Effect of Transcriptional Activators SoxS, RobA, and RamA on Expression of Multidrug Efflux Pump AcrAB-TolC in Enterobacter cloacae

Abstract

Control of membrane permeability is a key step in regulating the intracellular concentration of antibiotics. Efflux pumps confer innate resistance to a wide range of toxic compounds such as antibiotics, dyes, detergents, and disinfectants in members of the Enterobacteriaceae. The AcrAB-TolC efflux pump is involved in multidrug resistance in Enterobacter cloacae. However, the underlying mechanism that regulates the system in this microorganism remains unknown. In Escherichia coli, the transcription of acrAB is upregulated under global stress conditions by proteins such as MarA, SoxS, and Rob. In the present study, two clinical isolates of E. cloacae, EcDC64 (a multidrug-resistant strain overexpressing the AcrAB-TolC efflux pump) and Jc194 (a strain with a basal AcrAB-TolC expression level), were used to determine whether similar global stress responses operate in E. cloacae and also to establish the molecular mechanisms underlying this response. A decrease in susceptibility to erythromycin, tetracycline, telithromycin, ciprofloxacin, and chloramphenicol was observed in clinical isolate Jc194 and, to a lesser extent in EcDC64, in the presence of salicylate, decanoate, tetracycline, and paraquat. Increased expression of the acrAB promoter in the presence of the above-described conditions was observed by flow cytometry and reverse transcription-PCR, by using a reporter fusion protein (green fluorescent protein). The expression level of the AcrAB promoter decreased in E. cloacae EcDC64 derivates deficient in SoxS, RobA, and RamA. Accordingly, the expression level of the AcrAB promoter was higher in E. cloacae Jc194 strains overproducing SoxS, RobA, and RamA. Overall, the data showed that SoxS, RobA, and RamA regulators were associated with the upregulation of acrAB, thus conferring antimicrobial resistance as well as a stress response in E. cloacae. In summary, the regulatory proteins SoxS, RobA, and RamA were cloned and sequenced for the first time in this species. The involvement of these proteins in conferring antimicrobial resistance through upregulation of acrAB was demonstrated in E. cloacae.

INTRODUCTION

Enterobacter cloacae is an important nosocomial pathogen responsible for various infections, including sepsis, infections of the respiratory tract and urinary tract, wound infections, and meningitis. Multiple antibiotic-resistant strains have caused outbreaks of infections in hospitals, usually in settings where seriously ill patients are housed, such as intensive care units (ICUs). These pathogens of concern in an ICU setting cause significant morbidity and mortality, and infection management is complicated by resistance to multiple antibiotics (44).

Efflux pumps confer innate resistance to a wide range of toxic compounds, such as antibiotics, dyes, detergents, and disinfectants in members of the family Enterobacteriaceae (35, 36). Therefore, efflux pumps participate, at least partly, in the ability of bacteria to adapt to diverse environments, in drug resistance mechanisms, and in bacterial pathogenesis.

Control of membrane permeability is a key step in regulating the intracellular concentration of antibiotics in Enterobacteriaceae. The expression of porins and efflux pump components is jointly controlled by several positive global regulators, which respectively decrease or enhance transcription of specific genes such as acrAB and tolC, directly or via a regulation cascade. The acrAB regulation mechanism must be examined to understand the physiological role of AcrAB. In Escherichia coli, the expression of acrAB may be subjected to multiple levels of regulation, and it is locally modulated by the AcrR repressor (23). At a global level, AcrAB is controlled by stressful conditions and by regulators such as MarA, SoxS, and Rob (37, 40). In addition to identifying the MarA/SoxS/Rob family, George et al. (13) also identified and characterized RamA, a member of the AraC/XylS family, for its role in conferring multidrug resistance (MDR) in Klebsiella. Most recently, the ramA gene was identified in Enterobacter aerogenes, Enterobacter cloacae, and Salmonella Typhimurium, in which it may also be involved in MDR (9, 19, 52). These activators perform their function via binding to discrete but degenerate nucleotide sequences known as the mar-, rob-, or soxbox sequences located in the upstream region of regulated genes (including acrAB) (14, 18, 25, 26, 30). The expression of each regulator is controlled differently. In E. coli, the marA transcription is repressed by marR and derepressed through binding by compounds such as salicylate (45). The level of soxS is controlled by the activator soxR, in contrast to the repressor activity of marR on marA. Activation of soxR by superoxides and redox cycling compounds, such as paraquat, induces expression of soxS and the subsequent induction of the soxRS regulon (11). Rob is an abundant protein that is expressed constitutively (3), and its activity is enhanced by decanoate (40). Regulation of ramA is provided locally by ramR, presumably through prevention of RamR binding to an operator sequence near ramA and the subsequent relaxation of RamR repression at the ramA promoter (1, 38). Furthermore, induction of acrAB by indole is regulated through RamA, independently of MarA, SoxS, and Rob (30). The main goal of the present study was to investigate the regulation of acrAB, as well as its activator-encoding genes soxS, robA, and ramA isolated from E. cloacae, in the presence of different stress agents such as salicylate, decanoate, tetracycline, and paraquat. In order to define this regulation system, the effect of the inactivation or artificial overexpression of soxS, robA, and ramA on the expression of acrAB and the multidrug resistance phenotype of E. cloacae was also investigated.

MATERIALS AND METHODS

Strains, culture media, and plasmids.

The laboratory strains and plasmids used in the present study are listed in Table 1. The clinical isolates used were two clonally unrelated strains of Enterobacter cloacae isolated from two different patients: EcDC64 and Jc194 (33, 34). Both strains were isolated from a patient admitted to the A Coruña University Hospital (Spain). Escherichia coli strain TG1 was used for cloning procedures. All strains used in the study were maintained at −80°C in 15% (vol/vol) glycerol until use. The strains were grown on MacConkey agar plates (Becton Dickinson, Franklin Lakes, NJ), in Luria-Bertani (LB) broth, or on LB agar in the presence of 50 μg of ampicillin/ml, 25 μg of kanamycin/ml, 30 μg of chloramphenicol/ml, 4 μg of gentamicin/ml, or 30 μg of tetracycline/ml, when necessary. Salicylate (10 mM), paraquat (100 μM), or decanoate (10 mM) (all from Sigma-Aldrich, St. Louis, MO) were added to LB agar and broth media, as required.

Table 1.

