JOURNAL OF BACTERIOLOGY, Apr. 2007, p. 2996–3005 Vol. 189, No. 80021-9193/07/$08.00�0 doi:10.1128/JB.01819-06Copyright © 2007, American Society for Microbiology. All Rights Reserved.

The Transcriptional Regulators NorG and MgrA Modulate Resistanceto both Quinolones and �-Lactams in Staphylococcus aureus�

Que Chi Truong-Bolduc and David C. Hooper*Division of Infectious Diseases and Medical Services, Massachusetts General Hospital,

Harvard Medical School, Boston, Massachusetts 02114-2696

Received 4 December 2006/Accepted 25 January 2007

MgrA is a known regulator of the expression of several multidrug transporters in Staphylococcus aureus.We identified another regulator of multiple efflux pumps, NorG, by its ability, like that of MgrA, to bindspecifically to the promoter of the gene encoding the NorA efflux pump. NorG is a member of the familyof the GntR-like transcriptional regulators, and it binds specifically to the putative promoters of the genesencoding multidrug efflux pumps NorA, NorB, NorC, and AbcA. Overexpression of norG produces athreefold increase in norB transcripts associated with a fourfold increase in the level of resistance toquinolones. In contrast, disruption of norG produces no change in the level of transcripts of norA, norB,and norC but causes an increase of at least threefold in the transcript level of abcA, associated with afourfold increase in resistance to methicillin, cefotaxime, penicillin G, and nafcillin. Overexpression ofcloned abcA caused an 8- to 128-fold increase in the level of resistance to all four �-lactam antibiotics.Furthermore, MgrA and NorG have opposite effects on norB and abcA expression. MgrA acts as an indirectrepressor for norB and a direct activator for abcA, whereas NorG acts as a direct activator for norB anda direct repressor for abcA.

MgrA, a multifunctional MarR-like regulator, was first iden-tified by its ability to bind directly to the norA promoter, lead-ing to altered expression of norA, which encodes the NorAefflux pump in Staphylococcus aureus, an important pathogenresponsible for infections in hospital and community settings(28). Infections caused by S. aureus can be difficult to treatbecause of resistance to multiple antibiotics and multiple vir-ulence factors (2). One of the most interesting mechanismsof resistance in S. aureus is its ability to express severalmultidrug resistance efflux pumps, which constitute a majordefense against diverse families of toxins and antimicrobialagents (17, 18).

The complete genome of S. aureus N315 is 2.81 Mb inlength, containing genes predicted to encode 210 transporters.Sixty-seven of these transporters are predicted to be ATP de-pendent, representing 31% of the total. The majority of theATP-dependent transporters belong to the ABC family (63ABC transporters, or 94% of the total of 67). Among the 114(54.3%) secondary transporters energized by ion gradientsacross the membrane, there are 28 efflux pumps (24.6%) be-longing to the major facilitator superfamily (MFS) and fewerthat are members of other families. A similar distribution isfound in the genomes of S. aureus strains COL, Mu50, andNCTC 8325 (11, 19). NorB, NorC, and Tet38 are three newadditions to the MFS of transporters in S. aureus, and recentlyMepA, a multidrug resistance pump belonging to the multi-drug and toxin extrusion family, was identified (9, 26, 27).Efflux pumps can extrude a specific class of antibiotics such astetracyclines (TetK, TetL, and Tet38) or macrolides (MsrA) or

can extrude diverse unrelated compounds, such as quinolones,ethidium bromide, and cetrimide (NorA, NorB, and MepA) (1,9, 22, 26).

AbcA is an ATP-dependent transporter of the ABC family,members of which use the energy liberated by ATP hydrolysisrather than the energy generated by transmembrane ion gra-dients, which is used by the members of the MFS to extrudetheir substrates (3). AbcA was shown to participate in cell wallautolysis, but no relation was established between its overex-pression and resistance to �-lactam antibiotics (4, 6, 24). Thistransporter shares an overlapping promoter region with thestructural gene (pbpD) encoding the PBP4 protein, a transpep-tidase/carboxypeptidase, which is involved in cell wall synthesisand confers a decrease in sensitivity to �-lactam drugs (4). Theexpression of abcA and that of pbpD, however, appear to beindependent of each other and to require different regulatoryfactors. The transcription of abcA depends on the agr regula-tory system (24).

MgrA affects resistance to antibiotics by controlling the ex-pression of at least four efflux pumps, NorA, NorB, NorC, andTet38, which are responsible for decreases in susceptibility tohydrophilic (norfloxacin and ciprofloxacin) and hydrophobic(moxifloxacin and sparfloxacin) quinolones, tetracycline, andchemical compounds (ethidium bromide, cetrimide, and tetra-phenylphosphonium [TPP]) (26–28). In addition to modulatingthe expression of efflux transporters, MgrA also regulates au-tolytic activity and the expression of several virulence factors,including alpha-toxin, nuclease, protein A, and capsular poly-saccharides (7, 8, 14).

In this report, we have identified and characterized an ad-ditional regulatory factor, NorG, a new member of the GntR(gluconate regulatory protein) family that regulates expressionof the NorB and AbcA efflux pumps and affects resistance toboth quinolones and �-lactam antimicrobial agents. We have

* Corresponding author. Mailing address: Division of InfectiousDiseases, Massachusetts General Hospital, 55 Fruit Street, Boston MA02114-2696. Phone: (617) 726-3812. Fax: (617) 726-7416. E-mail:dhooper@partners.org.

� Published ahead of print on 2 February 2007.

2996

on August 20, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

further identified AbcA as a transporter that can confer resis-tance to �-lactams.

MATERIALS AND METHODS

Bacterial strains, plasmids, growth media, and other materials. Bacterialstrains and plasmids used in this study are listed in Table 1. S. aureus strains werecultivated in brain heart infusion (BHI) broth (Difco, Sparks, MD) at 37°C unlessotherwise stated. Escherichia coli strains were grown in Luria-Bertani (LB) me-dium. Lysostaphin was obtained from AMBI Products Inc., New York, NY;ciprofloxacin and moxifloxacin from Bayer Corp., Westhaven, CT; sparfloxacinfrom Parke-Davis Pharmaceutical Research Division, Ann Arbor, MI; and 2�-(4-ethoxyphenyl)-5-(4-methyl-1-piperazinyl)-2,5�-bi-1H-benzimidazole (Hoechst33342), norfloxacin, ethidium bromide, cetrimide, tetracycline, TPP, rhodamine,nafcillin, methicillin, penicillin G, cefotaxime, and chloramphenicol from SigmaChemical Co., St. Louis, MO. All primers used in this study were synthesized atthe Tufts University Core Facility, Boston, MA.

