Protein Modulator of Multidrug Efflux Gene Expression in Pseudomonas aeruginosa

Abstract

nalC multidrug-resistant mutants of Pseudomonas aeruginosa show enhanced expression of the mexAB-oprM multidrug efflux system as a direct result of the production of a ca. 6,100-Da protein, PA3719, in these mutants. Using a bacterial two-hybrid system, PA3719 was shown to interact in vivo with MexR, a repressor of mexAB-oprM expression. Isothermal titration calorimetry (ITC) studies confirmed a high-affinity interaction (equilibrium dissociation constant [K_D], 158.0 ± 18.1 nM) of PA3719 with MexR in vitro. PA3719 binding to and formation of a complex with MexR obviated repressor binding to its operator, which overlaps the efflux operon promoter, suggesting that mexAB-oprM hyperexpression in nalC mutants results from PA3719 modulation of MexR repressor activity. Consistent with this, MexR repression of mexA transcription in an in vitro transcription assay was alleviated by PA3719. Mutations in MexR compromising its interaction with PA3719 in vivo were isolated and shown to be located internally and distributed throughout the protein, suggesting that they impacted PA3719 binding by altering MexR structure or conformation rather than by having residues interacting specifically with PA3719. Four of six mutant MexR proteins studied retained repressor activity even in a nalC strain producing PA3719. Again, this is consistent with a PA3719 interaction with MexR being necessary to obviate MexR repressor activity. The gene encoding PA3719 has thus been renamed armR (antirepressor for MexR). A representative “noninteracting” mutant MexR protein, MexR_I104F, was purified, and ITC confirmed that it bound PA3719 with reduced affinity (5.4-fold reduced; K_D, 853.2 ± 151.1 nM). Consistent with this, MexR_I104F repressor activity, as assessed using the in vitro transcription assay, was only weakly compromised by PA3719. Finally, two mutations (L36P and W45A) in ArmR compromising its interaction with MexR have been isolated and mapped to a putative C-terminal α-helix of the protein that alone is sufficient for interaction with MexR.

Multidrug efflux systems of the resistance-nodulation-division (RND) family are broadly distributed among gram-negative bacteria, where they are, in many instances, important determinants of intrinsic and/or acquired antimicrobial resistance (48). Comprised of an inner membrane drug-proton antiporter (the RND component), an outer membrane channel-forming component, and a periplasmic membrane fusion protein (40), these pumps function to capture periplasmic and possibly cytosolic substrates and deliver them to the cell exterior (2, 42, 49). In Pseudomonas aeruginosa, seven RND-type pumps have been described to date (36, 46), although the major efflux determinant of intrinsic multidrug resistance and the best studied of these pumps in P. aeruginosa is MexAB-OprM (45, 46). In addition to numerous medically relevant antimicrobials (47), MexAB-OprM also exports a variety of dyes and detergents (25, 62, 64), inhibitors of fatty acid biosynthesis (58), biocides (10), organic solvents (26, 27), homoserine lactones associated with quorum sensing (14, 43), and possibly a virulence factor(s) (18). It is far from clear, therefore, that antimicrobial efflux is the intended function of this system.

The mexAB-oprM efflux operon is negatively regulated by MexR, the product of a gene upstream of and divergently transcribed from the efflux genes (50). MexR is a member of the MarR family of regulators (68) and binds as a dimer (28) to two sites in the mexR-mexA intragenic region, near mexR and overlapping promoters for both mexR and mexAB-oprM (1, 13, 54, 57). Mutations in mexR are associated with the increased mexAB-oprM expression and concomitant multidrug resistance of so-called nalB mutants (20, 55, 63). Mutations in two additional repressor genes, nalC (also known as PA3721) (8, 29, 63) and nalD (also known as PA3574) (60), also yield increased mexAB-oprM expression and multidrug resistance. While NalD regulates mexAB-oprM expression directly by binding upstream of mexAB-oprM at a second promoter more mexA proximal than the MexR-binding site (39), NalC regulates efflux gene expression indirectly as a result of its direct control of a two-gene operon, PA3720-PA3719 (8). Upregulation of PA3720-PA3719 in nalC mutants is, in fact, responsible for the increased mexAB-oprM expression in such mutants, with PA3719 alone able to effect this increase (8). PA3719 encodes a protein with a molecular mass of ca. 6,100 Da having no homology to any known or predicted gene products in the GenBank databases. Intriguingly, overproduction of PA3719 (from an expression vector or in a nalC mutant) also substantially increases MexR protein levels (8), although the latter clearly no longer represses efflux gene expression. One possibility, then, is that PA3719 somehow modulates MexR repressor activity by interacting directly with this repressor, thereby alleviating repression of both mexAB-oprM and mexR (hence the increased MexR levels in a nalC mutant). We provide here in vivo and in vitro data in support of a PA3719-MexR interaction that modulates MexR repressor activity and in doing so facilitates expression of this efflux operon.

MATERIALS AND METHODS

Bacterial strains and growth conditions.

The bacterial strains and plasmids used in this study are listed in Table 1. Bacterial cells were cultivated in Luria broth (Miller's Luria broth base [Difco] and 2 g of NaCl per liter of H₂O) and agar (Luria broth containing 1.5% [wt/vol] agar), supplemented with appropriate antibiotics when necessary, at 37°C. Plasmid pDSK519 and its derivatives were maintained with 50 μg/ml (in Escherichia coli) or 100 μg/ml (in P. aeruginosa) kanamycin. Plasmid pMMB206 and its derivatives were maintained with 10 μg/ml (in E. coli) or 150 μg/ml (in P. aeruginosa) chloramphenicol. Plasmid pBluescript II SK(+) and its derivatives were maintained in E. coli using 100 μg/ml ampicillin. Plasmids pDD1 (pET-Dest42::mexR), pDD2 (pTYB12::PA3719), and pDD3 (pET-Dest42:: mexR_I104F) were maintained in E. coli with 50 μg/ml ampicillin. A mexR deletion was introduced into nalC strain K1454 using plasmid pRSP75 as described previously (63) and was confirmed using colony PCR (59) with primers MexRF and MexRR as described below.

TABLE 1.

