MarA-Like Regulator of Multidrug Resistance in Yersinia pestis

Abstract

MarA47_Yp from Yersinia pestis, showing 47% identity to Escherichia coli MarA in its N terminus, caused resistance to antibiotics and to organic solvents when expressed in both E. coli and Y. pestis. Resistance was linked to increased expression of the AcrAB multidrug efflux pump. In four of five spontaneous multidrug-resistant mutants of Y. pestis independently selected by growth on tetracycline, the marA47_Yp gene was overexpressed. The findings suggest that marA47_Yp is a marA ortholog in Y. pestis.

Multiple-drug resistance in microorganisms is commonly acquired through plasmids, transposons, or integrons specifying different genes for resistance (12). Alternatively, mutations in chromosomal genes may produce resistance to a wide variety of antibiotics and other toxic substances. One such chromosomal locus, the mar (multiple antibiotic resistance) locus in Escherichia coli and other enteric bacteria, results in resistance to multiple antibiotics, oxidative stress agents, organic solvents, and disinfectant products (1, 2, 6, 7, 10, 27, 29).

In E. coli, the mar locus encodes the marRAB operon (5), specifying MarR, which represses the operon by binding to marO (15), and MarA, which positively regulates the operon and affects expression of the Mar phenotype and many other chromosomal genes through both activation and repression (4, 16, 21, 24-26). marB specifies a small putative protein of unknown function (5). Multidrug resistance is principally caused by MarA-mediated overexpression of the acrAB efflux system (19).

E. coli MarA is a member of the XylS/AraC family of transcriptional activators which contain two helix-turn-helix motifs (9). Most larger members (>250 amino acids [aa]) also possess an effector domain at either the N or C terminus of the protein (9). MarA, a 129-amino-acid protein, lacks the effector domain. MarA control of other chromosomal genes occurs via binding to the “marbox,” a 20-bp degenerate MarA binding site (3, 11).

The genus Yersinia contains 11 species, of which 3 are human pathogens: Y. pestis, Y. pseudotuberculosis, and Y. enterocolitica (20). Yersinia pestis is the causative agent of bubonic and pneumonic plague. The majority of Y. pestis strains contain three plasmids involved in virulence, as well as the ∼102-kb chromosomal pathogenicity locus pgm, specifying genes for a Yersinia-specific iron siderophore and its receptor and hemin adsorption (20). The report of a multidrug-resistant strain (8) and the possible use of Y. pestis as a vehicle for biological warfare have caused increased public health concern.

We asked whether Y. pestis, like other members of the Enterobacteriaceae, contained chromosomal loci, such as marA, involved in multidrug resistance. We identified a Yersinia ortholog of MarA, gene YPO2243 (designated marA47_Yp), which produced multidrug resistance in E. coli and in Y. pestis when overexpressed in each host.

MATERIALS AND METHODS

Bacterial strains, plasmids, and growth conditions.

Table 1 shows the bacterial strains and plasmids used in this study. The Y. pestis strain without the pgm pathogenicity locus was approved by the University Institutional Biosafety Committee for studies at the BL2 level. The E. coli strains were grown at 37°C in Luria-Bertani (LB) broth. The Y. pestis strains were grown in brain heart infusion broth (BHI) (Difco) at 26°C and BHI with 5 mM CaCl₂ at 37°C. Ampicillin (100 μg/ml), kanamycin (50 μg/ml), and tetracycline (5 μg/ml) were added when necessary. The plasmid pJP105 (22) was used to clone the gene marA47_Yp behind the lac promoter, regulated by the plasmid-borne lac repressor. The plasmid pMB102 (21) bearing E. coli marA was used as a positive control. The lacZ promoter was induced with isopropyl-β-d-thiogalactopyranoside (IPTG) at a final concentration of 0.2 mM. pJPBH was the vector control (24).

TABLE 1.

