Transcriptional analysis of blaNDM-1 and copy number alteration under carbapenem stress

New Delhi metallo beta-lactamase is known to compromise carbapenem therapy and leading to treatment failure. However, their response to carbapenem stress is not clearly known. Here, we have investigated the transcriptional response of blaNDM-1 and plasmid copy number alteration under carbapenem exposure. Three blaNDM-1 harboring plasmids representing three incompatibility types (IncFIC, IncA/C and IncK) were inoculated in LB broth with and without imipenem, meropenem and ertapenem. After each 1 h total RNA was isolated, immediately reverse transcribed into cDNA and quantitative real time PCR was used for transcriptional expression of blaNDM-1. Horizontal transferability and stability of the plasmids encoding blaNDM-1 were also determined. Changes in copy number of blaNDM-1 harboring plasmids under the exposure of different carbapenems were determined by real time PCR. Clonal relatedness among the isolates was determined by pulsed field gel electrophoresis. Under carbapenem stress over an interval of time there was a sharp variation in the transcriptional expression of blaNDM-1 although it did not follow a specific pattern. All blaNDM-1 carrying plasmids were transferable by conjugation. These plasmids were highly stable and complete loss was observed between 92nd to 96th serial passages when antibiotic pressure was withdrawn. High copy number of blaNDM-1 was found for IncF type plasmids compared to the other replicon types. This study suggests that the single dose of carbapenem pressure does not significantly influence the expression of blaNDM-1 and also focus on the stability of this gene as well as the change in copy number with respect to the incompatible type of plasmid harboring resistance determinant.


Background
Since the discovery of New Delhi metallo-β-lactamase (bla NDM ) in 2008 from a Swedish patient of Indian origin in New-Delhi, India [1], this enzyme is known for several reasons including treatment failure, emergence of new variants and lateral transfer of the gene coding this enzyme within diverse host range of Gram negative bacilli [2,3]. The bla NDM is known for its ignominious nature being linked with other resistance determinants along with various mobile elements like plasmid, insertion sequences & transposons which facilitates its horizontal dissemination [2,4]. In many studies bla NDM-1 was found to be associated with ISAba125 [2,5]. However, there were also reports of other insertion elements like ISCR1, ISCR16, IS26, IS1, ISEc33 and IS903 associated with this gene [5]. Additionally, the transposons Tn3 and Tn125 were reported to be linked with this resistance determinant and horizontal transfer of bla NDM-1 is often facilitated by plasmids of IncF, IncA/C, IncL/M, IncH, IncN and more recently by IncX type [6]. Among Enterobacteriaceae, bla NDM was detected in Escherichia coli in many countries worldwide (Australia, France, Germany, Japan, UK and the USA) [7]. E. coli is the most common pathogen associated with nosocomial and community acquired infections and also been considered as a potent host for this resistance determinant [7]. Dissemination of bla NDM-1 through E. coli has become a global concern [8] and also in India there were several reports of NDM-producing E. coli in all parts of the country [8][9][10][11][12][13][14]. Treatment of infections with NDM-producers is restricted due to their multidrug resistance phenotype [15]. Several studies have highlighted the hydrolytic activity of NDM-1 to carbapenems [2,16]. However, it is not known how bacteria harboring this resistance gene will respond when carbapenem therapy is initiated to a patient. The present study was designed to investigate the transcriptional response of bla NDM-1 in vitro under single dose carbapenem exposure, and also to investigate the transmission dynamics within clinical isolates of Escherichia coli in a single center study from India.

Transcriptional expression analysis of bla NDM-1
Transcriptional expression of bla NDM-1 in response to imipenem, meropenem and ertapenem stress was determined by inoculating the organisms harboring bla NDM-1 in Luria Bertani broth (Hi-media, Mumbai, India) with and without antibiotics. Antibiotic concentration used was 1 μg/ml. For a period of 16 h, total RNA was isolated at the interval of 1 h using Qiagen RNease Mini Kit (Qiagen, Germany), immediately reverse transcribed into cDNA by using QuantiTect ® reverse transcription kit (Qiagen, Germany). The cDNA was quantified by Picodrop (Pico 200, Cambridge, UK) and real time PCR was performed using Power Sybr Green Master Mix (Applied Biosystem, Warrington, UK) in Step One Plus real time detection system (Applied Biosystem, USA) using a set of primer (NDM-F 5 / -GGGCAGTCGCTTC CAACGGT-3 / and NDM-RT-R 5 / -CGACCGGCAG GTTGATCTCC-3 / ). The relative expression of bla NDM-1 in each interval with and without carbapenem pressure was determined by ΔΔC t method [19]. Relative quantification was done using a transformant (E. coli DH 5 α harboring bla NDM ; P EC-611 ) grown for 4 h without any antibiotic pressure.

