Phenotypic evidence for inducible multiple antimicrobial resistance in Salmonella choleraesuis

Abstract

Multiple antimicrobial resistance (MAR) in Salmonella choleraesuis is becoming a major concern. It has been demonstrated that a MAR phenotype can be induced in Escherichia coli and other members of the Enterobacteriaceae by exposing the isolates to salicylates, various antimicrobials, or organic solvents used to combat and control bacterial infection. Therefore the purpose of the present study was to determine whether this mar A-associated MAR-phenotype is inducible in S. choleraesuis. Isolates used in the present study were able to withstand toxic effects of the organic solvent cyclohexane naturally, or following exposure to the inducing compounds salicylate, tetracycline, or chloramphenicol. All isolates possessed fragments of mar A with the predicted size of 408 bp when amplified using mar A-specific primers by PCR. The resulting PCR products that were sequenced revealed that amplified S. choleraesuis mar A was 99% and 85% homologous to the published Salmonella typhimurium and E. coli mar A sequences respectively. Minimum inhibitory concentrations of tetracycline (P<0.08), chloramphenicol (P<0.001), rifampin (P<0.08), and nalidixic acid (P<0.001) against cyclohexane-tolerant mutants were significantly increased when compared with wild-type S. choleraesuis. Northern hybridization signals for both mar A and acr B were increased in the induced isolates when compared to uninduced controls while sox S expression did not change between induced and uninduced cultures. The results suggest that mar A is present in S. choleraesuis and a MAR-phenotype is inducible in S. choleraesuis presumably due to the overexpression of mar A and acr B and not to the overexpression of sox S.

Salmonella choleraesuis, marA, Multiple antimicrobial resistance

1 Introduction

Salmonella choleraesuis is a host-adapted, facultative, intracellular pathogen of swine [1] and is the cause of one of several bacterial diseases associated with swine respiratory syndrome (SRD). SRD has been estimated to account for annual losses totaling over $210 million in production costs alone in the United States [2]. The disease complex associated with SRD includes, atrophic rhinitis, pleuritis, pleuropneumonia, and pneumonia [2]. The primary approach for treatment and control of SRD and associated bacterial infections is the use of antimicrobial agents; however, it has been shown that, in the case of S. choleraesuis infections leading to SRD, the bacteria are becoming increasingly resistant to multiple antibiotics [2].

Several mechanisms have been identified that contribute to the acquisition and dissemination of antimicrobial resistance by bacteria. Resistance can be acquired through natural mutations or resistance determinants located on R-plasmids, transposons, and integrons [3,4]. More recently, multiple antimicrobial resistance (MAR) has been found to be mediated by the bacterial chromosome [5]. This MAR system, initially identified in Escherichia coli, is attributed to mutations in an area of the genome located at 34 min on the E. coli genomic map, and has been termed the mar RAB operon [5]. The mar RAB operon consists of four known genes; mar C, mar R, mar A, and mar B under the control of two divergent promoters with marC being upstream of promoter one and the remaining genes downstream of promoter two [5,6]. The mar R and mar A regions encode the repressor and activator of the operon, respectively [5]. Mutations in either the operator region or mar R have been shown to lead to constitutive expression of the operon with subsequent increase in mar A expression [5]. A member of the ara C/xyl S family of transcriptional activators, mar A regulates a very large regulon consisting of at least 60 different genes [7]. Many genes within this regulon are associated with antimicrobial resistance including increased efflux (acr B) and decreased influx (omp F) of a variety of antimicrobial agents [5,6,8]. Expression of the operon has been demonstrated to be induced with a variety of compounds, including tetracycline, chloramphenicol, salicylate, and cyclohexane [5,6,8–14] leading to subsequent resistance to these compounds and other compounds used in the control of bacterial infections.

