Journal of Antimicrobial Chemotherapy Advance Access published September 4, 2014
J Antimicrob Chemother doi:10.1093/jac/dku340
Differential contribution of AcrAB and OqxAB efflux pumps to multidrug resistance and virulence in Klebsiella pneumoniae Suzanne Bialek-Davenet1–3, Jean-Philippe Lavigne4,5, Kathleen Guyot1,6, Noe´mie Mayer1, Re´gis Tournebize7,8, Sylvain Brisse2,3, Ve´ronique Leflon-Guibout1 and Marie-He´le`ne Nicolas-Chanoine1,6,9* 1
Service de Microbiologie, Hoˆpital Beaujon, AP-HP, Clichy, France; 2Institut Pasteur, Microbial Evolutionary Genomics, Paris, France; CNRS, UMR3525, Paris, France; 4INSERM U1047, Universite´ Montpellier 1, UFR de Me´decine, Nıˆmes, France; 5Laboratoire de Bacte´riologie, CHU Caremeau, Nıˆmes, France; 6Faculte´ de Me´decine Paris Diderot, Paris, France; 7Institut Pasteur, Pathoge´nie Microbienne Mole´culaire, Paris, France; 8INSERM U786, Paris, France; 9INSERM UMR 1149, Universite´ Paris 7, Paris, France
3
Received 13 March 2014; returned 30 May 2014; revised 15 July 2014; accepted 5 August 2014 Objectives: In Klebsiella pneumoniae, overexpression of the AcrAB efflux pump and the more recently described OqxAB efflux pump has been linked to an antibiotic cross-resistance phenotype, but the mechanisms of regulation are largely unknown. Moreover, while AcrAB has been shown to participate in K. pneumoniae virulence, the contribution of OqxAB has not yet been assessed. Methods: In the present study we investigated a K. pneumoniae clinical isolate (KPBj1 E+), displaying crossresistance to quinolones, chloramphenicol and cefoxitin, and its phenotypic revertant (KPBj1 Rev, susceptible to antibiotics) by using whole-genome sequencing, RT–PCR, complementation and a Caenorhabditis elegans virulence model. Results: We detected a point mutation in the oqxR repressor gene of KPBj1 E+, which overexpressed genes rarA, encoding a transcriptional regulator, and oqxB, but not acrB. Complementation with wild-type oqxR restored antibiotic susceptibility and normalized rarA and oqxB expression levels. Whole-genome sequencing showed that KPBj1 Rev had lost the entire rarA-oqxABR locus, situated close to an integration hot spot of phage P4. This large deletion seemed responsible for the significantly lower virulence potential of strain KPBj1 Rev compared with KPBj1 E+. Moreover, we found that KPBj1 E+ DacrB was significantly less virulent than its parental strain. Conclusions: This work demonstrates the role of the overexpression of efflux pump OqxAB, due to a mutation in gene oqxR, in the antibiotic resistance phenotype of a clinical isolate, and suggests that the presence of AcrAB, associated with overexpression of OqxAB, is required for high virulence potential. Keywords: K. pneumoniae, efflux systems, antimicrobial resistance
Introduction Klebsiella pneumoniae is an important pathogen responsible for a wide range of nosocomial infections.1 It can also cause invasive community-acquired infections, including liver abscess, pneumonia, severe bacteraemia and meningitis.2,3 Among the various mechanisms of resistance displayed by K. pneumoniae, efflux system overexpression has recently been shown to induce low-level cross-resistance to b-lactams (mainly cefoxitin) and to antibiotics belonging to other families, such as quinolones and chloramphenicol.4 The efflux system that has been the most extensively studied in K. pneumoniae and other members of the Enterobacteriaceae family, with regard to antibiotic resistance, is the pump AcrAB.5 – 11
However, the chromosome of K. pneumoniae harbours various other genes and operons encoding putative efflux systems, among which the rarA-oqxABR locus has recently been characterized.12 The oqxAB operon was originally described on plasmid pOLA52, carried by an Escherichia coli strain isolated from swine manure.13 It has been shown to encode an efflux pump conferring resistance to olaquindox, a growth enhancer widely used in pig farming, as well as to chloramphenicol and quinolones.14 Since then, the presence of oqxAB has been reported in K. pneumoniae isolates from various regions of the world.12,15 – 18 Moreover, Veleba et al.12 have recently described the two genes flanking the oqxAB operon in the chromosome of K. pneumoniae reference strain MGH 78578: rarA, encoding an AraC-type transcriptional activator, and oqxR, encoding a GntR-type
# The Author 2014. Published by Oxford University Press on behalf of the British Society for Antimicrobial Chemotherapy. All rights reserved. For Permissions, please e-mail: [email protected]
1 of 8
Downloaded from http://jac.oxfordjournals.org/ at University of California, San Diego on September 15, 2014
*Corresponding author. Service de Microbiologie, Hoˆpital Beaujon, 100 boulevard du Ge´ne´ral Leclerc, 92110 Clichy, France. Tel: +33-1-40-87-54-54; Fax: +33-1-40-87-05-50; E-mail: [email protected]
Bialek-Davenet et al.
Materials and methods Bacterial strains and plasmids All bacterial strains and plasmids used and/or constructed in this study are listed in Table 1.
Antibiotic susceptibility testing Susceptibility to nalidixic acid, ciprofloxacin, chloramphenicol, tigecycline and cefoxitin was determined in triplicate by the broth dilution method according to the guidelines of EUCAST.20 A difference between MICs was considered to be significant starting from two dilutions (i.e. a 4-fold MIC variation).
Sequence studies Genes rarA and oqxR, as well as the intergenic region situated between rarA and oqxA (inter-rar), containing the promoters of both genes, were amplified and sequenced using the primers listed in Table S1 (available as Supplementary data at JAC Online) and following the PCR and sequencing procedures described previously.21
Real-time RT– PCR for analysis of gene expression Real-time RT – PCR for analysis of the expression of genes rarA and oqxB was carried out and interpreted as described previously.21 Briefly, after RNA extraction and DNaseI pre-treatment, total RNA concentrations were determined using a NanoDrop Spectrophotometer and adjusted to 100 mg/L. Amplifications were performed in duplicate on a LightCycler using the one-step LightCyclerw RNA Master SYBR Green I kit (Roche Applied Science, Meylan, France). Data were analysed by the 22DDCt method; the rpoB gene was chosen as a reference and strain ATCC 13883T as a calibrator. All the primers used are listed in Table S1.
Cloning and transformation of the oqxR gene The 677 bp wild-type oqxR gene of strain ATCC 13883T [amplified using primers oqxRF and oqxRR (Table S1)] was cloned into pSC-A-amp/kan by using the StrataClone PCR Cloning Kit (Agilent Technologies, Massy, France). After verification of the construct by sequencing, the recombinant plasmid was electroporated into competent strain KPBj1 E+ and the transformants were selected on LB agar plates containing 50 mg/L kanamycin. The empty plasmid pSC-A-amp/kan-E1 21 was used as a negative control.
Gene replacement experiments Inactivation of the chromosomal acrB gene and replacement of a native oqxR allele by a mutated one were both achieved following a strategy adapted from Datsenko and Wanner.22 Briefly, recipient strains were transformed with plasmid pKOBEG199, containing the l Red region under the control of the arabinose-inducible ParaB promoter, the araC gene and a tetracycline resistance cassette.23 Concurrently, appropriate linear DNA fragments were synthesized. For acrB inactivation, an FRTflanked kanamycin resistance cassette (kan) was amplified from the template plasmid pKD4 using primers acrBForFRTKm and acrBRevFRTKm (Table S1) containing 76 –79 nt extensions homologous to the N-terminal and C-terminal flanking regions of the acrB gene, respectively. The PCR product was digested with DpnI (New England BioLabs, Evry, France) in order to remove the remaining template plasmid, and purified using the
Table 1. List of bacterial strains used in this study Strain or plasmid Strains Escherichia coli OP50 K. pneumoniae ATCC 13883T KPBj1 E+ KPBj1 Rev KPBj1 M3 Lev KPBj1 E+ DacrB KPBj1 E+ ToqxR-ATCC KPBj1 E+ T0 LM21 LM21 RoqxR-KPBj1 E+ Plasmids pSC-A-amp/kan-oqxR-ATCC pSC-A-amp/kan-E1 pKOBEG199
Description
avirulent control strain
34
wild-type reference strain MDR clinical isolate antibiotic-susceptible phenotypic revertant of KPBj1 E+ ramR-deleted mutant derived from KPBj1 Rev acrB-deleted strain derived from KPBj1 E+, KANr KPBj1 E+ complemented with plasmid pSC-A-amp/kan-oqxR-ATCC, AMPr, KANr KPBj1 E+ complemented with plasmid pSC-A-amp/kan-E1, AMPr, KANr wild-type clinical strain LM21 with the oqxR allele of strain KPBj1 E+
44 4 19 37 this work this work this work 35 this work
pSC-A-amp/kan plasmid vector containing the cloned wild-type oqxR gene, AMPr, KANr empty control plasmid vector, AMPr, KANr pBR322 derivative containing the l Red region and araC gene of pKOBEG, TETr
this work 21 23
TET, tetracycline; KAN, kanamycin; AMP, ampicillin.
2 of 8
Reference or source
Downloaded from http://jac.oxfordjournals.org/ at University of California, San Diego on September 15, 2014
transcriptional repressor. Their results suggest that rarA upregulates, whereas oqxR down-regulates, the expression of the oqxAB efflux pump operon.12 We have previously shown that three K. pneumoniae clinical isolates, presenting an antibiotic resistance pattern indicative of the overexpression of an efflux system and transcription levels of gene acrB either equal to or only slightly increased compared with those of their spontaneous phenotypic revertants, were significantly more virulent than the latter in a Caenorhabditis elegans model.19 Our aim in the present study was to investigate the possible role of the rarA-oqxABR locus in the antibiotic resistance phenotype and in the increased virulence potential of one of those three isolates and to get insight into the regulatory mechanisms of this novel efflux pump.
Role of efflux pumps in Klebsiella pneumoniae resistance and virulence
Feeding behaviour assays For occupancy assays, each bacterial strain was cultured overnight in LB medium, spotted as a circular lawn on NGM plates and dried at room temperature for 20 min. Approximately 20 L4 nematodes were carefully placed in the centre of the bacterial lawn. The number of worms inside or outside each lawn was counted after 16 h. The number of bacteria within the C. elegans digestive tract was obtained as previously described.34 Briefly, worms were picked at 72 h and the surface bacteria were removed by washing the worms twice in M9 medium containing 25 mg/L gentamicin. The nematodes were then mechanically disrupted in M9 medium containing 1% Triton X-100. Serial dilutions of the sample were then plated on LB agar and the colonies were counted after 24 h. For both types of experiment, at least three replicates were performed for each condition.
