Comparative genomic analysis of Klebsiella pneumoniae subsp. pneumoniae KP617 and PittNDM01, NUHL24835, and ATCC BAA-2146 reveals unique evolutionary history of this strain

Background Klebsiella pneumoniae subsp. pneumoniae KP617 is a pathogenic strain that coproduces OXA-232 and NDM-1 carbapenemases. We sequenced the genome of KP617, which was isolated from the wound of a Korean burn patient, and performed a comparative genomic analysis with three additional strains: PittNDM01, NUHL24835 and ATCC BAA-2146. Results The complete genome of KP617 was obtained via multi-platform whole-genome sequencing. Phylogenetic analysis along with whole genome and multi-locus sequence typing of genes of the Klebsiella pneumoniae species showed that KP617 belongs to the WGLW2 group, which includes PittNDM01 and NUHL24835. Comparison of annotated genes showed that KP617 shares 98.3 % of its genes with PittNDM01. Nineteen antibiotic resistance genes were identified in the KP617 genome: blaOXA-1 and blaSHV-28 in the chromosome, blaNDM-1 in plasmid 1, and blaOXA-232 in plasmid 2 conferred resistance to beta-lactams; however, colistin- and tetracycline-resistance genes were not found. We identified 117 virulence factors in the KP617 genome, and discovered that the genes encoding these factors were also harbored by the reference strains; eight genes were lipopolysaccharide-related and four were capsular polysaccharide-related. A comparative analysis of phage-associated regions indicated that two phage regions are specific to the KP617 genome and that prophages did not act as a vehicle for transfer of antimicrobial resistance genes in this strain. Conclusions Whole-genome sequencing and bioinformatics analysis revealed similarity in the genome sequences and content, and differences in phage-related genes, plasmids and antimicrobial resistance genes between KP617 and the references. In order to elucidate the precise role of these factors in the pathogenicity of KP617, further studies are required. Electronic supplementary material The online version of this article (doi:10.1186/s13099-016-0117-1) contains supplementary material, which is available to authorized users.


Background
Klebsiella pneumoniae is a Gram-negative, non-motile, encapsulated, facultative anaerobic bacterium, which belongs to the family Enterobacteriaceae. K. pneumoniae is found in the normal flora of the mouth, skin, and intestines; however, this bacterium may act as an opportunistic pathogen, causing severe nosocomial infections such as septicemia, pneumonia, and urinary tract infections in hospitalized and immune-comprised patients with chronic ailments [1,2].
Beta-lactam antibiotics, used as therapeutic agents against a broad range of bacteria, bind to the

Open Access
Gut Pathogens *Correspondence: todaewon@gmail.com † Taesoo Kwon and Young-Hee Jung contributed equally to this work ‡ Won Kim and Dae-Won Kim contributed equally to the work 3 Division of Biosafety Evaluation and Control, Korea National Institute of Health, Cheongju 363-951, Republic of Korea Full list of author information is available at the end of the article penicillin-binding protein and inhibit biosynthesis of the bacterial cell membrane. However, the extended spectrum β-lactamases (ESBLs) and carbapenemases confer resistance to penicillin, cephalosporins, or carbapenem [3,4]. The β-lactamases are divided into four classes on the basis of the Ambler scheme: class A (Klebsiella pneumoniae carbapenemase, KPC; imipenem-hydrolyzing β-lactamase, IMI; Serratia marcescens enzyme, SME; Serratia fonticola carbapenemase, SFC), class B (Verona integron-encoded metallo-β-lactamase, VIM; imipenem-resistant Pseudomonas, IMP; New Delhi metallo-β-lactamase, NDM), class C (AmpC-type β-lactamase, ACT; cephamycinhydrolyzing β-lactamase, CMY), and class D (oxacillinase, OXA) [5] are composed of transposon, cassettes, and integrons and transferred within and between species by HGT (horizontal gene transfer). Numerous carbapenemase-producing bacteria similarly harbor drug resistance genes that are transferred to other strains by horizontal gene transfer [6,7]; infections caused by such multi-drug-resistant bacteria are difficult to treat [8]. The emergence of the novel carbapenemase NDM-1 (the New Delhi metallo-β-lactamase) is of great concern, as no therapeutic agents are available to treat infections caused by NDM-1-producing bacterial strains [9]. NDM-1-producing K. pneumoniae strains were first isolated from a Swedish patient who had travelled to India in 2009 [10]. Since then, NDM-1 has been reported to be produced by various species of Enterobacteriaceae, such as K. pneumoniae, Escherichia coli, Enterobacter spp. and Acinetobacter spp., in numerous countries [11].
The carbapenem-hydrolyzing β-lactamase OXA-232, which was first reported in E. coli and two K. pneumoniae strains [12], belongs to the OXA-48-like family. Carbapenemase-producing Gram-negative bacteria are often multi-drug resistant [13]. K. pneumoniae isolates that coproduce OXA-48-like β-lactamase and NDM-1 have been isolated in numerous countries [14][15][16]. Recently, K. pneumoniae isolates coproducing two carbapenemases, bla NDM-1 and bla  , have been identified in several countries; of these, two isolates originating in India were recovered in the USA and Korea, in January 2013, and sequenced [16,17] but not studied yet the characteristics in the context of genomic contents by comparing these isolates. In the present study, we performed a comparative analysis of the genomes of these isolates.

