Genomic characterization of an emerging blaKPC-2 carrying Enterobacteriaceae clinical isolates in Thailand

The rapidly increasing prevalence of carbapenem-resistant Enterobacteriaceae (CRE) over the past decade has increased concern in healthcare facilities and the impact on public health. The prevalence of blaKPC (KPC) in Thailand remains very low; only blaKPC-13 has been described previously. This study is the first to describe the characteristics of blaKPC-2-carrying Klebsiella pneumoniae, Escherichia coli, and Enterobacter asburiae in Thailand. The prevalence rate of blaKPC-2-carrying isolates was 0.13% among CRE isolates in our study. Based on carbapenem susceptibility testing, K. pneumoniae C1985 was resistant to meropenem and ertapenem, E. coli C1992 was resistant to meropenem, imipenem, and ertapenem, and E. asburiae C2135 was only resistant to imipenem. K. pneumoniae C1985 carried blaKPC-2, blaSHV-11, fosA, oqxA, and oqxB, while E. coli C1992 contained blaKPC-2 and mdf(A) and E. asburiae C2135 harbored blaKPC-2, blaACT-2, and qnrE1. The genetic features of blaKPC-2 in the 3 isolates revealed identical rearrangement and flanking regions. Analysis of genomic sequences from these 3 isolates revealed that the sequence types of K. pneumoniae C1985, E. coli C1992, and E. asburiae C2135 were ST4008, ST7297, and ST1249, respectively. The 3 blaKPC-2 isolates were from individual living cases. Two cases were colonization for K. pneumoniae C1985 and E. asburiae C2135 and the third case was hospital-acquired infection of E. coli C1992. Although the prevalence of blaKPC-2-carrying CRE is relatively low in this study, continued surveillance and close monitoring are warranted. In addition, prompt or early detection of CRE and strict implementation of infection control are essential to prevent outbreaks or rapid spread in hospitals.


Genomic characterization of bla KPC-2 -carrying isolates.
In September 2015, we established a surveillance network to detect carbapenemase genes in CRE isolates from hospitals in 11 provinces in Thailand consisting of: Surin, Udonthani, Sakhon Nakhon, Nakhon Phanom, Roi Et, Bueng Kan, Mukdahan, Tak, Phayao, Chumporn, and Surat Thani. The network is still operational. Of 2,245 CRE isolates, 3 isolates harboring bla KPC (0.13%) were detected using multiplex PCR. These isolates were from a hospital in Surin province, northeastern Thailand. Biochemical identification revealed 2 isolates were K. pneumoniae (strains C1985 and C2135) and the other isolate was E. coli (strain C1992). However, matrix-assisted laser desorption ionization-time of flight mass spectrometry (MALDI-TOF MS Biotyper, Bruker, Daltonics, Germany) identification revealed isolates no. C1985, C1992, and C2135 were K. pneumoniae, E. coli, and E. asburiae, respectively. Comparison of genomic sequences using KmerFinder and the ANI calculator demonstrated that isolate C2135 was Enterobacter asburiae, while isolate C1985 was K. pneumoniae and isolate C1992 was E. coli.
Analysis of genomic sequences from these 3 isolates revealed the sequence types (STs) of K. pneumoniae C1985, E. coli C1992, and E. asburiae C2135 were ST4008, ST7297, and ST1249, respectively. K. pneumoniae ST4008 and E. asburiae ST1249 are novel STs. Analysis based on the MLST database revealed that K. pneumoniae (C1985) ST4008 was a double-locus variant of gapA and tonB for ST399, ST3630 and ST3413, mdh and tonB for ST1037, phoE and rpoB for ST1451, phoE and tonB for ST2258, and gapA and rpoB for ST3612. The phylogenetic tree constructed using concatenated sequences of the 7 MLST genes showed that ST4008 was very close to ST1451 (Fig. 2). E. coli (C1992) ST7297 was closely related to ST4948 by a single locus variant (mdh) and this ST was a double locus variant (gyrB and icd) of ST2805 concordant with phylogeny analysis (Fig. 2). E. asburiae (C2135) ST1249 was closely related to E. cloacae ST384 with a single locus variant (dnaA) and this ST had no double-locus variant with any of the STs in the MLST database (Fig. 2). A triple locus variant of ST1249 was found in ST23 and ST637. The phylogenetic tree indicated that ST1249 was closely related to ST384 (Fig. 2).