Bacterial strains and plasmids

Strain or plasmid	Characteristic(s)	Reference
E. coli
AG100	uncB+ argE3 thi-1 rpsL xyl mtl galK supE44	10
AG112	Spontaneous marR mutant of AG100	18
AG100B	AG100 acrR replaced with kanamycin cassette	19
AG100Kan	AG100 marCORAB replaced with kanamycin cassette	13
AG100A	AG100 acrAB replaced with kanamycin cassette	19
N8453	ΔlacU169 strain GC4468 deleted of mar (Δ39 kb), soxS, and rob; chloramphenicol resistant (Tn9)	16
Y. pestis
EV7651F	Parental Y. pestis strain with deleted pgm locus; contains 70-kb virulence plasmid	17
Plasmids
pJP105	pBR322 lacI lacZp::soxS (tetracycline susceptible)	22
pMB102	E. coli marA replacing soxS in pJP105 between BamHI and HindIII	21
pRU1	marA47Yp cloned into pJP105 using BamHI and HindIII sites	This study
pJPBH	pJP105 digested with BamHI and HindIII to eliminate soxS, followed by blunt ending and religation	24

Cloning of marA47_Yp.

The primers A47YPF (CCCAAGCTTGGGAATAAAATGATGAGTGAAGACATATTG) and A447YPR (CGGGATCCCGGCGTTAATTCACTGCCATCA) were used at a 50-pmol final concentration to amplify the marA47_Yp coding region beginning with the first putative ATG. The PCR used the Taq polymerase (Invitrogen) and conditions recommended by the manufacturer. marA47_Yp was directionally cloned into pJP105 between the BamHI and HindIII sites using a 1:1 vector to insert ratio and T4 DNA ligase (New England Biolabs), creating the plasmid pRU1.

Drug susceptibility testing.

Drug susceptibilities of strains were determined using E-tests (gift of AB Biodisk, Solna, Sweden). The E. coli strains were tested by incubation at 37°C overnight on LB agar plates with or without IPTG. Y. pestis was examined after incubation for 2 days on BHI agar plates for growth at 26°C or on BHI agar plates with 5 mM CaCl₂ for growth at 37°C with or without IPTG.

Organic solvent tolerance.

An overnight culture was diluted 1/100 in respective medium (LB or BHI) containing 10 mM MgSO₄ and grown at either 37°C or 26°C to an A₅₃₀ of 0.2. After 10-fold dilutions with phosphate-buffered saline, 5 μl was spotted onto agar plates containing 10 mM MgSO₄ and allowed to dry. The plates were flooded with 6 to 8 ml of organic solvent (hexane, cyclohexane, and pentane) under a fume hood and placed in a sealed container. The E. coli cultures were incubated at 30°C overnight and the Y. pestis at room temperature for 2 days.

Isolation of Y. pestis spontaneous mutants.

Approximately 4 × 10⁴ cells of an overnight culture of Y. pestis Ev7651F were inoculated into 15 tubes of fresh BHI broth (2 ml) and incubated at 26°C overnight to an A₆₀₀ of 0.7 to 0.8. One hundred fifty microliters from each tube was plated on BHI agar plates containing 5 μg/ml tetracycline and incubated at 26°C. Over 4 to 5 days, newly appearing colonies were marked on each day. Approximately 75 mutants growing on tetracycline plates were obtained. Five were selected for further study.

RT-PCR.

Early-log-phase cultures of Y. pestis EV7651F/RU1, grown at 26°C and 37°C, were further grown in the presence or absence of 0.2 mM IPTG for ∼3 h. RNA from the spontaneous mutants was isolated from cultures grown at both 26°C and 37°C without IPTG. RNA isolated using a QIAGEN RNA isolation kit was DNase treated. Reverse transcription (RT)-PCRs were performed using ∼500 ng of total RNA. The RT reaction was performed using the Superscript III first-strand synthesis system (Invitrogen) and following the manufacturer's protocol. A minus RT enzyme reaction of each sample served as a negative control.

Real-time PCR.