Bacterial strains and plasmids used in the present study

Strain or plasmid	Features (resistance marker)	Source or reference
Strains
E. cloacae
EcDC64	MDR phenotype strain overexpressing the AcrAB-TolC efflux pump	33, 34
Jc194	Clinical isolate with basal efflux pump expression and a more susceptible resistance profile than strain EcDC64	33, 34
EcΔsoxS::Km	EcDC64 with the soxS gene disrupted with a kanamycin resistance marker	This study
EcΔrobA::Km	EcDC64 with the robA gene disrupted with a kanamycin resistance marker	This study
EcΔramA::Km	EcDC64 with the ramA gene disrupted with a kanamycin resistance marker	This study
E. coli
BL21(DE3)	E. coli cell suitable for transformation and protein purification	Invitrogen
TG1	Susceptible E. coli strain used for cloning procedures	Invitrogen
Plasmids
pACYC184	Cloning vector (chloramphenicol and tetracycline)	7, 39
pAcGFP1	Cloning vector (ampicillin)	Clontech
pUCP24	Cloning vector (gentamicin)	50
pKOBEG	Red helper plasmid (chloramphenicol)	8
pGFP	pACYC184 (chloramphenicol) containing the acrAB promoter fused to the gene coding for GFP	This study
pCR-BluntII-TOPO	Cloning vector (kanamycin)	Invitrogen
pTsoxS	pCR-BluntII-TOPO (kanamycin) containing the soxS gene under the control of its own promoter	This study
pTrobA	pCR-BluntII-TOPO (kanamycin) containing the robA gene under the control of its own promoter	This study
pTramA	pCR-BluntII-TOPO (kanamycin) containing the ramA gene under the control of its own promoter	This study
pUCsoxS	pUCP24 (gentamicin) containing the soxS gene under the control of its own promoter	This study
pUCrobA	pUCP24 (gentamicin) containing the robA gene under the control of its own promoter	This study
pUCramA	pUCP24 (gentamicin) containing the ramA gene under the control of its own promoter	This study
pGEM-t	Cloning vector (ampicillin)	Promega
pET-28	Expression vector (kanamycin)	Novagen
pETM-44	Modified pET-24d expression vector (kanamycin)	EMBL

PCR amplification, sequencing, and cloning of soxS, robA, and ramA efflux pump regulatory genes.

Genomic DNA from EcDC64 was extracted from overnight cultures at 37°C by use of a genomic DNA purification kit (Promega Corp., Madison, WI). The oligonucleotides used to isolate, amplify, and clone the efflux pump regulatory genes are listed in Table 2. To isolate and amplify soxS, robA, and ramA, oligonucleotides were designed on the basis of the previously reported nucleotide sequence of Enterobacter sp. strain 638 complete genome (GenBank accession code CP000653). The upstream region of these genes was amplified by using two consecutive random and specific primers. The partially random primer arb.1F (5′-GGCCACGCGTCGACTAGTACNNNNNNNNNNACGCCC, where N represents A, T, G, or C) was used in the first PCR to amplify anonymous fragments of DNA in the upstream region of the regulator genes. PCR conditions consisted of 94°C for 3 min, followed by 5 cycles of 94°C for 30 s, 30°C for 30 s, and 72°C for 1 min, followed by 30 cycles of 94°C for 30 s, 45°C for 30 s, and 72°C for 1 min, with a final 5-min extension step at 72°C. The DNA fragments generated were then used as a template for the second PCR in which the primer arb.2F (5′-GGCCACGCGTCGACTAGTAC) was used. The reaction was performed under the same conditions as the first PCR, except annealing was carried out at 55°C. The PCR product was purified from the gel, by use of a Geneclean kit (MP Biomedicals, Ohio), and was then sequenced. Sequencing was carried out with a Taq DyeDeoxi- terminator cycle sequencing kit in an automatic DNA sequencer (377 ABI Prism; Perkin-Elmer). Specific oligonucleotides (Table 2) were designed from the sequences obtained to amplify each regulator with its own promoter. Amplicons were then cloned into the pCR-Blunt II-TOPO cloning vector, according to the manufacturer's instructions (Invitrogen Corp., Carlsbad, CA), to generate the recombinant plasmids named pTsoxS, pTrobA, and pTramA (Table 1).

Table 2.

Primers used in this study

Primer	Genea	Procedure	Sequence (5′–3′)b
sox-F	soxS*	Cloning	ATGTCNCATCAGCARATWATTCAG
sox-R	soxS	Cloning	TTAGTTGAGCTGGTGGCGGTA
soxUp	soxS	Cloning	CAGGGCCACCAGCGTGTGAATA
rob-F	robA	Cloning	ATGGATCAGGCTGGAATTATT
rob-R	robA	Cloning	TTAACGACGTACCGGAATCAG
robUp	robA	Cloning	GGTGACGTTTTTAACGTCCGGATCG
ram-F	ramA	Cloning	ATGATCAATCAGGAAGCTGA
ram-R	ramA*	Cloning	TCAGTGSGYRCGRCTGTG
ramUp	ramA	Cloning	TTGTGCTGGCGAAACATACC
soxSKmF	soxS	Knockout	ACCATTTCGAATAGCCCGACTTCCTGGCGACCACATCAATGTTCAACGGCTGGTCGATATGTTCATCAATCCATTCAATACATATGAATATCCTCCTTAG
soxSKmR	soxS	Knockout	AGTGCCCTTCCGGCAATACGCCGAACGCGTCGCCGATGGTGGCCAGAGGAATGCCGATACGTTGCGCAATTTTAATGATC TGTGTAGGCTGGAGCTGCTTCG
robAKmF	robA	Knockout	GGCTATTCCAAGTGGCATCTGCAAAGGATGTTCAAGGATGTCACCGGTCATGCTATCGGTGCCTATATTCGCGCACGTCGTGTGTAGGCTGGAGCTGCTTCG
robAKmR	robA	Knockout	AACGCGCATCTGATGGCGGAACTCGGAGATCTGCTCCAGGGAGCAGGAGTAGCTCTGCGTGGTGCCGACCAGGTGCGTTT CATATGAATATCCTCCTTAG
ramAKmF	ramA	Knockout	CTACACCAGCCGTTACGCATCGAAGAAATTGCCCGCCACGCGGGTTACTCAAAATGGCATTTACAGCGGCTGTTTATGCA TGTGTAGGCTGGAGCTGCTTCG
ramAKmR	ramA	Knockout	TGCGGGTAAACGTCTGCTGCGAGTCAAACCCGTAGCGCAGGCAGATGTCGTACACCCGCTCGTCTGACTCACGCAGATCG CATATGAATATCCTCCTTAG
pacrA-F	acrAB promoter	Cloning	TCAGCAACGAACTCACATTTATG
pacrA-R	acrAB promoter	Cloning	TAACCCTCTGTTTTTGTTCATATGT
pacrAgfp-F	pacrA::gfp construction	Cloning	CCAACGAACTCACATTTATG
pacrAgfp-F	pacrA::gfp construction	Cloning	AGTCGCGGCCGCTCACTTGTACAGC
soxSpET44F	soxS	Cloning	acatgtCGCATCAGCAAATTATTCAG
soxSpET44F	soxS	Cloning	aagcttTTAGTTGAGCTGGTGGCGGT
ramApET44F	ramA	Cloning	ccatggTCAATCAGGAAGCTGAAGGG
ramApET44R	ramA	Cloning	aagcttTCAGTGCGCGCGGCTGTG
robApET28F	robA	Cloning	catatgGATCAGGCTGGAATTAT
robApET28F	robA	Cloning	ctcgagTTAACGACGTACCGGAATCA
acrA RT-F	acrA	RT-PCR	GCCTCTGGCGGTCGTTCTGAT
acrA RT-R	acrA	RT-PCR	AGAGGTTCGGATTTGAGCGTCAC
acrB RT-F	acrB	RT-PCR	GTGAGCGTCGAGAAATCGTCCA
acrB RT-R	acrB	RT-PCR	TACGGCTGATGGCGTCCTTCAT
rpoB RT-F	rpoB*	RT-PCR	CAGCCGCGAYCAGGTTGACTACA
rpoB RT-R	rpoB	RT-PCR	GACGCACCGCAGGATACCACCTG
soxS RT-F	soxS	RT-PCR	GCGCAGGTTACTGCTGGCAGCG
soxS RT-R	soxS	RT-PCR	GCGGCGAAACACGCGGGAAA
robA RT-F	robA	RT-PCR	CGCGCGGCCCATTCTTGATA
robA RT-R	robA	RT-PCR	GAGTGCCGGCGTCAACGAGA
ramA RT-F	ramA	RT-PCR	CAAAGGCGAAAGTCTGGGGCG
ramA RT-R	ramA	RT-PCR	CCCGTAGCGCAGGCAGATATCG