MIC determinations. MICs of quinolones, �-lactams, TPP, rhodamine,ethidium bromide, cetrimide, and Hoechst 33342 were carried out by serial agardilution on BHI agar. All plates were incubated at 30°C or 37°C for 24 h beforereading. Determinations of MICs of quinolones, �-lactams, and other chemicalcompounds for transformants containing pQT13 or pQT14 were done on BHIagar containing 5 �g tetracycline per ml to ensure maintenance of the plasmid,with incubation at 30°C.

RNA analysis. Total S. aureus RNA was prepared by extraction from lyso-staphin-treated cells grown to exponential phase at 37°C or 30°C, using theRNeasy minikit (QIAGEN, Valencia, CA). The concentration of RNA wasdetermined spectrophotometrically as the absorbance at 260 nm. For Northernblot analysis, 10 �g of total RNA was electrophoresed through a 0.9% agarose–0.66 M formaldehyde gel in morpholinepropanesulfonic acid (MOPS) and blot-ted onto Hybond-N� membranes as previously described (26, 28). DNA probeswere amplified from the ISP794 chromosome and labeled with psoralen for thedetection of specific transcripts, using the Northern maxi kit (Ambion, Inc.,Austin, TX) as recommended by the manufacturer. Blots were hybridized withprobes overnight at 42°C, washed, and autoradiographed with Kodak X-Omatfilm. The reverse transcription-PCR (RT-PCR) analyses were performed usingthe SuperScript one-step RT-PCR kit (Invitrogen Inc.) with 10 picograms of totalRNA as the template. Primers for norB (5�-GAAGATAGTTTCAATACAGA-3� and 5�-ATTATAAATGATAGGATGAA-3�) generated a 370-bp amplicon.The running conditions were 1 cycle for 30 min at 45°C; 1 cycle for 2 min at 94°C;30 cycles for 45 s at 94°C, 45 s at 48°C, and 30 s at 72°C; and 1 cycle for 10 minat 72°C. The 16S rRNA was used as an internal control to normalize the RT-PCRdata as described previously (5, 27).

Cloning and overexpression of norG. To clone the norG gene, primers basedon flanking sequences (NCTC8325, Oklahoma University) were synthesized bythe Tufts University Core Facility (Boston, MA). A 1,321-bp fragment wasamplified by PCR from S. aureus ISP794 chromosomal DNA with sense primer

5�-ATGGACAGCTGATGAAGATA-3�) (the PstI site is underlined) and anti-sense primer 5�-CGAATTAGAATTCTTGTTTTAA-3� (the EcoRI site is un-derlined), which generated flanking PstI and EcoRI sites, respectively. Theamplified norG gene was digested with PstI and EcoRI, ligated into the PstI andEcoRI sites of the plasmid pGEM3-zf(�) to yield pGEM3-zf(�)-norG, andintroduced into E. coli DH5�. Plasmids extracted from ampicillin-resistant col-onies were screened for the norG fragment insertion by restriction endonucleasedigest patterns and confirmed by DNA sequencing.

To generate a plasmid for overexpression of norG in S. aureus, the norG genewas amplified by PCR from S. aureus ISP794 chromosomal DNA with senseprimer 5�-ATGGAGGATCCATGAAGATA-3� (the BamHI site is underlined)and antisense primer 5�-CGAATTAGAATTCTTGTTTTAA-3� (the EcoRI siteis underlined), which generated flanking BamHI and EcoRI sites, respectively.The amplified norG gene was digested with BamHI and EcoRI and ligated intothe BamHI and EcoRI sites of the temperature-sensitive shuttle plasmid pSK950to yield pQT13. This plasmid was then electroporated into S. aureus RN4220(8325 r�) to generate transformants, and the structure of pQT13 in S. aureus wasconfirmed by restriction mapping. Electrocompetent ISP794 was then trans-formed with this plasmid isolated from RN4220. Tetracycline-resistant coloniesisolated at 30°C were confirmed to have intact pQT13 by restriction mapping.

Construction of an abcA overexpressor. The abcA gene was amplified by PCRfrom S. aureus ISP794 chromosomal DNA with sense primer 5�-GGATCCTTAATCTGTTAATTTTTGA-3� (the BamHI site is underlined) and antisenseprimer 5�-GAATTCATGAAACGAGAAAATCCAT-3� (the EcoRI site is un-derlined). The amplified abcA gene was digested with BamHI and EcoRI andligated into the BamHI and EcoRI sites of the temperature-sensitive shuttleplasmid pSK950 to yield pQT14. This plasmid was electroporated into S. aureusRN4220 (8325 r�), reextracted, and then introduced into ISP794 by electropo-ration. Tetracycline-resistant colonies isolated at 30°C were confirmed to haveintact pQT14 by restriction mapping.

Construction of a norG mutant by allelic exchange. The 800-bp DNA fragmentcontaining the cat gene was amplified from plasmid pLI50 using primers catpvu1and catpvu2 (23, 28). The PCR product was digested with PvuII and then ligatedinto an EcoRV site within the putative norG coding region of plasmid pGEM-3zf(�)-norG. The resultant plasmid containing the 2.1-kb norG::cat was sub-cloned into the temperature-sensitive shuttle plasmid pCL52.2 to yieldpCL52.2-(norG::cat). The allelic exchange procedure was then carried out asdescribed previously(28). pCL52.2-(norG::cat) was first introduced into RN4220by electroporation, and chloramphenicol-resistant (5 �g/ml) colonies of RN4220were grown at 30°C in the presence of 5 �g/ml tetracycline and used for reiso-lation of pCL52.2-(norG::cat), which was then electroporated into ISP794.ISP794 harboring pCL52.2-(norG::cat) was grown in BHI broth with tetracycline(3 �g/ml) at 30°C, diluted 1:1,000 in fresh medium, and propagated at 42°C for24 h. The culture was diluted and grown again at 30°C without selection for 48 h.Chloramphenicol-resistant, tetracycline-sensitive colonies, representing possibledouble-crossover events, were tested for cat insertion into norG by PCR andsequencing. To construct the mgrA::cat norG::cat double mutant, we carried outa second allelic exchange using the same plasmid construct pCL52.2-(norG::cat)and QT1 (mgrA::cat) as the recipient. Since QT1 already had one chromosomalcopy of the cat gene, we increased the chloramphenicol concentration to 10�g/ml for the selection of the double mutant. DNA sequencing was performed toconfirm the presence of the chromosomal insertion of the norG:: cat andmgrA::cat genes.