Bacterial strains and plasmids

Strain or plasmid	Relevant characteristicsa	Reference or source
E. coli strains
DH5α	φ80dlacZΔM15 endA1 recA1 hsdR17 (rK− mK+) supE44 thi-1 gyrA96 relA1 F− Δ(lacZYA-argF)U169	5
SU202	lexA71::Tn5 sulA211 sulA::lacZ Δ(lacIPOZYA)169 F′ lacIqlacZΔM15::Tn9	12
P. aeruginosa strains
K767	PAO1 prototroph	32
K1491	K767 ΔmexR	63
K1454	K767 nalC	8
K2519	K1454 ΔmexR	This study
K2276	K1454 ΔPA3719	This study
Plasmids
pBluescript II SK(+)	Cloning vector; Apr	Stratagene
pLC30	pBluescript II SK(+)::mexR	This study
pMS604	LexA1-87WT-Fos zipper fusion; ori (ColE1) Tcr	12
pLK452	pMS604::mexR WT	1
pLC60	pMS604::mexR L35P	This study
pLC61	pMS604::mexR I104F	This study
pLC62	pMS604::mexR M112T	This study
pLC63	pMS604::mexR L135F	This study
pLC64	pMS604::mexR L28P	This study
pLC65	pMS604::mexR L75P	This study
pDP804	LexA1-87408-Jun zipper fusion; ori (P15A) Apr	12
pSF001	pDP804::PA3719 WT	This study
pSF002	pDP804::PA3719 W45A	This study
pSF003	pDP804::PA3719 L36P	This study
pRK001	pDP804::PA3719 S25-Y53b	This study
pDSK519	Broad-host-range cloning vector; Kmr	23
pLC23	pDSK519::PA3719	8
pLC66	pDSK519::mexR WT	This study
pLC67	pDSK519::mexR L35P	This study
pLC68	pDSK519::mexR I104F	This study
pLC69	pDSK519::mexR M112T	This study
pLC70	pDSK519::mexR L135F	This study
pLC71	pDSK519::mexR L28P	This study
pLC72	pDSK519::mexR L75P	This study
pMMB206	Broad-host-range, low-copy-number cloning vector; Cmr	38
pSF004	pMMB206::PA3719 WT	This study
pSF005	pMMB206::PA3719 W45A	This study
pSF006	pMMB206::PA3719 L36P	This study
pET-Dest42	His6 tag expression vector; Apr	Invitrogen
pDD1	pET-Dest42::mexR	This study
pTYB12	Intein fusion expression vector	New England Biolabs
pDD2	pTYB12::PA3719	This study
pDD3	pET-Dest42::mexRI104F	This study

^aAp^r, ampicillin resistant; Tc^r, tetracycline resistant; Km^r, kanamycin resistant; Cm^r, chloramphenicol resistant; WT, wild type. Mutated residues in the MexR or PA3719 proteins encoded by the plasmids are indicated.

^bN-terminally truncated PA3719 comprised of residues S25 to Y53.

DNA protocols.

Standard protocols were used for restriction endonuclease digestion, ligation, transformation, electroporation (E. coli), and agarose gel electrophoresis, as described by Sambrook and Russell (56). Electroporation of plasmids into P. aeruginosa was carried out as described previously (9). Ligation mixtures were typically applied to Millipore type VM 0.05-μm filter disks over distilled H₂O and incubated for 30 min at room temperature prior to electroporation. Genomic DNA of P. aeruginosa was extracted by following the protocol of Barcak et al. (7). Plasmids were isolated from E. coli DH5α and E. coli SU202 using a QIAprep Spin MiniPrep kit (QIAGEN, Inc., Chatsworth, CA). DNA fragments used for cloning were excised from agarose gels using Prep-A-Gene (Bio-Rad Laboratories, Richmond, CA) in accordance with the manufacturer's instructions. PCR products were purified using a QIAquick PCR purification kit (QIAGEN). Oligonucleotides were chemically synthesized by Cortec DNA Services Inc., Kingston, Ontario, Canada, and nucleotide sequencing was carried out by ACGT Corp., Toronto, Ontario, Canada, or by Agencourt, Beverly, MA.

Plasmids.

The mexR gene was cloned into plasmid pDSK519 following amplification of the gene by PCR using primers MexRF (5′-GATCGGATCCCATTAGGTTTACTCGGCCAAACC-3′; BamHI site underlined) and MexRR (5′-GATCGAATTCCGCCAGTAAGCGGATACCTG-3′; EcoRI site underlined) in a reaction mixture (50 μl) containing 1 μg of PAO1 strain K767 chromosomal DNA as the template, 2.5 U of Vent DNA polymerase (New England Biolabs, Mississauga, Ontario, Canada), 0.2 mM of each deoxynucleoside triphosphate, 1× ThermoPol buffer (New England Biolabs), 30 pmol of each primer, 2 mM MgSO₄, and 10% (vol/vol) dimethyl sulfoxide (DMSO). The reaction mixture was subjected to an initial 3-min denaturation step at 95°C, followed by 30 cycles of 45 s at 95°C, 30 s at 60°C, and 45 s at 72°C before a 5-min elongation at 72°C. PCR products were purified, digested with BamHI and EcoRI, and cloned into pDSK519 and pBluescript II SK(+), yielding pLC66 and pLC30, respectively. For use in isothermal calorimetry (ITC) studies MexR was expressed with a C-terminal His₆ tag from plasmids provided in Invitrogen Gateway cloning and expression kits (Invitrogen, Carlsbad, CA). Briefly, the mexR gene was amplified by PCR using Accuprime GC-rich DNA polymerase and primers MexRsrp-For (5′-GGGGACAAGTTTGTACAAAAAAGCAGGCTTCGAAGGCTTCGAAGGAGATAGAACCATGAACTACCCCGTGAATCCC-3′) and MexRsrp-Rev (5′-GGGGACCACTTTGTACAAGAAAGCTGGGTCAATATCCTCAAGCGGTTGCGC-3′) and inserted into pDONOR221 using the BP recombinase according to the manufacturer's instructions (Invitrogen). The mexR gene was then transferred from pDONOR221 to pET-Dest42 (yielding pDD1) using the LR recombinase, again according to the manufacturers’ instructions,. For ITC studies of MexR_I104F, plasmid pDD3 carrying mexR_L104F was constructed by introduction of a point mutation into the mexR gene of pDD1 using the Stratagene QuickChange mutagenesis system and a protocol provided by the manufacturer (Stratagene, La Jolla, CA).

The PA3719 gene was cloned into the two-hybrid vector pDP804 following amplification of the gene from plasmid pLC23 using PCR and primers PDP-PA3719F (5′-TCAGCTCGAGCTCGAGATGTCCCTGAACACTCCG-3′; tandem XhoI sites underlined) and PDP-PA3719R (5′-TCATAGATCTAGATCTTGCGCGGATTCTGATAGCT-3′; tandem BglII sites underlined). Reaction mixtures (50 μl) containing 100 ng of pLC23, 0.6 μM of each primer, 0.2 mM of each deoxynucleoside triphosphate, 2 mM MgSO₄, and 1 U of Pfu DNA polymerase (Fermentas Life Sciences, Burlington, Ontario, Canada) in 1× Pfu DNA polymerase buffer were heated at 98°C for 45 s and then subjected to 25 cycles of 98°C for 45 s, 62°C for 45 s, and 72°C for 1 min, followed by 10 min at 72°C. The PA3719-carrying PCR product was purified (see below), digested with XhoI and BglII, and cloned into pDP804, yielding pSF001. A New England Biolabs IMPACT intein fusion kit was used for cloning and expression of PA3719 (as an intein fusion) for use in ITC studies. PCR amplification of the gene was carried out using primers PA3719srp-For (5′-GGCGAAGCGGCCGCATGTCCCTGAACACTCCGC-3′; NotI site underlined) and PA3719srp-Rev (5′-CGTGGCTCGAGTCAGTAGAAGTGCTCGCCG-3′; XhoI site underlined) and Accuprime GC-rich DNA polymerase, and the gene was cloned into the kit-provided plasmid pTYB12, yielding the PA3719-intein fusion-expressing vector pDD2.

