Abstract
Carbapenemase-producing Klebsiella pneumoniae pose an increasing risk for healthcare facilities worldwide. A continuous monitoring of ST distribution and its association with resistance and virulence genes is required for early detection of successful K. pneumoniae lineages. In this study, we used WGS to characterize MDR blaOXA-48-positive K. pneumoniae isolated from inpatients at the University Medical Center Göttingen, Germany, between March 2013 and August 2014.
Closed genomes for 16 isolates of carbapenemase-producing K. pneumoniae were generated by single molecule real-time technology using the PacBio RSII platform.
Eight of the 16 isolates showed identical XbaI macrorestriction patterns and shared the same MLST, ST147. The eight ST147 isolates differed by only 1–25 SNPs of their core genome, indicating a clonal origin. Most of the eight ST147 isolates carried four plasmids with sizes of 246.8, 96.1, 63.6 and 61.0 kb and a novel linear plasmid prophage, named pKO2, of 54.6 kb. The blaOXA-48 gene was located on a 63.6 kb IncL plasmid and is part of composite transposon Tn1999.2. The ST147 isolates expressed the yersinabactin system as a major virulence factor. The comparative whole-genome analysis revealed several rearrangements of mobile genetic elements and losses of chromosomal and plasmidic regions in the ST147 isolates.
Single molecule real-time sequencing allowed monitoring of the genetic and epigenetic microevolution of MDR OXA-48-producing K. pneumoniae and revealed in addition to SNPs, complex rearrangements of genetic elements.
Introduction
The spread of carbapenemase-producing Klebsiella pneumoniae in hospitals is continuously increasing and has emerged as a major global health threat. Carbapenemases are typically plasmid-encoded and are either of the serine-β-lactamase type, such as KPC and OXA-48, or of the metallo-β-lactamase type, such as VIM and NDM.1
A recent multicentre study revealed that OXA-48 is currently the most frequent carbapenemase in K. pneumoniae isolates from German hospitals.2 Previous molecular epidemiological data indicated that OXA-48-producing K. pneumoniae are frequently found in North African countries, in the Middle East and in south European Mediterranean countries, but less often in North and South America.3–5 More than 150 clonal groups (CGs) of K. pneumoniae are currently defined based on whole-genome analyses.6 Five widely distributed CGs, namely CG258, CG14/15, CG17/20, CG43 and CG147, are responsible for 72% of all reported K. pneumoniae outbreaks in which ST was performed.7 Owing to the focus of many of these studies on drug resistance, the number of reported isolates associated with resistance is high.
Spread of carbapenemases occurs either by clonal expansion and transmission of successful lineages or by horizontal gene transfer of carbapenemase-encoding plasmids to distinct K. pneumoniae strains or to other Enterobacteriaceae,8,9 The blaOXA-48 gene is often part of a composite transposon Tn1999 variant and associated with IncL plasmids, irrespective of the MLST.10 In recent years, WGS was successfully applied to characterize carbapenemase-producing K. pneumoniae, with the majority of studies describing KPC-producing K. pneumoniae.11–14 A comprehensive K. pneumoniae database of closed genomes is necessary for a complete understanding of the genome plasticity of these organisms and can significantly improve the tracking of MDR isolates. Here, we report the results of whole-genome analysis using single molecule real-time (SMRT) sequencing of blaOXA-48-positive, MDR K. pneumoniae isolates from the University Medical Center, Göttingen, Germany.
Methods
Bacterial isolates and susceptibility testing
Details of bacterial species determination and susceptibility testing are given in the legends of Tables S1 and S2 (available as Supplementary data at JAC Online).