DNA mobility shift analysis. Primers designed to amplify the putative pro-moter regions of norA, norB, norC, norG, abcA, pbpD, tet38, and mgrA are listedin Table 2. One of the primers was biotinylated at the Tufts University CoreFacility (Boston, MA). The gel mobility shift assay was carried out using theLightShift chemiluminescent EMSA kit (Pierce, Rockford, IL), as recommendedby the manufacturer. The biotin-labeled DNA was incubated with the indicatedamount of cell extract or purified proteins from S. aureus in 20 �l of bindingbuffer (10 mM HEPES [pH 8], 60 mM KCl, 4 mM MgCl2, 0.1 mM EDTA, 0.1mg/ml of bovine serum albumin, 0.25 mM dithiothreitol) containing 1 �g ofpoly(dI-dC), 200 ng of sheared herring sperm DNA, and 10% glycerol. Thereaction mixture was incubated for 20 min at room temperature and analyzed by5% nondenaturing polyacrylamide gel electrophoresis (PAGE). For the compe-tition assays, a 100-fold excess of specific or nonspecific unlabeled DNA wasadded to the reaction mixture prior to the incubation.

Identification of the NorG protein from cell extracts. Cell extracts collectedfrom 1 liter of overnight culture of S. aureus QT1 were used to purify the NorGprotein, as previously described for purification of MgrA protein (28). The150-bp biotinylated norA DNA fragment was immobilized on magnetic beadswith covalently coupled streptavidin (Dynabeads M-280; Dynal) according to themanufacturer’s protocol. DNA bound to beads was incubated with protein ex-

TABLE 1. Bacterial strains and plasmids used in this study

Strain orplasmid Genotype or relevant characteristic(s) Reference

or source

S. aureusRN4220 8325-4 r� 10ISP794 8325-4 pig-131 25QT1 ISP794 mgrA::cat 28QT11 ISP794 norG::cat This studyQT12 ISP794 mgrA::cat norG::cat This study

E. coli DH5� F� �80dlacZ�M15 �(lacZYA-argF)U169deoR recA1 endA1 phoA hsdR17(rK

�

mK�) supE44� thi-1 gyrA96 relA1

Gibco-BRL

PlasmidspGEM3-zf(�) 2.9-kb E. coli cloning vector; Apr PromegapLI50 Shuttle cloning vector (Apr Cmr) 23pCL52.2 Temperature-sensitive E. coli-S. aureus

shuttle vector23

pSK950 10.5-kb plasmid carrying the attP siteof phage L54a, replicon of pE194;Tcr Emr (S. aureus)

15

pQT13 pSK950-norG This studypQT14 pSK950-abcA This study

VOL. 189, 2007 NorG REGULATORY PROTEIN 2997

on August 20, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

tract in binding buffer containing herring sperm DNA (200 ng) for 20 min atroom temperature. Beads were washed twice with binding buffer containingherring DNA and twice with binding buffer without DNA. Proteins were theneluted in binding buffer containing 0.5 M NaCl. Eluted proteins were dialyzedagainst water, concentrated, and separated by sodium dodecyl sulfate (SDS)-PAGE. The 47-kDa protein was blotted onto a polyvinylidene difluoride mem-brane for N-terminal amino acid sequencing by the Edman degradation method(Tufts Core Facility, Boston, MA).

Purification of NorG protein. The norG gene was subcloned into the plasmidpTrcHisA (Invitrogen, Carlsbad, CA) to yield pTrcHisA-norG and then intro-duced into E. coli BL21. The purification of histidine-tagged NorG was carriedout as recommended by the manufacturer. E. coli BL21 harboring pTrcHisA-norG was grown to mid-log phase in LB medium, at which time IPTG (isopropyl-�-D-thiogalactopyranoside) (1 mM) was added to the culture. After 3 h, the cellswere harvested by centrifugation and then resuspended in 20 mM sodium phos-phate buffer, pH 7.4. The cells were lysed with lysozyme (0.02%) and thencentrifuged (100,000 g) for 90 min. The supernatant was applied to a nickelaffinity column (iminodiacetic acid-Sepharose-Ni) (Amersham Pharmacia Bio-tech, Uppsala, Sweden) and then washed with buffer (20 mM Tris-HCl [pH 7.4],150 mM NaCl, 5% glycerol) supplemented with concentrations of imidazoleincreasing from 10 to 60 mM. NorG protein was eluted with 100 mM imidazole.The homogeneity of the eluted protein was verified by SDS-PAGE.

RESULTS

Identification of norG. We determined the presence of aprotein in addition to MgrA that binds to the norA gene pro-moter, by the pattern of gel shift of a 150-bp DNA fragmentcontaining the norA promoter DNA, with and without incuba-tion with cell extracts of strain QT1 (mgrA), which lacks MgrA.Using magnetic beads coupled to the 150-bp DNA fragmentcontaining the norA promoter as an affinity agent, we incu-bated cell extracts of QT1 and eluted a single protein that wasbound, in a manner similar to that that led to the isolation ofMgrA from wild-type cell extracts (28). After separation bySDS-PAGE, the protein band was transferred to a polyvinyl-idene difluoride membrane, and N-terminal amino acid se-quencing identified the sequence KIPPQRQLATQY, whichmatched with that encoded by a 1,321-bp open reading frame(ORF) designated SA0104 in the genome of S. aureus N315.This ORF is predicted to encode a protein of the FadR sub-family of the GntR-like family of regulators. Based on its role

in regulation of resistance to quinolones and �-lactams asoutlined below, we named this ORF norG (Fig. 1).

NorG binds to efflux pump promoters. We cloned the norGgene into plasmid pTrcHisA, a His tag expression vector. Afterinduction with IPTG (1 mM) and purification using nickelaffinity chromatography, we isolated a protein of �50 kDa (47kDa predicted for native NorG plus 3 kDa for the His tagregion) to �95% apparent homogeneity by SDS-PAGE (datanot shown). The purified protein was then used to perform gelshift assays of DNA fragments containing the putative promot-ers of several genes known to encode efflux pumps.

