A family of anti-Bacteroidales peptide toxins wide-spread in the human gut microbiota

Abstract

Bacteria often produce antimicrobial toxins to compete in microbial communities. Here we identify a family of broad-spectrum peptide toxins, named bacteroidetocins, produced by Bacteroidetes species. We study this toxin family using phenotypic, mutational, bioinformatic, and human metagenomic analyses. Bacteroidetocins are related to class IIa bacteriocins of Gram-positive bacteria and kill members of the Bacteroidetes phylum, including Bacteroides, Parabacteroides, and Prevotella gut species, as well as pathogenic Prevotella species. The bacteroidetocin biosynthesis genes are found in horizontally acquired mobile elements, which likely allow dissemination within the gut microbiota and may explain their wide distribution in human populations. Bacteroidetocins may have potential applications in microbiome engineering and as therapeutics for polymicrobial diseases such as bacterial vaginosis and periodontal disease.

Fig. 1

Agar spot overlay assays demonstrating broad-spectrum toxin production by four Bacteroides strains. The four toxin-producing strains are shown at the top and were cultured on plates overnight and removed, then the plates were overlaid with top agar containing the strains shown on the left to test for growth inhibition. A dark spot indicates growth inhibition of the overlaid strain. a Six Bacteroides and Parabacteroides strains used in the overlays. b The toxin-producing strains used in the overlays. c Non-Bacteroidetes gut strains from three different phyla used in the overlays

Fig. 2

Identification and analysis of bacteriocin biosynthesis regions. a Agar overlay assays analyzing growth inhibition activity of BoCL02T12C04 and four transposon mutants against B. thetaiotaomicron VPI-5482 and B. vulgatus ATCC 8482. b ORF maps of a four-gene genetic region common to BoCL02 and BtCL15 showing the insertion sites of the transposon mutants (vertical blue arrows) that abrogated toxin production. Gene names of the BtCL15 strain beginning with the prefix EH213_0 are shown inside each gene. A similar genetic region from the BvCL01 and BvCL14 strains is shown below and the percent similarity of the proteins of these two regions is shown. Putative functions or properties of the proteins encoded by each gene are indicated. c Agar overlay assays analyzing the effects of deletion of three genes in the toxin biosynthesis region on the ability of BtCL15 to inhibit the growth of the two sensitive strains. d Agar overlay assays testing for the ability of the genes encoding the five transmembrane protein (p1845) or the thiol oxidoreductase (p1843) to protect B. thetaiotaomicron VPI-5482 when expressed in trans. e Agar overlay assay analyzing the ability of wild-type BtCL15 and the putative peptide bacteriocin gene deletion (Δ1844) to inhibit the growth of Δ1844. f Agar overlay assay analyzing the production of toxin from E. coli expressing the four-gene region (EH213_1845-1842) from an IPTG inducible promoter

Fig. 3

Analysis of the peptides produced from wild-type bacteria. a Alignment of the bacteriocin-like peptide of BvCL01/BvCL14 with that of BoCL02/BtCL15. The blue N-terminal sequence is the leader peptide predicted to be cleaved after the GG motif. Four cysteine residues in each peptide are highlighted in pink. b Alignment of the BoCL02/BtCL15 peptide with the class IIa bacteriocin pediocin from Pediococcus spp.. A conserved motif in class IIa bacteriocins of Gram-positive bacteria is underlined. c Western immunoblot analysis of the migration of reduced and non-reduced peptides from BvCL01 or BtCL15 probed with an antiserum generated to the synthesized peptides. d Results of LC–MS/MS analysis of the peptides from BvCL01/BvCL14 or BtCL15 showing the number of times each residue was detected by the number above. Yellow highlights indicate regions that were not detected by LC–MS/MS analysis. Source data are provided as a Source Data File

Fig. 4

Analysis of killing activity by synthesized 3825 peptide. a Growth inhibition assays by the 3825 peptide (bacteroidetocin A) when 1 µg is applied in a 5 µl spot on top of an agar overlay containing the indicated strains. b Growth inhibition assays in broth culture when 2 µg of bacteroidetocin A (red) or water control (blue) is added to the indicated strain. All data represent the average of biological triplicates, ±SEM. c Colony forming unit (cfu/ml) quantification of viable bacteria following four hours in growth medium with bacteroidetocin A (red) compared to the starting cfu/ml (gray), or after four hours growth with added water (blue). Biological triplicates were performed for the growth assay with added bacteroidetocin A. Error bars represent standard error of the mean. d Agar overlay assays demonstrating the amount of bacteroidetocin A necessary to inhibit the growth of the overlay strains. e Titration of 3825 peptide in broth to determine the minimal inhibitory concentration for growth inhibition of B. thetaiotaomicron VPI 5482 and P. goldsteinii CL02. Source data are provided as a Source Data File

