Population genomics and antimicrobial resistance in Corynebacterium diphtheriae

Abstract

Background Corynebacterium diphtheriae, the agent of diphtheria, is a genetically diverse bacterial species. Although antimicrobial resistance has emerged against several drugs including first-line penicillin, the genomic determinants and population dynamics of resistance are largely unknown for this neglected human pathogen. Methods Here, we analyzed the associations of antimicrobial susceptibility phenotypes, diphtheria toxin production, and genomic features in C. diphtheriae. We used 247 strains collected over several decades in multiple world regions, including the 163 clinical isolates collected prospectively from 2008 to 2017 in France mainland and overseas territories. Results Phylogenetic analysis revealed multiple deep-branching sublineages, grouped into a Mitis lineage strongly associated with diphtheria toxin production and a largely toxin gene-negative Gravis lineage with few toxin-producing isolates including the 1990s ex-Soviet Union outbreak strain. The distribution of susceptibility phenotypes allowed proposing ecological cutoffs for most of the 19 agents tested, thereby defining acquired antimicrobial resistance. Penicillin resistance was found in 17.2% of prospective isolates. Seventeen (10.4%) prospective isolates were multidrug-resistant (≥ 3 antimicrobial categories), including four isolates resistant to penicillin and macrolides. Homologous recombination was frequent (r/m = 5), and horizontal gene transfer contributed to the emergence of antimicrobial resistance in multiple sublineages. Genome-wide association mapping uncovered genetic factors of resistance, including an accessory penicillin-binding protein (PBP2m) located in diverse genomic contexts. Gene pbp2m is widespread in other Corynebacterium species, and its expression in C. glutamicum demonstrated its effect against several beta-lactams. A novel 73-kb C. diphtheriae multiresistance plasmid was discovered. Conclusions This work uncovers the dynamics of antimicrobial resistance in C. diphtheriae in the context of phylogenetic structure, biovar, and diphtheria toxin production and provides a blueprint to analyze re-emerging diphtheria.

Ongoing diphtheria outbreak in Yemen: a cross-sectional and genomic epidemiology study

“…Full susceptibility or very low levels of penicillin resis tance were observed for all Yemen diphtheria isolates, which did not carry the the pbp2m gene encoding for penicillin binding protein identified in C diphtheriae in 2020. 22 Amoxicillin and erythromycin were active against all isolates, as were many other tested molecules. Resistance to ciprofloxacin (but not to moxifloxacin) was observed in two isolates, which both had an Asp93Gly substitution in the quinolone resistancedetermining region of their gyrA gene.…”

Treat Me Well or Will Resist: Uptake of Mobile Genetic Elements Determine the Resistome of Corynebacterium striatum

“…In corynebacteria, different ARGs associated with mobile genetic elements, including plasmids, integrons (In), insertion sequences (IS), and transposons (Tn), have been reported [ 39 , 43 , 56 , 57 ]. In C. striatum, resistance is mainly determined by Tn, IS, and plasmids.…”

Genomic epidemiology and strain taxonomy ofCorynebacterium diphtheriae

“…While this work was in progress, additional C. diphtheriae genomes were published from Austria [27], the USA [23], Yemen [9], France [11], Spain [28] and India [29]. We included these 162 genomes in a joint phylogenetic analysis with the 440 above genomes (Figure…”

“…The allele call rate of the cgMLST scheme loci was determined using posteriorly published genomes from Austria [27], the USA [23], Yemen [9], France [11], Spain [28] and India [29] and was found to be 98.6%  1.5%. We also calculated the call rate using 337 additional, prospectively collected genome sequences from the French NRC for diphtheria (unpublished).…”

“…Diphtheria toxin gene tox-positive isolates were observed in 19 sublineages distributed on the Southern hemisphere of the tree presented in Figure 1, including the sublineage that includes the vaccine strain PW8, and in 5 sublineages in the Northern hemisphere of the tree, including the one that comprised strain NCTC13129 [11]. These two hemispheres of the tree are identifiable by the distribution of the spuA gene, involved in starch utilization [54]: the upper lineage was spuA-positive with the exception of six spuA-negative sublineages, whereas in sharp contrast, the lower lineage was almost entirely spuA negative, with ST5 being the single exceptional spuA-positive sublineage.…”