OUP user menu

Relative contribution of target gene mutation and efflux to varying quinolone resistance in Irish Campylobacter isolates

Deborah Corcoran, Teresa Quinn, Leslie Cotter, Séamus Fanning
DOI: http://dx.doi.org/10.1016/j.femsle.2005.09.019 39-46 First published online: 1 December 2005

Abstract

The contribution of target gene mutations and active efflux to varying levels of quinolone resistance in Irish Campylobacter isolates was studied. The Thr-86-Ile modification of GyrA did not correlate with the level of quinolone resistance. The efflux pump inhibitor Phe-Arg-β-Naphthylamide (PAβN) had no effect on the MICs to ciprofloxacin. In contrast, a PAβN sensitive efflux system contributed to the low-level nalixidic acid resistance phenotype. The lack of effect of PAβN in high-level resistant nalidixic isolates may be attributable to mutations identified in the CmeB efflux pump of these isolates. PAβN may have limited diagnostic value in the assessment of the contribution of efflux pump activity to ciprofloxacin resistance in Campylobacter.

Keywords
  • Campylobacter
  • gyrA mutation
  • Efflux pumps
  • Efflux pump inhibitor Phe-Arg-β-Naphthylamide

1 Introduction

Campylobacter jejuni and Campylobacter coli are considered to be major causative agents of acute diarrheal disease in humans in the world. Macrolides and fluoroquinolones are the antimicrobial agents used in the treatment of severe infections [1]. Since the early 1990s, there has been a significant increase in the prevalence of quinolone resistance amongst Campylobacte r isolated from broiler sources [2]. This is now recognised as a major health problem in many European countries, including Ireland [3,4].

Resistance to quinolones in Campylobacter has mainly been attributed to a mutation in the gyrA-encoding subunit of the DNA gyrase, resulting in a Thr-86-Ile amino acid substitution [3,5] although other amino acid substitutions in the quinolone resistance determining region (QRDR) of the gyrA have also been documented [4,6]. Varying levels of quinolone resistance have been associated with the Thr-86-Ile mutation [5,710] suggesting that other factors may contribute to the resistance phenotype. The resistance nodulation cell division (RND) efflux pump CmeABC, has been reported to contribute to the intrinsic [1113] and acquired [8,14,15] resistance of C. jejuni and C. coli to various antimicrobial agents including quinolones. However, the interplay between gyrA mutations and efflux in the genesis of quinolone resistance in Campylobacter still remains unclear.

In this study, we investigated the contribution of both target gene mutations and efflux pump activity to quinolone resistance in Irish Campylobacter isolates.

2 Materials and methods

2.1 Bacterial strains and antibiotic susceptibility testing

Three isolates C. coli CIT-405, C. jejuni CIT-423 C. jejuni CIT-424, that were obtained from a poultry surveillance programme, and one human clinical isolate C. jejuni CIT-428, isolated in the James Connolly Memorial Hospital, Dublin, Ireland were used in this study. These were chosen on the basis that they demonstrated varying levels of quinolone resistance. Two reference strains NCTC 11168 (C. jejuni, human isolate) and NCTC 11366 (C. coli, porcine isolate) were also included. All isolates were cultivated under microaerophilic conditions on Mueller–Hinton (MH) agar plates (Difco, Becton Dickinson, France) supplemented with 5% (v/v) sterile horse blood (TCS Biosciences, UK). Antimicrobial susceptibility testing was carried out by E-test (AB BIODISK, Sweden). The following breakpoints were used: ciprofloxacin > 2 μg/ml and nalidixic acid > 16 μg/ml.

2.2 Chromosomal DNA template isolation

All isolates were sub-cultured onto Preston agar consisting of a Campylobacter agar base (Oxoid Ltd., Basingstoke, UK) containing 5% (v/v) lysed horse blood and Campylobacter modified selective supplement (Oxoid) and cultured microaerophilically for 48 h. Chromosomal DNA was purified by initially suspending the culture in 1 ml 0.85% (w/v) NaCl and washing twice in this solution. It was then treated with formaldehyde to inhibit DNase activity [16]. DNA was extracted as previously described [17]. DNA concentration was measured spectrophotometrically at 260 and 280 nm. The integrity of the isolated DNA was then assessed by conventional agarose gel electrophoresis in a 1.5% (w/v) agarose gel.

