Global Landscape of Clostridioides Difficile Phylogeography, Antibiotic Susceptibility, and Toxin Polymorphisms by Post-Hoc Whole-Genome Sequencing from the MODIFY I/II Studies

Abstract

Introduction: Clostridioides (Clostridium) difficile infection, the leading cause of healthcare-associated diarrhea, represents a significant burden on global healthcare systems. Despite being a global issue, information on C. difficile from a global perspective is lacking. The aim of this study is to model the global phylogeography of clinical C. difficile.

Methods: Using samples collected from the MODIFY I and II studies (NCT01241552, NCT01513239), we performed whole-genome sequencing of 1501 clinical isolates including 37 novel sequence types (STs), representing the largest worldwide collection to date.

Results: Our data showed ribotypes, multi-locus sequence typing clades, and whole-genome phylogeny were in good accordance. The clinical C. difficile genome was found to be more conserved than previously reported (61% core genes), and modest recombination rates of 1.4-5.0 were observed across clades. We observed a significant continent distribution preference among five C. difficile clades (Benjamini-Hochberg corrected Fisher's exact test P < 0.01); moreover, weak association between geographic and genetic distance among ribotypes suggested sources beyond healthcare-related transmission. Markedly different trends of antibiotic susceptibility among lineages and regions were identified, and three novel mutations (in pyridoxamine 5'-phosphate oxidase family protein: Tyr130Ser, Tyr130Cys, and a promoter SNP) associated with metronidazole-reduced susceptibility were discovered on a nim-related gene and its promotor by genome-wide association study. Toxin gene polymorphisms were shown to vary within and between prevalent ribotypes, and novel severe mutations were found on the tcdC toxin regulator protein.

Conclusion: Our systematic characterization of a global set of clinical trial C. difficile isolates from infected individuals demonstrated the complexity of the genetic makeup of this pathogenic organism. The geographic variability of clades, variability in toxin genes, and mutations associated with antibiotic susceptibility indicate a highly complex interaction of C. difficile between host and environment. This dataset will provide a useful resource for validation of findings and future research of C. difficile.

Keywords: Antibiotic susceptibility; CDI; Clostridioides (Clostridium) difficile; Evolution; Metronidazole; Moxifloxacin; Phylogeography; Rifaximin; Toxin; Vancomycin.

Fig. 1

Maximum-likelihood phylogenetic tree with MIC information of four drugs. Circular presentation of the recombination-free ML phylogeny (midpoint rooted in order to display the structure clearly, five isolates were removed as outliers by Gubbins). MLST clades are labeled on the outer circle; predominant (top 1 for clade 3, 4, 5; top 3 for clade 2; top 5 for clade 1) ribotypes are highlighted in the inner circle. The MICs of MOX, RIF, MET, and VAN for each isolate are presented as heatmaps plotted on the phylogenic tree. The color key for the four outer circles indicates the drug sensitivity of isolates as per EUCAST guidelines: red, resistant or reduced susceptibility; green, sensitive. Branch lengths are indicative of the estimated substitution rate per variable site. CDI C. difficile infection, EUCAST European Committee on Antimicrobial Susceptibility Testing, MET metronidazole, MIC minimum inhibitory concentration, ML maximum-likelihood, MLST multi-locus sequence typing, MOX moxifloxacin, RIF rifaximin, VAN vancomycin

Fig. 2

Genetic diversity of all major ribotypes. Each purple dot represents a PWSNP (left y-axis) between two isolates within a ribotype. Each red triangle is the nucleotide diversity (π; right y-axis). The number of samples is provided below the corresponding ribotype. Ribotypes are arranged in order of ascending nucleotide diversity (π). PWSNP pairwise single nucleotide polymorphism

Fig. 3

Geographic distribution of ribotypes and clades. In panels a, b, and c, each pie chart represents a collection of isolates from a region (country/province/city). The size of the pie chart corresponds to the sample size collected. The proportion of MLST clades is presented in each pie chart; the most predominant ribotype in each clade is provided in parentheses in the color key. The panels show: a global prevalence data; b prevalence data for individual states in the USA; c prevalence data for individual European countries (except Russia). In panel d, the size of the circles corresponds to the number of isolates collected. MLST multi-locus sequence typing

Fig. 4

Prevalence of resistance based on in-vitro data (MIC) and in silico prediction (CARD). a Reduced antibiotic susceptibility/resistance rate across ribotypes, clades, and geographic region. Heatmap of proportions of antibiotic-reduced susceptibility/resistant samples for VAN, MET, RIF, and MOX are plotted from top to bottom against ribotypes, color-coded according to clade or continent of sampling; values are shown for proportions > 0.5. b The presence of resistance determinants, identified using the Comprehensive Antibiotic Resistance Database and literature reports, is indicated with a red line. The resistance determinants are indicated by gene name (top) and Antibiotic Resistance Ontology accession number (bottom), and the corresponding antibiotic class and clade are indicated by the color-coded X and Y axes, respectively. The heatmaps represent the proportions of antibiotic-reduced susceptibility/resistant samples; values are shown for proportions > 0.5. EUCAST European Committee on Antimicrobial Susceptibility Testing, MET metronidazole, MIC minimum inhibitory concentration, MOX moxifloxacin, RIF rifaximin, VAN vancomycin

Fig. 5

Variant types of tcdC. VTs with a stop-gain mutation or deletion in the tcdC gene. Each VT has two sequences; the upper sequence is the nucleotide sequence, and the lower sequence is the corresponding amino acid sequence. “-” indicates a deletion; “.” indicates presence of the same base or residue as the reference gene; color indicates bases or residues that differ from the reference gene; “*” with black background indicates a stop codon. VT variant type