The real-time PCR assay was carried out using the Quantitect SYBR Green PCR kit (QIAGEN). The reactions followed the kits' protocol. The primers either to acrAB_Yp (SACABF-GCCGGTGATCGCCTGATTA and SACABR-ATGGTCGGATATTGCGCTAC), marA47_Yp (SA47F-CTATATCCGTGGGCGAGTGT and SA47R-GCTTGATTTCCGGCGTATAA), or YP16s (S16VF-CAGAAGAAGCACCGGCTAAC and S16VR-CGGGGATTTCACATCTGACT) rRNA were used for the real-time PCR carried out in the MX4000 multiplex quantitative PCR system from Stratagene. The annealing temperature for the reactions was 55°C for 30 s. Three individual readings were taken at annealing temperature during plateau and dissociation analysis. Each reaction was performed at least twice. The amount of 16S rRNA from each strain was used to normalize mRNA levels. The C_T (threshold cycle) of each gene from the amplification plot was used to calculate the ΔC_T (ΔC_T = C_T of gene X of parental EV7651F minus C_T of gene X of the mutant). The quantity of gene expression was given by q = 2Δ^C_T. The degree of difference in expression of gene X with respect to 16S rRNA was calculated to determine the relative expression of gene X in the mutant compared to that in the wild type.

Western blot analysis.

Protein extracts of E. coli and Y. pestis strains were separated by gel electrophoresis and blotted to membranes for Western analysis using anti-E. coli AcrA antibody (gift of H. Zgurskaya, University of Oklahoma), following published methodology (28).

RESULTS AND DISCUSSION

Search for MarA and AcrAB orthologs in Y. pestis genome.

A marRAB locus similar to that in E. coli was not found in the genome sequence of Y. pestis CO92 (www.sanger.ac.uk), but a number of coding sequences showing homology to the individual genes, marR and marA, were detected. Genes for four orthologs of the MarA protein were found on the chromosome and one each on the 70-kb virulence and the 100-kb plasmids of Y. pestis. Orthologs of the E. coli membrane efflux pumps AcrAB, AcrEF, and EmrEF were also identified using the respective proteins as query sequences.

Cloning of the marA ortholog of Y. pestis.

The attenuated Y. pestis strain EV7651F was used to clone potential marA orthologs. The coding sequence for YPO1737 (36% identity to E. coli MarA) was first chosen because of its size similarity (128 aa) to the E. coli MarA protein. The gene, cloned into plasmid pJP105 downstream of the IPTG-inducible promoter, did not produce any antibiotic resistance in E. coli AG100 or Y. pestis EV7651F when induced by IPTG (data not shown). This lack of activity of a putative MarA homolog was reported earlier for E. coli ykgA, which did not produce multidrug resistance when overexpressed in E. coli (14).

The product of the YPO2243 coding sequence (which we designated marA47_Yp), though double in size (297 aa), was next chosen because of the relatively high identity of its N-terminal half to E. coli MarA (47%) and SoxS (46%) and comparatively lower identity to E. coli Rob (32%) (which is similar in size). The MarA47_Yp protein has the two helix-turn-helix motifs at its N terminus rather than the C terminus as found for AraC. Figure 1 shows the alignment of the N-terminal region of the MarA47_Yp protein with E. coli MarA and other related proteins using the ClustalW program. The phylogenetic tree analysis of proteins examined in Fig. 1 places MarA47_Yp next to E. coli MarA (data not shown).

FIG. 1.

Alignment of MarA47_Yp with MarA and other members of the AraC family. The ClustalW program was used for the alignment of MarA47_Yp with other members of AraC family. The shaded and open horizontal boxes represent the helix-turn-helix (HTH) motifs of MarA. The gray shaded residues are the hydrophobic core of the HTH motif, and the black shaded residues determine the sequence specificity. Asterisks mark the residues that interact with the phosphate backbone group of DNA as determined from the structure of E. coli MarA (23). EMarA, E. coli MarA; YPMarA47, Y. pestis ortholog of MarA; RamA, MarA ortholog from Klebsiella pneumoniae; ESoxS, E. coli SoxS; ERob, E. coli Rob; PqrA, MarA ortholog from Proteus vulgaris; EAraC, E. coli AraC.

marA47_Yp expression results in multiple drug resistance and organic solvent tolerance in both E. coli and Y. pestis.

