Outer Membrane Protein OmpW Participates with Small Multidrug Resistance Protein Member EmrE in Quaternary Cationic Compound Efflux

Abstract

In Escherichia coli, the small multidrug resistance (SMR) transporter protein EmrE confers host resistance to a broad range of toxic quaternary cation compounds (QCC) via proton motive force in the plasma membrane. Biologically produced QCC also act as EmrE osmoprotectant substrates within the cell and participate in host pH regulation and osmotic tolerance. Although E. coli EmrE is one of the most well-characterized SMR members, it is unclear how the substrates it transports into the periplasm escape across the outer membrane (OM) in Gram-negative bacteria. We tested the hypothesis that E. coli EmrE relies on an unidentified OM protein (OMP) to complete the extracellular release of its QCC. Eleven OMP candidates were screened using an alkaline phenotypic growth assay to identify OMP involvement in EmrE-mediated QCC efflux. E. coli single-gene deletion strains were transformed with plasmid-carried copies of emrE to detect reduced-growth and rescued-growth phenotypes under alkaline conditions. Among the 11 candidates, only the ΔompW strain showed rescued alkaline growth tolerance when transformed with pEmrE, supporting the corresponding protein's involvement in EmrE osmoprotectant efflux. Coexpression of plasmids carrying the ompW and emrE genes transformed into the E. coli ΔompW and ΔemrE strains demonstrated a functional complementation restoring the original alkaline loss-of-growth phenotype. Methyl viologen drug resistance assays of pEmrE and pOmpW plasmid-complemented E. coli ΔompW and wild-type strains found higher host drug resistance than with other plasmid combinations. This study confirms our hypothesis that the porin OmpW participates in the efflux of EmrE-specific substrates across the OM.

INTRODUCTION

In Gram-negative bacteria, toxins and antimicrobial agents must traverse both the outer membrane (OM) and plasma membrane (PM) to gain entry into the cell. One of many effective cellular resistance strategies involves the extrusion of the compound itself from the cell by the activities of various single and multicomponent transporters/channels, which may reside in one or both membranes (as reviewed in references 1 and 2). In general, Gram-negative multidrug efflux transport systems fall into two main categories (for a review of these systems, refer to references 3 and 4), multipartite efflux systems composed of proteins that span both the PM and OM membranes, such as the well-characterized resistance-nodulation-cell division (RND) efflux system AcrAB-TolC (as reviewed in reference 5), or independent membrane efflux systems, each of which transports compounds across its respective membrane, such as the general diffusion porins OmpF or OmpC in the OM (reviewed in reference 6) or the metabolite and toxin extrusion (MATE) system member MdfA in the PM (reviewed in reference 7). The multidrug resistance transporter ethidium multidrug resistance protein E (EmrE), a member of the small multidrug resistance (SMR) protein family, relies on secondary active antiport of drugs in exchange for protons across the PM. It is currently believed to function independently, and as implied by the family name, members of the SMR protein family are short in length (EmrE is 110 amino acids in length), spanning the PM as four alpha-helical transmembrane (TM) segments (8, 9). The active site of EmrE has been shown to involve a highly conserved, negatively charged glutamate residue (E14) located in the first TM segment, which participates in both substrate and proton binding (10–13). Despite their small size, SMR proteins confer bacterial host resistance to a broad range of chemically diverse quaternary cationic compounds (QCC) that are used agriculturally, medically, or industrially as cationic surfactants, antiseptics, herbicides, and lipophilic dyes, such as tetraphenylphosphonium, benzalkonium, methyl viologen, and ethidium bromide, respectively (8). In addition to QCC efflux, EmrE also transports osmoprotectants, namely, betaine and choline, indicating a role in host osmotic tolerance (14). The findings of this same study demonstrated that emrE gene expression decreased Escherichia coli alkaline pH tolerance, resulting in significant reduction-in-growth phenotypes under minimal medium conditions (8).

