Extensive genome analysis identifies novel plasmid families in Clostridium perfringens

Abstract

Globally, the anaerobic bacterium Clostridium perfringens causes severe disease in a wide array of hosts; however, C. perfringens strains are also carried asymptomatically. Accessory genes are responsible for much of the observed phenotypic variation and virulence within this species, with toxins frequently encoded on conjugative plasmids and many isolates carrying up to 10 plasmids. Despite this unusual biology, current genomic analyses have largely excluded isolates from healthy hosts or environmental sources. Accessory genomes, including plasmids, also have often been excluded from broader scale phylogenetic investigations. Here we interrogate a comprehensive collection of 464 C. perfringens genomes and identify the first putative non-conjugative enterotoxin (CPE)-encoding plasmids and a putative novel conjugative locus (Bcp) with sequence similarity to a locus reported from Clostridium botulinum. We sequenced and archived 102 new C. perfringens genomes, including those from rarely sequenced toxinotype B, C, D and E isolates. Long-read sequencing of 11 C. perfringens strains representing all toxinotypes (A-G) identified 55 plasmids from nine distinct plasmid groups. Interrogation of the 464 genomes in this collection identified 1045 plasmid-like contigs from the nine plasmid families, with a wide distribution across the C. perfringens isolates. Plasmids and plasmid diversity play an essential role in C. perfringens pathogenicity and broader biology. We have expanded the C. perfringens genome collection to include temporal, spatial and phenotypically diverse isolates including those carried asymptomatically in the gastrointestinal microbiome. This analysis has resulted in the identification of novel C. perfringens plasmids whilst providing a comprehensive understanding of species diversity.

Keywords: Clostridium perfringens; genome; horizontal gene transfer; phylogeny; plasmids; toxins.

Fig. 1.

Distribution of C. perfringens isolates across species, clades, toxinotypes and phylogeny, where scale is branch length which is defined as base pair substitutions per sequence site. Heatmaps showing (a) the number of isolates in each clade per source and toxinotype and (b) the number of isolates from each source per toxinotype. Significantly overrepresented groups are indicated, as determined by Fisher’s exact test, *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001, *****P<0.00001. (c) Core gene phylogeny of 464 C . perfringens strains with toxinotype (outer ring), source (inner ring) and strains sequenced in this study (pink triangles).

Fig. 2.

Toxin, AMR and MGE genes are widespread throughout the C. perfringens phylogeny. Genome distance tree generated from the whole genome sequences of 464 C. perfringens isolates, where scale is branch length which is defined as base pair substitutions per sequence site. Each strain is annotated with clade (branch colours and highlights), and their toxin, antimicrobial resistance (AMR) genes and mobile genetic elements (MGEs) (boxes right of the tree).

Fig. 3.

Alignments of sequenced plasmid to the three known plasmid types, Tcp, Pcp and pIP404. BRIG diagrams depicting the 24 Tcp plasmids (outer coloured rings) that have similarity to the well-studied pCW3 plasmid (inner ring annotation), eight Pcp plasmids (outer coloured rings) which shared identity with pCP13 (inner ring annotation) and six pIP404 plasmids (outer coloured rings) that share sequence with pIP404 (inner ring annotation).

Fig. 4.

Alignment of Bcp group plasmids. An alignment of the Bcp group of plasmids p58875_8 and pJGS1984_5 and the plasmid-encoding contig pPS49_13 that contain virD4, virB4 and mobC genes that are similar to those in the C. botulinum pCDC3875 (blue text).

Fig. 5.

Alignment of small plasmids and genetic arrangement of the phage-like plasmid. An alignment of the small plasmids and the phage-like plasmid. Each plasmid contains a rep gene (yellow arrow). Four subgroups of small plasmids formed based on the similarity of their rep genes: subgroup 1 (red box), subgroup 2 (blue box), subgroup 3 (green box) and subgroup 4 (yellow box).

Fig. 6.

Identification of a third conjugation locus, Bcp. Maximum-likelihood trees of the (a) VirB4 and (b) VirD4 conjugation proteins identified as being encoded by the closed plasmids, where scale is branch length which is defined as base pair substitutions per sequence site. Each was categorized as being from the Tcp conjugation group (purple), the Pcp conjugation group (green) or the novel Bcp conjugation group (blue), where coloured bubbles represent sequences with at least 40 % sequence identity for VirB4 and 20 % for VirD4.

Fig. 7.

Categorization of plasmid-encoding contigs of C. perfringens. (a) Similarity to known VirD4, VirB4 and Rep proteins allowed for categorization of plasmid-contigs into seven groups: Tcp, Pcp, Bcp, Phage-like, pIP404, small and unclassified. (b) The number of each type of plasmid contig per clade. Overrepresented groups, Fisher’s exact test: *P<0.05, **P<0.01, *****P<0.00001.