^a*, The oligonucleotide degenerated where R is A or G,Y is C or T, S is C or G, W is T or A, and N is A, T, C, or G.

^bNucleotides shown in lowercase italics indicaterestriction sites for cloning.

Construction of knockout strains.

Disruption of the soxS, robA, and ramA genes in strain EcDC64 was performed by the method described by Datsenko and Wanner (10), with some modifications. The Red helper pKOBEG (kindly donated by J. M. Ghigo, Institut Pasteur, Paris, France) (Table 1) is a low-copy-number plasmid that contains a gene for chloramphenicol resistance, a temperature-sensitive origin of replication, and the Red system, which comprises an exonuclease and the β and γ functions of phage λ. The pKOBEG plasmid (Table 1) was introduced in the strain EcDC64 by heat shock, and transformants were selected on LB agar with chloramphenicol, after incubation for 24 h at 30°C. One transformant carrying the Red helper plasmid was made electrocompetent. A selectable resistance gene was amplified by PCR from genomic DNA by use of primers including 5′ extensions with homology for the soxS, robA, and ramA genes listed in Table 2. The PCR product was used to disrupt the soxS, robA, and ramA genes of strain EcDC64 by electroporation. Electroporation (25 μF, 200 Ω, and 2.5 kV) of the electrocompetent strains was carried out according to the manufacturer's instructions (Bio-Rad Laboratories, Madrid, Spain), with 50 μl of a cell suspension and 1 μg of the purified and dialyzed PCR products. Dialysis was performed in order to remove the salts of purified PCR products using 0.025-μm-pore-size nitrocellulose membranes (Millipore, Billerica, MA). Shocked cells were added to 1 ml of LB broth, incubated overnight at 30°C, spread on LB agar containing kanamycin, and incubated for 24 h at 30°C. The mutant strains were then grown on LB agar containing kanamycin for 24 h at 44°C and incubated overnight at 30°C on LB agar containing kanamycin and chloramphenicol in order to test for the loss of the helper plasmid.

Construction of transcriptional reporter fusion.

The oligonucleotides listed in Table 2 were designed to amplify a fragment of 526 bp that contained 506 bp upstream and 21 bp downstream of the ATG of the acrAB operon from E. cloacae EcDC64. The PCR product was double digested with BamHI and NcoI and fused to the gene encoding the green fluorescent protein (GFP; acgfp1) in the pACGFP1 vector (Table 1). The fragment containing the transcriptional fusion was amplified with specific oligonucleotides (Table 2), purified and digested with HindIII and BamHI, and then ligated with the T4 DNA ligase (Promega) into a similarly digested pACYC184 expression vector to generate the recombinant plasmid named pGFP (Table 1). Fusion was sequenced in order to verify the DNA sequence.

Cloning procedures for complementation assays.

Recombinant plasmids were first constructed and then introduced in the knockout strains for complementation studies. For construction of these plasmids, universal primers M13 and M13Rv were used to amplify the soxS, robA, and ramA genes under the control of their own promoter regions from the pTsoxS, pTrobA, and pTramA recombinant vectors, respectively. The amplified DNA was purified, digested, and then ligated with the T4 DNA ligase (Promega) into the pUCP24 expression vector (Table 1), previously digested with the same enzymes. The PCR fragments containing soxS and ramA genes were cloned between BamHI and EcoRI sites, and the fragment containing the robA gene was cloned between the HindIII and PstI sites. The accuracy of the construct was checked by restriction analysis.

Antibiotic susceptibility testing.

The susceptibility of various strains to the following antibiotics was determined by the standard disk diffusion method (Becton Dickinson) on Mueller-Hinton agar (Tables 3 and 4): telithromycin, ciprofloxacin, erythromycin, chloramphenicol, and tetracycline. The susceptibility to these antibiotics was also determined in the presence of salicylate (10 mM), decanoate (10 mM), paraquat (100 μM), or tetracycline (the MICs for the strains tested were determined, after a gradual, stepwise increase in the exposure to tetracycline) (Table 4).

Table 3.

Antibiotic susceptibility profiles^a

Strain and stimulus	Antibiotic susceptibility (inhibition zone diam [mm])
	Erythromycin	Tetracycline	Telithromycin	Ciprofloxacin	Chloramphenicol
Jc194
None	10	25	13	34	24
SAL	0	18	9	28	21
DEC	0	23	11	32	24
PQ	0	18	8	27	17
TET	0	11	8	27	11
EcDC64
None	0	0	9	32	25
SAL	0	0	0	28	21
DEC	0	0	8	32	25
PQ	0	0	0	27	18
TET	0	0	0	32	24
EcΔrobA
None	10	26	14	38	24
SAL	0	21	11	28	22
DEC	10	26	13	46	33
PQ	0	18	11	27	17
TET	0	14	10	26	16
EcΔsoxS
None	9	23	14	34	24
SAL	0	18	11	29	25
DEC	8	22	12	32	25
PQb	10	31	17	42	28
TET	0	16	11	29	15
EcΔramA
None	10	25	15	37	25
SAL	0	19	12	28	23
DEC	9	22	12	31	27
PQ	0	18	12	32	18
TET	0	24	17	40	26

^aAntibiotic susceptibility profiles, expressed as the diameters of the inhibition zones, were determined by the standard disk diffusion method for the bacterial isolates in the presence of various stimuli: 10 mM sodium salicylate (SAL), 10 mM sodium decanoate (DEC), 0.1 mM paraquat (PQ), or 4 μg of tetracycline/ml (TET). A value of “0” means that an inhibition zone around the disk (6 mm) was not detected.