After incubation of NorG with the 150-bp DNA fragmentcontaining the norA promoter, a clear shift was shown in theDNA mobility pattern on agarose gels, a shift that was abol-ished in the presence of a 100-fold excess of unlabeled norApromoter DNA but not with a 100-fold excess of herring DNA.These data indicated that NorG bound specifically to the norApromoter fragment, as expected based on the affinity purifica-tion procedure used for its identification (Fig. 2A).

Similar experiments were carried out using DNA fragmentscontaining the putative promoters of norB, norC, and tet38,which encode efflux pumps, as well as the putative promoter ofnorG itself. In addition, the two identified promoters of mgrAwere amplified together on a 200-bp fragment (8). As previ-ously reported, norB may have two promoters, P1 and P2, withP2 located upstream of the ORF SA1272, encoding a putativealanine dehydrogenase, and P1 directly upstream of norB itself(26) (Fig. 2C).

Purified NorG-mediated DNA fragment shifts were foundwith both the P1 and P2 putative promoters of norB, shown asseparate DNA fragments in gel shift assays (Table 2 and Fig.2B and C), since P1 and P2 were separated by three ORFstotaling approximately 5 kb. Similar promoter DNA band shiftpatterns were found associated with the promoters of norC andnorG, but no change in DNA fragment mobility was detectedwith the promoters of tet38 and mgrA. Interestingly, purifiedMgrA protein also caused a shift in mobility of the putative

TABLE 2. Primers used to amplify putative promoter regions

PromoterPrimer DNA

length (bp)Orientation Sequence (5�33�)

norA Sense TGCAATTTCATATGATCAATCCC 150Antisense AGATTGCAATTCATGCTAAATATT

norB P1 Sense ATAAGGTAAGATAACTAGCA 150Antisense ATCTCTATTTGCCTCCCTATA

norB P2 Sense ACTCCTCTTTCATTTGTGAA 150Antisense GAGGTAAACTAAATATCAAC

norC Sense GCAGCTGTGATGGTGA 150Antisense ATTTCATTCATGTTAGTTA

norG Sense ATCACTCCTTTAATTATGT 150Antisense ATTCCATAAAACCAATCC

mgrA (P1 and P2)a Sense CGAATTCATTCATGATTT 200Antisense AAAGTTGATTGTTTATTAA

abcA Sense CACCTTAATTAAATTGTTGA 138Antisense CTCAAAGTCTTTATATTAAA

pbpD Sense TGTTTGTACGAAGAGCAA 138Antisense CTATCATATCATATACTTTC

tet38 Sense TTGGATGCGTATGGGTATTT 150Antisense TAATCATCTACACCAATGAC

a The P1 and P2 promoters of the mgrA gene are both located on the same 200-bp DNA fragment.

2998 TRUONG-BOLDUC AND HOOPER J. BACTERIOL.

on August 20, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

norG promoter fragment, suggesting that MgrA could affectexpression of norG (Fig. 3A). The DNA fragment shifts wereeach shown to be specific by competition experiments using100-fold excesses of specific and nonspecific unlabeled DNAs(data not shown), as was done to establish the specificity of theinteraction of NorG with the norA promoter.

The abcA gene, which encodes a putative ABC transporter,is transcribed divergently from pbpD, the structural gene en-coding the penicillin-binding protein PBP4. The publishedoverlapping promoter region abcA-pbpD as well as the specificregions harboring either abcA or pbpD promoters were ampli-fied and labeled with biotin and used in the gel mobility shiftassays. The inverted repeat that was shown previously to affectthe expression of abcA and pbpD was present in both DNAfragments (4) (Table 2 and Fig. 3C). NorG bound to the abcApromoter and the overlapping region, but no binding to theputative pbpD promoter was detected. MgrA also bound to theabcA promoter and the overlapping region but not to the pbpDpromoter (Fig. 3B). We performed the competition experi-ments using 100-fold excesses of specific and nonspecific un-labeled DNAs, which demonstrated the specificity of the in-

teraction of NorG and MgrA with the abcA promoter (data notshown).

Phenotype of norG mutants. We constructed a norG knock-out mutant, QT11, by allelic exchange of a disrupted copy ofnorG::cat for the wild-type copy of norG on the chromosome ofstrain ISP794 (28). We also constructed a norG mgrA doublemutant, QT12, using the same allelic exchange procedure withstrain QT1 (mgrA::cat). QT12 carried two copies of the catgene on its chromosome and was able to be constructed by itsability to grow in the presence of 10 �g chloramphenicol perml, a concentration at which neither QT1 nor QT11 grew. Thegrowth curves of the two mutants QT11 and QT12 are similarto that of the wild-type ISP794 (data not shown). We deter-mined the MICs of quinolones, �-lactams, and dyes forISP794, QT1, QT11, and QT12. The MICs of the quinolones(norfloxacin, ciprofloxacin, moxifloxacin, and sparfloxacin)showed no change for mutant QT11 (norG::cat) relative to itsparent strain. Mutant QT12 (norG::cat mgrA::cat), however,showed an increase of fourfold for the four quinolones testedcompared to those of the wild-type ISP794 but twofold lessthan those of the single mutant QT1 (mgrA). Thus, intact norG

FIG. 1. Nucleotide and amino acid sequences of the 1,321 bp of S. aureus DNA containing the norG gene from ISP794 (complete sequenceshown). The �35 and �10 sequences of the putative promoter region and the putative helix-turn-helix (H-T-H) region are shown. The codingregion of norG, preceded by a putative ribosome-binding site (RBS), is demarcated by the ATG start codon and the TAA stop codon.

VOL. 189, 2007 NorG REGULATORY PROTEIN 2999

on August 20, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

is needed for the full effect of mutation in mgrA on resistanceto quinolones.

We also determined the MICs of four �-lactams (nafcillin,penicillin G, methicillin, and cefotaxime) and five dyes (cet-rimide, TPP, rhodamine, ethidium bromide, and Hoechst33342) for all strains. Interestingly, QT11 (norG) and QT12(norG mgrA), respectively, showed fourfold and twofold in-creases in the MICs of the four �-lactams, while QT1 (mgrA)was twofold more sensitive to these �-lactams than the wild-type strain ISP794. Both QT11 and QT12 showed increases inMICs of dyes, but QT12 was twofold more resistant thanQT11. The MIC data are summarized in Table 3.