Whole-genome sequencing
Statistical analysis
Total DNA of strains KPBj1 E+ and KPBj1 Rev was extracted using the QIAamp DNA Blood Midi Kit (Qiagen). Genomic sequencing was performed by GATC Biotech (Konstanz, Germany) on an Illumina HiSeq-2000 sequencer, using Nextera technology for library construction and a 2×100 nt paired-end strategy. All reads were pre-processed to remove low-quality or artefactual nucleotides. First, all nucleotides occurring at the 5′ and 3′ ends and supported by a Phred quality score ,30 were trimmed off using Sickle (github.com/najoshi/sickle). Second, contaminant oligonucleotides (i.e. library adapters) were detected and trimmed off using AlienTrimmer.24 Third, reads shorter than 60 nt after the above cleaning steps were discarded, as well as those containing .20% nucleotides with a Phred score ,30. Finally, the program fqDuplicate (ftp.pasteur. fr/pub/gensoft/projects/fqtools) was used to discard every duplicate read pair. Resulting reads were assembled using clc_assembler from the CLC Genomics Workbench analysis package (www.clcbio.com/products/ clc-genomics-workbench/). All contigs of size ≥500 nt were reordered and reoriented, using as reference the genomic sequence of strain MGH 78578, with Mauve Contig Mover.25 The reordered contigs were subsequently imported into the Bacterial Isolate Genome Sequence Database (BIGSdb) platform26 and analysed for the presence of known efflux systems and regulators. Alleles corresponding to the seven MLST genes27 were also extracted from both genomic sequences in order to determine the ST of the strains. Furthermore, the two genomes were annotated and compared using the RAST server.28 Finally, the reads corresponding to strain KPBj1 Rev were mapped against the contig containing the rarA-oqxABR locus in the genome assembly of strain KPBj1 E+ using BWA29 and SAMtools,30 and the read alignments on the reference sequence were visualized using Artemis.31 Comparison of the genomic region surrounding the rarA-oqxABR locus in strains KPBj1 E+, KPBj1 Rev and KCTC 224232 was achieved using Easyfig version 2.1.33
To compare the entire survival curves in nematode killing assays, Cox regression was used. To perform pairwise comparison between two different strains, we used a log rank test. The analysis was carried out using SPSS 6.1.1 (SPSS Inc., Chicago, IL, USA).
Nematode killing assays The virulence potential of K. pneumoniae strains was assessed in the C. elegans model, as previously described.19 C. elegans mutant line Fer-15 was provided by the Caenorhabditis Genetics Center. Overnight cultures of bacterial strains in nematode growth medium (NGM) were harvested, centrifuged, washed once and suspended in PBS at pH 7.0 at a concentration of 105 cfu/mL. NGM agar plates were inoculated with 10 mL of bacterial suspension and incubated at 378C for 8210 h. Plates were brought back to room temperature and seeded with L4 stage worms (20230 per plate). Plates were then incubated at 258C and scored each day for live worms under a stereo microscope (Leica MS5). At least three replicates repeated five times were performed for each selected strain and avirulent strain E. coli OP50 was used as a control. Lethal time 50% (LT50) corresponded to time (days) required to kill 50% of the nematode population.
Results Contribution of AcrAB to the efflux phenotype of strain KPBj1 E1 In order to determine whether the AcrAB efflux pump was responsible for the antibiotic resistance phenotype of KPBj1 E+, we constructed an acrB deletion mutant of this clinical isolate. The MICs of antibiotics belonging to different families and known to be affected by efflux system overexpression were measured for the acrB-deleted and parental strains (Table 2). The results show that the acrB knockout conserved its efflux phenotype, since the MIC values of all the antibiotics tested were not significantly different from those measured for strain KPBj1 E+. Consequently, the MDR phenotype of this strain is unrelated to AcrAB and is probably due to overexpression of another efflux pump.
Genetic determinants of the antibiotic resistance phenotype of strain KPBj1 E1 Subsequently, we wanted to test the implication of the recently characterized rarA-oqxABR locus in the MDR phenotype of KPBj1 E+. For this purpose, we measured in this strain the expression levels of genes acrB, oqxB and rarA. Compared with the reference strain ATCC 13883T, KPBj1 E+ considerably overexpressed oqxB (89.7-fold) and rarA (1537-fold) (Table 2). Moreover, in contrast to what had been found previously, KPBj1 E+ did not seem to overexpress, even slightly, gene acrB (Table 2). Subsequently, amplification and sequencing of genes oqxR and rarA, as well as of the rarA-oqxA intergenic region, were carried out in strains KPBj1 Rev, KPBj1 E+ and ATCC 13883T. In strain KPBj1 Rev, the PCRs using the primers rarA+intF and rarA+intR, on the one hand, and oqxRF and oqxRR, on the other hand, were negative, suggesting deletion of these loci. The sequences of the rarA gene of strains KPBj1 E+ and ATCC 13883T displayed only some silent polymorphisms. As regards the oqxR gene, we detected a T305G mutation specific to strain KPBj1 E+, which was responsible for the replacement of Val-102, located in the C-terminal effectorbinding and oligomerization domain of the OqxR protein, by Gly.
3 of 8
Downloaded from http://jac.oxfordjournals.org/ at University of California, San Diego on September 15, 2014
QIAquick PCR Purification Kit (Qiagen, Courtaboeuf, France). For oqxR allelic replacement, a 175 bp fragment was amplified from the DNA of strain KPBj1 E+ using primers oqxR-175-F and oqxR-175-R (Table S1). Finally, the PCR products were electroporated into competent cells carrying pKOBEG199 in which the expression of l Red recombinase had been induced by arabinose, and transformants were selected on LB agar plates containing either 50 mg/L kanamycin, for acrB inactivation, or 16 mg/L chloramphenicol, for oqxR allelic replacement. The correct insertion of the kan cassette, leading to acrB gene knockout, was verified by PCR amplification using the primer pairs acrBupF/kt and acrBdoR/k2 on the one hand (expected to be positive) and acrBupF/acrBupR and acrBdoR/ acrBdoF on the other hand (expected to be negative) (Table S1). Similarly, allelic replacement of gene oqxR was checked by PCR and sequencing. Moreover, we obtained pKOBEG199-free transformants following successive subcultures on LB plates without tetracycline antibiotic.
JAC
Bialek-Davenet et al.
Table 2. Antibiotic susceptibility profiles and relative transcription levels of the acrB, oqxB and rarA genes for the studied K. pneumoniae strains MIC (mg/L)
Gene expression (fold change)
NAL
CIP
CHL
TGC
FOX
acrB
oqxB
rarA
ATCC 13883T KPBj1 E+ KPBj1 E+ DacrB KPBj1 E+ ToqxR-ATCC KPBj1 E+ T0 KPBj1 Rev KPBj1 M3 Lev LM21 LM21 RoqxR-KPBj1 E+
4 128 128 4 128 4 32 4 128
0.06 0.5 0.5 0.03 0.5 0.03 0.25 0.03 0.25
4 128 128 4 128 4 32 4 128
1 2 1 2 2 1 4 1 1
4 8 4 2 8 2 32 4 16
1 0.57+0.00 undetectable 0.50+0.00 0.81+0.04 1.15+0.15 3.74+0.37 0.36+0.07 0.79+0.29
1 89.7+0.9 74.9+26.6 3.48+3.07 77.7+0.0 ND ND 1.07+0.34 67.2+13.4
1 1537+602 2787+964 1.92+1.67 1756+334 ND ND 2.06+2.11 2251+942
NAL, nalidixic acid; CIP, ciprofloxacin; CHL, chloramphenicol; TGC, tigecycline; FOX, cefoxitin; ND, not determined. For gene expression levels, the results of at least two different experiments are indicated as mean+SD.
Complementation with the wild-type oqxR gene We hypothesized that the Val-102Gly mutation identified in OqxR was responsible for the loss of its repressor function and for the subsequent overexpression of the OqxAB efflux pump, resulting in an MDR phenotype. Transformation of KPBj1 E+ with pSC-A-amp/kan-oqxR-ATCC considerably lowered the transcription level of gene oqxB and the MICs of nalidixic acid, ciprofloxacin, chloramphenicol and cefoxitin, but not that of tigecycline (Table 2). Furthermore, introduction of the wild-type oqxR gene resulted in a decrease in rarA expression level, suggesting that OqxR is also a repressor of rarA. In contrast, complementation with control plasmid pSC-A-amp/kan-E1 had no significant effect on either gene expression levels or antibiotic susceptibilities (Table 2).
Allelic replacement of the oqxR gene In order to definitely demonstrate that the point mutation detected in the oqxR gene of strain KPBj1 E+ was the cause of the efflux phenotype of this strain, we performed an allelic replacement experiment: the oqxR allele originally present in the antibiotic susceptible strain LM2135 was replaced by the mutated allele of strain KPBj1 E+. This allelic change resulted in a sharp increase in the MICs of the antibiotic tested (nalidixic acid, ciprofloxacin, chloramphenicol and cefoxitin) except for tigecycline, as well as in the transcription levels of genes rarA and oqxB (Table 2).
Genomic studies The genome sequences of strains KPBj1 E+ and KPBj1 Rev were assembled into 277 and 934 contigs, respectively. Gene-by-gene analysis revealed that both strains belonged to the same sequence type, ST461, which is quite unusual. Genes acrR, marA, marR, ramA, ramR, soxR, soxS, rob, sdiA, fis and envR, coding for efflux system regulators, were detected in the two genomes with exactly the same nucleotide sequence, confirming our unpublished Sanger sequencing data. Genes oqxA, oqxB and rarA were present only in strain KPBj1 E+, whereas gene oqxR was truncated (148/480 bp) in KPBj1 Rev. Mapping of the reads corresponding to strain KPBj1 Rev against the sequence of KPBj1 E+, taken as a reference, showed the deletion of a region of
4 of 8
nearly 17.4 kb from the KPBj1 Rev genome. The deleted region comprised genes oqxA, oqxB and rarA, as well as a cluster of genes lying downstream of rarA and corresponding to a putative NAD(P)H-flavin oxidoreductase, a putative LysR-family transcriptional regulator, a putative plasmid-related protein and a number of putative phage P4-related proteins (Figure 1).