Basic molecular characterization
The presence of carbapenemase genes (blaNDM, blaOXA-48, blaVIM, blaIMP and blaKPC) was tested by PCR as previously described.15,16 PCR for MLST and subsequent sequence data analysis was performed according to the protocols and tools given on the following web site: http://bigsdb.pasteur.fr/klebsiella/klebsiella.html.17,18 MEGA7 software was used for the construction of an MLST-based unweighted pair group method using average linkages (UPGMA) dendrogram;19 and a minimum spanning tree was constructed using PHYLOViZ.20
Genome sequencing
High molecular weight DNA was prepared using Qiagen Genomic Tip/100G (Qiagen, Hilden, Germany) according to the manufacturer’s instructions. An SMRTbell™ template library was prepared according to the instructions from Pacific Biosciences (Menlo Park, CA, USA) following the Procedure & Checklist—10 kb Template Preparation Using BluePippin™ Size-Selection System. Briefly, for preparation of 15 kb libraries 8 μg genomic DNA was sheared using g-tubes™ from Covaris (Woburn, MA, USA) according to the instructions of the manufacturer. DNA was end-repaired and ligated overnight to hairpin adapters applying components from the DNA/Polymerase Binding Kit P6 from Pacific Biosciences. Reactions were carried out according to the manufacturer’s instructions. BluePippin™ Size-Selection to 4 kb was performed according to the manufacturer’s instructions (Sage Science, Beverly, MA, USA). Conditions for annealing of sequencing primers and binding of polymerase to purified SMRTbell™ template were assessed with the calculator in RS Remote (Pacific Biosciences). SMRT sequencing was carried out on the PacBio RSII (Pacific Biosciences) taking one 240 min movie for each SMRT cell. In total one SMRT cell per strain was run, only for strain 829763-1 were two SMRT cells run. Short read sequencing was performed from the same DNA on the HiSeq 2500 (llumina Inc., San Diego, CA, USA) for 100 cycles in both directions to obtain ∼4 Mio reads per sample.
Genome assembly and bioinformatics analysis
SMRT cell data were assembled independently using the RS_HGAP_Assembly.3 protocol included in SMRT Portal version 2.3.0 using default parameters. The assemblies yielded a circular chromosome for each strain and up to six extrachromosomal elements, comprising plasmids as well as linear phages. The validity of the assembly was checked using the RS_Bridgemapper.1 protocol. Each replicon was circularized independently; in particular, artificial redundancies at the ends of the contigs were removed. The chromosome was adjusted to dnaA as the first gene. Finally, each genome was corrected for indel errors by a mapping of Illumina reads on to finished genomes using the Burrows–Wheeler Alignment tool21 with subsequent variant and consensus calling using VarScan 2.22 A consensus concordance of QV60 could be confirmed for all genomes. Finally, automated genome annotation was generated using Prokka 1.823 as well as the NCBI prokaryotic genome annotation pipeline.24 Annotation of non-coding RNAs was made taking the Rfam database (http://rfam.xfam.org/) into account.25 Plasmid incompatibility groups were identified using the PlasmidFinder 1.3 tool.26 Integrons have been analysed using the integron identification tool INTEGRALL (http://integrall.bio.ua.pt/).27 The genome sequences were deposited in NCBI GenBank and accession numbers are listed in Tables S3 and S4.
Core-genome MLST (cgMLST) and SNP analysis
Allele variants for cgMLST analysis were identified by uploading assembled genomes to the K. pneumoniae section of the Institut Pasteur MLST and whole-genome MLST (wgMLST) database (http://bigsdb.pasteur.fr/).6 The selected locus scheme was cgMLST_KpI. Allele variants of all eight ST147 isolates were transferred into a spreadsheet and manually curated for genes not present in all eight genomes. For SNP analysis, genomes were pairwise aligned with the progressive Mauve algorithm in the Geneious software package v.8.1.8.28 The aligned sequences were searched for disagreements and each disagreement was manually inspected for SNP identity. SNP-based phylogeny was investigated with the CSI-Phylogeny-1.4 tool at the web site of the Center for Genomic Epidemiology (https://cge.cbs.dtu.dk/services/CSIPhylogeny/).29
Virulence and resistance genes
Antimicrobial resistance (AMR) genes were identified using the ResFinder 2.1 tool (https://cge.cbs.dtu.dk/services/ResFinder/) with an identity threshold of >96%.30K. pneumoniae virulence genes were identified with the aid of the K. pneumoniae section of the Institut Pasteur MLST and wgMLST database (http://bigsdb.pasteur.fr).