^bThe concentration of paraquat used was 50 μM.

Table 4.

Antibiotic susceptibility profiles^a

Antibiotic	Susceptibility profile (inhibition zone diam [mm]) for:
	E. cloacae EcDC64										E. cloacae Jc194
	EcDC64	EcΔsoxS	EcΔsoxS(pUCsoxS)	EcΔsoxS(pUCP24)	EcΔrobA	EcΔrobA(pUCrobA)	EcΔrobA(pUCP24)	EcΔramA	EcΔramA(pUCramA)	EcΔramA(pUCP24)	Jc194	Jc194(pTsoxS)	Jc194(pTrobA)	Jc194(pTramA)	Jc194(pTΦ)b
ERY	0	9	0	9	10	0	10	10	0	11	10	0	0	0	8
TET	0	23	13	25	26	14	26	25	14	25	25	0	0	0	20
TEL	9	14	9	19	14	10	18	15	11	18	13	0	0	0	14
CIP	32	34	22	38	38	21	39	37	23	40	34	25	26	23	32
CHL	25	24	12	23	24	12	25	25	15	25	24	0	0	8	21

^aAntibiotic susceptibility profiles, expressed as diameters of the inhibition zones, were determined by the standard disk diffusion method for the bacterial isolates indicated. Abbreviations: ERY, erythromycin; TET, tetracycline; TEL, telithromycin; CIP, ciprofloxacin; CHL, chloramphenicol. A value of “0” means that an inhibition zone around the disk (6 mm) was not detected.

^bpCR-BluntII-TOPO (pTΦ) is the empty cloning vector.

Flow cytometry assays.

To examine the effect of transcriptional regulators SoxS, RobA, and RamA on the multidrug efflux pump AcrAB-TolC in E. cloacae, expression experiments were performed with GFP as a reporter. Single colonies of each bacterial strain (harboring the pGFP plasmid) were inoculated into 10 ml of LB broth containing appropriate amounts of the selected antibiotics. After overnight incubation at 37°C and 180 rpm, the cultures were diluted 1:100 in LB medium. The cells were then incubated at 37°C with continuous shaking until they reached an optical density at 600 nm (OD₆₀₀) of 0.4 and were then diluted 1:100 in 2 ml of 0.9% saline solution. Induction experiments were performed to test the effects of sodium salicylate (SAL), sodium decanoate (DEC), paraquat (PQ), and tetracycline (TET) on acrAB expression. A 10 mM concentration of SAL, 10 mM DEC, 0.1 mM PQ, or the MIC of TET was added to the cultures. Fluorescence analysis was performed with a FACScan cytometer, and 50,000 cells were measured for each sample. The values obtained were calculated as fluorescence units relative to the control strain containing the reporter plasmid grown under the same conditions.

Real-time RT-PCR experiments.

Real-time reverse transcription-PCR (RT-PCR) was carried out to determine the expression levels of the soxS, ramA, and robA regulator genes and the efflux components acrA and acrB. Specific primers designed from soxS, ramA, robA, acrA, and acrB sequences (GenBank accession code JQ727666 for soxS gene; JQ727667 for robA gene and JQ727668 for ramA gene) are listed in Table 2. Total RNA was isolated with a High-Pure RNA isolation kit (Roche Diagnostics GmbH, Mannheim, Germany) according to the manufacturer's instructions. The concentration of RNA was determined spectrophotometrically.

RNA (1 μg) was reverse transcribed into single-stranded cDNA by use of a Transcriptor first-strand cDNA synthesis kit (Roche Diagnostics) according to the manufacturer's instructions. cDNA was quantified by real-time PCR amplification with specific primers (Table 2) by use of a LightCycler 480 SYBR green I master kit and a LightCycler 480 instrument (both from Roche Diagnostics) with an initial incubation of 95°C for 10 min, followed by 45 cycles of 10 s at 95°C, 20 s at 60°C, and 10 s at 72°C. In all cases, the expression levels were standardized relative to the transcription levels of rpoB (a housekeeping gene) for each isolate.

Cloning, overexpression, and protein purification.

Full-length soxS, robA, and ramA genes were amplified by PCR from genomic DNA of EcDC64 strain with the primers listed in Table 2. A DNA fragment corresponding to the robA gene was cloned into the pET-28 expression vector (Novagen) between NdeI and XhoI restriction sites. The soxS and ramA genes were cloned to MBP (maltose-binding protein) into the pETM-44 expression vector (modified pET-24d, EMBL-made vector by Arie Geerlof) between NcoI and HindIII restriction sites. Fusion proteins with a His tag at the N-terminal region were expressed in conventional E. coli strain BL21(DE3) cells grown in LB medium supplemented with 40 μg of kanamycin/ml. Cells were grown at 37°C until the OD₆₀₀ reached 0.6. Protein expression was induced by the addition of IPTG (isopropyl-β-d-thiogalactopyranoside; Calbiochem) to the culture, to a final concentration of 0.5 mM, followed by incubation for 5 to 6 h at 37°C. Cell cultures were harvested by centrifugation (6,000 × g, 20 min, 4°C). Cell pellets were resuspended in 10 ml of lysis buffer (10 mM Tris-HCl [pH 8.0], 150 mM NaCl) and sonicated in a Misonix S4000 sonicator. Bacterial lysates were pelleted at 30,000 rpm for 1 h at 4°C in a 60Ti rotor (Beckman). The supernatant was loaded onto a His-Trap HP 5-ml column (GE Healthcare-Amersham Biosciences) equilibrated in binding buffer (10 mM Tris-HCl [pH 8.0], 5 mM imidazole, 500 mM NaCl). The target proteins were eluted with 50 ml of elution buffer (10 mM Tris-HCl [pH 7.0], 150 mM imidazole, 500 mM NaCl).

DNA mobility shift assays.

