Tracking of Antibiotic Resistance Transfer and Rapid Plasmid Evolution in a Hospital Setting by Nanopore Sequencing

Abstract

Infections with multidrug-resistant bacteria often leave limited or no treatment options. The transfer of antimicrobial resistance genes (ARG) carrying plasmids between bacterial species by horizontal gene transfer represents an important mode of expansion of ARGs. Here, we demonstrate the application of Nanopore sequencing in a hospital setting for monitoring transfer and rapid evolution of antibiotic resistance plasmids within and across multiple species. In 2009, we experienced an outbreak with extensively multidrug-resistant Pseudomonas aeruginosa harboring the carbapenemase-encoding bla_IMP-8 gene. In 2012, the first Citrobacter freundii and Citrobacter cronae strains harboring the same gene were detected. Using Nanopore and Illumina sequencing, we conducted comparative analysis of all bla_IMP-8 bacteria isolated in our hospital over a 6-year period (n = 54). We developed the computational platform plasmIDent for Nanopore-based characterization of clinical isolates and monitoring of ARG transfer, comprising de novo assembly of genomes and plasmids, plasmid circularization, ARG annotation, comparative genome analysis of multiple isolates, and visualization of results. Using plasmIDent, we identified a 40-kb plasmid carrying bla_IMP-8 in P. aeruginosa and C. freundii, verifying the plasmid transfer. Within C. freundii, the plasmid underwent further evolution and plasmid fusion, resulting in a 164-kb megaplasmid, which was transferred to C. cronae Multiple rearrangements of the multidrug resistance gene cassette were detected in P. aeruginosa, including deletions and translocations of complete ARGs. In summary, plasmid transfer, plasmid fusion, and rearrangement of the ARG cassette mediated the rapid evolution of opportunistic pathogens in our hospital. We demonstrated the feasibility of near-real-time monitoring of plasmid evolution and ARG transfer in clinical settings, enabling successful countermeasures to contain plasmid-mediated outbreaks.IMPORTANCE Infections with multidrug-resistant bacteria represent a major threat to global health. While the spread of multidrug-resistant bacterial clones is frequently studied in the hospital setting, surveillance of the transfer of mobile genetic elements between different bacterial species was difficult until recent advances in sequencing technologies. Nanopore sequencing technology was applied to track antimicrobial gene transfer in a long-term outbreak of multidrug-resistant Pseudomonas aeruginosa, Citrobacter freundii, and Citrobacter cronae in a German hospital over 6 years. We developed a novel computational pipeline, pathoLogic, which enables de novo assembly of genomes and plasmids, antimicrobial resistance gene annotation and visualization, and comparative analysis. Applying this approach, we detected plasmid transfer between different bacterial species as well as plasmid fusion and frequent rearrangements of the antimicrobial resistance gene cassette. This study demonstrated the feasibility of near-real-time tracking of plasmid-based antimicrobial resistance gene transfer in hospitals, enabling countermeasures to contain plasmid-mediated outbreaks.

Keywords: IMP-8; Nanopore; Nanopore sequencing; Pseudomonas aeruginosa; antimicrobial resistance; genome assembly; horizontal gene transfer; long read; pathoLogic; plasmIDent; plasmid-mediated resistance; plasmids; surveillance studies.

FIG 1

Schematic diagram of the data analysis workflow used in this study. The pathoLogic platform was created using the Nextflow (39) environment to chain different tools and scripts, represented here as circular nodes. Connecting lines indicate data flow between the separate processes; dashed lines show tools that are not directly included in the pipeline and need manual data handling. In pathoLogic, the assembly step (*) can be performed by Unicycler (16), Canu (40), miniasm (41), hybridSPAdes (15), or flye (42, 43). ORF, open reading frame.

FIG 2

(A) Timeline of isolation of bla_IMP-8 Gram-negative bacteria in the hemato-oncology department over 6 years. Bars represent isolates from patients and the length of their stay in the hospital. Patient 21 was seen only in the outpatient department, marked with an “O.” Environmental isolates are marked with an “X” at the date of isolation. The introduction of a rectal screening program is marked with a black arrow. (B) Overview of plasmids with relevance to the evolution of the bla_IMP-8 plasmid found in P. aeruginosa, C. freundii, and C. cronae. (C) Maximum likelihood phylogeny of Citrobacter species included in the study (n = 9). The Citrobacter freundii strains formed two clusters, Cf1 (n = 5) and Cf2 (n = 3). Strains of cluster Cf2 harbored a chromosomal transposon region (black triangle) homologous to the regions of plasmid C. C. cronae clustered with the closely related C. werkmanii NBRC105721 and DSM17579 strains. The scale bar shows the expected number of nucleotide changes per site. PA, P. aeruginosa; CF, Citrobacter freundii; CC, Citrobacter cronae.

FIG 3

Detailed alignment of plasmids A, B, and C. Plasmid A (blue, outer circle) harbored a multidrug resistance cassette that included bla_IMP-8 on a class 1 integron and shows high GC content (inner circle, green). Plasmid B (green, outer circle) harbored no bla_IMP-8 resistance gene and shows lower GC content (red). Plasmid C (red) was the largest and comprised plasmid A, plasmid B, duplicated regions 1 and 2 (black arrows), and a unique extension by two Tn3 elements in one fusion region. Highlighted in blue are hits in the IS finder database annotated to the transposon or IS family level. Parts of Tn3 family transposons are present in different locations of the plasmids. The class 1 integron consists of 9 AMR genes, including those encoding aminoglycosides, beta-lactams, and sulfonamides. Additional AMR genes and a mercury resistance operon (23) are present within the duplicated region. The resistance gene translocations are displayed in a schematic manner.

FIG 4

(A) Coverage plot based on short-read Illumina data mapped against reference plasmid C found in C. cronae. Plasmid C (red bar) comprises of the sequences of plasmids A (blue bar) and B (green bar) fused by two transposon-rich regions. Red lines indicate breakpoints, which are characterized by the absence of reads spanning the breakpoint. White areas indicate the absence of coverage and hence the absence of sequence in a given isolate, which in some cases could have been the result of a deletion event or might indicate boundaries between scaffolds. (B) Comparison of the resistance gene cassette configurations of RSC1 and RSC2, showing putative AMR gene translocation events and a deletion of qacH.

FIG 5

Concept of plasmid evolution and transmission across three bacterial species. P. aeruginosa bla_IMP-8 was isolated approximately 2.5 years prior to the first isolation of Citrobacter species harboring bla_IMP-8, leading to the hypothesis of a transfer of plasmid A from P. aeruginosa to Citrobacter species. Occurrences of the C. freundii and C. cronae bla_IMP-8 genes started at the same time; thus, the timeline does not suggest a specific direction of the transfer. However, the existence of the transposon region in the chromosome of C. freundii cluster Cf2 (marked with a black triangle) makes it the most likely host of the merging of plasmid A and B, which are linked by two copies of the transposon region. Solid arrows represent the transmission sequence resulting from this hypothesis.