The 3′ end of PA3719 encoding residues S25 to Y53 was cloned into the two-hybrid vector pDP804 as a XhoI-BglII fragment following the annealing of two XhoI- and BglII-flanked 110-bp oligonucleotides corresponding to both strands of the S25- to Y53-encoding region of PA3719. Annealing was achieved by incubating 800 pmol of each oligonucleotide in T4 ligation buffer at 95°C for 3 min and allowing the reaction mixture to cool at room temperature for 5 h, after which the 3′ PA3719 DNA was purified using a gel and PCR cleanup kit (Promega).

Random mutagenesis.

PCR-based random mutagenesis (53) of mexR was carried out using primers RanMuMexRForward (5′-GAGCGGTGACCATGAACTACCCCGTGAATCC-3′; BstEII site underlined) and RanMuMexRReverse (5′-GAGGCTCGAGTTAAATATCCTCAA-GCGGTTG-C-3′; XhoI site underlined). The reaction mixtures (50 μl) contained 100 ng of plasmid pLC30 as the template, 2.5 U of Taq DNA polymerase (New England Biolabs, Mississauga, Ontario, Canada), 1× ThermoPol buffer (New England Biolabs), 0.2 mM of each deoxynucleoside triphosphate (except dATP or dCTP [40 μM]), 30 pmol of each primer, 2 mM MgSO₄, and 10% (vol/vol) DMSO. The PCR mixtures were subjected to an initial 3-min denaturation step at 95°C, followed by 29 cycles of 45 s at 95°C, 30 s at 60°C, and 45 s at 72°C before a 5-min elongation at 72°C. The PCR products were then digested with BstEII and XhoI and cloned into pLK452 (pMS604::mexR) purified so that it was free of the wild-type mexR-containing BstEII-XhoI fragment (i.e., the wild-type gene was replaced with mutagenized mexR). PCR-based random mutagenesis of PA3719 was carried out with primers PDP-PA3719F and PDP-PA3719R (see above) as described above for mexR, omitting DMSO, using pLC23 as the template, and including MnCl₂ at a final concentration of 25 to 150 μM. The reaction conditions were the same as those described above for PCR of mexR. PCR products were purified, digested with XhoI and BglII, and cloned into pSF001 (pDP840::PA3719) purified so that it was free of the wild-type PA3719-containing XhoI-BglII fragment of this vector (i.e., the wild-type PA3719 gene of pSF001 was replaced with mutagenized PA3719).

Mutant mexR genes (screened using the two-hybrid assay [see below]) were cloned into plasmid pDSK519 following their amplification, from plasmid pMS604 derivatives harboring them, with primers RanMexRForward (5′-GAGCGGATCCATGAACTACCCCGTGAATCC-3′; BamHI site underlined) and RanMexR Reverse (5′-GAGGGAATTCTTAAATATCCTCAAG-CGGTTGC-3′; EcoRI site underlined). The reaction mixtures (50 μl) contained 1 μg of pMS604 derivative as the template, 2.5 U of Pfu DNA polymerase (New England Biolabs, Mississauga, Ontario, Canada), 1× Pfu buffer (New England Biolabs), 0.2 mM of each deoxynucleoside triphosphate, and 30 pmol of each primer. The PCR mixtures were subjected to an initial 3-min denaturation step at 98°C, followed by 29 cycles of 45 s at 98°C, 30 s at 65°C, and 45 s at 72°C before a 5-min elongation at 72°C. Following purification and digestion (with BamHI and EcoRI) the mexR-carrying fragments were cloned into appropriately digested pDSK519 and mobilized into P. aeruginosa using a previously described triparental mating procedure (69), with transconjugants selected on LB agar containing kanamycin (750 μg/ml) and imipenem (0.5 μg/ml) (to counterselect E. coli).

Wild-type and mutant (screened using the two-hybrid assay [see below]) PA3719 genes were cloned into plasmid pMMB206 following their amplification, from plasmid pDP804 derivatives harboring them, with primers PA3719SD-F (5′-GACTGAATTCGAATTCACGGGGATGAACTGGTGACGCCGCCATGTCCCTGAACACTCCG-3′; tandem EcoRI sites underlined) and PA3719DS-R (5′-GACTAAGCTTAAGCTTTGCGCGGATTCTGATAGCT-3′; tandem HindIII sites underlined). The reaction mixtures were formulated as described above for PA3719 cloning into pDP804, except that Vent DNA polymerase (2 U; New England Biolabs, Ltd., Pickering, Ontario, Canada) and its buffer replaced Pfu and MgSO₄ was included at a concentration of 1 mM. The reaction mixtures were also heated as described above, with the exception that an annealing temperature of 65°C (not 62°C) was used. Following purification and digestion (with EcoRI and HindIII) the PA3719-carrying fragments were cloned into appropriately restricted pMMB206. The resultant plasmids were then introduced into P. aeruginosa via electroporation.

Bacterial two-hybrid system.

To assess an interaction between MexR and PA3719 in vivo and the impact of mutations (or truncations) on this interaction, mexR-carrying pMS604 (or its mutated derivatives) and PA3719-carrying pDP804 (or its mutated or truncated derivatives) were electroporated into E. coli strain SU202 and plated onto 1% (wt/vol) lactose-MacConkey agar containing ampicillin (100 μg/ml) and tetracycline (10 μg/ml) as described previously (53). Any interaction between MexR and PA3719 sequences encoded by the various mexR- and PA3719-containing pMS604 and pDP804 derivatives was observable as a lack of or reduction in β-galactosidase activity in SU202 (i.e., pale pink to white colonies on lactose-MacConkey agar), while the absence of an interaction yielded β-galactosidase activity (red colonies) in this reporter strain (53). These results were confirmed using a more quantitative β-galactosidase assay, as described below. In screening randomly mutagenized mexR or PA3719 genes for mutations compromising the MexR-PA3719 interaction, the ligation mixtures producing the pLK452 and pSF001 derivatives that carry randomly PCR-mutagenized mexR and PA3719, respectively (see above), were electroporated directly into E. coli SU202, which was then plated onto 1% (wt/vol) lactose-MacConkey agar containing ampicillin and tetracycline and screened as described above. In all instances, whole-cell protein extracts were prepared and screened (using immunoblotting with anti-LexA antibodies [see below]) for the production of LexA-MexR and LexA-PA3719 fusions, to ensure production of wild-type and mutant versions of these proteins.