Although EmrE serves as the archetypical member of the SMR protein family, the nature of its transport mechanism remains uncertain and is a topic of interest and debate (mechanisms are reviewed in reference 8). Isogenic overexpression of emrE in E. coli confers host resistance to a variety of toxic QCC (8, 9), suggesting that EmrE functions independently in the PM. Support for this notion can be drawn from observations that SMR genes have demonstrated both chromosomal and lateral inheritance, where genes encoding EmrE homologues are frequently transmitted on mobile genetic elements (integrons) and resistance plasmids (15). Currently, the fate of QCC transported into the periplasm by EmrE is unknown, with one of two possibilities suggested for E. coli. The first method involves escape of QCC from the periplasm through the OM, involving an unidentified channel or porin. Alternatively, EmrE may be reliant on another assisting transporter system(s), like the multipartite AcrAB-TolC efflux complex, and delivers QCC to this efflux system. Evidence supporting the involvement of an as yet unidentified OM protein (OMP) is far more likely, since E. coli strains lacking acrAB-tolC system genes rely on other functional multidrug resistance transporters, including EmrE (16, 17).

The purpose of this study was to confirm if EmrE relies on the presence of an OMP to complete the extracellular efflux of its substrates in Gram-negative bacteria. To accomplish this, an alkaline pH-based phenotypic E. coli growth assay was selected as a screening method to identify any OMP gene(s) capable of rescuing EmrE-induced loss-of-growth phenotypes observed in minimal medium at pH 9 developed in a previous study (14). This phenotypic screening method provided an ideal growth assay to screen single OMP gene deletions in E. coli transformed with pEmrE. The screen was based on the observation that overexpression of emrE in E. coli grown at high pH causes a loss-of-growth phenotype due to the intracellular loss of osmoprotectants, such as betaine (14). If an OMP participates in osmoprotectant efflux with EmrE by completing substrate removal from the cell, the deletion of this OMP should restore growth under alkaline conditions by preventing osmoprotectant loss. Eleven OMP genes (Table 1) were selected for screening to identify their potential involvement in EmrE-mediated efflux of osmoprotectant and QCC methyl viologen. Using this screening approach, in combination with methyl viologen resistance (MV) E. coli growth assays, we have determined that EmrE relies on the presence of an OMP, specifically OmpW, to complete its extracellular substrate efflux across the OM.

TABLE 1.

E. coli K-12 single-gene knockout Keio collection mutants used for this study (18)

Strain	Gene deletion	Genotype	Kanamycin resistance
BW25113		F− Δ(araD-araB)567 ΔlacZ4787(::rrnB-3) λ− rph-1 Δ(rhaD-rhaB)568 hsdR514	No
JW0531	emrE	F− Δ(araD-araB)567 ΔlacZ4787(::rrnB-3) ΔemrE750::kan λ− rph-1 Δ(rhaD-rhaB)568 hsdR514	Yes
JW0731	pal	F− Δ(araD-araB)567 ΔlacZ4787(::rrnB-3) λ− Δpal-790::kan rph-1 Δ(rhaD-rhaB)568 hsdR514	Yes
JW0799	ompX	F− Δ(araD-araB)567 ΔlacZ4787(::rrnB-3) λ− ΔompX786::kan rph-1 Δ(rhaD-rhaB)568 hsdR514	Yes
JW0912	ompF	F− Δ(araD-araB)567 ΔlacZ4787(::rrnB-3) λ− ΔompF746::kan rph-1 Δ(rhaD-rhaB)568 hsdR514	Yes
JW0940	ompA	F− Δ(araD-araB)567 ΔlacZ4787(::rrnB-3) λ− ΔompA772::kan rph-1 Δ(rhaD-rhaB)568 hsdR514	Yes
JW1248	ompW	F− Δ(araD-araB)567 ΔlacZ4787(::rrnB-3) λ− ΔompW764::kan rph-1 Δ(rhaD-rhaB)568 hsdR514	Yes
JW1371	ompN	F− Δ(araD-araB)567 ΔlacZ4787(::rrnB-3) λ− ΔompN740::kan rph-1 Δ(rhaD-rhaB)568 hsdR514	Yes
JW2047	wza	F− Δ(araD-araB)567 ΔlacZ4787(::rrnB-3) λ− Δwza-760::kan rph-1 Δ(rhaD-rhaB)568 hsdR514	Yes
JW2203	ompC	F− Δ(araD-araB)567 ΔlacZ4787(::rrnB-3) λ− ΔompC768::kan rph-1 Δ(rhaD-rhaB)568 hsdR514	Yes
JW3474	slp	F− Δ(araD-araB)567 ΔlacZ4787(::rrnB-3) λ− Δslp-761::kan rph-1 Δ(rhaD-rhaB)568 hsdR514	Yes
JW3698	bglH	F− Δ(araD-araB)567 ΔlacZ4787(::rrnB-3) λ− rph-1 ΔbglH751::kan Δ(rhaD-rhaB)568 hsdR514	Yes
JW5503	tolC	F− Δ(araD-araB)567 ΔlacZ4787(::rrnB-3) λ− ΔtolC732::kan rph-1 Δ(rhaD-rhaB)568 hsdR514	Yes

