- Email: [email protected]
Genome sequencing and characterization of an extensively drug-resistant sequence type 111 serotype O12 hospital outbreak strain of Pseudomonas aeruginosa
Accepted Article
Received Date: 07-Oct-2013 Revised Date: 20-Dec-2013 Accepted Date: 07-Jan-2014 Article Type: Original Article - E-only Category: Bacteriology
Running Title: Pseudomonas aeruginosa ST111 outbreak strain genome
Title: Genome sequencing and characterization of an XDR ST111 serotype O12 hospital outbreak strain of Pseudomonas aeruginosa
Adam A. Witney1, Katherine A. Gould1, Cassie F. Pope2, Frances Bolt2, Neil G. Stoker1, Marc D. Cubbon3, Christina R. Bradley4. Adam Fraise4, Aodhán S. Breathnach2, Philip D. Butcher1, Tim D. Planche1,2, Jason Hinds1.
1
Division of Clinical Sciences, St George’s University of London, Cranmer Terrace,
SW17 0RE. 2
Department of Microbiology, St George’s Healthcare NHS Trust, Blackshaw Road,
London, SW17 0QT. 3
Brighton and Sussex University Hospitals NHS Trust, Brighton, UK.
This article has been accepted for publication and undergone full peer review but has not been through the copyediting, typesetting, pagination and proofreading process, which may lead to differences between this version and the Version of Record. Please cite this article as doi: 10.1111/1469-0691.12528 This article is protected by copyright. All rights reserved.
Accepted Article
4
Hospital Infection Research Laboratory, Queen Elizabeth Hospital Birmingham, UK.
Corresponding author: Dr Adam A Witney, Division of Clinical Sciences, St George’s, University of London, Cranmer Terrace, London, SW17 0RE. UK. Tel: +44 208 725 0698 Email: [email protected]
1. Abstract A series of extensively drug-resistant (XDR) isolates of Pseudomonas aeruginosa from two outbreaks in UK hospitals were characterized by whole genome sequencing (WGS). Although these isolates were resistant to antibiotics other than colistin, we confirmed that they are still sensitive to disinfectants. The sequencing confirmed that isolates in the larger outbreak were serotype O12 and further revealed that they belonged to the ST111 sequence type which is a major epidemic strain of P. aeruginosa throughout Europe. As this is the first reported sequence of an ST111 strain, the genome was examined in depth, focusing particularly on antibiotic resistance and potential virulence genes, and on the reported regions of genome plasticity. High degrees of sequence similarity were discovered between outbreak isolates collected from recently infected patients, sinks, an isolate from the sewer and an historical isolate, suggesting that the ST111 strain has been endemic in the hospital for many years. The ability to translate easily from outbreak
This article is protected by copyright. All rights reserved.
Accepted Article
investigation to detailed genome biology using the same data demonstrates the flexibility of WGS application in a clinical setting.
2. Introduction An outbreak investigation into the spread of multidrug-resistant Pseudomonas aeruginosa susceptible only to colistin, in two UK hospitals was previously described (1). Recent guidelines redefine these isolates as XDR (2), due to susceptibility to only one class of antibiotic (colistin). The outbreaks occurred in university teaching hospitals in the South of England, one in London (outbreak 1) and one on the South coast (outbreak 2). In outbreak 1, 85 hospital-wide cases were identified between 2005 and 2011, and during outbreak 2, four cases occurred in a single ward between 2009 and 2010. An environmental investigation was conducted to identify potential sources of infection and inform control measures that might reduce spread.
Outbreak 1 consisted of sporadic outbreaks separated by long infection-free periods. This outbreak was caused by a serotype O12 strain that caused severe infection, usually in immune-compromised patients. As reported previously (1) the overall case fatality during outbreak 1 was 34/85 (40%); in patients with bacteraemia the case fatality was as high as 14/18 (78%). Haematology patients were more likely to have bacteraemia (6/7 cases). During outbreak 2 all cases had bacteraemia and there were no deaths attributed to XDR P. aeruginosa. The potential of waste water systems to act as a reservoir for XDR P. aeruginosa was reported in both outbreaks. The UK reference laboratory reported the strain associated with outbreak 1 as being a unique type confined to hospital 1 using Variable Number Tandem Repeat (VNTR)
This article is protected by copyright. All rights reserved.
Accepted Article
analysis (3). However, genome sequencing revealed the outbreak strain was sequence type ST111, one of a small number of major epidemic clones that are widespread in Europe, and for example was recently reported as the major metalloβ-lactamase-producing P. aeruginosa in the Netherlands (4). We describe the clinical application of whole-genome sequencing (WGS) in these two hospital outbreaks to discover the relationships between circulating patient and environmental strains that may inform infection control measures. Furthermore, greater understanding of the biology of the XDR strain gained from the genome sequence may provide useful insights into the pathogenicity and guide future treatment or disinfectant options. To our knowledge, this is the first report of a genome sequence of an ST111 or an XDR strain of P. aeruginosa. We describe mutations and elements associated with antibiotic resistance and compare the genome sequence to a collection of clinical and environmental isolates from both outbreaks and other published genomes.
3. Results 3.1 Isolate phenotyping All isolates of P. aeruginosa that were collected from outbreak 1 from June 2005 onwards were resistant to carbapenems. These isolates were sent for typing at a reference laboratory and 74/85 were all reported as serotype O12, PCR-positive for the blaVIM-2 β-lactamase gene, and identical or very similar by PFGE and/or VNTR typing (1). The other 11 isolates were all different and deemed to be unrelated sporadic cases of different resistant P. aeruginosa strains. A small number were single-locus variants (SLVs) by VNTR typing, which we considered as most likely
This article is protected by copyright. All rights reserved.
Accepted Article
being the same strain with minor sequence variations leading to an SLV. The reference laboratory also reported this VNTR type as being unique to this hospital. In order to characterize this strain type, PA38182 (designated as the primary isolate because it was from a current patient (Patient A)), as well as eight control samples were examined further (see Table 1).
•
Four environmental isolates from outbreak 1 were included: two P. aeruginosa isolates (PA48922, PA46218) from patient A's sink (isolated 3 weeks apart), a P. putida isolate (PP48922) that was isolated at the same time from the sink, as there is evidence that P. putida can act as a reservoir of mobile elements for P. aeruginosa (5), and a P. aeruginosa isolate obtained from the sewer using a Moore's swab (PASwab).
•
One P. aeruginosa isolate (PA128572) of the same VNTR-type isolated from a patient (Patient B) in Hospital 1 six years earlier.
•
One XDR PA isolate (PA99736) from Patient C at the same hospital with a different VNTR-type.
•
Two XDR P. aeruginosa isolates (PA97498, PA156066) from outbreak 2 - one clinical and one environmental.
These isolates were characterized by VNTR (Table 1), and MIC data for the outbreak 1 representative isolate PA38182 is shown in Table 2A. As these isolates were susceptible to only one class of antibiotics, their susceptibility to disinfectants is critical for infection control purposes. We therefore carried out tests with a variety of disinfectants (Table 2B). These showed that the isolates from
This article is protected by copyright. All rights reserved.
Accepted Article
outbreak 1 were similar in sensitivity to a susceptible control strain, with all being sensitive at practical concentrations.
3.2 Sequence analysis WGS was carried out using an Ion Torrent Personal Genome Machine generating between 0.5 - 2 million reads per sample, with an average length of 120 bp. Sequence data were analysed both by reference mapping and de novo assembly approaches. The genome sequence-inferred serotypes were determined by BLAST comparison against publicly available O-antigen gene sequences (see Methods). An identity of greater than 99% across the entire 25kb O12 locus confirmed the reference laboratory's designation of the index isolate (PA38182) as serotype O12, in addition to all other isolates related to the outbreak, including the six year old isolate. The unrelated isolate from Hospital 1 (PA99736) was inferred as serotype O6. Multi-locus sequence typing (MLST) has been used to type P. aeruginosa strains across the world (6), and has the advantage of ease of data comparison between studies. MLST derived from genome sequence identified the outbreak cluster isolates from Hospital 1 as ST111, whereas the VNTR unrelated isolate (PA99736) was ST155, and the two Hospital 2 isolates were ST654 (Table 1).
3.3 Phylogenetic reconstruction In order to determine epidemiological links between the isolates, the phylogenetic relatedness of the isolates was determined in two ways: using core genome sequence alignment and SNP-based reconstruction. Whole genome phylogenetic This article is protected by copyright. All rights reserved.
Accepted Article
reconstruction was performed using the core genomes of the isolates sequenced here in the context of 35 genomes available in the public databases (Figure. 1A). The ST111 and ST654 isolates were more related to each other than other sequences in the public databases; however the number of publicly available P. aeruginosa genome sequences available is limited and the strains diverse. The ST155 isolate clustered tightly with other ST155 strains from China. The previously sequenced serotype O12 strain (PA7) is a taxonomic outlier (7) and is significantly diverged from the ST111 O12 isolates sequenced in this study (Supplementary Figure 1).
Therefore PA7 was removed from the tree, as was an unpublished
sequence similar to PA7, VRFPA01 (Genbank: AOBK01000001.1) in order to aid visualisation of the remaining relationships.
In order to investigate further the relationships between the ST111 isolates, SNP sites were identified with respect to a reference sequence. Sites where all isolate sequences were identical but differed only from the reference were ignored, leaving only variant sites between the isolates. Phylogenetic reconstruction of the ST111 outbreak isolates based on 24 variant sites identified by mapping to the LESB58 reference genome (RefSeq: NC_011770; several reference sequences were used independently due to the lack of a previously sequenced ST111 reference and showed a consistent relationship (data not shown)) suggested that the sewer isolate is very closely related to both the clinical and the associated sink isolates collected (Figure 1B). Also, the similarity of the historical isolate (PA128572), suggests that this strain has been present in the hospital for at least 6 years.
This article is protected by copyright. All rights reserved.
Accepted Article
3.4 Genome analysis of XDR ST111 isolates Sequence reads of the XDR ST111 isolates were both mapped to other published sequences and assembled de novo. The isolates were highly similar, but the invasive clinical isolate PA38182 was used as the primary isolate for ST111 in the sequence analysis. For this isolate 460 contigs (> 500bp) totalling 7 Mb were obtained, and no major deletions compared to the reference strains were identified except those accessory elements detailed below. Despite being a global epidemic strain of P. aeruginosa, there is no previously reported genome sequence of an ST111 strain. Therefore a detailed study of the PA38182 genome content was conducted, focusing particularly on antibiotic resistance and virulence elements. 3.4.1 Antibiotic Resistance The MICs in Table 2A show that ST111 is clinically resistant to most antibiotics. Analysis of the genome sequence of ST111 identifies key genetic mutations and the presence of multiple genes implicated in resistance to a wide range of pharmaceutical agents. 3.4.1.1 Quinolones Amino acid changes in the gyrA (Thr83Ile) and parC (Ser87Leu) genes of ST111 are most likely responsible for the resistance to the fluoroquinolones as described previously (8). 3.4.1.2 Aminoglycosides Several aminoglycoside resistance encoding genes were identified; ST111 contains a pair of aacA29a / aacA29b genes (aminoglycoside-6’N-acetyltransferase), most likely surrounding the blaVIM-2 gene, as described previously (9). Also, an aminoglycoside phosphotransferase (aph) IIb, a FadE36 aph and a further putative
This article is protected by copyright. All rights reserved.
Accepted Article
aph gene were identified. The ST654 isolates did not contain the aacA29a / aacA29b genes but did contain copies of the same aph genes as ST111; furthermore it also contained an ant(4')-IIb gene and two genes annotated as associated with streptomycin resistance. Isolate ST155 has limited specific aminoglycoside specific genes containing copies of only the aph-IIb and the putative aph described above. 3.4.1.3 β-lactamases The ubiquitous class C β-lactamase gene, ampC, showed no apparent mutations, neither did ampG. However, mutations in the ampC regulator genes, ampD (G148A) and ampR (E283G), were present but it is not known what role these may play. Class B metallo-β-lactamase (MBL) genes, providing resistance to β-lactams except monobactams, are commonly located on mobile elements.