SDS-polyacrylamide gel electrophoresis and immunoblotting.

Whole-cell extracts were prepared as described previously (52), electrophoresed on 20% (wt/vol) sodium dodecyl sulfate (SDS)-polyacrylamide gels, and transferred to Immobilon-P polyvinylidene difluoride membranes (Millipore) (61). Membranes were probed with monoclonal anti-LexA antibodies (Invitrogen) as described previously (61).

β-Galactosidase assay.

E. coli SU202 derivatives containing pDP804 (or its derivatives) and pMS604 (or its derivatives) were grown in LB medium supplemented with ampicillin (100 μg/ml) and tetracycline (10 μg/ml) for approximately 18 h at 37°C. Cultures were then diluted 1:49 into fresh antibiotic-supplemented medium and incubated at 37°C until the optical density at 600 nm was 0.8 to 1.0 before they were assayed for β-galactosidase activity as described previously (35).

Susceptibility testing.

The antibiotic susceptibility of P. aeruginosa was assessed using a previously described broth dilution assay (21), and the data are reported as the MIC for each agent tested.

Expression and purification of MexR and PA3719.

Wild-type MexR used in mobility shift assays was expressed in and purified from E. coli BL21(DE3) as described previously (28). For use in ITC studies, MexR and MexR_I104F were expressed in E. coli BL21(DE3)* Star (Invitrogen) from plasmids pDD1 and pDD3, respectively. Briefly, Luria broth (4 liters) supplemented with ampicillin was inoculated 1:99 with an overnight culture of plasmid-carrying E. coli BL21(DE3)* Star, and the culture was grown to an optical density at 600 nm of ∼0.5 and induced with IPTG (final concentration, 1 mM) for 4 h at 37°C. Cells were harvested by centrifugation at 7,500 × g for 10 min at 4°C, resuspended in 10 ml of lysis buffer (20 mM sodium phosphate [pH 7.2], 500 mM NaCl, 10 mM imidazole, 1× Roche Complete EDTA-free protease inhibitor cocktail), and lysed by three consecutive passes through a French pressure cell at 20,000 lb/in². Protein purification was performed on a HisTrap HP column according to the manufacturer's instructions (Amersham Biosciences, Piscataway, NJ). Purified MexR proteins were quantitated using the Bradford assay and were verified by A₂₈₀ measurement after denaturation with 6 M guanidinium-HCl. PA3719 for use in ITC was expressed (as an intein fusion) from the pTYB12::PA3719 vector, pDD2, also in E. coli BL21 (DE3)* Star as described above except for the use of IPTG at a concentration of 0.5 mM and subsequent growth at 18°C for 18 h. Purification of the PA3719-intein fusion protein was performed according to the manufacturer's instructions (New England Biolabs). The intein-cleaved PA3719 protein retained an AGHMTSSRVDGGR N-terminal extension and was quantitated as described above for MexR. All proteins were >98% pure as determined by SDS-polyacrylamide gel electrophoresis.

ITC.

MexR, MexR_I104F, and PA3719 were purified as described above. The protein samples were buffer exchanged into ITC assay buffer (20 mM Tris-HCl [pH 8.0], 200 mM NaCl) using a combination of an Amicon 8200 stirred concentration cell with a YM1 ultrafiltration membrane to concentrate the samples to 10 ml, followed by dialysis for 8 h at room temperature. Proteins were quantitated using the Bradford assay and were verified by A₂₈₀ measurement after denaturation in 6 M guanidinium-HCl. ITC experiments were carried out using a VP-ITC Microcal calorimeter (Microcal Inc., Studio City, CA). Solutions were degassed for 15 min prior to experiments using a vacuum degasser. Calorimetric titrations were carried out with wild-type MexR protein at a concentration of 16.2 μM in the sample cell, ITC assay buffer in the reference cell, and PA3719 at a concentration of 107.9 μM in the automated syringe. The injection sequence consisted of a single 4-μl injection followed by a series of 28 injections of 10 μl, separated by 300 s of equilibration time. Calorimetric titrations with the site-directed variant MexR_I104F were performed as described above except that MexR_I104F was present at a concentration of 70 μM in the sample cell and PA3719 was present at a concentration of 350 μM in the syringe. The mixing speed was 307 rpm, the cell temperature was 25°C, and the reference power was 5 μcal/s, with an initial delay of 60 s. The saturation levels of the interaction at the concentrations tested were approximately 2.9-fold for wild-type MexR and 2.2-fold for MexR_I104F, giving PA3719/MexR molar ratios of approximately 1.45 and 1.1, respectively, by the end of the titrations. C values of the titrations were calculated using the following equation: C = K_A × M_Tot × n, where K_A is the association rate constant, M_Tot is the concentration of the binding partner in the cell being titrated, and n is the stoichiometry of the interaction. The data were fitted by a single-site model using the ORIGIN ITC software provided by Microcal to obtain a least-squares estimate of K_A (in M⁻¹). The equilibrium dissociation constant (K_D) was calculated using the equation K_D = 1/K_A.

Mobility shift assay.

The mobility shift assay was carried out in 10 μl of 20 mM Tris-HCl (pH 7.5)-50 mM NaCl as described previously (28), using purified MexR and PA3719 synthesized (with the N-terminal Met) by C S Bio Company, Inc. (Menlo Park, CA) (www.csbio.com). The proteins (individually and together in various amounts) were mixed with target DNA (28-mer MexR operator [28]; 30 μM) at room temperature prior to gel loading.

In vitro transcription assay.

An 836-bp DNA fragment containing a phage T7 promoter at its 5′ end and mex DNA beginning 251 bp upstream of mexA (including the MexR-binding site) and extending 585 bp into mexA was amplified from P. aeruginosa PAO1 genomic DNA by PCR using Accuprime GC-rich DNA polymerase (Invitrogen, Carlsbad, CA) and primers A (5-GAAATTAATACGACTCACTATAGGGTAAATGTGGTTGATCCAGTCAAC-3′; T7 promoter underlined) and B (5′-CGGGTCGAGCTGTTGCACGGTGGCCATCGCGTTGGC-3′). The PCR product was resolved using agarose (1%, wt/vol) gel electrophoresis, purified using a QIAGEN gel extraction kit (QIAGEN, Valencia, CA), and quantitated using a nanodrop spectrophotomoter (Nanodrop, Wilmington, DE). In vitro transcription from this template was conducted using a Megascript T7 RNA polymerase-based in vitro transcription kit (Ambion, Austin, TX) according to the manufacturers’ instructions, with the exception that the amount of template was reduced to 250 ng (final concentration, 22.7 nM) and one-half the suggested amounts of T7 polymerase enzyme mixture and nucleotides were used. Control experiments using the kit-provided pTRI-Xef-1 template were carried out according to the manufacturer's instructions. To assess the impact of the MexR or MexR_I104F repressors on transcription from these templates, the purified proteins were added at a final concentration of 1 μM in 20-μl reaction mixtures. To assess the impact of PA3719 on MexR/MexR_I104F repressor activity in this assay, PA3719 was added at various concentrations to reaction mixtures containing 1 μM of each repressor. Control reactions were also carried out with PA3719 alone. In all instances, reactions were allowed to proceed for 3 h at 37°C, after which the RNA product was purified using a QIAGEN RNeasy RNA purification kit (QIAGEN, Valencia, CA), incorporating on-column DNase digestion for 30 min. Purified RNA products were quantitated using a nanodrop spectrophotometer, mixed with 2× Novex Tris-borate-EDTA-urea sample buffer, heated to 70°C for 3 min, separated by electrophoresis in Novex 6% Tris-borate-EDTA-urea gels (Invitrogen, Carlsbad, CA), and visualized by ethidium bromide staining.