Isolation and serotyping of strains
In January 2013, a 32-year-old man was hospitalized in the Intensive Care Unit of a general hospital in Seoul, Korea, two days after suffering burns during a visit to India. K. pneumoniae was isolated from his wound and another patient in the same room became infected with the same strain [18]. The K. pneumoniae isolate was identified as the KP617 strain belonging to the sequence type (ST)14, and found to coproduce NDM-1 and OXA-232, which conferred resistance to ertapenem, doripenem, imipenem, and meropenem (MICs: >32 mg/L). The K. pneumoniae strains PittNDM01 [17], NUHL24835 [19], and ATCC BAA-2146 [20] were used as reference strains for comparative genomic analysis.

Library preparation and whole-genome sequencing
Whole-genome sequencing of KP617 was performed using three platforms: Illumina-HiSeq 2500, PacBio RS II, and Sanger sequencing (GnC Bio: Daejeon, Republic of Korea) [16]. Sanger sequencing was used for the construction of a physical map of the genome.

Genome assembly and annotation
A hybrid assembly was performed using the Celera Assembler (version 8.2) [21] and a fosmid paired-end sequencing map was used to confirm the assembly. The final assembly was revised using proovread (version 2.12) [22]. An initial annotation of the KP617 genome was generated using the RAST (Rapid Annotation using Subsystem Technology, version 4.0) server pipeline [23]. The genomes of three K. pneumoniae strains, PittNDM01, NUHL24835, and ATCC BAA-2146, were annotated using the RAST server pipeline. In order to compare the total coding sequences (CDSs) of KP617 with those of the three K. pneumoniae strains, the sequence-based comparison functionality of the RAST server was utilized.

Comparison of genomic structure
The chromosome and plasmids of KP617 and the reference strains were compared using Easyfig (version 2.2.2) [30]. Whole-genome nucleotide alignments were generated using BLASTN to identify syntenic genes. The syntenic genes and genomic structures were visualized using Easyfig. A stand-alone BLAST algorithm was used to analyze the structure of the genes of interest, i.e. the OXA-232-and NDM-1 carbapenemase-encoding genes.

Analysis of virulence factors and phage-associated regions
The virulence factor-encoding genes were searched against the virulence factor database (VFDB) [32] using BLAST with an e-value threshold of 1e-5. Homologous virulence factor genes with a BLAST Score Ratio (BSR) of ≥0.4 were selected. The BSR score was calculated using our in-house scripts. Phage-associated regions in the genome sequences of the four K. pneumoniae strains were predicted using the PHAST server [33]. Three scenarios for the completeness of the predicted phageassociated regions were defined according to how many genes/proteins of a known phage the region contained: intact (≥90 %), questionable (90-60 %), and incomplete (≤60 %).

Quality assurance
Genomic DNA was purified from a pure culture of a single bacterial isolate of KP617. Potential contamination of the genomic library by other microorganisms was assessed using a BLAST search against the non-redundant database.  Table  S1), 342 putative ORFs on plasmid 1, and 9 putative ORFs on plasmid 2 were identified.

General features
Comparison of KP617 and the reference strains based on sequence similarity (percent identity ≤80) showed that 32 genes are unique for KP617, and that most of the functional genes of this strain are also conserved in the reference strains. The genes unique to the KP617 strain, such as the SOS-response repressor and protease LexA (EC 3.4.21.88), integrase, and phage-related protein were identified as belonging to the genome of the prophage Salmonella phage SEN4 (GenBank accession: NC_029015). When the KP617 genome was compared with that of the PittNDM01 strain, which represents the closest neighbor of the former strain on the phylogenetic tree (Figs. 3a, b), 94 genes showed a percent similarity of below 80; most of these were phage protein-encoding genes. These results indicate that the presence of prophage DNA is an important feature of the KP617 genome.

