Within-patient plasmid dynamics in Klebsiella pneumoniae during an outbreak of a carbapenemase-producing Klebsiella pneumoniae

Introduction Knowledge of within-patient dynamics of resistance plasmids during outbreaks is important for understanding the persistence and transmission of plasmid-mediated antimicrobial resistance. During an outbreak of a Klebsiella pneumoniae carbapenemase-producing (KPC) K. pneumoniae, the plasmid and chromosomal dynamics of K. pneumoniae within-patients were investigated. Methods During the outbreak, all K. pneumoniae isolates of colonized or infected patients were collected, regardless of their susceptibility pattern. A selection of isolates was short-read and long-read sequenced. A hybrid assembly of the short-and long-read sequence data was performed. Plasmid contigs were extracted from the hybrid assembly, annotated, and within patient plasmid comparisons were performed. Results Fifteen K. pneumoniae isolates of six patients were short-read whole-genome sequenced. Whole-genome multi-locus sequence typing revealed a maximum of 4 allele differences between the sequenced isolates. Within patients 1 and 2 the resistance gene- and plasmid replicon-content did differ between the isolates sequenced. Long-read sequencing and hybrid assembly of 4 isolates revealed loss of the entire KPC-gene containing plasmid in the isolates of patient 2 and a recombination event between the plasmids in the isolates of patient 1. This resulted in two different KPC-gene containing plasmids being simultaneously present during the outbreak. Conclusion During a hospital outbreak of a KPC-producing K. pneumoniae isolate, plasmid loss of the KPC-gene carrying plasmid and plasmid recombination was detected within the isolates from two patients. When investigating outbreaks, one should be aware that plasmid transmission can occur and the possibility of within- and between-patient plasmid variation needs to be considered.

Introduction detected in a urine sample of a patient (patient 2) admitted on the surgical ward. This event followed the earlier repatriation of a patient (the index patient 1) from an Italian hospital on the 11 th of September 2017, who was found to be colonized with KPC-KP two days after admission to the intensive care unit (Fig 1). Because carbapenem-resistant Enterobacteriaceae have been practically absent in this hospital so far and because patient 2 was also admitted on the intensive care unit previous to the detection of a KPC-KP in the patient's urine sample (3)(4)(5)(6)(7) October), the finding was considered suspect for nosocomial transmission (Fig 1). An outbreak management team was formed and in both the surgical ward and intensive care unit patient contacts were screened for KPC-KP carriage (Fig 1). During this outbreak, a total of 6

PLOS ONE
patients were colonized (n = 4) or infected (n = 2) with a KPC-KP. In all patients, for every specimen wherein a K. pneumoniae isolate was obtained, an isolate, regardless of the susceptibility pattern, was collected and stored at -80C˚using Microbank TM . Species identification was performed using the Bruker MALDI Biotyper TM (BD Diagnostics, MD, USA), and antimicrobial susceptibility testing was performed using the Phoenix TM platform (BD Diagnostics, MD, USA) and EUCAST breakpoints v.9.0. [13]. Every specimen, from which a K. pneumoniae isolate was obtained that was measured susceptible to meropenem and/or imipenem, was additionally inoculated on a CHROMagar TM KPC plate (CHROMagar, Paris, France).

Short-read whole-genome sequencing
A selection of K. pneumoniae isolates was sequenced on an Illumina MiSeq using Nextera XT chemistry (Illumina, San Diego, United States) and assembled with SPAdes v. 3.9.1 [14]. The selection was made in a way that: at least each K. pneumoniae isolate with a distinct susceptibility pattern and at least one K. pneumoniae isolate per patient per specimen type was sequenced. A distinct susceptibility pattern was defined as a four-fold difference in minimal inhibitory concentration (MIC) in any of the following antibiotics: amoxicillin-clavulanic acid, ceftriaxone, ceftazidime, meropenem, ciprofloxacin, and gentamicin. Before sequencing isolates were regrown and plated on a CHROMagarTM KPC plate (CHROMagar, Paris, France) when measured resistant to meropenem and on sheep blood agar when measured susceptible to meropenem. Plates were incubated for 18 to 24 hours at 35 to 37C˚. The DNA isolation and sequencing protocol are described in the S1 Data. The following quality control criteria for acceptable assemblies were used: coverage: �20; number of scaffolds: �1000; N50: �15.000 bases and maximum scaffold length: �50.000 bases.

Short-read whole genome analysis
Whole-genome MLST (wgMLST) (core and accessory genome) was performed for all sequenced isolates using Ridom SeqSphere+, version 4.1.9 (Ridom, Münster, Germany). Species-specific typing schemes were used as described by Kluytmans-van den Bergh et al. [15]. The all-to-all pairwise genetic difference was calculated between the isolates by counting the total number of allele differences in the wgMLST typing scheme and by dividing the total number of allele differences in the wgMLST typing scheme by the total number of shared alleles in the wgMLST typing scheme, ignoring pairwise missing values. The phylogenetic tree was visualized using iTOL v5.5.1 [16]. The genomes of the sequenced isolates were uploaded to the online bioinformatic tools ResFinder v.3.1 and PlasmidFinder v.2.0 (Center for Genomic Epidemiology, Technical University of Denmark, Lingby, Denmark) [17,18]. Acquired resistance genes were called when at least 60% of the length of the best matching gene in the ResFinder database was covered with a sequence identity of at least 90%. Plasmid replicon genes were called when at least 60% of the sequence length of the replicon gene in the Plasmid-Finder database was covered with a sequence identity of at least 80%.