RESULTS

MexR-PA3719 interaction.

nalC multidrug-resistant mutants of P. aeruginosa show enhanced expression of mexAB-oprM that is dependent upon increased expression of PA3719, the second gene of the NalC repressor-regulated PA3720-PA3719 operon (8). One way in which PA3719 might increase mexAB-oprM expression is by interacting with and modulating the activity of MexR, the primary repressor of mexAB-oprM expression and a target for mutation in so-called nalB multidrug-resistant mutants (63). To assess an interaction between MexR and PA3719, then, a bacterial two-hybrid approach was employed (12), in which the PA3719 and mexR genes were cloned in frame to coding sequences for the DNA-binding domain of LexA on plasmids pDP804 and pMS604, respectively, and introduced into E. coli SU202 carrying a chromosomal lacZ gene under the control of a LexA operator. LexA binding to its operator and subsequent repression of lacZ in this strain require prior dimerization of the LexA-binding domains encoded by pDP804 and pMS604, necessitating interaction of the PA3719 and MexR sequences fused to the LexA DNA-binding domains of these vectors. As such, lack of β-galactosidase activity is a measure of the PA3719-MexR interaction. The two-hybrid vectors also contain sequences encoding Jun and Fos zipper motifs (which are known to interact) fused to lexA, such that E. coli SU202 carrying these vectors demonstrates substantial repression of lacZ (Table 2). The unaltered vectors thus provide a positive control for the system, although the Jun and Fos zipper-encoding sequences will be disrupted upon cloning of PA3719 and/or mexR sequences, making lacZ repression dependent upon the PA3719-MexR interaction. As shown in Table 2, individual cloning of mexR or PA3719 into pMS6054 or pDP804 provided substantial β-galactosidase activity in SU202 as a result of the absence of an appropriate interacting partner in these vectors to effect lacZ repression. When mexR-carrying pMS604 and PA3719-carrying pDP804 were introduced into SU202, however, the β-galactosidase levels were reduced substantially (Table 2), consistent with lacZ repression and thus a MexR-PA3719 interaction. A MexR-PA3719 interaction was subsequently confirmed in vitro using ITC analysis (Fig. 1A). Integration of the raw data (peaks of heat released upon binding of PA3719 to wild-type MexR) gave an estimated stoichiometry of 0.54 ± 0.01 mol of PA3719 per mol of MexR, corresponding to one PA3719 binding site per MexR dimer in solution and a K_D of 158.0 ± 18.1 nM.

TABLE 2.

β-Galactosidase activity of E. coli SU202 isolates expressing wild-type and mutant MexR and PA3719 proteins in the LexA-based two-hybrid system

Two-hybrid plasmid combinationa		β-Galactosidase activity (Miller units)b
pDP804 derivative	pMS604 derivative
pDP804	pMS604	535
pDP804	mexR WT	5,856
PA3719 WT	pMS604	7,987
PA3719 WT	mexR WT	54
PA3719 WT	mexR L35P	2,699
PA3719 WT	mexR I104F	2,769
PA3719 WT	mexR M112T	2,778
PA3719 WT	mexR L135F	2,509
PA3719 WT	mexR L28P	2,499
PA3719 WT	mexR L75P	3,312
PA3719 W45A	mexR WT	3,359
PA3719 L36P	mexR WT	1,804
PA3719 S25-Y53	mexR WT	23

^aWT, wild type.

^bThe data are representative of at least three independent determinations.

FIG. 1.

Binding of PA3719 to MexR and MexR_I104F measured by ITC. (A) ITC titration of wild-type MexR (16.2 μM) with 10-μl injections of 107.9 μM PA3719. The upper panel shows the raw data for heat released by the binding of PA3719 to MexR expressed in microcalories per second versus time in minutes, while the lower panel shows the peak integration for each injection expressed in kilocalories per mole of PA3719 versus the molar ratio of MexR to PA3719 in solution. The line shows the fit of the data to a single-site model. The least-squares estimate for N (stoichiometry) was 0.54 ± 0.01 PA3719 binding site per MexR molecule or one binding site per MexR dimer, the least-squares estimate for K_A was 6.29 × 10⁶ ± 0.72 × 10⁶ M⁻¹, and the least-squares estimate for K_D was 158.0 ± 18.1 nM. (B) ITC titration of MexR_I104F (70 μM) with 10-μl injections of 350 μM PA3719. The least-squares estimate for N was 0.63 ± 0.01, the least-squares estimate for K_A was 1.17 × 10⁶ ± 0.21 × 10⁶ M⁻¹, and the least-squares estimate for K_D was 853.2 ± 151.1 nM.

PA3719 modulates MexR DNA binding and repressor activity.

MexR regulates mexAB-oprM expression by binding to an operator region that overlaps the major promoter of this efflux operon (13, 39). Gel shift assays involving a synthetic mexR operator sequence confirmed this MexR-DNA binding (Fig. 2A, lane 3). Addition of PA3719 yielded, as expected, a MexR-PA3719 complex (Fig. 2B, lane 6) and reduced MexR binding to operator DNA (Fig. 2A, lane 6), and this became more pronounced (more complex, less DNA binding) as more PA3719 was added (Fig. 2, lane 7). The smear of DNA apparently comigrating with the MexR-PA3719 complex (Fig. 2A, lane 7) is likely not indicative of DNA-bound MexR-PA3719, as one would not expect a DNA-free PA3719-MexR complex (Fig. 2, lane 5) to show the same mobility as a PA3719-MexR-DNA complex, just as DNA-free MexR and its DNA-bound form migrate very differently (Fig. 2B, compare lanes 2 and 3). While these data suggest that PA3719 modulates MexR operator binding and, ultimately, MexR-mediated repression of mexAB-oprM, in the absence of some explanation for the apparently anomalous migration of the MexR operator DNA in the presence of the PA3719-MexR complex we cannot unequivocally make this claim. To assess whether PA3719 does modulate MexR repression of mexAB-oprM, a system for in vitro transcription of mexA was established by first engineering a phage T7 promoter upstream of sequences encompassing the MexR-binding site and extending into mexA and then using T7 RNA polymerase to drive mexA transcription from this template (Fig. 3A, lane 1). While addition of MexR clearly reduced mexA transcription in this assay (Fig. 3A, lane 2), subsequent addition of PA3719 at a concentration equal to its K_D (Fig. 3A, lane 3) or threefold greater than its K_D (Fig. 3A, lane 4) interfered with MexR repressor activity (i.e., mexA transcription increased), ultimately restoring mexA transcription to levels seen without the repressor (Fig. 3A, compare lane 4 and with lane 1). As expected, PA3719 alone did not impact mexA transcription (Fig. 3B, compare lanes 2 and 3 with lane 1) and MexR did not impact transcription of a control template lacking a MexR-binding site (Fig. 3A, compare lanes 6 and 7 with lane 5).

