strain-specific qPCR. Significance was determined by ANOVA across all time points for each strain (2-078382-3, p = 0.016; 20793–3, p = 0.0048; 20656-2-1, p = 0.048). Tukey’s multiple comparison test results are shown comparing 24 and 48 h time points to the inoculum (* = p<0.05, ** = p<0.01). Each data point is an independent measurement. (B) Growth curve of NCTC 9343 and individual ETBF strains. One representative experiment is shown. (C) Agar diffusion assay. Supernatant obtained from a culture of strain NCTC 9343 was untreated (a) PK treated (b), heat treated (c), or trypsin treated (d) and then added to wells in plates seeded with the ETBF indicator strain 2-078382-3: (D) Quantitation of agar diffusion assays in which culture supernatants from strain NCTC 9343 WT or NCTC 9343 ΔtssC were added to ETBF strain 2-078382-3, 20793–3, 20656-2-1, or 86-5443-2-2. Supernatants were untreated, heat-treated, protease treated, or fractionated by molecular weight. A Student’s t-test was used to demonstrate a significant loss of activity (*, p<0.05). The red circles indicate a complete loss of activity.

Fig 6. ETBF strains exhibit enhanced biofilm formation.

(A) B. fragilis strains forming biofilms in microtiter plates were stained with crystal violet and quantified at A₅₇₀. A single strain or a combination of three ETBF strains (2-078382-3, 20793–3, and 20656-2-1) was grown in BHIS medium, in co-culture with strain NCTC 9343, or in an equal mixture of BHIS medium and NCTC 9343 culture supernatant. Data from at least three independent experiments are shown along with the standard deviations. Asterisks indicate significant changes as determined by unpaired t-test: * p<0.05, **p<0.01, ***p<0.001.