ST111 isolates in
Norway and Sweden were reported to have both the blaVIM-2 MBL and associated aminoglycoside acetyltransferase genes (aacA29a / aacA29b) on a single cassette (9); our ST111 isolate contains the same genes, however we were unable to deduce the arrangement from the assembly, likely due to the high level of identity between aacA29a and aacA29b flanking the blaVIM-2 gene, causing contig breaks either side of the blaVIM-2 gene. The ST654 isolates described in (9) had a cassette carrying blaVIM-2 MBL, the strA and strB streptomycin resistance genes, and aadB but no aacA29a or aacA29b genes. This is consistent with our ST654 sequence except no aadB gene was found, suggesting a variant in this isolate. 3.4.1.4 Efflux pumps Efflux pumps are potentially responsible for resistance to a wide range of agents in P. aeruginosa, and indeed the ST111 isolate contains copies of the mexAB-oprM,
This article is protected by copyright. All rights reserved.
Accepted Article
mexCD-oprJ, mexEF-oprN, mexHI-opmD operons and their respective regulator genes mexR, nfxB, mexT and mexG. Additionally mexMN, mexVW and mexXY sequences were identified. More specialised efflux pumps were also identified. These included triABC (10) and mexJKL operons (11), shown to be associated with triclosan resistance, and the czcCBA operon, shown to confer heavy metal resistance (12). These appeared to be contained within integrated phages or integrons. Additionally the copper-induced mexPQ-opmE efflux pump (13) was present. Other P. aeruginosa efflux systems identified were muxABC-opmB (14), emrAB-opmG (15) and an SMR (small multidrug resistance family of proteins) that confer resistance to a wide variety of quaternary ammonium compounds through a proton-drug efflux antiport mechanism. 3.4.1.5 Colistin All these isolates are currently sensitive to colistin (polymyxin E). Initial studies into the genetic basis of colistin resistance have implicated the 2-component systems pmrAB and phoPQ genes in P. aeruginosa, Salmonella enterica serovar Typhimurium and Acinetobacter baumannii (16–18), presumably by altering expression of the genes that are directly responsible. Our studies confirmed a lack of alterations in either the pmrAB or phoPQ genes in the isolates studied here. 3.4.2 Virulence Factors Comparison to the Virulence Factor Database (http://www.mgc.ac.cn/VFs/) shows that most of the common virulence factors are present in the ST111 isolates studied here. Notably the type III secretion system equivalent to PA1690-PA1725 locus in PAO1 is present along with the genes for the effectors, exoenzymes ExoS and ExoT and adenylate cyclase ExoY. (ST654 carries the same effectors but ST155 is
This article is protected by copyright. All rights reserved.
Accepted Article
missing the exoY gene). The ST111 ExoS protein has an S121N change relative to published sequences. The Type IV pilus biosynthesis cluster is present as is the associated twitching motility protein locus. 3.4.3 The ST111 Accessory Genome Studies of the P. aeruginosa genome show it to have a conserved core genome, with variable regions scattered throughout. These regions have been referred to as Regions of Genome Plasticity (RGPs) (19) by including 'any region of at least four contiguous ORFs that are missing in at least one of the genomes analyzed'. While some of these RGPs appear to be rare deletions, others appear to be sites where frequent changes occur, through additions and losses within the region, or through replacement of the entire region. Some of the genes or sets of genes can involve potential virulence genes, and include phages, integrons carrying Type IV secretion systems, proteins encoding resistance to heavy metals, organic solvents or acriflavin, or efflux pumps.
Supplementary Table 1 includes a detailed analysis of the accessory genome in isolate PA38182. Several accessory elements appeared novel to P. aeruginosa as shown by BLAST similarity to the non-redundant database at NCBI: • RGP2 contains an araC family regulator gene as part of a ~5kb region which has no BLAST hits to any P. aeruginosa sequences in NCBI’s non-redundant database, but shows 98% similarity to the P. putida genome sequenced in this study. • RGP5 contains an integrated conjugative element including a type II secretion system and a DNA topoisomerase III gene, both also novel to P. aeruginosa.
This article is protected by copyright. All rights reserved.
Accepted Article
• Other integrated conjugative elements containing Trb conjugal protein loci were also found in RGP26, RGP36 and RGP75, the latter two also containing the copG gene, a putative metal transporting P-type ATPase, thought to confer copper resistance (20). Furthermore, the element contained within RGP75 has been sequenced in P. aeruginosa strain NCGM2.S1 previously, as well as having 99% identity to the P. putida sequenced in this study. • Other examples of accessory elements carrying resistance genes can be seen in RGP29, where a phage like element contains the czcCBA heavy metal efflux pump system and RGP56 contains the putative multidrug-efflux transporter also equivalent to PLES_22791 in LESB58. • A Type IV pilus operon can be found in RGP41, which has a high level of similarity to the that found in RGP7 of PA7 (7). • RGP42 is the site for another novel phage, this containing a zona occludens toxin highly similar to PA0726, found in RGP5 of PAO1. • Comparing to previously described genomic islands, evidence of complete or remnant PAGI-7 (RGP88), PAGI-8 (RGP86) and PAGI-9 (RGP89) sequences was also seen (21). All the ST111 isolates sequenced in this study carried the same accessory components except for the presence of a phage in RGP87, which was variably present in the isolates; only the historical clinical isolate PA128572 and the recent environmental isolates PA48922 and PA46218 contained the phage (Figure 2). However, no obvious phenotypic association to the presence of this phage could be identified.
This article is protected by copyright. All rights reserved.
Accepted Article
4. Discussion Whole genome sequencing has been demonstrated to be a very powerful tool in diagnostic microbiology (22–24) with the potential to provide results in a clinically relevant time frame. This is especially true for epidemiological typing which can guide infection control investigation of outbreaks. In this study, WGS was applied to a selection of isolates of P. aeruginosa that were collected during an investigation into two hospital outbreaks (1). We have shown that in these outbreaks, the environmental and patient isolates of P. aeruginosa that are causing significant morbidity are highly related within each hospital. This provides further evidence of the potential of the waste water system acting as a reservoir for infection.
Identification of ST111 The isolates from outbreak 1 had previously been shown to be closely related by VNTR typing and were thought to be specific to the originating hospital. However, WGS analysis showed that the strain belongs to the global epidemic type ST111, which is thought to be responsible for many of the MDR P. aeruginosa cases around the world (4,9). We report the first full genome sequence of an ST111 strain. ST111 isolates have been identified in the UK previously by MLST (25), however MLST typing is not routinely performed on incoming isolates by the UK reference laboratory. This demonstrates that VNTR typing has limited utility in isolation without comparison to MLST and that sequence-based typing schemes must be adopted to use in tandem with or replacing VNTR typing.
This article is protected by copyright. All rights reserved.
Accepted Article
Sequencing selected genes from strain PA7, which is also serotype O12, has previously shown that it is significantly divergent from the worldwide MDR O12 clone (27). Sequencing of the O12 ST111 isolates in this study provides more complete genomic evidence of how dissimilar these two O12 strain types are.
Water colonisation Blockages account for the most significant contamination of clinical areas from the waste water system (1). Therefore, the implication of the waste water system as the source of the XDR P. aeruginosa in 2011 led to the implementation of many control measures in hospital 1 to reduce contamination of clinical areas from blockages, leaks and splashback events (1). Subsequently only infrequent XDR P. aeruginosa colonisation
events
and
clinical
infections
have
occurred
in
both
immunocompromised and non-immunocompromised individuals associated with the main hospital wing.
Indeed, since April 2012 there have been only four P. aeruginosa ST111 clinical infections in patients who have spent time on the General Intensive Care Unit (GICU), consisting of two infections in 2012 and two infections in 2013. On both occasions ST111 strains were isolated simultaneously from the sink U-bends and basins in patient rooms and from the sluice room. The fact that in this report we recently recovered from the sewer an isolate of ST111 highly related to those from current patient and environmental isolates and from a historical patient isolate, is further evidence that the waste water system probably acts as a reservoir for nosocomial transmission of XDR P. aeruginosa in this
This article is protected by copyright. All rights reserved.
Accepted Article
hospital. However, sequencing of other ST111 isolates from a separate source would give us a much greater understanding of the level diversity within this strain population. The recurring nature of these infections over a lengthy period of time suggests that this is more likely than transfer of XDR P. aeruginosa from colonised patients to the waste water system. Furthermore, consideration must be given to the downstream environmental impact of drug resistant P. aeruginosa, which has been shown to be found in wastewater treatment plants downstream of hospitals elsewhere (26).
The ST111 genome In addition to the epidemiological information generated by WGS, genetic markers of drug resistance and virulence can be identified and potentially used to guide better treatments
for
clinical
case
management.
Specific
genetic
markers
for
fluoroquinolones, aminoglycosides and β-lactams resistance were identified in our study; indeed Akasaka et al (28) suggested that the pattern of gyrA mutations identified in our ST111 isolates would confer susceptibility to the fluoroquinolone Sitofloxacin, but we were unable to obtain this drug to test. Indeed, rapid genotyping for these SNPs across these specific gyrA regions could be a useful clinical indicator for antibiotic choice. The wide-ranging specificities of the efflux pump machinery in P. aeruginosa are striking, but make full antibiotic and disinfectant genotypic to phenotype predictions much more complicated, requiring more experimental information on the regulation of these operons and specificities of their receptor molecules. Transcriptional regulation of these pumps would likely explain the susceptibility to disinfectants such
This article is protected by copyright. All rights reserved.
Accepted Article
as Triclosan despite the presence of pumps associated with resistance. In addition to the efflux pumps, the presence of genes responsible for copper resistance, heavy metal resistance, and acriflavine resistance shows that this organism is heavily adapted to live in harsh environmental conditions such as waste water systems. Phage integration appears important for the carriage of some of these elements and indeed, metagenomic studies have previously shown the presence of various phagerelated resistance markers in activated sludge from wastewater treatment plants (29). The presence of a phage in some but not all of the ST111 isolates is difficult to interpret due to the small number of isolates sequenced, but may simply reflect the transient nature of phages within these genomes or may be due to phage loss in the isolation process.
One P. putida isolate was sequenced out of curiosity, because there has been speculation that this species could act as a reservoir for horizontal gene transfer. One element was identified within PA38182 RGP2 with a high level of nucleotide similarity (98%) to the P. putida genome; this sequence has been seen in other sequenced genomes of P. putida but no other P. aeruginosa sequences within the public databases, suggesting that this element may have been transferred within the context of this outbreak but equally it may just have transferred between species in the past and the parent strains of P. aeruginosa containing it have not yet been identified and sequenced. A second element with high levels of similarity to the P. putida sequence was identified within RGP75 of PA38182, but this has been previously sequenced in both P. putida and P. aeruginosa sequences in the public databases, suggesting any transfer between species may have occurred previously.
This article is protected by copyright. All rights reserved.
Accepted Article
In this study, isolates from two separate UK hospital outbreaks were sequenced and compared, in order to identify any genetic evidence for the difference in mortality rates; however, within known virulence elements, only the S121N change in ExoS could be found, a site identified previously in an investigation of ExoS virulence, but its role could not be conclusively determined (30). P. aeruginosa isolates studied here have all been shown to be XDR and sensitive to only one antibiotic, colistin, in addition to disinfectant agents. However, colistin resistance has been described in patients with cystic fibrosis (31,32), therefore it is essential to monitor this organism carefully and to clinically manage treatment appropriately.
Development of further resistance would result in an organism
refractory to antibiotic treatment within the hospital and which could also lead to loss of susceptibility to disinfectants making infection control and environmental management within the hospital difficult. We believe that routine WGS of all isolates within a hospital environment would define the gene pool for resistance enabling both rapid molecular epidemiological investigation of outbreaks and enable surveillance for antibiotic resistance, thus informing good antibiotic stewardship - a major international governmental imperative.