MATERIALS AND METHODS

Materials and strains used in this study.

A list of all E. coli strains and genotypes used in this study is provided in Table 1. All E. coli strains used in alkaline phenotypic growth screens, complementation, and methyl viologen resistance assays were provided by the National BioResource Project E. coli K-12 single-gene-knockout Keio Collection (18). Nucleotide primers used for PCR cloning were obtained from Integrated DNA Technology (Iowa, USA) and are listed in Table 2. All chemicals used in the preparation of media and chemical solutions for molecular biology experiments and resistance assays were supplied by Sigma (Missouri, USA), EMD (Darmstadt, Germany), or BD Biosciences (New Jersey, USA).

TABLE 2.

Plasmids used and modified in this study and primers used for cloning OMP genes into the multiple cloning site of pMS119EHC

Vector name	Antibiotic resistance	Gene cloned	Vector origin/reference	Primer sequence(s) used for OMP gene cloning
pMS119EHA	Amp	bla	21
pMS119EHC	Cm	cat	This study	DraI restriction fragment from pCA24N containing cat operon
pCA24N	Cm	cat	22
pEmrE	Amp	emrE	20
pOmpW	Cm	ompW	This study	Forward, 5′ATATTCTAGAAGGAGAAATAATATGAAAAAGTTAACAGTG 3′
				Reverse, 5′ATATAAGCTTTTAAAAACGATATCCTGCTGAGAACATAAA 3′
pOmpA	Cm	ompA	This study	Forward, 5′ATATTCTAGAAGGAGAAATAATATGAA 3′
				Reverse, 5′ATATAAGCTTTTAAGCCTGCGGCTGAG 3′
pLamB	Cm	lamB	This study	Forward, 5′ATATTCTAGAAGGAGAAATAATATGAT 3′
				Reverse, 5′ATATAAGCTTTTACCACCAGATTTCCA 3′

Cloning and plasmids used in this study.

All plasmids used and constructed for this study are listed in Table 2. The E. coli strain DH5α was used for all cloning procedures. Standard protocols were used for bacterial transformations and PCR amplification as described in reference 19, both methods being involved in cloning OMP genes from E. coli BW25113 (W3110) genomic DNA. Plasmid isolations were performed using e.Z.N.A. MiniPrepII kits (Omega BioTek, USA). All cloning experiments were performed using the plasmid pMS119EHA (with an ampicillin [Amp] resistance marker, bla), which contained the cloned emrE gene (constructed in a previous study [20]). This expression plasmid permits lactose-inducible expression of cloned genes using a PtacI-inducible promoter and rrnB (Trp) transcription termination system in pMS119EHA (21). This expression system was selected because it exhibited “leaky” nontoxic expression of a hexahistidinyl-tagged emrE gene without the induction by isopropyl β-d-1-thiogalactopyranoside (IPTG) (14). As a result, IPTG was not used to induce expression for this work.