Plasmid RU1, bearing the cloned marA47_Yp gene, was transformed into E. coli and Y. pestis hosts. When induced by IPTG, marA47_Yp produced multiple drug resistance and organic solvent tolerance in AG100 (Table 2) with values similar to those for AG100 bearing E. coli marA on the same plasmid vector (pMB102). The vector pJPBH served as a negative control. In addition, E. coli strain AG112, which constitutively expresses marA because of an inactivating mutation in marR, also expressed multidrug resistance as expected. Neither transformant showed resistance to aminoglycosides.

TABLE 2.

Drug susceptibilities and organic solvent tolerances of E. coli strains with or without marA or marA47_Yp

Straina	MIC of drug (μg/ml)b												Growth on solventc
	Tet		Dox		Rif		Chl		Nor		Nal		Hexane		Cyclohexane		Pentane
	−	+	−	+	−	+	−	+	−	+	−	+	−	+	−	+	−	+
AG100	1.0	1.0	1.5	1.5	7.5	7.5	4.0	4.0	.08	.08	3.5	3.5	+++	+++	−	−	−	−
AG112	6.0	6.0	6.0	6.0	9.0	9.0	10	10	.25	0.25	6.0	6.0	+++	+++	+++	+++	++	++
AG100/v	1.0	1.0	1.5	1.5	7.5	7.5	4.0	4.0	.08	.08	3.5	3.5	+++	+++	−	−	−	−
AG100/102	1.0	2.0	1.5	2.5	7.5	8.0	4.0	6.0	.08	0.15	3.5	4.5	+++	+++	−	+++	−	−
AG100/RU1	1.0	2.0	1.5	2.5	7.5	8.0	4.0	6.0	.08	0.15	3.5	4.5	+++	+++	−	+++	−	+
AG100Kan	1.0	1.0	1.5	1.5	7.5	7.5	4.0	4.0	.08	.08	3.0	3.0	+++	+++	−	−	−	−
AG100Kan/v	1.0	1.0	1.5	1.5	7.5	7.5	4.0	4.0	.08	.08	3.0	3.0	+++	+++	−	−	−	−
AG100Kan/102	1.0	2.0	1.5	2.5	7.5	8.0	4.0	5.0	.08	0.15	3.0	4.0	+++	+++	−	++	−	−
AG100Kan/RU1	1.0	2.0	1.5	2.5	7.5	8.0	4.0	5.0	.08	0.15	3.0	4.0	+++	+++	−	++	−	+
AG100A	0.3	0.3	0.3	0.3	6.0	6.0	1.5	1.5	ND	ND	ND	ND	−	−	−	−	−	−
AG100A/v	0.3	0.3	0.3	0.3	6.0	6.0	1.5	1.5	ND	ND	ND	ND	−	−	−	−	−	−
AG100A/102	0.3	0.3	0.3	0.3	6.0	6.0	1.5	1.5	ND	ND	ND	ND	−	−	−	−	−	−
AG100A/RU1	0.3	0.3	0.3	0.3	6.0	6.0	1.5	1.5	ND	ND	ND	ND	−	−	−	−	−	−
N8453	0.5	0.5	0.6	0.6	7.0	7.0	ND	ND	.01	0.04	3.0	3.0	−	−	−	−	−	−
N8453/v	0.5	0.5	0.6	0.6	7.0	7.0	ND	ND	.01	0.04	3.0	3.0	−	−	−	−	−	−
N8453/102	0.5	0.8	0.6	1.5	7.0	8.0	ND	ND	.01	0.05	3.0	4.0	−	+++	−	+	−	−
N8453/RU1	0.5	1.0	0.6	1.6	7.0	8.0	ND	ND	.01	0.05	3.0	4.0	−	+++	−	++	−	+

^aAG100, wild type; AG112, marR mutant; AG100Kan, Δmar locus; AG100A, ΔacrAB; N8453, Δmar, sox rob; 102, plasmid pMB102; RU1, plasmid pRU1; v, pJPBH (vector control). See Table 1 for strain descriptions. N8453 has a chloramphenicol resistance transposon.