The rarA-oqxABR locus in complete public K. pneumoniae genomes Among the nine complete K. pneumoniae genomes available at the NCBI genome sequence repository (www.ncbi.nlm.nih.gov/ genome/) as of 17 December 2013, eight contained the oqxA, oqxB, oqxR and rarA genes. The rarA-oqxABR locus was absent from the genome of strain HS11286, an MDR clinical isolate from China.36 In only one of the eight remaining genomes, that of strain KCTC 2242, an isolate used for industrial 2,3-butanediol production,32 a group of genes annotated as putative phage P4-related proteins was detected. However, this gene cluster was homologous but not identical to that present in the genome of strain KPBj1 E+. A comparison of the genomic region surrounding the rarA-oqxABR locus in strain KCTC 2242 and in the two strains sequenced in the present study is depicted in Figure 1. In the genome of strain JM45, another cluster of phagic origin, having no homology with those found in strains KCTC 2242 and KPBj1 E+, was inserted after gene smpB. In the six other complete public K. pneumoniae genomes, either a completely different gene cluster was present (n¼ 2) or smpB was adjacent to the gene encoding a putative plasmid-related protein (n¼ 4). In all the genomes considered, a BLAST search identified a sequence corresponding to the tmRNA gene (ssrA; locus tag KP1_6100 in the genome of strain NTUH K2044) immediately downstream of gene smpB.
Virulence studies In order to determine the respective roles of AcrAB and OqxAB in virulence, several K. pneumoniae strains, differentially expressing both efflux pumps, were tested in the C. elegans model. Consistent with previous findings,19 strain KPBj1 E+ was significantly more virulent than both strain KPBj1 Rev (P, 0.001) and strain ATCC 13883T (P, 0.001), which showed the same virulence profile
Downloaded from http://jac.oxfordjournals.org/ at University of California, San Diego on September 15, 2014
Strain
Role of efflux pumps in Klebsiella pneumoniae resistance and virulence
JAC
(Figure 2). Deletion of the acrB gene led to a decrease in the virulence level of strain KPBj1 E+ DacrB (LT50 4.0+0.2 days) compared with strain KPBj1 E+ (LT50 3.0+0.2 days) (P ¼ 0.007) (Figure 2). Furthermore, the same effect was observed for strain KPBj1 M3 Lev (LT50 5.4+0.2 days), an in vitro selected MDR mutant of KPBj1 Rev overexpressing acrB due to a deletion of the ramR gene,37 in comparison with its parental strain (LT50 4.6+0.2 days) (P, 0.001) (Figure 2). To exclude the possibility that the variations in the observed sensitivity of nematodes could be due to an alteration of their feeding behaviour, we performed an occupancy test in which nematodes were placed in the centre of a bacterial lawn and the fractions of worms inside and outside the lawn were quantified after overnight incubation. None of the bacterial strains tested in this study induced strong avoidance behaviour and there was no significant difference in the fraction of nematodes on the bacterial lawn between the different conditions tested (data not shown). We also measured the ingestion and proliferation of bacteria in the digestive tract of nematodes for each condition. Bacterial counts in the digestive tract of the nematodes were similar whatever the strain tested, showing that all the bacteria were actually ingested.
Discussion One of the aims of this work was to explain, at the molecular level, the MDR phenotype of the clinical isolate KPBj1 E+, which had
previously been shown to slightly overexpress the acrB gene, compared with strain ATCC 13883T.19 However, the results obtained in the present study did not confirm this observation and it seems to us that the AcrAB efflux pump is, in fact, not overexpressed in KPBj1 E+. Consistently, the MIC values of nalidixic acid, ciprofloxacin, chloramphenicol, cefoxitin and tigecycline did not change in the acrB-deleted KPBj1 E+ mutant. This suggests that these antibiotics, which are well-known substrates of AcrAB in different enterobacterial species,38 – 41 are probably extruded by another efflux pump in K. pneumoniae. Indeed, in KPBj1 E+ we observed high levels of transcription of gene oqxB, as well as of the regulator rarA gene, compared with the wild-type ATCC 13883T strain. Sequencing of the oqxR gene revealed a point mutation leading to the amino acid substitution Val-102Gly in the OqxR protein of strain KPBj1 E+. This substitution is different from those reported by Veleba et al.,12 who described for the first time the rarA and oqxR loci in MDR clinical isolates of K. pneumoniae. Complementation of strain KPBj1 E+ with the wild-type oqxR gene dramatically decreased the expression levels of genes oqxB and rarA, and this result suggests that OqxR is a repressor not only of the oqxAB efflux operon, but also of rarA. In contrast to us, Veleba et al.12 did not observe any change in rarA expression levels in their clinical isolates complemented with wild-type oqxR; however, those isolates could harbour additional genetic alterations responsible for overexpression of rarA. Furthermore, in our study the acrB expression level of isolates overexpressing rarA
5 of 8
Downloaded from http://jac.oxfordjournals.org/ at University of California, San Diego on September 15, 2014
Figure 1. The rarA-oqxABR locus in some K. pneumoniae genomes. Schematic representation of the region encompassing the rarA-oqxABR locus in one K. pneumoniae public genome (strain KCTC 2242)32 compared with the two genomes sequenced in the present study. The large arrows indicate the locations and orientations of the different coding sequences, with annotations displayed below or above the arrows. The black vertical bar represents the limits of contigs in the genomic sequence of strain KPBj1 Rev. The BLAST identity scale is shown in the lower right corner, with the matches ranging from 100% (darkest) to 82% (lightest). 1, abortive infection phage resistance protein; 2, hypothetical protein; 3, phage polarity suppression protein; 4, phage late gene regulator; 5, phage capsid and scaffold protein; 6, hypothetical protein; 7, phage transcriptional regulator; 8, putative phage immunity repressor protein; 9, hypothetical protein; 10, hypothetical phage protein; 11, bacteriophage P4 DNA primase.
Bialek-Davenet et al.
9.0 8.0 7.0
** *
* *
5.0 4.0 3.0 2.0 1.0 0.0 OP50
ATCC 13883T
KPBj1 E+
KPBj1 Rev
KPBj1 M3 Lev KPBj1 E+ DacrB
Figure 2. Kinetics of killing of C. elegans infected by the studied K. pneumoniae strains. LT50, corresponding to time (days) required to kill 50% of the nematode population, is represented for each of the studied strains as the mean+SD of five experiments. E. coli strain OP50 was used as a control. Statistical significance of the differences observed is indicated as follows: *P,0.001; **P¼0.007.
was not elevated, which is also different from the findings of Veleba et al.12 and suggests that the role of RarA as a regulator of acrB is unclear. In order to definitely demonstrate that the point mutation found in the oqxR gene of strain KPBj1 E+ is sufficient to explain its MDR phenotype, an allelic replacement experiment was performed in strain LM21, a wild-type K. pneumoniae clinical isolate. As expected, substitution of the native oqxR allele by the mutated one derepressed the expression of genes oqxB and rarA in strain LM21 RoqxR-KPBj1 E+. With regard to antibiotic susceptibility, transformation of KPBj1 E+ with pSC-A-amp/kan-oqxR-ATCC decreased the MICs of nalidixic acid, ciprofloxacin, chloramphenicol and cefoxitin at least 4-fold, but did not significantly affect the MIC of tigecycline. By using isogenic strains in which the rarA-oqxABR locus is absent and the AcrAB pump is either normally expressed or overexpressed, we showed that the MICs of all of the antibiotics tested were increased in the variant overproducing AcrAB. All these results definitively indicate that nalidixic acid, ciprofloxacin, chloramphenicol and cefoxitin are substrates of both the OqxAB efflux pump and the AcrAB efflux pump, while tigecycline is a substrate of only the AcrAB efflux pump.42 Analysis of the genomic sequences of strains KPBj1 E+ and KPBj1 Rev, which is a phenotypic revertant of the former strain that had spontaneously recovered susceptibility to several antibiotics, yielded some interesting observations. First, it confirmed the clonal relatedness of these two strains, which has previously been shown by RAPD (random amplified polymorphic DNA) analysis.19 Second, the genome of KPBj1 Rev was shown to harbour a deletion of nearly 17.4 kb, encompassing genes oqxA, oqxB and rarA and two-thirds of gene oqxR, as well as a large cluster of genes lying downstream of rarA and comprising a number of putative phage P4-related proteins, when compared with the
6 of 8
corresponding region in the genome of strain KPBj1 E+. Among the nine complete K. pneumoniae genomes publicly available at the time period of our study, only one contained a putative phage P4-related gene cluster inserted in the vicinity of the rarA-oqxABR locus, whereas three others harboured a different phage or gene cluster. Gene ssrA, situated in this genomic region just downstream of smpB and encoding a tmRNA, is a known hot spot of integration of mobile genetic elements in Enterobacteriaceae.43 We suppose that the phage has been excised from the genome of KPBj1 E+, accidentally removing the neighbouring genes and thus leading to strain KPBj1 Rev, previously considered a spontaneous revertant of KPBj1 E+ because of its loss of the cross-resistance phenotype. In our previous study of K. pneumoniae virulence assessed in the C. elegans model, we surmised that the difference in virulence level between strains KPBj1 E+ and KPBj1 Rev was due to a slight variation in the expression of gene acrB.19 However, since we have shown in the present study that KPBj1 E+ in fact does not overexpress acrB, our current hypothesis is that the high virulence of this strain is related to the overexpression of genes oqxB and/or rarA. With regard to the role of AcrAB in K. pneumoniae virulence, we observed that the acrB deletion mutant was significantly less virulent than its parental strain, KPBj1 E+. Our results are consistent with a previous study by Padilla et al.,11 which also demonstrated that the inactivation of gene acrB attenuated the virulence potential of the highly pathogenic K. pneumoniae strain 52145. However, quite unexpectedly, the virulence level of KPBj1 M3 Lev, an acrB-overexpressing in vitro-selected mutant of KPBj1 Rev, was lower than that of its parental strain. From this we conclude that acrB overexpression does not have by itself a positive effect on virulence, although the low virulence of strain KPBj1 M3 Lev might also be due to the fact that this strain carries a
Downloaded from http://jac.oxfordjournals.org/ at University of California, San Diego on September 15, 2014
LT50 (days)
6.0
Role of efflux pumps in Klebsiella pneumoniae resistance and virulence
deletion of the entire ramR gene.37 In order to definitely address these questions, further studies are needed and are currently in progress. To conclude, we found that overexpression of the OqxAB efflux pump, resulting from a point mutation in the repressor oqxR gene, was responsible for the MDR phenotype and for the increased virulence potential of one K. pneumoniae clinical isolate. As this pump is widely distributed in K. pneumoniae genomes,12,15 – 18 but not necessarily overexpressed, it would be interesting to determine the extent to which it participates in clinically relevant antimicrobial resistance and in pathogenicity of nosocomial and community-acquired isolates.