Results and discussion
Between March 2013 and August 2014, 16 MDR K. pneumoniae isolates from 12 patients were identified at the Institute for Medical Microbiology (Göttingen, Germany; Table S1). PCR analysis revealed that 15 isolates were blaOXA-48-positive and one was blaNDM-1-positive. MLST indicated that 8 of 15 isolates were in ST147, an ST that is widely distributed in the Middle East, Europe and Asia7 and that has been frequently reported to be associated with carbapenem resistance. Epidemics of KPC-2 and VIM-producing ST147 isolates have been extensively reported in Greece.31,32 OXA-48-producing K. pneumoniae of ST147 have been described before in Mediterranean countries and in Asia.33,34 Other STs detected were ST101 (three isolates from two patients), ST11 (two isolates from one patient), ST23, ST15 and ST395 (Figure 1a). In previous studies, STs 11, 15, 101 and 395 were found to be associated with OXA-48 in middle European countries, such as France.35,36 Seven of our study isolates originate from immigrants/refugees from the Middle East/Northern Africa, whereas nine isolates were cultured from autochthonous German citizens (Figure 1b). However, of the ST147 isolates, only one isolate originates from the Middle East (from a 22 year old male refugee); the remaining ST147 isolates were isolated from autochthonous German patients hospitalized on different surgical wards of the University Medical Center (Göttingen, Germany), aged between 51 and 82 years. One patient was female and the seven remaining patients were male (Table S1).
(a) MLST sequence-based UPGMA tree. All 16 K. pneumoniae isolates included in our study have been MLS-typed and MEGA7 was used to calculate the depicted UPGMA dendrogram. Optimal tree with the sum of branch length = 0.01042393 is shown. Dendrogram was drawn to scale, with branch lengths in the same units as those of evolutionary distances used to infer the phylogenetic tree. Evolutionary distances were computed using the Maximum Composite Likelihood method and are in units of the number of base substitutions per site. (b) MLST-based minimum-spanning tree. PHYLOViZ-generated minimum-spanning tree is based on the allele number matrix of the gene loci included in the K. pneumoniae MLST scheme. Grey indicates an isolate originating from an autochthonous German citizen, while black indicates an isolate originating from an immigrant from the Middle East/northern Africa. White numbers in the circles indicate the MLST sequence type. Black numbers along the linking lines indicate the absolute distance.
PFGE, cgMLST and SNP analysis
PFGE was applied on the ST147 isolates and the identical macrorestriction patterns obtained (only one band difference to the isolate Kp_Goe_822579 from the refugee) indicated that all eight ST147 isolates belonged to one clone (Figure S1). SMRT sequencing was performed on the ST147 isolates and sequence data were analysed first by cgMLST using the K. pneumoniae section of the Institut Pasteur MLST and wgMLST database. The allele variants of the 1222 common genes were 100% identical for the eight ST147 isolates. These isolates are thus indistinguishable by cgMLST, which further supports a clonal origin. To investigate the population structure of the ST147 isolates in higher depth, chromosomal SNP analysis was applied. The genomes were pairwise aligned with the progressive Mauve algorithm in the Geneious software package (v.8.1.8) and the aligned sequences were searched for disagreements, followed by manual inspection of each potential SNP. Seven isolates were found to be highly similar and differed by only one to eight SNPs. The first ST147 isolate Kp_Goe_822579 from March 2013 showed a higher degree of variation, with 18–21 SNPs in pairwise comparisons. The microevolution of ST147 isolates was further investigated with the SNP-based CSI-Phylogeny-1.4 tool hosted at the web site of the Center for Genomic Epidemiology. The SNP-based phylogeny confirms the increased phylogenetic distance of isolate Kp_Goe_822579 (Figure S2). Given the 14 month distance between sampling of isolate Kp_Goe_822579 and the other seven ST147 isolates, the number of identified SNPs is in good agreement with the previously estimated ∼10 mutations per year per K. pneumoniae genome.9