Transformation and Conjugation assay
Transformation was performed by heat shock method [15] using E. coli DH 5 α as a recipient and the transformants were selected on Luria Bertani agar (Hi-Media, Mumbai, India) containing 0.25 μg/ml of imipenem. Conjugation experiment was carried out using bla NDM-1 harboring clinical strains as donors and a streptomycin resistant E. coli recipient strain B (Genei, Bangalore, India). The MIC of clinical isolates against streptomycin was pre-determined to optimize the agar for selection of transconjugants. Both the donor and recipient cells were cultured in Luria Bertani Broth (Hi-Media, Mumbai, India) till it reach an O.D. of 0.8-0.9 at A 600 . Cells were mixed at 1:5 donor-to-recipient ratios and transconjugants were selected on agar plates containing imipenem (0.25 μg/ml) and streptomycin (1000 μg/ml). The E. coli strain B is chromosomally resistant to streptomycin which can grow on media containing streptomycin at a concentration of 1000 μg/ml. However, the donors although resistant to aminoglycoside had the minimum inhibitory concentration ranging from 100-200 μg/ml. Therefore, selection of transformants in 1000 μg/ml rules out false selection of donor strains. The accuracy of conjugation was further cross checked by typing all the transconjugants by enterobacterial repetitive intergenic consensus PCR [20] and pulsed field gel electrophoresis using Xba1 restriction enzyme.
Plasmid stability analysis of parent strains and transformants was done by serial passage method for consecutive 100 days at 1:1000 dilutions without any antibiotic pressure [23]. After each passage, 1 ml of the culture was diluted in normal saline (1:1000) and 40 μl of the diluted sample was spread on to the LB agar plate. After overnight incubation, 50 colonies from the agar plates were randomly picked and subjected to phenotypic detection of MBL and further confirmed by PCR assay for the presence of bla NDM-1 using primers (NDM-F 5 / -GGGCAGTCGCTTCCAACGGT-3 / and NDM-R 5 / -GTAGTGCTCAGTGTCGGCAT-3 / .

Copy number determination of plasmid encoding bla NDM-1
Clinical isolates of Escherichia coli harboring bla NDM-1 carried by plasmids of incompatibility groups IncFIC, IncA/C or IncK were selected for determining the copy number under exposure of different concentrations of carbapenem antibiotics. Single colony of each incompatibility type was inoculated into LB broth containing 0.5 μg/ml, 1 μg/ml, 2 μg/ml and 4 μg/ml of each imipenem, meropenem and ertapenem and also without any antibiotic (considered as a reference), was incubated at 37°C for 5-6 h until the OD reached 0.9 at A 600 . Transformants with different bla NDM-1 carrying plasmid types (IncFIC, A/C & K) were used as control (without any antibiotic pressure). Plasmid DNA was extracted using QIAprep Spin Miniprep Kit (Qiagen, Germany). Quantitative Real Time PCR was performed using Step One Plus real time detection system (Applied Biosystem, USA) to estimate the relative copy number of bla NDM-1 for different concentrations of each antibiotic for three different incompatibility types. The copy number of bla NDM-1 within the wild type plasmid of different incompatibility types were also determined to know the type of Inc group where copy number of bla NDM-1 gene was maintained in high number. Quantitative real time PCR reaction was carried out using 10 μl of SYBR® Green PCR Master Mix (Applied Biosystem, Warrington, UK), 4 ng plasmid DNA as template and 3 μl of each primer (10 Picomol) in a 20 μl reaction under a reaction condition of initial denaturation at 94°C for 5 min, 40 cycles of denaturation 94°C for 20 s, annealing 52°C for 40 s and extension at 72°C for 30 s. The relative fold change was measured by ΔΔCT method and Ct value of each sample was normalized against a housekeeping gene rpsel of E. coli [19].

Antimicrobial susceptibility testing and MIC determination
Antibiotic susceptibility of bla NDM-1 harboring parent strains, transformants and transconjugants were determined by Kirby Bauer disc-diffusion method including piperacillin-tazobactam (100/10 μg), co-trimoxazole (25 μg), amikacin (30 μg), gentamicin (10 μg), ciprofloxacin (5 μg), polymyxin B (300units), netilmicin (30 μg), carbenicillin (100 μg), tigecycline(30 μg) and faropenem (5 μg) (Hi-Media, Mumbai, India). MICs of imipenem, meropenem, ertapenem, cefepime, aztreonam, gentamicin, amikacin, ciprofloxacin, piperacillintazobactam & polymixin-B were determined for parent strains harboring bla NDM-1, as well as transformants and transconjugants by agar dilution method. Each stock solution for the corresponding antibiotic was made at 1 mg/ml concentration in nuclease free water and was stored at −80°C. The quality control for stock solution was checked each time against E. coli ATCC 25922. The result of the susceptibility testing was interpreted as per CLSI guidelines [24]. However, for polymyxin B, faropenem and carbenicillin, the organisms were considered as non susceptible if the MIC value was higher and diameter of the zone of inhibition was lower than the values given in CLSI guidelines for respective antibiotics against E. coli ATCC 25922.
Typing of bla NDM-1 harboring isolates All bla NDM-1 harboring E. coli isolates were typed by pulsed field gel electrophoresis (PFGE), genomic DNA was prepared in agarose blocks and digested with the restriction enzyme Xba1 (Promega, Madison, USA) and the DNA fragments were separated with a CHEF-DR III (Bio-Rad, USA) for 24 h at 6 V/cm with a pulses at 120 0 angle in a 10-40 s pulse time [25].