The mar RAB operon and the associated genes have been found to be widespread among members of the Enterobacteriaceae including members of the genus, Salmonella[15–18]. However, a study by Cohen et al. suggested that not all members of this family including S. choleraesuis possessed mar RAB operon, [17] yet mar RAB operon has been shown to be conserved in S. typhimurium with similar functions as seen in E. coli[13]. A review of S. choleraesuis isolates from pigs presented to the bacteriology laboratory at the Animal Disease Diagnostic Laboratory (ADDL), Purdue University indicated that these isolates are resistant to multiple antibiotics including those due to overexpression of mar A. These findings suggest that multiple antibiotic resistance in S. choleraesuis could be related to the mar RAB operon. The purpose of the present study was to determine the presence and similarities of the mar A gene in S. choleraesuis isolates and determine if MAR among S. choleraesuis isolates was inducible.

2 Materials and methods

2.1 Bacterial strains, media, and general reagents

Fifty-three S. choleraesuis isolates were obtained from the bacteriology laboratory at the ADDL. The bacteria were isolated from pigs with pneumonia or acute death presented to the ADDL for necropsy and identified as S. choleraesuis through serotype analysis. E. coli strains AG100 (not expressing mar A) and AG112 (mar-mutant of AG100 expressing mar A) were obtained from Dr. Stuart Levy's laboratory at the Center for Adaptation Genetics and Drug Research, Tufts School of Medicine and used as negative and positive controls, respectively. All wild-type and mar-mutant isolates were grown in Luria–Bertani (LB) broth at 30°C and maintained on LB agar slants at 4°C. Antibiotics (Sigma, St. Louis, MO, USA) used in the gradient plate assays were added to LB broth to the following concentrations: tetracycline, 256 µg ml⁻¹; chloramphenicol, 128 µg ml⁻¹; rifampin, 128 µg ml⁻¹; and nalidixic acid, 128 µg ml⁻¹. Salicylate (Sigma) used in the induction of mar A expression assays was added to LB broth to a final concentration of 5 mM. While, tetracycline and chloramphenicol were added to LB agar to a final concentration of 4 µg ml⁻¹. Paraquat (Sigma) was added to LB broth to a final concentration of 10 mM.

2.2 Selection of potential marA overexpressing mutants

Potential mar A overexpressing mutants were selected using an organic solvent tolerance assay as previously described [12]. In round one all 53 isolates were initially assayed for growth in the presence of cyclohexane. Potential mutants were chosen as those isolates that grew in the presence of cyclohexane. To determine if mar A could be induced, isolates (negative for growth in cyclohexane in the first round) were grown in LB broth supplemented with low concentrations of salicylate, tetracycline, or chloramphenicol. After overnight incubation at 30°C, cultures were diluted to an OD₅₃₀ of 0.2 and organic solvent tolerance assays were repeated. Isolates that obtained cyclohexane resistance following induction were chosen as potential mar A-overexpressing mutants.

2.3 Genetic analysis of marA

Bacterial DNA was extracted from isolated colonies of S. choleraesuis by using standard procedures [19]. Primers used in PCR assays were designed using the published mar RAB sequence of S. typhimurium[13] (GenBank accession number U54468). StmarA-F (AGAGGTATGACGATGTCCAGACG) corresponding to nucleotides 1427–1450 and StmarA-R (CGACTTGGTACCGTTGTGATCAAACA) corresponding to nucleotides 1827–1801 of the published sequence were synthesized (Genosys, The Woodlands, TX, USA) and used as the upstream and downstream primers, respectively. 50-µl PCR reaction mixtures containing 26.5 µl ddH₂O, 5 µl 10×Taq polymerase reaction buffer B (20 mM Tris–HCl (pH 8.0), 100 mM KCl, 0.1 mM EDTA, 1 mM DTT, 50% glycerol, 0.5% Nonidet®, 0.5% Tween®-20, Promega, Madison, WI, USA), 5 µl MgCl₂ (Promega), 1 µl each of forward and reverse primers, 0.5 µl of Taq polymerase (Promega) and 5 µl of bacterial DNA were amplified in a 96-well thermocycler (Tetrad Thermocycler, MJ Research, Waltham, MA, USA) for 1 min at 95°C followed by 25 cycles of 30 s at 94°C, 30 s at 59°C, and 60 s at 72°C followed by a single cycle for 5 min at 72°C. PCR products were run on a 1.5% agarose gel, stained with ethidium bromide and visualized using ultraviolet light (Fisher, Chicago, IL, USA).

