Whole Genome Analysis of the Extensively Drug Resistant Clinical Klebsiella Pneumonia Isolates from the Military Hospital in Vietnam

Background The over using of antibiotics in hospitals, communities and agriculture has raised drug resistance in Vietnam. Recently, Klebsiella pneumoniae (K.pneumoniae) emerged as the main agent causing hospital infections because of the alarming resistance to carbapenem and high mortality. The purpose of this study was to identify the antibiotic resistance pattern, gene carbapenemase expression and mutation that related to drug resistance mechanisms of K.pneumoniae isolated from clinical samples in the Military Hospital 103 in Vietnam. Seventeen extensively drug resistant (XDR) strains were completed whole genome sequence, analyzed by bioinformatics, and conrmed using PCR and Sanger sequence methods. K.pneumoniae showed resistance to almost fteen antibiotics with very high rates of over 70%, only sensitive to colistin 100%. There is a correlation between carpabenem resistance and XDR status in the K.pnemoniae group. Analysis Multilocus sequence typing (MLST) reported the total of 9/17 strains belong to ST15, after that ST11 (4/17), ST656 (2/17), ST 16 (1/17) and ST395 (1/17). Most of the K.pneumoniae isolates harbored the widely distributed ESBL genes, e.g. NDM, SHV, OXA, KPC, TEM and CTX-M, in which, the most prevalent carbapenemase genes were NDM-1, NDM-4, OXA-48, OXA-181 and KPC-2. Especially, two strains demonstrated resistance to a wide range of antibiotics and showed the novel drug resistance gene mutation, namely Pro187Ala in NDM-4 gene and Gly25Ser in KPC-2 gene of MH16-335M and MH17-011M, respectively. Our study adds new and signicant DNA sequence data on K.pneumoniae strains and demonstrates the value of whole-genome sequencing in characterizing multidrug resistance in clinical isolates. MIC of the antibiotics carbapenem was measured for all isolates using Etest (bioM'erieux) according to the manufacturer’s instructions. The results were used to classify the isolates as either susceptible, intermediate or resistant to the tested antibiotics, based on CLSI (Clinical and Laboratory Standard Institute) breakpoints [23]. Antibiotics used as antibiotics include: Ampicillin (AM), Ampicillin/ Clavulanic acid (AMC), Piperacillin/tazobactam (TZP), Cefotaxime (CTX), Ceftazidime (CAZ), Cefepime (FEP), Amikacin (AN), Gentamicin (GM), Ciprooxacin (CIP), Noroxacin Fosfomycin (FOS), Nitrofurantoin (NIT), Sulfamethoxazole/Trimethoprim (SXT), Ertapenem (ETP), Imipenem (IPM), Meropenem (MEM), and Doripenem (DOR). genomes from K.pneumoniae strains the ANI (Average Nucleotide Identity) data, a 16S rRNA-based subspecies analysis conducted to determine the genetic relationships of the strains using in research. Fig. phylogenetic species of all K.pneumoniae isolates closely related in a group of K.pneumoniae strains ATCC_13883,11296, 13884 (77–97 bootrap values), and to the family Enterobacteriaceae. of K.pneumoniae isolated K.pneumoniae Antimicrobial susceptibilities were performed for samples by the Vitek 2 system and the AST-GN card, E-test, and Broth microdilution (BMD) methods. The isolates were screened against with 18 antibiotics. The results showed that all K.pneumoniae isolates were still susceptible to colistin (100%); highly resistant to ampicillin, sulfamethoxazole/trimethoprim, and ertapenem (82.4–94.1%); and completely resistant (100%) to ampicillin, piperacillin/tazobactam, cefotaxime, certapenem, cefepime, ciprooxacin, and noroxacin.