We thank C. Forestier for kindly providing K. pneumoniae strain LM21.
Funding The research leading to these results has received support from the Innovative Medicines Joint Undertaking under Grant Agreement No. 115525, resources which are composed of financial contributions from the European Union’s seventh framework programme (FP7/2007-2013) and European Federation of Pharmaceutical Industries and Associations companies in kind contribution.
Transparency declarations None to declare.
Supplementary data Table S1 is available as Supplementary data at JAC Online (http://jac. oxfordjournals.org/).
References 1 Podschun R, Ullmann U. Klebsiella spp. as nosocomial pathogens: epidemiology, taxonomy, typing methods, and pathogenicity factors. Clin Microbiol Rev 1998; 11: 589– 603. 2 Keynan Y, Rubinstein E. The changing face of Klebsiella pneumoniae infections in the community. Int J Antimicrob Agents 2007; 30: 385– 9. 3 Decre´ D, Verdet C, Emirian A et al. Emerging severe and fatal infections due to Klebsiella pneumoniae in two university hospitals in France. J Clin Microbiol 2011; 49: 3012– 4. 4 Pages JM, Lavigne JP, Leflon-Guibout V et al. Efflux pump, the masked side of b-lactam resistance in Klebsiella pneumoniae clinical isolates. PLoS One 2009; 4: e4817. 5 Okusu H, Ma D, Nikaido H. AcrAB efflux pump plays a major role in the antibiotic resistance phenotype of Escherichia coli multiple-antibioticresistance (Mar) mutants. J Bacteriol 1996; 178: 306– 8. 6 Piddock LJ, White DG, Gensberg K et al. Evidence for an efflux pump mediating multiple antibiotic resistance in Salmonella enterica serovar Typhimurium. Antimicrob Agents Chemother 2000; 44: 3118– 21.
9 Perez A, Canle D, Latasa C et al. Cloning, nucleotide sequencing, and analysis of the AcrAB-TolC efflux pump of Enterobacter cloacae and determination of its involvement in antibiotic resistance in a clinical isolate. Antimicrob Agents Chemother 2007; 51: 3247– 53. 10 Sanchez-Cespedes J, Vila J. Partial characterisation of the acrAB locus in two Citrobacter freundii clinical isolates. Int J Antimicrob Agents 2007; 30: 259–63. 11 Padilla E, Llobet E, Domenech-Sanchez A et al. Klebsiella pneumoniae AcrAB efflux pump contributes to antimicrobial resistance and virulence. Antimicrob Agents Chemother 2010; 54: 177–83. 12 Veleba M, Higgins PG, Gonzalez G et al. Characterization of RarA, a novel AraC family multidrug resistance regulator in Klebsiella pneumoniae. Antimicrob Agents Chemother 2012; 56: 4450– 8. 13 Hansen LH, Johannesen E, Burmolle M et al. Plasmid-encoded multidrug efflux pump conferring resistance to olaquindox in Escherichia coli. Antimicrob Agents Chemother 2004; 48: 3332– 7. 14 Hansen LH, Jensen LB, Sorensen HI et al. Substrate specificity of the OqxAB multidrug resistance pump in Escherichia coli and selected enteric bacteria. J Antimicrob Chemother 2007; 60: 145–7. 15 Park KS, Kim MH, Park TS et al. Prevalence of the plasmid-mediated quinolone resistance genes, aac(6 ′ )-Ib-cr, qepA, and oqxAB in clinical isolates of extended-spectrum b-lactamase (ESBL)-producing Escherichia coli and Klebsiella pneumoniae in Korea. Ann Clin Lab Sci 2012; 42: 191– 7. 16 Yuan J, Xu X, Guo Q et al. Prevalence of the oqxAB gene complex in Klebsiella pneumoniae and Escherichia coli clinical isolates. J Antimicrob Chemother 2012; 67: 1655– 9. 17 Rodriguez-Martinez JM, Diaz de Alba P, Briales A et al. Contribution of OqxAB efflux pumps to quinolone resistance in extended-spectrumb-lactamase-producing Klebsiella pneumoniae. J Antimicrob Chemother 2013; 68: 68– 73. 18 Perez F, Rudin SD, Marshall SH et al. OqxAB, a quinolone and olaquindox efflux pump, is widely distributed among multidrug-resistant Klebsiella pneumoniae isolates of human origin. Antimicrob Agents Chemother 2013; 57: 4602– 3. 19 Bialek S, Lavigne JP, Chevalier J et al. Membrane efflux and influx modulate both multidrug resistance and virulence of Klebsiella pneumoniae in a Caenorhabditis elegans model. Antimicrob Agents Chemother 2010; 54: 4373– 8. 20 EUCAST. Determination of minimum inhibitory concentration (MICs) of antibacterial agents by broth dilution. Clin Microbiol Infect 2003; 9: 1 –7. 21 Bialek-Davenet S, Marcon E, Leflon-Guibout V et al. In vitro selection of ramR and soxR mutants overexpressing efflux systems by fluoroquinolones as well as cefoxitin in Klebsiella pneumoniae. Antimicrob Agents Chemother 2011; 55: 2795– 802. 22 Datsenko KA, Wanner BL. One-step inactivation of chromosomal genes in Escherichia coli K-12 using PCR products. Proc Natl Acad Sci USA 2000; 97: 6640– 5. 23 Balestrino D, Haagensen JA, Rich C et al. Characterization of type 2 quorum sensing in Klebsiella pneumoniae and relationship with biofilm formation. J Bacteriol 2005; 187: 2870 –80. 24 Criscuolo A, Brisse S. AlienTrimmer: a tool to quickly and accurately trim off multiple short contaminant sequences from high-throughput sequencing reads. Genomics 2013; 102: 500– 6. 25 Rissman AI, Mau B, Biehl BS et al. Reordering contigs of draft genomes using the Mauve aligner. Bioinformatics 2009; 25: 2071– 3.
7 Mazzariol A, Zuliani J, Cornaglia G et al. AcrAB efflux system: expression and contribution to fluoroquinolone resistance in Klebsiella spp. Antimicrob Agents Chemother 2002; 46: 3984 –6.
26 Jolley KA, Maiden MC. BIGSdb: scalable analysis of bacterial genome variation at the population level. BMC Bioinformatics 2010; 11: 595.
8 Pradel E, Pages JM. The AcrAB-TolC efflux pump contributes to multidrug resistance in the nosocomial pathogen Enterobacter aerogenes. Antimicrob Agents Chemother 2002; 46: 2640– 3.
27 Diancourt L, Passet V, Verhoef J et al. Multilocus sequence typing of Klebsiella pneumoniae nosocomial isolates. J Clin Microbiol 2005; 43: 4178– 82.
7 of 8
Downloaded from http://jac.oxfordjournals.org/ at University of California, San Diego on September 15, 2014
Acknowledgements
JAC
Bialek-Davenet et al.
28 Aziz RK, Bartels D, Best AA et al. The RAST server: rapid annotations using subsystems technology. BMC Genomics 2008; 9: 75. 29 Li H, Durbin R. Fast and accurate short read alignment with Burrows-Wheeler transform. Bioinformatics 2009; 25: 1754 –60. 30 Li H, Handsaker B, Wysoker A et al. The sequence alignment/Map format and SAMtools. Bioinformatics 2009; 25: 2078–9. 31 Carver T, Harris SR, Berriman M et al. Artemis: an integrated platform for visualization and analysis of high-throughput sequence-based experimental data. Bioinformatics 2012; 28: 464–9. 32 Shin SH, Kim S, Kim JY et al. Complete genome sequence of the 2,3-butanediol-producing Klebsiella pneumoniae strain KCTC 2242. J Bacteriol 2012; 194: 2736– 7.
34 Garsin DA, Sifri CD, Mylonakis E et al. A simple model host for identifying Gram-positive virulence factors. Proc Natl Acad Sci USA 2001; 98: 10892– 7. 35 Favre-Bonte S, Joly B, Forestier C. Consequences of reduction of Klebsiella pneumoniae capsule expression on interactions of this bacterium with epithelial cells. Infect Immun 1999; 67: 554–61. 36 Liu P, Li P, Jiang X et al. Complete genome sequence of Klebsiella pneumoniae subsp. pneumoniae HS11286, a multidrug-resistant strain isolated from human sputum. J Bacteriol 2012; 194: 1841– 2.
8 of 8
38 Visalli MA, Murphy E, Projan SJ et al. AcrAB multidrug efflux pump is associated with reduced levels of susceptibility to tigecycline (GAR-936) in Proteus mirabilis. Antimicrob Agents Chemother 2003; 47: 665–9. 39 Ruzin A, Visalli MA, Keeney D et al. Influence of transcriptional activator RamA on expression of multidrug efflux pump AcrAB and tigecycline susceptibility in Klebsiella pneumoniae. Antimicrob Agents Chemother 2005; 49: 1017– 22. 40 Piddock LJ. Clinically relevant chromosomally encoded multidrug resistance efflux pumps in bacteria. Clin Microbiol Rev 2006; 19: 382–402. 41 Nikaido H, Takatsuka Y. Mechanisms of RND multidrug efflux pumps. Biochim Biophys Acta 2009; 1794: 769– 81. 42 Veleba M, Schneiders T. Tigecycline resistance can occur independently of the ramA gene in Klebsiella pneumoniae. Antimicrob Agents Chemother 2012; 56: 4466– 7. 43 Williams KP. Traffic at the tmRNA gene. J Bacteriol 2003; 185: 1059– 70. 44 Skerman VBD, Mc Gowan V, Sneath PHA. Approved lists of bacterial names. Int J Syst Bacteriol 1980; 30: 225– 420.
Downloaded from http://jac.oxfordjournals.org/ at University of California, San Diego on September 15, 2014
33 Sullivan MJ, Petty NK, Beatson SA. Easyfig: a genome comparison visualizer. Bioinformatics 2011; 27: 1009 –10.
37 Bialek-Davenet S, Leflon-Guibout V, Tran Minh O et al. Complete deletion of the ramR gene in an in vitro-selected mutant of Klebsiella pneumoniae overexpressing the AcrAB efflux pump. Antimicrob Agents Chemother 2013; 57: 672–3.
