Inhibition of stationary phase respiration impairs persister formation in E. coli

Abstract

Bacterial persisters are rare phenotypic variants that temporarily tolerate high antibiotic concentrations. Persisters have been hypothesized to underlie the recalcitrance of biofilm infections, and strategies to eliminate these cells have the potential to improve treatment outcomes for many hospital-treated infections. Here we investigate the role of stationary phase metabolism in generation of type I persisters in Escherichia coli, which are those that are formed by passage through stationary phase. We find that persisters are unlikely to derive from bacteria with low redox activity, and that inhibition of respiration during stationary phase reduces persister levels by up to ∼1,000-fold. Loss of stationary phase respiratory activity prevents digestion of endogenous proteins and RNA, which yields bacteria that are more capable of translation, replication and concomitantly cell death when exposed to antibiotics. These findings establish bacterial respiration as a prime target for reducing the number of persisters formed in nutrient-depleted, non-growing populations.

Figure 1. Stationary phase metabolic activity and persistence.

(a) Overnight cultures (24 h) were stained with RSG for 1 h, and then segregated with FACS into four quantiles, A, B, C and D, comprising 10, 40, 40 and 10% of total population. (b) Segregated cells diluted in fresh LB were treated with 200 μg ml⁻¹ ampicillin or 5 μg ml⁻¹ ofloxacin for 5 h to enumerate the persister levels. These treatment conditions produced biphasic killing as depicted in Supplementary Figure 2A. (c) Shortly after sorting, segregated samples with known number of cells were plated on LB agar to count the c.f.u. Culturable cell fraction is the ratio of the number of c.f.u. to total number of cells as determined by FACS. (d) Overnight cultures with MO001 cells where mCherry protein was expressed with 1 mM IPTG were stained with RSG, and then A and D subpopulations were segregated with FACS. Unstained control of mCherry-positive cells is provided in Supplementary Figure 7A. (e,f) Segregated subpopulations (A:grey, D:black) were diluted in fresh LB without IPTG and then cultured for 2.5 h at 37 °C with shaking. At t=0 h, and 2.5 h, samples were analysed with flow cytometry to quantify the mCherry protein at single cell level. (g,h) The number of total cells and non-growing cells in these cultures at t=2.5 h were enumerated with counting beads. Non-growing cells retained their mCherry levels, whereas growing cells had reduced mCherry levels due to cell division. (i,k) Overnight cultures with E. coli cells carrying pQE-80LmCherry (without IPTG) were stained with RSG, and then A and D subpopulations were segregated with FACS. Cells were diluted in fresh LB with 1 mM IPTG, and mCherry was measured with flow cytometry at t=0 (black), and 10 min (grey). The fold changes in fluorescence for both A and D subpopulations after 10 min of culturing are shown in k. ‘*' signifies significant differences for comparisons to subpopulation A (P-value<0.05, t-test). At least three biological replicates were performed for each experimental condition. Each data point was denoted by mean value±s.e.

Figure 2. Impact of inhibition of stationary phase respiration on type I persister levels.

Cultures at t=6 h or 22 h were treated with 1 mM KCN or transferred to an anaerobic chamber. At t=24 h, cultures were washed to remove the chemical inhibitors and diluted (100-fold) in fresh LB and treated with ampicillin or ofloxacin. c.f.u. ml⁻¹ levels were monitored for 5 h during the treatments. Note that for controls, the cells were treated with solvent, H₂O. ‘*' signifies significant differences for comparisons with the control group, which was untreated, aerobic cultures here (P-value<0.05, t-test). At least three biological replicates were performed for each experimental condition. Each data point was denoted by mean value±s.e.

Figure 3. Dissolved oxygen, non-growing cell abundance and protein expression levels following inhibition of stationary phase respiration.

(a) Percentages of dissolved oxygen concentrations in cell cultures with respect to saturated media were determined after treatment with KCN. (b,c) Overnight cultures of MO001 cells where mCherry was expressed with 1 mM IPTG were similarly treated with KCN or transferred to an anaerobic chamber at t=6 h. At t=24 h, cells were washed and diluted in fresh LB, and the non-growing cells were enumerated at t=2.5 h with flow cytometry. (d,e) Overnight cultures with E. coli cells carrying pQE-80Lgfp (without IPTG) were treated with KCN or transferred to an anaerobic chamber at t=6 h. At t=24 h, cells were washed and diluted in fresh LB with the inducer, and GFP expression was monitored within 10 min with flow cytometry. Note that for controls, the cells were treated with solvent, H₂O. ‘*' signifies significant differences for comparisons with the untreated group (P-value<0.05, t-test). At least three biological replicates were performed for each experimental condition. Each data point was denoted by mean value±s.e.

Figure 4. Persister levels in strains overexpressing catalases and superoxide dismutases and cultures respiring anaerobically.

(a) sodA, sodB, katE and katG were overexpressed at t=6 h in overnight cultures using pQE-80L plasmid, and at t=24 h, cells were washed to remove the inducer and diluted in fresh media for persister assay. (b) Wild-type (WT) cell cultures at t=6 h were treated with 40 mM NaNO₃ and/or 1 mM KCN and/or transferred to an anaerobic chamber. At t=24 h, cultures were washed to remove the chemicals and diluted (100-fold) in fresh LB and treated with ampicillin or ofloxacin aerobically. c.f.u. levels were monitored for 5 h during the treatments. ‘*' signifies significant differences for comparisons to the control groups, which are empty vector or aerobic culturing with NO₃⁻ (P-value<0.05, t-test). At least three biological replicates were performed for each experimental condition. Each data point was denoted by mean value±s.e.

Figure 5. RNA integrity, protein levels and degradation, and cells size of stationary phase cells.

(a–c,e,f) Cell cultures at early stationary phase (t=6 h) were treated with 1 mM KCN or transferred to an anaerobic chamber. At t=24 h, cells were pelleted for RNA, protein, and microscope analyses. For controls, untreated overnight cultures (t=24 h) and early stationary phase cultures (t=6 h) were used. (a,b) RNA quality was determined with a bioanalyzer using an RNA 6000 Nano kit. The degradation of rRNA was assessed with RNA integrity values which range from 10 (intact) to 1 (totally degraded). (c) Cells were sonicated and the protein content in the supernatant was determined with Bradford assays. (d) Before KCN treatment at t=6 h, the inducer for gfp expression was removed in the cultures with the cells carrying pQE-80Lgfp–ssrA. After the KCN treatment, GFP levels were measured. Background fluorescence was determined using cells with empty vectors. (e,f) Phase-contrast images of fixed cells were taken using a microscope, and cell size (fold change relative to 24 h untreated overnight cultures) were determined with ImageJ. (g,h) KCN treatment was performed at t=9 h. At t=24 h, microscope images were taken, and ampicillin and ofloxacin persister levels were determined. ‘*' signifies significant differences for comparisons to control groups (P-value<0.05, t-test). At least three biological replicates were performed for each experimental condition. Each data point was denoted by mean value±s.e.