Introduction
In the near decades, the problem of concern has been the rapid increase of strains of bacteria resistant to many new antibiotics, especially resistant strains of Gram-negative bacteria such as Acinetobacter baumannii, Pseudomonas aeruginosa and Klebsiella pneumoniae [1]. Currently, antimicrobial resistance of Gram-negative bacilli is a persistent issue [2], and becoming more and more di cult to treat [3]. Over time, resistance develops with all antibiotics and spreads.
The increasing clinical incidence of antibiotic resistant bacteria is a serious issue for global health care. Understanding the characterization of antibiotic resistance determinants at the genomic level plays a critical role in potentially controlling the spread of multidrug resistant pathogens [4].
The rapid dissemination of multiple drug resistance (MDR) and XDR or pan drug resistant (PDR) K.pneumoniae has recently recognized by the CDC as an urgent public health threat, thus requiring immediate and aggressive actions [5]. In immunocompromised individuals, neonates and the elderly, K.pneumoniae can cause severe hospital or community acquired infections (Brisse et al., 2009). The current antimicrobial resistance and susceptibility of K.pneumoniae bacteria have been observed clinically [6]. Some clinical K.pneumoniae isolates are resistant to all available antibiotics [7]. The acquisition of antibiotic resistance genes and intrinsic resistance of bacteria has consequently limited treatment options for infections caused by K.pneumoniae. Currently, K.pneumoniae strains producing Extended Spectrum Beta-Lactamases (ESBLs) and carbapenemases have spread globally [8]. The frequent outbreak of nosocomial infections due to extended spectrum β-lactamase of K.pneumoniae that is attributed to multiple drug resistance mechanisms [9]. The mechanisms involved when β-lactamase hydrolyzes β-lactam antibiotics or inactivates β-lactams have been investigated [10]. However, no precise molecular biological mechanisms underlying the antimicrobial drug resistance of K.pneumoniae have been reported [11]. In the previous several studies indicated that antibiotic resistance genes originate in microorganisms and subsequently integrate into the genome of other pathogens through transduction and/or transformation to produce a multitude of different antibiotic resistance strains that are especially dangerous for the course of treatment for patients [12][13][14].
Antimicrobial resistance genes in K.pneumoniae is also associated with high rates of morbidity and mortality in clinical patients [15]. Therefore, understanding the mechanisms involved in the antibiotic resistance of K.pneumoniae are highly required to warrant the development of novel antibiotics or alternative treatment strategies to decrease mortality of patients infected with K.pneumoniae.
Carbapenem is the widest antimicrobial spectrum and stable activity to against bacteria that produce AmpC β-lactamase and ESBL antibiotic enzymes, therefore, this antibiotic is considered last treatment options in the treatment of infections caused by multiple antibiotics resistant gram-negative bacteria [16].
However, the increasing of carbapenem resistant (CRE) isolates parallelly with their global distribution highlights the rapid dissemination of MDR genes to other highly virulent nosocomial pathogens [17]. Carbapenemase genes are frequently encoded on plasmids or other mobile genetic elements that can potentially be carrying additional resistance genes against other classes of antibiotics [18]. Besides, the major contributing factor to the rapid global spread of carbapenem resistance is transmissible plasmids that carry carbapenemase genes [19,20], namely, K.pneumonia carbapenemase (KPC), New Delhi metallo-blactamase (NDM), and oxacillinase (OXA) type carbapenemases [21]. The other genetic mechanisms of carbapenem resistance include e ux pumps, altered function or expression of porins and penicillin-binding proteins (PBPs) [22].
Whole-genome sequencing (WGS) to be established as a viable option to antimicrobial susceptibility characterization of the clinical isolates, the association between the resistance genotype and phenotype, and the characterization of carbapenem resistance factors and to assess the risk for horizontal transfer and dissemination of carbapenemase genes. This is especially important to many carbapenem-resistant pathogens which often exhibit complex phenotypic pro les with resistance to multiple classes of antibiotics. In addition, the use of such high throughput technologies to study pathogens has brought a breakthrough in surveillance, monitoring and diagnostics in public health. In this study, we used WGS in understanding antimicrobial resistance, identifying resistance determinants and predicting genotype resistance in K.pneumoniae isolated from patients who were admitted to the Military Hospital 103 (MH103), a military central hospital in Northern Vietnam.