FIG. 2.

Electrophoretic mobility shift assay demonstrating disruption of the MexR-operator DNA complex in lieu of formation of a MexR-PA3719 complex as visualized by (A) ethidium bromide staining of the MexR operator DNA and (B) Coomassie brilliant blue staining of the protein. The protein components indicated, MexR (0.6 nmol) and PA3719 (0.6 nmol [+] or 1.2 nmol [++]), and MexR operator DNA (5′-ATTTTAGTTGACCTTATCAACCTTGTTT-3′; 0.3 nmol) (28) were incubated in 10-μl (final volume) mixtures and subjected to nondenaturing polyacrylamide (12%, wt/vol) gel electrophoresis under reducing conditions at 77 V for 3 h at room temperature. Note that PA3719 has a high pI and migrates in the opposite direction when it is not in a complex with MexR. Thus, a control experiment run with PA3719 and operator DNA only shows no PA3719 and free, unshifted DNA (data not shown).

FIG. 3.

PA3719 and MexR control of in vitro mexA transcription. (A) Impact of PA3719 on MexR repression of mexA transcription. Phage T7 RNA polymerase-based in vitro transcription reactions were initiated from a 836-bp PCR-generated template (250 ng) carrying a T7 promoter and mexA sequences extending from the MexR operator to 585 bp into the mexA gene (lanes 1 to 4). The results for control reactions with a T7 promoter-containing DNA template, pTRI-Xef, provided with the T7 in vitro transcription kit, are shown in lanes 5 to 7. Reactions were carried out at 37°C for 3 h, and RNA was purified and subjected to electrophoresis prior to visualization with ethidium bromide. MexR (1 μM) and PA3719 (150 nM [+¹] or 500 nM [+^2]) were included in the reaction mixtures as indicated. Equal volumes of RNA were loaded in each instance. The following amounts of RNA were produced: lane 1, 4.2 μg; lane 2, 1.1 μg; lane 3, 2.0 μg; lane 4, 4.2 μg; lane 5, 1.7 μg; lane 6, 1.6 μg; and lane 7, 1.6 μg. (B) Impact of PA3719 on mexA transcription. In vitro transcription of the mex (lanes 1 to 3) and pTRI-Xef (lanes 4 to 6) templates was carried out as described above with or without PA3719 (500 nM [+²] or 2 μM [+³]), as indicated. The following amounts of RNA were produced: lane 1, 4.4 μg; lane 2, 4.5 μg; lane 3, 4.6 μg; lane 4, 4.6 μg; lane 5, 4.6 μg; and lane 6, 4.4 μg. (C) Differential impact of PA3719 on the repressor activities of MexR and MexR_I104F. In vitro transcription of the mex (lanes 1 to 5) and pTRI-Xef (lanes 6 to 8) templates was carried out as described above with or without PA3719 (500 nM) and MexR/MexR_I104F (1 μM), as indicated. The following amounts of RNA were produced: lane 1, 4.2 μg; lane 2, 0.8 μg; lane 3, 3.8 μg; lane 4, 0.9 μg; lane 5, 1.9 μg; lane 6, 6.6 μg; lane 7, 6.5 μg; and lane 8, 6.4 μg. RNA was quantitated prior to loading, and the values reported reflect the amount loaded and not the total amount produced. Because reaction, recovery, and loading volumes were kept constant throughout, these values accurately reflect the relative transcript levels produced in each instance.

MexR mutations compromising PA3719 interaction.

To assess the details of the MexR-PA3719 interaction, including identification of residues or regions of MexR important for or involved in this interaction, mutations in mexR compromising this interaction were selected using the two-hybrid system mentioned above. Thus, mexR-carrying pMS604 was mutagenized and introduced into pDP804::PA3719-carrying SU202, and LacZ⁺ colonies (i.e., red colonies) were selected on MacConkey agar; a defect in the MexR-PA3719 interaction was expected to obviate lacZ repression in the SU202 reporter strain. Following confirmation that putative mutant MexR proteins (as fusions to LexA) defective in the PA3719 interaction were expressed (to eliminate further study of mutations compromising MexR production or stability, which would also compromise the interaction with PA3719 and obviate lacZ repression in SU202), mexR genes were sequenced and the genes carrying single point mutations were saved for further study. Six mutant mexR genes (from more than 40 noninteracting mutants originally selected and sequenced) producing a stable MexR (-LexA) product (Fig. 4A) carried single mutations (L28P, L35P, L75P, I104F, M112T, L135F) and were shown in β-galactosidase assays to be defective in PA3719 interaction (i.e., the β-galactosidase activity increased relative to that of wild-type MexR) (Table 2). These mutations, while distributed throughout MexR (L28P, helix 1; L35P, loop between helices 1 and 2; L75P, helix 4; and I104F, M112T, and L135P, helix 5), occurred at generally interior sites of the protein (Fig. 5) and are unlikely to define a PA3719-binding site or identify MexR residues involved directly in binding PA3719. More likely, these mutations impact MexR protein conformation and thereby PA3719 binding.

FIG. 4.

Whole-cell anti-LexA immunoblot of E. coli SU202 expressing mutant versions of (A) MexR-LexA and (B) PA3719-LexA fusions defective in the MexR-PA3719 interaction. (A) E. coli SU202 carrying pSF001 (pDP804::PA3719_WT) and pMS604 derivatives expressing mutant MexR proteins (lane 1, L35P; lane 2, I104F; lane 3, M112T; lane 4, L135F; lane 5, L28P; lane 6, L75P). Note that the band below MexR is a cross-reactive product and not PA3719-LexA, which migrates lower in the gel, as indicated on the right. (B) E. coli SU202 carrying pLK452 (pMS604::mexR_WT) and pDP804 derivatives expressing wild-type (lane 4) or mutant (lane 5, W45A; lane 6, L36P) PA3719 proteins. Lane 1, pDP804 alone; lane 2, pMS604 alone; lane 3, pMS604::mexR and pDP804. The positions of the Jun-LexA (lanes 1 and 3) and Fos-LexA (lane 2) fusions encoded by the native pDP804 and pMS604 vectors, respectively, are indicated. PA3719_L36P (PA3719*) migrates anomalously on gels (compare lanes 4 and 6). Equal loading of all wells was verified by running duplicate gels that were stained with Coomassie brilliant blue.

