Molecular Evolution of Extensively Drug-Resistant (XDR) Pseudomonas aeruginosa Strains From Patients and Hospital Environment in a Prolonged Outbreak

Abstract

In this study, we aimed to elucidate a prolonged outbreak of extensively drug-resistant (XDR) Pseudomonas aeruginosa, at two adjacent hospitals over a time course of 4 years. Since all strains exhibited a similar antibiotic susceptibility pattern and carried the carbapenemase gene bla_VIM, a monoclonal outbreak was assumed. To shed light on the intra-hospital evolution of these strains over time, whole genome sequence (WGS) analysis of 100 clinical and environmental outbreak strains was employed. Phylogenetic analysis of the core genome revealed the outbreak to be polyclonal, rather than monoclonal as initially suggested. The vast majority of strains fell into one of two major clusters, composed of 27 and 59 strains, and their accessory genome each revealed over 400 and 600 accessory genes, respectively, thus indicating an unexpectedly high structural diversity among phylogenetically clustered strains. Further analyses focused on the cluster with 59 strains, representing the hospital from which both clinical and environmental strains were available. Our investigation clearly shows both accumulation and loss of genes occur very frequently over time, as reflected by analysis of protein enrichment as well as functional enrichment. In addition, we investigated adaptation through single nucleotide polymorphisms (SNPs). Among the genes affected by SNPs, there are a multidrug efflux pump (mexZ) and a mercury detoxification operon (merR) with deleterious mutations, potentially leading to loss of repression with resistance against antibiotics and disinfectants. Our results not only confirm WGS to be a powerful tool for epidemiologic analyses, but also provide insights into molecular evolution during an XDR P. aeruginosa hospital outbreak. Genome mutation unveiled a striking genetic plasticity on an unexpectedly high level, mostly driven by horizontal gene transfer. Our study adds valuable information to the molecular understanding of "real-world" Intra-hospital P. aeruginosa evolution and is a step forward toward more personalized medicine in infection control.

Keywords: SNPs; WGS; extensive drug resistance; functional enrichment; mercury; protein enrichment; single-nucleotide polymorphisms; whole genome sequencing.

Figure 1

Project workflow summary flowchart. XDR-PA, extensively drug-resistant Pseudomonas aeruginosa; HQ, high quality; WGS, whole genome sequence; SNP, single nucleotide polymorphism.

Figure 2

Core genome phylogeny of the 100 XDR P. aeruginosa outbreak strains included in this study, taking recombinational events into account. Out of these 100 strains, 23 strains (all from cluster 2) were excluded by the Gubbins algorithm because they were genetically identical with one of the other strains. Thus, only 36 strains of cluster 2 (constituted of a total of 59 strains) are shown. Grouping into clusters 1–8 is indicated in translucent colors: (i) two major clusters 1 and 2 (green and orange, respectively), and (ii) six small clusters 3, 4, 5, 6, 7, and 8 (red, gray, yellow, violet, turquoise, and blue, respectively). Sampling source (hospital A and B) is indicated for each strain in non-translucent colors (dark green and dark orange, respectively). As evident from the branches in the far left of the figure, the phylogenetic tree shows two major clusters, designated clusters 1 and 2, alongside with the six additional small clusters, designated cluster 3–8. In the middle left of the figure, the color coding (green/orange) indicates in which of the two hospitals (hospital A/B, respectively) the individual strains were sampled. In the middle and to the right of the figure, the colored barcode lines (red and blue) represent predicted recombinations, either shared by multiple isolates through common descent (red blocks), or occurring on terminal branches which are unique to individual isolates (blue blocks). In the far right of the figure, the source of the strains [P, patient (bold) or E, environment (italic)] and the MLST type are indicated. MLST, multi-locus sequence typing; XDR, extensively drug-resistant.

Figure 3

Accessory genome of cluster 2. All strains of cluster 2 (n = 59) are shown. Heatmap visualization with gene presence indicated by black bars and gene absence by white bars. Frames contain genes pertaining to AG group I, II, and III (in green, violet, and blue, respectively) and are presented in magnification inserts (B–D) next to the heatmap (A). An overview of AG groups and gene blocks is given in the table insert (D). (A) X axis: accessory genes are consecutively numbered, beginning with 1 through 474. Left Y axis: indication of (i) the source of the strains [P, patient (bold) or E, environment (italic)] (“P/E”), (ii) the MLST type (“MLST”), (iii) day of sampling (“sample day”) consecutively counted with the day of sampling of the oldest strain in this study (strain ID 57) set to 1, (iv) strain identification number (“strain ID”), chronologically ordered by timepoint of sampling descending from early to late, and (v) AG groups I, II and III (“AG group”). Right Y axis: assignment of the strains to three time groups corresponding to early, middle and late time periods of sampling (yellow, orange and red, respectively). (B–D) Magnification inserts: AG group I, II and III (in green, violet, and blue, respectively) with indication of gene blocks 1–0 and the number of genes contained within these. Missing gene blocks are framed in the respective color, whereas additional gene blocks are filled in the respective color. Genes not belonging to one of the gene blocks are filled in black color. For AG groups I and II, some additional strains (without genes belonging to one of the gene blocks) are displayed in the magnification inserts for comparative visualization and are filled in black color. Abbreviations: AG group, accessory genome group; MLST, multi-locus sequence typing.

Figure 4

(A) Protein enrichment of cluster 2, Venn diagram visualization of proteins shared between time groups corresponding to early, middle and late time periods of sampling (yellow, orange and red, respectively) during the course of the outbreak. (B) Functional enrichment of GO terms within cluster 2, visualization of GO term names overrepresented or underrepresented in time groups corresponding to early, middle and late time periods of sampling during the course of the outbreak. Enrichment (overrepresentation): ↑ (green), depletion (underrepresentation): ↓ (red). GO, gene ontology.