In addition to SNPs, the Kp_Goe_822579 genome differs by several larger deletions, insertions and inversions when compared with the other ST147 genomes. The most relevant difference is an 11.6 kb genomic region that contains AMR genes and mobile genetic elements at position 3.1 Mb in the Kp_Goe_822579 genome. This region is absent in the other seven ST147 genomes (Figure S3). A BLASTN search revealed that this element shares 100% identity to a region within the 362 kb K. pneumoniae plasmid pKpn23412-362 (NZ_CP011314) from an ESBL-producing ST514 isolate from a German hospital in 2011.37 The backbone of this plasmid is still present in all ST147 isolates as a 246.8 kb plasmid (see the section entitled Extrachromosomal elements of the ST147 isolates). Furthermore, Kp_Goe_822579 differs from all other ST147 isolates by the reverse orientation of a 63 kb genomic region at position 1.9 Mb. This region contains a putative integrative and conjugative element (ICE) with a high pathogenicity island (HPI) similar to HPI-ICEKp1 described in 2008 from Taiwan.38 The HPI-ICE of the ST147 isolates is located adjacent to an asparagine tRNA gene and carries the yersinabactin system for iron acquisition, comprising the yersinabactin receptor gene fyuA, the siderophore genes irp1 and irp2 and the yersiniabactin siderophore cluster ybtAEPQSTUX. The genes for conjugative transfer are adjacent (virB1 to virB11; mobB, mobC). However, the central region of HPI-ICEKp1 is missing in the ST147 isolates. This region contains several virulence genes, such as the iroN-iroB-iroC-iroD operon involved in iron uptake, the vagC-vagD operon and the regulator of the mucoid phenotype rmpA.38 In conclusion, the higher degree of diversity of Kp_Goe_822579 leaves two possibilities: Kp_Goe_822579 is a direct progenitor of the seven clonal ST147 isolates, or less likely, it is a subclone from a common progenitor that has evolved independently from the other isolates.
Extrachromosomal elements of the ST147 isolates
Seven of the eight ST147 isolates carried four plasmids, with sizes of 246.8, 96.1, 63.6 and 61.0 kb. The blaOXA-48 genes were located on conjugative 63.6 kb IncL plasmids and are part of a 5.6 kb Tn1999.2 composite transposon (Figure 2). This type of transposon contains two copies of the insertion sequence IS1999. The copy that is located proximal to blaOXA-48 is truncated by the insertion of an IS1R element and referred to as ΔIS1999.39 The 63.6 kb plasmids display >99.9% sequence identity to the 63.6 kb conjugative plasmid pE71T (accession no. KC335143), isolated from a clinical K. pneumoniae isolate from Ireland.39 The Tn1999.2 sequence of the first ST147 isolate Kp_Goe_822579 is 100% identical to the Tn1999.2 sequence of plasmid pE71T; all other ST147 isolates possess a single SNP in Tn1999.2 within the right inverted repeat of the IS1R element (TAAGTTGG→TATGTTGG).
Plasmids of ST147 isolates. Seven of the eight ST147 isolates carry four plasmids, respectively, with sizes of 246.8, 96.1, 63.6 and 61.0 kb. The 63.6 kb plasmid carries the blaOXA-48 gene as part of a Tn1999.2 transposon. Isolate Kp_Goe_825795 lacks the 96.1 and 63.6 kb plasmid but carries a 117 kb plasmid. This novel plasmid is composed of the 96.1 kb plasmid backbone, in which a 21 kb region from the 63.6 kb blaOXA-48-carrying plasmid was integrated at an IS1R site. Since the recombination event occurred at the IS1R element that is part of Tn1999.2, the resulting new Tn1999 variant lacks the proximal IS1999 region. Plasmid backbones of the 246.8 and 61.0 kb plasmids as well as sequence regions integrated into these plasmids are indicated. Class 1 integron (In363) in the 61.0 kb plasmid is indicated in red.