5. Methods Susceptibility testing, serotyping, PFGE and VNTR Serotyping, PFGE and VNTR typing were performed by the Health Protection Agency, Colindale, UK (now The Public Health England, Laboratory for Healthcare Associated Infection (LHCAI)). Susceptibility testing using BSAC disc diffusion methods was described previously (1).
This article is protected by copyright. All rights reserved.
Accepted Article
Susceptibility to various disinfectants/antiseptics was measured by standard methods (33) for P. aeruginosa by measuring the minimum inhibitory concentration (MIC). The test strain for assessing bactericidal efficacy of disinfectants, strain NCTC 6749 was included. Each disinfectant was serially diluted by doubling dilutions. Three replicates were performed at each dilution and the modal value was taken.
Isolate preparation and DNA preparation Isolates were obtained from several environmental and patient sources (Table 1). The sewer sample was isolated according to Moore et al (34), briefly a gauze swab was placed in a sewer outlet pipe fed from three buildings on the hospital site for 48 hours. Fragments of the swab were incubated in selective broth (Nutrient Broth, Oxoid, Basingstoke, UK) containing 5 mg/ml vancomycin, 10 mg/ml meropenem overnight at 37°C, followed by single colony selection and overnight growth on Columbia blood agar plates (Oxoid, Basingstoke, UK). Single colonies were selected and species verified using API NE (BioMerieux, Marcy l'Etoile, France). Isolates were incubated overnight under aerobic conditions at 37°C on Columbia blood agar plates. Single colonies were subsequently grown for a second time overnight under the same conditions before DNA extraction using the FastDNA SPIN Kit for soil (MP Biomedicals, USA).
DNA Sequencing Whole genome sequencing was performed using the Ion Torrent Personal Genome Machine (PGM) (Life Technologies, Paisley, UK). 5 µg of genomic DNA was fragmented by sonication using a BioRupter UD-200 and the library prepared using
This article is protected by copyright. All rights reserved.
Accepted Article
the Ion Fragment Library Kit (Life Technologies) according to the manufacturer’s instructions. Briefly, fragmented DNA was end-repaired, ligated to Ion-compatible adapters, size-selected for optimum insert length and then nick-translated and amplified. Size-selection was carried out using a 2% E-Gel SizeSelect Agarose Gel (Invitrogen, UK) to produce a library with a median fragment size of approximately 200bp. The DNA library was clonally amplified on Ion Sphere Particles™ (ISP) by emulsion PCR using the Ion Xpress Template kit, and sequenced on an Ion 316 chip using the Ion Sequencing Kit v 2.0 which generates an average read length of approximately 100 bp.
Sequence Analysis Sequence
quality
was
assessed
using
Fastqc
(http://www.bioinformatics.babraham.ac.uk/projects/fastqc/). Reads were trimmed to a maximum length of 150bp using seqtk (https://github.com/lh3/seqtk). Sequence assembly was performed using MIRA v3.9.4 (35) with default parameters for Ion Torrent data. Automated gene annotation was performed using Prokka v1.6 (http://www.vicbioinformatics.com/software.prokka.shtml).
Manual
analysis
and
sequence inspection were performed using the Artemis and ACT genome visualisation tools (36,37).
In silico serotype and MLST analyses The serotype was inferred in silico by similarity searching a database of O-antigen gene sequences of serotypes O1-O20 using BLASTN (38) (Genbank accessions: AC104719, AC104720, AC104721, AC104722, AC104723, AC104724, AC104725,
This article is protected by copyright. All rights reserved.
Accepted Article
AC104726, AC104727, AC104728, AC104729, AC104730, AC104731, AC104732, AC104733, AC104734, AC104735, AC104736, AC104737, AC104738, AC104739). MLST types were determined by a method involving mapping reads to the PAO1 reference genome (RefSeq: NC_002516), and then subsequent BLAST comparison of consensus sequences to the P. aeruginosa MLST database.
Phylogenetic Analysis Whole genome phylogenetic reconstruction was performed as previously described (39); genomes were aligned with Mugsy (40), core genome sequences extracted with mothur (41) and maximum likelihood inference performed using RAxML v7.4.2 (42) with a GTR model of nucleotide substitution and a GAMMA model of rate heterogeneity, branch support values were determined using 1000 bootstrap replicates.
SNP based phylogenetic reconstruction was performed using a method similar to (22). Briefly, reads were mapped to the LESB58 reference genome (RefSeq: NC_011770; (43)) using TMAP v3.4.0 (Ion Torrent, USA) and alignments sorted with Picard v1.76 (http://picard.sourceforge.net/). Per base alignment statistics were generated with samtools mpileup (v0.1.18) (44) and variants called using bcftools v0.1.17-dev. Variant sites were filtered based on the following criteria; mapping quality (MQ) above 30, site quality score (QUAL) above 30, at least 4 reads covering each site with at least 2 reads mapping to each strand but with maximum depth of coverage 200, at least 75% of reads supporting site (DP4), allelic frequency (AF1) of 1. Sites which failed these criteria in any strain were removed from the analysis.
This article is protected by copyright. All rights reserved.
Accepted Article
Also, sites corresponding to known phage or genomic islands were removed. Phylogenetic reconstruction was performed using RAxML v7.4.2 (42) as above and branch SNP counts estimated using ancestral sequence reconstruction performed with PAML v4 (45).
6. Sequence data Sequence data has been deposited in the European Nucleotide Archive with study accession number PRJEB4573.
7. Acknowledgements: TP and CP thank St George’s Healthcare NHS Trust Charitable Trustees for funding to cover the costs of sequencing. The authors thank Public Health England for the VNTR, PFGE and serotyping data. The authors declare no conflicts of interest.
8. References: 1.
Breathnach AS, Cubbon MD, Karunaharan RN, Pope CF, Planche TD. Multidrug-resistant Pseudomonas aeruginosa outbreaks in two hospitals: association with contaminated hospital waste-water systems. Journal of Hospital Infection. 2012 Sep;82(1):19–24.
2.
Magiorakos A-P, Srinivasan A, Carey RB, Carmeli Y, Falagas ME, Giske CG, et al. Multidrug-resistant, extensively drug-resistant and pandrug-resistant bacteria: an international expert proposal for interim standard definitions for acquired resistance. Clinical Microbiology and Infection. 2012;18(3):268–81.
3.
Turton JF, Matos J, Kaufmann ME, Pitt TL. Variable number tandem repeat loci providing discrimination within widespread genotypes of Acinetobacter baumannii. Eur J Clin Microbiol Infect Dis. 2009 May 1;28(5):499–507.
This article is protected by copyright. All rights reserved.
Accepted Article
4.
Van der Bij AK, Van der Zwan D, Peirano G, Severin JA, Pitout JDD, Van Westreenen M, et al. Metallo-β-lactamase-producing Pseudomonas aeruginosa in the Netherlands: the nationwide emergence of a single sequence type. Clinical Microbiology and Infection. 2012;18(9):E369–E372.
5.
Juan C, Zamorano L, Mena A, Albertí S, Pérez JL, Oliver A. Metallo-betalactamase-producing Pseudomonas putida as a reservoir of multidrug resistance elements that can be transferred to successful Pseudomonas aeruginosa clones. The Journal of antimicrobial chemotherapy. 2010 Mar;65(3):474–8.
6.
Maiden MCJ, Bygraves JA, Feil E, Morelli G, Russell JE, Urwin R, et al. Multilocus sequence typing: A portable approach to the identification of clones within populations of pathogenic microorganisms. Proc Natl Acad Sci U S A. 1998 Mar 17;95(6):3140–5.
7.
Roy PH, Tetu SG, Larouche A, Elbourne L, Tremblay S, Ren Q, et al. Complete genome sequence of the multiresistant taxonomic outlier Pseudomonas aeruginosa PA7. PloS one. 2010 Jan;5(1):e8842.
8.
Jalal S, Wretlind B. Mechanisms of quinolone resistance in clinical strains of Pseudomonas aeruginosa. Microb Drug Resist. 1998;4(4):257–61.
9.
Samuelsen O, Toleman M a, Sundsfjord A, Rydberg J, Leegaard TM, Walder M, et al. Molecular epidemiology of metallo-beta-lactamase-producing Pseudomonas aeruginosa isolates from Norway and Sweden shows import of international clones and local clonal expansion. Antimicrobial agents and chemotherapy. 2010 Jan;54(1):346–52.
10. Mima T, Joshi S, Gomez-Escalada M, Schweizer HP. Identification and Characterization of TriABC-OpmH, a Triclosan Efflux Pump of Pseudomonas aeruginosa Requiring Two Membrane Fusion Proteins. J Bacteriol. 2007 Nov;189(21):7600–9. 11. Chuanchuen R, Narasaki CT, Schweizer HP. The MexJK Efflux Pump of Pseudomonas aeruginosa Requires OprM for Antibiotic Efflux but Not for Efflux of Triclosan. J Bacteriol. 2002 Sep;184(18):5036–44. 12. Leedjarv A, Ivask A, Virta M. Interplay of Different Transporters in the Mediation of Divalent Heavy Metal Resistance in Pseudomonas putida KT2440. J Bacteriol. 2008 Apr;190(8):2680–9. 13. Teitzel GM, Geddie A, De Long SK, Kirisits MJ, Whiteley M, Parsek MR. Survival and Growth in the Presence of Elevated Copper: Transcriptional Profiling of Copper-Stressed Pseudomonas aeruginosa. J Bacteriol. 2006 Oct;188(20):7242–56. This article is protected by copyright. All rights reserved.
Accepted Article
14. Mima T, Kohira N, Li Y, Sekiya H, Ogawa W, Kuroda T, et al. Gene cloning and characteristics of the RND-type multidrug efflux pump MuxABC-OpmB possessing two RND components in Pseudomonas aeruginosa. Microbiology. 2009 Nov 1;155(11):3509–17. 15. Muller JF, Stevens AM, Craig J, Love NG. Transcriptome Analysis Reveals that Multidrug Efflux Genes Are Upregulated To Protect Pseudomonas aeruginosa from Pentachlorophenol Stress. Appl Environ Microbiol. 2007 Jul;73(14):4550– 8. 16. Moskowitz SM, Ernst RK, Miller SI. PmrAB, a Two-Component Regulatory System of Pseudomonas aeruginosa That Modulates Resistance to Cationic Antimicrobial Peptides and Addition of Aminoarabinose to Lipid A. J Bacteriol. 2004 Jan;186(2):575–9. 17. Sun S, Negrea A, Rhen M, Andersson DI. Genetic Analysis of Colistin Resistance in Salmonella enterica Serovar Typhimurium. Antimicrob Agents Chemother. 2009 Jun;53(6):2298–305. 18. Beceiro A, Llobet E, Aranda J, Bengoechea JA, Doumith M, Hornsey M, et al. Phosphoethanolamine Modification of Lipid A in Colistin-Resistant Variants of Acinetobacter baumannii Mediated by the pmrAB Two-Component Regulatory System▿. Antimicrob Agents Chemother. 2011 Jul;55(7):3370–9. 19. Mathee K, Narasimhan G, Valdes C, Qiu X, Matewish JM, Koehrsen M, et al. Dynamics of Pseudomonas aeruginosa genome evolution. Proceedings of the National Academy of Sciences of the United States of America. 2008 Feb;105(8):3100–5. 20. Gutiérrez-Barranquero JA, Vicente A de, Carrión VJ, Sundin GW, Cazorla FM. Recruitment and Rearrangement of Three Different Genetic Determinants into a Conjugative Plasmid Increase Copper Resistance in Pseudomonas syringae. Appl Environ Microbiol. 2013 Feb 1;79(3):1028–33. 21. Battle SE, Rello J, Hauser AR. Genomic islands of Pseudomonas aeruginosa. FEMS microbiology letters. 2009 Jan;290(1):70–8. 22. Harris SR, Cartwright EJ, Török ME, Holden MT, Brown NM, Ogilvy-Stuart AL, et al. Whole-genome sequencing for analysis of an outbreak of meticillinresistant Staphylococcus aureus: a descriptive study. The Lancet Infectious Diseases. 2012 Nov;3099(12):1–7. 23. Köser CU, Ellington MJ, Cartwright EJP, Gillespie SH, Brown NM, Farrington M, et al. Routine Use of Microbial Whole Genome Sequencing in Diagnostic and Public Health Microbiology. Rall GF, editor. PLoS Pathogens. 2012 Aug;8(8):e1002824. This article is protected by copyright. All rights reserved.