The construction of a chloramphenicol (Cm)-resistant pMS119EHC expression vector was necessary for complementation growth experiments. This involved replacing the bla gene from pMS119EHA with the Cm resistance cat gene from the pCA24 plasmid using DraI restriction sites (22). The E. coli genes ompW, lamB, and ompA were amplified by PCR using E. coli BW25113 (W3110) genomic DNA with the primer pairs listed in Table 2. Amplicons were cloned into the multiple cloning site of the newly constructed pMS119EHC plasmid at XbaI and HindIII restriction sites, and all gene sequences were verified by DNA sequencing from the upstream PtacI to the downstream rrnB (Trp) transcription termination regions (Eurofins MWG Operon, Ebersberg, Germany). Plasmid constructs pLamB, pOmpA, and pOmpW were used for growth phenotype complementation screens of E. coli ΔompW (JW1248) (Table 2).

Alkaline growth phenotype screens of candidate OMP mutants.

To perform the phenotype screen, 11 E. coli OMP mutants and the wild-type (BW25113) strain (Table 1) were transformed with pEmrE or the control vector pMS119EHA (Table 2). These plasmid-transformed E. coli strains were inoculated into lysogeny broth (LB) medium (1% [wt/vol] yeast extract, 0.5% [wt/vol] tryptone, and 0.5% [wt/vol] NaCl) with 100 μg/ml Amp and grown for 16 h (overnight) at 37°C. Overnight cultures were diluted 10⁻³ into 3.0 ml of phosphate-buffered (pH value of 7, 8, or 9) minimal 9 salts (M9) medium (1.3% [wt/vol] NaH₂PO4 · 7H₂O, 0.3% [wt/vol] K₂HPO4, 0.05% [wt/vol] NaCl, 0.1% NH₄Cl, 1.6 × 10⁻⁵% [wt/vol] MgSO₄, 9.0 × 10⁻⁷% [wt/vol] CaCl₂, and 0.00015% [wt/vol] thiamine) supplemented with 0.01% (wt/vol) glucose. Buffered M9 culture 10⁻³ dilutions of all 12 E. coli strains, transformed with either pEmrE or pMS119EHA, were grown in shaking incubators at 37°C, and optical density (OD) measurements were taken at 600 nm (OD₆₀₀) after 16 h of incubation. The 16-h time point was selected based on the observed optimal growth curve endpoint differences as previously described for the wild-type strain, BW25113 (WT) (14). Mean 16-h OD₆₀₀ measurements were calculated from five independently inoculated pH 7 to 9 M9 phenotype growth screen trials, and pairwise Student's t test calculations were performed between empty vector (pMS119EHA)- and pEmrE-containing strains and/or between pEmrE-transformed WT and OMP deletion strains to determine significant differences (P ≤ 0.005). Percent growth recovery values were calculated using mean 16-h OD₆₀₀ values for each plasmid-transformed strain set at a particular pH M9 growth condition by dividing the mean pEmrE-ΔOMP strain 16-h OD₆₀₀ by the mean pMS119EHA-ΔOMP strain 16-h OD₆₀₀ (with both strains grown in the same M9 medium at pH 7, 8, or 9), multiplied by 100. Any surveyed ΔOMP gene deletion strains that rescued growth at a minimum of 80% or higher than wild-type values were deemed to indicate positive OMP candidates identified in the screen. The minimum 80% threshold value used to identify OMP candidates from this screen was determined using standard deviation values from 14-h to 16-h mean OD₆₀₀ growth curve experiments performed with pEmrE-transformed WT strains to set minimum percent confidence intervals according to standard errors. Only one candidate emerged from this pH-based phenotypic growth screen, the gene deletion strain JW1248 (ΔompW), and ompW was the sole experimental focus in the remaining sections below.