Long-read whole genome sequencing
A selection of the isolates was long-read sequenced on a MinION sequencer using the FLO-MIN106D flow cell and the Rapid Barcoding Sequencing Kit SQK RBK004 according to the standard protocol provided by the manufacturer (Oxford Nanopore Technologies, Oxford, United Kingdom). The selection was made in a way that in each patient in which more than one isolate was short-read sequenced all isolates with a unique plasmid replicon content were long-read sequenced. Short-and long-read sequencing was performed from extracted DNA of the same regrown culture.

Hybrid short-and long-read plasmid analysis
A hybrid assembly of long-read and short-read sequence data was performed using Unicycler v.0.8.4 [19]. The genomes created using the hybrid assembly were uploaded to the online bioinformatic tools ResFinder v.3.1 and PlasmidFinder v.2.0 (Center for Genomic Epidemiology, Technical University of Denmark, Lingby, Denmark) [17,18]. Contigs created by the hybrid assembly that were smaller than 1000kb and that contained plasmid replicons were extracted from the assembly graph using BANDAGE v.0.8.1 [20]. All extracted plasmid contigs were annotated using Prokka v. 1.13.3 [21]. A pan-genome was constructed and pairwise comparisons were performed of plasmids between isolates of the same patient using BLAST+ v.2.6.0. (identity cut-off 95%) and Gview v.1.7. via the Gview webserver (https://server.gview.ca/) [22,23].

Accession numbers
Generated raw reads were submitted to the European Nucleotide Archive (ENA) of the European Bioinformatics Institute (EBI) under the study accession number: PRJEB35018 (link to data: https://www.ebi.ac.uk/ena/data/view/PRJEB35018).

Klebsiella pneumoniae collection
During the outbreak period, a total of 35 K. pneumoniae isolates (patient 1: n = 15, patient 2: n = 13, patient 3: n = 4 and patient 4-6: n = 1) with two distinct susceptibility patterns were collected (Fig 2; S1 Table). The two distinct susceptibility patterns were detected in isolates cultured from patient 2 only and were based on differences in MIC for amoxicillin-clavulanic acid, ceftriaxone, ceftazidime, and meropenem ( Table 1). In the specimens containing a K. pneumoniae isolate susceptible to meropenem, no growth was detected on the CHROMagar TM KPC plate. Fifteen isolates were selected to be sequenced: patient 1: n = 4, patient 2: n = 7, patient 3-6: n = 1 (Fig 2; Table 1).

Short-read whole genome analysis
Short-read WGS was performed on 15 K. pneumoniae isolates of 6 patients (Fig 2). Despite MIC testing revealing two distinct susceptibility patterns, using wgMLST the maximum number of allele differences detected between the various isolates was 4 (0.09%)(S1 Fig; S2 Table). Moreover, in none of the pairwise comparisons of the sequenced isolates did the number of allele differences exceed the limit of clonal relatedness (smaller or equal to 0.45%) as defined by Kluytmans-van den Bergh et al. (S1 Fig) [15]. The acquired resistance gene content did differ most notably with four isolates not containing a blaKPC gene in the Whole Genome Assembly (WGA) ( Table 2), explaining the difference in antimicrobial susceptibility profile seen between the isolates. Moreover, 2 isolates contained a tet(A) gene not detected in any of the other genomes ( Table 2). Plasmid replicon content also differed between the isolates: three isolates contained one plasmid replicon gene, three isolates contained two plasmid replicon genes and nine isolates contained three plasmid replicon genes. The difference in plasmid replicon-and acquired resistance gene-content was shown between isolates collected from the same patient (both in isolates from patient 1 and patient 2) (Tables 1 and 2).

Hybrid short-and long-read plasmid analysis
Four isolates were selected, based on within-patient plasmid replicon content differences, to be long-read sequenced: KP1 and KP3 of patient 1 and KP5 and KP9 of patient 2 (Fig 2).