FIG. 5.

Location in MexR of mutations compromising interaction with PA3719. Structural details of the MexR dimer (PDB accession no. 1LNW [http://www.ncbi.nlm.nih.gov]) are shown, with sticks highlighting the residues that, when mutated, compromised the interaction with PA3719. Residues are labeled on only one monomer.

Interestingly, four of these six mutations (L35P, I104F, M112T, L135F) did not impact MexR repressor activity, at least to any appreciable extent, as shown by the ability of plasmid pDSK519 derivatives carrying the corresponding cloned mutant mexR genes to reduce multidrug resistance in the ΔmexR P. aeruginosa strain K1491 to an extent comparable to that seen with the cloned wild-type mexR gene (Table 3). These results were confirmed directly with a representative noninteracting MexR mutant protein, MexR_I104F, which was able to repress mexA transcription in the in vitro transcription assay described above (Fig. 3C, compare lanes 1 and 4). The four above-mentioned mutant MexR proteins also showed substantial mexAB-oprM repressor activity in the ΔmexR nalC strain K2519, as evidenced by their ability to reduce resistance to MexAB-OprM antimicrobial substrates in this strain when they were supplied in trans (Table 3). This was in contrast to wild-type MexR, whose mexAB-oprM repressor activity was largely ameliorated in a nalC background (the cloned gene yielded little or no change in resistance in K2519, similar to the vector control) (Table 3). Since nalC strains show increased expression of PA3719 that is essential for elevated mexAB-oprM expression and associated multidrug resistance (8), these data are consistent with a PA3719-MexR interaction ameliorating MexR-DNA binding (Fig. 2) and thus repression of mexAB-oprM (Table 3). As such, mutant MexR proteins defective in the PA3719 interaction are unresponsive to PA3719 modulation and hence are able to bind and repress mexAB-oprM in the presence of PA3719, at least in vivo. To test this in vitro, a representative PA3719-binding-deficient MexR mutant protein (MexR_I104F) was purified and examined for PA3719 interaction, again using ITC analysis. As expected, the mutant protein had a lower affinity for PA3719 than wild-type MexR, characterized by a 5.4-fold decrease in the K_D (K_D, 853.2 ± 151.1 nM) (Fig. 1B). Consistent with these results, while PA3719 effectively reversed wild-type MexR repression of mexA transcription in vitro (Fig. 3C, compare lanes 2 and 3), it had only a modest impact on MexR_I104F repression of mexA transcription, providing only a slight increase in transcription (Fig. 3C, compare lanes 4 and 5).

TABLE 3.

Influence of mexR and PA3719 mutations on the antibiotic susceptibility of P. aeruginosa

Strain	MexR protein expressed	PA3719 protein expressed	MIC (μg/ml) of:
			Carbenicillin	Chloramphenicol	Novobiocin	Norfloxacin	Nalidixic acid
K767	Wild type	None	64	32	128	0.5	NDc
K1491	None	None	512	256	1,024	2	ND
K2519	None	Wild type	512	256	1,024	2	ND
K1491(pDSK519)a	None	None	512	128	1,024	2	ND
K1491(pLC66)	Wild type	None	32	16	128	0.5	ND
K1491(pLC67)	L35P	None	64	16	128	0.5	ND
K1491(pLC68)	I104F	None	128	32	128	0.5	ND
K1491(pLC69)	M112T	None	128	32	128	0.5	ND
K1491(pLC70)	L135F	None	64	16	128	0.5	ND
K1491(PLC71)	L28P	None	512	128	1,024	2	ND
K1491(pLC72)	L75P	None	512	128	1,024	2	ND
K2519(pDSK519)a	None	Wild type	512	128	1,024	2	ND
K2519(pLC66)	Wild type	Wild type	256	128	512	2	ND
K2519(pLC67)	L35P	Wild type	64	32	64	0.5	ND
K2519(pLC68)	I104F	Wild type	64	32	128	1	ND
K2519(pLC69)	M112T	Wild type	64	16	64	0.5	ND
K2519(pLC70)	L135F	Wild type	128	64	128	1	ND
K2519(PLC71)	L28P	Wild type	512	128	1,024	2	ND
K2519(pLC72)	L75P	Wild type	512	128	1,024	2	ND
K2276	Wild type	None	32	ND	ND	ND	128
K2276(pMMB206)b	Wild type	None	32	ND	ND	ND	128
K2276(pSF001)	Wild type	Wild type	128	ND	ND	ND	512
K2276(pSF002)	Wild type	W45A	32	ND	ND	ND	128
K2276(pSF003)	Wild type	L36P	32	ND	ND	ND	128

^aPlasmid pDSK519 derivatives encoding the MexR proteins indicated were introduced into P. aeruginosa strains K1491 (ΔmexR) and K2519 (nalC ΔmexR), and antimicrobial susceptibility was assessed. PA3719 is expressed only in K2519, as a result of the nalC mutation.

^bPlasmid pMMB206 derivatives encoding the PA3719 proteins indicated were introduced into P. aeruginosa strain K2276 (nalC ΔPA3719), and antimicrobial susceptibility was assessed.

^cND, not determined.

Mutations in PA3719 compromising interaction with MexR.

To identify residues of PA3719 important for MexR interaction, the PA3719-carrying pDP804 plasmid was mutagenized as described above for mexR and introduced into pMS604-carrying E. coli SU202, and again, LacZ⁺ colonies were selected on MacConkey agar. Despite the fact that dozens of putative noninteracting PA3719 mutants were screened, only two bore single point mutations in PA3719 (yielding L36P and W45A changes in the protein) (Table 2). Immunoblotting confirmed production of pDP804-encoded PA3719_L36P (fused to LexA), which migrated anomalously on SDS-polyacrylamide gel electrophoresis gels (Fig. 4B, lane 6), although the W45A version was not detected (Fig. 4B, lane 5). Still, control studies with pDP804-encoded wild-type PA3719 revealed that the protein (as a LexA fusion) was unstable and, as a result, undetectable in immunoblots in the absence of MexR (data not shown). One interpretation of these results is that an interaction with MexR stabilizes the PA3719 (-LexA) protein in this system and mutations in PA3719 that compromise this interaction (or the absence of its interacting partner) render the protein unstable. Interestingly, both L36 and W45 occur within the only predicted (using PSIPRED [http://bioinf.cs.ucl.ac.uk/psipred/]) secondary structure of note in the protein (an α-helix that extends from D31 to D46), which is likely involved in the MexR interaction. Indeed, an N-terminally truncated PA3719 derivative encompassing this region (S25 to Y53) was able to interact with MexR in the two-hybrid assay (Table 2).