Plasmid complementation alkaline growth assays.

Based on the experimental results generated by the alkaline pH growth phenotype screens, the E. coli deletion strain JW1248 (ΔompW) was selected for plasmid complementation analysis. The ΔompW (JW1248) strain was cotransformed with the pOmpW and pEmrE plasmids to determine if the EmrE-induced loss-of-growth phenotype conferred by transformation could be produced by reintroducing ompW back into ΔompW strains under alkaline (pH 9) M9 growth conditions. All buffered M9 media contained 100 μg/ml Amp and 34 μg/ml Cm to maintain both cotransformed plasmids in each strain tested. All alkaline M9 growth phenotype complementation assays of cotransformed ΔompW strains and reverse complementation of cotransformed ΔemrE strains were set up and monitored as described for pH 7 to 9 M9 growth phenotype experiments, where mean 16-h OD₆₀₀ values from five independent inoculation culturing experiments were measured. Statistical analysis of growth differences between pEmrE and pOmpW cotransformants and control cotransformants were determined using a Student two-tailed t test, and P values of ≤0.005 were considered to be significantly different. Percent growth recovery calculations similar to those described above for pH phenotypic assays were used to evaluate complementation assays by adjusting the following formula: [(mean 16-h OD₆₀₀ for pEmrE or OMP-carrying-plasmid cotransformant strain)/(mean 16-h OD₆₀₀ for strain with control plasmids pMS119EHA and pMS119EHC)] multiplied by 100%. The percent growth recovery values for empty vector-containing strains were calculated by dividing mean 16-h OD₆₀₀ values by the mean 16-h OD₆₀₀ values for the growth of the same strains lacking plasmids. Since the pMS119EHA plasmid (derived from pBR322) (21) is present in low copy number (20 to 30 copies/cell) (23) and expression (14, 24) from each cotransformant is approximately half the amount of that for single pEmrE transformants, the cutoff value for significant complementation was set to 60% based on the 30% growth recovery values observed for single pEmrE transformants in the WT.

Methyl viologen resistance assays.

Plasmid-cotransformed E. coli strain QCC MIC plating assays were performed using the EmrE substrate, methyl viologen dichloride (MV) (25, 26) to confirm if the combination of pEmrE and pOmpW coexpression increased host drug tolerance. Plasmid-cotransformed WT (BW25113) and ΔompW (JW1248) E. coli strains used in complementation assays were inoculated onto M9 agar (M9 medium with 1.5% [wt/vol] agar) with increasing concentrations of MV, ranging from 5 μg/ml to 200 μg/ml, in addition to growth tolerance control plates lacking MV. Spot dilutions of overnight LB (with 100 μg/ml Amp and 34 μg/ml Cm) cultures of E. coli strains cotransformed with the pEmrE (Amp) and pOmpW (Cm) plasmids and appropriate empty control vector combinations were spotted (5 μl) at 10¹ to 10⁻⁴ dilutions onto MV containing M9 agar media. All pH 7 M9 agar plates contained 100 μg/ml Amp and 34 μg/ml Cm antibiotics to maintain both cotransformed plasmids. M9 medium was selected instead of using LB plates to limit the influence of other compounds present in undefined LB medium so that only the effect of MV on plasmid-cotransformed-strain resistance was observed. Spotted colony growth was measured as “+” or “−” colony formation after 16, 24, and 48 h of incubation at 37°C, and the MIC of MV for each cotransformant strain was based on results obtained from five independent plating assay trials. Statistical analysis was performed to calculate the significance of growth results, and P values of <0.005 were deemed to show significant difference according to Student's t test calculations.

RESULTS

Only ompW gene deletion rescues the loss-of-alkaline-growth phenotype induced by pEmrE.