Accepted Article
24. Köser CU, Holden MTG, Ellington MJ, Cartwright EJP, Brown NM, Ogilvy-Stuart AL, et al. Rapid whole-genome sequencing for investigation of a neonatal MRSA outbreak. New England Journal of Medicine. 2012;366(24):2267–75. 25. Curran B, Jonas D, Grundmann H, Pitt T, Dowson CG. Development of a Multilocus Sequence Typing Scheme for the Opportunistic Pathogen Pseudomonas aeruginosa. J Clin Microbiol. 2004 Dec;42(12):5644–9. 26. Slekovec C, Plantin J, Cholley P, Thouverez M, Talon D, Bertrand X, et al. Tracking Down Antibiotic-Resistant Pseudomonas aeruginosa Isolates in a Wastewater Network. PLoS One [Internet]. 2012 Dec 19 [cited 2013 Jul 23];7(12). Available from: http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3526604/ 27. Pirnay J, Bilocq F, Pot B, Cornelis P, Zizi M, Eldere JV, et al. Pseudomonas aeruginosa Population Structure Revisited. Population (English Edition). 2009;4(11). 28. Akasaka T, Tanaka M, Yamaguchi A, Sato K, Hemother ANAGC. Type II Topoisomerase Mutations in Fluoroquinolone-Resistant Clinical Strains of Pseudomonas aeruginosa Isolated in 1998 and 1999 : Role of Target Enzyme in Mechanism of Fluoroquinolone Resistance. Society. 2001;45(8):2263–8. 29. Parsley LC, Consuegra EJ, Kakirde KS, Land AM, Harper WF, Liles MR. Identification of Diverse Antimicrobial Resistance Determinants Carried on Bacterial, Plasmid, or Viral Metagenomes from an Activated Sludge Microbial Assemblage. Appl Environ Microbiol. 2010 Jun;76(11):3753–7. 30. Ferguson MW, Maxwell JA, Vincent TS, da Silva J, Olson JC. Comparison of the exoS Gene and Protein Expression in Soil and Clinical Isolates of Pseudomonas aeruginosa. Infection and Immunity. 2001 Apr 1;69(4):2198–210. 31. Schülin T. In vitro activity of the aerosolized agents colistin and tobramycin and five intravenous agents against Pseudomonas aeruginosa isolated from cystic fibrosis patients in southwestern Germany. J Antimicrob Chemother. 2002 Feb 1;49(2):403–6. 32. Clifton I, Peckham D, Conway S. 10 years of use of colistin in a regional cystic fibrosis centre. Journal of Cystic Fibrosis Medicine. 2005;202:S54. 33. McDonnel GE. Antisepsis, Disinfection, and Sterilization: Types, Action, and Resistance. ASM Press; 2007. 378 p. 34. Moore B, Perry EL, Chard ST. A survey by the sewage swab method of latent enteric infection in an urban area. J Hyg (Lond). 1952 Jun;50(2):137–56.
This article is protected by copyright. All rights reserved.
Accepted Article
35. Chevreux B, Wetter T, Suhai S. Genome Sequence Assembly Using Trace Signals and Additional Sequence Information. Computer Science and Biology: Proceedings of the German Conference on Bioinformatics. 1999;99:45–56. 36. Carver TJ, Rutherford KM, Berriman M, Rajandream M-A, Barrell BG, Parkhill J. ACT: the Artemis comparison tool. Bioinformatics. 2005 Aug 15;21(16):3422–3. 37. Rutherford K, Parkhill J, Crook J, Horsnell T, Rice P, Rajandream M-A, et al. Artemis: sequence visualization and annotation. Bioinformatics. 2000 Oct 1;16(10):944–5. 38. Altschul SF, Madden TL, Schaffer AA, Zhang J, Zhang Z, Miller W, et al. Gapped BLAST and PSI-BLAST: a new generation of protein database search programs. Nucleic Acids Res. 1997 Sep 1;25(17):3389–402. 39. McNally A, Cheng L, Harris SR, Corander J. The Evolutionary Path to Extraintestinal Pathogenic, Drug-Resistant Escherichia coli Is Marked by Drastic Reduction in Detectable Recombination within the Core Genome. Genome Biol Evol. 2013;5(4):699–710. 40. Angiuoli SV, Salzberg SL. Mugsy: fast multiple alignment of closely related whole genomes. Bioinformatics. 2011 Feb 1;27(3):334–42. 41. Schloss PD, Westcott SL, Ryabin T, Hall JR, Hartmann M, Hollister EB, et al. Introducing mothur: Open-Source, Platform-Independent, CommunitySupported Software for Describing and Comparing Microbial Communities. Appl Environ Microbiol. 2009 Dec;75(23):7537–41. 42. Stamatakis A. RAxML-VI-HPC: maximum likelihood-based phylogenetic analyses with thousands of taxa and mixed models. Bioinformatics (Oxford, England). 2006 Nov;22(21):2688–90. 43. Winstanley C, Langille MGI, Fothergill JL, Kukavica-Ibrulj I, Paradis-Bleau C, Sanschagrin F, et al. Newly introduced genomic prophage islands are critical determinants of in vivo competitiveness in the Liverpool Epidemic Strain of Pseudomonas aeruginosa. Genome research. 2009 Jan;19(1):12–23. 44. Li H, Handsaker B, Wysoker A, Fennell T, Ruan J, Homer N, et al. The Sequence Alignment/Map format and SAMtools. Bioinformatics. 2009 Aug 15;25(16):2078–9. 45. Yang Z. PAML 4: Phylogenetic Analysis by Maximum Likelihood. Mol Biol Evol. 2007 Aug 1;24(8):1586–91.
This article is protected by copyright. All rights reserved.
Accepted Article
Table 1. Pseudomonas spp. isolates sequenced in this study.
Strain name
Source1
1 PA38182
Patient blood2
A 14 2011
Feb 11, 3, 4, 3, 2, 2, O12 5, 4, 7
ST1 11
17,5,5,4,4 ,4,3
2 PA48922
Patient sink2
A 01 2011
Mar 11, 3, 4+, 3, 2, 2, O124 -, 4, 7
ST1 11
17,5,5,4,4 ,4,3
3 PP48922
Patient sink2
A 01 2011
Mar n/a (P. putida)
n/a
4 PA46218
Patient sink2
A 21 2011
Feb 11, 3, 4+, 3, 2, 2, O124 5, 4, 11
ST1 11
17,5,5,4,4 ,4,3
5 PASwab
Sewer2
06 2011
Oct 11, 3, 4, 3, 2, 2, O124 5, 4, 7
ST1 11
17,5,5,4,4 ,4,3
6 PA128572
Patient blood2
B 12 2006
Aug 11, 3, 4, 3, 2, 2, O124 5, 4, 7
ST1 11
17,5,5,4,4 ,4,3
7 PA99736
Patient blood2
C 02 2011
May 11, 4, 5, 2, 3, 2, - O64 , 4, 12
ST1 55
28,5,36,3, 3,13,7
8 PA97498
Patient blood3
02 2010
Aug 11, 3, 2, -, 3, 1, 9, 3, 10
5
ST6 54
17,5,26,3, 4,4,26
9 PA156066
Patient bathroom3
04 2011
May 11, 3, 2, -, 3, 1, 9, 3, 10
5
ST6 54
17,5,26,3, 4,4,26
Date
VNTR-type
1
Seroty pe
n/a
MLS MLST profile6 T4
See section 3.1 for isolate descriptions; 2 Hospital 1; 3 Hospital 2; 4 inferred from sequence data; 5 undeterminable from sequence data; 6 MLST genes are: acs, aro, gua, mut, nuo, pps, trp
This article is protected by copyright. All rights reserved.
Accepted Article
Table 2A. Minimum inhibitory concentrations of antibiotics for a representative isolate of ST111 (isolate PA38182).
Antibiotic Class Antibiotic
MIC (mg/L)
Ciprofloxacin
>8
Moxifloxacin
>32
Amikacin
64
Gentamycin
>32
Tobramycin
>32
Monobactam
Azetreonam
16
Cephalosporin
Ceftazidine
16
Carbapenems
Imipenem
>128
Meropenem
>32
Piperacillin
64
Fluoroquinolone
Aminoglycoside
Penicillin Penicillin inhibitor
/
β-lactamase Piperacillin-tazobactam
64
Polymyxin
Colistin
2
Carboxypenicillin
Ticaricillin / clavulanate
64
Tetracycline
Minocycline
32
Glycylcycline
Tigecycline
16
/ β-lactamase inhibitor
This article is protected by copyright. All rights reserved.
Accepted Article
Table 2B. Minimum inhibitory concentrations of disinfectants. Isolates PA38182, PA128572 and PASwab were tested. In addition, the standard test strain for assessing bactericidal efficacy of disinfectants, strain NCTC 6749 was included.
Disinfectant Povidone
Initial concentration
MIC obtained NCTC 6749
PA128572
PA38182
PASwab
10%
1 in 8
1 in 8
1 in 8
1 in 8
4%
1 in 4
1 in 4
1 in 4
1 in 4
1000 ppm
1 in 2
1 in 2
1 in 2
1 in 2
2%
1 in 8
1 in 8
1 in 8
1 in 4
1100 ppm
1 in 8
1 in 8
1 in 16
1 in 16
1500 ppm
1 in 32
1 in 32
1 in 64
1 in 64
1%
1 in 2
1 in 2
1 in 2
1 in 2
iodine Chlorhexidine NaDCC Glutaraldehyde Chlorine dioxide Peracetic acid Triclosan
This article is protected by copyright. All rights reserved.
Accepted Article
Figure 1A. Phylogenetic relatio onship between P. aeruginosa isolates seque enced in this study in the context of 35 geno ome sequences from public databases (see su upplementary Table 2 for accessions); based on core genome alignments. Sequence derive ed MLST type included in square brackets. PA7 was excluded as an outlier. 1) Outbreak 1 strains, 2) Outbreak 2 strains, 3) unrelated strain from Hospital 1.
1
2
3
This article is protected by copyright. All rights reserved.
Accepted Article
Figure 1B. Phylogenetic reconstruction from 24 ST111-specific variant sites identified by comparing isolate read data with h the reference genome LESB58 (RefSeq: NC_ _011770). The tree is rooted to NC_011770. Asterisks marks branch with 100% bootstrap su upport values. Ancestral site reconstruction was w used to estimate the number of SNP changes on branches; branches with significant support values are labelled with the estimatted number of SNPs out of the total number of SNP sites identified when mapped against LESB58, PAO1, UCBPP-PA14 (RefSeq: NC_011 1770, NC_002516, NC_008463) respectively. T The scale bar indicates estimated mutations per p site. +/- refers to presence/absence of pu utative phage (see Figure 2). See section 3.1 for isolate descriptions.
[6yo strain] +
[Patient A sin nk] +
*
[Patien nt A sink] +
[Patient A] -
[Sewer swab] -
This article is protected by copyright. All rights reserved.
Accepted Article
Figure 2. Read coverage acrross differential phage region within ST111 isolates. A) predicted genes, putative pha age genes coloured red; B) PA128572; C C) PA48922; D)PA46218; E) PA38182; F) PASwab.