Electrophoretic mobility shift assays (EMSAs) were performed as previously described (2). The DNA promoter of the acrA gene was amplified by PCR from E. cloacae EcDC64 genomic DNA by use of suitable oligonucleotide primers (Table 2), and the purified PCR fragment was cloned into the pGEM-T vector (Promega). The presence of the desired promoter was confirmed by sequencing the DNA plasmid with the T7 and SP6 primers of the pGEM-T vector. A DNA probe was prepared by PCR amplification, with one of the primers labeled with digoxigenin (DIG) at its 5′ end and was then purified. DNA-protein reaction mixtures (20 μl) containing 25 ng of a DIG-labeled DNA probe and SoxS, RamA, or RobA proteins were incubated for 30 min at 37°C in EMSA buffer containing 20 mM Tris-HCl (pH 8), 50 mM KCl, 5% (vol/vol) glycerol, 1 μg of bulk carrier sperm salmon DNA, 0.5 mM 1,4-dithiothreitol, and 0.1 mg of bovine serum albumin per ml. DNA-protein complexes were visualized by separation on a 5% nondenaturing polyacrylamide gel (40 mM Tris-acetate [pH 8.0]) at 90 V for 3 h and were then transferred to a Biodine B nylon membrane (Pall Gelman Laboratory). DIG-labeled DNA-protein complexes were detected according to the manufacturer's protocol (Roche).

Nucleotide sequence accession codes.

The GenBank accession codes for the nucleotide sequences determined in the present study are as follows: JQ727666 for the soxS gene, JQ727667 for the robA gene, and JQ727668 for the ramA gene.

RESULTS

Stressing compounds such as salicylate, decanoate, tetracycline, and paraquat modify the resistance phenotype of E. cloacae.

To analyze the effect of salicylate, decanoate, tetracycline, and paraquat on the resistance profile of E. cloacae, two clinical isolates of E. cloacae were used: EcDC64 and Jc194. E. cloacae EcDC64 was isolated from a patient admitted to the ICU of the A Coruña University Hospital (northwestern Spain). The bacterial isolate displayed a MDR phenotype overexpressing the AcrAB-TolC efflux pump and lower permeability than strain Jc194 (with basal efflux pump expression level) (33, 34).

The resistance profile to several antibiotic families known to be good substrates for efflux pumps (such as macrolides, tetracyclines, or quinolones) was determined in both strains, by the standard disk diffusion method, in the presence or absence of the different compounds studied (Table 3).

The effect of these compounds on the resistance profile of E. cloacae was evaluated, and the results revealed that the susceptibility to a number of antibiotics belonging to different antibiotic families was reduced in both wild-type (WT) strains in response to the compounds. When the bacteria were incubated with salicylate and paraquat, a decrease in the susceptibility to all of the antibiotics was observed in both WT strains (Table 3). The effect of tetracycline on antibiotic resistance of EcDC64 was not clearly established because of the high level of tetracycline resistance shown by this clinical isolate. However, this antibiotic was the strongest inducer of resistance in E. cloacae strain Jc194. The inhibition zones were reduced for all antibiotics, which indicate a significant increase in the resistance of this strain to the antibiotics.

Sodium decanoate had a modest effect on the antibiotic resistance of strain Jc194. The susceptibility to all antibiotics, except chloramphenicol, was slightly decreased. However, the resistance profile of EcDC64 was not modified by incubation of this strain with sodium decanoate. The intensity of the effect differed depending on the type of molecule (Table 3).

Analysis of acrAB expression in the presence of different stressing compounds.

Flow cytometry assays were performed with the WT strain Jc194, which showed a basal expression level of the AcrAB-TolC efflux pump and a susceptible profile of resistance. Fluorescence data for E. cloacae Jc194 cultures containing the fusion acrAB promoter-GFP gene were measured in the presence of 10 mM sodium salicylate, 10 mM sodium decanoate, 4 μg of tetracycline/ml, and 0.1 mM paraquat. The results obtained are shown in Fig. 1A. All of the compounds tested induced an increase in the fluorescence intensity, indicating that the AcrAB efflux pump was upregulated in the presence of these compounds. The greatest increase in fluorescence was observed when E. cloacae strain Jc194 was incubated with tetracycline. A similar result was obtained in response to salicylate, although the increase in intensity was slightly lower. The compound that triggered the lowest expression of acrAB was sodium decanoate (Fig. 1A). To confirm the results of the acrAB expression levels obtained by flow cytometry experiments with GFP as a reporter, real-time RT-PCR assays were also performed. The results of the RT-PCR assays were found to be consistent with the flow cytometry data (Fig. 1B).

Fig 1.

(A) Induction of the acrAB operon measured by growing E. cloacae Jc194 containing the reporter plasmid pGFP in the presence of 10 mM sodium salicylate (SAL), 10 mM sodium decanoate (DEC), 0.1 mM paraquat (PQ), or 4 μg of tetracycline/ml (TET). The values shown are relative fluorescence units, comparative to the control strain, Jgfp, grown without the compounds. The bars show the average values from triplicate assays. P < 0.05 in all cases. (B) RT-PCR analysis of acrA and acrB gene expression in E. cloacae Jc194 in the presence of 10 mM sodium salicylate (SAL), 10 mM sodium decanoate (DEC), 0.1 mM paraquat (PQ), or 4 μg of tetracycline/ml (TET). The bars show the average values for triplicate assays. The relative expression is calculated as 2^−ΔCT, where ⁻ΔCT is the ratio of the crossing points target value to the crossing point reference value. The target is the strain indicated, whereas the reference is E. cloacae Jc194 in all cases.

Role of SoxS, RobA, and RamA regulating AcrAB-TolC efflux pump in response to different signals.

The soxS, robA, and ramA genes from EcDC64 were first amplified by high-fidelity PCR, cloned into the pCR-BluntII-TOPO cloning vector, and finally sequenced. Sequence analysis showed that genes soxS, robA, and ramA from E. cloacae EcDC64, which were 327, 870, and 375 bp long, respectively, encoded proteins containing 108, 289, and 124 amino acids, respectively. These proteins showed a high level of similarity to their homologues in other members of the Enterobacteriaceae family. When the sequences obtained were compared to the recently released genome of E. cloacae subsp. cloacae ATCC 13047 (GenBank accession code NC_014121), the amino acid identities were 100% for SoxS and 99% for RamA. RobA displayed 97% amino acid identity with the respective homologue in E. cloacae subsp. cloacae ATCC 13047. Moreover, the upstream region of these genes was amplified by using two consecutive random and specific primers, as described above. Fragments of various lengths (between 600 and 800 bp) were obtained. Analysis of the upstream region of soxS revealed the presence of the soxR gene. Assembly of the sequence of the entire locus revealed an organization similar to that described for other Enterobacteriaceae, with divergently transcribed soxR and soxS genes separated by an intergenic sequence of 98 bp. Amplification of the upstream region of ramA gene yielded an amplicon of ca. 700 bp containing the partial coding region of the romA gene.