A single PCR fragment corresponding to mar A gene was purified using QIAquick PCR purification kit (Qiagen, Valencia, CA, USA) and digested with Ban II restriction enzyme (Life Technologies, Grand Island, NY, USA) according to protocols recommended by the manufacturers. Digested product was analyzed on a 1.5% agarose gel as described above. Purified PCR product from a single reaction was cloned into the pCR-II cloning vector (Invitrogen, Carlsbad, CA, USA) according to manufacturer's recommendations and sequenced (University of California Davis, Davis, CA, USA). Sequence analysis and comparison was carried out using Multiple Alignment Construction and Analysis software developed at the National Center for Biotechnology Information (NCBI, Bethesda, MD, USA).

2.4 Minimum inhibitory concentration (MIC) determination

MIC was determined using a gradient plate method according to the procedures described previously [4]. Bacterial cultures of wild-type and potential mar A-overexpressing mutant isolates grown overnight at 30°C were diluted to an OD₅₃₀= 0.2 and swabbed side-by-side onto the plates and incubated overnight at 30°C.

2.5 Northern hybridization and densitometry

Three isolates from the uninduced, salicylate, and chloramphenicol induced isolates were chosen for northern analysis. Total RNA from exponential-phase cultures of test isolates was prepared by using TRIreagent (MRC, Inc., Cincinnati, OH, USA) according to manufacturer's recommendations. Total RNA was electrophoresed and blotted to nylon membranes as described elsewhere [19]. Probe templates were prepared by PCR using primers specific for mar A, acr B, and sox S and labeled with ³²P using the HighPrime kit (Roche Molecular Biochemicals, Indianapolis, IN, USA) according to manufacturer's recommendations. Membranes were prehybridized for 30 min at 55°C, hybridization was carried out overnight at 55°C, and washes carried out at room temperature three times for 20 min each using the SuperHyb hybridization and prehybridization/wash solutions (MRC, Inc.). Following washes membranes were exposed to X-ray film and developed using standard procedures [19]. Densitometry analysis of hybridization signals was done with ScionImage (NIH) software and is the average fold difference (after induction/before induction) among wild-type and induced mutants for each inducing method and for each gene assayed.

3 Results

3.1 Organic solvent tolerance related to MAR

Eighteen of the 53 S. choleraesuis isolates were able to grow in the presence of the organic solvent cyclohexane prior to exposure to any inducing agents used in this study. Of the 33 isolates, which were initially sensitive to cyclohexane, nine were able to grow in the presence of cyclohexane following induction with salicylate and six were able to grow in the presence of cyclohexane following induction with chloramphenicol. An additional nine were induced by both salicylate and chloramphenicol and a single isolate was inducible with both tetracycline and chloramphenicol (data not shown). Nine of the 33 cyclohexane sensitive isolates were uninducible by salicylate, chloramphenicol, or tetracycline (data not shown).

3.2 Detection and sequence analysis of marA

A mar A gene fragment with the expected size of 408 bp was successfully amplified from the DNA extracted from all 53 S. choleraesuis isolates. The PCR fragment of mar A gene was cut by Ban II restriction enzyme at nucleotide 239 leaving two fragments of the expected size, 239 bp and 169 bp, respectively. Sequence analysis and alignment showed that the nucleotides of the mar A gene fragment of S. choleraesuis differed from that of Salmonella typhimurium mar A (GenBank accession number U54468) by only 1 bp and was 85% homologous to E. coli mar A (GenBank accession number M96235) as previously published.

