Fig 1.
Screening platform (SAFIRE) used to identify EPMs.
(A) Schematic of screening methodology. (B) Representative micrographs of infected macrophages from DMSO-treated wells. Upper left is a field with 522 macrophages; remaining images are the indicated channels zoomed in on the boxed region. Scale bars are 50 μm. (C) Representative micrographs of infected macrophages treated with rifampicin [2 μg/mL] or of uninfected macrophages treated with DMSO. (D) Distribution of B-scores and p-values for 14,400 compounds from the Maybridge HitFinderTM v11 library. The locations of the three hit compounds (EPM30, EPM35 and EPM43) and of chloramphenicol (Cm), which was identified from the library, are shown.
Table 1.
Bacterial strains.
Fig 2.
Three compounds may modulate bacterial efflux pump activity.
(A, B) Salmonella was incubated with Hoechst 33342 and the indicated compound. (A) Fluorescence was normalized to the maximum of heat-killed bacteria (100%). Mean and SEM from three biological replicates. (B) Maximum fluorescence over 60 minutes of exposure normalized to the maximum fluorescence of heat-killed bacteria (100%). The EC50s in the key were established using a nonlinear four-parameter fit. Mean and SEM from three biological replicates, each performed in duplicate. (C, D) Nitrocefin [100 μM] hydrolysis in a membrane-compromised E. coli strain treated with the indicated concentrations of drugs. Hydrolysis increases when efflux is blocked. (C) Absorbance 486 nm of RAM121 E. coli. Mean and SD are representative of three independent biological replicates. (D) Slope of the linear region of the A486 plot from at least three experiments. Data is normalized to A486/minute. ** p < 0.001, **** p < 0.0001 by one-way ANOVA with Dunnett’s post-test. (E) Structures of the three EPMs.
Fig 3.
The three hit compounds decrease bacterial load of Salmonella in mammalian cells.
(A-C) Monitoring of bacterial load by GFP (SAFIRE) or CFU in RAW264.7 cells. (A) Representative micrographs of cells in 96-well plates infected with sifB::GFP Salmonella. Two hours after infection cells were treated with the indicated compound [25 μM] for 16 hours. (B) Dose response curve for SAFIRE and (C) CFU; keys includes IC50 values. (D-E) Monitoring of bacterial load by GFP (SAFIRE) in HeLa cells infected with Salmonella expressing rpsM::GFP and treated with the indicated compound [25 μM]. (F) Monitoring of bacterial load by CFU in BMDMs treated with compounds [25 μM]. Mean and SEM from three independent biological replicates. The nonlinear curve fitting (B, C) is constrained using uninfected cells as the minimum and DMSO-treated cells as the maximum. (E, F) * p < 0.05; ** p < 0.01, *** p < 0.001 compared to DMSO by one-way ANOVA with Dunnett’s post-test.
Fig 4.
The hit compounds reduce the survival of MDR Salmonella in macrophages.
Monitoring of bacterial load by CFU in RAW264.7 cells infected for two hours with the strain of Salmonella shown and then treated with the indicated compound for 16 hours, followed by macrophage lysis and plating for CFU. A) SL1344 Wild Type Salmonella, B) Clinical Salmonella isolate S10801, (C) SL1344 with macAB::kan, acrAB::kan or tolC::cm, respectively. Geometric mean of four biological replicates. Upper lines, mean CFU/well of wild-type SL1344 with DMSO treatment; lower lines, limit of detection. * p < 0.05, ** p < 0.01; *** p < 0.001, **** p < 0.0001 relative to DMSO, one-way ANOVA with Dunnett’s post-test.
Fig 5.
EPMs block efflux of ethidium bromide and Nile red.
Salmonella were incubated with either ethidium bromide or Nile red without glucose, treated with compound, exposed to glucose, and then monitored for fluorescence. Data for each sample were normalized to the initial fluorescence (100%). Dose response curves are (A) ethidium bromide at 28 minutes after glucose addition, (B) Nile red at 7 minutes after glucose addition, and (C) Nile red after washout of compound 7 minutes after glucose addition. Mean of at least two biological replicates performed in duplicate. * p < 0.05; ** p < 0.01; *** p < 0.001; ****p < 0.0001 compared to DMSO + glucose by one-way ANOVA and Dunnett’s multiple comparison post-test.
Fig 6.
