nalD Encodes a Second Repressor of the mexAB-oprM Multidrug Efflux Operon of Pseudomonas aeruginosa

Abstract

The Pseudomonas aeruginosa nalD gene encodes a TetR family repressor with homology to the SmeT and TtgR repressors of the smeDEF and ttgABC multidrug efflux systems of Stenotrophomonas maltophilia and Pseudomonas putida, respectively. A sequence upstream of mexAB-oprM and overlapping a second promoter for this efflux system was very similar to the SmeT and TtgR operator sequences, and NalD binding to this region was, in fact, demonstrated. Moreover, increased expression from this promoter was seen in a nalD mutant, consistent with NalD directly controlling mexAB-oprM expression from a second promoter.

Pseudomonas aeruginosa is an opportunistic human pathogen that demonstrates an innate resistance to multiple classes of antimicrobials (5), a property explained in part by the operation of multidrug efflux systems of the resistance-nodulation division (RND) family (19, 20). Several RND-type multidrug efflux systems have been described in P. aeruginosa (20), although the major system contributing to intrinsic multidrug resistance is encoded by the mexAB-oprM operon (4, 10, 21). MexAB-OprM exports a variety of clinically used antimicrobials (including several classes of antibiotics and biocides) (19, 20) but, as well, several additional agents with antimicrobial activity (e.g., dyes [9], detergents [34], and organic solvents such as aromatic hydrocarbons [11]) and acylhomoserine lactones (AHLs) associated with quorum sensing (3, 18). AHLs play a role in cell density-dependent expression of a number of virulence factors in P. aeruginosa and, thus, the activity of this efflux system can influence virulence (6). The observation that MexAB-OprM overproduction can compromise bacterial “fitness” (29) and that pump-overproducing mutants of P. aeruginosa can be selected in animal models of infection in the absence of antimicrobial selection (7) also point to in vivo functions for this pump independent of antimicrobial efflux and resistance.

Hyperproduction of MexAB-OprM has been documented in so-called nalB mutants (13) carrying lesions in the mexR gene (13, 26, 35, 38) encoding a repressor of mexAB-oprM expression (22, 35). MexR is a member of the MarR family of regulators (16) and binds as a dimer (12) to two sites in the mexR-mexA intragenic region, near mexR and overlapping promoters for mexR and mexAB-oprM (2). MexAB-OprM hyperexpression also occurs independently of mutations in mexR; these so-called nalC mutants (1, 13, 35) carry a mutation in a gene (PA3721, also known as nalC) that encodes a TetR family repressor of an adjacent two-gene operon, PA3720-PA3719 (1). It is, in fact, the increased expression of PA3719 that results from disruption of the nalC repressor gene that promotes mexAB-oprM hyperexpression (1), apparently as a result of PA3719 modulation of MexR repressor activity (L. Cao, S. Fraud, D. Daigle, M. Wilke, A. Pacey, C. Dean, N. Strynadka, and K. Poole, submitted for publication). Recently, mutations in yet a third gene, nalD, also encoding a TetR family repressor-like protein, have been described that are responsible for mexAB-oprM hyperexpression in multidrug-resistant P. aeruginosa (32). We show here that NalD binds to a second promoter upstream of mexAB-oprM, consistent with it directly repressing efflux gene expression, and nalD mutations thus affording multidrug resistance via derepression of mexAB-oprM.

