Bioluminescent Pseudomonas aeruginosa and Escherichia coli for whole-cell screening of antibacterial and adjuvant compounds

Abstract

Continued efforts to discover new antibacterial molecules are critical to achieve a robust pre-clinical pipeline for new antibiotics. Screening of compound or natural product extract libraries remains a widespread approach and can benefit from the development of whole cell assays that are robust, simple and versatile, and allow for high throughput testing of antibacterial activity. In this study, we created and validated two bioluminescent reporter strains for high-throughput screening, one in Pseudomonas aeruginosa, and another in a hyperporinated and efflux-deficient Escherichia coli. We show that the bioluminescent strains have a large dynamic range that closely correlates with cell viability and is superior to conventional optical density (OD₆₀₀) measurements, can detect dose-dependent antibacterial activity and be used for different drug discovery applications. We evaluated the assays' performance characteristics (signal to background ratio, signal window, Z' robust) and demonstrated their potential utility for antibiotic drug discovery in two examples. The P. aeruginosa bioluminescent reporter was used in a pilot screen of 960 repurposed compound libraries to identify adjuvants that potentiate the fluoroquinolone antibiotic ofloxacin. The E. coli bioluminescent reporter was used to test the antibacterial activity of bioactive bacterial supernatants and assist with bioassay-guided fractionation of the crude extracts.

Fig. 1

Growth of P.a lux + produces luminescence which closely correlates with OD₆₀₀ and viable bacterial counts. (a) Luminescence and OD₆₀₀ over time in P.a lux + cultures grown in a 96 well black plate incubated in a plate reader incubation chamber (Cytation 5). Cultures were started at an OD₆₀₀ ~ 0.05, and luminescence (●) and OD₆₀₀ (o) were measured simultaneously every 30 min. (b) Luminescence and absorbance correlation. The respective RLU and OD₆₀₀ measured in each well for T = 0 to 6 h for panel (a) was plotted to estimate the correlation between the two, with each data point representing a single replicate. (c) Luminescence in P.a lux + cultures grown in a 96 well black plate incubated in a shaking incubator (Infors Multitron Pro). Cultures were started at an OD₆₀₀ ~ 0.05 and RLU was read every 2 h. (d) Luminescence and absorbance correlation. Cells, grown for 6 h in 96 well black plates, were serially diluted, following which their luminescence (●) and OD₆₀₀ (o) was plotted against viability. Data is shown as mean ± SD for (a) (n = 6) and (c) (n = 15), and as mean ± SEM for (d) (n = 6) but error bars are too small to be visualized. # Indicates a non-zero RLU value. See Figure S1 for more details.

Fig. 2

Pilot screen using P.a lux + for adjuvants that potentiate ofloxacin. (a) and (b) The screen assay performance was evaluated with LB medium (control), 1 µg/mL ofloxacin, 1.5 µM PMBN and [1 µg/mL ofloxacin + 1.5 µM PMBN]. RLU was measured at (a) 0 h and (b) 6 h. For (a) and (b), each data point represents a single well, with n ≥ 8 replicate from a representative experiment. (c) Pilot screen with 960 compounds. Compounds at 100 µM (single replicate) were tested for adjuvant activity in the presence of 0.25 µg/mL ofloxacin while controls included LB medium, 0.25 µg/mL ofloxacin, 1.5 µM PMBN, and [0.25 µg/mL ofloxacin and 1.5 µM PMBN]. A total of 91 hits were identified with > 50% RLU inhibition compared to 0.25 µg/mL ofloxacin alone. Each data point represents a single well. (d) Counterscreen of the 41 hits, with compounds tested alone at 100 µM. Data is shown as mean ± range (n = 2 independent replicate). A total of 8 compounds < 20% RLU inhibition (represented by grey bar) compared to LB + ofloxacin alone were identified.

Fig. 3

Dose–response curves of 8 hits that potentiate ofloxacin against P. aeruginosa. A 10-point IC₅₀ dose–response for (a) entrectnib (ENT), tetrabromobenzotriazole (TBB), SB-408124 (SB) and ARRY-520 (ARRY) and (b) balofloxacin (BAL), LGD-4033(LGD), LY2886721 (LY) and vonoprazan/ TAK-438 (VON) were determined in the presence of 0.25 µg/mL ofloxacin. Curves were fitted to the equation Log (inhibitor) vs. Response-Variable slope (four parameters). Data points and error bars represent the mean ± SEM (n = 3 biological replicates from a representative experiment).

Fig. 4

E.c.-lux + successfully produces luminescence, with a high dynamic range. (a) Luminescence and OD₆₀₀ over time in E.c.-lux + cultures grown in a 96 well black plate incubated in a plate reader incubation chamber (Cytation 5). Cultures were started at an OD₆₀₀ ~ 0.05, and luminescence (●) and OD₆₀₀ (o) were measured simultaneously every hour. Cells grown for 14 h in 96 well black plates, were serially diluted, after which (b) their luminescence (●) OD₆₀₀ (o) was plotted against viability and, (c) the luminescence and absorbance correlation was determined. (d) E.c.-lux + was grown over 14 h in 96 well black plates with and without 100 µg/mL ampicillin. Luminescence readings were taken every hour and plotted against time. Data for (a), (b), and (c) are presented as mean ± SEM (n = 9 biological replicates from 3 experiments) but error bars are too small to be visualized. For (d), each data point represent a replicate, with data pooled from 3 independent experiments. # Indicates a non-zero RLU value. See Figure S6 for more details.

Fig. 5

E.c.-lux + responds to antibacterial challenge in a dose dependent manner. E.c.-lux + cells were sub-cultured in LB Lennox and grown in the presence of decreasing doses of (a) tetracycline (1 µg/mL – 0.00012 µg/mL), (b) ofloxacin (0.12 µg/mL – 1.5 × 10^-5 µg/mL), (c) colistin (4 µg/mL – 0.00049 µg/mL), and (d) supernatant (50% v/v – 0.006% v/v). The RLU was measured following 14 h of incubation in a shaking incubator and IC₅₀ curves were determined by non-linear fitting of the data using the equation log inhibitor concentration versus response with a variable slope. All data is presented as mean ± SEM from n = 12 biological replicates pooled from 3 experiments.

Fig. 6

E.c.-lux + can detect differences in activity from MNAAK extract fractions. The bioactive MNAAK crude extract was fractionated using Prep-TLC. E.c.-lux + cells were tested against fractions at a concentration of 64 µg/mL and 0.3125% MeOH in LB Lennox. The RLU was measured following 14 h of incubation in a shaking incubator. An unfractionated MNAAK crude extract and untreated LB Lennox medium were used as positive and negative controls respectively. Dashed line indicates 50% inhibition relative to the untreated LB control. Data is presented as mean ± SEM from n = 9 biological replicates pooled from 3 experiments.