To validate our hypothesis that an OMP is associated with EmrE-mediated substrate efflux, an initial pH-based phenotypic growth screen of Keio collection E. coli strains carrying single OMP gene deletions (18) was performed in M9 medium. If a specific OMP is involved in the extracellular efflux of an osmoprotectant in alkaline-buffered minimal medium, the deletion of that OMP gene in an E. coli strain overaccumulating the EmrE protein should rescue the loss-of-growth phenotype by preventing its complete release from the cell.

The selection of 11 E. coli OMP single-gene deletions (Table 1) was determined by pore- and/or channel-forming potential based on OM proteomic experiments (27) and/or their known association with host multidrug resistance (3, 6). To conduct the alkaline M9 growth phenotype screen, all 11 E. coli OMP gene deletion strains and the WT strain, which served as a control for phenotype reproducibility, were transformed with the pMS119EHA and pEmrE plasmids. To identify if any OMP candidates were associated with EmrE-mediated osmoprotection under pH 7 to 9 M9 growth conditions, the relative differences in mean 16-h OD₆₀₀ values (representing cell growth) were determined between the same strain transformed with pMS119EHA or pEmrE to provide the percent growth recovery. Based on the outcome of these growth experiments at pH 7 to 9, only pH 9 ΔOMP growth screens produced significant and reliable differences in percent growth recovery values and growth phenotypes (Fig. 1; see also Fig. S1 in the supplemental material). Hence, only the results for pH 9 ΔOMP screens will be discussed in the following sections. One notable exception was the result obtained for the Δpal deletion strain, which failed to grow above pH 8. As a result, percent growth recovery values for this mutant at pH 8 are shown in Fig. 1. The results of the ΔOMP alkaline growth phenotype screen are provided in Fig. 1 and show that only the ΔompW mutation significantly rescues the loss-of-growth phenotype caused by emrE overexpression by more than 30% in comparison to results for the WT and all other ΔOMP strains tested (Fig. 1). To ensure that 16 h was an appropriate time frame to determine the optimal phenotype (based on previous work [14]), growth curve experiments were performed with pMS119EHA- or pEmrE-transformed ΔompA, ΔompW, and WT strains. Similar to previous study findings, optimal phenotypic growth differences in pH 9 M9 medium were most significant after 14 h of growth and reconfirmed that mean OD₆₀₀ measurements at 24 h were statistically similar between all transformed strains (Fig. 2). Hence, this alkaline phenotypic growth screening method identified that ΔompW in E. coli rescued the loss-of-growth phenotype caused by EmrE osmoprotectant efflux, and it was selected for further plasmid complementation analysis to confirm its involvement.

FIG 1.

Growth recovery screens of 11 plasmid-transformed E. coli ΔOMP strains in alkaline pH 9 M9 media after 16 h at 37°C. The percentages of growth recovery, shown as a bar chart, represent the change in the mean 16-h OD₆₀₀ (growth) of each strain transformed with pEmrE compared to results for the vector pMS119EHA. Percent growth recovery values shown for Δpal strain screens reflect pH 8 M9 growth endpoint values, since this strain failed to grow at pH 9. All strains except the ΔompW pEmrE transformant showed statistically significant reductions (P ≤ 0.005) in growth between the pEmrE-transformed strain and the empty control vector (pMS119EHA)-containing strain.

FIG 2.

Growth curve experiments with E. coli WT, ΔompA, and ΔompW single-gene deletion strains transformed with the pMS119EHA and pEmrE plasmids. Mean OD₆₀₀ values (y axis) represent the growth of cultures measured every 2 h in pH M9 medium at 37°C over a 24-h time frame (x axis) for the E. coli WT (A), ΔompA (B), or ΔompW (C) strain transformed with pMS119EHA (black circles) or pEmrE (gray squares). Error bars represent the standard errors determined for each mean OD₆₀₀ value.