PLOS ONE
Within-patient plasmid dynamics in Klebsiella pneumoniae during an outbreak Hybrid assembly of short-and long-read sequences revealed 10 contigs of which, based on size, 6 were assumed to be of plasmid origin ( Table 3). Despite the fact that only one IncFII(k) replicon was detected in the short-read whole-genome assembly of isolate KP1, the hybrid assembly revealed that the IncFII(K) plasmid replicon was actually present in two separate plasmids in the KP1 isolate: an IncFII(K) and IncFIB(K) replicon plasmid and an IncFII(K) and IncFIB(pQil) replicon plasmid. Within patient 1, GView BLAST analysis revealed the 3.1 plasmid to be a recombinant plasmid resulting from a recombination event between 18 CDS of plasmid 1.1 (containing the blaKPC gene), 50 CDS present in plasmid 1.1 and plasmid 1.2, 56 CDS of plasmid 1.2 only combined with an introduction of a tet(A) gene containing transposon (Fig 3A). Moreover, it revealed a loss of major parts of the plasmid content between isolates KP1 and KP3 without affecting the isolates susceptibility pattern. In patient 2, the loss of the entire 5.1 blaKPC gene-containing plasmid was observed between isolates KP5 and KP9 (Table 3). Additionally, a 26.786 bp deletion occurred in plasmid 5.2 when compared to plasmid 9.1 resulting in loss of 34 coding sequences among which were the antibiotic resistance genes aac(6')Ib-cr, blaOXA-1, and a catB3-like gene (Fig 3B; Table 3).

Discussion
The plasmid replicon content of the first sequenced blaKPC containing isolate of each patient during the hospital outbreak was similar between the different patients. However, during the outbreak within both patient 1 and patient 2 the plasmid replicon contents highly varied. This variation in plasmid replicon content was partially the result of plasmid loss observed in both patients, leading to a distinct susceptibility pattern in the isolates of one of these patients. A previous study also described plasmid loss during long time colonization in K. pneumoniae [9]. However, the present study also includes isolates with all distinct resistance patterns revealing the loss of a blaKPC containing plasmid within an outbreak setting. The similar replicon content of the plasmids 1.1, 1.2 and 5.1 would suggest that incompatibility between these plasmids [24,25] was the cause of the plasmid loss observed in the isolates of patient 1 and 2. Besides plasmid loss, within-patient 1 acquisition of a tet(A) containing transposon was detected in the plasmid content of the KP3 isolate when compared to the plasmid content of isolate KP1. Thus not only gene loss was revealed but also the acquisition of genetic elements in the plasmid content of the isolates collected during this outbreak. The plasticity of the plasmid content in bacterial isolates observed in this study has been described before both in vitro as in vivo [7,9,26]. However, recent reports also describe plasmids which remain highly stable [27,28]. This suggests that in vivo plasmid stability is likely the result of an interplay between host factors, plasmid content and the different plasmids composing the plasmid content of a bacterial isolate, possibly resulting in either a highly stable or unstable plasmid content. The hybrid assembly revealed that the 3.1 plasmid was the result of a recombination event between the 1.1 and the 1.2 plasmid occurring in the KP3 isolate only (and possibly the KP4 isolate) and not in the KP5 isolate. These recombination events between different plasmids in the same isolates have also been described in other studies [9,26]. However, this is to the best of our knowledge the first study to describe within patient blaKPC gene-containing plasmid recombination and loss during an hospital outbreak. This recombination event led to two Table 2. Acquired resistance gene-and plasmid replicon-content of whole-genome assembly of the sequenced K. pneumoniae isolates. different blaKPC plasmids occurring within one patient and plasmid transmission of these two different plasmids could have occurred during this outbreak. Several studies have already reported blaKPC-plasmid transmission between different isolates during outbreaks [11,29]. Distinguishing outbreak related from non-outbreak related plasmids based on sequence data is essential for using molecular data to confirm blaKPC-containing plasmid transmission in outbreaks. Our findings suggest that when investigating plasmid transmission during outbreaks, the possibility of within-patient plasmid variation needs to be considered. Therefore, it could well be that transmission of blaKPC-containing IncF plasmids within hospital outbreaks cannot be dismissed based on sequence dissimilarity between the different plasmids investigated. Complicating the investigation of plasmid transmission during hospital outbreaks even when the antibiotic susceptibility pattern is not altered. The present study has some limitations. The first was that in a culture of a specific specimen, colonies of the same morphology were not routinely isolated and stored. Therefore, possible subpopulations were not detected. Despite this, in the cultures in which a K. pneumoniae Table 3. Size, gene-, acquired antimicrobial resistance gene-and plasmid replicon-content of the plasmid contigs created with the hybrid assemblies.  was isolated that was measured susceptible to carbapenem, no growth was observed on the CHROMagar KPC TM suggesting that no resistant subpopulations were present in these specimens. Based on susceptibility pattern differences, specimen type of isolation, and plasmid replicon content a selection of the isolates were sequenced therefore plasmid variations that did not influence the susceptibility pattern, specimen type of isolation, and plasmid replicon content might go undetected. Moreover, since only 1 isolate was collected in patients 3-6 no conclusions can be drawn regarding longitudinal plasmid variation in these patients. Concluding, during a hospital outbreak of a blaKPC producing K. pneumoniae isolate plasmid loss of the blaKPC carrying plasmid and plasmid recombination was detected in two patients. When investigating outbreaks wherein plasmid transmission can occur, the possibility of within-and between-patient plasmid variation needs to be considered.

Patient isolate Plasmid replicon and acquired resistance gene content whole-genome assembly
Supporting information S1 Data. Supplementary materials and methods. (DOCX) S1