J Antimicrob Chemother doi:10.1093/jac/dku340
Differential contribution of AcrAB and OqxAB efflux pumps to multidrug resistance and virulence in Klebsiella pneumoniae Suzanne Bialek-Davenet1–3, Jean-Philippe Lavigne4,5, Kathleen Guyot1,6, Noe´mie Mayer1, Re´gis Tournebize7,8, Sylvain Brisse2,3, Ve´ronique Leflon-Guibout1 and Marie-He´le`ne Nicolas-Chanoine1,6,9* 1
Service de Microbiologie, Hoˆpital Beaujon, AP-HP, Clichy, France; 2Institut Pasteur, Microbial Evolutionary Genomics, Paris, France; CNRS, UMR3525, Paris, France; 4INSERM U1047, Universite´ Montpellier 1, UFR de Me´decine, Nıˆmes, France; 5Laboratoire de Bacte´riologie, CHU Caremeau, Nıˆmes, France; 6Faculte´ de Me´decine Paris Diderot, Paris, France; 7Institut Pasteur, Pathoge´nie Microbienne Mole´culaire, Paris, France; 8INSERM U786, Paris, France; 9INSERM UMR 1149, Universite´ Paris 7, Paris, France
3
Received 13 March 2014; returned 30 May 2014; revised 15 July 2014; accepted 5 August 2014 Objectives: In Klebsiella pneumoniae, overexpression of the AcrAB efflux pump and the more recently described OqxAB efflux pump has been linked to an antibiotic cross-resistance phenotype, but the mechanisms of regulation are largely unknown. Moreover, while AcrAB has been shown to participate in K. pneumoniae virulence, the contribution of OqxAB has not yet been assessed. Methods: In the present study we investigated a K. pneumoniae clinical isolate (KPBj1 E+), displaying crossresistance to quinolones, chloramphenicol and cefoxitin, and its phenotypic revertant (KPBj1 Rev, susceptible to antibiotics) by using whole-genome sequencing, RT–PCR, complementation and a Caenorhabditis elegans virulence model. Results: We detected a point mutation in the oqxR repressor gene of KPBj1 E+, which overexpressed genes rarA, encoding a transcriptional regulator, and oqxB, but not acrB. Complementation with wild-type oqxR restored antibiotic susceptibility and normalized rarA and oqxB expression levels. Whole-genome sequencing showed that KPBj1 Rev had lost the entire rarA-oqxABR locus, situated close to an integration hot spot of phage P4. This large deletion seemed responsible for the significantly lower virulence potential of strain KPBj1 Rev compared with KPBj1 E+. Moreover, we found that KPBj1 E+ DacrB was significantly less virulent than its parental strain. Conclusions: This work demonstrates the role of the overexpression of efflux pump OqxAB, due to a mutation in gene oqxR, in the antibiotic resistance phenotype of a clinical isolate, and suggests that the presence of AcrAB, associated with overexpression of OqxAB, is required for high virulence potential. Keywords: K. pneumoniae, efflux systems, antimicrobial resistance
Introduction Klebsiella pneumoniae is an important pathogen responsible for a wide range of nosocomial infections.1 It can also cause invasive community-acquired infections, including liver abscess, pneumonia, severe bacteraemia and meningitis.2,3 Among the various mechanisms of resistance displayed by K. pneumoniae, efflux system overexpression has recently been shown to induce low-level cross-resistance to b-lactams (mainly cefoxitin) and to antibiotics belonging to other families, such as quinolones and chloramphenicol.4 The efflux system that has been the most extensively studied in K. pneumoniae and other members of the Enterobacteriaceae family, with regard to antibiotic resistance, is the pump AcrAB.5 – 11
However, the chromosome of K. pneumoniae harbours various other genes and operons encoding putative efflux systems, among which the rarA-oqxABR locus has recently been characterized.12 The oqxAB operon was originally described on plasmid pOLA52, carried by an Escherichia coli strain isolated from swine manure.13 It has been shown to encode an efflux pump conferring resistance to olaquindox, a growth enhancer widely used in pig farming, as well as to chloramphenicol and quinolones.14 Since then, the presence of oqxAB has been reported in K. pneumoniae isolates from various regions of the world.12,15 – 18 Moreover, Veleba et al.12 have recently described the two genes flanking the oqxAB operon in the chromosome of K. pneumoniae reference strain MGH 78578: rarA, encoding an AraC-type transcriptional activator, and oqxR, encoding a GntR-type
# The Author 2014. Published by Oxford University Press on behalf of the British Society for Antimicrobial Chemotherapy. All rights reserved. For Permissions, please e-mail: [email protected]
1 of 8
Downloaded from http://jac.oxfordjournals.org/ at University of California, San Diego on September 15, 2014
*Corresponding author. Service de Microbiologie, Hoˆpital Beaujon, 100 boulevard du Ge´ne´ral Leclerc, 92110 Clichy, France. Tel: +33-1-40-87-54-54; Fax: +33-1-40-87-05-50; E-mail: [email protected]
Bialek-Davenet et al.
Materials and methods Bacterial strains and plasmids All bacterial strains and plasmids used and/or constructed in this study are listed in Table 1.
Antibiotic susceptibility testing Susceptibility to nalidixic acid, ciprofloxacin, chloramphenicol, tigecycline and cefoxitin was determined in triplicate by the broth dilution method according to the guidelines of EUCAST.20 A difference between MICs was considered to be significant starting from two dilutions (i.e. a 4-fold MIC variation).
Sequence studies Genes rarA and oqxR, as well as the intergenic region situated between rarA and oqxA (inter-rar), containing the promoters of both genes, were amplified and sequenced using the primers listed in Table S1 (available as Supplementary data at JAC Online) and following the PCR and sequencing procedures described previously.21
Real-time RT– PCR for analysis of gene expression Real-time RT – PCR for analysis of the expression of genes rarA and oqxB was carried out and interpreted as described previously.21 Briefly, after RNA extraction and DNaseI pre-treatment, total RNA concentrations were determined using a NanoDrop Spectrophotometer and adjusted to 100 mg/L. Amplifications were performed in duplicate on a LightCycler using the one-step LightCyclerw RNA Master SYBR Green I kit (Roche Applied Science, Meylan, France). Data were analysed by the 22DDCt method; the rpoB gene was chosen as a reference and strain ATCC 13883T as a calibrator. All the primers used are listed in Table S1.
Cloning and transformation of the oqxR gene The 677 bp wild-type oqxR gene of strain ATCC 13883T [amplified using primers oqxRF and oqxRR (Table S1)] was cloned into pSC-A-amp/kan by using the StrataClone PCR Cloning Kit (Agilent Technologies, Massy, France). After verification of the construct by sequencing, the recombinant plasmid was electroporated into competent strain KPBj1 E+ and the transformants were selected on LB agar plates containing 50 mg/L kanamycin. The empty plasmid pSC-A-amp/kan-E1 21 was used as a negative control.
Gene replacement experiments Inactivation of the chromosomal acrB gene and replacement of a native oqxR allele by a mutated one were both achieved following a strategy adapted from Datsenko and Wanner.22 Briefly, recipient strains were transformed with plasmid pKOBEG199, containing the l Red region under the control of the arabinose-inducible ParaB promoter, the araC gene and a tetracycline resistance cassette.23 Concurrently, appropriate linear DNA fragments were synthesized. For acrB inactivation, an FRTflanked kanamycin resistance cassette (kan) was amplified from the template plasmid pKD4 using primers acrBForFRTKm and acrBRevFRTKm (Table S1) containing 76 –79 nt extensions homologous to the N-terminal and C-terminal flanking regions of the acrB gene, respectively. The PCR product was digested with DpnI (New England BioLabs, Evry, France) in order to remove the remaining template plasmid, and purified using the
Table 1. List of bacterial strains used in this study Strain or plasmid Strains Escherichia coli OP50 K. pneumoniae ATCC 13883T KPBj1 E+ KPBj1 Rev KPBj1 M3 Lev KPBj1 E+ DacrB KPBj1 E+ ToqxR-ATCC KPBj1 E+ T0 LM21 LM21 RoqxR-KPBj1 E+ Plasmids pSC-A-amp/kan-oqxR-ATCC pSC-A-amp/kan-E1 pKOBEG199
Description
avirulent control strain
34
wild-type reference strain MDR clinical isolate antibiotic-susceptible phenotypic revertant of KPBj1 E+ ramR-deleted mutant derived from KPBj1 Rev acrB-deleted strain derived from KPBj1 E+, KANr KPBj1 E+ complemented with plasmid pSC-A-amp/kan-oqxR-ATCC, AMPr, KANr KPBj1 E+ complemented with plasmid pSC-A-amp/kan-E1, AMPr, KANr wild-type clinical strain LM21 with the oqxR allele of strain KPBj1 E+
44 4 19 37 this work this work this work 35 this work
pSC-A-amp/kan plasmid vector containing the cloned wild-type oqxR gene, AMPr, KANr empty control plasmid vector, AMPr, KANr pBR322 derivative containing the l Red region and araC gene of pKOBEG, TETr
this work 21 23
TET, tetracycline; KAN, kanamycin; AMP, ampicillin.
2 of 8
Reference or source
Downloaded from http://jac.oxfordjournals.org/ at University of California, San Diego on September 15, 2014
transcriptional repressor. Their results suggest that rarA upregulates, whereas oqxR down-regulates, the expression of the oqxAB efflux pump operon.12 We have previously shown that three K. pneumoniae clinical isolates, presenting an antibiotic resistance pattern indicative of the overexpression of an efflux system and transcription levels of gene acrB either equal to or only slightly increased compared with those of their spontaneous phenotypic revertants, were significantly more virulent than the latter in a Caenorhabditis elegans model.19 Our aim in the present study was to investigate the possible role of the rarA-oqxABR locus in the antibiotic resistance phenotype and in the increased virulence potential of one of those three isolates and to get insight into the regulatory mechanisms of this novel efflux pump.
Role of efflux pumps in Klebsiella pneumoniae resistance and virulence
Feeding behaviour assays For occupancy assays, each bacterial strain was cultured overnight in LB medium, spotted as a circular lawn on NGM plates and dried at room temperature for 20 min. Approximately 20 L4 nematodes were carefully placed in the centre of the bacterial lawn. The number of worms inside or outside each lawn was counted after 16 h. The number of bacteria within the C. elegans digestive tract was obtained as previously described.34 Briefly, worms were picked at 72 h and the surface bacteria were removed by washing the worms twice in M9 medium containing 25 mg/L gentamicin. The nematodes were then mechanically disrupted in M9 medium containing 1% Triton X-100. Serial dilutions of the sample were then plated on LB agar and the colonies were counted after 24 h. For both types of experiment, at least three replicates were performed for each condition.