In order to determine the different pathways activating the MDR acrAB-mediated phenotype, mRNA was extracted from E. cloacae Jc194 incubated in the presence of the compounds listed above, and the expression levels of the soxS, robA, and ramA genes were measured by real-time RT-PCR (Fig. 2). The results confirmed a high level of soxS expression under superoxide stress and showed that the sodium decanoate activates robA expression (Fig. 2). Furthermore, the ramA gene was overexpressed in E. cloacae Jc194 in the presence of tetracycline and salicylate (Fig. 2). Moreover, the expression level of marA gene was measured in the presence of salicylate and tetracycline, and the results obtained confirmed a high level of expression of marA gene by salycilate (data not shown).

Fig 2.

RT-PCR analysis of soxS, robA, and ramA gene expression in E. cloacae Jc194 in the presence of 10 mM sodium salicylate (SAL), 10 mM sodium decanoate (DEC), 0.1 mM paraquat (PQ), or 4 μg of tetracycline/ml (TET). The bars show the average values for triplicate assays. The relative expression was calculated as 2^−ΔCT, where ⁻ΔCT is the ratio of the crossing points target value to the crossing point reference value. The target is the strain indicated, whereas the reference is in all cases E. cloacae Jc194.

Induction of acrAB by SoxS, RobA, and RamA in E. cloacae Jc194.

To determine the effect of the regulatory proteins SoxS, RobA, and RamA on the expression level of acrAB, fluorescence data of E. cloacae Jc194 cultures containing the acrAB promoter fused to the gene encoding for GFP protein were measured and quantified by flow cytometry.

The WT strain Jc194 harboring the recombinant plasmid pGFP (Table 1) was used as an experimental control that showed basal fluorescence intensity (fluorescence value = 1). Each of the genes encoding regulatory proteins was cloned under the control of their own promoter in the pCR-BluntII-TOPO cloning vector and introduced into the control strain Jc194 harboring the recombinant plasmid pGFP (Table 1) in order to measure the fluorescence intensity emitted when each of the regulators was overexpressed.

The results obtained by flow cytometry experiments showed that the regulatory proteins SoxS, RobA, and RamA induced the acrAB promoter, thus leading to a significant increase in fluorescence intensity (Fig. 3A). To confirm the results, the mRNA levels of acrA and acrB genes were measured by RT-PCR. The results of the RT-PCR assays are consistent with the flow cytometry data (Fig. 3B).

Fig 3.

(A) Induction of acrAB operon due to ramA, soxS, and robA overexpression in E. cloacae Jc194 containing the reporter plasmid pGFP (see Materials and Methods), Jgfp. The values shown are fluorescence units relative to the Jgfp control. The bars show the average values from triplicate assays. P < 0. 05 in all cases. (B) RT-PCR analysis of acrA and acrB gene expression in E. cloacae Jc194 overexpressing ramA, soxS, and robA. The bars show the average values from triplicate assays. Relative expression is calculated as 2^−ΔCT, where ⁻ΔCT is the ratio of the crossing points target value to the crossing point reference value. The target is the indicated strains, whereas the reference is E. cloacae Jc194 in all cases.

Influence of SoxS, RobA, and RamA on acrAB expression in E. cloacae EcDC64.

In order to understand the role of the regulatory proteins SoxS, RobA and RamA in the MDR phenotype mediated by the AcrAB-TolC efflux pump in EcDC64 strain, soxS, robA, and ramA genes were inactivated in the wild-type EcDC64 strain to generate mutant strains EcΔsoxS, EcΔrobA, and EcΔramA. These knockout derivates were transformed with the recombinant plasmid (pGFP) carrying the transcriptional fusion of acrAB promoter-GFP gene. The WT EcDC64 strain was also transformed with pGFP to control for fluorescence intensity. Flow cytometry assays were performed with the WT strain and the knockout derivates harboring the pGFP recombinant plasmid, and fluorescence data were measured and recorded to investigate the effect of each regulator on the expression level of acrAB from E. cloacae EcDC64. The results obtained are shown in Fig. 4A. Inactivation of the robA gene, which is constitutively expressed (3), led to a significant reduction in the fluorescence intensity. The fluorescence intensity emitted by EcΔrobA was reduced by half relative to that observed in the WT strain EcDC64. However, when soxS and ramA genes were inactivated, the fluorescence intensity remained almost unchanged in comparison with that in the WT strain.

Fig 4.

(A) Level of expression of acrAB in knockout strains from E. cloacae EcDC64 and their complemented strains. The values shown are fluorescence units relative to the EcDC64 WT strain containing the reporter plasmid, pGFP. The bars show the average values from triplicate assays. *, P < 0.05. (B) RT-PCR analysis of acrA and acrB gene expression in knockout strains from E. cloacae EcDC64 and their complemented strains. The bars show the average values from triplicate assays. The relative expression is calculated as 2^−ΔCT, where ⁻ΔCT is the ratio of the crossing points target value to the crossing point reference value. The target is the indicated strains, whereas the reference is E. cloacae EcDC64 in all cases.

Complementation assays were performed to confirm the role of these regulatory proteins on acrAB expression. The soxS, robA, and ramA genes were amplified under the control of their own promoter regions from the WT EcDC64 strain and cloned into the pUCP24 vector (conferring resistance to gentamicin) in order to transform strains EcΔsoxS, EcΔrobA, and EcΔramA, respectively. The fluorescence activity of these cultures containing the pGFP recombinant plasmid was measured and quantified by flow cytometry. As expected, the fluorescence intensity was increased to various degrees when the regulatory genes were expressed from a plasmid into the knockout derivates (Fig. 4A). Real-time RT-PCR results confirmed the fluorescence data obtained by flow cytometry assays (Fig. 4B).

Role of the soxS, robA, and ramA regulatory proteins on the antimicrobial resistance profile of E. cloacae.

The resistance profile to several antibiotic families known as good substrates for efflux pumps (such as macrolides, tetracyclines, or quinolones) was determined in both strains by the standard disk diffusion method (Table 4). EcDC64 showed greater resistance or lower susceptibility to the antibiotics tested than strain Jc194, which is consistent with lower expression of AcrAB-TolC. Different approaches were used to determine the role of the regulatory proteins (SoxS, RobA, and RamA) on the efflux-mediated resistance profile.

First, the genes encoding regulatory proteins were overexpressed in E. cloacae isolate Jc194. The strains overexpressing soxS, robA, and ramA became resistant to erythromycin, tetracycline, and telithromycin and showed lower susceptibility to ciprofloxacin. The clinical isolate Jc194 was resistant to chloramphenicol when soxS and robA were overexpressed and showed reduced susceptibility to this antibiotic as a result of ramA overexpression. The strain harboring the empty cloning vector showed no significant changes in the resistance profile relative to the WT strain (Table 4).

