Occurrence of Diverse Antimicrobial Resistance Determinants in Genetically Unrelated Biocide Tolerant Klebsiella pneumoniae

Nosocomial infections due to Klebsiella pneumoniae is a significant problem in health care settings worldwide. In this study, we examined the antimicrobial susceptibility, genetic profiles and mechanisms of antibiotic resistance in K. pneumoniae isolates of Indian origin. To our knowledge this is the first report demonstrating the high prevalence of β-lactamases, aminoglycoside modifying enzymes, quinolone resistance genes besides demonstrating the involvement of active efflux in K. pneumoniae Indian isolates. This study has enabled us to correlate the phenotypic and genotypic characteristics in K. pneumoniae, providing an important base for continued monitoring and epidemiological studies of this emerging nosocomial pathogen in Indian hospitals.


Introduction
Klebsiella pneumoniae is a rapidly emerging nosocomial pathogen that causes wide variety of infections from mild to severe including bacteraemia, pneumonia, meningitis, urinary tract infections (UTIs), wound infections and neonatal septicemia [1]. Infectious diseases society of America has reported this organism to be amongst top five causing major diseases and mortality in the United States [2]. According to the documented reports, Klebsiella spp. is responsible for 8% of total hospital acquired infections (HAIs) in the USA and Europe and is at 8 th position in a list of most important nosocomial pathogen in hospitals [3].A hallmark of K. pneumoniae infection is the difficulty to treat illness because of its high-level multiple intrinsic antibiotic resistance trait, which is responsible for decreased effectiveness of antibacterial therapy and thus remains as a great threat for clinicians [4].
Resistant strains of K. pneumoniae are reported in several endemic and epidemic hospital infection outbreaks [5].A study by the National Healthcare Safety Network in the USA for over a 12 month period from 2006-2007 revealed that K. pneumoniae was the causative agent for 5.8% of HAIs, and 7.7% of catheter-associated UTIs [6]. The same study showed that 21.2% of those K. pneumoniae isolates were resistant to cephalosporins and 10.1% exhibited

Materials and Methods
Bacteriology and clonality K. pneumoniae strains were collected from Chennai and Chandigarh during the period 2012-2013 and were streaked on Klebsiella agar selective media from Hi Media. Based on the growth of culture on selective agar, the isolates were confirmed to be K. pneumoniae. To determine the lineage to which these strains belong, we standardized BOX-PCR and ERIC-PCR in this study to determine the clonality among the isolates [27].

Determination of Minimum Inhibitory Concentration (MIC)
1. Concentrations of antibiotics tested. MIC was determined for the clinical isolates by the agar diffusion method. Luria-Bertani (LB) plates containing various concentrations of the drug were prepared and the overnight cultures were freshly inoculated to achieve the OD 600nm of 0.1 and culture was then spotted on each LB plate having different concentration of the antimicrobials. The plates were then incubated at 37˚C for 16 h and then growth was monitored [29,30]. In this study we used 18 antibiotics from different classes viz. β-lactams (ampicillin, carbenicillin, ceftazidime, cefotaxime), aminoglycosides (amikacin, gentamicin, kanamycin, neomycin, streptomycin), quinolones (ciprofloxacin, nalidixic acid, norfloxacin), and others (erythromycin, chloramphenicol, polymyxin B, rifampicin, tetracycline and trimethoprim) at different concentrations (0.5 μg/ml, 4 μg/ml, 16 μg/ml, 64 μg/ ml, 256 μg/ml and 1024 μg/ml). We calculated the MIC from inhibition of growth in plates with antibiotics with respect to control plate having no added supplement in it.