Whole-genome sequencing
Statistical analysis
Total DNA of strains KPBj1 E+ and KPBj1 Rev was extracted using the QIAamp DNA Blood Midi Kit (Qiagen). Genomic sequencing was performed by GATC Biotech (Konstanz, Germany) on an Illumina HiSeq-2000 sequencer, using Nextera technology for library construction and a 2×100 nt paired-end strategy. All reads were pre-processed to remove low-quality or artefactual nucleotides. First, all nucleotides occurring at the 5′ and 3′ ends and supported by a Phred quality score ,30 were trimmed off using Sickle (github.com/najoshi/sickle). Second, contaminant oligonucleotides (i.e. library adapters) were detected and trimmed off using AlienTrimmer.24 Third, reads shorter than 60 nt after the above cleaning steps were discarded, as well as those containing .20% nucleotides with a Phred score ,30. Finally, the program fqDuplicate (ftp.pasteur. fr/pub/gensoft/projects/fqtools) was used to discard every duplicate read pair. Resulting reads were assembled using clc_assembler from the CLC Genomics Workbench analysis package (www.clcbio.com/products/ clc-genomics-workbench/). All contigs of size ≥500 nt were reordered and reoriented, using as reference the genomic sequence of strain MGH 78578, with Mauve Contig Mover.25 The reordered contigs were subsequently imported into the Bacterial Isolate Genome Sequence Database (BIGSdb) platform26 and analysed for the presence of known efflux systems and regulators. Alleles corresponding to the seven MLST genes27 were also extracted from both genomic sequences in order to determine the ST of the strains. Furthermore, the two genomes were annotated and compared using the RAST server.28 Finally, the reads corresponding to strain KPBj1 Rev were mapped against the contig containing the rarA-oqxABR locus in the genome assembly of strain KPBj1 E+ using BWA29 and SAMtools,30 and the read alignments on the reference sequence were visualized using Artemis.31 Comparison of the genomic region surrounding the rarA-oqxABR locus in strains KPBj1 E+, KPBj1 Rev and KCTC 224232 was achieved using Easyfig version 2.1.33
To compare the entire survival curves in nematode killing assays, Cox regression was used. To perform pairwise comparison between two different strains, we used a log rank test. The analysis was carried out using SPSS 6.1.1 (SPSS Inc., Chicago, IL, USA).
Nematode killing assays The virulence potential of K. pneumoniae strains was assessed in the C. elegans model, as previously described.19 C. elegans mutant line Fer-15 was provided by the Caenorhabditis Genetics Center. Overnight cultures of bacterial strains in nematode growth medium (NGM) were harvested, centrifuged, washed once and suspended in PBS at pH 7.0 at a concentration of 105 cfu/mL. NGM agar plates were inoculated with 10 mL of bacterial suspension and incubated at 378C for 8210 h. Plates were brought back to room temperature and seeded with L4 stage worms (20230 per plate). Plates were then incubated at 258C and scored each day for live worms under a stereo microscope (Leica MS5). At least three replicates repeated five times were performed for each selected strain and avirulent strain E. coli OP50 was used as a control. Lethal time 50% (LT50) corresponded to time (days) required to kill 50% of the nematode population.
Results Contribution of AcrAB to the efflux phenotype of strain KPBj1 E1 In order to determine whether the AcrAB efflux pump was responsible for the antibiotic resistance phenotype of KPBj1 E+, we constructed an acrB deletion mutant of this clinical isolate. The MICs of antibiotics belonging to different families and known to be affected by efflux system overexpression were measured for the acrB-deleted and parental strains (Table 2). The results show that the acrB knockout conserved its efflux phenotype, since the MIC values of all the antibiotics tested were not significantly different from those measured for strain KPBj1 E+. Consequently, the MDR phenotype of this strain is unrelated to AcrAB and is probably due to overexpression of another efflux pump.
Genetic determinants of the antibiotic resistance phenotype of strain KPBj1 E1 Subsequently, we wanted to test the implication of the recently characterized rarA-oqxABR locus in the MDR phenotype of KPBj1 E+. For this purpose, we measured in this strain the expression levels of genes acrB, oqxB and rarA. Compared with the reference strain ATCC 13883T, KPBj1 E+ considerably overexpressed oqxB (89.7-fold) and rarA (1537-fold) (Table 2). Moreover, in contrast to what had been found previously, KPBj1 E+ did not seem to overexpress, even slightly, gene acrB (Table 2). Subsequently, amplification and sequencing of genes oqxR and rarA, as well as of the rarA-oqxA intergenic region, were carried out in strains KPBj1 Rev, KPBj1 E+ and ATCC 13883T. In strain KPBj1 Rev, the PCRs using the primers rarA+intF and rarA+intR, on the one hand, and oqxRF and oqxRR, on the other hand, were negative, suggesting deletion of these loci. The sequences of the rarA gene of strains KPBj1 E+ and ATCC 13883T displayed only some silent polymorphisms. As regards the oqxR gene, we detected a T305G mutation specific to strain KPBj1 E+, which was responsible for the replacement of Val-102, located in the C-terminal effectorbinding and oligomerization domain of the OqxR protein, by Gly.
3 of 8
Downloaded from http://jac.oxfordjournals.org/ at University of California, San Diego on September 15, 2014
QIAquick PCR Purification Kit (Qiagen, Courtaboeuf, France). For oqxR allelic replacement, a 175 bp fragment was amplified from the DNA of strain KPBj1 E+ using primers oqxR-175-F and oqxR-175-R (Table S1). Finally, the PCR products were electroporated into competent cells carrying pKOBEG199 in which the expression of l Red recombinase had been induced by arabinose, and transformants were selected on LB agar plates containing either 50 mg/L kanamycin, for acrB inactivation, or 16 mg/L chloramphenicol, for oqxR allelic replacement. The correct insertion of the kan cassette, leading to acrB gene knockout, was verified by PCR amplification using the primer pairs acrBupF/kt and acrBdoR/k2 on the one hand (expected to be positive) and acrBupF/acrBupR and acrBdoR/ acrBdoF on the other hand (expected to be negative) (Table S1). Similarly, allelic replacement of gene oqxR was checked by PCR and sequencing. Moreover, we obtained pKOBEG199-free transformants following successive subcultures on LB plates without tetracycline antibiotic.
JAC
Bialek-Davenet et al.
Table 2. Antibiotic susceptibility profiles and relative transcription levels of the acrB, oqxB and rarA genes for the studied K. pneumoniae strains MIC (mg/L)
Gene expression (fold change)
NAL
CIP
CHL
TGC
FOX
acrB
oqxB
rarA
ATCC 13883T KPBj1 E+ KPBj1 E+ DacrB KPBj1 E+ ToqxR-ATCC KPBj1 E+ T0 KPBj1 Rev KPBj1 M3 Lev LM21 LM21 RoqxR-KPBj1 E+
4 128 128 4 128 4 32 4 128
0.06 0.5 0.5 0.03 0.5 0.03 0.25 0.03 0.25
4 128 128 4 128 4 32 4 128
1 2 1 2 2 1 4 1 1
4 8 4 2 8 2 32 4 16
1 0.57+0.00 undetectable 0.50+0.00 0.81+0.04 1.15+0.15 3.74+0.37 0.36+0.07 0.79+0.29
1 89.7+0.9 74.9+26.6 3.48+3.07 77.7+0.0 ND ND 1.07+0.34 67.2+13.4
1 1537+602 2787+964 1.92+1.67 1756+334 ND ND 2.06+2.11 2251+942
NAL, nalidixic acid; CIP, ciprofloxacin; CHL, chloramphenicol; TGC, tigecycline; FOX, cefoxitin; ND, not determined. For gene expression levels, the results of at least two different experiments are indicated as mean+SD.
Complementation with the wild-type oqxR gene We hypothesized that the Val-102Gly mutation identified in OqxR was responsible for the loss of its repressor function and for the subsequent overexpression of the OqxAB efflux pump, resulting in an MDR phenotype. Transformation of KPBj1 E+ with pSC-A-amp/kan-oqxR-ATCC considerably lowered the transcription level of gene oqxB and the MICs of nalidixic acid, ciprofloxacin, chloramphenicol and cefoxitin, but not that of tigecycline (Table 2). Furthermore, introduction of the wild-type oqxR gene resulted in a decrease in rarA expression level, suggesting that OqxR is also a repressor of rarA. In contrast, complementation with control plasmid pSC-A-amp/kan-E1 had no significant effect on either gene expression levels or antibiotic susceptibilities (Table 2).
Allelic replacement of the oqxR gene In order to definitely demonstrate that the point mutation detected in the oqxR gene of strain KPBj1 E+ was the cause of the efflux phenotype of this strain, we performed an allelic replacement experiment: the oqxR allele originally present in the antibiotic susceptible strain LM2135 was replaced by the mutated allele of strain KPBj1 E+. This allelic change resulted in a sharp increase in the MICs of the antibiotic tested (nalidixic acid, ciprofloxacin, chloramphenicol and cefoxitin) except for tigecycline, as well as in the transcription levels of genes rarA and oqxB (Table 2).
Genomic studies The genome sequences of strains KPBj1 E+ and KPBj1 Rev were assembled into 277 and 934 contigs, respectively. Gene-by-gene analysis revealed that both strains belonged to the same sequence type, ST461, which is quite unusual. Genes acrR, marA, marR, ramA, ramR, soxR, soxS, rob, sdiA, fis and envR, coding for efflux system regulators, were detected in the two genomes with exactly the same nucleotide sequence, confirming our unpublished Sanger sequencing data. Genes oqxA, oqxB and rarA were present only in strain KPBj1 E+, whereas gene oqxR was truncated (148/480 bp) in KPBj1 Rev. Mapping of the reads corresponding to strain KPBj1 Rev against the sequence of KPBj1 E+, taken as a reference, showed the deletion of a region of
4 of 8
nearly 17.4 kb from the KPBj1 Rev genome. The deleted region comprised genes oqxA, oqxB and rarA, as well as a cluster of genes lying downstream of rarA and corresponding to a putative NAD(P)H-flavin oxidoreductase, a putative LysR-family transcriptional regulator, a putative plasmid-related protein and a number of putative phage P4-related proteins (Figure 1).