A B
C
D E
F
This article is protected by copyright. All rights reserved.
Received Date: 07-Oct-2013 Revised Date: 20-Dec-2013 Accepted Date: 07-Jan-2014 Article Type: Original Article - E-only Category: Bacteriology
Running Title: Pseudomonas aeruginosa ST111 outbreak strain genome
Title: Genome sequencing and characterization of an XDR ST111 serotype O12 hospital outbreak strain of Pseudomonas aeruginosa
Adam A. Witney1, Katherine A. Gould1, Cassie F. Pope2, Frances Bolt2, Neil G. Stoker1, Marc D. Cubbon3, Christina R. Bradley4. Adam Fraise4, Aodhán S. Breathnach2, Philip D. Butcher1, Tim D. Planche1,2, Jason Hinds1.
1
Division of Clinical Sciences, St George’s University of London, Cranmer Terrace,
SW17 0RE. 2
Department of Microbiology, St George’s Healthcare NHS Trust, Blackshaw Road,
London, SW17 0QT. 3
Brighton and Sussex University Hospitals NHS Trust, Brighton, UK.
This article has been accepted for publication and undergone full peer review but has not been through the copyediting, typesetting, pagination and proofreading process, which may lead to differences between this version and the Version of Record. Please cite this article as doi: 10.1111/1469-0691.12528 This article is protected by copyright. All rights reserved.
Accepted Article
4
Hospital Infection Research Laboratory, Queen Elizabeth Hospital Birmingham, UK.
Corresponding author: Dr Adam A Witney, Division of Clinical Sciences, St George’s, University of London, Cranmer Terrace, London, SW17 0RE. UK. Tel: +44 208 725 0698 Email: [email protected]
1. Abstract A series of extensively drug-resistant (XDR) isolates of Pseudomonas aeruginosa from two outbreaks in UK hospitals were characterized by whole genome sequencing (WGS). Although these isolates were resistant to antibiotics other than colistin, we confirmed that they are still sensitive to disinfectants. The sequencing confirmed that isolates in the larger outbreak were serotype O12 and further revealed that they belonged to the ST111 sequence type which is a major epidemic strain of P. aeruginosa throughout Europe. As this is the first reported sequence of an ST111 strain, the genome was examined in depth, focusing particularly on antibiotic resistance and potential virulence genes, and on the reported regions of genome plasticity. High degrees of sequence similarity were discovered between outbreak isolates collected from recently infected patients, sinks, an isolate from the sewer and an historical isolate, suggesting that the ST111 strain has been endemic in the hospital for many years. The ability to translate easily from outbreak
This article is protected by copyright. All rights reserved.
Accepted Article
investigation to detailed genome biology using the same data demonstrates the flexibility of WGS application in a clinical setting.
2. Introduction An outbreak investigation into the spread of multidrug-resistant Pseudomonas aeruginosa susceptible only to colistin, in two UK hospitals was previously described (1). Recent guidelines redefine these isolates as XDR (2), due to susceptibility to only one class of antibiotic (colistin). The outbreaks occurred in university teaching hospitals in the South of England, one in London (outbreak 1) and one on the South coast (outbreak 2). In outbreak 1, 85 hospital-wide cases were identified between 2005 and 2011, and during outbreak 2, four cases occurred in a single ward between 2009 and 2010. An environmental investigation was conducted to identify potential sources of infection and inform control measures that might reduce spread.
Outbreak 1 consisted of sporadic outbreaks separated by long infection-free periods. This outbreak was caused by a serotype O12 strain that caused severe infection, usually in immune-compromised patients. As reported previously (1) the overall case fatality during outbreak 1 was 34/85 (40%); in patients with bacteraemia the case fatality was as high as 14/18 (78%). Haematology patients were more likely to have bacteraemia (6/7 cases). During outbreak 2 all cases had bacteraemia and there were no deaths attributed to XDR P. aeruginosa. The potential of waste water systems to act as a reservoir for XDR P. aeruginosa was reported in both outbreaks. The UK reference laboratory reported the strain associated with outbreak 1 as being a unique type confined to hospital 1 using Variable Number Tandem Repeat (VNTR)
This article is protected by copyright. All rights reserved.
Accepted Article
analysis (3). However, genome sequencing revealed the outbreak strain was sequence type ST111, one of a small number of major epidemic clones that are widespread in Europe, and for example was recently reported as the major metalloβ-lactamase-producing P. aeruginosa in the Netherlands (4). We describe the clinical application of whole-genome sequencing (WGS) in these two hospital outbreaks to discover the relationships between circulating patient and environmental strains that may inform infection control measures. Furthermore, greater understanding of the biology of the XDR strain gained from the genome sequence may provide useful insights into the pathogenicity and guide future treatment or disinfectant options. To our knowledge, this is the first report of a genome sequence of an ST111 or an XDR strain of P. aeruginosa. We describe mutations and elements associated with antibiotic resistance and compare the genome sequence to a collection of clinical and environmental isolates from both outbreaks and other published genomes.
3. Results 3.1 Isolate phenotyping All isolates of P. aeruginosa that were collected from outbreak 1 from June 2005 onwards were resistant to carbapenems. These isolates were sent for typing at a reference laboratory and 74/85 were all reported as serotype O12, PCR-positive for the blaVIM-2 β-lactamase gene, and identical or very similar by PFGE and/or VNTR typing (1). The other 11 isolates were all different and deemed to be unrelated sporadic cases of different resistant P. aeruginosa strains. A small number were single-locus variants (SLVs) by VNTR typing, which we considered as most likely
This article is protected by copyright. All rights reserved.
Accepted Article
being the same strain with minor sequence variations leading to an SLV. The reference laboratory also reported this VNTR type as being unique to this hospital. In order to characterize this strain type, PA38182 (designated as the primary isolate because it was from a current patient (Patient A)), as well as eight control samples were examined further (see Table 1).
•
Four environmental isolates from outbreak 1 were included: two P. aeruginosa isolates (PA48922, PA46218) from patient A's sink (isolated 3 weeks apart), a P. putida isolate (PP48922) that was isolated at the same time from the sink, as there is evidence that P. putida can act as a reservoir of mobile elements for P. aeruginosa (5), and a P. aeruginosa isolate obtained from the sewer using a Moore's swab (PASwab).
•
One P. aeruginosa isolate (PA128572) of the same VNTR-type isolated from a patient (Patient B) in Hospital 1 six years earlier.
•
One XDR PA isolate (PA99736) from Patient C at the same hospital with a different VNTR-type.
•
Two XDR P. aeruginosa isolates (PA97498, PA156066) from outbreak 2 - one clinical and one environmental.
These isolates were characterized by VNTR (Table 1), and MIC data for the outbreak 1 representative isolate PA38182 is shown in Table 2A. As these isolates were susceptible to only one class of antibiotics, their susceptibility to disinfectants is critical for infection control purposes. We therefore carried out tests with a variety of disinfectants (Table 2B). These showed that the isolates from
This article is protected by copyright. All rights reserved.
Accepted Article
outbreak 1 were similar in sensitivity to a susceptible control strain, with all being sensitive at practical concentrations.
3.2 Sequence analysis WGS was carried out using an Ion Torrent Personal Genome Machine generating between 0.5 - 2 million reads per sample, with an average length of 120 bp. Sequence data were analysed both by reference mapping and de novo assembly approaches. The genome sequence-inferred serotypes were determined by BLAST comparison against publicly available O-antigen gene sequences (see Methods). An identity of greater than 99% across the entire 25kb O12 locus confirmed the reference laboratory's designation of the index isolate (PA38182) as serotype O12, in addition to all other isolates related to the outbreak, including the six year old isolate. The unrelated isolate from Hospital 1 (PA99736) was inferred as serotype O6. Multi-locus sequence typing (MLST) has been used to type P. aeruginosa strains across the world (6), and has the advantage of ease of data comparison between studies. MLST derived from genome sequence identified the outbreak cluster isolates from Hospital 1 as ST111, whereas the VNTR unrelated isolate (PA99736) was ST155, and the two Hospital 2 isolates were ST654 (Table 1).
3.3 Phylogenetic reconstruction In order to determine epidemiological links between the isolates, the phylogenetic relatedness of the isolates was determined in two ways: using core genome sequence alignment and SNP-based reconstruction. Whole genome phylogenetic This article is protected by copyright. All rights reserved.
Accepted Article
reconstruction was performed using the core genomes of the isolates sequenced here in the context of 35 genomes available in the public databases (Figure. 1A). The ST111 and ST654 isolates were more related to each other than other sequences in the public databases; however the number of publicly available P. aeruginosa genome sequences available is limited and the strains diverse. The ST155 isolate clustered tightly with other ST155 strains from China. The previously sequenced serotype O12 strain (PA7) is a taxonomic outlier (7) and is significantly diverged from the ST111 O12 isolates sequenced in this study (Supplementary Figure 1).
Therefore PA7 was removed from the tree, as was an unpublished
sequence similar to PA7, VRFPA01 (Genbank: AOBK01000001.1) in order to aid visualisation of the remaining relationships.
In order to investigate further the relationships between the ST111 isolates, SNP sites were identified with respect to a reference sequence. Sites where all isolate sequences were identical but differed only from the reference were ignored, leaving only variant sites between the isolates. Phylogenetic reconstruction of the ST111 outbreak isolates based on 24 variant sites identified by mapping to the LESB58 reference genome (RefSeq: NC_011770; several reference sequences were used independently due to the lack of a previously sequenced ST111 reference and showed a consistent relationship (data not shown)) suggested that the sewer isolate is very closely related to both the clinical and the associated sink isolates collected (Figure 1B). Also, the similarity of the historical isolate (PA128572), suggests that this strain has been present in the hospital for at least 6 years.
This article is protected by copyright. All rights reserved.
Accepted Article
3.4 Genome analysis of XDR ST111 isolates Sequence reads of the XDR ST111 isolates were both mapped to other published sequences and assembled de novo. The isolates were highly similar, but the invasive clinical isolate PA38182 was used as the primary isolate for ST111 in the sequence analysis. For this isolate 460 contigs (> 500bp) totalling 7 Mb were obtained, and no major deletions compared to the reference strains were identified except those accessory elements detailed below. Despite being a global epidemic strain of P. aeruginosa, there is no previously reported genome sequence of an ST111 strain. Therefore a detailed study of the PA38182 genome content was conducted, focusing particularly on antibiotic resistance and virulence elements. 3.4.1 Antibiotic Resistance The MICs in Table 2A show that ST111 is clinically resistant to most antibiotics. Analysis of the genome sequence of ST111 identifies key genetic mutations and the presence of multiple genes implicated in resistance to a wide range of pharmaceutical agents. 3.4.1.1 Quinolones Amino acid changes in the gyrA (Thr83Ile) and parC (Ser87Leu) genes of ST111 are most likely responsible for the resistance to the fluoroquinolones as described previously (8). 3.4.1.2 Aminoglycosides Several aminoglycoside resistance encoding genes were identified; ST111 contains a pair of aacA29a / aacA29b genes (aminoglycoside-6’N-acetyltransferase), most likely surrounding the blaVIM-2 gene, as described previously (9). Also, an aminoglycoside phosphotransferase (aph) IIb, a FadE36 aph and a further putative
This article is protected by copyright. All rights reserved.
Accepted Article
aph gene were identified. The ST654 isolates did not contain the aacA29a / aacA29b genes but did contain copies of the same aph genes as ST111; furthermore it also contained an ant(4')-IIb gene and two genes annotated as associated with streptomycin resistance. Isolate ST155 has limited specific aminoglycoside specific genes containing copies of only the aph-IIb and the putative aph described above. 3.4.1.3 β-lactamases The ubiquitous class C β-lactamase gene, ampC, showed no apparent mutations, neither did ampG. However, mutations in the ampC regulator genes, ampD (G148A) and ampR (E283G), were present but it is not known what role these may play. Class B metallo-β-lactamase (MBL) genes, providing resistance to β-lactams except monobactams, are commonly located on mobile elements.
