Genome-wide comparison of Corynebacterium diphtheriae isolates from Australia identifies differences in the Pan-genomes between respiratory and cutaneous strains

Abstract

Background: Corynebacterium diphtheriae is the main etiological agent of diphtheria, a global disease causing life-threatening infections, particularly in infants and children. Vaccination with diphtheria toxoid protects against infection with potent toxin producing strains. However a growing number of apparently non-toxigenic but potentially invasive C. diphtheriae strains are identified in countries with low prevalence of diphtheria, raising key questions about genomic structures and population dynamics of the species. This study examined genomic diversity among 48 C. diphtheriae isolates collected in Australia over a 12-year period using whole genome sequencing. Phylogeny was determined using SNP-based mapping and genome wide analysis.

Results: C. diphtheriae sequence type (ST) 32, a non-toxigenic clone with evidence of enhanced virulence that has been also circulating in Europe, appears to be endemic in Australia. Isolates from temporospatially related patients displayed the same ST and similarity in their core genomes. The genome-wide analysis highlighted a role of pilins, adhesion factors and iron utilization in infections caused by non-toxigenic strains.

Conclusions: The genomic diversity of toxigenic and non-toxigenic strains of C. diphtheriae in Australia suggests multiple sources of infection and colonisation. Genomic surveillance of co-circulating toxigenic and non-toxigenic C. diphtheriae offer new insights into the evolution and virulence of pathogenic clones and can inform targeted public health actions and policy. The genomes presented in this investigation will contribute to the global surveillance of C. diphtheriae both for the monitoring of antibiotic resistance genes and virulent strains such as those belonging to ST32.

Keywords: Diphtheria; Molecular epidemiology; Pan-genome analysis; Vaccine preventable disease; Virulence; Whole genome sequencing.

Fig. 1

Core phylogenetic tree with Local Health District (LHD), sequence type (ST), disease site and toxin gene presence marked. Coloured shaded blocks highlight sub-clades identified by ST and date isolated. Image prepared with iTOL [22]

Fig. 2

Maximum likelihood tree based on genome-wide SNP detection of reads mapped to reference NCTC13129. Branch lengths correspond to numbers of nucleotide substitutions per site. The heatmap shows Spa pilus gene clusters when compared to the reference NCTC11329 with high homology shown in yellow, absence or poor homology shown in blue

Fig. 3

The gene groups unique to sub-clades according to pan-genome analysis with core genome phylogenetic tree (a). The top panel (b) shows a single representative nucleotide sequence inferred for each gene of the pangenome. The middle panel (c) displays presence (blue) or absence (white) of blocks relative to genes and contigs in the pan-genome and metadata on disease site and health region (LHD). Disease site is classified as respiratory (green), cutaneous (orange) and blood (purple). There were four LHD regions identified as LHD-1 (orange), LHD-2 (red), LHD-3 (yellow) and LHD-4 (purple). Unique gene groups found in defined sub-clades have been circled and numbered accordingly; gene group I (red) - transposable elements and other proteins found in sub-clade 1; gene group II (orange) - 11 genes containing transposons, unique outer membrane proteins and a phenazine biosynthesis protein (PhzF); gene group III (green) – genes unique to sub-clade 2 containing tetO; gene group IV - sul1, genes for a fimbrial subunit type 1, sulphur carrying protein (ThiS), inner membrane transporter protein (RhIA) and VRR-NUC domain; gene group IV (purple) – genes unique to sub-clade 6 containing sdpA and sdpB, an integrase, von Willebrand factor type A domain protein and a putative transposon Tn552, all shown to be part of a NRPS/ PKS module and another NRPS module containing homologs of mbtB, irtA and irtB. The image was prepared using Phandango [21]

Fig. 4

The type 1 PKS cluster of C. diphtheriae. The collagen binding protein is indicated by the red circle and corresponds to the same green gene in each cluster. The homology of the cluster is indicated for a selection of well characterised C. diphtheriae genomes on Genbank, with the closest homology across the cluster. Siderophore modules were compared and image generated using antiSMASH [26]