The rarA-oqxABR locus in complete public K. pneumoniae genomes Among the nine complete K. pneumoniae genomes available at the NCBI genome sequence repository (www.ncbi.nlm.nih.gov/ genome/) as of 17 December 2013, eight contained the oqxA, oqxB, oqxR and rarA genes. The rarA-oqxABR locus was absent from the genome of strain HS11286, an MDR clinical isolate from China.36 In only one of the eight remaining genomes, that of strain KCTC 2242, an isolate used for industrial 2,3-butanediol production,32 a group of genes annotated as putative phage P4-related proteins was detected. However, this gene cluster was homologous but not identical to that present in the genome of strain KPBj1 E+. A comparison of the genomic region surrounding the rarA-oqxABR locus in strain KCTC 2242 and in the two strains sequenced in the present study is depicted in Figure 1. In the genome of strain JM45, another cluster of phagic origin, having no homology with those found in strains KCTC 2242 and KPBj1 E+, was inserted after gene smpB. In the six other complete public K. pneumoniae genomes, either a completely different gene cluster was present (n¼ 2) or smpB was adjacent to the gene encoding a putative plasmid-related protein (n¼ 4). In all the genomes considered, a BLAST search identified a sequence corresponding to the tmRNA gene (ssrA; locus tag KP1_6100 in the genome of strain NTUH K2044) immediately downstream of gene smpB.
Virulence studies In order to determine the respective roles of AcrAB and OqxAB in virulence, several K. pneumoniae strains, differentially expressing both efflux pumps, were tested in the C. elegans model. Consistent with previous findings,19 strain KPBj1 E+ was significantly more virulent than both strain KPBj1 Rev (P, 0.001) and strain ATCC 13883T (P, 0.001), which showed the same virulence profile
Downloaded from http://jac.oxfordjournals.org/ at University of California, San Diego on September 15, 2014
Strain
Role of efflux pumps in Klebsiella pneumoniae resistance and virulence
JAC
(Figure 2). Deletion of the acrB gene led to a decrease in the virulence level of strain KPBj1 E+ DacrB (LT50 4.0+0.2 days) compared with strain KPBj1 E+ (LT50 3.0+0.2 days) (P ¼ 0.007) (Figure 2). Furthermore, the same effect was observed for strain KPBj1 M3 Lev (LT50 5.4+0.2 days), an in vitro selected MDR mutant of KPBj1 Rev overexpressing acrB due to a deletion of the ramR gene,37 in comparison with its parental strain (LT50 4.6+0.2 days) (P, 0.001) (Figure 2). To exclude the possibility that the variations in the observed sensitivity of nematodes could be due to an alteration of their feeding behaviour, we performed an occupancy test in which nematodes were placed in the centre of a bacterial lawn and the fractions of worms inside and outside the lawn were quantified after overnight incubation. None of the bacterial strains tested in this study induced strong avoidance behaviour and there was no significant difference in the fraction of nematodes on the bacterial lawn between the different conditions tested (data not shown). We also measured the ingestion and proliferation of bacteria in the digestive tract of nematodes for each condition. Bacterial counts in the digestive tract of the nematodes were similar whatever the strain tested, showing that all the bacteria were actually ingested.
Discussion One of the aims of this work was to explain, at the molecular level, the MDR phenotype of the clinical isolate KPBj1 E+, which had
previously been shown to slightly overexpress the acrB gene, compared with strain ATCC 13883T.19 However, the results obtained in the present study did not confirm this observation and it seems to us that the AcrAB efflux pump is, in fact, not overexpressed in KPBj1 E+. Consistently, the MIC values of nalidixic acid, ciprofloxacin, chloramphenicol, cefoxitin and tigecycline did not change in the acrB-deleted KPBj1 E+ mutant. This suggests that these antibiotics, which are well-known substrates of AcrAB in different enterobacterial species,38 – 41 are probably extruded by another efflux pump in K. pneumoniae. Indeed, in KPBj1 E+ we observed high levels of transcription of gene oqxB, as well as of the regulator rarA gene, compared with the wild-type ATCC 13883T strain. Sequencing of the oqxR gene revealed a point mutation leading to the amino acid substitution Val-102Gly in the OqxR protein of strain KPBj1 E+. This substitution is different from those reported by Veleba et al.,12 who described for the first time the rarA and oqxR loci in MDR clinical isolates of K. pneumoniae. Complementation of strain KPBj1 E+ with the wild-type oqxR gene dramatically decreased the expression levels of genes oqxB and rarA, and this result suggests that OqxR is a repressor not only of the oqxAB efflux operon, but also of rarA. In contrast to us, Veleba et al.12 did not observe any change in rarA expression levels in their clinical isolates complemented with wild-type oqxR; however, those isolates could harbour additional genetic alterations responsible for overexpression of rarA. Furthermore, in our study the acrB expression level of isolates overexpressing rarA
5 of 8
Downloaded from http://jac.oxfordjournals.org/ at University of California, San Diego on September 15, 2014
Figure 1. The rarA-oqxABR locus in some K. pneumoniae genomes. Schematic representation of the region encompassing the rarA-oqxABR locus in one K. pneumoniae public genome (strain KCTC 2242)32 compared with the two genomes sequenced in the present study. The large arrows indicate the locations and orientations of the different coding sequences, with annotations displayed below or above the arrows. The black vertical bar represents the limits of contigs in the genomic sequence of strain KPBj1 Rev. The BLAST identity scale is shown in the lower right corner, with the matches ranging from 100% (darkest) to 82% (lightest). 1, abortive infection phage resistance protein; 2, hypothetical protein; 3, phage polarity suppression protein; 4, phage late gene regulator; 5, phage capsid and scaffold protein; 6, hypothetical protein; 7, phage transcriptional regulator; 8, putative phage immunity repressor protein; 9, hypothetical protein; 10, hypothetical phage protein; 11, bacteriophage P4 DNA primase.
Bialek-Davenet et al.
9.0 8.0 7.0
** *
* *
5.0 4.0 3.0 2.0 1.0 0.0 OP50
ATCC 13883T
KPBj1 E+
KPBj1 Rev
KPBj1 M3 Lev KPBj1 E+ DacrB
Figure 2. Kinetics of killing of C. elegans infected by the studied K. pneumoniae strains. LT50, corresponding to time (days) required to kill 50% of the nematode population, is represented for each of the studied strains as the mean+SD of five experiments. E. coli strain OP50 was used as a control. Statistical significance of the differences observed is indicated as follows: *P,0.001; **P¼0.007.
was not elevated, which is also different from the findings of Veleba et al.12 and suggests that the role of RarA as a regulator of acrB is unclear. In order to definitely demonstrate that the point mutation found in the oqxR gene of strain KPBj1 E+ is sufficient to explain its MDR phenotype, an allelic replacement experiment was performed in strain LM21, a wild-type K. pneumoniae clinical isolate. As expected, substitution of the native oqxR allele by the mutated one derepressed the expression of genes oqxB and rarA in strain LM21 RoqxR-KPBj1 E+. With regard to antibiotic susceptibility, transformation of KPBj1 E+ with pSC-A-amp/kan-oqxR-ATCC decreased the MICs of nalidixic acid, ciprofloxacin, chloramphenicol and cefoxitin at least 4-fold, but did not significantly affect the MIC of tigecycline. By using isogenic strains in which the rarA-oqxABR locus is absent and the AcrAB pump is either normally expressed or overexpressed, we showed that the MICs of all of the antibiotics tested were increased in the variant overproducing AcrAB. All these results definitively indicate that nalidixic acid, ciprofloxacin, chloramphenicol and cefoxitin are substrates of both the OqxAB efflux pump and the AcrAB efflux pump, while tigecycline is a substrate of only the AcrAB efflux pump.42 Analysis of the genomic sequences of strains KPBj1 E+ and KPBj1 Rev, which is a phenotypic revertant of the former strain that had spontaneously recovered susceptibility to several antibiotics, yielded some interesting observations. First, it confirmed the clonal relatedness of these two strains, which has previously been shown by RAPD (random amplified polymorphic DNA) analysis.19 Second, the genome of KPBj1 Rev was shown to harbour a deletion of nearly 17.4 kb, encompassing genes oqxA, oqxB and rarA and two-thirds of gene oqxR, as well as a large cluster of genes lying downstream of rarA and comprising a number of putative phage P4-related proteins, when compared with the
6 of 8
corresponding region in the genome of strain KPBj1 E+. Among the nine complete K. pneumoniae genomes publicly available at the time period of our study, only one contained a putative phage P4-related gene cluster inserted in the vicinity of the rarA-oqxABR locus, whereas three others harboured a different phage or gene cluster. Gene ssrA, situated in this genomic region just downstream of smpB and encoding a tmRNA, is a known hot spot of integration of mobile genetic elements in Enterobacteriaceae.43 We suppose that the phage has been excised from the genome of KPBj1 E+, accidentally removing the neighbouring genes and thus leading to strain KPBj1 Rev, previously considered a spontaneous revertant of KPBj1 E+ because of its loss of the cross-resistance phenotype. In our previous study of K. pneumoniae virulence assessed in the C. elegans model, we surmised that the difference in virulence level between strains KPBj1 E+ and KPBj1 Rev was due to a slight variation in the expression of gene acrB.19 However, since we have shown in the present study that KPBj1 E+ in fact does not overexpress acrB, our current hypothesis is that the high virulence of this strain is related to the overexpression of genes oqxB and/or rarA. With regard to the role of AcrAB in K. pneumoniae virulence, we observed that the acrB deletion mutant was significantly less virulent than its parental strain, KPBj1 E+. Our results are consistent with a previous study by Padilla et al.,11 which also demonstrated that the inactivation of gene acrB attenuated the virulence potential of the highly pathogenic K. pneumoniae strain 52145. However, quite unexpectedly, the virulence level of KPBj1 M3 Lev, an acrB-overexpressing in vitro-selected mutant of KPBj1 Rev, was lower than that of its parental strain. From this we conclude that acrB overexpression does not have by itself a positive effect on virulence, although the low virulence of strain KPBj1 M3 Lev might also be due to the fact that this strain carries a
Downloaded from http://jac.oxfordjournals.org/ at University of California, San Diego on September 15, 2014
LT50 (days)
6.0
Role of efflux pumps in Klebsiella pneumoniae resistance and virulence
deletion of the entire ramR gene.37 In order to definitely address these questions, further studies are needed and are currently in progress. To conclude, we found that overexpression of the OqxAB efflux pump, resulting from a point mutation in the repressor oqxR gene, was responsible for the MDR phenotype and for the increased virulence potential of one K. pneumoniae clinical isolate. As this pump is widely distributed in K. pneumoniae genomes,12,15 – 18 but not necessarily overexpressed, it would be interesting to determine the extent to which it participates in clinically relevant antimicrobial resistance and in pathogenicity of nosocomial and community-acquired isolates.
