Crp/cAMP Regulation of Persister Cell Formation in the Stationary Phase.

E. coli K-12 MG1655 WT and mutant cells at early (t=5 h) and late (t=24 h) stationary phases were transferred to fresh medium with antibiotics for persister cell quantification. At time points 0, 2, 6, 8, and 20 h, 1 mL of the treated culture was washed with 1X phosphate-buffered saline (PBS) to remove antibiotics. It was then serially diluted and plated on an agar plate to count the colony-forming units (CFUs). (A) Persister levels of ampicillin-treated culture with an antibiotic concentration of 200 μg/mL. (B) Persister levels of ofloxacin-treated culture with an antibiotic concentration of 5 μg/mL. (C) Persister levels of gentamicin-treated culture with an antibiotic concentration of 50 μg/mL. The number of biological replicates is n=4 for all panels. Biphasic kill curves were generated using a non-linear model (see Materials and Methods). Statistical significance tests were conducted using F-statistics (*P < 0.05, **P < 0.01, ****P < 0.0001). The data for each time point represent the mean value ± standard deviation.

The effect of Crp/cAMP on Persister Cell Metabolism during Stationary Phase.

(A) MS analysis of E. coli K-12 MG1655 WT, Δcrp, and ΔcyaA at early (t=5 h) and late (t=24 h) stationary phases. Unsupervised hierarchical clustering was applied to standardized metabolic data. Each column represents a biological replicate. n=4. (B) Pathway enrichment analysis was conducted using MetaboAnalyst (Lu et al., 2023). Upregulated and downregulated pathways of the Δcrp strain compared to WT in the late stationary growth phase were provided in this figure. (C, D) Pathway enrichment maps comparing metabolites of the TCA cycle, pentose phosphate metabolism, glycolysis, gluconeogenesis, and pyruvate metabolism in Δcrp versus WT for early and late stationary phase conditions, respectively. Circle size corresponds to the ratio of normalized metabolite intensities between mutant and control cells. Blue (P ≤ 0.05 for dark blue; 0.05 < P < 0.10 for light blue) and red (P ≤ 0.05 for dark red; 0.05 < P < 0.10 for light red) indicate significantly downregulated or upregulated metabolites in the mutant compared to the control. White signifies no significant difference. n=4 for all panels. ESP: Early stationary phase, LSP: Late stationary phase.

Validation of Crp/cAMP-Mediated Metabolic State in Persister Cells Through Proteomics Analysis.

Pathway enrichment analysis was conducted in STRING (Szklarczyk et al., 2021, 2023) for upregulated (panel A) and downregulated (panel B) proteins. Genes highlighted in red are linked with the upregulated protein networks, while genes in blue, gray, and purple correspond to those in the downregulated protein network. The visual network in STRING illustrates protein interactions. In evidence mode, color in the network represents the interaction evidence of data support, derived from curated databases, experimental data, gene neighborhood, gene fusions, co-occurrence, co-expression, protein homology, and text mining (Szklarczyk et al., 2021, 2023).

The Role of Crp/cAMP in Non-Growing Cell Formation.

(A, B) Flow cytometry histograms depict mCherry expression in E. coli K-12 MG1655 WT, Δcrp, and ΔcyaA at early (t=5 h) and late (t=24 h) stationary phases, respectively. Cells containing an IPTG-inducible mCherry expression system were cultivated with IPTG. After washing and dilution of early and late stationary phase cells in IPTG-free fresh media, fluorescence was tracked in non-growing and growing cells for 2.5 hours. The panel is a representative biological replicate. Consistent results were seen across all 3 biological replicates. (C) Growth curves of WT, Δcrp, and ΔcyaA cultures were determined using flow cytometry to calculate lag and doubling times. Lag times were calculated using the "Microbial lag phase duration calculator" (Opalek et al., 2022). Doubling times were computed using the formula td=Δt/(3.3xLog10(N/No)). n=3. *Statistical significance observed between control and mutant strains (P < 0.05, 2-tailed t-test). The data for each time point represent the mean value ± standard deviation.

Crp/cAMP-Mediated Metabolic State of Persister Cells.