3.3 MIC

S. choleraesuis isolates that grew in the presence of cyclohexane, without prior induction, showed a significant increase (P<0.001) in average MICs of chloramphenicol and nalidixic acid when compared to the wild-type parent isolates (Table 1). The average MIC of tetracycline and rifampin were higher in cyclohexane tolerant mutant isolates when compared to their respective wild-type isolates; however, the difference was not significant at P<0.001 (Table 1). The 18 salicylate-induced cyclohexane tolerant isolates showed a significant increase (P<0.001) in resistance to chloramphenicol when MIC of wild-type and mutants were compared (Table 1). Also, the salicylate-induced mutants displayed a higher average MIC to tetracycline, rifampicin, and nalidixic acid than the wild-type isolates; however, these differences were not significant at P<0.001 (Table 1). The 16 chloramphenicol-induced cyclohexane tolerant isolates had a significantly increased (P<0.001) average MIC to chloramphenicol and nalidixic acid when compared to the wild-type isolates; however while resistance to tetracycline and rifampin increased it was not significant at P<0.001 (Table 1). The one isolate induced by tetracycline and tolerant to cyclohexane was significantly (P<0.001) more resistant to all antimicrobials tested than the wild-type isolate (data not shown). The average MIC of the eight isolates that were cyclohexane sensitive and were not induced when exposed to salicylate, tetracycline, or chloramphenicol were 112 µg ml⁻¹ for tetracycline, 10 µg ml⁻¹ for chloramphenicol, 22 µg ml⁻¹ for rifampin, and 15 µg ml⁻¹ for nalidixic acid (data not shown). MICs of control strain E. coli AG100 (not expressing mar A) to tetracycline, chloramphenicol, rifampicin, and nalidixic acid, were significantly lower (P<0.001) than those of control strain E. coli AG102 (constitutively expressing mar A) (Table 1).

1

Average MIC (µg ml⁻¹) for uninduced, salicylate, and chloramphenicol induced, cyclohexane tolerant isolates

	Uninduced (n=18)		Fold increase	Salicylate-induced (n=18)		Fold increase	Chlor.-induced (n=16)		Fold increase	AG100	AG102	Fold increase
Before (−)/after (+) induction	−	+		−	+		−	+		NI	NI
Tetracycline	83	114	1.4	119	147	1.2	105	155	1.5	2	33	16.5
Chloramphenicol	3	21	7	2	12	6	3	26	8.7	13	59	4.5
Rifampin	26	42	1.6	23	32	1.4	25	36	1.4	31	39	1.3
Nalidixic acid	9	25	2.8	16	37	2.3	14	115	8.2	7	17	2.4

Average results of three repeated assays.

Includes isolates induced by both sal/chl.

Includes isolates induced by both chl/sal and chl/tet; NI = not induced.

Non-mar A expressing E. coli negative control.

mar A expressing E. coli positive control.

Statistically significant increase between wild-type and mutant isolates at P<0.08.

Statistically significant increase between wild-type and mutant isolates at P<0.001.

3.4 Northern hybridization

The expression levels of mar A and acr B were increased in all three isolates that grew in the presence of cyclohexane prior to induction (Fig. 1). Similar results were seen when the isolates were grown in the presence of salicylate and chloramphenicol with concomitant increase in acr B expression (Fig. 1). Densitometric analysis of Northern hybridization signals indicates a 2.3–8.1-fold increase in mar A and 2.7–10.5-fold increase in acr B signal densities following growth in the presence of cyclohexane or after induction with salicylate or chloramphenicol (Fig. 2). Signal densities for sox S levels did not change for any isolates with the exception of a single isolate induced with salicylate that caused a two-fold (±1.0) increase in sox S signal density (Fig. 2).

1

Northern hybridization analysis of RNA transcripts obtained from cyclohexane tolerant isolates prior to induction (A) or following induction with salicylate (B) or chloramphenicol (C). Three isolates from each group as well as AG100 (−) and AG112 (+) controls were selected for analysis. Signals obtained using a mar A, acr B, or sox S probes are indicated.

2

Average fold increase in hybridization signals using densitometry analysis. Bars represent the average hybridization signal densities obtained from cyclohexane tolerant isolates prior to induction (A) or following induction with salicylate (B) or chloramphenicol (C) when probed with mar A, acr B, or sox S. Error bars indicate the standard deviation calculated from the densities of the three hybridization signals as seen in Fig. 2 for each inducing method.