We thank C. Forestier for kindly providing K. pneumoniae strain LM21.
Funding The research leading to these results has received support from the Innovative Medicines Joint Undertaking under Grant Agreement No. 115525, resources which are composed of financial contributions from the European Union’s seventh framework programme (FP7/2007-2013) and European Federation of Pharmaceutical Industries and Associations companies in kind contribution.
Transparency declarations None to declare.
Supplementary data Table S1 is available as Supplementary data at JAC Online (http://jac. oxfordjournals.org/).
References 1 Podschun R, Ullmann U. Klebsiella spp. as nosocomial pathogens: epidemiology, taxonomy, typing methods, and pathogenicity factors. Clin Microbiol Rev 1998; 11: 589– 603. 2 Keynan Y, Rubinstein E. The changing face of Klebsiella pneumoniae infections in the community. Int J Antimicrob Agents 2007; 30: 385– 9. 3 Decre´ D, Verdet C, Emirian A et al. Emerging severe and fatal infections due to Klebsiella pneumoniae in two university hospitals in France. J Clin Microbiol 2011; 49: 3012– 4. 4 Pages JM, Lavigne JP, Leflon-Guibout V et al. Efflux pump, the masked side of b-lactam resistance in Klebsiella pneumoniae clinical isolates. PLoS One 2009; 4: e4817. 5 Okusu H, Ma D, Nikaido H. AcrAB efflux pump plays a major role in the antibiotic resistance phenotype of Escherichia coli multiple-antibioticresistance (Mar) mutants. J Bacteriol 1996; 178: 306– 8. 6 Piddock LJ, White DG, Gensberg K et al. Evidence for an efflux pump mediating multiple antibiotic resistance in Salmonella enterica serovar Typhimurium. Antimicrob Agents Chemother 2000; 44: 3118– 21.
9 Perez A, Canle D, Latasa C et al. Cloning, nucleotide sequencing, and analysis of the AcrAB-TolC efflux pump of Enterobacter cloacae and determination of its involvement in antibiotic resistance in a clinical isolate. Antimicrob Agents Chemother 2007; 51: 3247– 53. 10 Sanchez-Cespedes J, Vila J. Partial characterisation of the acrAB locus in two Citrobacter freundii clinical isolates. Int J Antimicrob Agents 2007; 30: 259–63. 11 Padilla E, Llobet E, Domenech-Sanchez A et al. Klebsiella pneumoniae AcrAB efflux pump contributes to antimicrobial resistance and virulence. Antimicrob Agents Chemother 2010; 54: 177–83. 12 Veleba M, Higgins PG, Gonzalez G et al. Characterization of RarA, a novel AraC family multidrug resistance regulator in Klebsiella pneumoniae. Antimicrob Agents Chemother 2012; 56: 4450– 8. 13 Hansen LH, Johannesen E, Burmolle M et al. Plasmid-encoded multidrug efflux pump conferring resistance to olaquindox in Escherichia coli. Antimicrob Agents Chemother 2004; 48: 3332– 7. 14 Hansen LH, Jensen LB, Sorensen HI et al. Substrate specificity of the OqxAB multidrug resistance pump in Escherichia coli and selected enteric bacteria. J Antimicrob Chemother 2007; 60: 145–7. 15 Park KS, Kim MH, Park TS et al. Prevalence of the plasmid-mediated quinolone resistance genes, aac(6 ′ )-Ib-cr, qepA, and oqxAB in clinical isolates of extended-spectrum b-lactamase (ESBL)-producing Escherichia coli and Klebsiella pneumoniae in Korea. Ann Clin Lab Sci 2012; 42: 191– 7. 16 Yuan J, Xu X, Guo Q et al. Prevalence of the oqxAB gene complex in Klebsiella pneumoniae and Escherichia coli clinical isolates. J Antimicrob Chemother 2012; 67: 1655– 9. 17 Rodriguez-Martinez JM, Diaz de Alba P, Briales A et al. Contribution of OqxAB efflux pumps to quinolone resistance in extended-spectrumb-lactamase-producing Klebsiella pneumoniae. J Antimicrob Chemother 2013; 68: 68– 73. 18 Perez F, Rudin SD, Marshall SH et al. OqxAB, a quinolone and olaquindox efflux pump, is widely distributed among multidrug-resistant Klebsiella pneumoniae isolates of human origin. Antimicrob Agents Chemother 2013; 57: 4602– 3. 19 Bialek S, Lavigne JP, Chevalier J et al. Membrane efflux and influx modulate both multidrug resistance and virulence of Klebsiella pneumoniae in a Caenorhabditis elegans model. Antimicrob Agents Chemother 2010; 54: 4373– 8. 20 EUCAST. Determination of minimum inhibitory concentration (MICs) of antibacterial agents by broth dilution. Clin Microbiol Infect 2003; 9: 1 –7. 21 Bialek-Davenet S, Marcon E, Leflon-Guibout V et al. In vitro selection of ramR and soxR mutants overexpressing efflux systems by fluoroquinolones as well as cefoxitin in Klebsiella pneumoniae. Antimicrob Agents Chemother 2011; 55: 2795– 802. 22 Datsenko KA, Wanner BL. One-step inactivation of chromosomal genes in Escherichia coli K-12 using PCR products. Proc Natl Acad Sci USA 2000; 97: 6640– 5. 23 Balestrino D, Haagensen JA, Rich C et al. Characterization of type 2 quorum sensing in Klebsiella pneumoniae and relationship with biofilm formation. J Bacteriol 2005; 187: 2870 –80. 24 Criscuolo A, Brisse S. AlienTrimmer: a tool to quickly and accurately trim off multiple short contaminant sequences from high-throughput sequencing reads. Genomics 2013; 102: 500– 6. 25 Rissman AI, Mau B, Biehl BS et al. Reordering contigs of draft genomes using the Mauve aligner. Bioinformatics 2009; 25: 2071– 3.
7 Mazzariol A, Zuliani J, Cornaglia G et al. AcrAB efflux system: expression and contribution to fluoroquinolone resistance in Klebsiella spp. Antimicrob Agents Chemother 2002; 46: 3984 –6.
26 Jolley KA, Maiden MC. BIGSdb: scalable analysis of bacterial genome variation at the population level. BMC Bioinformatics 2010; 11: 595.
8 Pradel E, Pages JM. The AcrAB-TolC efflux pump contributes to multidrug resistance in the nosocomial pathogen Enterobacter aerogenes. Antimicrob Agents Chemother 2002; 46: 2640– 3.
27 Diancourt L, Passet V, Verhoef J et al. Multilocus sequence typing of Klebsiella pneumoniae nosocomial isolates. J Clin Microbiol 2005; 43: 4178– 82.
7 of 8
Downloaded from http://jac.oxfordjournals.org/ at University of California, San Diego on September 15, 2014
Acknowledgements
JAC
Bialek-Davenet et al.
28 Aziz RK, Bartels D, Best AA et al. The RAST server: rapid annotations using subsystems technology. BMC Genomics 2008; 9: 75. 29 Li H, Durbin R. Fast and accurate short read alignment with Burrows-Wheeler transform. Bioinformatics 2009; 25: 1754 –60. 30 Li H, Handsaker B, Wysoker A et al. The sequence alignment/Map format and SAMtools. Bioinformatics 2009; 25: 2078–9. 31 Carver T, Harris SR, Berriman M et al. Artemis: an integrated platform for visualization and analysis of high-throughput sequence-based experimental data. Bioinformatics 2012; 28: 464–9. 32 Shin SH, Kim S, Kim JY et al. Complete genome sequence of the 2,3-butanediol-producing Klebsiella pneumoniae strain KCTC 2242. J Bacteriol 2012; 194: 2736– 7.
34 Garsin DA, Sifri CD, Mylonakis E et al. A simple model host for identifying Gram-positive virulence factors. Proc Natl Acad Sci USA 2001; 98: 10892– 7. 35 Favre-Bonte S, Joly B, Forestier C. Consequences of reduction of Klebsiella pneumoniae capsule expression on interactions of this bacterium with epithelial cells. Infect Immun 1999; 67: 554–61. 36 Liu P, Li P, Jiang X et al. Complete genome sequence of Klebsiella pneumoniae subsp. pneumoniae HS11286, a multidrug-resistant strain isolated from human sputum. J Bacteriol 2012; 194: 1841– 2.
8 of 8
38 Visalli MA, Murphy E, Projan SJ et al. AcrAB multidrug efflux pump is associated with reduced levels of susceptibility to tigecycline (GAR-936) in Proteus mirabilis. Antimicrob Agents Chemother 2003; 47: 665–9. 39 Ruzin A, Visalli MA, Keeney D et al. Influence of transcriptional activator RamA on expression of multidrug efflux pump AcrAB and tigecycline susceptibility in Klebsiella pneumoniae. Antimicrob Agents Chemother 2005; 49: 1017– 22. 40 Piddock LJ. Clinically relevant chromosomally encoded multidrug resistance efflux pumps in bacteria. Clin Microbiol Rev 2006; 19: 382–402. 41 Nikaido H, Takatsuka Y. Mechanisms of RND multidrug efflux pumps. Biochim Biophys Acta 2009; 1794: 769– 81. 42 Veleba M, Schneiders T. Tigecycline resistance can occur independently of the ramA gene in Klebsiella pneumoniae. Antimicrob Agents Chemother 2012; 56: 4466– 7. 43 Williams KP. Traffic at the tmRNA gene. J Bacteriol 2003; 185: 1059– 70. 44 Skerman VBD, Mc Gowan V, Sneath PHA. Approved lists of bacterial names. Int J Syst Bacteriol 1980; 30: 225– 420.
Downloaded from http://jac.oxfordjournals.org/ at University of California, San Diego on September 15, 2014
33 Sullivan MJ, Petty NK, Beatson SA. Easyfig: a genome comparison visualizer. Bioinformatics 2011; 27: 1009 –10.
37 Bialek-Davenet S, Leflon-Guibout V, Tran Minh O et al. Complete deletion of the ramR gene in an in vitro-selected mutant of Klebsiella pneumoniae overexpressing the AcrAB efflux pump. Antimicrob Agents Chemother 2013; 57: 672–3.
