Microevolution and genomic epidemiology of the diphtheria-causing zoonotic pathogen Corynebacterium ulcerans

Abstract

Corynebacterium ulcerans is an emerging zoonotic pathogen which causes diphtheria-like infections. Although C. ulcerans is found in multiple domestic and wild animal species, most human cases are linked with pets. Our ability to decipher cross-host species transmission dynamics and to understand the emergence of clinically relevant clones (e.g., diphtheria toxin-positive) is currently hampered by a limited knowledge of C. ulcerans strain diversity and genome evolution. Here, we explore the genomic population structure and evolution of C. ulcerans with 582 isolates from diverse hosts and geographical locations. A newly developed core genome genotyping scheme captures the population structure of C. ulcerans both at deep and shallow phylogenetic levels, uncovering its main sublineages and offering high strain subtyping resolution for epidemiological surveillance. Additionally, we reveal the diversity and distribution of the diphtheria toxin gene (tox), and those of its associated mobile elements. Considering the entire Corynebacterium diphtheriae Species Complex, we find four diphtheria toxin families, five tox-prophage families, and a novel tox-carrying genetic element. We show that some toxin families are shared across Corynebacterium species, revealing tox-prophage cross-species transfer. Our work enhances knowledge on the ecology and evolution of C. ulcerans and provides a genomic framework for tracking the dissemination of emerging sublineages.

Fig. 1. Phylogenetic tree of 582 Corynebacterium ulcerans genomes from human and animal hosts.

This maximum-likelihood tree was inferred from the core gene set and rooted using a Corynebacterium pseudotuberculosis outgroup (collapsed). Colored leaves indicate the host species (human, dog, cat) or host group of origin (farm animals, wild animals). External strips show clonal groups (CG) and sublineages (SL), with their respective number inherited from the seven-gene MLST nomenclature for the most important ones (with white standing for not assigned). For all tox+ isolates, the associated mobile genetic element (carrying the tox gene) is shown in the outermost strip, when its identification was possible (e.g., except for genomes that were too fragmented).

Fig. 2. Phylogenetic trees and isolation timeline of two major Corynebacterium ulcerans clonal groups (CG331, CG583).

The two maximum-likelihood phylogenetic trees were inferred from CG331 (A) and CG583 (B) isolates and rooted based on the complete C. ulcerans phylogenetic tree (Fig. 1). The first external strip shows the majority Sequence Type (ST) in dark yellow (ST331 for CG331, and ST325 for CG583) and other STs in lighter yellow (ST889 and ST933 for CG331, and ST583 and ST782 for CG583). The second and third external strips show the toxin alleles and the toxin-associated mobile genetic elements, respectively; C Timeline of isolation of CG331 and CG583 isolates.

Fig. 3. Genetic heterogeneity of Corynebacterium ulcerans isolates from sporadic isolates, case clusters, and cryptic clusters.

A Density plot showing the pairwise distances (i.e., number of allelic differences) between core genome multilocus sequence typing (cgMLST) profiles among sporadic isolates, cryptic clusters and case clusters; B Pairwise distances within each case cluster (n = 16), excluding those with only one genome available (Supplementary Table S1); C Pairwise distances within each cryptic cluster (n = 37, excluding FR_cryptic_17, Fig. S3), defined as single-linkage groups with ≤25 allele mismatches and no associated epidemiological data. In (B and C), each dot represents the pairwise distance between two isolates, whereas each vertical black line indicates the mean for each cluster. D The major cryptic cluster detected in this work (FR_cryptic_02, sequence type ST331), shows a strong geographical association with the Northern regions of France and temporal spread with progressive dissemination to neighboring regions over time.

Fig. 4. Toxin families of the Corynebacterium diphtheriae Species Complex (CdSC).

This maximum-likelihood phylogenetic tree (midpoint rooting) was inferred from all toxin alleles gathered from every tox-carrying species of the CdSC: C. diphtheriae, C. ulcerans, C. ramonii, C. silvaticum, and C. pseudotuberculosis. Toxin alleles cluster into four main toxin families, named after their main bacterial host species: Diphtheriae Toxin Family (DTF; dark blue), Ulcerans Toxin Family (UTF; dark green), Silvaticum Toxin Family (STF; dark purple), and Pseudotuberculosis Toxin Family (PTF; light purple). Some alleles are unique to one bacterial host species (e.g., STF, PTF, and the majority of DTF alleles), whereas others are found in more than one species (e.g., DTF tox_2 and tox_3, and eight alleles of the UTF).

Fig. 5. Distribution of toxin alleles among sublineages and mobile genetic elements in Corynebacterium ulcerans.

A Bar plot showing the main tox alleles of C. ulcerans and their distribution among sublineages (SL); NA: not assigned. B Bar plot of the main tox alleles detected in C. ulcerans and the mobile genetic elements (MGE) carrying them; unknown MGE are due to high genome fragmentation; PAI pathogenicity island; ILE integron-like element.

Fig. 6. Toxin alleles of the Corynebacterium diphtheriae Species Complex (CdSC) and their association with tox-prophage families.

Phylogenetic classifications of C. ulcerans (n = 176) and C. diphtheriae (n = 232) tox-prophages (right) and their tox alleles (left), with main alleles shown in color. Prophage families (PF) are indicated on the internal branches of the corresponding subtree in the prophage phylogenetic tree (PF1–PF5), in which leaves are colored according to the bacterial host species of detection. The phage integrase type (1–3) is depicted on the right-hand side of the tox allele as shades of gray.