Phylogenetic analysis
The whole-genome phylogenetic analysis indicated that KP617 is evolutionarily close to PittNDM01 and NUHL24835, and that the strains belong to the WGLW2 group. However, KP617 was found to be evolutionarily distant from ATCC BAA-2146 (Fig. 3). Concordantly, MLST-based phylogenetic analysis revealed that while KP617, PittNDM01, and NUHL24835 belong to the same group [sequence type (ST)14], ATCC BAA-2146 belongs to the HS11286 group, ST 11 [20]. The only difference between the whole-genome phylogenetic tree and the MLST-based phylogenetic tree was the divergence time within the same group; MLST-based phylogeny did not reveal the minor details of genomic evolution such as the divergence between KP617, PittNDM01 and NUHL24835 in the whole-genome phylogeny. The difference was attributed to horizontal gene transfer in regions not covered by the MLST genes.

Comparison of genome structures
The comparison of genomic structures of the chromosome indicated the presence of highly conserved structures in the KP617, NUHL24835, and PittNDM01 strains (Fig. 4a). Interestingly, a 1-Mb region (233,805-1,517,597) of the KP617 chromosome was inverted relative to its arrangement in the chromosome of PittNDM01 (1,500,972-225,619). Despite this inversion, KP617 and PittNDM01 exhibited a lower substitution rate (score 20) than NUHL24835 (score 30) (Fig. 3). However, the chromosomal structure of the ATCC BAA-2146 strain, which consisted of two large inverted regions, was significantly different from that of the other strains. In addition, a 71 Kb inversion was found in the sequence of plasmid 1 of KP617 (18,633-90,686) relative to plasmid 1 of PittNDM01 (91,507-19,453); however, the two plasmids were highly homologous to each other (Fig. 4b).

Antimicrobial resistance genes
Nineteen antibiotic resistance genes were identified in the genome of KP617, 39 in the genome of PittNDM01, 29 in that of ATCC BAA-2146, and nine in that of the NUHL24385 strain ( Table 2). The β-lactam resistance genes in the KP617 genome were bla OXA-1 and bla  in the chromosome, bla NDM-1 in plasmid 1, and bla OXA-232 in plasmid 2; however, genes conferring resistance to colistin and tetracycline were not found ( Table 2). Plasmid 2 of KP617, which includes the OXA-232-encoding gene, consists of a 6141-bp sequence; the sequence of this plasmid was identical to that of plasmid 4 of PittNDM01 (100 % coverage and similarity) and the plasmid of E. coli (coverage: 100 %, similarity: 99.9 %). Plasmid 2 of KP617, plasmid 4 of PittNDM01 and E. coli Mob gene cluster (GenBank accession: JX423831) [12] carried the OXA-232-encoding gene, and pKF-3 of K. pneumoniae carried the OXA-181-encoding gene. However, pKF-3 was identical to plasmid 2 of KP617, except in that the insertion sequence ISEcp1 was inserted upstream of OXA-181 and included in the transposon Tn2013 [12,34]. The structure of plasmid 1 (273,628 bp in size) of the KP617 strain was similar to that of plasmid 1 (283,371 bp in size) of PittNDM01. A region of about 40 kb in size within plasmid 1 of the KP617 strain, which included the NDM-1-encoding gene, was composed of various resistance genes such as aadA2, armA, aac(3″)-VI, dfrA12, msrE, mphE, sul1 and qnrB1, and identical (coverage: 100 %, homology: 100 %) to a 40-kb sequence of plasmid 1 of PittNDM01 (Fig. 4b). Adjacent to the NDM-1-encoding gene, a region of about 70 kb in size was inverted in plasmid 1 of KP617 relative to plasmid 1 of PittNDM01. In addition, the OXA-1-encoding gene was identified in PittNDM01 but not in KP617. Transposases were found in a part of the NDM-1-encoding gene cluster (about 10 kb) in plasmid 1 of KP617. Gram-negative bacteria are known to possess a diverse range of transposases; moreover, the sequence of the NDM-1-encoding gene cluster includes a transposon [35,36]. The partial, or complete, transfer of NDM-1-harboring plasmids between K. pneumoniae and E. coli, via conjugation, has been shown to result in the emergence of strains resistant to several antimicrobial agents [11,32,36,37].
Following the initial identification of NDM-1 in a K. pneumoniae isolate from a patient who had travelled to India in 2008, most NDM-1-producing K. pneumoniae isolates have been recovered from patients associated with India; however, in some cases, these strains have been isolated from patients with no history of travelling abroad, or any association with India [38]. These observations suggest that the transfer of the NDM-1-and OXA-232-harboring plasmids between Gram-negative bacteria has resulted in the spread of carbapenem resistance and emergence of strong carbapenem-resistant strains outside the Indian subcontinent.