Disruption of the nalD gene (PA3574) was shown previously to increase mexAB-oprM expression and multidrug resistance, consistent with NalD functioning to negatively regulate expression of this efflux system (32). Consistent with this, too, introduction of the nalD gene (on plasmid pMLS003 [Table 1 ]) into P. aeruginosa strain K870 (22) reduced resistance to representative MexAB-OprM antimicrobial substrates, including carbenicillin (MIC reduced from 64 to 32 μg/ml) and chloramphenicol (MIC reduced from 32 to 8 μg/ml). Interestingly, the nalD gene product shows substantial similarity to the SmeT (28) and TtgR (37) repressors of the smeDEF and ttgABC RND-type multidrug efflux operons of Stenotrophomonas maltophilia and Pseudomonas putida, respectively (SmeT, 34% identity; TtgR, 36% identity), with sequence conservation being highest in the first 55 amino acids of the proteins (Fig. 1A). SmeT, TtgR, and NalD are members of the TetR family of regulatory proteins that typically bind target DNA through a prototypical N-terminal helix-turn-helix (HTH) domain (23) and, indeed, such a domain is predicted to occur at the same place within the N termini of each of these proteins (Fig. 1A). Moreover, the predicted HTH motifs of these proteins are highly similar (13 of 21 HTH residues are exact matches in all three proteins [Fig. 1A]), suggesting that they may recognize the same or related nucleotide sequences in their respective target DNAs. TtgR and SmeT have been shown to bind upstream of their respective target efflux operons, with a TtgR-binding site precisely defined and shown to encompass canonical −35/−10 ttgABC promoter sequences (37) (Fig. 1B). Although an actual binding SmeT site has not been delineated, a binding site has been proposed that similarly overlaps the smeDEF promoter region (28) (Fig. 1B). Intriguingly, the TtgR- and SmeT-binding sequences are very similar (Fig. 1B), and a related sequence was identified upstream of the mexAB-oprM operon (23/30 matches with the TtgR- and SmeT-binding sequences) (Fig. 1B). Moreover, this sequence overlaps a putative second promoter for mexAB-oprM (P_II) (Fig. 1B) downstream (i.e., more mexAB-oprM proximal) from the primary mexAB-oprM promoter (P_I) (Fig. 1B) identified previously and shown to be regulated by the MexR repressor (2). This earlier study also provided evidence for mexAB-oprM expression from a second, more mexAB-oprM-proximal promoter, and a second set of canonical −35 and −10 sequences was identified closer to mexAB-oprM, although an exact transcription start site was not identified (2).

TABLE 1.

Bacterial strains and plasmids

Strain or plasmid	Descriptiona	Reference
E. coli BL21(DE3)	F−ompT rB− mB−; DE3 is a lambda derivative carrying lacI and T7 RNA polymerase genes under placUV5 control	36
P. aeruginosa strains
K767	P. aeruginosa PAO1 (prototroph)	14
K870	Spontaneous Strr derivative of K767	22
K2346	K870 nalD::mini-Tn5-tet	32
K2543	K767 nalD::mini-Tn5-tet	This study
K2568	K767 ΔmexR	This study
K2569	K2543 ΔmexR	This study
Plasmids
pDSK519	Broad-host-range cloning vector; IncQ Kmr	8
pLC66	pDSK519::mexR	Cao et al., submitted
pMLS003	pDSK519::nalD	32
pK18MobSacB	Broad-host-range gene replacement vector; sacB Kmr	31
pYM029	pK18MobSacB::nalD::mini-Tn5-tet	This study
pMP190	Broad-host-range, low-copy-number lacZ fusion vector; Camr Strr	24
pYM025	pMP190::PI	This study
pYM026	pMP190::PIIb	This study
pYM030	pMP190::PIIc	This study
pYM031	pMP190::PI+PIId	This study
pRSP75	pEX18Tc::ΔmexR	35

^aStr^r, streptomycin resistant; Km^r, kanamycin resistant; Cam^r, chloramphenicol resistant.

^bRegion cloned into pMP190 extends 60 bp upstream of a putative mexAB-oprM promoter P_II −35 region; identified as PII-1 in Fig. 1B.

^cRegion cloned into pMP190 extends 110 bp upstream of a putative mexAB-oprM promoter P_II −35 region; identified as PII-2 in Fig. 1B.

^dEntire mexR-mexA intergenic region cloned into pMP190.

FIG. 1.