4 Discussion

The association of inducible resistance to multiple antimicrobial agents and organic solvents with overexpression of mar A has been well documented in members of the Enterobacteriaceae. However, a report by Cohen et al. [17] indicated that while mar A was associated with inducible MAR in S. typhimurium, it was not found to be so in the closely related S. choleraesuis serotype. This result led us to question whether a MAR phenotype was inducible in S. choleraesuis and if so could it be associated with mar A.

As opposed to work done by Cohen et al., polymerase chain reaction, restriction digestion, and sequence alignment indicates that the mar A gene is present in S. choleraesuis and is genetically similar to previously published mar A sequences [5,13]. A single base pair substitution was seen when S. choleraesuis mar A was compared to S. typhimurium mar A; however, the deduced amino acid sequence between them did not change, suggesting that S. choleraesuis mar A may function the same as in S. typhimurium.

One hallmark of mar A-associated MAR is inducibility with unrelated antimicrobials and organic solvents [6]. This induction ultimately leads to increased resistance to these same antimicrobials and solvents. Based on this phenomenon the ability of an organism to grow in the presence of cyclohexane, which is normally toxic, can be used to screen for potential mar A-overexpressing mutants. We were able to determine that some of the isolates used in this research could withstand the killing effects of cyclohexane without any prior induction of mar A. Further, we were able to show that cyclohexane-sensitive isolates could be made cyclohexane-resistant by growing them in the presence of known mar A-inducing agents, salicylate, chloramphenicol, or tetracycline. These results are similar to previously published data on mar A inducibility and organic solvent tolerance [12] and suggest involvement of mar A.

Interestingly, nine isolates never grew in the presence of cyclohexane following induction. It is known that salicylate interacts directly with MarR to derepress the operon [6]; however, the exact mechanism of derepression by chloramphenicol, tetracycline, and organic solvents is unknown, although it has been shown that at high concentrations (>10 mM) tetracycline can also bind to MarR. Further, it has been shown that certain MarR homologs (MexR) are unresponsive to salicylate [6]. It seems plausible that those S. choleraesuis isolates unresponsive to induction may possess homologs of MarR that differ in their ability to interact with the inducing agent. Research into this possibility is currently underway.

The effect of mar A expression in bacterial cells is a two- to four-fold increase in resistance of that cell to various antimicrobial agents and organic solvents when compared to the uninduced wild-type cell. The upregulation of the acr B and mic F genes is concomitant with mar A induction and both play important roles in the mechanisms of the MAR phenotype [8]. It has been shown that AcrB is a primary component of an efflux pump mechanism responsible for energy-dependent efflux of antimicrobial agents and solvents [20]. We have shown that S. choleraesuis cyclohexane-resistant mutants with or without prior induction overexpress mar A and acr B using Northern hybridization assays. Subsequently, these cells also had higher MICs of tetracycline (P<0.08), chloramphenicol (P<0.001), rifampin (P<0.08), and nalidixic acid (P<0.001) when compared to their cyclohexane sensitive, uninduced wild-type parent cells.

It has been demonstrated that a second operon, sox RS can also regulate a similar set of genes upon induction, which can lead to increased antimicrobial resistance [12]. Often, both sox S and mar A are upregulated simultaneously, or depending upon the inducing agent one or the other may be induced. We were able to determine through Northern hybridization analysis that none of the cyclohexane resistant mutants expressed sox S to any appreciable level before or after induction. This suggests that mar A and not sox S is mediating the increase in MICs seen among these isolates.

It seems reasonable, based on the results of the present study, that mar A is present in S. choleraesuis isolates, functions as an activator of the mar RAB operon, and regulates downstream genes associated with subsequent MAR.

Acknowledgements

The authors would like to thank Dr. David White, Food and Drug Administration, Center for Veterinary Medicine, for the scientific discussions regarding this research and Dr. Stuart Levy, Center for Adaptation Genetics and Drug Research, Tufts School of Medicine, for the gracious gift of the AG100 and AG102 mar A control strains used in this research.