Bacterial transformation and plating on selective media
Plasmid DNA was extracted from the isolates by a method described previously [31], except that for plasmid isolation, 0.75ml of culture and twice the suggested volume of all three solutions were used. E. coli JM109 electro competent cells, prepared according to the manufacturer's recommendations (Bio-Rad Laboratories, Richmond, Calif.), were transformed with 30 ng of plasmid preparations as described before [32].
Genomic DNA isolation, PCR amplification, sequence analysis Presence of resistance determinant genes in the chromosome of the clinical isolates were detected by PCR based approach. The genomic DNA from all the isolates was extracted using the CTAB method and was used as a template all the PCRs performed. Clonal relatedness was determined by ERIC-PCR using primers ERIC-1R: ATGTAAGCTCCTGGGGATTCAC and ERIC-2: AAGTAAGTGACTGGGGTGAGCG and products were analyzed on gel. Resistance genes were determined by using specific sets of gene specific primers [33][34][35][36][37][38][39]. We checked for presence of ESBL genes by performing PCR for bla SHV , bla VEB and bla GES; metallo β lactamases genes like bla SIM , bla VIM , bla GIM , bla IMP , and bla SPM; genes investigated also included carbapenemase genes bla OXA-23, bla OXA-24 , bla OXA-58 and bla OXA-69 . Presence of genes for conferring resistance to aminoglycosides like aadA1, aphA3, aphA6, rmtB, rmtC, rmtD, strA, strB were also determined. We also checked for the presence of quinolone resistance genes qnrA qnrB, qnrS and some other genes like catIII, catB2, catB3, cmlB and CusABC efflux pump (epF: 5'-GATATGGATCTGGTGCCGAAATA-3' and epR: 5' ACTGCGAGGCGTCTTTAATC-3'). Elution of the amplicons was done as described previously. Sequencing was carried out by following the manufacturer's instructions (Big Dye Terminator kit; Applied Biosystems) with an ABI Prism 310 instrument. The identities of the sequences were established through a database search by using the BLAST program [40,41].
Growth inhibition assay to elucidate the active efflux activity The assays were performed as described previously [29]. Cultures in the mid-exponential phase were inoculated into LB broth containing antimicrobials either alone or with the efflux pump inhibitors and the growth profile of the cultures at 37˚C and was analyzed by measuring the absorbance at OD600 nm periodically at a regular interval thereafter using Synergy H1 Hybrid microplate reader (BioTek Instruments Inc., Winooski VT). In this study, we used efflux based substrates antibiotics {kanamycin 16 μg/ml}, {acriflavine 8µg/ml}, and disinfectants such as {benzalkonium chloride 3.2µg/ml} as substrates. The 2, 4-dinitrophenol (DNP), carbonyl cyanide 3-chloro phenyl hydrazone (CCCP), verampamil (VER) and reserpine (RES) were used as efflux pump inhibitors {2.5 μg/ml} in this study. We analysed efflux activity as the difference in growth profile of the strains with the presence and absence of drug along with presence and absence of different inhibitors in independent experiments.

Analysis of outer membrane protein profiles
Outer membrane proteins were purified by the method as described previously [42], with some slight modification. Cells were harvested by centrifugation (6,000 × g for 15 min) and were suspended in 50 mM potassium phosphate buffer (pH 7.2) and sonicated for 5 mins. The cells were crushed by sonication and the unbroken cells were removed by centrifugation (10,000 × g for 10 min). The crude envelope fraction was collected from the supernatant by centrifugation at 105,000 x g for 1 h at 4˚C. The pellet containing the crude envelope fraction was treated with 0.5% (wt/vol) sarkosyl (Sigma) solution for 30 mins to selectively solubilize the inner membrane part. The insoluble outer membrane fraction (OMP) was recovered as pellet by centrifugation at 105,000 x g for 1 h at 4˚C. The pellet was resuspended and stored at -20˚C until used. Protein contents of membrane preparations were determined by the method of bicinchoninic acid (BCA) method (Pierce BCA protein assay kit) with bovine serum albumin (BSA) (Sigma) as standard.

Estimation of biofilm formation
In this study, the overnight cultures were freshly inoculated to achieve an initial OD 600nm of 0.1, and then incubated at 37˚C for 16 h and were observed for the growth and measured as OD 600nm then the culture were discarded and washed carefully without disturbing possible biofilm formation along the wall. Tubes were then filled with 0.1% crystal violet and incubated for 30 min. Later stain was discarded, washed and dried. Now biofilm was dissolved using 1 ml 33% acetic acid and OD was measured at 570nm [29]. Biofilm formation was calculated as the ratio of OD 570nm /OD 600nm.