Alterations in gene expression in norG mutants. We carriedout Northern blotting using RNAs extracted from ISP794,QT1, QT11, and QT12, which were hybridized with biotin-

labeled probes generated from 200-bp internal regions of ei-ther abcA or pbpD (encoding PBP4). The level of abcA tran-scripts from QT11 was at least threefold greater than that fromQT12 and was almost undetectable for ISP794 and QT1 (Fig.4A). In contrast, no increases in the level of transcription ofpbpD in QT11 and QT12 were detected (data not shown). Inorder to confirm these findings, RT-PCRs were carried outusing the same primers designed for amplification of the twoprobe DNAs used for Northern hybridizations. We found asimilar increase (threefold) in the intensity of the abcA ampli-con for QT11 compared to that of QT12, confirming the resultsof the Northern blot assays (data not shown).

Because the low copy number of the norA, norB, and norCtranscripts in the wild-type strain ISP794 made it difficult toperform Northern blotting, we performed RT-PCR assays to

FIG. 2. (A) Gel mobility shift analyses of the interactions of the crude cell extracts (CE) from ISP794 and QT1 and the purified NorG proteinwith the biotinylated 150-bp norA promoter fragment. Competing unlabeled herring sperm DNA (nonspecific) and norA promoter DNA (specific)were used to determine the specificity of promoter binding. The amount of labeled DNA used was 2 ng per reaction. The amount of protein usedwas 50 ng per reaction for the NorG protein and 100 ng per reaction for the crude cell extracts. (B) Gel mobility shift analyses of the interactionsof the crude cell extracts from ISP794 and the purified NorG protein with the biotinylated 150-bp norB P1 and P2 promoter fragments.(C) Schematic representation of the positions of norB and the three adjacent ORFs on the S. aureus N315 published genome (11). The two putativepromoters and the putative rho-dependent terminator are indicated.

3000 TRUONG-BOLDUC AND HOOPER J. BACTERIOL.

on August 20, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

detect whether any differences occurred in transcript levelsbetween strains. The 16S rRNA was used as an internal controlas described previously (5, 27). RT-PCRs were carried out fornorA, norB, and norC in ISP794 (parent strain), QT11, andQT12. An increase of twofold was found associated with norBand norC amplicons in QT12 (norG mgrA), an increase thatwas similar to that with the mgrA mutation alone (26, 27). Incontrast, there was no change in the levels of these transcriptsfor QT11 (norG) relative to those for the wild-type strain. Nochange in levels of norA mRNA was detected in either mutant(data not shown).

Alterations in gene expression and phenotype by overex-pression of norG. We cloned the 1,321-bp norG gene intoplasmid pSK950 to generate pQT13. This plasmid was intro-duced first into RN4220 and then into ISP794 to study theeffects of overproduction of norG. As expected, in Northernblots there was an increase (threefold) in the level of norG

transcripts in ISP794(pQT13) compared to ISP794 (Fig. 4A).In complementation experiments pQT13 was introduced intomutant QT11 (norG::cat), resulting in an increase of norGmRNA and a decrease in abcA mRNA. RT-PCR assays usingspecific primers of an internal region of 200 bp of the norA,norB, and norC genes showed a 2.5-fold increase in norB tran-script levels in ISP794(pQT13) compared to ISP794 (Fig. 4B).In contrast, no differences in the levels of norA and norCtranscripts were detected between the two strains.

We then determined the MICs of quinolones, �-lactams,and dyes for ISP794 and ISP794(pQT13). The MICs of nor-floxacin, ciprofloxacin, moxifloxacin, and sparfloxacin showedan increase of fourfold for ISP794(pQT13), while no changeoccurred for the other drugs tested (Table 3). The transformantQT11(pQT13) showed a MIC profile for quinolones, �-lactams,and dyes that was identical to that of ISP794(pQT13) and in-creased for quinolones relative to ISP794. The presence of

FIG. 3. (A) Gel mobility shift analyses of the interactions of the purified NorG and MgrA proteins with the biotinylated promoters of mgrA,norG, and norC. (B) Gel mobility shift analyses of the interactions of the crude cell extracts from ISP794 and QT1 and the purified NorG and MgrAproteins with the biotinylated 150-bp abcA promoter. (C) Schematic representation of the overlapping promoter region abcA-pbpD as publishedby Domanski et al. (4). Primers designed to amplify abcA DNA generated a fragment that included the inverted repeat located 8 bp downstreamfrom the abcA transcriptional start (�1). This region was shown to be essential for the expression of both genes. The �35 and �10 consensussequences of the abcA and pbpD promoters are underlined and/or in bold. The inverted repeat region is underlined and in bold. The asteriskindicates the transcription start site for abcA. The boxed DNA region indicates the abcA promoter that was used in the gel mobility shift bindingassay. The pbpD promoter region used for the gel mobility shift assay contains the inverted repeat as well as the �10 and �35 regions (underlined).The inverted repeat is located at a distance of 46 bp from the transcription start site of the pbpD gene (T).

VOL. 189, 2007 NorG REGULATORY PROTEIN 3001

on August 20, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

norG overexpression from the plasmid also did not affect thegrowth rate (data not shown).

Overexpression of abcA causes increased resistance to�-lactams. Susceptibility to �-lactams was not affected by ex-pression of the genes encoding NorA, NorB, NorC, and Tet38.To assess whether the overexpression of abcA seen in QT11(norG) contributed to the �-lactam resistance phenotype ofthis strain, we cloned the 1,728-bp abcA gene into the plasmidpSK950 to generate plasmid pQT14. This plasmid was intro-duced into RN4220 and then into the parental strain ISP794.We assessed the overexpression of abcA from the plasmidconstruct pQT14 by Northern blotting. There was an increaseof fourfold for abcA mRNA of ISP794(pQT14) compared tothat of ISP794 or QT11, with or without pQT13 (Fig. 4A). Wethen determined the MICs of quinolones, �-lactams, and dyesfor ISP794 and ISP794(pQT14). ISP794(pQT14), relative toplasmid-free ISP794, showed increases in MICs of 128-fold fornafcillin, 64-fold for penicillin, and 8-fold for methicillin andcefotaxime, with the magnitude of the increase correlating withthe hydrophobicity of the �-lactam side chains (Table 3).ISP794(pQT14) also showed increases in MICs of eightfold forTPP, fourfold for rhodamine, and twofold for ethidium bro-mide and the Hoechst 33342 dye. No change in the MICs ofcetrimide or quinolones was detected. The presence of abcAoverexpression from the plasmid also did not affect the bacte-rial growth rate (data not shown). Thus, in the norG mutant,�-lactam resistance is attributable at least in part to overex-pression of abcA.