Identification of a putative DNA-binding motif in NalD (A) and a putative NalD-binding site upstream of mexAB-oprM (B) that impacts mexAB-oprM expression (C). (A) Multiple sequence alignment of NalD and the SmeT (GenBank accession number CAC87048) and TtgR (GenBank accession number AAK15050) repressors performed online using ClustalW (http://www.ebi.ac.uk/clustalw/). Exact matches (*) and conserved changes (:) are indicated. Putative DNA-binding helix-turn-helix regions were identified using the online Motif HTH program (http://npsa-pbil.ibcp.fr/cgi-bin/npsa_automat.pl?page = npsa_hth.html) available from the Pôle Bioinformatique Lyonnais website (http://pbil.univ-lyon1.fr/) and are highlighted in boldface italics. (B) Nucleotide sequence of the mexR-mexA intergenic region. Canonical −35/−10 regions for a confirmed mexAB-oprM-distal (−35 [I]/−10 [I]; P_I) and a putative mexAB-oprM-proximal (−35 [II]/−10 [II]; P_II) promoter are highlighted in italicized boldface text. The translational start sites for mexR and mexA are indicated (underlined in boldface text), as are two MexR-binding sites (boxed sequence) overlapping P_I. mexA-distal (delineated by open arrowheads) and -proximal (delineated by filled arrowheads) sequences used in electrophoretic mobility shift assays are also indicated. Promoter P_I- and P_II-containing sequences cloned into the lacZ transcriptional fusion vector pMP190 are delineated with arrows labeled PI and PI-t (for promoter P_I) and either PII-1 and PII-t or PII-2 and PII-t (for promoter P_II). A transcriptional start site identified downstream of the P_II −10 region using RACE (C) is boxed. Sequences from the smeDEF (smeD) and ttgABC (ttgA) promoter regions of S. maltophilia and P. putida, respectively, that encompass putative (smeD) or known (ttgA) binding sites for the cognate repressors SmeT and TtgR are shown below the mexR-mexA intergenic sequence aligned with the putative mexAB-oprM promoter P_II (mexA). Exact matches for all three promoter sequences are shown in uppercase text. The site of a mutation (C→T) yielding increased mexAB-oprM expression and multidrug resistance in a previous paper (22) is highlighted by an asterisk. (C) 5′-RACE product (lane 1) from pLC66-carrying NalD⁻ P. aeruginosa strain K2543 obtained with a 5′-RACE kit-provided 5′-end primer (20-mer) and mexA-specific primer annealing 220 bp into mexA. A transcript initiating from P_II would yield a product of ca. 320 bp. Lane 2, 100-bp ladder.

One possibility, then, is that NalD functions to regulate mexAB-oprM expression directly, by controlling expression of the efflux genes from this second promoter. Indeed, a previous study (25) noted that a C-T transition mutation within the putative NalD-binding region (Fig. 1B) increased mexAB-oprM expression and antibiotic resistance, as would be expected were this region targeted by NalD. To assess NalD binding to the second mexAB-oprM promoter region, attempts were made to construct and purify a polyhistidine (His)-tagged NalD protein for use in an electrophoretic mobility shift assay (EMSA). A His-tagged version of NalD was engineered by amplification of the nalD gene (also known as PA3574) from plasmid pMLS003 using PCR (17) and cloning it into plasmid pET23a (Novagen, Madison, WI) to yield pLC80 (primers and parameters available upon request). Following introduction of pLC80 into the expression strain Escherichia coli BL21(DE3), NalD-His expression induced with 1 mM isopropyl-β-d-thiogalactopyranoside (IPTG; in a 5-ml LB culture supplemented with ampicillin) was measured as described before (33). Immunoblotting (33) using anti-His antibodies (Santa Cruz Biotechnology, Inc., Santa Cruz, CA) confirmed NalD-His expression in whole-cell extracts (24) of pLC80-carrying E. coli BL21(DE3), although levels seen were insufficient to warrant attempts at purification. Crude soluble extracts have, however, been employed previously in assessing protein-DNA interactions in mobility shift assays (28) and, so, soluble whole-cell extracts were prepared from log-phase E. coli BL21(DE3) cells carrying pLC80 (25 ml) and induced with IPTG (1 mM) following disruption of cells by sonication (2). Following centrifugation (300,000 × g; 15 min), the NalD-His-containing supernatant was recovered and the protein concentration was determined using the BCA protein assay kit (Pierce, Rockford, IL).