K.pneumoniae sample collection
The seventeen K.pneumoniae strains analysed in the present study were isolated between 2015 and 2017 in the Military Hospital 103, and were selected due to their extensively drug resistant phenotypes. The strains originated from inpatients in hospital, who were sampled due to signs of infection. The identi cation of K.pneumoniae isolate was done by analyzing colony morphology in special culture media, microscopic examination, performing biochemical testing, and Vitek 2 system (bioMérieux, Marcy l'Étoile, France).
Resistant antibiotic phenotype identi cation of K.pneumoniae strains Testing of antibiotics susceptibility and carbapenemase producing capacity used the Vitek 2 compact system and E-test method. Antimicrobial susceptibilities were performed for 17 K.pneumoniae isolates using the Vitek 2 system and the AST-N204 card (bioMérieux, Marcy l'Étoile, France  DNA extraction and species identi cation of K.pneumoniae strains Total bacterial DNA was extracted from the bacteria that enriched on blood agar medium and incubating overnight. The DNA was extracted using a GeneJET Genomic DNA Puri cation Kit (Thermo Fisher Scienti c) according to the manufacturer guideline. The quality and quantity of DNA were assessed using NanoDrop ND 1000 spectrophotometer, Qubit 2.0 Fluorometer (Thermo Fisher Scienti c), Agilent 2100 Bioanalyzer system (Agilent), visualization on an agarose gel 1% and yield concentration reached 200 ng / uL before sending DNA for sequencing. For the con rm of WGS data, DNA samples, primers of gene targets were prepared for PCR according to the protocol description. The 16S rRNA gene of seventeen isolates was sequenced in the ABI3500 system (Applied Biosystem). The isolate sequences in this study, and 23 published reference sequences including twenty two Enterobacteriae strains and one out group strain were aligned using MUSCLE algorithms. The maximum likelihood tree with T92 + G substitution model and bootstrap 1000 were built to determine the genetic relationship. All analysis processes were performed using MEGA6 software [24].
Whole genome sequencing setup All isolates were index tagged and sequenced at the National Institute of Infectious Diseases (Tokyo, Japan), using the HiSeq4000 system (Illumina) with the High Output Reagent Kit (300 cycles) and the MinION sequencer (Oxford Nanopore Technologies) with the R9.4.1 owell. The Nextera XT DNA Library Preparation Kit (Illumina) was used for preparing Illumina sequencing (150 bp paired-end, insert size of 500-900 bp), and the library for MinION sequencing was prepared using the Rapid Sequencing Kit (SQK-RAD004) according to the manufacturer recommendations. Basecalling for MinION reads was carried out by Guppy v.3.6.1.