DISCUSSION

NorG directly activates the expression of norB, encoding theNorB efflux pump. NorG belongs to the GntR family of regu-lators. GntR is itself a repressor of the gluconate operon inBacillus subtilis. GntR-like regulators possess a helix-turn-helix

region, an N-terminal DNA-binding domain, and a highly vari-able C-terminal domain, which contains the effector-bindingsites and an oligomerization region. Based on the heterogene-ity of the C-terminal regions of this family, NorG was classifiedin the FadR subfamily, which consists of proteins with anall-helical C-terminal domain, often involved in the regulationof various metabolic pathways or oxidized substrates in aminoacid metabolism (21). NorG protein bound specifically to norA,norB, and norC promoters, but the transcription level of onlynorB was increased when norG was overexpressed. These datacorrelated with a fourfold increase in MICs of moxifloxacinand sparfloxacin for the strain overexpressing norG from plas-mid pQT13 and are consistent with the previously reportedsubstrate profile of the NorB efflux pump (26).

As reported in our previous study, upstream of norB arethree ORFs, encoding a putative amino acid permease (N315-SA1270), a putative threonine deaminase (N315-SA1271), anda putative alanine dehydrogenase (N315-SA1272), all of whichshowed at least a threefold increase in their transcript levels inmicroarrays of QT1 (mgrA) compared to those of parentalISP794 (26). The intergenic region between SA1272 andSA1273 contained a putative rho-dependent terminator and aputative promoter, P2, in addition to the putative promoter P1directly upstream of the norB gene. NorG bound specifically toboth putative norB promoters.

The pattern of DNA gel band shifts showed that MgrAbound specifically to the norG promoter and bound lessstrongly to the norB promoter (26). In contrast, NorG boundspecifically to the norB promoter but not to the mgrA pro-moter. These data taken together with our earlier microarraydata showing a threefold increase in the transcripts of norGand norB in QT1 (mgrA) suggested that NorG regulates aminoacid metabolism via expression of ORFs SA1270 to SA1272coordinately with the expression of the NorB transporter,which transports small hydrophobic molecules, such as moxi-

TABLE 3. Susceptibilities of strains to quinolones and other agents

Antimicrobial or dyeMIC (�g/ml) for straina: Side chain

partitioncoefficientbISP794 QT1 QT11 QT12 ISP794(pQT13) ISP794(pQT14)

QuinolonesNorfloxacin 0.5 4 0.5 2 2 0.5Ciprofloxacin 0.25 2 0.25 1 1 0.25Moxifloxacin 0.06 0.5 0.06 0.25 0.25 0.06Sparfloxacin 0.125 1 0.06 0.5 0.5 0.125

�-LactamsNafcillin 0.5 0.25 2 1 0.5 64 4,200Penicillin G 0.03 0.01 0.125 0.06 0.03 2 270Methicillin 4 2 16 8 4 32Cefotaxime 0.5 0.25 2 1 0.5 4 6

Chemical compounds and dyesCetrimide 0.5 2 0.5 2 0.5 0.5TPP 0.5 2 2 4 0.5 8Rhodamine 1 2 1 2 0.5 4Ethidium bromide 2 4 4 8 2 4Hoechst 33342 1 4 2 4 2 2

a Strain QT1, ISP794 mgrA::cat; QT11, ISP794 norG::cat; QT12, ISP794 mgrA::cat norG::cat. pQT13, construct pSK950-norG; pQT14, construct pSK950-abcA. Strainsharboring plasmids pQT13 and pQT14 were grown in the presence of tetracycline (5 �g/ml) at 30°C.

b The side chain partition coefficients were determined by Nikaido et al. (16) and represent the calculated octanol-water partition coefficients of the 6-substituentsand the 7-substituents of the �-lactams.

3002 TRUONG-BOLDUC AND HOOPER J. BACTERIOL.

on August 20, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

floxacin and sparfloxacin, across the cytoplasmic membrane.The exact role of NorB, if any, in amino acid metabolism ortransport is not known, but its coordinated expression with thatof components of certain metabolic pathways may reflect aresponse to environmental conditions in which removal of tox-ins by efflux and changes in amino acid metabolism are bothadvantageous to cell survival.

We postulate that MgrA acts as an indirect repressor of norBvia repression of the expression of NorG, which acts as a direct

activator of transcription of norB. A similar regulatory systemhas been demonstrated for the farAB-encoded efflux pump inNeisseria gonorrhoeae, which confers resistance to antibacterialfatty acids. The farAB operon was indirectly activated by theMtrR regulator, which is a direct repressor of FarR, a repres-sor of the farAB system (12). The role of NorG in regulation ofexpression of norA- and norC-encoded efflux pumps remains tobe determined. The ability of the protein to bind to bothpromoters without an apparent effect on gene transcript levelsin vivo suggests participation of other proteins or other path-ways in this regulation. This hypothesis is supported by ourfinding that disruption of the norG gene did not lead to anydetectable effects on the transcription of norA or norC in mu-tant QT11 (norG).

NorG represses the expression of genes involved in �-lac-tam susceptibility and resistance. The norG::cat mutant hadno change in quinolone resistance phenotype, but this mutantshowed a fourfold increase in the MICs of �-lactam drugs,including nafcillin, penicillin G, methicillin, and cefotaxime. ByNorthern blotting and RT-PCR assays, we found that the tran-scription level of abcA increased fourfold in this mutant. AbcAis an ATP-dependent transporter, involved in cell autolysis,and its transcription is dependent on the agr regulatory system(3, 24). pbpD is the structural gene encoding the transpeptidasePBP4 of S. aureus, a native low-molecular-weight penicillin-binding protein that participates in the synthesis of highlycross-linked muropeptide components of the cell wall (13).Although abcA and pbpD are divergently transcribed, theyshare an intergenic region with overlapping promoters. Aninverted repeat region of 26 bp that plays an important role inthe expression of these two genes was found 8 bp from thetranscription initiation point (�1) of abcA, while it was at adistance of 46 bp from the transcription initiation point ofpbpD (4). Schrader-Fischer and Berger-Bachi (24) found noconnection between resistance to methicillin and AbcA expres-sion in their studies, but the cloned abcA structural gene inthose experiments lacked approximately 4% of its full se-quence. Exposure of cells to methicillin, however, led to anincrease in expression of abcA transcripts (24). We found anincrease in abcA transcripts without a change in pbpD tran-scripts in the norG mutant, strengthening the earlier sugges-tion that abcA and pbpD are regulated differently (6). Consis-tent with this notion, NorG bound to the putative abcApromoter, including the inverted repeat region, in a specificmanner, but no binding occurred in the DNA region bearingthe putative pbpD promoter (data not shown).