EPMs do not disrupt bacterial inner or outer membranes.
(A,B) Salmonella treated with DMSO or EPMs [100 μM] but not CCCP [1 mM] acquire TMRM staining within 30 minutes. (A) Representative data from one of three independent experiments. (B) Median fluorescence intensity from three experiments normalized to unstained control (0). (C) Disk diffusion assays; the radius of the zone of growth inhibition after 16 hours of exposure to compound across a dose range. Black lines, semilog fit for the combined antibiotic data; gray lines, semilog fit for CCCP and PAβN; dotted lines, limit of detection (disk radius). Average of two measurements from each image captured from one experiment representative of two independent experiments. (D,E) Nitrocefin access to the periplasm as monitored by nitrocefin [100 μM] hydrolysis in the presence of the indicated concentrations of compounds. (D) Absorbance 486 nm of bla+ Salmonella normalized to bla- Salmonella. Data is representative of at least two independent biological replicates. (E) Slope of the linear region of the A486 plot from at least three experiments. Data is normalized to A486/minute. * p < 0.05, *** p < 0.001, **** p < 0.0001 by one-way ANOVA with Dunnett’s post-test.
Fig 7.
EPM35 and EPM43 interact with the efflux pump AcrB.
(A, B) Representative ITC for the binding of (A) EPM35 and (B) EPM43 to E. coli AcrB. Each peak in the upper graphs corresponds to the injection of 2 μL of 100 μM of the EPM in buffer containing 20 mM Na-HEPES (pH 7.5), 0.05% DDM and 5% DMSO into the reaction containing 10 μM of E. coli monomeric AcrB in buffer containing 20 mM Na-HEPES (pH 7.5), 0.05% DDM and 5% DMSO. The lower graphs show the cumulative heat of reaction displayed as a function of injection number. The solid line is the least-square fit to the experimental data.
Table 2.
Binding of inhibitors by AcrB.
Fig 8.
EPMs synergize with antimicrobial peptides.
Salmonella was grown in M9-based defined media in the presence of (A-D) polymyxin B (5 μg/ml; 1/8 MIC) or (E-H) LL37 (5 μg/ml; 1/8 MIC) and EPMs (PAβN, 500 μM; EPMs, 25 μM). Mean and SD of triplicate samples from one representative experiment of three independent biological replicates. DMSO, polymyxin B and LL37 curves repeat across graphs.
Fig 9.
EPMs sensitize bacteria to outer membrane permeabilization by antimicrobial peptides.
Nitrocefin access to the periplasm in the presence of a non-permeabilizing concentration of polymyxin B [1 μg/ml] and the indicated concentrations of EPMs, quantitated as in Fig 6E. * p < 0.05, ** p < 0.01 by one-way ANOVA with Dunnett’s post-test.
Table 3.
MICs (μg/mL) of AcrB substrates in combination with EPMs.
Fig 10.
EPMs enhance activity of erythromycin and ciprofloxacin against Salmonella in macrophages.
RAW 264.7 macrophages were infected with Salmonella and treated with a dose range of (A) erythromycin or (B) ciprofloxacin and with DMSO or the indicated concentration of an EPM. At 18 hours post infection samples were processed for fluorescence microscopy as described. Data were normalized to treatment with DMSO and antibiotic vehicle (100%). Key: black, DMSO; red, EPM30; green, EPM35; blue, EPM43; gray, calculated additivity of the antibiotic and the corresponding EPM using the formula (100 –([percent inhibition EPM] + [percent inhibition antibiotic])), where percent inhibition is calculated as 100 –[percent of DMSO]. Data are mean + SEM of three biological replicates. * p < 0.05; ** p < 0.01; *** p < 0.001; **** p < 0.0001 of EPM treatment versus calculated additivity by one-way ANOVA with Sidak’s multiple comparison test.
Fig 11.
EPM35 and EPM43 block efflux of Nile red from ESKAPE MDR clinical isolates.
(A-D) Defined strains obtained from BEI resources were examined for Nile red retention after glucose addition in the presence of the indicated compound. Data for each sample were normalized to the initial fluorescence (100%). Dose response curves at seven minutes after glucose addition. Mean of three biological replicates performed in duplicate. * p < 0.05; ** p < 0.01; *** p < 0.001; ****p < 0.0001 compared to DMSO + glucose by one-way ANOVA and Dunnett’s multiple comparison post-test.