Novel Inducers of the Expression of Multidrug Efflux Pumps That Trigger Pseudomonas aeruginosa Transient Antibiotic Resistance

Abstract

The study of the acquisition of antibiotic resistance (AR) has mainly focused on inherited processes, namely, mutations and acquisition of AR genes. However, inducible, noninheritable AR has received less attention, and most information in this field derives from the study of antibiotics as inducers of their associated resistance mechanisms. Less is known about nonantibiotic compounds or situations that can induce AR during infection. Multidrug resistance efflux pumps are a category of AR determinants characterized by the tight regulation of their expression. Their contribution to acquired AR relies in their overexpression. Here, we analyzed potential inducers of the expression of the chromosomally encoded Pseudomonas aeruginosa clinically relevant efflux pumps, MexCD-OprJ and MexAB-OprM. For this purpose, we developed a set of luxCDABE-based P. aeruginosa biosensor strains, which allows the high-throughput analysis of compounds able to modify the expression of these efflux pumps. Using these strains, we analyzed a set of 240 compounds present in Biolog phenotype microarrays. Several inducers of the expression of the genes that encode these efflux pumps were found. The study focused in dequalinium chloride, procaine, and atropine, compounds that can be found in clinical settings. Using real-time PCR, we confirmed that these compounds indeed induce the expression of the mexCD-oprJ operon. In addition, P. aeruginosa presents lower susceptibility to ciprofloxacin (a MexCD-OprJ substrate) when dequalinium chloride, procaine, or atropine are present. This study emphasizes the need to study compounds that can trigger transient AR during antibiotic treatment, a phenotype difficult to discover using classical susceptibility tests.

Keywords: Biolog; MDR efflux pumps; Pseudomonas aeruginosa; transient antibiotic resistance.

FIG 1

Effect of known inducers of mexAB and mexCD expression in luminescence emitted by mexAB and mexCD bioreporter strains. The luminescence and OD₆₀₀ values for PAO1 CTX-lux::PmexAB and PAO1 CTX-lux::PmexCD were measured in the presence of 0.136 μg/ml of H₂O₂ and 10 μg/ml benzalkonium chloride, respectively. Fold changes in the luminescence emitted by strains grown in the presence of the inducers versus strains grown without any inducer are shown. Error bars represent the standard deviations of three independent replicates. As shown, the known inducer compounds increase the production of luminescence of the bioreporter strains.

FIG 2

Effects of different compounds on the expression of either mexAB-oprM or mexCD-oprJ. The figure shows the normalized luminescence values produced by PAO1 CTX-lux::PmexAB and PAO1 CTX-lux::PmexCD in the presence of four different concentrations of 240 compounds from Biolog plates. The outliers of the boxplot represent the conditions under which there was potential overexpression or repression of the genes encoding the studied efflux pumps: values greater than 1.53 for PAO1 CTX-lux::PmexAB and 1.73 for PAO1 CTX-lux::PmexCD indicate overexpression, while values less than 0.7 for PAO1 CTX-lux::PmexAB and 1.1 for PAO1 CTX-lux::PmexCD indicate repression.

FIG 3

Structures of inducers of mexCD-oprJ expression. The structures of atropine, procaine, and dequalinium chloride were obtained from the ChemSpider database.

FIG 4

Effect of dequalinium chloride, procaine, and atropine on the expression of mexCD-oprJ. The figure shows the normalized luminescence values produced by reporter strain PAO1 CTX-lux::PmexCD in the presence of 10 μg/ml of dequalinium chloride, 2 mg/ml of procaine, and 2 mg/ml of atropine. The luminescence was normalized to that produced by PAO1 CTX-lux in the presence of inducer. As shown, the expression of mexCD-oprJ is induced by the three tested compounds. Error bars represent the standard deviations of three independent replicates.

FIG 5

Analysis of mexCD-oprJ expression by real-time PCR in the presence of and after the removal of dequalinium chloride, procaine, or atropine. (A) mexC expression after 90 min of incubation with 10 μg/ml of dequalinium chloride, 2 mg/ml of procaine, or 2 mg/ml of atropine or without inducer. The nfxB* strain grown in the absence of any inducer was used as a control of overexpression. As shown, the expression of mexCD-oprJ is induced by the three tested compounds. (B, C, and D) mexC expression 30, 60, and 120 min after the removal of the inducers (–I) and compared to the expression level after 90 min of induction (+I) with 10 μg/ml of dequalinium chloride (B), 2 mg/ml of procaine (C), and 2 mg/ml of atropine (D). As shown, the overexpression of mexCD-oprJ is reduced 30 min after the removal of procaine and atropine to levels similar to those for the untreated bacterial levels; 30 and 60 min after removal of dequalinium chloride, the expression level is reduced to 10 times the untreated bacterial levels and to 2 times the untreated bacterial levels after 120 min. The fold changes were calculated regarding the expression of untreated P. aeruginosa PAO1. Each represented value is the average of three biological replicates. Statistically significant differences regarding PAO1 untreated were calculated with a Student t test for paired samples assuming equal variances (*, P < 0.05; **, P < 0.005; ***, P < 0.0005).

FIG 6

Effect of inducers of P. aeruginosa growth in the presence of ciprofloxacin. The growth curves of P. aeruginosa PAO1 (green), nfxB* (blue), and of nfxB*ΔmexD (red) in LB medium containing 0.25 μg/ml ciprofloxacin with (light) or without (dark) 10 μg/ml of dequalinium chloride (A), 2 mg/ml of procaine (B), or 2 mg/ml of atropine (C) are shown. As indicated, wild-type P. aeruginosa grows better in the presence of ciprofloxacin when the inducers are added. Each represented OD₆₀₀ value is the average of three biological replicates.

FIG 7

Checkerboard analysis for ciprofloxacin and mexCD-oprJ inducer compounds. Checkerboard analyses were performed with P. aeruginosa PAO1 wild-type (A) and nfxB* ΔmexD (B) strains. Wells with bacterial growth are represented in blue, and wells in which there was no growth are represented in white. Inducer compounds showed an antagonistic effect with ciprofloxacin in strain PAO1, but no antagonistic effect was observed for the mexD-deficient strain.

FIG 8

Effect of MexCD-OprJ on the susceptibility of P. aeruginosa to mexCD-oprJ inducers. PAO1 wild-type (blue) and nfxB* ΔmexD (red) strains were grown in LB as a control (A) and in the presence of 4 mg/ml of atropine (B), 4 mg/ml of procaine (C), or 20 μg/ml of dequalinium chloride (D). As shown, the absence of MexCD-OprJ increases P. aeruginosa susceptibility to procaine. Each represented OD₆₀₀ value is the average of three biological replicates.