Oxidative stress tolerance assay
We performed oxidative disc diffusion assay to know the oxidative stress tolerance of the clinical isolates. Hydrogen peroxide (H 2 O 2 ) was used to generate oxidative stress in our current study. The test cultures were grown overnight and then diluted to make OD 600nm equivalent to 0.1. The cultures were spread on the LB plates and Whatman filter paper discs soaked in different concentrations of H 2 O 2 (0%, 0.1%, 1%, 3%, or 10%) are placed at appropriate distances from one another and incubated at 37˚C for 16 h. After incubation, the zone of inhibition around every H 2 O 2 disc was measured and compared [30].

Precipitation test and Motility Assay
For the precipitation test, 1 ml cultures were centrifuged at 4000g for 2 mins and left in standing condition for 15 mins, later visually tested for mucoidy or no -mucoidy supernatant and dense pellet. The motility assays were done with representative strains where LB grown K. pneumoniae cultures (OD 600nm = 1.0) were inoculated with a toothpick on LB agar plates with 0.25%, 0.45% and 0.7% agar and incubated for 12 hrs at 37˚C. In this growth medium, bacteria can swim through the soft agar and produce a halo. The diameter of the halo is a measure of the ability to swarming motility. And some bacteria use to have twitching motility which is also detected using staining of bacterial cells in lower bottom of media as zone of twitching motility [30].
isolates used in this study were diverse, clonally distinct and could broadly classified into 12 different clusters (Fig 1B).

Multidrug resistance profiles of Indian K. pneumoniae isolates
We analyzed the MDR pattern of isolates and found that many were multidrug resistant. We classified the isolates into groups based on their antibiotic resistance profile ( Fig 2B, Table 1). Interestingly 4.4% of isolates showed resistance to all the 25 antibiotics used and alarmingly high percentage (31%) of isolates were resistant to 20 or more antibiotics.  (Fig 3A).

Role of plasmids in Indian K. pneumoniae isolates
Plasmids are extra chromosomal genetic elements capable of self replication and are easily transmissible. Plasmids are known to carry genes required for sustaining in adverse conditions and also genes responsible for various drug resistance. Clinical isolates were studied for the presence of plasmid. The 63% of strains were found to have multiple plasmids of various sizes ( Fig 3B).Upon transforming the plasmids in E. coli JM109 electro competent cells, we found transformants on ampicillin, streptomycin, tetracycline containing antibiotic plates; work is in progress to know their precise contribution in conferring antimicrobial resistance in K. pneumoniae from Indian scenario.

Diverse antibiotic resistance determinants in K. pneumoniae
With the precise information about the antibiotic resistance profile of these strains it was necessary to know the resistance genes present in the genome of the pathogen. Therefore the genomic DNA was isolated and gene specific primers were used to amplify the respective genes.
• The β-lactamases are known as a group of enzymes having the ability to hydrolyze β-lactam antibiotics thereby conferring resistance to the micro organisms. In our study, bla SHV gene was present in 26%, bla VEB was present in 8.8%, bla GIM was present in 15.5%, and bla SIM was detected in 6% of the collected strains. Interestingly, while bla OXA-24 and bla OXA-58 was not found in this collection, the presence of bla OXA-23 was detected in 30% of strains and bla OXA-69 was detected in 15% of the strains ( Fig 3C, Table 2).
• The aminoglycoside adenyl transferase aadA1 was detected in 30% of the isolates, aadB was detected in 3% of these strains, aphA3 was detected in 17% of the isolates, aphA6 was detected in 6% of the isolates. The rRNA methyl transferase alleles for conferring resistance to aminoglycosides e.g.rmtB was detected in 11% of the isolates, rmtC was detected in 13% of the isolates, and surprisingly rmtD was not found in the isolates. Linked genes strA-strB present on mobile elements of pathogenic organisms is main contributor for resistance to streptomycin. The presence of strA was detected in 20% of these isolates and the presence of strB was detected in 13% of these isolates ( Fig 3C, Table 3).
• The presence of qnr alleles contributes to the plasmid mediated resistance against quinolone group of antibiotics. The presence of qnrA was detected in 10% of the isolates. The presence of qnrB was detected in 20% and presence of qnrS was detected in 8.8% of the isolates ( Fig  3C, Table 3).   (Fig 4A, ii).