Fig. 5

Ability of bacteroidetocin A to inhibit the growth of Prevotella strains. a Growth inhibition assays by bacteroidetocin A when 1 µg is applied in a 5 µl spot on top of an agar overlay containing the indicated strains. b Growth inhibition assays in broth culture when 2 µg of baceroidetocin A (red), or water control (blue) is added to the indicated strain. c Agar overlay assays demonstrating the amount of bacteroidetocin A necessary to inhibit the growth of the overlay strains. Source data are provided as a Source Data File

Fig. 6

Colonization level of bacteroidetocin B-producing and bacteroidetocin B mutant strain (BtCL15Δ1844) in the gnotobiotic mouse gut. a Colonization levels of gnotobiotic mice monoassociated with BtCL15 or BtCL15Δ1844. Fecal samples were collected after 6 days. The mean cfu/ml recovered from groups of three mice of each sex is plotted with the SEM as error bars; P value > 0.9999 for both groups of mice. The statistical analysis was performed using a two-tailed Mann–Whitney test as calculated by Prism version 8.0.1 for 64-bit Windows (GraphPad Software, San Diego, CA). b Western immunoblots of feces of mice monoassociated with BtCL15 or BtCL15Δ1844 for 6 days using 1844 (bacteroidetocin B) antiserum. M1–M3 and M7–M9 were monoassociated with wild-type BtCL15 and M4–M6 were monoassociated with BtCL15Δ1844. Source data are provided as a Source Data File

Fig. 7

Analysis of the extended genetic regions of the bacteroidetocin A and B biosynthesis genes. a Left - Names of all strains in our genome collection with the bacteroidetocin A biosynthesis region. Right - A 26-kb region common to all strains with the bacteroidetocin A biosynthesis genes (colored). A sliding window plot of the G+C content of the region is shown above. b Top - Names of all strains in our genome collection with the bacteroidetocin B biosynthesis region. Top orf map - A 96-kb integrative and conjugative element (ICE) common to all strains with the bacteroidetocin B biosynthesis genes (colored). Genes typical of ICE involved in mobility, excision and transfer are colored green. Middle panel – a similar genetic region present in the genome of B. caccae CL03T12C61. This region lacks the ~19-kb region present in the bacteroidetocin B ICE. Dotted line delineate the regions of divergence. The extent and percent DNA identity between regions is indicated with horizontal lines. Bottom panel – a similar ICE from strain B. uniformis CL03T12C37 that has a distinct ~26-kb central portion. A sliding window plot of the G+C content of the region is shown below

Fig. 8

Additional peptides encoded by Bacteroidetes genomes with bacteroidetocin-like properties. a Alignment of peptides that were retrieved by tblastn searches of the Bacteroidetes/Chlorobi superphylum genomes using bacteroidetocin A and B (including leader sequences) as queries. The predicted GG cleavage site is highlighted yellow. Cysteine residues are highlighted pink, and residues conserved in at least four of the five peptides are highlighted in turquoise. Pr- Prevotella, ole – oleiciplenus, col – colorans. b Peptide sequences of potential bacteroidetocins retrieved using tblastn repeatedly in an iterative manner. Double GG cleavage motif is highlighted yellow and cysteine residues are highlighted pink. c ORF maps of the 19 genetic regions of each of the non-redundant bacteroidetocin-like peptides. The bacteroidetocin-like genes are colored red, TM protein genes are colored light blue, thiol oxidoreductase genes are colored gold, and ABC transporter/peptidase genes are colored yellow. Genes colored white are distinct from the four genes of the bacteroidetocin A and B regions. d Toxin dot assays of synthesized peptides bacteroidetocin A (left), the Prevotella oralis ATCC 33269 HMPREF0663_ORF3 peptide (bacteroidetocin C) (middle), and the Prevotella. oryzae DSM 17970 XylorDRAFT_0090 peptide (bacteroidetocin D) (right). 1 µg of peptide was applied in a 5 µl spot on top of an agar overlay containing the indicated strains