AbcA and resistance to �-lactams. Cloning and overexpres-sion of the complete abcA gene on plasmid pQT14 resulted inincreased �-lactam resistance. Thus, it appears that �-lactamsmay be substrates of AbcA and that the �-lactam resistancephenotype of a norG mutant can be attributed at least in partto overexpression of abcA.

AbcA is classified in the group A family of ATP-dependenttransporters, whose members have two membrane-spanningdomains and an ATP-binding domain in a single polypeptide.Exporting substrates is the principal function of this family(24). AbcA shows similarity with efflux transporters LmrA ofLactococcus lactis (65% amino acid similarity) and MsbA ofEscherichia coli (57% amino acid similarity). The substrateprofiles of these two transporters both include the dyes

FIG. 4. (A) Northern blot analyses of RNAs isolated from thespecified S. aureus strains. The same amount of RNA (10 �g) wasloaded in each lane, and loading was verified by ethidium bromidestaining of 16S rRNA before RNAs were transferred onto a nylonmembrane and hybridized with specific probes (abcA or norG).(B) RT-PCR using RNA extracted at exponential phase. Each reactionused 10 picograms of total RNA as the template and primers specificto an internal region of norB. We used 16S rRNA as an internal controlfor the RT-PCR assays as described previously (27). Photographs ofethidium bromide-stained gels were scanned and analyzed using theNIH Scion Image program as described previously (5).

VOL. 189, 2007 NorG REGULATORY PROTEIN 3003

on August 20, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

ethidium bromide and Hoechst 33342 (20). Overexpression ofabcA from pQT14 also produced increases in the MICs of thedyes, TPP, rhodamine, ethidium bromide, and Hoechst 33342.In E. coli, MsbA is an essential transporter involved in thetrafficking of lipids, including lipid A. Since the targets of�-lactams are extracellular transpeptidases, a role of multidrugtransporters located in the cytoplasmic membrane in �-lactamresistance might be unexpected. LmrA in L. lactis, however,has been previously shown to confer resistance to lipophilic butnot hydrophilic �-lactam antibiotics, suggesting a link betweenthe ability of the antibiotics to partition into the cytoplasmicmembrane and LmrA-mediated drug resistance (29). Similarly,the multidrug efflux pump AcrAB of Salmonella enterica sero-var Typhimurium confers resistance to �-lactams with morelipophilic side chains, possibly due to side chain partitioning inthe membrane (16). The observation that AbcA confers resis-tance to �-lactam drugs with a preference toward the morelipophilic ones such as nafcillin or penicillin G supports thishypothesis. Further study of the transport properties of AbcAis ongoing.

AbcA was previously shown to be involved in cell autolysisand is under the control of the agr regulatory system (24). Inour early microarray experiments leading to the identificationof the NorB, NorC, and Tet38 efflux pumps (26, 27), we alsoobserved a fivefold increase in abcA mRNA in a strain over-expressing mgrA compared to that of the wild-type ISP794 anda specific binding of MgrA to the overlapping promoter abcA-pbpD (data not shown), suggesting that MgrA is a direct acti-vator for the expression of abcA. In this study, we demon-strated that abcA was also under the control of NorG and alsoaffects resistance to �-lactams. Taken together, our data sug-gest that AbcA is oppositely regulated by MgrA and NorG.

Conclusions. Multiple regulators affect the expression of avariety of efflux pumps that alter antimicrobial susceptibility inS. aureus. Thus far two regulators, NorG and MgrA, have beenshown to bind the norA promoter, a property that led to theiridentification (28). In contrast to MgrA, NorG binds morestrongly to the putative norB promoters P1 and P2, as well asto the norC putative promoter. NorG also binds specifically toits own putative promoter and the promoter of abcA. Further-more, MgrA and NorG have opposite effects on norB and abcAexpression. MgrA behaves as an indirect repressor for norBand a direct activator for abcA, whereas NorG behaves as adirect activator for norB and a direct repressor for abcA. Themultiplicity of staphylococcal efflux pumps and the complexityof their regulation imply that such pumps are highly importantto the physiology of S. aureus and likely contribute to its abilityto survive in diverse environments, an ability that underlies itsfacility in colonizing, persisting in, and causing disease in mam-malian hosts.

ACKNOWLEDGMENT

This work was supported in part by grant R01-AI23988 from theU.S. Public Health Service, National Institutes of Health (to D.C.H.).

REFERENCES

1. Chopra, I., P. M. Hawkey, and M. Hinton. 1992. Tetracyclines, molecularand clinical aspects. J. Antimicrob. Chemother. 29:245–277.

2. Cunha, B. A. 2002. Strategies to control antibiotic resistance. Semin. Respir.Infect. 17:250–258.

3. Domanski, T. L., and K. W. Bayles. 1995. Analysis of Staphylococcus aureus

genes encoding penicillin-binding protein 4 and an ABC-type transporter.Gene 167:111–113.

4. Domanski, T. L., B. L. M. De Jonge, and K. W. Bayles. 1997. Transcriptionanalysis of the Staphylococcus aureus gene encoding penicillin-binding pro-tein 4. J. Bacteriol. 179:2651–2657.

5. Fournier, B., Q. C. Truong-Bolduc, X. Zhang, and D. C. Hooper. 2001. Amutation in the 5� untranslated region increases stability of norA mRNA,encoding a multidrug resistance transporter of Staphylococcus aureus. J.Bacteriol. 183:2367–2371.

6. Henze, U. U., and B. Berger-Bachi. 1996. Penicillin-binding protein 4 over-production increases �-lactam resistance in Staphylococcus aureus. Antimi-crob. Agents Chemother. 40:2121–2125.