2.3 PCR amplification and sequencing of the QRDR region of the gyrA gene and the cmeRABC tripartite efflux pump

A 673 bp fragment of the QRDR was amplified and sequenced from all isolates using the primers GZgyrA5 and GZgyrA6 (Table 1) as previously described [18]. Fragments of the cmeA, cmeB and cmeC genes were amplified using the primers outlined in Table 1. The cmeB amplicon was also sequenced using the same primers. The complete transcriptional regulatory gene cmeR (633 bp) was amplified within a 973 bp fragment and sequenced using the novel primers CmeRF and CmeRR (Table 1). PCR reactions were carried out in a final volume of 100 μl containing 200 ng of genomic DNA, 10 μl of 10× PCR buffer (100 mM Tris–HCl [pH 9.0], 500 mM KCl, 1% (v/v) Triton X-100), 2.5 mM MgCl2, 0.2 mM each dNTP, 2.5 U Taq DNA polymerase (Promega, Madison, WI), and 25 pmol each of the forward and reverse primers (MWG-Biotech, Germany). After an initial denaturation of 5 min at 95 °C, amplification was performed over 30 cycles with each consisting of 95 °C for 1 min, relevant annealing temperature (Table 1) for 1 min and 72 °C for 1 min with a final extension of 7 min at 72 °C. Amplification was carried out in a DNA Thermal Cycler (Biosciences). PCR amplified products were resolved by electrophoresis in a 1.5% (w/v) agarose gel in 1× Tris-Borate-EDTA buffer containing 0.5 μg/ml ethidium bromide, and imaged using a Gel Doc 2000 (Bio-Rad, Hercules, CA). PCR products were purified using a QIAquick PCR purification kit (Qiagen, GmbH, Germany) and sequenced commercially by MWG Biotech (Ebersberg, Germany).

View this table:
Table 1

Primers used for PCR and sequencing

Name and sequenceAnnealing temperature (°C)Amplicon size (bp)Reference
QRDR
GZgyrA5, 5′-ATT TTT AGC AAA GAT TCT GAT-3′ [18]
GZgyrA6, 5′-CCA TAA ATT ATT CCA CCT GT-3′50673 [18]
cmeRABC operon
CmeRF, 5′-GCA GGA GAA CAA GTC AAA-3′This study
CmeRR, 5′-GCT GCA AGC AGT GAG TAA-3′51973This study
CmeAF, 5′-TTG ATG GCT AAG GCA ACT TTC-3′ [11]
CmeAR, 5′-CTC CAA TTT CTT TAA CTT CGC-3′53771 [11] a
CmeBF, 5′-GAC GTA ATG AAG GAG AGC CA-3′ [12]
CmeBR, 5′-CTG ATC CAC TCC AAG CTA TG-3′521070 [12]
CmeCF, 5′-GCT TGG AGC CTT ATC TTG GGA-3′ [11] a
CmeCR, 5′-TGG CTC TTG CTT GAG CAA GTT-3′57571This study
  • aPrimer sequence modified.

2.4 DNA sequence analysis and protein secondary structure prediction

The BLAST suite of programmes was used to compare DNA sequences to GenBank sequence databases. Nucleotide and amino acid alignments were subsequently generated by CLUSTALW http://www.ebi.qc.uk/clustalw. Protein secondary structure prediction was analysed using programs such as NNPREDICT and SOUSHI available on line at http://www.hgmp.mrc.ac.uk/GenomeWeb/prot-2-struct.html.

2.5 Accumulation assays

Accumulation of ethidium bromide (EtdBr) was measured by fluorescent spectrophotometry as previously described [19,20]. Cells were pelleted and washed twice with 20 mM Hepes buffer (pH 7.0). The cells were then re-suspended in same buffer to give a final absorbance of 0.3 at 600 nm. EtdBr was added to a final concentration of 2 μg/ml, to 200 μl aliquots of the bacterial suspension. Fluorescence of the cell suspension was used as an indicator of EtdBr accumulation (fluorescence of EtdBr increases upon binding with DNA) and was recorded at excitation and emission wavelengths of 530 and 600 nm, respectively. Seven minutes after EtdBr addition (when steady-state accumulation was observed) efflux pump inhibitors verapamil (100 μg/ml) [20] carbonyl cyanide m-chlorophenyl hydrazone (CCCP, 21 μg/ml) [21] and l-phenylalanine-l-arginine-β-naphthylamide (PAβN, 20 μg/ml) were added in individual reactions and fluorescence recorded for a further 13 min. The concentration of PAβN used throughout this study was previously determined to have the greatest effect on the MICs to quinolones in the reference strains (data not shown).

2.6 Effect of efflux pump inhibition on MICs

To examine the contribution of efflux pump activity to the quinolone resistance phenotypes, the MICs of ciprofloxacin and nalidixic acid were determined using E-test strips, in the presence and absence of the efflux pump inhibitor (PAβN, 20 μg/ml). This experiment was repeated on three separate occasions. Concentrations of PAβN up to 64 μg/ml had no visible affect on bacterial growth.