The ca. 200-bp target DNA to be used in gel shift assays, encompassing the promoter P_II region upstream of mexA but downstream of the previously identified MexR-binding sites (Fig. 1B), was amplified from the chromosome of P. aeruginosa PAO1 strain K767 (14) using PCR. A control DNA (200 bp) upstream of the P_II region and overlapping the MexR-binding region was similarly PCR amplified. The gel shift assay was performed using an EMSA kit (E33075; Molecular Probes, Inc., Invitrogen) according to the manufacturer's instructions. Briefly, 40 ng target DNA was incubated with increasing amounts of NalD-His-containing crude soluble extract (250 ng, 500 ng, and 1,000 ng) for 20 min at room temperature in a 15-μl reaction mixture containing 1× binding buffer (750 mM KCl, 0.5 mM dithiothreitol, 0.5 mM EDTA, 50 mM Tris-HCl, pH 7.4). Following the addition of EMSA gel-loading solution, mixtures were separated by electrophoresis on a nondenaturing 8% (wt/vol) polyacrylamide gel in 0.5× TBE buffer (22 mM Tris-HCl, 22 mM boric acid, 0.5 mM EDTA, pH 8.0) (27), and gels were stained with 1× SYBR Green EMSA nucleic acid stain. DNA was then visualized using digital photography with an S6656 SYPRO photographic filter. As seen in Fig. 2, extracts prepared from E. coli cells expressing NalD-His from pLC80 clearly shifted the mexA-proximal DNA fragment (Fig. 2B, panel I, lanes 2 to 4), while extracts prepared from the control strain carrying the pET23a vector without insert did not (Fig. 2B, panel II, lanes 2 to 4). This is consistent with NalD binding to sequences proximal to mexA. The NalD-containing extract did not, however, shift the mexA-distal fragment that encompasses the P_I promoter region (Fig. 2B, panel III, lanes 2 to 4), indicating that NalD binding in the first instance was specific and that NalD directly regulates mexAB-oprM expression. As such, mexAB-oprM hyperexpression in nalD strains is explainable by loss of NalD repression of efflux gene expression.

FIG. 2.

NalD expression (A) and binding to sequences upstream of mexAB-oprM (B). (A) Western immunoblot of E. coli BL21(DE3) carrying pET23a (lanes 1 and 2) or pLC80 (pET23a::nalD) (lanes 3 and 4) with (lanes 2 and 4) or without (lanes 1 and 3) IPTG induction developed using an anti-His antibody. (B) Mobility shift assay in which soluble protein extracts (lane 1, 0 ng; lane 2, 250 ng; lane 3, 500 ng; lane 4, 1,000 ng) prepared from E. coli BL21(DE3) cells carrying pLC80 (panels I and III) or pET23a (panel II) were incubated with DNA fragments encompassing the mexA-proximal (panels I and II) or -distal (panel III) portions of the mexR-mexA intergenic region. The mexA-proximal and -distal regions are delineated in Fig. 1B by filled and open arrowheads, respectively.