Complementation by pEmrE and pOmpW reproduces the loss-of-alkaline-growth phenotype.

Based on the outcome of the alkaline M9 phenotypic growth screening experiments, the ΔompW (JW1248) gene deletion strain was selected for complementation assays to confirm involvement of OmpW in osmoprotectant efflux induced by EmrE. The expectation was that if the ompW gene was in fact responsible for restoring host alkaline tolerance in M9, due to the efflux of osmoprotectant substrates by EmrE, its reintroduction into the ΔompW strain containing pEmrE would restore the loss-of-growth phenotype by complementing the deleted OMP. To ensure that the complementation was specific to emrE and ompW expression, plasmids carrying an unrelated ompA or lamB OMP gene (pOmpA or pLamB) were also cotransformed with pEmrE in ΔompW strains. The results from ΔompW strain plasmid complementation growth assays in pH 9 M9 medium after 16 h of incubation are shown in Fig. 3A. Based on the outcome of the assay, only cotransformation of ΔompW strains with pEmrE and pOmpW resulted in a statistically significant reduction in growth compared to results for all other plasmid pair cotransformation combinations, confirming that reintroduction of ompW with emrE reproduced the loss-of-growth phenotype (47% ± 7%) under alkaline growth conditions (Fig. 3A). Our confidence in this finding is high, since a 92% complementation was achieved specifically for ΔompW pOmpW and pEmrE transformants and fell within the ≤60% complementation cutoff value. Transformations with either pOmpA or pLamB and pEmrE did not significantly alter percent growth recovery values (Fig. 3A). A reverse complementation experiment was also performed for the E. coli ΔemrE (JW0531) strain cotransformed with pairwise combinations of empty vectors pEmrE and pOmpW, shown in Fig. 3B. Similar to the outcome of ΔompW complementation assays, only cotransformation of pEmrE and pOmpW in the ΔemrE strain specifically resulted in reduced percent growth recovery values (58% ± 7.6%), again below the 60% cutoff (Fig. 1 and 3B). Hence, both plasmid complementation assays confirmed that only the complementation of both emrE and ompW in a ΔemrE or ΔompW strain reproduced the loss-of-growth phenotype due to the efflux of osmoprotectants by both membrane proteins when the strains were grown under alkaline conditions.

FIG 3.

Plasmid complementation of the E. coli ΔompW (A) or ΔemrE (B) strain cultured using alkaline (pH 9) M9 growth phenotype assays. Both panels show the percent growth recovery of plasmid-cotransformed single-gene deletions strains as bar charts, reflecting the change in mean 16-h OD₆₀₀ (growth) between the pEmrE-transformed strain and the pMS119EHA strain. Statistically significant reductions in percent growth recovery values (P ≤ 0.005) were observed only for pEmrE and pOmpW in both panels and are indicated by asterisks. Dashed lines on each chart indicate the cutoff value (60%) used to determine significant complementation.

ΔompW and WT strains transformed with both pEmrE and pOmpW conferred host resistance to the highest concentrations of methyl viologen.