Results
During the study period (March-September), 17 isolates were obtained carrying bla NDM-1 , collected from different clinical samples mostly associated with surgical wound infection from surgery ward of the hospital (Table 1). Transcriptional expression of bla NDM-1 with or without carbapenem stress is shown in Fig. 1. It was observed that at the initial stage, under meropenem pressure the transcriptional level of bla NDM-1 was low. However, there was a sharp increase from 12 th hour of incubation for meropenem and ertapenem (approximately 2 fold and 4 fold respectively), whereas imipenem did not cause any alteration in transcriptional response. Overall the transcriptional expression did not show any specific pattern of response.
Plasmids carrying bla NDM-1 were selected in the medium containing imipenem and could be conjugatively transferred from all 17 clinical E. coli isolates into recipient E. coli strain B. The transformation experiment revealed that the size of the transferable plasmids was approximately of 50-60 kb. Replicon typing showed that FIC was the predominant replicon type (n = 7) followed by A/C (n = 4) and K (n = 3) whereas 3 isolates were untypeable ( Table 1). The copy number of bla NDM-1 was found to be variable. The copy number of bla NDM-1 gene within IncFIC and IncA/C type of plasmids showed an increasing trend when increasing concentrations of imipenem and meropenem were added whereas for ertapenem, the case was reverse (Figs. 2 & 3). For IncK type plasmids, the copy number of bla NDM-1 consistently raised when meropenem concentration was increased whereas with the increasing concentration of imipenem and ertapenem, the copy number of bla NDM-1 reduced (Fig. 4). The overall copy number of F-Inc type was six fold higher compared to IncA/C and K type (Fig. 5). Complete loss of plasmids for all the isolates containing bla NDM-1 was observed between 92 nd to 96 th serial passages when antibiotic pressure was withdrawn.

Discussion
Resistance to carbapenems due to the production of New Delhi metallo-β-lactamase among enterobacterial isolates has become a very common phenomenon and the expansion of bla NDM-1 among the members of Enterobacteriaceae is increasing and in consequence this resistance determinant has been reported across the globe [26]. Earlier studies demonstrated that the subinhibitory concentrations of antibiotics interfere the expression of the genes, colonization and motility of the cell [27]. Therefore, we have investigated the transcriptional response of NDM-1 against carbapenem antibiotics below the inhibitory concentration level. Under the pressure of imipenem, no significant change was observed in the pattern of transcriptional level for 16 h duration, which is in contrast to the previous report of Liu et al. 2012 [28], as they reported that under the pressure of imipenem bla NDM-1 gene was expressed (0.83 times higher) than that of the control. In this study, a possible down regulated expression of bla NDM-1 took place under the exposure of meropenem, however to support our data no existing literature is available till date. This study has pointed that no specific or defined transcriptional response is initiated for bla NDM-1 when carbapenem stress is created and the overall response is partially chaotic. Thus, there could be other inducing factors which trigger its response in order to synthesis this carbapenemase. The study isolates showed resistance to almost all the antibiotics especially high rate of polymyxin resistance was also observed. The emergence of different E. coli clones with pulsotype 2 as the most common, indicates a possible clonal spread but different replicon types within this clone are uncommon and require further detailed analyses in future studies. Plasmids encoding bla NDM-1 gene were successfully transferred to the recipient E. coli strain B by conjugation indicating potential horizontal transmission through diverse incompatible plasmid types such as IncFIC, IncA/C and IncK in this hospital setting. Association of bla NDM-1 with IncK type of plasmid in the present study is not commonly reported as coexisting data recorded recent spread of bla NDM-1 in India has been associated with IncA/C type, IncF1/FII-type, or unknown types of plasmids [7]. An earlier study [29] suggested that copy number of bla NDM-1 is affected by the concentration of imipenem. In contrast we observed that plasmid copy number is not only related with high concentration of imipenem but also depends on the replicon type of the bla NDM-1 carrying plasmids. This could be supported by the high copy number of bla NDM-1 within IncF type plasmids compared to the other replicon types (e.g. IncA/C or Inc K).

Conclusion
The expression of bla NDM-1 could predict the bacterial response in different time interval when a single carbapenem exposure is applied. Additionally, this study could underscore that irrespective of plasmid types, bla NDM-1 is highly stable within a host of clinical origin. However, it was also evident from this study that different Inc types of plasmids have a specific pattern in copy number alteration under concentration gradient carbapenem stress. Thus, the study came up with epidemiological knowledge of a stable bla NDM-1 mediated carbapenem resistance in E. coli and further investigation is required to evaluate the risk for their dissemination in health care systems in this geographical part of the world.

Additional files
Additional file 1: Table S1. Antimicrobial profile of bla NDM-1 harboring