ST111 isolates in
Norway and Sweden were reported to have both the blaVIM-2 MBL and associated aminoglycoside acetyltransferase genes (aacA29a / aacA29b) on a single cassette (9); our ST111 isolate contains the same genes, however we were unable to deduce the arrangement from the assembly, likely due to the high level of identity between aacA29a and aacA29b flanking the blaVIM-2 gene, causing contig breaks either side of the blaVIM-2 gene. The ST654 isolates described in (9) had a cassette carrying blaVIM-2 MBL, the strA and strB streptomycin resistance genes, and aadB but no aacA29a or aacA29b genes. This is consistent with our ST654 sequence except no aadB gene was found, suggesting a variant in this isolate. 3.4.1.4 Efflux pumps Efflux pumps are potentially responsible for resistance to a wide range of agents in P. aeruginosa, and indeed the ST111 isolate contains copies of the mexAB-oprM,
This article is protected by copyright. All rights reserved.
Accepted Article
mexCD-oprJ, mexEF-oprN, mexHI-opmD operons and their respective regulator genes mexR, nfxB, mexT and mexG. Additionally mexMN, mexVW and mexXY sequences were identified. More specialised efflux pumps were also identified. These included triABC (10) and mexJKL operons (11), shown to be associated with triclosan resistance, and the czcCBA operon, shown to confer heavy metal resistance (12). These appeared to be contained within integrated phages or integrons. Additionally the copper-induced mexPQ-opmE efflux pump (13) was present. Other P. aeruginosa efflux systems identified were muxABC-opmB (14), emrAB-opmG (15) and an SMR (small multidrug resistance family of proteins) that confer resistance to a wide variety of quaternary ammonium compounds through a proton-drug efflux antiport mechanism. 3.4.1.5 Colistin All these isolates are currently sensitive to colistin (polymyxin E). Initial studies into the genetic basis of colistin resistance have implicated the 2-component systems pmrAB and phoPQ genes in P. aeruginosa, Salmonella enterica serovar Typhimurium and Acinetobacter baumannii (16–18), presumably by altering expression of the genes that are directly responsible. Our studies confirmed a lack of alterations in either the pmrAB or phoPQ genes in the isolates studied here. 3.4.2 Virulence Factors Comparison to the Virulence Factor Database (http://www.mgc.ac.cn/VFs/) shows that most of the common virulence factors are present in the ST111 isolates studied here. Notably the type III secretion system equivalent to PA1690-PA1725 locus in PAO1 is present along with the genes for the effectors, exoenzymes ExoS and ExoT and adenylate cyclase ExoY. (ST654 carries the same effectors but ST155 is
This article is protected by copyright. All rights reserved.
Accepted Article
missing the exoY gene). The ST111 ExoS protein has an S121N change relative to published sequences. The Type IV pilus biosynthesis cluster is present as is the associated twitching motility protein locus. 3.4.3 The ST111 Accessory Genome Studies of the P. aeruginosa genome show it to have a conserved core genome, with variable regions scattered throughout. These regions have been referred to as Regions of Genome Plasticity (RGPs) (19) by including 'any region of at least four contiguous ORFs that are missing in at least one of the genomes analyzed'. While some of these RGPs appear to be rare deletions, others appear to be sites where frequent changes occur, through additions and losses within the region, or through replacement of the entire region. Some of the genes or sets of genes can involve potential virulence genes, and include phages, integrons carrying Type IV secretion systems, proteins encoding resistance to heavy metals, organic solvents or acriflavin, or efflux pumps.
Supplementary Table 1 includes a detailed analysis of the accessory genome in isolate PA38182. Several accessory elements appeared novel to P. aeruginosa as shown by BLAST similarity to the non-redundant database at NCBI: • RGP2 contains an araC family regulator gene as part of a ~5kb region which has no BLAST hits to any P. aeruginosa sequences in NCBI’s non-redundant database, but shows 98% similarity to the P. putida genome sequenced in this study. • RGP5 contains an integrated conjugative element including a type II secretion system and a DNA topoisomerase III gene, both also novel to P. aeruginosa.
This article is protected by copyright. All rights reserved.
Accepted Article
• Other integrated conjugative elements containing Trb conjugal protein loci were also found in RGP26, RGP36 and RGP75, the latter two also containing the copG gene, a putative metal transporting P-type ATPase, thought to confer copper resistance (20). Furthermore, the element contained within RGP75 has been sequenced in P. aeruginosa strain NCGM2.S1 previously, as well as having 99% identity to the P. putida sequenced in this study. • Other examples of accessory elements carrying resistance genes can be seen in RGP29, where a phage like element contains the czcCBA heavy metal efflux pump system and RGP56 contains the putative multidrug-efflux transporter also equivalent to PLES_22791 in LESB58. • A Type IV pilus operon can be found in RGP41, which has a high level of similarity to the that found in RGP7 of PA7 (7). • RGP42 is the site for another novel phage, this containing a zona occludens toxin highly similar to PA0726, found in RGP5 of PAO1. • Comparing to previously described genomic islands, evidence of complete or remnant PAGI-7 (RGP88), PAGI-8 (RGP86) and PAGI-9 (RGP89) sequences was also seen (21). All the ST111 isolates sequenced in this study carried the same accessory components except for the presence of a phage in RGP87, which was variably present in the isolates; only the historical clinical isolate PA128572 and the recent environmental isolates PA48922 and PA46218 contained the phage (Figure 2). However, no obvious phenotypic association to the presence of this phage could be identified.
This article is protected by copyright. All rights reserved.
Accepted Article
4. Discussion Whole genome sequencing has been demonstrated to be a very powerful tool in diagnostic microbiology (22–24) with the potential to provide results in a clinically relevant time frame. This is especially true for epidemiological typing which can guide infection control investigation of outbreaks. In this study, WGS was applied to a selection of isolates of P. aeruginosa that were collected during an investigation into two hospital outbreaks (1). We have shown that in these outbreaks, the environmental and patient isolates of P. aeruginosa that are causing significant morbidity are highly related within each hospital. This provides further evidence of the potential of the waste water system acting as a reservoir for infection.
Identification of ST111 The isolates from outbreak 1 had previously been shown to be closely related by VNTR typing and were thought to be specific to the originating hospital. However, WGS analysis showed that the strain belongs to the global epidemic type ST111, which is thought to be responsible for many of the MDR P. aeruginosa cases around the world (4,9). We report the first full genome sequence of an ST111 strain. ST111 isolates have been identified in the UK previously by MLST (25), however MLST typing is not routinely performed on incoming isolates by the UK reference laboratory. This demonstrates that VNTR typing has limited utility in isolation without comparison to MLST and that sequence-based typing schemes must be adopted to use in tandem with or replacing VNTR typing.
This article is protected by copyright. All rights reserved.
Accepted Article
Sequencing selected genes from strain PA7, which is also serotype O12, has previously shown that it is significantly divergent from the worldwide MDR O12 clone (27). Sequencing of the O12 ST111 isolates in this study provides more complete genomic evidence of how dissimilar these two O12 strain types are.
Water colonisation Blockages account for the most significant contamination of clinical areas from the waste water system (1). Therefore, the implication of the waste water system as the source of the XDR P. aeruginosa in 2011 led to the implementation of many control measures in hospital 1 to reduce contamination of clinical areas from blockages, leaks and splashback events (1). Subsequently only infrequent XDR P. aeruginosa colonisation
events
and
clinical
infections
have
occurred
in
both
immunocompromised and non-immunocompromised individuals associated with the main hospital wing.
Indeed, since April 2012 there have been only four P. aeruginosa ST111 clinical infections in patients who have spent time on the General Intensive Care Unit (GICU), consisting of two infections in 2012 and two infections in 2013. On both occasions ST111 strains were isolated simultaneously from the sink U-bends and basins in patient rooms and from the sluice room. The fact that in this report we recently recovered from the sewer an isolate of ST111 highly related to those from current patient and environmental isolates and from a historical patient isolate, is further evidence that the waste water system probably acts as a reservoir for nosocomial transmission of XDR P. aeruginosa in this
This article is protected by copyright. All rights reserved.
Accepted Article
hospital. However, sequencing of other ST111 isolates from a separate source would give us a much greater understanding of the level diversity within this strain population. The recurring nature of these infections over a lengthy period of time suggests that this is more likely than transfer of XDR P. aeruginosa from colonised patients to the waste water system. Furthermore, consideration must be given to the downstream environmental impact of drug resistant P. aeruginosa, which has been shown to be found in wastewater treatment plants downstream of hospitals elsewhere (26).
The ST111 genome In addition to the epidemiological information generated by WGS, genetic markers of drug resistance and virulence can be identified and potentially used to guide better treatments
for
clinical
case
management.
Specific
genetic
markers
for
fluoroquinolones, aminoglycosides and β-lactams resistance were identified in our study; indeed Akasaka et al (28) suggested that the pattern of gyrA mutations identified in our ST111 isolates would confer susceptibility to the fluoroquinolone Sitofloxacin, but we were unable to obtain this drug to test. Indeed, rapid genotyping for these SNPs across these specific gyrA regions could be a useful clinical indicator for antibiotic choice. The wide-ranging specificities of the efflux pump machinery in P. aeruginosa are striking, but make full antibiotic and disinfectant genotypic to phenotype predictions much more complicated, requiring more experimental information on the regulation of these operons and specificities of their receptor molecules. Transcriptional regulation of these pumps would likely explain the susceptibility to disinfectants such
This article is protected by copyright. All rights reserved.
Accepted Article
as Triclosan despite the presence of pumps associated with resistance. In addition to the efflux pumps, the presence of genes responsible for copper resistance, heavy metal resistance, and acriflavine resistance shows that this organism is heavily adapted to live in harsh environmental conditions such as waste water systems. Phage integration appears important for the carriage of some of these elements and indeed, metagenomic studies have previously shown the presence of various phagerelated resistance markers in activated sludge from wastewater treatment plants (29). The presence of a phage in some but not all of the ST111 isolates is difficult to interpret due to the small number of isolates sequenced, but may simply reflect the transient nature of phages within these genomes or may be due to phage loss in the isolation process.
One P. putida isolate was sequenced out of curiosity, because there has been speculation that this species could act as a reservoir for horizontal gene transfer. One element was identified within PA38182 RGP2 with a high level of nucleotide similarity (98%) to the P. putida genome; this sequence has been seen in other sequenced genomes of P. putida but no other P. aeruginosa sequences within the public databases, suggesting that this element may have been transferred within the context of this outbreak but equally it may just have transferred between species in the past and the parent strains of P. aeruginosa containing it have not yet been identified and sequenced. A second element with high levels of similarity to the P. putida sequence was identified within RGP75 of PA38182, but this has been previously sequenced in both P. putida and P. aeruginosa sequences in the public databases, suggesting any transfer between species may have occurred previously.
This article is protected by copyright. All rights reserved.
Accepted Article
In this study, isolates from two separate UK hospital outbreaks were sequenced and compared, in order to identify any genetic evidence for the difference in mortality rates; however, within known virulence elements, only the S121N change in ExoS could be found, a site identified previously in an investigation of ExoS virulence, but its role could not be conclusively determined (30). P. aeruginosa isolates studied here have all been shown to be XDR and sensitive to only one antibiotic, colistin, in addition to disinfectant agents. However, colistin resistance has been described in patients with cystic fibrosis (31,32), therefore it is essential to monitor this organism carefully and to clinically manage treatment appropriately.
Development of further resistance would result in an organism
refractory to antibiotic treatment within the hospital and which could also lead to loss of susceptibility to disinfectants making infection control and environmental management within the hospital difficult. We believe that routine WGS of all isolates within a hospital environment would define the gene pool for resistance enabling both rapid molecular epidemiological investigation of outbreaks and enable surveillance for antibiotic resistance, thus informing good antibiotic stewardship - a major international governmental imperative.