(A) GFP reporter plasmid introduced into E. coli K-12 MG1655 WT, Δcrp, and ΔcyaA cells to monitor SQR gene activity. Flow cytometry was used to detect activity at early (t=5 h) and late (t=24 h) stationary phases. The panel on the left represents a biological replicate, and the results are consistent across all 3 replicates, as demonstrated in the panel on the right. Statistical significance observed between control and mutant groups (*P < 0.05, **P < 0.01, ***P < 0.001, 2-tailed t-test). (B) Redox activities of E. coli K-12 MG1655 WT, Δcrp, and ΔcyaA cells were measured at early (t=5 h) and late (t=24 h) stationary phases by flow cytometry using a RSG dye. This dye fluoresces green after reduction by bacterial reductases. A representative biological replicate is shown (left), with consistent results across all 5 replicates (right). Statistical significance observed between control and mutant groups (*P < 0.05, **P < 0.01, 2-tailed t-test). (C) E. coli cells with integrated mCherry expression system used to validate cellular respiration. Cells were diluted into fresh media and treated with ampicillin (200 μg/mL) for 20 hours. Flow cytometry measured the red fluorescence of intact surviving cells. A representative biological replicate is shown, with consistent results across all 3 replicates. (D) RSG levels of cells (carrying the mCherry expression system) at exponential phase (t=3 h); cells before ampicillin treatment; non-lysed (intact) cells after 20 hours of ampicillin treatment; and untreated cells after 20 hours of culturing. A representative biological replicate is shown (left), with consistent results across all 4 replicates (right). Statistical significance observed between intact antibiotic-treated cells and others (*P < 0.05, **P < 0.01, 2-tailed t-test). (E) High-throughput screening of mutants from the Keio collection. The mutant strains selected are associated with central metabolism. Stationary phase cells were diluted 100-fold in fresh medium and treated with ampicillin (200 μg/mL) or ofloxacin (5 μg/mL) for 20 hours. Treated cultures were washed, serially diluted, and plated on agar plates to quantify CFUs. (F) Genes related to the TCA cycle, ETC, ATP synthase, glycolysis, and pentose phosphate pathway (PPP) were knocked out and then treated with ampicillin (200 μg/mL) or ofloxacin (5 μg/mL) to enumerate CFUs. n=4. Biphasic kill curves were generated using a non-linear model. Statistical significance tests were conducted using F-statistics (*P < 0.05, and **P < 0.01). Each data point represents the mean value ± standard deviation.

cAMP concentrations normalized to cell numbers for E. coli K-12 MG1655 WT, Δcrp, and ΔcyaA.

The cAMP levels were measured in late stationary phase cultures at 450 nm using the Cyclic AMP XP® Assay Kit (Cell Signaling Technology). n=4. Statistical significance was observed between control and mutant strains (*P < 0.05, ***P < 0.001, One-way ANOVA with Dunnett’s multiple comparisons test). The data for each time point represent the mean value ± standard deviation.

Genetic perturbation of Crp/cAMP enhanced PcyaA promoter activity, resulting in increased gfp expression.

Overnight cultures of E. coli K-12 MG1655 WT, Δcrp, and ΔcyaA strains harboring the pMSs201 plasmid, which encodes green fluorescent protein (GFP) under the control of the PcyaA promoter, were diluted 1:1000 into fresh LB medium and incubated at 37°C with shaking at 250 rpm for 24 hours. Cells at the late stationary phase were then collected, diluted in 1X PBS, and analyzed by flow cytometry. n=4. Statistical significance was observed between control and mutant strains (****P < 0.0001, One-way ANOVA with Dunnett’s multiple comparisons test). The data for each time point represent the mean value ± standard deviation.

Growth curves of E. coli K-12 MG1655 WT, Δcrp, and ΔcyaA.

Optical densities of cell cultures at 600 nm (OD600) were measured every hour using a plate reader. n=3. The data for each time point represent the mean value ± standard deviation.

Normalized cAMP concentrations of E. coli K-12 MG1655 WT.

The cAMP concentrations were measured in growth cultures at the indicated time points. First, the cAMP concentrations were normalized to the number of cells. Subsequently, the data were further normalized based on the time point 0 to mitigate errors associated with batch-to-batch assay kit variations and to capture the trend in cAMP levels across the time points. n=8. Statistical significance was observed between time points (**P < 0.01, One-way ANOVA with Dunnett’s multiple comparisons test). The data for each time point represent the mean value ± standard error.

Persister levels of cells carrying the Crp expression system.

Cells at early (t=5 h) and late (t=24 h) stationary phases were transferred to fresh media with antibiotics for persister cell quantification. At time points 0, 2, 6, 8, and 20 h, 1 mL of the treated culture underwent two washes with 1X PBS to remove antibiotics. It was then serially diluted and plated on an agar plate to count the CFUs. (A) Persister levels of ampicillin-treated culture (200 μg/mL). (B) Persister levels of ofloxacin-treated cultures (5 μg/mL). (C) Persister levels of gentamicin-treated culture (50 μg/mL). n=4. Biphasic kill curves were generated using a non-linear model (see Materials and Methods). Statistical significance tests were conducted using F-statistics (**P < 0.01 and ****P < 0.0001). The data for each time point represent the mean value ± standard deviation.

Persister levels of E. coli K-12 MG1655 WT, Δcrp, and ΔcyaA strains at normalized antibiotic concentrations.

Panels show survival following treatment with (A) ampicillin, (B) ofloxacin, and (C) gentamicin. After treatment, samples were washed six times with 1X PBS to minimize antibiotic carryover. Antibiotic concentrations were normalized to MICs to ensure valid comparisons across strains (The concentrations of 33× MIC for ampicillin and 100× MIC for ofloxacin and gentamicin match those used in Figure 1). n=4. Statistical significance was observed between control and mutant strains (****P < 0.0001, One-way ANOVA with Dunnett’s multiple comparisons test). The data for each time point represent the mean value ± standard deviation.