Con rmation and mutation analysis of antibiotic resistance genes in K.pneumoniae isolates using PCR and Sanger sequencing The distribution of genes associated with antimicrobial resistance was identi ed for all K.pneumoniae genomes. Acquired genes associated with antimicrobial resistance were identi ed using the ResFinder database (Kleinheinz et al., 2014). The ResFinder database included genes that may confer resistance to the following antibiotics: aminoglycosides, beta-lactams, colistin, uoroquinolone, fosfomycin, fusidic acid, glycopeptide, macrolide, lincosamide, and streptogramin B (MLS), nitroimidazole, oxazolidinone, phenicol, rifampicin, sulphonamide, tetracycline, and trimethoprim. The presence of carbapenem resistance related genes such as NDM-1, NDM-4, KPC-2, OXA-181, and OXA-48 in 17 K.pneumoniae isolates was examined by PCR with speci c primers (Table 1). Mutations within the NDM-1, NDM-4, KPC-2, OXA-181, and OXA-48 genes that are known to confer resistance to carbapenem were also investigated using Sanger sequencing.  Based on the genomes from the K.pneumoniae strains published on the ANI (Average Nucleotide Identity) data, a 16S rRNA-based subspecies analysis was conducted to determine the genetic relationships of the strains using in research. The Fig. 1     Assembling and annotating genome sequence The cleaned and paired reading datasets were used to assemble the de novo genome using the CLC Genomics Workbench v.9.0 (Qiagen), resulting in K.pneumoniae genome sizes were 5,533,931 bp to 5,907,064 bp in length; N50 is from 44,277 to 198,000 bp and the number of contig contents from 132 (MH17-556D) to 377 (MH15-258M). In silico MLST typing was performed on all assemblies using the previously described Pasteur system [30]. All of K.pneumoniae isolates analyzed, nine were assigned to ST101, four were assigned to ST15, two were assigned to ST656, and one sample such as MH17-010D or MH15-208H was assigned to ST16 and ST395, respectively (Table 4). In silico antimicrobial resistance genes pro ling From the analysis of genomic annotation using the DFAST server (https://dfast.nig.ac.jp), antimicrobial resistance genes and function were detected using ResFinder v.4.0 and Patric PlasmidFinder v.2.1, respectively (http://www.genomicepidemiology.org). Genes responsible for conferring aminoglycoside resistance, beta-lactam resistance, uoroquinolone, aminocyclitol, folate pathway antagonist, macrolide, rifamycin, and tetracycline resistance were presented across all genomes of K.pneumoniae (Table 5).   Supplementary Table 1. Brie y, uoroquinolone resistance genes (oqxA, oqxB, and aac(6')-Ib-cr) were identi ed within every K.pneumoniae genome. Genes resistance for aminoglycoside and fosfomycin resistance were identi ed in 100% of K.pneumoniae genomes as aac(6')-Ib-cr and fosA, respectively. Additionally, other aminoglycoside resistance genes were shown such as rmtB, aac(6')-Ib/Ib3, aac(3)-IIa/IId, aadA, aph(6)-Id, and aph(3'')-Ib. A total of 8 different beta-lactam genes was shown in isolates (i.e., CTX-M, SHV, OXA, KPC, NDM, TEM, DHA, and LAP).

Analysis of molecular characterization of the relevant carbapenem resistance genes using PCR and Sanger sequencing
There is little knowledge about carbapenemase-producing K.pneumoniae in Vietnam. The con rm characterization of presumptive carbapenem resistant genes from WGS analysis were used polymerase chain reaction (PCR) and Sanger sequencing. Out of 17 K.pneumoniae that showed resistance against the test antibiotics, 94.1% isolates were con rmed to harbour carbapenem resistance genes and the result is shown in Table 5. The mutations in carbapenem resistance genes (NDM-1, NDM-4, KPC-2, OXA-181, and OXA-48) were screened by DNA sequencing. Results of analysis and comparison of the sequence of the KPC-2 gene encoding for KPC-2 enzyme of K.pneumoniae isolates with the reference K.pneumoniae genome data (NG 049253.1) using bioinformatics software, indicated the missens change in Gly25er (acid amin 25 Glycine changed to a Serine) in MH17-011M (Fig. 2). And the mutation in the location of NDM-4 gene was Pro187Ala (amino acid 187 Proline changed to an Alanine) in MH16-335M (Fig. 3). Other OXA genes were not shown any mutation in K.pneumoniae isolate sequences (Fig. 4).