5. Methods Susceptibility testing, serotyping, PFGE and VNTR Serotyping, PFGE and VNTR typing were performed by the Health Protection Agency, Colindale, UK (now The Public Health England, Laboratory for Healthcare Associated Infection (LHCAI)). Susceptibility testing using BSAC disc diffusion methods was described previously (1).
This article is protected by copyright. All rights reserved.
Accepted Article
Susceptibility to various disinfectants/antiseptics was measured by standard methods (33) for P. aeruginosa by measuring the minimum inhibitory concentration (MIC). The test strain for assessing bactericidal efficacy of disinfectants, strain NCTC 6749 was included. Each disinfectant was serially diluted by doubling dilutions. Three replicates were performed at each dilution and the modal value was taken.
Isolate preparation and DNA preparation Isolates were obtained from several environmental and patient sources (Table 1). The sewer sample was isolated according to Moore et al (34), briefly a gauze swab was placed in a sewer outlet pipe fed from three buildings on the hospital site for 48 hours. Fragments of the swab were incubated in selective broth (Nutrient Broth, Oxoid, Basingstoke, UK) containing 5 mg/ml vancomycin, 10 mg/ml meropenem overnight at 37°C, followed by single colony selection and overnight growth on Columbia blood agar plates (Oxoid, Basingstoke, UK). Single colonies were selected and species verified using API NE (BioMerieux, Marcy l'Etoile, France). Isolates were incubated overnight under aerobic conditions at 37°C on Columbia blood agar plates. Single colonies were subsequently grown for a second time overnight under the same conditions before DNA extraction using the FastDNA SPIN Kit for soil (MP Biomedicals, USA).
DNA Sequencing Whole genome sequencing was performed using the Ion Torrent Personal Genome Machine (PGM) (Life Technologies, Paisley, UK). 5 µg of genomic DNA was fragmented by sonication using a BioRupter UD-200 and the library prepared using
This article is protected by copyright. All rights reserved.
Accepted Article
the Ion Fragment Library Kit (Life Technologies) according to the manufacturer’s instructions. Briefly, fragmented DNA was end-repaired, ligated to Ion-compatible adapters, size-selected for optimum insert length and then nick-translated and amplified. Size-selection was carried out using a 2% E-Gel SizeSelect Agarose Gel (Invitrogen, UK) to produce a library with a median fragment size of approximately 200bp. The DNA library was clonally amplified on Ion Sphere Particles™ (ISP) by emulsion PCR using the Ion Xpress Template kit, and sequenced on an Ion 316 chip using the Ion Sequencing Kit v 2.0 which generates an average read length of approximately 100 bp.
Sequence Analysis Sequence
quality
was
assessed
using
Fastqc
(http://www.bioinformatics.babraham.ac.uk/projects/fastqc/). Reads were trimmed to a maximum length of 150bp using seqtk (https://github.com/lh3/seqtk). Sequence assembly was performed using MIRA v3.9.4 (35) with default parameters for Ion Torrent data. Automated gene annotation was performed using Prokka v1.6 (http://www.vicbioinformatics.com/software.prokka.shtml).
Manual
analysis
and
sequence inspection were performed using the Artemis and ACT genome visualisation tools (36,37).
In silico serotype and MLST analyses The serotype was inferred in silico by similarity searching a database of O-antigen gene sequences of serotypes O1-O20 using BLASTN (38) (Genbank accessions: AC104719, AC104720, AC104721, AC104722, AC104723, AC104724, AC104725,
This article is protected by copyright. All rights reserved.
Accepted Article
AC104726, AC104727, AC104728, AC104729, AC104730, AC104731, AC104732, AC104733, AC104734, AC104735, AC104736, AC104737, AC104738, AC104739). MLST types were determined by a method involving mapping reads to the PAO1 reference genome (RefSeq: NC_002516), and then subsequent BLAST comparison of consensus sequences to the P. aeruginosa MLST database.
Phylogenetic Analysis Whole genome phylogenetic reconstruction was performed as previously described (39); genomes were aligned with Mugsy (40), core genome sequences extracted with mothur (41) and maximum likelihood inference performed using RAxML v7.4.2 (42) with a GTR model of nucleotide substitution and a GAMMA model of rate heterogeneity, branch support values were determined using 1000 bootstrap replicates.
SNP based phylogenetic reconstruction was performed using a method similar to (22). Briefly, reads were mapped to the LESB58 reference genome (RefSeq: NC_011770; (43)) using TMAP v3.4.0 (Ion Torrent, USA) and alignments sorted with Picard v1.76 (http://picard.sourceforge.net/). Per base alignment statistics were generated with samtools mpileup (v0.1.18) (44) and variants called using bcftools v0.1.17-dev. Variant sites were filtered based on the following criteria; mapping quality (MQ) above 30, site quality score (QUAL) above 30, at least 4 reads covering each site with at least 2 reads mapping to each strand but with maximum depth of coverage 200, at least 75% of reads supporting site (DP4), allelic frequency (AF1) of 1. Sites which failed these criteria in any strain were removed from the analysis.
This article is protected by copyright. All rights reserved.
Accepted Article
Also, sites corresponding to known phage or genomic islands were removed. Phylogenetic reconstruction was performed using RAxML v7.4.2 (42) as above and branch SNP counts estimated using ancestral sequence reconstruction performed with PAML v4 (45).
6. Sequence data Sequence data has been deposited in the European Nucleotide Archive with study accession number PRJEB4573.
7. Acknowledgements: TP and CP thank St George’s Healthcare NHS Trust Charitable Trustees for funding to cover the costs of sequencing. The authors thank Public Health England for the VNTR, PFGE and serotyping data. The authors declare no conflicts of interest.
8. References: 1.
Breathnach AS, Cubbon MD, Karunaharan RN, Pope CF, Planche TD. Multidrug-resistant Pseudomonas aeruginosa outbreaks in two hospitals: association with contaminated hospital waste-water systems. Journal of Hospital Infection. 2012 Sep;82(1):19–24.
2.
Magiorakos A-P, Srinivasan A, Carey RB, Carmeli Y, Falagas ME, Giske CG, et al. Multidrug-resistant, extensively drug-resistant and pandrug-resistant bacteria: an international expert proposal for interim standard definitions for acquired resistance. Clinical Microbiology and Infection. 2012;18(3):268–81.
3.
Turton JF, Matos J, Kaufmann ME, Pitt TL. Variable number tandem repeat loci providing discrimination within widespread genotypes of Acinetobacter baumannii. Eur J Clin Microbiol Infect Dis. 2009 May 1;28(5):499–507.
This article is protected by copyright. All rights reserved.
Accepted Article
4.
Van der Bij AK, Van der Zwan D, Peirano G, Severin JA, Pitout JDD, Van Westreenen M, et al. Metallo-β-lactamase-producing Pseudomonas aeruginosa in the Netherlands: the nationwide emergence of a single sequence type. Clinical Microbiology and Infection. 2012;18(9):E369–E372.
5.
Juan C, Zamorano L, Mena A, Albertí S, Pérez JL, Oliver A. Metallo-betalactamase-producing Pseudomonas putida as a reservoir of multidrug resistance elements that can be transferred to successful Pseudomonas aeruginosa clones. The Journal of antimicrobial chemotherapy. 2010 Mar;65(3):474–8.
6.
Maiden MCJ, Bygraves JA, Feil E, Morelli G, Russell JE, Urwin R, et al. Multilocus sequence typing: A portable approach to the identification of clones within populations of pathogenic microorganisms. Proc Natl Acad Sci U S A. 1998 Mar 17;95(6):3140–5.
7.
Roy PH, Tetu SG, Larouche A, Elbourne L, Tremblay S, Ren Q, et al. Complete genome sequence of the multiresistant taxonomic outlier Pseudomonas aeruginosa PA7. PloS one. 2010 Jan;5(1):e8842.
8.
Jalal S, Wretlind B. Mechanisms of quinolone resistance in clinical strains of Pseudomonas aeruginosa. Microb Drug Resist. 1998;4(4):257–61.
9.
Samuelsen O, Toleman M a, Sundsfjord A, Rydberg J, Leegaard TM, Walder M, et al. Molecular epidemiology of metallo-beta-lactamase-producing Pseudomonas aeruginosa isolates from Norway and Sweden shows import of international clones and local clonal expansion. Antimicrobial agents and chemotherapy. 2010 Jan;54(1):346–52.
10. Mima T, Joshi S, Gomez-Escalada M, Schweizer HP. Identification and Characterization of TriABC-OpmH, a Triclosan Efflux Pump of Pseudomonas aeruginosa Requiring Two Membrane Fusion Proteins. J Bacteriol. 2007 Nov;189(21):7600–9. 11. Chuanchuen R, Narasaki CT, Schweizer HP. The MexJK Efflux Pump of Pseudomonas aeruginosa Requires OprM for Antibiotic Efflux but Not for Efflux of Triclosan. J Bacteriol. 2002 Sep;184(18):5036–44. 12. Leedjarv A, Ivask A, Virta M. Interplay of Different Transporters in the Mediation of Divalent Heavy Metal Resistance in Pseudomonas putida KT2440. J Bacteriol. 2008 Apr;190(8):2680–9. 13. Teitzel GM, Geddie A, De Long SK, Kirisits MJ, Whiteley M, Parsek MR. Survival and Growth in the Presence of Elevated Copper: Transcriptional Profiling of Copper-Stressed Pseudomonas aeruginosa. J Bacteriol. 2006 Oct;188(20):7242–56. This article is protected by copyright. All rights reserved.
Accepted Article
14. Mima T, Kohira N, Li Y, Sekiya H, Ogawa W, Kuroda T, et al. Gene cloning and characteristics of the RND-type multidrug efflux pump MuxABC-OpmB possessing two RND components in Pseudomonas aeruginosa. Microbiology. 2009 Nov 1;155(11):3509–17. 15. Muller JF, Stevens AM, Craig J, Love NG. Transcriptome Analysis Reveals that Multidrug Efflux Genes Are Upregulated To Protect Pseudomonas aeruginosa from Pentachlorophenol Stress. Appl Environ Microbiol. 2007 Jul;73(14):4550– 8. 16. Moskowitz SM, Ernst RK, Miller SI. PmrAB, a Two-Component Regulatory System of Pseudomonas aeruginosa That Modulates Resistance to Cationic Antimicrobial Peptides and Addition of Aminoarabinose to Lipid A. J Bacteriol. 2004 Jan;186(2):575–9. 17. Sun S, Negrea A, Rhen M, Andersson DI. Genetic Analysis of Colistin Resistance in Salmonella enterica Serovar Typhimurium. Antimicrob Agents Chemother. 2009 Jun;53(6):2298–305. 18. Beceiro A, Llobet E, Aranda J, Bengoechea JA, Doumith M, Hornsey M, et al. Phosphoethanolamine Modification of Lipid A in Colistin-Resistant Variants of Acinetobacter baumannii Mediated by the pmrAB Two-Component Regulatory System▿. Antimicrob Agents Chemother. 2011 Jul;55(7):3370–9. 19. Mathee K, Narasimhan G, Valdes C, Qiu X, Matewish JM, Koehrsen M, et al. Dynamics of Pseudomonas aeruginosa genome evolution. Proceedings of the National Academy of Sciences of the United States of America. 2008 Feb;105(8):3100–5. 20. Gutiérrez-Barranquero JA, Vicente A de, Carrión VJ, Sundin GW, Cazorla FM. Recruitment and Rearrangement of Three Different Genetic Determinants into a Conjugative Plasmid Increase Copper Resistance in Pseudomonas syringae. Appl Environ Microbiol. 2013 Feb 1;79(3):1028–33. 21. Battle SE, Rello J, Hauser AR. Genomic islands of Pseudomonas aeruginosa. FEMS microbiology letters. 2009 Jan;290(1):70–8. 22. Harris SR, Cartwright EJ, Török ME, Holden MT, Brown NM, Ogilvy-Stuart AL, et al. Whole-genome sequencing for analysis of an outbreak of meticillinresistant Staphylococcus aureus: a descriptive study. The Lancet Infectious Diseases. 2012 Nov;3099(12):1–7. 23. Köser CU, Ellington MJ, Cartwright EJP, Gillespie SH, Brown NM, Farrington M, et al. Routine Use of Microbial Whole Genome Sequencing in Diagnostic and Public Health Microbiology. Rall GF, editor. PLoS Pathogens. 2012 Aug;8(8):e1002824. This article is protected by copyright. All rights reserved.