As shown in Fig. 3, based on a reference genome-based single nucleotide polymorphism (SNP) K. pneumoniae C1985 was related to K. pneumoniae strain 3189STDY5864938 (ST2258) isolated from human blood in Pakistan. E. coli C1992 was more related to E. coli strain 1011 than strains 31111 and MH17-008M. The strain 1011 was isolated from urine of patient with urinary infection in Germany, whereas strains 31111 and MH17-008M were  www.nature.com/scientificreports www.nature.com/scientificreports/ from Vietnam, having been isolated from human feces and blood, respectively. E. asburiae C2135 isolate was related to E. asburiae strain ENIPBJ-CG1; this isolate was from a bone marrow transplant patient in China.

Discussion
Due to a rise in CRE globally, there are gradually fewer options for treatment. Carbapenemases is a major key mechanism for resistance to carbapenems and other β-lactam antimicrobial agents. The prevalence of CRE and the carbapenemase species such as KPC, NDM, IMP, and OXA-48 family enzymes is dependent on geographic region. KPC and its variants are mostly found in the USA, South America, and European countries 7 . NDM is concentrated in Asia (India, Pakistan, Bangladesh, China), and Southeast Asian countries 7 . The OXA-48-like enzymes are mainly found in European countries (France, Germany, the Netherlands, Italy, and the United Kingdom), in the Middle East (Turkey), and in Mediterranean countries including North Africa (mainly Morocco, Tunisia, Egypt, and Libya) 7 . In Thailand, bla NDM , bla OXA48-like and bla IMP-14 are frequently detected in clinical Enterobacteriaceae isolates 3,6 .
However, the prevalence of bla KPC (KPC) in Thailand remains very low; only bla KPC-13 has been described previously 4,8 , with the later study reporting rates of bla KPC-13 -carrying isolates of 0.02% among Enterobacteriaceae and 1.7% among CRE isolates. This current study is the second report on KPC in Thailand and we discovered bla KPC-2 for the first time in Thai patients. The prevalence rate of bla KPC-2 -carrying isolates was 0.13% among CRE isolates (n = 2,245) in our study, which was lower than in the previous report 8 . Other reports showed high prevalence rates of bla KPC-2 -carrying Enterobacteriaceae of 14.6% in Taiwan 9 , 71.4% in Eastern China 10 , 25.3% in Australia 11 , and about 40-65% in Greece 12 . The bla KPC-2 -carrying patients in the current study were separately admitted in different wards, indicating that there was no relationship or contact among them. Indeed, the genetic organization of bla KPC-2 was similar among the isolates even though they were different species. We hypothesized that these isolates may have received this gene from the same source. However, complete plasmid sequences should be done and aligned to confirm this hypothesis. In addition, further study should investigate for ability of bla KPC-2 conjugative transferable and enzyme kinetics.
The bla KPC -carrying K. pneumoniae strains can be characterized by their clonality, in that the majority of the strains circulating globally belong to clonal complex (CC) 258 as defined by MLST. The most common ST is ST258 in the USA, but other STs in CC258 such as ST11, ST340, ST437, and ST512 predominate in countries outside the USA 1 . In Singapore, bla KPC-2 -carrying K. pneumoniae were associated with non-ST258 isolates, namely ST11, ST14, ST17, ST23, ST48, ST65, ST70, ST86, ST231, ST273, ST792, and ST1117 13 . Vargas et al.
(2019) reported 40% of K. pneumoniae ST25 (CC65) carried bla KPC-2 in Argentina. In China, bla KPC-2 -carrying ST11 in K. pneumoniae isolates were the most frequent 10,14 . K. pneumoniae ST86 was reported to be carrying bla KPC-2 in Canada 15 . Our study was different from those reports, because we revealed a novel ST4008 of K. pneumoniae harboring bla KPC-2 . This ST was closely related to ST1451, a strain susceptible to carbapenem and isolated from human blood in North America (https://bigsdb.pasteur.fr/cgi-bin/bigsdb/bigsdb.pl?db=pubmlst_kleb-siella_seqdef&page=query). SNP phylogeny demonstrated that K. pneumoniae C1985 (ST4008) is related to strain 3189STDY5864938 from Pakistan; this is ST2258. MLST revealed that our isolate was closely related to ST1451; however, the genomic sequence of ST1451 is not available in the database yet, so comparison to these genomes is not possible at this time.
