Sigma S-dependent antioxidant defense protects stationary-phase Escherichia coli against the bactericidal antibiotic gentamicin

Abstract

Stationary-phase bacteria are important in disease. The σ(s)-regulated general stress response helps them become resistant to disinfectants, but the role of σ(s) in bacterial antibiotic resistance has not been elucidated. Loss of σ(s) rendered stationary-phase Escherichia coli more sensitive to the bactericidal antibiotic gentamicin (Gm), and proteomic analysis suggested involvement of a weakened antioxidant defense. Use of the psfiA genetic reporter, 3'-(p-hydroxyphenyl) fluorescein (HPF) dye, and Amplex Red showed that Gm generated more reactive oxygen species (ROS) in the mutant. HPF measurements can be distorted by cell elongation, but Gm did not affect stationary-phase cell dimensions. Coadministration of the antioxidant N-acetyl cysteine (NAC) decreased drug lethality particularly in the mutant, as did Gm treatment under anaerobic conditions that prevent ROS formation. Greater oxidative stress, due to insufficient quenching of endogenous ROS and/or respiration-linked electron leakage, therefore contributed to the greater sensitivity of the mutant; infection by a uropathogenic strain in mice showed this to be the case also in vivo. Disruption of antioxidant defense by eliminating the quencher proteins, SodA/SodB and KatE/SodA, or the pentose phosphate pathway proteins, Zwf/Gnd and TalA, which provide NADPH for ROS decomposition, also generated greater oxidative stress and killing by Gm. Thus, besides its established mode of action, Gm also kills stationary-phase bacteria by generating oxidative stress, and targeting the antioxidant defense of E. coli can enhance its efficacy. Relevant aspects of the current controversy on the role of ROS in killing by bactericidal drugs of exponential-phase bacteria, which represent a different physiological state, are discussed.

FIG 1

The ΔrpoS mutant of E. coli is more sensitive to Gm than the wild type (WT). Data presented here and subsequently are for stationary-phase cells. E. coli K-12, strain BW25113 was used, except where otherwise indicated. Gm was used at a concentration of 16 μg/ml with incubation at 37°C. Solid bars show mean log CFU counts per ml after Gm treatment for 24 h; striped bars represent untreated controls. The Student t test comparison is between drug-treated and untreated cells of the same strain. **, P < 0.01.

FIG 2

Gm treatment results in greater ROS levels in the ΔrpoS mutant. Solid bars represent Gm-treated cells, and striped bars represent untreated controls. (A) Activation of SOS response with and without Gm treatment in the wild type and the ΔrpoS mutant containing a single copy of the sfiA-lacZ fusion at 24 h, as monitored by β-galactosidase activity. (B) Representative differential interference contrast micrographs of cells of the two strains with and without 24-h Gm treatment (magnification, ×1,000; number of cells examined, ca. 10,000 of each strain). (C) Effect of 24-h Gm treatment on mean relative fluorescence units (RFU) of 3′-(p-hydroxyphenyl) fluorescein (HPF) in cells of the two strains. (D) Effect of Gm treatment on H₂O₂ production by the two strains. H₂O₂ was measured in Gm-treated cells by Amplex Red at 2, 4, and 6 h; the area under the concentration-time curve for H₂O₂ at these time points is plotted and represents total cellular H₂O₂ generation during this time; a shorter treatment time was used for this measurement because intracellular H₂O₂ may be subject to decomposition (26). (E) Effect of 24-h Gm treatment on protein carbonylation in the two strains, as detected by slot blot analysis. Student's t test was used to compare Gm-treated and untreated cells of the same strain (*, P < 0.05; **, P < 0.01; ***, P < 0.001) and untreated cells of the wild type and the mutant (††, P < 0.01; †††, P < 0.001).

FIG 3

Coadministration of N-acetyl cysteine (5 mg/ml) (A) and anaerobiosis (B) decrease Gm lethality in stationary-phase cells in both the wild type and the ΔrpoS mutant but to a greater extent in the latter. Bars show survival after 24 h of Gm treatment under the specified conditions. Asterisks indicate P values in comparisons of NAC treatment or anoxic effects within the same strains. *, P < 0.05; ***, P < 0.001.

FIG 4

Deletion of ROS quencher proteins or PPP enzymes renders E. coli more sensitive to Gm. Effect of Gm treatment was determined as described in the legend of Fig. 1 on strains lacking ROS quencher proteins (ΔkatE ΔsodA and ΔsodA ΔsodB mutants) or PPP enzymes (Δzwf Δgnd and ΔtalA mutants). Solid bars represent the effect of Gm treatment; bars with stripes represent untreated controls. The effect of the drug on the wild type is reproduced from Fig. 1 for reference. Asterisks indicate P values for Gm-treated and untreated cells of the same strain. *, P < 0.05; **, P < 0.01; ***, P < 0.001.

FIG 5

Gm treatment generates greater oxidative stress in mutants lacking quencher or PPP proteins than in the wild type. (A) Activation of SOS response at 24 h, as monitored by measuring β-galactosidase activity after Gm treatment in the quencher (ΔsodA ΔsodB mutant) or PPP (Δzwf Δgnd mutant) class of mutants containing a single copy of the sfiA-lacZ fusion. (B) Effect of 24-h Gm treatment on protein carbonylation in the ΔsodA ΔsodB, Δzwf Δgnd, and ΔtalA mutants as detected by slot blot analysis. Data on the wild type are reproduced from Fig. 2 for comparison. Asterisks indicate P values for Gm-treated and untreated cells of the same strain (*, P < 0.05; **, P < 0.01; ***, P < 0.001), and daggers represent comparisons between untreated cells of the wild type and the mutant (††, P < 0.01; †††, P < 0.001).

FIG 6

Gm compromises ability of the ΔrpoS mutant of UPEC (AMG1) to colonize female mouse bladder. Bladder infections in mice were initiated using stationary-phase wild-type or ΔrpoS mutant bacteria of the AMG1 strain. The mice were treated with 0, 0.5, 5, or 50 μg of Gm (A) or after N-acetyl cysteine (NAC; 10 mg/kg) administration (B). Data are presented as box-and-whisker plots which depict maximum, 75th percentile, median, 25th percentile, and minimum values of each group. For the experiment depicted in panel B, NAC was given as a single 10 mg/kg intraperitoneal dose 1 h before infection. Gm was administered at the time of infection and every 2 h thereafter for five doses of 0.5 μg. For statistical analysis, the data were transformed into log₁₀ values to equalize group variance. Student's t test was performed using the transformed data (*, P < 0.05; **, P < 0.01, between groups as indicated).

FIG 7

Gm increases electron leakage from the respiratory chain. Relative fluorescence units (RFU) of alamarBlue are shown for the membrane fraction of cell extracts of stationary-phase E. coli (BW25113) containing rotenone without (CE, for cell extract) and with (CE+Gm) Gm. ***, P < 0.001. See the text for further details.