A gut-derived Streptococcus salivarius produces the novel nisin variant designated nisin G and inhibits Fusobacterium nucleatum in a model of the human distal colon microbiome

Abstract

Fusobacterium nucleatum is a human pathogen associated with intestinal conditions including colorectal cancer. Screening for gut-derived strains that exhibit anti-F. nucleatum activity in vitro revealed Streptococcus salivarius DPC6487 as a strain of interest. Whole-genome sequencing of S. salivarius DPC6487 identified a nisin operon with a novel structural variant designated nisin G. The structural nisin G peptide differs from the prototypical nisin A with respect to seven amino acids (Ile4Tyr, Ala15Val, Gly18Ala, Asn20His, Met21Leu, His27Asn, and His31Ile), including differences that have not previously been associated with a natural nisin variant. The nisin G gene cluster consists of nsgGEFABTCPRK with transposases encoded between the nisin G structural gene (nsgA) and nsgF, notably lacking an equivalent to the nisI immunity determinant. S. salivarius DPC6487 exhibited a narrower spectrum of activity in vitro compared to the nisin A-producing Lactococcus lactis NZ9700. Nisin G-producing S. salivarius DPC6487 demonstrated the ability to control F. nucleatum DSM15643 in an ex vivo model colonic environment while exerting minimal impact on the surrounding microbiota. The production of this bacteriocin by a gut-derived S. salivarius, its narrow-spectrum activity, and its anti-F. nucleatum activity in a model colonic environment indicates that this strain merits further attention with a view to harnessing its probiotic potential.IMPORTANCEFusobacterium nucleatum is a human pathogen associated with intestinal conditions, including colorectal cancer, making it a potentially important therapeutic target. Bacteriocin-producing probiotic bacteria demonstrate the potential to target disease-associated taxa in situ in the gut. A gut-derived strain Streptococcus salivarius DPC6487 was found to demonstrate anti-F. nucleatum activity, which was attributable to a gene encoding a novel nisin variant designated nisin G. Nisin G-producing S. salivarius DPC6487 demonstrated the ability to control an infection of F. nucleatum in a simulated model of the human distal colon while exerting minimal impact on the surrounding microbiota. Here, we describe this nisin variant produced by S. salivarius, a species that is frequently a focus for probiotic development. The production of nisin G by a gut-derived S. salivarius, its narrow-spectrum activity against F. nucleatum, and its anti-F. nucleatum activity in a model colonic environment warrants further research to determine its probiotic-related applications.

Keywords: Fusobacterium nucleatum; Streptococcus salivarius; antimicrobial; bacteriocin; colorectal cancer; ex vivo colon model; nisin.

Fig 1

An unknown antimicrobial is produced by S. salivarius DPC6487. Deferred antagonism assay whereby S. salivarius DPC6487 demonstrated antimicrobial activity against F. nucleatum DSM15643 (A). The presence of a 3404.59 Da mass was revealed by MALDI-TOF MS (B). The antimicrobial activity of S. salivarius DPC6487 CFS was assessed against Lactobacillus delbrueckii subsp. bulgaricus DPC5383 after subjection to heat (C), pH (D), and proteinase K (E).

Fig 2

Comparison of the nisin G gene cluster found in the genome of S. salivarius DPC6487 with nisin A, nisin H, nisin S, nisin U, nisin P, nisin O, and salivaricin D gene clusters. (A) Neighbor-joining (NJ) tree and multiple sequence alignment of nisin G and other nisin variants. Branches on the NJ tree are labeled with protein and producer isolate names, with Nisin G (NisG) highlighted. Salivaricin D (SalD) produced by S. salivarius 5M6c and Kunkecin A (KunA) produced by Apilactobacillus kunkeei FF30-6 were included as they are considered “nisin like.” Internal nodes are labeled and colored by bootstrap values, based on 100 rounds of bootstrapping. Branch length (x-axis) indicates the number of amino acid substitutions per site. Nis = Nisin; Sal = Salivaricin; Kun = Kunkecin. In the multiple sequence alignment (MSA), substituted positions with respect to NisinA are highlighted (Ile4Tyr, Ala15Val, Gly18Ala, Asn20His, Met21Leu, His27Asn, and His31Ile). (B) Figures shown in nisin A, nisin H, and salivaricin D gene clusters represent percentage amino-acid identity to nisin G gene homologs. Where no percentage identity is indicated, no homolog was identified in the nisin G cluster. NS, no similarity. Nisin accessions (reference is provided where accession is not linked to primary source): Nisin A, ABN45880; Nisin Z, ABV64387; Nisin F, ABU45463; Nisin Q, ADB43136; Nisin H, AKB95119; Nisin J (41); Nisin U, Q2QBT0; Nisin U2, ABO32538; Nisin P (42); Nisin O (43); Kunkecin A (44), Salivaricin D, and AEX55166. Nisin E (18) and Nisin S (17).

Fig 3

Quantification of F. nucleatum in colon model wells at 0, 6, and 24 hours. F. nucleatum numbers were significantly decreased in colon model wells that had been simultaneously inoculated with S. salivarius DPC6487 compared to wells inoculated with F. nucleatum DSM15643 alone. Mean F. nucleatum copy numbers and SDs for each treatment were derived from three colon model wells at each timepoint.

Fig 4

Taxonomic α-diversity (A), β-diversity (B), and relative abundances (C) across timepoints (T_0, T_6, and T_24). (A) Violin plots presenting Inverse Simpson Index values (α-diversity) and Berger-Parker Index values (dominance) at each timepoint. P values refer to Wilcoxon rank-sum tests and are adjusted for multiple testing. (B) PCoA of robust Aitchison distance matrices illustrates beta diversity of relative abundances of microbial taxa between groups. (C) Heatmap presents relative abundances of species within each sample (column), faceted by sampled timepoint. The main heatmap shows species present at a cumulative abundance >5% across all samples, while the bottom two rows show abundances for Fusobacterium mortiferum and F. nucleatum, which were present at relative abundances <1%.

Fig 5

Comparison of microbiome diversity after 24 hours of treatment with S. salivarius DPC6487. (A) Violin plots comparing α-diversity metrics between wells with any S. salivarius DPC6487 exposure (wells inoculated with S. salivarius DPC6487 alone or with a combination of S. salivarius DPC6487 and F. nucleatum DSM15643) and no S. salivarius DPC6487 exposure (Control or inoculated with F. nucleatum DSM15643 alone). (B and C) PCoA of microbial taxonomic profiles between S. salivarius DPC6487-treated and untreated wells using robust Aitchison and Bray-Curtis dissimilarity matrices. (D and E) Non-metric multidimensional scaling (NMDS) ordination plots between S. salivarius DPC6487-treated and untreated wells using (C) unweighted UNIFRAC and (D) weighted UNIFRAC distances.

Fig 6

Functional profiling of samples after 24 hours. PCoA of functional profiles between S. salivarius DPC6487-treated and untreated wells using robust Aitchison distance matrices for (A) SUPER-FOCUS level 1 pathways, (B) SUPER-FOCUS level 2 pathways, (C) SUPER-FOCUS level 3 pathways, (D) unstratified HUMAnN4 profiles, and (E) stratified HUMANn4 profiles. (F) Contributional alpha diversity computed from stratified HUMANn4 profiles using Pielou’s evenness index, filtered to display the top 99 pathways without NA values, and samples clustered using Euclidean distances.