E. coli carrying bla KPC-2 has been reported in ST131, ST410, ST457, and ST5700 from Italy, Greece, China, Mexico, and Tunisia [16][17][18][19][20][21] , respectively, whereas our study involved detection in ST7297. This ST is closely related to ST4948 and both have the same ancestor similar with ST2805 according to the phylogenetic tree of concatenated MLST allelic sequences. SNP analysis revealed that E. coli C1992 (ST7297) was related to E. coli strain 1011 isolated from Germany; this strain is ST131. Due to there being no genomic sequence of ST4948 in the database yet, comparison between the genomes of these isolates is not possible at this time. Finally, bla KPC-2 was found in www.nature.com/scientificreports www.nature.com/scientificreports/ several species of Enterobacter such as E. asburiae, E. cloacae, and E. hormaechei [22][23][24] , similar to our study. SNP phylogeny revealed that E. asburiae in our study was related to the isolate from China.
In conclusion, we reported the emergence of bla KPC-2 -harboring Enterobacteriaceae isolates. Although the prevalence of this type of CRE is relatively low, continued surveillance and close monitoring are warranted. In addition, prompt or early detection of CRE and strict implementation of infection control are essential to prevent outbreaks or rapid spread in hospitals.

Methods
Bacterial isolates and case information. Three isolates (K. pneumoniae C1985, E. coli C1992, and E. asburiae C2135) each carrying bla KPC-2 were used to characterize this study. Clinical information of patients with these isolates was reviewed by the attending physicians at the hospital.
Microbiological methods. Presumptive conventional biochemical tests described elsewhere were used for presumptive K. pneumoniae and Enterobacter spp., identification, including indole production, methyl-red reaction, Voges-Proskauer reaction, citrate utilization, and urea hydrolysis 25 . An E. coli identification kit (U & V Holding, Thailand) was used for confirmation of E. coli according to the manufacturer's instruction. MALDI-TOF MS Biotyper (Bruker, Daltonics, Germany) was used to confirm E. coli, K. pneumoniae, and E. asburiae. Carbapenemase genes were detected using multiplex PCR 26 .
Whole genome sequencing and analysis. The whole genome sequencing of the three bla KPC-2 isolates was carried out on the Illumina platform with massively parallel sequencing Illumina technology at Beijing Novogene Bioinformatics Technology Co., Ltd. (Beijing, China). A-tailed, ligated to paired-end adaptors and PCR amplified with a 500 bp insert and a mate-pair library with an insert size of 5 kb were used for the library construction.
Illumina PCR adapter reads and low-quality reads from the paired-end and mate pair library were filtered by the step of quality control using the company compiling pipeline. All good quality paired reads were assembled using the SOAP de novo (http://soap.genomics.org.cn/soapdenovo.html) into a number of scaffolds. The assembly result was gap-filled and optimized by the company using software such as krskgf and gapclose.
Multilocus sequence typing (MLST) analysis for sequence types of E. coli, E. cloacae complex, and K. pneumoniae used MLST 2.0 (https://cge.cbs.dtu.dk/services/MLST/) 38 and Institut Pasteur MLST (https://bigsdb. pasteur.fr/klebsiella/klebsiella.html), respectively. Construction of phylogenetic trees for all STs in this study was performed via Phylogeny.fr 39 . The genomic comparison of these isolates was conducted using a reference genome-based single nucleotide polymorphism (SNP) strategy with BacWGSTdb 40 (http://bacdb.org/ BacWGSTdb/). Accession number. The genomic sequences assembled were deposited under the Bioproject accession number of PRJNA525849 with Sequence Read Archive (SRA) accession numbers of SRR8690013 (plasmid contigs) and SRR9722215 (chromosomal contigs) for K. pneumoniae C1985. The genomic sequence of E. coli C1992 was deposited under the Bioproject accession number of PRJNA525942 with SRA accession numbers of SRR8693447 (plasmid contigs) and SRR9722217 (chromosomal contigs), respectively. The genomic sequence of E. asburiae C2135 was deposited under the Bioproject accession number of PRJNA555780 with the SRA accession number of SRR9729532. ethical approval. This study was reviewed and approved by the Surin Hospital Ethics Review Board (ERB).
The medical records of these 3 cases were reviewed by attending physicians at the hospital using the clinical case record form approved by the ERB. The ERB waived requirement for informed consent as the study satisfied the conditions of the policy statement on ethical conduct for research involving humans. This study was conducted according to the principles expressed in the Declaration of Helsinki.