7. Ingavale, S., W. van Wamel, T. T. Luong, C. Y. Lee, and A. L. Cheung. 2005.Rat/MgrA, a regulator of autolysis, is a regulator of virulence genes inStaphylococcus aureus. Infect. Immun. 73:1423–1431.

8. Ingavale, S. S., W. van Wamel, and A. L. Cheung. 2003. Characterization ofRAT, an autolysis regulator in Staphylococcus aureus. Mol. Microbiol. 48:1451–1466.

9. Kaatz, G. W., F. McAleese, and S. M. Seo. 2005. Multidrug resistance inStaphylococcus aureus due to overexpression of a novel multidrug and toxinextrusion transport protein. Antimicrob. Agents Chemother. 49:1857–1864.

10. Kreiswirth, B. 1983. The toxic shock syndrome exotoxin structural gene isnot detectably transmitted by a prophage. Nature 305:709–712.

11. Kuroda, M., T. Ohta, I. Uchiyama, T. Baba, H. Yuzawa, I. Kobayashi, L. Cui,A. Oguchi, K. Aoki, Y. Nagai, J. Lian, T. Ito, M. Kanamori, H. Matsumaru,A. Maruyama, H. Murakami, A. Hosoyama, Y. Mizutani-Ui, N. K. Takahashi,T. Sawano, R. Inoue, C. Kaito, K. Sekimizu, H. Hirakawa, S. Kuhara, S.Goto, J. Yabuzaki, M. Kanehisa, A. Yamashita, K. Oshima, K. Furuya,C. Yoshino, T. Shiba, M. Hattori, N. Ogasawara, H. Hayashi, and K.Hiramatsu. 2001. Whole genome sequencing of methicillin-resistant Staph-ylococcus aureus. Lancet 357:1225–1240.

12. Lee, E. H., C. Rouquette-Loughlin, J. P. Folster, and W. M. Shafer. 2003.FarR regulates the farAB-encoded efflux pump of Neisseria gonorrhoeae viaan MtrR regulatory mechanism. J. Bacteriol. 185:7145–7152.

13. Leski, T. A., and A. Tomasz. 2005. Role of penicillin-binding protein 2(PBP2) in the antibiotic susceptibility and cell wall cross-linking of Staphy-lococcus aureus: evidence for the cooperative functioning of PBP2, PBP4,and PBP2A. J. Bacteriol. 187:1815–1824.

14. Luong, T. T., S. W. Newell, and C. Y. Lee. 2003. mgr, a novel global regulatorin Staphylococcus aureus. J. Bacteriol. 185:3703–3710.

15. Niemeyer, D. M., M. J. Pucci, J. A. Thanassi, V. K. Sharma, and G. L.Archer. 1996. Role of mecA transcriptional regulation in the phenotypicexpression of methicillin resistance in Staphylococcus aureus. J. Bacteriol.178:5464–5471.

16. Nikaido, H., M. Basina, V. Nguyen, and E. Y. Rosenberg. 1998. Multidrugefflux pump AcrAB of Salmonella typhimurium excretes only those �-lac-tam antibiotics containing lipophilic side chains. J. Bacteriol. 180:4686–4692.

17. Putman, M., H. W. Van Veen, J. E. Degener, and W. N. Konings. 2000.Antibiotic resistance: era of the multidrug pump. Mol. Microbiol. 36:772–773.

18. Putman, M., H. W. Van Veen, and W. N. Konings. 2000. Molecular proper-ties of bacterial multidrug transporters. Microbiol. Mol. Biol. Rev. 64:672–693.

19. Ren, Q., K. H. Kang, and I. T. Paulsen. 2004. TransportDB: a relationaldatabase of cellular membrane transport systems. Nucleic Acids Res. 32:D284–D288.

20. Reuter, G., T. Janvilisri, H. Venter, S. Shahi, L. Balakrishnan, and H. W.Van Veen. 2003. The ATP binding cassette multidrug transporter LmrA andlipid transporter MsbA have overlapping substrate specificities. J. Biol.Chem. 278:35193–35198.

21. Rigali, S., A. Derouaux, F. Giannotta, and J. Dusart. 2002. Subdivision of thehelix-turn-helix GntR family of bacterial regulators in the FadR, HutC,MocR, and YtrA subfamilies. J. Biol. Chem. 277:12507–12515.

22. Ross, J. I., E. A. Eady, J. H. Cove, and S. Baumberg. 1996. Minimal func-tional system required for expression of erythromycin resistance by msrA inStaphylococcus aureus RN4220. Gene 183:143–148.

23. Sau, S., J. Sun, and C. Y. Lee. 1997. Molecular characterization and tran-scriptional analysis of type 8 capsule genes in Staphylococcus aureus. J.Bacteriol. 179:1614–1621.

24. Schrader-Fischer, G., and B. Berger-Bachi. 2001. The AbcA transporter ofStaphylococcus aureus affects cell autolysis. Antimicrob. Agents Chemother.45:407–412.

25. Stahl, M. L., and P. A. Pattee. 1983. Confirmation of protoplast fusion-derived linkages in Staphylococcus aureus by transformation with protoplastDNA. J. Bacteriol. 154:406–412.

26. Truong-Bolduc, Q. C., P. M. Dunman, J. Strahilevitz, S. J. Projan, and D. C.

3004 TRUONG-BOLDUC AND HOOPER J. BACTERIOL.

on August 20, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

Hooper. 2005. MgrA is a multiple regulator of two new efflux pumps inStaphylococcus aureus. J. Bacteriol. 187:2395–2405.

27. Truong-Bolduc, Q. C., J. Strahilevitz, and D. C. Hooper. 2006. NorC, a newefflux pump regulated by MgrA of Staphylococcus aureus. Antimicrob.Agents Chemother. 50:1104–1107.

28. Truong-Bolduc, Q. C., X. Zhang, and D. C. Hooper. 2003. Characterization

of NorR protein, a multifunctional regulator of norA expression in Staphy-lococcus aureus. J. Bacteriol. 185:3127–3138.

29. Van Veen, H. W., K. Venema, H. Bolhuis, I. Oussenko, J. Kok, B. Poolman,A. J. M. Driessen, and W. N. Konings. 1996. Multidrug resistance mediatedby a bacterial homolog of the human multidrug transporter MDR1. Proc.Natl. Acad. Sci. USA 93:10668–10672.

VOL. 189, 2007 NorG REGULATORY PROTEIN 3005

on August 20, 2020 by guest

http://jb.asm.org/

Dow

nloaded from