Susceptibility towards hospital based biocides
Apart from antibiotics used for treatment, bacteria were found resistant to biocides which are regularly used in hospital as cleansing materials which led them stay on clinical set up for long and thus making patients vulnerable to nosocomial infections. For benzalkonium chloride, 60% of the isolates showed high MIC values ranging from 51.2 μg/ml to >102.4 μg/ml and another 36% of isolates were found intermediately resistant having MIC values in the range of In this study we grouped the isolates in three different resistance level for each antibiotic viz. strains with high MIC values (256 μg/ml to >1024 μg/ml) are grouped as highly resistant, intermediate MIC value (16 μg/ml to 64 μg/ml) as intermediately resistant and low MIC value (4 μg/ml or less) as sensitive. For this study, we used the following antibiotics-AMP: ampicillin, CAR: carbenicillin, CTR: ceftriaxone, CAZ: ceftazidime, AMK: amikacin, GEN: gentamicin, KAN: kanamycin NEO: neomycin, STR: streptomycin CIP: ciprofloxacin, NAL: nalidixic acid, NOR: norfloxacin, CHL: chloramphenicol, ERY: erythromycin, PMB: polymyxin B, RIF: rifampicin, TET: tetracycline, TMP: trimethoprim. The bar graph represents the mean of three independent experiments. (B) DNA gel picture of K. pneumoniae plasmids isolated by alkaline lysis method from respective strains showed multiple plasmids of different size. The MDR strains carried multiple plasmids which indicated plasmids might play role in drug resistance. (C) The prevalence to different resistance determinants in the K. pneumoniae isolates are shown here. In this study, we found high prevalence of β-lactamases (bla GES, bla VEB, bla SHV, bla VIM, bla GIM, bla IMP, bla SPM, bla SIM, bla oxa 23, bla oxa 24, bla oxa 24, bla oxa 58, bla OXA69 ) ; aminoglycosidases (aacA4, aadA1, aadB, aphA3, aphA6, rmtB, rmtC, rmtD, indicates presence of the respective gene in the particular strain; '-'indicates respective gene was not found in the particular strain Table 3. PCR based detection of other resistance determinants.

Strains Insertion elements
Aminoglycosidase Quinolones others

5'cs Class I int aadA1 aadB aphA3 aphA6 rmtB rmtC rmtD strA strB qnrA qnrB qnrS cat III catB2 catB3
indicates presence of the respective gene in the particular strain; '-'indicates respective gene was not found in the particular strain. For Domex a commercially available disinfectant 83% of isolates shown MIC of 0.4 μg/ml to 0.8 μg/ml i.e. and also 10% of them have MIC of 1.6 μg/ml or more and 6% of the isolates were susceptible showing MIC values of <0.1 μg/ml. For lizol another commercially available disinfectant 36% of isolates shown MIC of 0.4% to 0.8% of the solution i.e. and also 6% of them have MIC of 1.2% to 1.6% solution or more and 56% of the isolates were susceptible showing MIC values of 0.2% to <0.1% solution (Fig 4B).
Overall we found the clinical isolates in our collection were not only multidrug resistant but were also broadly biocide tolerant.

Role of Efflux activity in Antimicrobial resistance
In our study, to investigate the role of efflux system in drug resistance we performed growth inactivation assays with different antimicrobials and inhibitors which act by hindering active efflux activity. It was interesting to see that the MDR strain when grown in kanamycin as a substrate without the inhibitor, after >12 hrs of growth, showed 1.19-fold, 1.01-fold, 0.98fold and 11.83-fold increased growth when compared to its growth in presence of inhibitors CCCP, DNP, verampamil and reserpine respectively in independent experiments (Fig 5A, i).
When MDR strain was grown in acriflavine as a substrate without the inhibitor, after >12 hrs of growth, showed 1.13-fold, 0.96-fold, 0.97-fold and 6.4-fold increased growth when compared to its growth in presence of inhibitors CCCP, DNP, verampamil and reserpine respectively in independent experiments (Fig 5A, ii). When MDR strain was grown in benzalkonium chloride as a substrate without the inhibitor, after >12 hrs of growth, showed 1.16fold, 1.01-fold, 0.99-fold and 10.44-fold increased growth when compared to its growth in presence of inhibitors CCCP, DNP, verampamil and reserpine respectively in independent experiments (Fig 5A, iii). Overall, we found the clinical isolates in our collection use active efflux as the underlying molecular mechanism to mediate not only multidrug resistance but also biocide tolerance.

