Efflux pumps mediate changes to fundamental bacterial physiology via membrane potential

Abstract

Efflux pumps are well known to be an important mechanism for removing noxious substances such as antibiotics from bacteria. Given that many antibiotics function by accumulating inside bacteria, efflux pumps contribute to resistance. Efflux pump inactivation is a potential strategy to combat antimicrobial resistance, as bacteria would not be able to pump out antibiotics. We recently discovered that the impact of loss of efflux function is only apparent in actively growing cells. We demonstrated that the global transcriptome of Salmonella Typhimurium is drastically altered during slower growth leading to stationary-phase cells having a remodeled, less permeable envelope that prevents antibiotics entering the cell. Here, we investigated the effects of deleting the major efflux pump of Salmonella Typhimurium, AcrB, on global gene transcription across growth. We revealed that an acrB knockout entered stationary phase later than the wild-type strain SL1344 and displayed increased and prolonged expression of genes responsible for anaerobic energy metabolism. We devised a model linking efflux and membrane potential, whereby deactivation of AcrB prevents influx of protons across the inner membrane and thereby hyperpolarization. Knockout or deactivation of AcrB was demonstrated to increase membrane potential. We propose that the global transcription regulator ArcBA senses changes to the redox state of the quinol pool (linked to the membrane potential of the bacterium) and coordinates the shift from exponential to stationary phase via the key master regulators RpoS, Rsd, and Rmf. Inactivation of efflux pumps therefore influences the fundamental physiology of Salmonella, with likely impacts on multiple phenotypes.IMPORTANCEWe demonstrate for the first time that deactivation of efflux pumps brings about changes to gross bacterial physiology and metabolism. Rather than simply being a response to noxious substances, efflux pumps appear to play a key role in maintenance of membrane potential and thereby energy metabolism. This discovery suggests that efflux pump inhibition or inactivation might have unforeseen positive consequences on antibiotic activity. Given that stationary-phase bacteria are more resistant to antibiotic uptake, late entry into stationary phase would prolong antibiotic accumulation by bacteria. Furthermore, membrane hyperpolarization could result in increased generation of reactive species proposed to be important for the activity of some antibiotics. Finally, changes in gross physiology could also explain the decreased virulence of efflux mutants.

Keywords: AcrAB-TolC; ArcBA; RND efflux pump; energy metabolism; hyperpolarization; redox state.

Fig 1

Venn diagrams comparing gene expression in the (a) wild-type and (b) ΔacrB strains. The number of genes differentially expressed (log₂ fold change ≥1.5 and adjusted P value <0.05) for each comparison is shown.

Fig 2

Data analysis strategy to characterize differentially expressed genes. For each gene, transcript levels at 1 and 3 h were compared, and a log₂ fold change (log₂FC) value was calculated. The log₂ fold change for the ΔacrB mutant was subtracted from the WT log₂FC to give a Δlog₂FC value indicating the difference in regulation between the two strains. A positive Δlog₂ value indicates a gene is more upregulated or less downregulated in ΔacrB for the 1 h vs 3 h timepoints. A negative Δlog₂ value indicates that a gene is more upregulated or less downregulated in the wild type.

Fig 3

Model of the mechanism of ArcBA-mediated redox control of stationary-phase entry and the impact of an AcrB mutant. In the wild-type strain during rapid growth (a), high flux through the electron transport chain results in a reduced quinol pool (QH₂), which permits ArcBA to be active, repressing expression of rpoS, rsd, and rmf. RssB also targets RpoS protein for degradation. As growth slows (b), the quinol pool becomes more oxidized (Q), deactivating ArcB, permitting activation of rpoS, rsd, and rmf, triggering entry into stationary phase. In a ΔacrB mutant (c), proton motive force is higher for longer during growth, repressing rpoS, rsd, and rmf and delaying entry into stationary phase. As a result, exponential-phase physiology is extended (d), necessitating upregulation of genes involved in anaerobic energy metabolism (Nap, Nrf, and CCP shown), likely via the global regulator FNR. TO, terminal oxidase.

Fig 4

Gene expression data for the key stationary-phase regulators rsd, rmf, and rpoS. Log₂ fold-change values for each gene: 1 h vs 3 h in the wild type, 3 h vs 5 h in the wild type, 1 h vs 3 h in ΔacrB, and 3 h vs 5 h in ΔacrB. ΔLog₂ values for the 1 h vs 3 h and 3 h vs 5 h comparisons. ΔLog₂ values are calculated by log₂ fold change for ΔacrB minus log₂ fold change for the wild type. ID: locus name. Fold changes that are insignificant (P_adj >0.05) are marked with a dagger (†). ΔLog₂ fold change values derived from one or more non-significant log₂ fold-change values are also in white. Significant fold changes are color coded as shown in the key.

Fig 5

Determination of membrane potential. S. Typhimurium cells were grown in MOPS minimal medium, and samples were taken at 1-, 3-, and 5-h growth, whereupon membrane potential was determined using flow cytometry and DiSC₃(5). The red fluorescence, corresponding to membrane potential, is shown for each strain. The protonophore CCCP was used as a control to collapse membrane potential. (a) Strains SL1344 (wild type), ΔacrB, and acrB D408A. (b) Strain SL1344 in the absence and presence of pPAR1-SacrAB, expressing acrAB. Error bars show the standard deviation of replicate cultures for a representative experiment. Significance was tested for comparisons of (a) wild type and ΔacrB/acrB D408A strains and (b) the absence and presence of pPAR1-SacrAB for each timepoint using a paired t-test. *P ≤ 0.05. Addition of CCCP to each sample resulted in a significant (P ≤ 0.05) decrease in membrane potential.