Discussion
The tracking resistance patterns, changing the prescribing habits, and increasing the infection control are the best way in reducing the development of bacterium antibiotic resistance, improving and saving in treatment to patients and health care facilities. The nosocomial infections by K.pneumoniae are still prevalent, and may be more dangerous due to the rapid development and spread of antimicrobial resistance in hospital [31]. In this study, most of the XDR K.pneumoniae isolates were collected from male patients. This result was similarly with a survey data from health care centers in Nigeria, here K.pneumoniae infection was higher in males than females [32]. Akter et al. also reported that male patients had a higher risk to get Klebsiella infection than females J Akter, AMMA Chowdhury and MAI Forkan [33]. There may be a association in poor lifestyle choices, smoking and alcoholism between male and female [32]. In addition, most of K.pneumoniae in this study were isolated from patients aged 23 to 92 years old. In the previous studies revealed a greater number of K.pneumoniae isolates were obtained from patients aged among 40 to 65 years or over 70 years old [34,35]. The differences in age distribution of patients may be related to the response of the immune system, such as under 40 years have stronger immune systems, and more pressure to ght to K.pneumoniae [36]. Besides, an increasing age leads to increase the comorbid illness cause of a higher risk of K.pneumoniae infection [37]. These results also revealed that K.pneumoniae isolates were mainly isolated from ICU and respiratory specimens (namely is sputum samples). Ashurst and Dawson [38] emphasized that K.pneumoniae typically colonizes human mucosal surfaces of the oropharynx. For this reason, K.pneumoniae is considered to be the most common cause of hospital acquired pneumonia in the Vietnam [39]. The nding is in line to a study by Wang et al., which reported that the respiratory tract was the main infection site of K.pneumoniae in China [40].
Antimicrobial resistance bacteria are commonly associated with nosocomial infections cause of the overuse of antibiotics, and without monitoring or control [41]. In this study, most of K.pneumoniae was resisted to various antibiotics, fully resistant to ampicillin, piperacillin/tazobactam, cefotaxime, ceftazidime, cefepime, cipro oxacin, and nor oxacin, whereas colistin, amikacin, nitrofurantoin, and fosfomycin were being the most effective for K.pneumoniae. The nding was higher than other results, Cepas et al. reported that 40% of K.pneumoniae strains were resisted to cipro oxacin and amoxicillin-clavulanic acid [42]. Manjula et al. showed that 90.2% of K.pneumonia isolates were MDR, and high resistance to penicillin, cephalosporin, uoroquinolone, aminoglycoside, and sulfonamide [43]. These MDR microbes are causing a great challenge in controlling infections, thus it is very important to monitor and optimize antibiotic use through antibiotic management programs [44,45].
In the resistant phenotype of K.pneumoniae, the XDR phenotype showed resistance to most antibiotics, especially strong resistance to carbapenem, evidencing that they hold a variety of resistance mechanisms. They are not only secreting the enzyme that destroys carbapenem but also showing the resistance mechanisms such as the porin channel closes and the pump expels antibiotics out. Previous studies have shown that antibiotic consumption leads to selective pressure increasing β-lactam resistance in the genus Enterobacteriaceae bacteria [46]. However, XDR may also a result of gene circulation in the bacterial population. Currently in Vietnam, research on carbapenem resistant genotype of K.pneumoniae has not been conducted much. In this study, whole genome sequencing of XDR K.pneumoniae strains, analysis and con rmation of carbapenem resistance genes, also the relationship between carbapenem resistance and multi-drug resistance of the strains of K.pneumoniae were performed. This will provide a clearer and more indepth look at about the XDR capacity of these strains. Curently, millions of MDR genes have been sequenced that includes mostly bla genes (TEM, SHV, KPC, CTX-M, OXA, VIM, CMY, GIM, and SPM etc), drug modifying genes like catB, aadA, aacA, aph, as well as sul, arr3, dhfr, mcr-1 and vanA genes [47][48][49]. Antimicrobial resistance genes can be transferred to other bacterial species and thereafter can lead to the production of various enzymes in order to inactivate antimicrobial activities. The several members of Enterobacteriaceae