The 246.8 kb plasmid shows the highest sequence identity (80% query coverage; 99% identity) to pKpn23412-362, the same 362 kb plasmid from which an 11.6 kb region has integrated into the chromosome of isolate 822579. The 246.8 kb plasmid lost a large part of pKpn23412-362, including all 13 AMR genes present on pKpn23412-362.37 Instead, the 246.8 kb plasmid contains a 32 kb region that is not part of pKpn23412-362. This region encodes genes for several ion transporters (an H+/CL− antiporter, a divalent cation transporter and a potassium transporter) and metabolic enzymes (enolase and phosphoglycerate mutase). Furthermore, an undecaprenyl-diphosphatase, which is an enzyme that participates in peptidoglycan biosynthesis, a pyrophosphatase, and several stress-associated genes can be found in this region. Similar to pKpn23412-362, the 246.8 kb plasmid exhibits more than one replicon type [IncFIB (Mar) and IncHI1B], which should broaden its host range.
The 96.1 kb plasmid belongs to the class of IncA/C2-type MDR plasmids and shares 96% coverage and 99% identity to the K. pneumoniae type 1A/C2 self-transferable plasmid pKP-Gr642.40 It carries the ARI-B resistance island (floR, tetA, strA, strB, sul2), but not the ARI-A island.
The 61.0 kb IncR-type MDR plasmid carries genes encoding ESBL-type CTX-M-15 and β-lactamase TEM-1, the quinolone resistance gene qnrS1 and additional genes [strA, strB, tet(A) and dfrA1]. The dfrA1 gene and a gene cassette of unknown function are organized in a class 1 integron (In363), as analysed with the integron identification tool INTEGRALL. The 61.0 kb plasmid shares 51 kb of its sequence with the 90 kb K. pneumoniae plasmid pSg1-NDM (CP011839).41 However, various larger regions of pSg1-NDM were lost, including the blaNDM-1 gene. Instead, regions from two other K. pneumoniae plasmids were integrated into the pSg1-NDM backbone. A 4.5 kb region including 5 ORFs is >99% identical to the 258 kb plasmid pCAV1193-258 (CP013323).42 This region encodes a colicin V synthesis protein, a colicin V secretion protein, restriction endonuclease and VapBC toxin–antitoxin system. A 5.5 kb region including six ORFs, encoding two aminoglycoside-O-phosphotransferases, two IS5 family transposases and two hypothetical proteins, is >99% identical to the 50 kb plasmid pKP1780 (JX424614), which was previously characterized from a ST147 isolate.43
Interestingly, one of the eight ST147 isolates, Kp_Goe_825795, lacks the 63.6 kb and the 96.1 kb plasmid. Instead, this isolate carries a 117 kb plasmid that results from integration of a blaOXA-48 containing 21 kb region from the 63.6 kb plasmid into the 96.1 kb plasmid at an IS1R site. This region is flanked by two IS1R elements and is identical to Tn6237, a transposon that was previously shown able to integrate into the bacterial chromosome (Figure 2).44
Furthermore, all ST147 strains carry a 54.6 kb linear, extrachromosomal element that contains bacteriophage genes. This element exhibits 83% sequence similarity to pKO2, which was described as a linear plasmid prophage of Klebsiellaoxytoca (Figure S4).45 While the vast majority of bacteriophages integrate their prophage DNA into the bacterial chromosome, a few phages such as φKO2, N15, P1 and φ20 replicate in the prophage stage as low-copy-number plasmids. The sequence similarity suggests that the identified linear 54.6 kb element is the plasmid prophage DNA of a novel bacteriophage, which is named here φKp1. The φKp1 phage contains a 54 bp palindromic sequence, called telRL (site of telomere formation), which serves as a recognition site for a protelomerase that produces linearized double-stranded DNA with covalently closed hairpin telomeres from circular plasmid DNA.46 The phage plasmid pKp1 encodes apart from phage-related genes an adenine-methylase (rsrIM). Strikingly and in contrast to pKO2, the pKp1 plasmid encodes a 378 amino acid protein that exhibits high sequence similarity to the pore-forming domain of colicin A. Whether this protein is indeed able to form pores remains to be confirmed experimentally.