Since the majority of known EmrE substrates are antimicrobial QCC, it was essential to confirm if the newly identified OmpW also conferred QCC resistance when overaccumulated with EmrE. A previous study demonstrated that overexpression of ompW in Salmonella enterica serovar Typhimurium contributed to MV resistance (28). MV is a QCC herbicide and a well-known substrate of EmrE; therefore, this compound was selected for further drug resistance assays. To determine if both OmpW and EmrE contribute to MV resistance in E. coli, plasmid-cotransformed ΔompW and WT cultures were grown on M9 plates containing 5-fold-increasing concentrations of MV (5 μg/ml to 200 μg/ml) to determine the extent of host resistance conferred over a 16-h to 48-h incubation period. The MIC values determined from these resistance assays are summarized in Table 3. Our results indicate that after 16 h of incubation, only 10⁻⁴ dilutions of both WT and ΔompW strains transformed with pEmrE and pOmpW demonstrated the greatest MV MIC values (Table 3). At higher culture dilutions (10⁻¹ and 10⁻³), either strain transformed with pEmrE or pOmpW also showed a clear increase in MV MIC values in comparison to results for empty vector transformants. After 24 h, 10⁻⁴ dilutions of both WT and ΔompW strain cotransformants had high MV MIC values only for strains with plasmid pairings of pEmrE-pOmpW, as well as pEmrE and pMS119EHA, indicating that resistance conferred by the cotransformations became indistinguishable over time. Interestingly, 10⁻³ culture dilutions at 16 h and 24 h indicated that strains containing pOmpW and pMS119EHA also had high MV resistance values but at half the concentrations conferred by the pEmrE-pMS119EHC or pEmrE-pOmpW combination (Table 3). This is likely due to variations in the copy number of pOmpW relative to that of pMS119EHA/C or to pEmrE maintained within cells. A longer incubation period of 48 h resulted in no differences in MV resistance even at the lowest dilutions tested (10⁻⁵ dilutions), where MIC values ranged from 5 to 15 μg/ml for all cotransformants, except for strains with empty vector pairs (Table 3). Overall, the findings of these MV resistance assays confirmed that the presence of both emrE and ompW conferred host resistance to the highest concentrations of MV and that both membrane proteins participate in MV efflux.

TABLE 3.

MV MICs for culture dilutions of E. coli ΔompW and WT strain plasmid cotransformants grown on M9 medium agar^a

Strain tested	Plasmid-cotransformed combination tested	MIC (μg/ml) of spotted dilution at:
		16 h			24 h
		10−2	10−3	10−4	10−3	10−4	10−5
WT	pMS119EHA and pMS119EHC	0	0	0	0	0	0
	pEmrE and pMS119EHC	20	10	0	10	10	5
	pOmpW and pMS119EHA	5	0	0	5	0	0
	pEmrE and pOmpW	20	15	10	15	10	10
ΔompW	pMS119EHA and pMS119EHC	0	0	0	0	0	0
	pEmrE and pMS119EHC	20	10	0	15	10	10
	pOmpW and pMS119EHA	0	0	0	5	0	0
	pEmrE and pOmpW	20	15	5	15	10	10

^aBoldfaced values indicate important differences.

DISCUSSION

The application of an alkaline minimal medium growth phenotype screen, determined based on a previous study examining the osmotic tolerance of E. coli strains overaccumulating EmrE, provided a useful in vivo assay to design and identify a screen to detect if any ΔOMP gene deletion strain was involved in EmrE-mediated drug resistance. This approach served as an ideal screening method to identify whether an OMP was associated with EmrE efflux, since it avoided the use of QCC potentially transported by other redundant multidrug efflux systems that possess overlapping substrate profiles (such as AcrAB-TolC and MdfA) (9, 16). OmpW was one of the most likely candidates prior to these studies, based on its involvement in host MV resistance and osmotic tolerance in other Gram-negative bacteria (28, 29, 30, 31). Taken altogether, this study provides the first evidence of OMP association with EmrE-mediated host drug resistance. It also confirms our hypothesis that an OMP assists in the removal of EmrE substrates, such as osmoprotectants and MV.

Overall, the OM-PM association between EmrE and OmpW may represent important, conditionally active branches of the so-called “resistome” (16), particularly when considering the efflux of MV, since Gram-negative bacteria lacking the dominant QCC/antibiotic RND efflux system, AcrAB-TolC, showed increased resistance to MV through the activation of other multidrug efflux systems (16, 32). This finding also suggests that other multidrug transporters currently known to confer resistance independently in the PM may have unidentified OMP counterparts that could potentially be identified through phenotypic screening methods similar to those used herein. In conclusion, we have demonstrated a functional association between E. coli EmrE and OmpW, indicating that they both participate in the extracellular efflux of QCC and osmoprotectant substrates.