On the other hand, the lack of the regulatory protein SoxS led to an increase in susceptibility to all antibiotics tested, except for chloramphenicol, where the inhibition zone determined by disk diffusion test remained unchanged relative to that observed for the WT strain. The MDR phenotype of isolate EcDC64 was restored or even increased for all antibiotics tested, except for tetracycline. The overexpression of soxS in EcΔsoxS led to partial restoration of the resistance to tetracycline by EcΔsoxS.

The same trend was observed in the isogenic EcΔrobA and EcΔramA isolates. The lack of RobA and RamA proteins had a similar effect on the resistance of EcDC64 isolate, with some exceptions (Table 4). For some antibiotics, such as tetracycline and ciprofloxacin, the increase in susceptibility was greater than that observed for EcΔsoxS. Analysis of knockout isogenic derivates EcΔrobA and EcΔramA overexpressing the robA and ramA genes, respectively, showed that the resistance profile was similar to that of the EcDC64 clinical isolate, although the MDR phenotype was not fully restored, as also observed in EcΔsoxS. Interestingly, the knockout derivates EcΔsoxS, EcΔrobA and EcΔramA containing the expression vector pUCP24 without any insertion were slightly more susceptible to the antibiotics than the other knockout derivates (Table 4). Moreover, susceptibility testing was performed with WT strains Jc194 and EcDC64 and the knockout derivates from EcDC64 (Table 3) in order to evaluate the effect of the lack of each regulator on the resistance profile in the presence of stressing agents.

In general, EcΔsoxS showed increased antibiotic resistance in response to the presence of salicylate, sodium decanoate and tetracycline. However, the superoxide stress generated by paraquat did not stimulate an increase in the resistance in soxS defective strain EcDC64. The antibiotic susceptibility also increased when EcΔsoxS was incubated with paraquat (Table 3). The concentration used in this case was 50 μM, which allowed the growth of EcΔsoxS and susceptibility testing to be performed. Analysis of antibiotic resistance profile of EcΔrobA mutant revealed that sodium decanoate, unlike salicylate, paraquat, and tetracycline, was not able to induce an increase in the antibiotic resistance of the robA defective EcDC64 clinical isolate.

Finally, tetracycline only induced an increase in the resistance to erythromycin by the isogenic knockout EcΔramA. Susceptibility to tetracycline was almost unchanged and increased susceptibility to telithromycin, ciprofloxacin and chloramphenicol was observed (Table 3). Although salicylate also induces overexpression of ramA, EcΔramA showed an increased antibiotic resistance in response to salicylate, due to the presence of a functional MarA regulator.

SoxS, RobA, and RamA bind to upstream region of acrA gene.

The aforementioned results indicate that SoxS, RobA, and RamA activators play a major role in inducing acrAB expression in response to the presence of compounds such as salicylate, superoxides, decanoate, or antibiotics such as tetracycline. To understand the regulation of acrAB by these regulatory proteins, EMSAs with SoxS, RobA, and RamA were performed. Plasmids encoding the histidine-tagged proteins were constructed as described in Materials and Methods. SoxS and RamA proteins were purified after being fused to MBP (maltose-binding protein) to enhance their solubility. The MBP was also purified for use as a control. The upstream region of acrA gene was amplified by PCR, and the fragment was incubated with each of the proteins. The three proteins—SoxS, RobA, and RamA—from the E. cloacae EcDC64 isolate bound to the promoter of the acrA gene, as revealed by the shift in the gel migration, whereas MBP did not bind to the promoter (Fig. 5).

Fig 5.

Electrophoretic mobility of the DNA fragments containing the EcDC64 upstream region of acrA gene from E. cloacae EcDC64. Lane 1, DIG-labeled DNA probe; lane 2, DIG-labeled DNA probe and SoxS protein fused to MBP; lane 3, DIG-labeled DNA probe and RobA protein; lane 4, DIG-labeled DNA probe and RamA protein fused to MBP; lane 5, DIG-labeled DNA probe and MBP protein.

DISCUSSION

Multidrug efflux pumps are the major agents conferring drug resistance in bacteria. Several investigators have studied the important roles of the AcrAB-TolC efflux pump in bacterial drug resistance and virulence in Enterobacteriaceae (28, 35). The data currently available in E. coli and other Enterobacteriaceae show that multidrug efflux pumps are often expressed under precise and elaborate transcriptional control, including specific regulators such as AcrR, and global regulators such as MarA, SoxS, and Rob. In E. coli, the transcriptional activators belonging to the AraC/XylS family interact with AcrAB and effectively enhance efflux (15). Because the AcrAB-TolC plays a predominant role in the intrinsic resistance of E. cloacae to a wide range of antibiotics, dyes, detergents, and solvents, as well as in bacterial fitness and virulence (33, 34), study of its regulation is of great importance in understanding the action of antibiotics and bactericidal agents in E. cloacae and related organisms. The mar, sox, and rob regulons are well-characterized regulatory systems in E. coli (4, 21, 31, 51). It is thought that the influx and efflux in E. cloacae might be regulated in a similar way as in E. coli. Genomic analysis of E. cloacae EcDC64 revealed the presence of SoxS and RobA regulators, essentially identical to those in E. coli. The ramA regulon was found to show the same genetic organization as that observed in Citrobacter (41) and Klebsiella (5) species, in which the ramA locus is composed by romA and ramA genes, both controlled by the RamR repressor. The DNA-binding domains of SoxS, RobA, and RamA proteins from EcDC64 share a high level of sequence identity, which suggests that these proteins have overlapping specificity. These regulators activate the transcription of a large set of promoters, including the promoter of acrAB, through the direct binding to a degenerated and asymmetrical DNA sequence, known as the marbox sequence (24, 27). Analysis of the upstream region of the acrAB operon of E. cloacae (of both clinical isolates used in the present study) revealed a marbox sequence in a suitable position. The results presented here demonstrate the importance of SoxS, RobA, and RamA regulators in the modulation of AcrAB-TolC-mediated antibiotic resistance in E. cloacae.