Outer membrane protein profiles of Indian K. pneumoniae isolates
The cell envelope is the prime line for most outside stress conditions that may modify envelope components and thus bring an extra cytoplasmic stress response. A reduction in the permeation of antibiotics is generally related to a decrease in porin expression or an alteration in the porin structure. The role of outer membrane proteins were also analysed among the clinical isolates. The outer membrane proteins were isolated and checked on SDS-PAGE. Differential patterns of outer membrane proteins amongst the clinical strains could be observed and the proteins over expressed bands from multidrug resistant strains were identified. The over expressed bands may be part of efflux pumps, so we are now working towards the identification of those interestingly distinguished protein (Fig 5B).

Determination of presence of CusABC efflux pumps
With the proof of efflux activity through efflux pumps in the representative strains it is necessary to know the presence of efflux pumps. Therefore the genomic DNA of the isolates and gene specific primers were used to amplify the AcrAB-like CusABC efflux genes. Among the pool of strains used in this study 51% of strains were having RND efflux genes, while 33% percent of the strains also showed the presence of unique regulator cueR gene (data not shown), work are in progress to elucidate and delineate their impact in drug resistance in K. pneumoniae.

Oxidative stress behavior of Indian K. pneumoniae isolates
We performed oxidative disc diffusion assay to know the oxidative stress tolerance of the clinical isolates. Hydrogen peroxide (H 2 O 2 ) was used to generate oxidative stress in our current study. It was found that 3.3% of strains were resistant (diameter of zone of inhibition with 1% H 2 O 2 is 8 mm or less) toward oxidative stress and another 53% were intermediately resistant (diameter of zone of inhibition with 1% H 2 O 2 is 8 mm-13 mm) and the other 43.3% were less susceptible (diameter of zone of inhibition with 1% H 2 O 2 is 14 mm or more) (Fig 6A).

Motility and Biofilm formation by Indian K. pneumoniae isolates
We performed the biofilm assay using LB broth medium in glass tubes. From this study it is found that most of the strains formed biofilm (Fig 6B, i) and upon comparing with their antibiogram it was found that > 80% of the strains that had an ability to form strong biofilms were resistant (Fig 6B, ii). Strains were tested for their ability to form dense pellet by centrifugation, and none were found to form loose pellet, all strains formed dense pellet (Fig 7A). We performed motility assay to check for swarming motility (Fig 7B, i) and twitching motility ( Fig  7B, ii) among 15 representative isolates using soft agar plates of 0.25% agar, 0.45% agar and 0.7% agar. In this study KPN23, KPN27, KPN8, KPN12, KPN28, KPN3 isolates showed maximum swarming motility with >60 mm of zone of motility at 0.25% agar concentration.