2.7 Nucleotide sequence accession numbers

The nucleotide sequence data for the cmeR regulatory gene were submitted to GenBank and assigned the Accession Nos. AY594250 (C. coli CIT-405), AY714320 (C. jejuni CIT-423), AY714321 (C. jejuni CIT-424) and AY714323 (C. jejuni CIT-428).

3 Results

3.1 Antimicrobial susceptibility

Campylobacter coli CIT-405 exhibited low-level resistance to both ciprofloxacin and nalidixic acid (ciprofloxacin MIC=4.0 μg/ml, nalidixic acid MIC=32.0 μg/ml). C. jejuni CIT-424 exhibited intermediate-level of resistance to ciprofloxacin (MIC=12.0 μg/ml) and low-level resistance to nalidixic acid (MIC=32.0 μg/ml). C. jejuni CIT-423 and C. jejuni CIT-428 exhibited high-level resistance to both antimicrobial agents (ciprofloxacin MIC=32.0 μg/ml, nalidixic acid MIC=256.0 μg/ml). The reference strains were susceptible to both antimicrobials (Table 2).

View this table:
Table 2

Effect of PAβN on susceptibility to quinolones

StrainsMIC (μg/ml)GyrA mutation/s
NumberSourceCiprofloxacinNalidixic acid
−PAβN+PAβN−PAβN+PAβN
Study strains
C. coli CIT-405Poultry4.03.032.08.0Thr-86-Ile
C. jejuni CIT-423Poultry32.032.0256.0256.0Thr-86-Ile, Asn-203-Ser
C. jejuni CIT-424Poultry12.08.032.04.0
C. jejuni CIT-428Human32.032.0256.0256.0Thr-86-Ile, Asn-203-Ser
Reference strains
C. jejuni NCTC 11168Human0.0640.0322.00.75
C. coli NCTC 11366Porcine0.0940.0474.01.0

3.2 Analysis of gyrA gene mutations

Sequence analysis of the QRDR of gyrA revealed the Thr-86-Ile amino acid substitution in C. coli CIT-405, C. jejuni CIT-423 and C. jejuni CIT-428. An additional amino acid substitution was identified in C. jejuni CIT-423 and C. jejuni CIT-428 at the 203-codon (Asn-203-Ser). No amino acid substitutions were identified in C. jejuni CIT-424, or in either of the reference strains.

3.3 Efflux pump sequence analysis

Fragments of the cmeA, cmeB and cmeC genes were amplified. The 1070 bp fragment of the cmeB gene was sequenced and a BLAST search indicated that the deduced amino acid sequence of the cmeB fragments were similar to transmembrane proteins of other RND efflux systems (data not shown). An amino acid alignment (deduced from the 1070 bp DNA fragment) of the CmeB protein from the four isolates, compared to the reference strain NCTC C. jejuni 11168, identified seven amino acid substitutions in C. jejuni CIT-423 and C. jejuni CIT-428 (Fig. 1). All of the latter substitutions were identical. NNpredict identified a helix-turn-helix DNA binding motif in the CmeR protein of all four isolates, characteristic of transcriptional regulatory proteins (Fig. 2). Amino acid substitutions were noted in the helix-turn-helix region of CmeR in both C. coli CIT-405 and C. jejuni CIT-424 when compared to C. jejuni NCTC 11168 (Fig. 2).

Figure 1

CLUSTALW alignment of a fragment of the deduced CmeB amino acid sequence form C. jejuni NCTC 11168, C. coli CIT-405, C. jejuni CIT-423, C. jejuni CIT-424 and C. jejuni CIT-428. Arrows identify the positions of the seven amino acid substitutions, which are indicated in bold.

Figure 2

CLUSTALW alignment of the deduced CmeR amino acid sequence form C. jejuni NCTC 11168 (F81379), C. coli CIT-405 (AAT06777), C. jejuni CIT-423 (AAU43760), C. jejuni CIT-424 (AAU43761) and C. jejuni CIT-428 (AAU43763). Amino acid substitutions are indicated in bold. The position of the helix-turn-helix DNA binding motif of the cmeR genes are identified by the broken line flanked by the diamond shapes.

3.4 Accumulation studies

Addition of EtdBr to bacterial cell suspensions resulted in a steady state accumulation of the dye after 7 min. The subsequent addition of three known EPIs resulted in a marked increase in accumulation (Fig. 3). Active efflux was also demonstrated in the quinolone susceptible reference strains C. jejuni NCTC 11168 and C. coli NCTC 11366 (data not shown).