While NalD clearly binds downstream of MexR to sequences more proximal to mexAB-oprM, it remains a possibility that it simply provides for a second negative regulator of efflux gene expression from the distal and, to date, only confirmed promoter, P_I (Fig. 1B), rather than a repressor of expression from a putative second mexAB-oprM promoter, P_II. Evidence for a promoter downstream of P_I was subsequently sought by fusing this region to a promoterless lacZ gene on plasmid pMP190. The region beginning 60 bp upstream of a putative P_II −35 region and ending just before the mexA gene (PII-1) (Fig. 1B) was amplified from the chromosome of P. aeruginosa K767 by PCR and cloned into pMP190, and the resultant P_II-lacZ fusion vector (pYM026) was mobilized into wild-type strain K767 and a NalD⁻ derivative, K2543. The latter was constructed following cloning of a nalD::mini-Tn5-tet-containing PstI fragment from the nalD mutant strain K2346 (Table 1) into the gene replacement vector pK18mobsacB and its subsequent introduction into K767 as described previously (1). Neither of the pYM026-containing strains demonstrated more than minimal β-galactosidase activity (measured using Luria broth-grown log-phase cells as described in reference 15) (data not shown), consistent with little or no promoter activity and in apparent agreement with earlier results (30). Still, when a larger mexA-proximal fragment (extending further upstream to just after the P_I −10 region [i.e., PII-2]) (Fig. 1B) was fused to lacZ in pMP190 (yielding pYM030), β-galactosidase activity was again minimal in NalD⁺ K767 but was substantially increased (14-fold) in the NalD⁻ K2543 strain (Table 2). In contrast, a P_I-lacZ fusion constructed in pMP190 (pYM025) following PCR amplification of the P_I region (Fig. 1B) showed substantial activity in wild-type K767, consistent with P_I being active in wild-type P. aeruginosa, as highlighted previously (2), and its activity was not markedly changed (1.5-fold increase) in the NalD⁻ mutant (Table 2). As expected, the cloned nalD but not the mexR gene reduced P_II-mediated lacZ expression in K2543 (Table 2). These data are consistent with the presence of a mexA-proximal P_II promoter that is specifically NalD controlled. Failure to detect this activity in earlier studies (30) likely stems from its weak activity in the wild-type (i.e., NalD⁺) strain, where it appears to be almost wholly repressed.

TABLE 2.

Activity of mexAB-oprM P_I and P_II promoters in P. aeruginosa^a

Strainb	Plasmid	Promoterc	β-Galactosidase activity (Miller units) ± SDd
K767	—e	PI	232 ± 1
	—	PII	14 ± 4
	—	PI+PII	211 ± 43
K2543 (NalD−)	—	PI	360 ± 11
	—	PII	196 ± 7
	—	PI+PII	489 ± 17
K2568 (MexR−)	—	PI	771 ± 70
	—	PII	25 ± 16
	—	PI+PII	625 ± 127
K2569 (NalD− MexR−)	—	PI	826 ± 133
	—	PII	90 ± 63
	—	PI+PII	524 ± 109
K2543 (NalD−)f	pDSK519	PI	101 ± 23
		PII	55 ± 25
	pMLS003 (nalD)	PI	131 ± 48
		PII	16 ± 7
	pLC66 (mexR)	PI	46 ± 12
		PII	122 ± 31

^aThe indicated P. aeruginosa strains carrying plasmid pMP190 derivates pYM025 (P_I), pYM030 (P_II) and pYM031 (P_I+P_II) with the mexAB-oprM P_I and/or P_II promoter regions cloned upstream of the promoterless lacZ gene of this vector were grown to late log phase and assayed for β-galactosidase activity to obtain a measure of P_I and/or P_II promoter activity.

^bThe relevant phenotypes of the strains are highlighted in parentheses.

^cThe indicated mexAB-oprM promoter regions (delineated in Fig. 1B) were cloned into pMP190 to yield lacZ transcriptional fusions to assess P_I (pYM025), P_II (pYM030), or P_I plus P_II (pYM031) activity. The larger P_II promoter fragment, PII-2 (Fig. 1B), was used in these studies. For P_I plus P_II, the entire mexR-mexA intergenic region was cloned into pMP190.

^dData shown are the means of three independent experiments ± the standard deviation. The data have been corrected for background β-galactosidase activity, measured for control strains carrying pMP190 without promoter insert.

^e—, not applicable.