Discussion
Over the millennia, bacteria have evolved evasion strategies to overcome arrays of chemical and environmental assails including antimicrobial drugs [1]. The occurrence of nosocomial infections in hospital intensive care units due to K. pneumoniae currently ranges from 2% to 10% of all Gram-negative bacterial infections in Europe and account for about 2.5% of them in the United States [2]. K. pneumoniae is a rapidly emerging nosocomial pathogen that causes wide variety of infections from mild to severe including bacteremia, pneumonia, meningitis, urinary tract infections, wound infections and neonatal septicemia [1]. Infectious disease society of America has reported this organism to be amongst top five causing major diseases and mortality in the United States [5]. It has now become a major cause of hospital-acquired  infections worldwide due to its remarkable propensity to rapidly acquire resistance determinants to a wide range of antibacterial agents [6]. It has been included in the acronym ESKAPE which lists top six most notorious Gram negative bacteria [7]. The capacity of this pathogen to cause major outbreaks stems from two major characteristics; first one is antimicrobial pneumoniae strain with different antimicrobial compounds and efflux pump inhibitors is shown by OD 600nm at every 15 minutes time interval. Growth inactivation assays (GIAs) were done with a representative strain to know the impact on drug efflux capability using substrates KAN: kanamycin, ACRI: acriflavine and BENZ: benzalkonium chloride, in presence and absence of efflux pump inhibitors (CCCP: carbonyl cyanide 3-chlorophenylhydrazone, DNP: 2, 4 -dinitrophenol, VER: verapamil, and RES: reserpine). The graph represents the mean of three independent experiments. (B) Protein gel picture of OMP profile of representative K. pneumoniae strains. OMP was isolated from the strains using sarkosyl method and was checked on SDS-PAGE. K. pneumoniae isolates showed differential expression in our collection. Trends of Resistance Mechanisms in Klebsiella pneumoniae resistance and second one being the ability to persist in the environment for extended periods of time [22]. There has been no significant study till date that describes the variation in the clinical strains that are isolated and mechanisms of drug resistance displayed by them. Therefore, we performed this longitudinal study, in the present study we collected multiple clinical strains of K. pneumoniae and started from their clonal studies and after knowing that they belonged to distinct clonal classes we worked further.
We found the strains to be multidrug resistant as determined by Kirby Bauer disc diffusion method and their MIC values. We screened for the presence of various resistance determinants by PCR method. Like any Gram-negative bacteria, our isolates also harbored multiple β-lactamases which explains their resistant phenotype. It is generally considered that clinical strains In this study 0%, 0.1%, 1%, 3%, and 10% of H 2 O 2 were used as oxidative stress inducing agent. The diagram represents the mean of three independent experiments. (B) Biofilm formation by the K. pneumoniae isolates are shown here: i) tubes showing stained biofilm rings in few representative isolates ii) graphical presentation of biofilm formation among the isolates is shown as the ratio of OD 570nm and OD 600nm . The bar graph represents the mean of three independent experiments. doi:10.1371/journal.pone.0166730.g006 contain integrons because of the multi-drug resistance advantages that these mobile DNA elements facilitate [32]. Though class 1 integron was not found in our study, multiple genes (the AME's and aphA6) encoding resistance to clinically relevant antimicrobials [32], particularly aminoglycosides were found which indicate that the strains could have evolved with them to circumvent the regime of drug assault such as amikacin, gentamicin and streptomycin.
Studies on the mechanisms of quinolone resistance in Gram-negative bacteria such as E. coli, N.gonorrhoeae, V. cholerae and Haemophilus influenzae have elucidated that the alterations in the GyrA subunit of DNA gyrase are primarily responsible for the high level resistance to nalidixic acid [32]. In E. coli, mutations are clustered in the quinolone resistancedetermining region (QRDR) located between amino acids Ala-67 andGln-106, the most frequent being at codon 83 or at equivalent positions in other microorganisms [43]. Several reports have discussed the significance of alteration in ParC subunit of topoisomerase IV in enhancing resistance to fluoroquinolones. However we were unable to identify any mutations in the previously defined QRDR in our collection of nalidixic acid resistant isolates respectively. As efflux pumps are one of the main contributors to drug resistance, we screened for the presence of some efflux pumps by PCR and also established the activity of efflux pumps by performing the growth inactivation assay.
We also determined the resistance to biocides as they are the major surface elements to which the clinical strains are exposed to. So this study again proves that efflux pumps are not only conferring drug resistance but also mediates resistance to a variety of structurally unrelated compounds which led to therapy failure by increasing the MIC of antimicrobials. This active drug transport is involved in low intrinsic susceptibility, cross-resistance to chemically unrelated classes of molecules, and selection/acquisition of additional mechanisms of resistance. We were also able to detect the presence of protein kinases in these strains, and currently our research focuses on their biological relevance.
The goal of this study was to link the phenotypes and the resistance determinants in K. pneumoniae, giving a picture of the complex endemic nature in this geographic location. From these data, it is clear that the potential impact of K. pneumoniae infections on humans in Indian hospitals is significant. Taken together, prevalence of resistance determinants for the front line choice of drugs threatens the future continued clinical use of antibiotics. Therefore, the breadth of resistome, monotony of resistance mechanisms, evolution and emergence of antibiotic resistant populations needs to be clearly understood to allow the discovery of new therapeutic drugs.