Figure 3

Accumulation of the EtdBr dye in C. coli CIT-405, C. jejuni CIT-423, CIT-424 and CIT-428. EtdBr (2 μg/ml) was added to the bacterial cell suspension at time zero. Three efflux pump inhibitors: verapamil (100 μg/ml), carbonyl cyanide m-chlorophenyl hydrazone (CCCP, 21 μg/ml) and PAβN (20 μg/ml) were added at 7 min, as indicated by the arrow in the figure. Fluorescence units (y-axis) represent the accumulation of EtdBr. The experiment was replicated four times and each data point represents a mean value of all four experiments.

3.5 Effectiveness of the efflux pump inhibitor PAβN

PAβN had no effect on the MICs to ciprofloxacin in any of the isolates studied. PAβN caused four- and eightfold decreases in the MICs to nalidixic acid in C. coli CIT-405 (MIC 32.0 reduced to 8.0 μg/ml) and C. jejuni CIT-424 (MIC 32.0 reduced to 4.0 μg/ml). It had no effect on the MICs to nalidixic acid in the high-level quinolone resistant strains C. jejuni CIT-423 and C. jejuni CIT-428. The MIC values to ciprofloxacin and nalidixic acid in both the reference strains C. jejuni NCTC 11168 and C. coli NCTC 11366, were decreased two- and fourfold, respectively (Table 2).

4 Discussion

To date in Campylobacter varying levels of quinolone resistance (low to high) have been associated with the Thr-86-Ile mutation in the gyrA in the absence of any other mutations in the QRDR region of gyrA, gyrB and parC [5,14,22]. This suggests that quinolone resistance may result from an inter-play between mutations and other antibiotic resistance determining mechanisms such as reduced accumulation of the antibiotic as reported in other bacteria [23].

In this study, the Thr-86-Ile amino acid substitution in the QRDR of gyrA was identified in the three isolates C. coli CIT-405, C. jejuni CIT-423 and C. jejuni CIT-428, which displayed low (CIT-405) and high-level (CIT-423 and CIT-428) quinolone resistance. C. jejuni CIT-423 and C. jejuni CIT-428 harboured an additional amino acid substitution at the 203-codon (Asn-203-Ser). This substitution is not associated with ciprofloxacin resistance as it has been identified in both susceptible and resistant C. jejuni isolates [6]. However, a possible contribution of this substitution to high-level nalidixic acid resistance cannot be out-ruled. No amino acid substitutions were identified in the QRDR of C. jejuni CIT-424, which displayed intermediate ciprofloxacin and low-level nalidixic acid resistance. Similarly, a number of authors have documented the occurrence of quinolone resistance in Campylobacter in the absence of gyrA mutations [5,24,25].

Based on bioinformatic data, nine putative efflux pumps in addition to the CmeABC system have been identified in C. jejuni and C. coli [15]. Of these the CmeDEF system has been well characterised [13]. Insertional mutagenesis of these efflux pumps had no significant effect on the MICs to ciprofloxacin in ciprofloxacin-resistant isolates [15]. In contrast, insertional mutagenesis of the CmeB pump resulted in a marked increase in susceptibility to quinolones in quinolone-resistant isolates [14,15] suggesting that the CmeABC efflux system is the main contributor to quinolone resistance in Campylobacter.

In this study, we identified the presence of the tripartite cmeABC operon in all isolates by amplifying fragments of its three component genes. Furthermore, the functional operation of an RND type efflux pump was verified using an EtdBr efflux assay in the presence of known efflux pump inhibitors. The EPI PAβN, has been reported to restore the efficacy of quinolones in Gram-negative bacteria such as Pseudomonas aeruginosa, Salmonella enterica Typhimurium DT204 and Escherichia coli [2630]. However, in this study, PAβN had no effect on the MICs to ciprofloxacin in ciprofloxacin-resistant isolates or on the MICs to nalidixic acid in isolates showing high-level nalidixic acid resistance. In contrast, inhibition of efflux pump activity restored susceptibility in the two low-level nalidixic acid resistant isolates, C. jejuni CIT-424 and C. coli CIT-405 suggesting that efflux pump activity may contribute to nalidixic acid resistance in these isolates.

The expression of multidrug efflux pump systems is usually controlled by transcriptional regulators. Recently, CmeR has been shown by insertional mutagenesis to function as a transcriptional repressor for CmeABC [31]. In this study, we amplified and sequenced cmeR in all isolates. Mutations were identified in the substrate-binding region of CmeR of C. jejuni CIT-424 and C. coli CIT-405. Mutations within this region have been shown to result in a derepression of the cmeABC operon resulting in an over expression of cmeB [24]. This further supports our hypothesis that efflux pump activity may contribute to the low-level nalidixic acid resistance in these strains.

