The global dissemination of hospital clones of Enterococcus faecium

Abstract

Background: The hospital-adapted A1 group of Enterococcus faecium remains an organism of significant concern in the context of drug-resistant hospital-associated infections. How this pathogen evolves and disseminates remains poorly understood.

Methods: A large, globally representative collection of short-read genomic data from the hospital-associated A1 group of Enterococcus faecium was assembled (n = 973). We analysed, using a novel analysis approach, global diversity in terms of both the dynamics of the accessory genome and homologous recombination among conserved genes.

Results: Two main modes of genomic evolution continue to shape E. faecium: the acquisition and loss of genes, including antimicrobial resistance genes, through mobile genetic elements including plasmids, and homologous recombination of the core genome. These events lead to new clones emerging at the local level, followed by the erosion of signals of clonality through recombination, and in some identifiable cases producing new clonal clusters. These patterns lead to new, emerging lineages which are able to spread globally over relatively short timeframes.

Conclusions: The ability of A1 E. faecium to continually present new combinations of genes for potential selection suggests that controlling this pathogen will remain challenging but establishing a framework for understanding genomic evolution is likely to aid in tracking the threats posed by newly emerging lineages.

Fig. 1

Clustering among global A1 isolates. a Clustering of 973 E. faecium group A1 isolates based on the presence or absence of genes within the pan-genome identified using Panaroo. Labelled clusters are represented on a reduced dimensionality 2-D grid with member isolates coloured as shown in the legend. At a core-genome SNP level isolates within the same cluster are expected to share the same ancestry across the majority of their genomes. Using ChromoPainter, substantial core-genome admixture was detected in the 78 grey-shaded isolates, resulting in their exclusion from the designated pan-genome clusters. b The levels of admixture of the aforementioned 78 grey-shaded isolates in a (left) are contrasted with estimated admixture of 78 randomly chosen isolates from the remaining 895 isolates. The x-axis label shows the initial cluster assignment based on the pan-genome with the y-axis bars representing co-ancestry signals originating from other clusters, using the same colours as in a

Fig. 2

Cluster-associated genes. All 366 genes significantly associated (p < 0.05) with pan-genome clusters are depicted with chromosomal and plasmid-derived genes coloured blue and red respectively. Of the 366 cluster-associated genes, 129 genes occurred on a plasmid in at least one isolate. Genes (along the x-axis) are grouped by cluster (along the y-axis) as depicted in the legend. No gene was exclusively limited to any one cluster, with the largest complement of genes associated with cluster 5 (n = 107)

Fig. 3

Cluster 5: Evidence of local adaptation and regional dissemination. Maximum clade credibility tree of 129 cluster 5 isolates following masking of recombination. The first column to the right of the tree is coloured by sampled country (key left upper corner: AUS, Australia; BEL, Belgium; DEU, Germany; DNK, Denmark; ESP, Spain; EU, European Union; NLD, Netherlands). The subsequent column or heatmap depicts admixture events across the core genes of isolates coloured by “donating” cluster (key below heatmap). Similar background patterns of admixture are observed, accompanied by evidence of ongoing recombination at regional levels. Two such instances linked to Denmark and Germany are highlighted by boxes A and B respectively. Time scale is shown on the y-axis below the phylogeny

Fig. 4

Cluster 9. The bottom panel along the x-axis represents individual cluster 9 isolates (originating exclusively from Australia) with coloured vertical bars showing individual admixture signals across their core-genome by donor cluster. The dashed line depicts an admixture threshold of 10%. The top half of the panel reflects the proportions of admixed regions by donor country (left) and donor cluster (right). Overall, cluster 9 isolates share the majority of their genome co-ancestry (between 61% and 98%) with other cluster 9 isolates. In admixed regions, the best genomic match to the donor cluster originated from another Australian isolate in 53% of cases

Fig. 5

Dissemination routes of E. faecium across the world. Worldwide dissemination of genomes belonging to cluster 2 (n = 122) analysed using spatial, temporal and genomic data through BEAST v2.6 and visualised using SpreaD3. Coloured nodes on the map represent an isolate’s country of origin (as shown in the legend), while connecting lines are coloured by destination location. An expanded view is shown in panel a to show inferred spread of E. faecium isolates across Europe with complex links between and within countries. b E. faecium dissemination across the globe with surrounding circle sizes proportional to the number of lineages (isolates that share the same continuous line of ancestry) that occur in that location and captures the absolute and relative intensity of the local spread at any given point in time

Fig. 6

Antimicrobial Resistance. a Presence (red) or absence (white) of resistance genes grouped by isolates within Panaroo clusters along the y-axis. Genes are named along the x-axis with mutational resistance represented by gene names highlighted by 2 asterisks and grouped by AMR class. AG, aminoglycosides; LZ, linezolid; FI, folate inhibitors; FQ, fluoroquinolones; MLS, macrolide-lincosamide-streptogramin; DAP, daptomycin; TET, tetracycline; VAN, vancomycin; and CAT, chloramphenicol. b Heatmap of the average number of antimicrobial resistance genes resulting in resistance within isolates by country. No isolates were included from grey shades countries. c Circular dendrogram of the tetM gene with designated cluster by colour depicted in the outer ring. Evidence of clustering is still seen (e.g. clusters 7, 8 and 9) but extensive transfer of alleles between clusters by horizontal gene transfer is evident