Functional genomics approaches such as knockout studies or heterologous expression studies have been used to assess the effect of these transcriptional activators on acrAB expression. The soxS, robA, and ramA knockout strains derived from MDR E. cloacae EcDC64 were constructed. The soxS, robA, and ramA genes were also cloned into plasmids and overexpressed in the susceptible strain E. cloacae Jc194 under the control of their own promoter regions. The expression level of acrAB was measured by two different methods: flow cytometry using the acrAB promoter-GFP gene fusion as a reporter and real-time RT-PCR. In addition, susceptibility testing was performed with all of the different strains obtained (in which soxS, robA, and ramA were inactivated and overexpressed). Transcriptional activation of acrAB is the predominant cause of multidrug resistance in strains that overexpress MarA or the closely related global regulators SoxS, RobA, and RamA (6, 9, 12, 19, 20, 22, 32, 40, 43, 47, 51, 52). Chollet et al. (9) have described how an increase in ramA-mediated acrAB expression leads to an increase in resistance to tetracycline, chloramphenicol, quinolones, and β-lactams in E. aerogenes. In the same way, Hornsey et al. (17) described the emergence of AcrAB-mediated tigecycline resistance in E. cloacae through the RamA regulator. Our findings demonstrated that increased levels of soxS, robA, and ramA transcription activated acrAB expression, which led to an increase in antibiotic resistance in E. cloacae Jc194. The data indicated a correlation between the increased resistance to macrolides, tetracycline, ketolides, fluoroquinolones, and chloramphenicol and the expression levels of acrAB and soxS, robA, and ramA. The results of DNA mobility shift assays indicated that each of these three proteins directly control the expression of acrAB through binding to its upstream promoter region. Overall, the data indicated that SoxS, RobA, and RamA activators play a major role as activators of AcrAB-TolC expression. Overexpression of ramA caused the highest acrAB gene expression, followed by overexpression of SoxS and, to a lesser extent, RobA.

The differences in the acrAB expression levels were also illustrated by quinolone resistance in Jc194. The effects of RobA were generally weaker than those of RamA and SoxS for both antibiotic resistance and gene expression. In contrast to SoxS and RamA, RobA expression is constitutive (46), and it is inactive until activated by an induction signal (40, 42), which may explain the greater effect of the overexpressed soxS and ramA genes in the activation of the AcrAB-TolC efflux pump. On the other hand, the inactivation of the regulatory genes soxS, robA, and ramA in E. cloacae ECDC64 led to increased susceptibility to all antibiotics tested, except for chloramphenicol, in comparison to that of the WT strain, showing that the knockout acquires a similar antibiotic resistance profile to drug-susceptible strain Jc194. However, only inactivation of the robA gene had a slight effect on acrAB expression, which was reduced by half. These data, supported by the results obtained in the susceptibility analysis, suggest that RobA plays an important role in the antibiotic resistance of E. cloacae EcDC64. Although the inactivation of soxS and ramA did not affect the expression of acrAB, these proteins also affected the antibiotic resistance of isolate EcDC64. This may be due to the level of expression of soxS and ramA genes, which was significantly lower than that of the robA gene. Therefore, the effect of the inactivation of soxS and ramA on acrAB expression goes unnoticed when RobA is functional. In addition, these regulatory proteins control many other genes involved in antimicrobial resistance, which may explain the AcrAB-independent modification in the resistance profile of EcDC64 (27).

In E. coli, AcrAB-TolC is upregulated in response to different signals, such as aromatic weak acids (salicylate), superoxides (generated by paraquat), bile salts, fatty acids (decanoate), and tetracycline (49). These toxic compounds activate the transcription of global regulators, and they cause upregulation of the AcrAB-TolC efflux pump (11, 40, 45, 48). In the present study, we examined whether a similar mechanism occurs in E. cloacae. Indeed, the AcrAB-TolC system was upregulated by salicylate, decanoate, paraquat, and tetracycline, which affected the resistance profile of E. cloacae, as observed by a decreased susceptibility to a number of antibiotics. The results demonstrated that the induction profiles in E. cloacae are very similar to the induction profiles previously described in other species belonging to the Enterobacteriaceae family. The increased expression of acrAB was significantly correlated with the degree of resistance shown by E. cloacae Jc194 to different antibiotics, with some exceptions. Induction with tetracycline triggered the largest increase in resistance to all antibiotics tested. The lowest effect was observed after incubation of strain Jc194 with decanoate. The resistance to chloramphenicol was not modified in response to this compound. The increased levels of resistance to erythromycin, telithromycin, and ciprofloxacin, which were shown by Jc194 in response to salicylate, paraquat, and tetracycline, were similar, although the tetracycline had a greater effect on tetracycline and chloramphenicol resistance. Bacteria must distinguish among different stress conditions and respond in an appropriate manner. Each activator is regulated in response to a different signal, causing the final induction of acrAB. As described above, the AcrAB-TolC efflux pump in E. cloacae is upregulated by SoxS, RobA, and RamA, causing an efflux-mediated MDR phenotype in E. cloacae. The present data revealed that acrAB from E. cloacae is activated through different pathways depending on the stimulus applied. Salicylate induced ramA expression, in addition to activating marA transcription, as previously described in E. coli (45). soxS and robA expression was also moderately increased by salicylate in E. cloacae, unlike in Salmonella, in which robA is downregulated (16). The expression data supported by the susceptibility testing results suggest that salicylate induces activation of the AcrAB-TolC efflux pump, mainly through the MarA regulator. In E. coli, tetracycline induces marA, soxS, and robA expression, although the intensity of the effect differs (48). Tetracycline-mediated similar adaptation mechanisms of increased efflux in E. cloacae but, unlike E. coli, it activated ramA expression instead of marA. The effect of tetracycline on ramA expression was greater than the effect on soxS and robA expression, suggesting that the RamA activator plays a key role in the AcrAB regulation network. This hypothesis was supported by the antibiotic resistance profile shown by the ramA-defective E. cloacae EcDC64 isolate, in which antibiotic resistance was not induced by tetracycline. Furthermore, E. cloacae showed an oxidative stress response similar to that described in E. coli, in which soxS is highly overexpressed and robA is downregulated by paraquat (29), showing an increased resistance profile. However, paraquat increased the antibiotic susceptibility of soxS-defective E. cloacae EcDC64, which confirms the involvement of SoxS in the regulation of AcrAB under oxidative stress conditions. Rosenberg et al. (40) found that the induction of acrAB by decanoate requires Rob, but not MarA or SoxS, in E. coli. However, decanoate downregulated robA transcription and activated marA expression in Salmonella enterica serovar Typhimurium (16). We found a model for the acrAB regulatory network in E. cloacae similar to that in E. coli. Decanoate was not able to induce an increase in the antibiotic resistance of robA-defective EcDC64. Moreover, only RobA was upregulated by decanoate, which suggests that it mediates the regulation of AcrAB by free fatty acids such as decanoate, and it apparently helps the survival of E. cloacae by making it more resistant to antimicrobial agents.

In summary, the results presented here show that the AraC/XylS regulators SoxS, RobA, and RamA play an important role in efflux-mediated multidrug resistance in E. cloacae by increasing acrAB expression. It was further demonstrated that each activator is regulated in response to a different signal, causing the final induction of acrAB. Therefore, the AcrAB-TolC efflux pump in E. cloacae is activated through different pathways depending on the stimulus applied, thus conferring resistance to a variety of antimicrobial agents.