group also can against carbapenems when combined mutations in chromosomal porin to prevent accumulation of β-lactam agents in the bacteria [50]. This study was also determined the antimicrobial susceptibility gene lists (Supplementary Table 1), the presence and mutation of ve relevant carbapenem genes such as KPC-2, NDM-1 and 4, and OXA-48 and 181. The K.pneumoniae isolates had 93 difference resistance genenotyles separating in 10 antibiotic groups (e.g., aminocyclitol, aminoglycoside, beta-lactam, uoroquinolone, folate pathway antagonist, fosfomycin, macrolide, rifamycin, tetracycline, and under development groups). Among the resistance against carbapenems K.pneumoniae isolates, the results have shown a high percentage of harboring the KPC-2 gene (41.2%), NDM-4 gene (29.4%), NDM-1 gene (23.5%), and a low percentage of both OXA-48 and OXA-181 genes (only 1/17 strain, accounting 5.9%). These ndings were in accordance with the other reports that showed a low percentage of the OXA-48 gene in K.pneumoniae isolates [51,52], whereas a previous study in Romania was revealed that OXA-48 was the most predominant genotype, followed by NDM-1 and KPC-2 in the carbapenem resistance K.pneumoniae clinical isolates [53]. There are a global epidemiology of the spread of carbapenem resistant K.pneumoniae strains with mainly carbapenemase genes as KPC-2 in Bulgaria [54] In particularlly, the mutations in KPC-2 gene, a main gene is causing the carbapenem resistance of K.pneumoniae, has been identi ed. This mutation is located in position acid amin 25 Glycine changed to a Serine. Another mutation on NDM-4 gene also indicated as amino acid 187 Proline changed to an Alanine. These mutations have not been mentioned in previous studies in carbapenem resistant K.pneumoniae strains. Mutations in DNA can have a signi cant impact on the resistance phenotype and hence the amino acid substitutions detected in two K.pneumoniae isolates may result in high levels of carbarpenem resistance (Table 3). Although, the antibiotic resistance mechanism of these related genes has not been studied in this experiment, this helps researchers to localize and be more interested in these gene groups in other related studies. An interesting note in our study, of the two isolates containing mutations (MH16-335M and MH17-011M) were belongs to the multilocus sequence type ST15. The most frequent STs of carbapenem resistant K.pneumoniae are characterized by different strains that distributed geographically, namely, ST258 being predominant in Europe and USA, ST11 being most common in East Asia, whereas ST15 is a less frequently occurring strain but has been indicated in clinical cases within outbreaks worldwide [45]. In the speci c outbreaks, ST15 isolates can be highly homogenic, a high variability in antibiotic susceptibility and potential horizontal gene acquisition (Björn Berglund, 2019). K.pneumoniae ST15 strains had few reports from Asia or Vietnam previously, however, near time, Jiansheng Huang et al. was the rst description of nosocomial outbreaks caused by K.pneumoniae ST15 strains in China [61]. This may indicate the spread of ST15 strains between geographically close regions. These ndings may contribute to understanding the existence and spread of resistant bacteria such as strain ST15. Although ST15 strains are low in carbapenem resistance and virulence, it is important to note the importance of careful assessment of phenotypic and genetic characteristics for increasing anti infection treatment.

Conclusion
In conclusion, our systematic study encountered predicting a resistance gene pro le from whole genome data, the mutation in relating the cabarpenemase genes and other alterations in clinical antibiotic resistance K.pneumoniae isolates. The identi cation of speci c genomic variations of the drug resistant bacteria is important for the development of more effective pathogen control and treatment strategies. In the future, it is importance in continuing the description of the genetic mechanisms that associated resistant phenotype bacteria and their antibiotic resistant genotypes, namely, the discovery of novel genes, the altered role of intrinsic genes in multi drug resistance mechanism of bacteria.  KPC-2 protein sequence alignment for K.pneumoniae isolates with some known sequences. NDM protein sequence alignment for K.pneumoniae isolates with some known sequences. The amino acid 154M>L: pMet154Leu (Methionine instead of Leucine) was identi ed as the protein coding for the NDM-4 gene.

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download.