DISCUSSION

In light of MexR's role as a repressor of mexAB-oprM expression, earlier observations of increased production of both MexAB-OprM and MexR in nalC strains of P. aeruginosa were initially hard to explain (8). Clearly, MexR in such mutants was inactive as a repressor (hence, there was increased MexAB-OprM), and increased production of MexR reflected a loss of negative autoregulation. The current study confirms the loss of repressor activity of MexR in nalC strains and demonstrates that this results from an interaction between PA3719 (ArmR), whose expression increases in nalC mutants, and MexR. Thus, MexR is unique among MarR family regulators, which tend to bind or be modulated by chemical effectors (typically aromatic compounds) (15, 19, 51, 67) that also abrogate operator binding and thereby promote expression of MarR family-regulated genes (68). Indeed, while MarR DNA binding and thus repressor activity are modulated by salicylate, an agent known to induce expression of the MarR-regulated marRAB locus in E. coli (3), and the crystallized MarR repressor demonstrates two possible binding sites for this agent, salicylate does not modulate MexR DNA binding and at least one of the proposed salicylate-binding sites of MarR is absent in MexR (4). Clearly then, MexR differs from MarR (and other Mar family regulators) with regard to the mechanism by which its activity is modulated.

Protein effectors of regulatory proteins, including repressors, are not uncommon in bacteria and may reflect a need to integrate multiple signals in controlling regulatory protein activity and/or perhaps a need to respond to a physical signal(s). In certain nitrogen-fixing organisms of the gamma subgroup of the Proteobacteria, for example, expression of the nitrogen fixation or nif genes is under the control of the NifA activator, whose activity is, however, modulated by NifL in response to both oxygen and fixed nitrogen (31). Apparently, NifL negatively controls NifA activity via a stable protein-protein interaction (24, 37) that is modulated by redox changes (which NifL can sense), ligand binding, and interactions with other proteins (31). Similarly, redox and blue light control of expression of photosynthesis genes in Rhodobacter sphaeroides is mediated by a repressor, PpsR, whose activity is modulated by the AppA antirepressor that responds to redox and blue light and forms a complex with PpsR (33). As observed with PA3719 (ArmR) and MexR, AppA-PpsR complex formation abrogates PpsR binding to target DNA (33). In addition, the ComA response regulator and transcription factor for competence development in Bacillus subtilis is also negatively regulated, by RapC, whose binding to ComA inhibits the latter's DNA-binding activity (11). Still, in each of these examples, the effector proteins are large (NifL, 519 amino acids; AppA, 450 amino acids; RapC, 382 amino acids) and capable of ligand and/or cofactor binding (31, 33), something that PA3719 (ArmR), which consists of 52 amino acids, is unlikely to be capable of. Interestingly, a small protein effector/antirepressor reminiscent of ArmR has been described in Myxococcus xanthus, where it mediates light inducibility of carotenoid biosynthesis (30, 66). The 111-amino-acid protein, CarS, modulates the activity of a CarA repressor that controls expression of carotenoid biosynthetic genes (crtEBDC) by binding to CarA and preventing the latter's binding to its operator sequence in the crtEBDC promoter (30, 66). In contrast to AppA, however, CarS does not respond directly to light; rather, its expression is light inducible and mediated by the ECF sigma factor CarQ, whose release by the anti-sigma factor CarR occurs in response to light (17, 34). Other small protein effectors that have been described include SinI, a 56-amino-acid residue antagonist of the SinR negative regulator of sporulation in B. subtilis (6), and the 88-amino-acid Hpr and 85-amino-acid Crh effectors found in low-G+C-content gram-positive bacteria and Bacillus sp., respectively, which function as corepressors (with CcpA) in carbon catabolite regulation in these organisms (16, 22). While sinI expression itself (and SinI activity as an antirepressor) responds to unknown environmental stimuli that promote sporulation (6), the Hpr and Crh effectors are controlled by phosphorylation, with the phosphorylated forms of these proteins only able to bind CcpA and promote repressor binding to target promoters (16). PA3719 (armR) expression is under the control of the NalC repressor, and so modulation of NalC by some hitherto unknown signal(s) to effect PA3719 production may be sufficient for PA3719 to act on MexR, although we cannot rule out the possibility that some modification of the antirepressor might also be involved. In any case, the nature of the signal(s) to which NalC responds remains unknown, as do the reasons for MexAB-OprM recruitment under conditions where NalC repression of PA3719 (ArmR) is alleviated. As well, the functional significance of the PA3720 gene that forms an operon with the PA3719 gene is unclear; the deduced protein product of this gene shows no significant homology to any known protein, and no observable phenotype is associated with either overexpression or loss of this gene. One possibility, however, is that it modulates NalC repressor activity in response to the “signals” that require PA3719 and, ultimately, mexAB-oprM expression.

How PA3719 (ArmR) negatively impacts MexR DNA binding and repression of target gene (mexR and mexAB-oprM) expression is uncertain, although the identification of mutations in MexR that compromise its interaction with this effector provides some room for speculation. Three of the six mutations disrupting the PA3719 (ArmR) interaction (L35P, L75P, and I104F) occur in regions of MexR which, if substantially altered by PA3719 binding, would likely alter the structure and/or disposition of the DNA recognition helix α-4. Clearly, this would be expected to adversely impact MexR binding to its operator sequences in the mexR-mexAB-oprM intergenic region. Still, it is unclear whether PA3719 (ArmR) acts directly on the α-4 helix, thereby interfering with the MexR-DNA interaction, or indirectly by altering the relative positioning of the individual DNA-binding domains of the MexR homodimer. MexR operator binding probably requires that its recognition helices fit into successive major grooves in the DNA, and thus, the spacing between the two recognition helices in the MexR dimer must match the major groove spacing in the target DNA. Thus, one way in which MexR DNA binding might be modulated and its repressor activity controlled is by altering the spacing of the DNA-binding helices (28), as has been seen for other repressor proteins (e.g., TetR [41] and FadR [65]).

Mutations in MexR distant from α-4 also negatively impact PA3719 (ArmR) binding, and it is impossible to predict with any certainty how these mutations might impact MexR structure and thus compromise PA3719 (ArmR) binding. Significantly, four of six mutations compromising PA3719 (ArmR) binding did not adversely affect MexR repressor activity, indicating that the lack of PA3719 (ArmR) binding is not explained by a gross alteration in the MexR structure but is explained by local perturbations specifically impacting PA3719 (ArmR) binding. This in itself suggests that the DNA-binding and PA3719 (ArmR)-binding domains may, in fact, be separate entities in MexR. This is reminiscent of another MarR family regulator, OhrR, which is involved in organic hydroperoxide resistance in several bacteria and whose effector- and DNA-binding regions are also distinct (19), but it is in contrast to MarR (68) (and CarA [44]), whose effector- and DNA-binding sites overlap. Ongoing efforts to determine the structure of the PA3719-MexR complex should identify the site of interaction and thus provide insights into the mechanism of action of this protein modulator of MexR activity.