Accepted Article
24. Köser CU, Holden MTG, Ellington MJ, Cartwright EJP, Brown NM, Ogilvy-Stuart AL, et al. Rapid whole-genome sequencing for investigation of a neonatal MRSA outbreak. New England Journal of Medicine. 2012;366(24):2267–75. 25. Curran B, Jonas D, Grundmann H, Pitt T, Dowson CG. Development of a Multilocus Sequence Typing Scheme for the Opportunistic Pathogen Pseudomonas aeruginosa. J Clin Microbiol. 2004 Dec;42(12):5644–9. 26. Slekovec C, Plantin J, Cholley P, Thouverez M, Talon D, Bertrand X, et al. Tracking Down Antibiotic-Resistant Pseudomonas aeruginosa Isolates in a Wastewater Network. PLoS One [Internet]. 2012 Dec 19 [cited 2013 Jul 23];7(12). Available from: http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3526604/ 27. Pirnay J, Bilocq F, Pot B, Cornelis P, Zizi M, Eldere JV, et al. Pseudomonas aeruginosa Population Structure Revisited. Population (English Edition). 2009;4(11). 28. Akasaka T, Tanaka M, Yamaguchi A, Sato K, Hemother ANAGC. Type II Topoisomerase Mutations in Fluoroquinolone-Resistant Clinical Strains of Pseudomonas aeruginosa Isolated in 1998 and 1999 : Role of Target Enzyme in Mechanism of Fluoroquinolone Resistance. Society. 2001;45(8):2263–8. 29. Parsley LC, Consuegra EJ, Kakirde KS, Land AM, Harper WF, Liles MR. Identification of Diverse Antimicrobial Resistance Determinants Carried on Bacterial, Plasmid, or Viral Metagenomes from an Activated Sludge Microbial Assemblage. Appl Environ Microbiol. 2010 Jun;76(11):3753–7. 30. Ferguson MW, Maxwell JA, Vincent TS, da Silva J, Olson JC. Comparison of the exoS Gene and Protein Expression in Soil and Clinical Isolates of Pseudomonas aeruginosa. Infection and Immunity. 2001 Apr 1;69(4):2198–210. 31. Schülin T. In vitro activity of the aerosolized agents colistin and tobramycin and five intravenous agents against Pseudomonas aeruginosa isolated from cystic fibrosis patients in southwestern Germany. J Antimicrob Chemother. 2002 Feb 1;49(2):403–6. 32. Clifton I, Peckham D, Conway S. 10 years of use of colistin in a regional cystic fibrosis centre. Journal of Cystic Fibrosis Medicine. 2005;202:S54. 33. McDonnel GE. Antisepsis, Disinfection, and Sterilization: Types, Action, and Resistance. ASM Press; 2007. 378 p. 34. Moore B, Perry EL, Chard ST. A survey by the sewage swab method of latent enteric infection in an urban area. J Hyg (Lond). 1952 Jun;50(2):137–56.
This article is protected by copyright. All rights reserved.
Accepted Article
35. Chevreux B, Wetter T, Suhai S. Genome Sequence Assembly Using Trace Signals and Additional Sequence Information. Computer Science and Biology: Proceedings of the German Conference on Bioinformatics. 1999;99:45–56. 36. Carver TJ, Rutherford KM, Berriman M, Rajandream M-A, Barrell BG, Parkhill J. ACT: the Artemis comparison tool. Bioinformatics. 2005 Aug 15;21(16):3422–3. 37. Rutherford K, Parkhill J, Crook J, Horsnell T, Rice P, Rajandream M-A, et al. Artemis: sequence visualization and annotation. Bioinformatics. 2000 Oct 1;16(10):944–5. 38. Altschul SF, Madden TL, Schaffer AA, Zhang J, Zhang Z, Miller W, et al. Gapped BLAST and PSI-BLAST: a new generation of protein database search programs. Nucleic Acids Res. 1997 Sep 1;25(17):3389–402. 39. McNally A, Cheng L, Harris SR, Corander J. The Evolutionary Path to Extraintestinal Pathogenic, Drug-Resistant Escherichia coli Is Marked by Drastic Reduction in Detectable Recombination within the Core Genome. Genome Biol Evol. 2013;5(4):699–710. 40. Angiuoli SV, Salzberg SL. Mugsy: fast multiple alignment of closely related whole genomes. Bioinformatics. 2011 Feb 1;27(3):334–42. 41. Schloss PD, Westcott SL, Ryabin T, Hall JR, Hartmann M, Hollister EB, et al. Introducing mothur: Open-Source, Platform-Independent, CommunitySupported Software for Describing and Comparing Microbial Communities. Appl Environ Microbiol. 2009 Dec;75(23):7537–41. 42. Stamatakis A. RAxML-VI-HPC: maximum likelihood-based phylogenetic analyses with thousands of taxa and mixed models. Bioinformatics (Oxford, England). 2006 Nov;22(21):2688–90. 43. Winstanley C, Langille MGI, Fothergill JL, Kukavica-Ibrulj I, Paradis-Bleau C, Sanschagrin F, et al. Newly introduced genomic prophage islands are critical determinants of in vivo competitiveness in the Liverpool Epidemic Strain of Pseudomonas aeruginosa. Genome research. 2009 Jan;19(1):12–23. 44. Li H, Handsaker B, Wysoker A, Fennell T, Ruan J, Homer N, et al. The Sequence Alignment/Map format and SAMtools. Bioinformatics. 2009 Aug 15;25(16):2078–9. 45. Yang Z. PAML 4: Phylogenetic Analysis by Maximum Likelihood. Mol Biol Evol. 2007 Aug 1;24(8):1586–91.
This article is protected by copyright. All rights reserved.
Accepted Article
Table 1. Pseudomonas spp. isolates sequenced in this study.
Strain name
Source1
1 PA38182
Patient blood2
A 14 2011
Feb 11, 3, 4, 3, 2, 2, O12 5, 4, 7
ST1 11
17,5,5,4,4 ,4,3
2 PA48922
Patient sink2
A 01 2011
Mar 11, 3, 4+, 3, 2, 2, O124 -, 4, 7
ST1 11
17,5,5,4,4 ,4,3
3 PP48922
Patient sink2
A 01 2011
Mar n/a (P. putida)
n/a
4 PA46218
Patient sink2
A 21 2011
Feb 11, 3, 4+, 3, 2, 2, O124 5, 4, 11
ST1 11
17,5,5,4,4 ,4,3
5 PASwab
Sewer2
06 2011
Oct 11, 3, 4, 3, 2, 2, O124 5, 4, 7
ST1 11
17,5,5,4,4 ,4,3
6 PA128572
Patient blood2
B 12 2006
Aug 11, 3, 4, 3, 2, 2, O124 5, 4, 7
ST1 11
17,5,5,4,4 ,4,3
7 PA99736
Patient blood2
C 02 2011
May 11, 4, 5, 2, 3, 2, - O64 , 4, 12
ST1 55
28,5,36,3, 3,13,7
8 PA97498
Patient blood3
02 2010
Aug 11, 3, 2, -, 3, 1, 9, 3, 10
5
ST6 54
17,5,26,3, 4,4,26
9 PA156066
Patient bathroom3
04 2011
May 11, 3, 2, -, 3, 1, 9, 3, 10
5
ST6 54
17,5,26,3, 4,4,26
Date
VNTR-type
1
Seroty pe
n/a
MLS MLST profile6 T4
See section 3.1 for isolate descriptions; 2 Hospital 1; 3 Hospital 2; 4 inferred from sequence data; 5 undeterminable from sequence data; 6 MLST genes are: acs, aro, gua, mut, nuo, pps, trp
This article is protected by copyright. All rights reserved.
Accepted Article
Table 2A. Minimum inhibitory concentrations of antibiotics for a representative isolate of ST111 (isolate PA38182).
Antibiotic Class Antibiotic
MIC (mg/L)
Ciprofloxacin
>8
Moxifloxacin
>32
Amikacin
64
Gentamycin
>32
Tobramycin
>32
Monobactam
Azetreonam
16
Cephalosporin
Ceftazidine
16
Carbapenems
Imipenem
>128
Meropenem
>32
Piperacillin
64
Fluoroquinolone
Aminoglycoside
Penicillin Penicillin inhibitor
/
β-lactamase Piperacillin-tazobactam
64
Polymyxin
Colistin
2
Carboxypenicillin
Ticaricillin / clavulanate
64
Tetracycline
Minocycline
32
Glycylcycline
Tigecycline
16
/ β-lactamase inhibitor
This article is protected by copyright. All rights reserved.
Accepted Article
Table 2B. Minimum inhibitory concentrations of disinfectants. Isolates PA38182, PA128572 and PASwab were tested. In addition, the standard test strain for assessing bactericidal efficacy of disinfectants, strain NCTC 6749 was included.
Disinfectant Povidone
Initial concentration
MIC obtained NCTC 6749
PA128572
PA38182
PASwab
10%
1 in 8
1 in 8
1 in 8
1 in 8
4%
1 in 4
1 in 4
1 in 4
1 in 4
1000 ppm
1 in 2
1 in 2
1 in 2
1 in 2
2%
1 in 8
1 in 8
1 in 8
1 in 4
1100 ppm
1 in 8
1 in 8
1 in 16
1 in 16
1500 ppm
1 in 32
1 in 32
1 in 64
1 in 64
1%
1 in 2
1 in 2
1 in 2
1 in 2
iodine Chlorhexidine NaDCC Glutaraldehyde Chlorine dioxide Peracetic acid Triclosan
This article is protected by copyright. All rights reserved.
Accepted Article
Figure 1A. Phylogenetic relatio onship between P. aeruginosa isolates seque enced in this study in the context of 35 geno ome sequences from public databases (see su upplementary Table 2 for accessions); based on core genome alignments. Sequence derive ed MLST type included in square brackets. PA7 was excluded as an outlier. 1) Outbreak 1 strains, 2) Outbreak 2 strains, 3) unrelated strain from Hospital 1.
1
2
3
This article is protected by copyright. All rights reserved.
Accepted Article
Figure 1B. Phylogenetic reconstruction from 24 ST111-specific variant sites identified by comparing isolate read data with h the reference genome LESB58 (RefSeq: NC_ _011770). The tree is rooted to NC_011770. Asterisks marks branch with 100% bootstrap su upport values. Ancestral site reconstruction was w used to estimate the number of SNP changes on branches; branches with significant support values are labelled with the estimatted number of SNPs out of the total number of SNP sites identified when mapped against LESB58, PAO1, UCBPP-PA14 (RefSeq: NC_011 1770, NC_002516, NC_008463) respectively. T The scale bar indicates estimated mutations per p site. +/- refers to presence/absence of pu utative phage (see Figure 2). See section 3.1 for isolate descriptions.
[6yo strain] +
[Patient A sin nk] +
*
[Patien nt A sink] +
[Patient A] -
[Sewer swab] -
This article is protected by copyright. All rights reserved.
Accepted Article
Figure 2. Read coverage acrross differential phage region within ST111 isolates. A) predicted genes, putative pha age genes coloured red; B) PA128572; C C) PA48922; D)PA46218; E) PA38182; F) PASwab.
A B
C
D E
F
This article is protected by copyright. All rights reserved.