^bMinus or plus sign, absence or presence of 0.2 mM IPTG; Tet, tetracycline; Dox, doxycycline; Rif, rifampicin; Chl, chloramphenicol; Nor, norfloxacin; Nal, nalidixic acid; ND, not determined. Reproducibly observed increased MICs (compared to those for parental strains) are in boldface. Results are averages for experiments performed in triplicate.

^cGrowth compared to growth on medium without organic solvents: −, no growth; +, minimal growth; ++, moderate growth; +++, full growth.

The induction of marA47_Yp was also studied in the marRAB-deleted mutant AG100Kan and in a mar soxS rob triple deletion mutant, E. coli N8453. The expression of marA47_Yp in both AG100K and N8453 resulted in resistance to multiple antibiotics and organic solvent tolerance (Table 2) comparable to that observed with cloned E. coli marA in AG100Kan and N8453 (Table 2). Thus, marA47_Yp can function independently of the intrinsic mar, soxS, and rob genes of E. coli.

In Y. pestis EV7651 IPTG induction of marA47_Yp on RU1 led to multidrug resistance at both 26°C (the temperature at which Yersinia grows in fleas) and also at 37°C (temperature in mammalian hosts) (Table 3). Growth under the organic solvent hexane was also noted only in the presence of 0.2 mM IPTG (data not shown). Transformants of EV7651F bearing the vector alone showed no drug or organic solvent resistance.

TABLE 3.

Drug susceptibilities of Y. pestis EV7651F with vector or pRU1 at 26°C and 37°C (with 5 mM CaCl₂)^a

Antibiotic	Treatment with IPTG (0.2mM)	MIC of drug
		EV7651F/v		EV7651F/RU1
		26°C	37°C	26°C	37°C
Tet	−	1.2	1.9	1.1	2.4
	+	1.1	2.4	3.0	5.4
Dox	−	1.2	2.0	1.3	2.0
	+	1.0	2.0	3.5	4.4
Rif	−	4.2	3.4	4.8	3.8
	+	5.3	4.3	8.0	5.0
Chl	−	2.3	1.3	1.8	1.6
	+	2.0	1.5	4.0	4.4
Nor	−	0.1	0.2	0.1	0.3
	+	0.1	0.2	0.5	0.9
Nal	−	0.4	0.5	0.5	0.5
	+	0.4	0.6	2.6	1.3

^aStrains were assayed by the E-test in the absence (−) or presence (+) of IPTG. MICs of Y. pestis bearing marA47_Yp (on pRU1) in the presence of IPTG are shown in boldface. EV7651F/v, Y. pestis bearing pJPBH. Results are averages for four experiments. Antibiotic designations are as described for Table 2.

The marA47_Yp-associated multidrug resistance functions via the AcrAB efflux pump.

Cloned marA47_Yp on pRU1 transformed into E. coli AG100A (acrAB deletion mutant) neither conferred multidrug resistance nor exhibited tolerance to organic solvents. Western blot analysis of AG100/RU1, AG100Kan/RU1, and N8453/RU1 cell extracts showed increased expression of AcrA only in the presence of IPTG induction of marA47_Yp (data not shown). These results imply that MarA47_Yp, like MarA, confers multiple drug resistance through the AcrAB efflux pump in E. coli.

To determine whether the induction of marA47_Yp in Ev7651F/RU1 also results in expression of acrAB_Yp, real-time PCR was performed (Materials and Methods). The levels of marA47_Yp transcripts were 15-fold higher in the presence of 0.2 mM IPTG at both temperatures, while the acrAB_Yp transcript was increased fourfold by 0.2 mM IPTG at both 26°C and 37°C. No difference was observed in the expression of other potential efflux pump genes, acrEF_Yp, and emrAB_Yp in Ev7651F/RU1 transformants induced with 0.2 mM IPTG (data not shown). This result suggests that induction of multiple-drug resistance by marA47_Yp in Y. pestis also functions via the AcrAB efflux pump.

marA47_Yp expression is increased in spontaneous multiple-drug-resistant mutants of Y. pestis.

