Sodium Malonate Inhibits the AcrAB-TolC Multidrug Efflux Pump of Escherichia coli and Increases Antibiotic Efficacy

Abstract

There is an urgent need to find novel treatments for combating multidrug-resistant bacteria. Multidrug efflux pumps that expel antibiotics out of cells are major contributors to this problem. Therefore, using efflux pump inhibitors (EPIs) is a promising strategy to increase antibiotic efficacy. However, there are no EPIs currently approved for clinical use especially because of their toxicity. This study investigates sodium malonate, a natural, non-hazardous, small molecule, for its use as a novel EPI of AcrAB-TolC, the main multidrug efflux pump of the Enterobacteriaceae family. Using ethidium bromide accumulation experiments, we found that 25 mM sodium malonate inhibited efflux by the AcrAB-TolC and other MDR pumps of Escherichia coli to a similar degree than 50 μΜ phenylalanine-arginine-β-naphthylamide, a well-known EPI. Using minimum inhibitory concentration assays and molecular docking to study AcrB-ligand interactions, we found that sodium malonate increased the efficacy of ethidium bromide and the antibiotics minocycline, chloramphenicol, and ciprofloxacin, possibly via binding to multiple AcrB locations, including the AcrB proximal binding pocket. In conclusion, sodium malonate is a newly discovered EPI that increases antibiotic efficacy. Our findings support the development of malonic acid/sodium malonate and its derivatives as promising EPIs for augmenting antibiotic efficacy when treating multidrug-resistant bacterial infections.

Keywords: AcrAB-TolC; EPI; Escherichia coli; efflux pump inhibitor; malonate; malonic acid; multidrug efflux pump; sodium malonate.

Figure 1

PAβN and sodium malonate, but not sodium acetate, are EPIs that prevent ethidium bromide efflux in E. coli. (A) Ethidium bromide accumulation was measured as the increase in relative fluorescence units (RFUs) for 30 cycles in the parental and ΔacrB strains. Results are presented as the average ± SEM (n = 6; i.e., 6 biological replicates, each including 3 technical replicates). RFU measurements remain stable between cycles 15 and 30. Ethidium bromide is a substrate of the AcrAB-TolC MDR pump and therefore it accumulated intracellularly significantly more in the ΔacrB strain than in the parental strain at all time points measured (p < 0.05; denoted with * next to the “ΔacrB untreated” legend at the top of the figure). (B–D) Ethidium bromide accumulation was measured as the increase in relative fluorescence units (RFUs) observed at cycle 30 for each strain in the presence of increasing concentrations of PAβN (B), sodium malonate (C), and sodium acetate (D). Results are presented as the average ± SEM (n = 3; i.e., 3 biological replicates, each including 3 technical replicates) and are shown as the n-fold change normalized to the untreated parental. Statistically significant differences (p < 0.05) between the untreated and treated are shown as * or *, for the parental and ΔacrB strains, respectively, for panels (B,C) (no differences for treatments were found in either strain for sodium acetate in panel (D)).

Figure 2

Effect of the EPIs PAβN and sodium malonate individually and in combination on ethidium bromide efflux. Ethidium bromide accumulation was measured in E. coli parental and ΔacrB strains as the increase in relative fluorescence units (RFUs) observed at cycle 30 for each strain and treatment. Results are presented as the average ± SEM (n = 3; i.e., 3 biological replicates, each including 3 technical replicates) and are shown as the n-fold change normalized to the untreated parental. Statistically significant differences (p < 0.05) between the untreated and treated are shown as * or *, for the parental and ΔacrB strains, respectively. Sodium malonate did not significantly change the effect of PAβN in either strain (noted as n.s.).

Figure 3

Effect of the EPIs sodium malonate and PAβN in combination on the antimicrobial activity of EtBr measured by 3D-checkerboard assays. Sodium malonate increased the efficacy of EtBr at concentrations of 25 to 100 mM (n = 3 biological replicates, each with one technical replicate), and all changes were statistically significant (p < 0.05). PAβN at concentrations of 6.25 to 50 μM decreased by 2-fold the MIC of EtBr at concentrations of sodium malonate of 25–50 mM; PAβN at 100 μM only decreased the efficacy of the EtBr-100 mM sodium malonate combination. Data are presented as the average of 3 biological replicates, each including 1 technical replicate. All differences were statistically significant (p < 0.05).

Figure 4

Molecular docking characterization of ethidium bromide, minocycline, PAβN, and sodium malonate binding to AcrB. (A) Overall structure of the AcrB homotrimer. (B) AcrB monomer, regions containing the proximal and distal binding pockets according to ref. [41] are indicated as blue and red dotted ovals, respectively. (C) Zoom view of the proximal (blue dotted oval) and distal binding pockets (red dotted oval) of AcrB showing molecular docking results for the four ligands tested. Ligands are colored by atom, and carbon atoms are colored in cyan (ethidium bromide), purple (minocycline), PAβN (yellow), and green (sodium malonate). The scores for the docking result shown for each ligand were: EtBr: −9.6; minocycline: −8.2; PAβN: −10.2 (only one of two similar poses with the same score is shown); and sodium malonate: −4.3. The poses shown represent the highest score poses found to bind to the distal or proximal binding pockets. Full docking results for all ligands are provided in Figure 5.

Figure 5

Molecular docking characterization of ethidium bromide, minocycline, PAβN, and sodium malonate binding to AcrB, showing all top 8 results obtained from AutoDock Vina. (A) Ethidium bromide (cyan): all 8 poses bound very closely in the distal binding pocket with scores ranging from -9.6 (top score; also shown in Figure 4C) to −8.4. (B) PAβN (yellow): 7 poses bound very closely in the distal binding pocket with scores ranging from -10.2 (top score; also shown in Figure 4C) to −9.4. Only one pose (no. 8; score = −9.2) was found outside the distal binding pocket. (C) Minocycline (purple): poses were found both in the distal binding pocket (pose no. 2, score −8.2, also shown in Figure 4C; pose no. 5, score = −8.1; and pose no. 8, score = −7.8) and in other areas of the AcrB porter domain (pose no. 1, score = −8.4; pose no. 3, score = −8.2; pose no. 4, score = −8.1; pose no. 6, score = −8.0; and pose no. 7, score = −7.8. (D,E) Sodium malonate (green) poses are shown in two different AcrB orientations. No poses were found in the distal binding pocket, and only one pose (no. 4, score = −4.3; also shown in Figure 4C) was found in the proximal binding pocket. The other poses were found in different areas of the porter domain (poses no. 1, no. 2, no. 3, no. 5, no. 6, no. 7, and no. 8, which had scores of −4.5, −4.5, −4.4, −4.2, −4.2, −4.4 and −4.2, respectively). (F) Zoom view of sodium malonate (green) bound to the proximal binding pocket (pose no. 4), which was stabilized with three hydrogen bonds (in cyan) with threonine-87 (shown in purple above sodium malonate) and two hydrogen bonds with arginine-815 (shown in magenta).