^fResults for K2543 carrying pDSK519 and its derivatives (pMLS003 = pDSK519::nalD; pLC66 = pDSK519::mexR) in addition to the indicated P_I or P_II fusion vectors do not compare with the remaining data presented here because of the presence of two plasmids and, thus, the need to include two antibiotics in the growth medium.

Interestingly, the active NalD-controlled promoter region described above extends well into the second of two MexR-binding sites that overlap P_I, where MexR might be expected to bind on the lacZ fusion. Still, a pYM030-carrying ΔmexR K767 derivative (K2568) did not show any significant increase in β-galactosidase activity relative to the MexR⁺ K767 parent (although P_I-lacZ did increase, as expected) (Table 2), indicating that the mexA-proximal promoter is not MexR controlled. In contrast to results with NalD⁻ strain K2543, however, loss of nalD in the MexR⁻ strain only modestly (<4-fold) enhanced P_II activity (Table 2, compare K2569 and K2568). Again, P_I activity was unaffected by loss of nalD (Table 2). This suggests that the NalD-controlled mexA-proximal promoter is most active in MexR⁺ cells when MexR is bound to P_I (i.e., in vivo when expression from P_I is abrogated). Any reduction in P_II activity measured with the P_II-lacZ-containing pYM030 in MexR⁻ (versus MexR⁺) cells cannot be due to strong transcription from P_I interfering with transcription from P_II, since pYM030 lacks P_I. One possibility, then, is that a productive P_II conformation can only be achieved when MexR is bound to the P_I region. Consistent with this, overexpression of the cloned mexR gene (from pLC66) increased the P_II activity of pYM030 twofold in NalD⁻ strain K2543 (Table 2).

Given the sequence similarly of the putative P_II promoter (Fig. 1B) with the known promoters of smeDEF and ttgABC, including a region implicated in repressor binding in all three instances, it is tempting to suggest it is a de facto second promoter for mexAB-oprM. Still, the fact that a P_II promoter activity was observed only when sequence extending ca. >100 bp upstream of the putative P_II promoter was engineered into the pMP190 promoter-proving vector suggests that either the indicated P_II is not the NalD-controlled promoter for mexAB-oprM expression or that substantial upstream sequence is somehow needed for functional topology/conformation of this promoter. Using 5′ rapid amplification of cDNA ends (RACE) analysis as before (2) with mRNA isolated from the NalD⁻ P. aeruginosa strain K2543 carrying the mexR plasmid pLC66 (where maximal P_II activity was detectable and expression from P_I was repressed [Table 1]), a single cDNA corresponding to the 5′ end of a mexAB-oprM-specific mRNA was recovered (Fig. 1C). The size of the cDNA (ca. 300 bp) was consistent with transcription having been initiated from the P_II promoter and, indeed, DNA sequencing identified the transcription start site as downstream of the canonical −10 site of the predicted P_II promoter (Fig. 1B). Clearly, then, P_II is a second, NalD-regulated promoter for mexAB-oprM expression in P. aeruginosa.

Fusion of the entirety of the mexR-mexA intergenic region to lacZ on pMP190 (a P_I plus P_II-lacZ fusion; pYM031) yielded activities in K767 and K2543 consistent with P_I being solely responsible for mexAB-oprM expression in wild-type cells (where P_I activity = P_I + P_II) (Table 2), while both contribute in a NalD⁻ strain. Moreover, the ca. twofold increase in P_I plus P_II activity (i.e., total mexAB-oprM promoter activity) seen in NalD⁻ strain K2543 (compared with K767 [Table 2]) is in agreement with the known impact of a nalD mutation on mexAB-oprM expression and multidrug resistance (32). In wild-type MexR⁺ NalD⁺ cells, then, mexAB-oprM expression can be influenced by both regulators responding to their own particular signals that require or reflect a need for MexAB-OprM efflux activity. Thus, mexAB-oprM expression reflects the integration of multiple signals via two direct regulators, MexR and NalD, although the signals to which these regulators respond remain to be elucidated.