Resistome, antimicrobial susceptibility and virulence-associated genes
All ST147 isolates carried chromosomally located AMR genes coding for an SHV-11 β-lactamase as well as oqxAB and fosA, which confer resistance to quinolones and fosfomycin, respectively. These genes belong to the core genome of the KpI–III phylogroups.47 However, the majority of the AMR genes were located on plasmids (Figure 2). In summary, the resistome analysis of the ST147 isolates predicted genes conferring resistance to penicillins (blaTEM-1, blaSHV-11, blaOXA-1), cephalosporins (blaCTX-M-15), carbapenems (blaOXA-48), aminoglycosides [aac(6´)Ib-cr, strA, strB, aacA4], fluoroquinolones (qnrS1, oqxAB), fosfomycin (fosA), chloramphenicol (floR), sulphonamides (sul1, sul2), tetracyclines (tetA) and trimethoprim (dfrA1). The tested antimicrobial susceptibility profiles of the eight ST147 isolates are in complete accordance with the prediction from the resistome, except for isolate Kp_Goe_827026 (Table S2, MICs). Moreover, the subsequently performed modified Hodge test was negative for isolate 827026 in contrast to the isolates with the unmodified blaOXA-48 gene. Although this isolate carries blaOXA-48, it was found to be susceptible to carbapenems (except ertapenem). Sequence analysis revealed a point mutation in the blaOXA-48 gene. One G-to-A transition at nucleotide 75 of the blaOXA-48 coding sequence converts a TGG-Trp codon into a TGA stop codon (W25STOP).
The intermediate ertapenem resistance (3 mg/L) of Kp_Goe_827026 in the absence of a functional OXA-48 is most likely due to a mutated ompK36. Sequence analysis revealed a 6 bp insertion leading to a 134–135 GD insertion in OmpK36 in all ST147 isolates. This insertion is known to diminish uptake of certain carbapenems, such as doripenem.48
From the panel of tested antibiotics, gentamicin was the only one for which at least seven of the eight ST147 isolates showed susceptibility (<2 mg/L). However, the gentamicin MIC was considerably higher for isolate Kp_Goe_822579 (>8 mg/L), which might be due to the 11.6 kb genomic region that is present in this isolate, but was lost in all other ST147 isolates. This region contains the gene aac(6′)Ib-cr encoding an aminoglycoside N(6′)-acetyltransferase that might be responsible for the increased gentamicin resistance of isolate Kp_Goe_822579 (Figure S3).
Known K. pneumoniae virulence genes were identified with the aid of the K. pneumoniae section of the Institut Pasteur MLST and wgMLST database. The ST147 isolates carried the previously described yersinabactin system and the type 3 fimbriae cluster mrkABCDF. According to the number of identified virulence factors, the ST147 isolates appear to be of moderate virulence; high virulence strains possess multiple iron acquisition systems and express the mucoid phenotype regulator gene rmpA.47
In conclusion, in this study, we followed the genetic microevolution of MDR OXA-48-producing K. pneumoniae of ST147 and could identify insertions/deletions of chromosomal regions and rearrangements of mobile genetic elements and plasmids in addition to SNPs. A continuous monitoring of ST distribution and its association with resistance and virulence genes is required for early detection of successful K. pneumoniae lineages.
Acknowledgements
We thank Simone Severitt, Nicole Heyer, Sybille Müller-Bertling and Girstin Ganske for excellent technical assistance as well as Jolantha Swiderski for bioinformatics assistance in genome assembly. We are grateful to Richard J. Roberts (New England Biolabs, Ipswich, MA, USA) for performing the REBASE analysis.
Funding
The work was supported by grant no. 105-4-5-08.001 of the German Federal Ministry of Science and Education through the German Center of Infection Research (DZIF) to J. O.
Transparency declarations
None to declare.
Supplementary data
Tables S1 to S4 and Figures S1 to S4 are available as Supplementary data at JAC Online.
References
Author notes
These authors contributed equally to this study.
These authors share senior authorship.