Virulence factors
Klebsiella pneumoniae, a significant pathogen of human hosts, causes urinary tract infections, pneumonia, septicemia, and soft tissue infections [1]. The clinical features of K. pneumoniae infections depend on the virulence factors expressed by the infecting strain [39]. Therefore, we investigated the virulence factors of the present strain and compared these with those of KP617 and the reference strains. A BLAST search was performed against VFDB to identify 117 virulence factors harbored by the KP617 strain (Table 3). All 117 virulence genes of KP617 were also harbored by the reference strains; KP617 did not possess any unique virulence factors. The PittNDM01 strain was also found to possess no unique virulence factors; however, NUHL24835 and ATCC BAA-2146 possessed 3 and 7 unique virulence factors, respectively. The 117 virulence genes of KP617 were classified into 31 the following categories: Iron uptake (30 genes), Immune evasion (12 genes), Endotoxin (11 genes), Adherence (10 genes), Fimbrial adherence determinants (8 genes), Toxin (7 genes), Antiphagocytosis (6 genes), Regulation (5 genes), Acid resistance (3 genes), Anaerobic respiration (2 genes), Cell surface components (2 genes) and Secretion system (2 genes). Among the 117 virulence genes identified, 8 genes were lipopolysaccharide [40]-related genes and 4 genes were capsular polysaccharide [41]-related.
KP617 and PittNDM01 were found to possess two virulence factors that were not present in the other two strains: invasion (encoded by ail, attachment invasion locus protein) [42] and Iron uptake (encoded by fyuA, Yersiniabactin siderophore) [43].

Phage-associated regions
Prophages contribute to the genetic and phenotypic plasticity of their bacterial hosts [44] and act as vehicles for the transfer of antimicrobial resistance genes [45] or virulence factors [46]. Six phage-associated regions (KC1-KC5) of the KP617 chromosome and one phageassociated region (KP1) in plasmid 1 of the KP617 strain were identified using the PHAST algorithm (Table 4). With regard to the reference strains, six phage-associated regions were identified in the PittNDM01 strain, six in NUHL24835, and 12 in ATCC BAA-2146.
Three of the six phages, KC1, KC2 and KC3, in the KP617 strain were intact, whereas the remaining prophages were incomplete (KC5 and KP1) or questionable (KC4) and had a low PHAST score of below 90. Based on the sequence similarity of their genomes,  KP617 and PittNDM01 were found to have high similarity to each other (Figs. 2, 3a, b). Concordantly, the profile of prophage DNA in their genomes, as determined via a BLAST search, was similar, and the two strains shared four of the six prophages, whereas two phage regions, KC2 (Entero_HK140) and KC3 (Salmon_SEN4), were specific to the KP617 genome. Furthermore, it was found that one phage-associated region of KP617, namely KC2 (Entero_HK140), exhibited a high similarity to the phage-associated region of the NUHL24835 strain, NC1, with 60 % query coverage and 99 % identity. It should be noted that the strains compared in the present study, i.e.
KP617 and the reference strain, ATCC BAA-2146, had no prophages in common. Investigation of the antimicrobial resistance genes harbored by the strains, which was performed using ResFinder, and comparison with the prophage-associated region, as predicted using PHAST, did not reveal the presence of a prophage-delivered beta-lactamaseencoding gene in the KP617 genome, indicating that prophages did not act as a vehicle for the transfer of antimicrobial resistance genes in this strain. This finding is consistent with previous observations that betalactamase-encoding genes are borne by transposons [35, 36]. Bacteriophages are applicable to phage therapy. In particular, bacteriophages have been used as a potential therapeutic agent to treat patients infected with multidrug resistant bacteria [47] and have been used for serological typing for diagnostic and epidemiological typing in K. pneumoniae [48]. However, because we did