Five mutants chosen from independently selected early-log-phase cultures of EV7561F plated on BHI plates containing 5 μg/ml of tetracycline (see Materials and Methods) showed increased levels of resistance to multiple antibiotics. The mutants displayed differences in organic solvent tolerance (Table 4).

TABLE 4.

Drug susceptibilities, organic solvent tolerances, and real-time PCR data for Y. pestis EV7651F spontaneous tetracycline-selected mutants at 26°C and 37°C (with 5mM CaCl₂)

Strain	MIC (μg/ml) of druga												Growth with organic solventa		Gene expressionb
	Tet		Dox		Rif		Chl		Nor		Nal		Hexane	Cyclohexane	marA47Yp		acrABYp
	26°C	37°C	26°C	37°C	26°C	37°C	26°C	37°C	26°C	37°C	26°C	37°C	37°C	37°C	26°C	37°C	26°C	37°C
EV7651F (WTc)	1.4	2.0	0.9	2.3	4.0	4.3	1.8	1.7	0.1	0.4	0.5	0.6	−	−
Mutant 127	3.7	3.3	8.0	6.0	6.0	6.0	8.0	1.8	0.4	0.2	3.7	0.7	+++	++	++	++	+	+
Mutant 128	5.3	3.7	8.0	4.7	8.3	6.7	12	0.9	0.5	0.8	5.0	0.3	+++	++	++	++	+	+
Mutant 11	3.0	6.0	7.3	13.3	5.3	4.7	9.3	5.3	0.3	0.5	1.7	0.5	+++	+++	+++	++	+	+
Mutant 55	2.7	2.7	4.0	6.0	6.0	5.3	3.7	2.5	0.1	0.3	0.7	0.5	−	−	−	−	+	+
Mutant 102	4.7	3.3	7.3	4.3	6.7	6.0	6.7	2.0	0.7	0.2	2.7	0.6	+++	+++	++	+++	++	++

^aFor antibiotic designations and organic solvent descriptions, see Table 2. Results represent averages for three experiments. Reproducibly observed increased MICs compared to those for the wild type are presented in boldface. Mutants 127 and 128 appeared on day 2, mutant 11 on day 3, mutant 55 on day 4, and mutant 102 on day 5 of tetracycline selection.

^bExpression was assayed by real-time PCR (see Materials and Methods). −, ≤1; +, 2- to 9-fold; ++, 10- to 50-fold; +++, >50-fold. No increase in expression was found for acrEF_Yp or emrAB_Yp at either temperature (data not shown).

^cWT, wild type.

Real-time PCR demonstrated overexpression of acrAB_Yp transcription in all five mutants. All except mutant 55 showed increased expression of marA47_Yp at both 26°C and 37°C (Table 4). There was no direct correlation between the level of marA47_Yp and acrAB_Yp expression. No change was detected (by real-time PCR) in the levels of acrEF_Yp and emrAB_Yp expression for any of the mutants compared to those for parental strain EV7651F (data not shown). The enhanced acrAB_Yp expression was confirmed by an increase in the AcrA protein seen in Y. pestis mutants by Western blot analysis using antibody to E. coli AcrA (Fig. 2). The relatively smaller increase in AcrA in mutant 55 was consistent with its relatively lower level of resistance.

FIG. 2.

AcrA expression in multidrug-resistant mutants of Y. pestis. AcrA expression was determined by Western blot assay using anti-AcrA antibodies (see Materials and Methods). E. coli AG100A deleted of acrAB and E. coli AG100B deleted of acrR (overexpressing AcrAB) served as negative and positive controls, respectively.

The isolation of spontaneous multiple-drug resistance mutants of Y. pestis with associated increased expression of marA47_Yp further supports the designation of marA47_Yp as a functional ortholog of marA in Yersinia.