Diversity of the Genomes and Neurotoxins of Strains of Clostridium botulinum Group I and Clostridium sporogenes Associated with Foodborne, Infant and Wound Botulism

Abstract

Clostridium botulinum Group I and Clostridium sporogenes are closely related bacteria responsible for foodborne, infant and wound botulism. A comparative genomic study with 556 highly diverse strains of C. botulinum Group I and C. sporogenes (including 417 newly sequenced strains) has been carried out to characterise the genetic diversity and spread of these bacteria and their neurotoxin genes. Core genome single-nucleotide polymorphism (SNP) analysis revealed two major lineages; C. botulinum Group I (most strains possessed botulinum neurotoxin gene(s) of types A, B and/or F) and C. sporogenes (some strains possessed a type B botulinum neurotoxin gene). Both lineages contained strains responsible for foodborne, infant and wound botulism. A new C. sporogenes cluster was identified that included five strains with a gene encoding botulinum neurotoxin sub-type B1. There was significant evidence of horizontal transfer of botulinum neurotoxin genes between distantly related bacteria. Population structure/diversity have been characterised, and novel associations discovered between whole genome lineage, botulinum neurotoxin sub-type variant, epidemiological links to foodborne, infant and wound botulism, and geographic origin. The impact of genomic and physiological variability on the botulism risk has been assessed. The genome sequences are a valuable resource for future research (e.g., pathogen biology, evolution of C. botulinum and its neurotoxin genes, improved pathogen detection and discrimination), and support enhanced risk assessments and the prevention of botulism.

Keywords: Botulism; Clostridium botulinum; Clostridium sporogenes; Foodborne; Genomes; Infant; Neurotoxins; Wound.

Figure 2

Phylogeny of genomes of C. botulinum Group I and C. sporogenes. Two major lineages (C. botulinum Group I (shown in black) and C. sporogenes (shown in red)) were identified. The phylogenetic tree was created by comparison of core single nucleotide polymorphisms identified using ParSNP program [70], Treegraph v2 [71], MEGA7 [72] and Figtree were used to annotate and visualize the phylogenetic tree. Accessory genes were identified using BLAST using reference sequences for the ha and orfX neurotoxin cluster configurations. The distance bar (0.05) represents the number of nucleotide substitutions per site for a given branch, based on the number of SNPs found in the core genome. Details of individual sequences and strains are given in Tables S1 and S2, where they appear in the same order as in this Figure. Newly sequenced strains are indicated (in black) as “new” in the right-hand column.

Figure 1

Geographical locations and heat map of C. botulinum Group I and C. sporogenes strains included in the present study. A total of 359 strains were attributed to a geographical location. Further details about individual isolates are given in Table S1. For each specified country, the detected botulinum neurotoxin genes for isolates are indicated (in black; with NT = no toxin gene), and the types of botulism are shown in colour (N = non-clinical; I = infant botulism; U = unknown; F = foodborne botulism; W = wound botulism).

Figure 3

Phylogeny of sub-types of botulinum neurotoxin type A. The toxin protein sequences were aligned with the Muscle module of MEGA7 [72] algorithm, and the phylogenetic tree was generated using the Neighbour-Joining method. Variants of five type A neurotoxin sub-types were identified. Scale bar represents the number of amino acid substitutions per site. A total of 271 amino acid sequences were analysed. White blocks represent absence of information regarding source. Black blocks represent reference neurotoxins [45]. Further details about individual sequences and isolates are given in Table S3, where they appear in the same order as in this Figure. The sources of the isolates are given in Table S1.

Figure 4

Phylogeny of sub-types of botulinum neurotoxin type B. The toxin protein sequences were aligned with the Muscle module of MEGA7 [72] algorithm, and the phylogenetic tree was generated using the Neighbour-Joining method. Variants of six type B neurotoxin sub-types were identified. Scale bar represents the number of amino acid substitutions per site. A total of 239 amino acid sequences were analysed. White blocks represent absence of information regarding source. Black blocks represent reference neurotoxins [45]. Further details about individual sequences and isolates are given in Table S3, where they appear in the same order as in this Figure. The sources of the isolates are given in Table S1.

Figure 5

Phylogeny of sub-types of botulinum neurotoxin type F. The toxin protein sequences were aligned with the Muscle module of MEGA7 [72] algorithm, and the phylogenetic tree was generated using the Neighbour-Joining method. Variants of five type F neurotoxin sub-types were identified. Scale bar represents the number of amino acid substitutions per site. A total of 39 amino acid sequences were analysed. White blocks represent absence of information regarding source. Black blocks represent reference neurotoxins [45]. Further details about individual sequences and isolates are given in Table S3, where they appear in the same order as in this Figure. The sources of the isolates are given in Table S1.

Figure 6

Pan-genome comparison of the C. botulinum Group I and C. sporogenes lineages (Figure 2). (A) Graphical representation of feature distribution linked to the SNP-based phylogenetic tree (Figure 2). (B) The lineage-specific genes are distributed throughout the genome, as shown using representative genomes for the C. botulinum Group I lineage (ATCC 19397) and the C. sporogenes lineages (NCIMB 10696). Dark blue circles represent C. botulinum Group I lineage features, and red circles represent C. sporogenes lineage features.