The effects of PAβN on quinolone resistance in C. jejuni and C. coli are controversial. Payot et al. [8] reported that the efflux pump inhibitor PAβN caused a significant decrease of the MIC of enrofloxacin in 27 of the 38 C. coli isolates examined and restored susceptibility in 15 isolates. Enrofloxacin is a fluoroquinolone licensed for use in animals that is catabolized to ciprofloxacin the fluoroquinolone used for human therapy. Furthermore, resistance and susceptibility to enrofloxacin reflects that of ciprofloxacin in Campylobacter [22]. An earlier study by Payot et al. [22] showed negligible effects of PAβN on nalidixic acid, ciprofloxacin and enrofloxacin resistance. Mamelli et al. [32] reported that PAβN had no effect on the degree of susceptibility to ciprofloxacin in Campylobacter. Recently, Ruiz et al. [33] reported that PAβN did not produce a significant decrease in the MIC of ciprofloxacin and the novel and more potent fluoroquinolones, levofloxacin and moxifloxacin. In agreement with this study, PAβN has been shown to have a greater efficacy in lowering the MIC to nalidixic acid than ciprofloxacin in a number of other bacteria [3436]. PAβN has also been shown to inhibit macrolide efflux in erythromycin-sensitive and resistant strains of C. jejuni and C. coli [8,32,37]. Furthermore, it is efficient in inhibiting EtdBr efflux as evidenced by the increase in EtdBr accumulation in this study following addition of PAβN to bacterial cell suspensions. Thus, it would appear that PAβN is more efficient in decreasing macrolide, EtdBr and nalidixic acid efflux than fluoroquinolones. This suggests that different antibiotics may have different binding sites on the CmeB pump as has been described for AcrB [38] and that inhibition by PAβN may be binding site specific.

The influence of PAβN on low-level nalidixic acid resistant isolates together with the apparent lack of effect on isolates showing high-level nalidixic acid resistance in this study prompted us to examine the cmeB gene for the presence of mutations. Mutations in the homologous Bmr pump in Bacillus subtilis have been shown to alter the response of the pump to efflux pump inhibitors including reserpine even though it continued to actively export antibiotics [39]. It is tempting to speculate that the CmeB amino acid substitutions observed in the high-level nalidixic acid resistant isolates C. jejuni CIT-423 and C. jejuni CIT-428 may have selectively prevented the interaction of PAβN with the nalidixic acid binding site thereby decreasing its efficacy in inhibiting nalidixic acid efflux. This requires further investigation.

In conclusion, efflux pump activity is the main contributor to low-level nalidixic acid resistance in these isolates. The lack of effect of PAβN in high-level resistant nalidixic isolates may be attributable to mutations identified in the CmeB efflux pump of these isolates. A possible contribution of the Asn-203-Ser mutation in the QRDR of the gyrA to high-level nalidixic acid resistance cannot be out-ruled. PAβN had no effect on the MIC to ciprofloxacin, therefore the use of PAβN alone for efflux pump recognition may be limited, particularly in relation to fluroquinolone resistance. Consequently, a variety of efflux inhibitors should be used to determine their effect(s) on an isolates efflux phenotype. Studies to date attributing fluoroquinolone resistance to gyrA mutations in Campylobacter on the basis of lack of effect of PAβN should be re-evaluated using an efflux pump inhibitor proven to be active against fluoroquinolones.

Acknowledgements

The authors acknowledge the financial support provided by grants from the Irish Government Food Institutional Research Measure (FIRM)-00/R&D/G/4 and the Food Safety Authority of Ireland (FSAI)-86/FS. We also thank our colleagues at the Centre for Food Safety, University College Dublin and the Molecular Diagnostic Unit, Cork Institute of Technology for technical support.

References

  1. [1].
  2. [2].
  3. [3].
  4. [4].
  5. [5].
  6. [6].
  7. [7].
  8. [8].
  9. [9].
  10. [10].
  11. [11].
  12. [12].
  13. [13].
  14. [14].
  15. [15].
  16. [16].
  17. [17].
  18. [18].
  19. [19].
  20. [20].
  21. [21].
  22. [22].
  23. [23].
  24. [24].
  25. [25].
  26. [26].
  27. [27].
  28. [28].
  29. [29].
  30. [30].
  31. [31].
  32. [32].
  33. [33].
  34. [34].
  35. [35].
  36. [36].
  37. [37].
  38. [38].
  39. [39].
View Abstract