Agar plates showing E. coli K-12 MG1655 WT, Δcrp, and ΔcyaA strains following treatment with antibiotics at normalized concentrations.

After treatment, cells were washed and subjected to 10-fold serial dilutions, then plated on agar. Plates were incubated for (A) 16 h, (B) 48 h, and (C) 72 h to assess colony formation over time. The panel is a representative biological replicate. Consistent results were seen across all 3 biological replicates.

Deletion of crp and cyaA reduces ampicillin and ofloxacin persistence in the hipA7 strain.

Persistence levels were assessed by exposing E. coli K-12 MG1655 WT, hipA7, hipA7Δcrp, and hipA7ΔcyaA strains in the late stationary phase to the specified antibiotics, followed by CFU quantification at designated time points. (A) Ampicillin (200 µg/ml), (B) Ofloxacin (5 µg/ml) and (C) Gentamicin (50 µg/ml). n=4. Statistical significance was observed between control and mutant strains (****P < 0.0001, One-way ANOVA with Dunnett’s multiple comparisons test). The data for each time point represent the mean value ± standard deviation.

The pathway enrichment analysis comparing WT and Δcrp strains during the early stationary phase cultures.

The analysis was conducted using MetaboAnalyst, with a threshold ratio (Δcrp/WT) set at ≤ 0.5 for downregulation and ≥ 2 for upregulation. (A) Upregulated and (B) Downregulated pathways in the Δcrp strain compared to WT during the early stationary growth phase (ESP).

The pathway enrichment analysis for the WT strain.

The analysis was conducted using MetaboAnalyst, with a threshold ratio (LSP/ESP) set at ≤ 0.5 for downregulation and ≥ 2 for upregulation. (A) Upregulated and (B) Downregulated pathways of the WT strain in the late stationary growth phase (LSP) compared to the WT strain in the early stationary phase (ESP).

The pathway enrichment analysis for the Δcrp strain.

The analysis was conducted using MetaboAnalyst, with a threshold ratio (LSP/ESP) set at ≤ 0.5 for downregulation and ≥ 2 for upregulation. (A) Upregulated and (B) Downregulated pathways of the Δcrp strain in the late stationary growth phase (LSP) compared to the Δcrp strain in the early stationary phase (ESP).

The MS analysis of proteins from both WT and Δcrp strains at the late stationary phase.

The proteomic data were subjected to unsupervised hierarchical clustering. Each column in the figure represents a biological replica. n=3.

Persister levels of E. coli WT, Δcrp, and ΔcyaA cells with the integrated mCherry expression system.

Late stationary phase cultures (t=24 h) were transferred to fresh media and treated with ampicillin (200 μg/mL), ofloxacin (5 μg/mL), and gentamicin (50 μg/mL) for 20 hours. Subsequently, 1 mL of the treated culture underwent two washes with 1X PBS to remove antibiotics. It was then serially diluted and plated on an agar plate to count the CFUs. The levels of ofloxacin and gentamicin persisters in the mutant strains were below the limit of detection. n=4. Statistical significance was observed between control and mutant strains (***P < 0.001, ****P < 0.0001, Two-way ANOVA with Tukey’s multiple comparisons test). The data for each time point represent the mean value ± standard deviation.

Non-growing cell levels in the E. coli strain carrying the Crp expression system.

(A, B) Flow cytometry histograms depict mCherry expression in Δcrp+pUA66-crp, and Δcrp+pUA66 EV at early (t=5 h) and late (t=24 h) stationary phases, respectively. Cells containing an IPTG-inducible mCherry expression system were cultivated with IPTG. After washing and dilution of early and late stationary phase cells in IPTG-free fresh media, fluorescence was tracked in non-growing and growing cells for 2.5 hours. The panel is a representative biological replicate. Consistent results were seen across all 3 biological replicates.

RSG staining control for bacterial metabolic activities.

Exponential phase (t=3 h) cells were stained with 1 μM RSG for 10 minutes at 37°C before analyzing by flow cytometry. Unstained cells and cells treated with 20 μM CCCP + 1 μM RSG were used as control. CCCP was expected to reduce cellular redox activities. The panel is a representative biological replicate. Consistent results were seen across all 3 biological replicates.

Intact (non-lysed) cell levels of E. coli WT, Δcrp, and ΔcyaA cells with the integrated mCherry expression system.

mCherry positive cells were diluted into fresh media and treated with ampicillin (200 μg/mL) for 20 hours. Flow cytometry was used to measure the red fluorescence of intact surviving cells at time points t=0, 1, 2, 4, 6 and 20 h. A representative biological replicate is shown, with consistent results across all 3 replicates.

VBNC levels of E. coli WT, Δcrp, and ΔcyaA cells with the integrated mCherry expression system.

Flow cytometry was employed to quantify intact surviving cells. Persister cells were quantified by plating the cells on agar media. Viable but nonculturable (VBNC) cells were enumerated by subtracting persister levels from the intact cell levels. n=4. Statistical significance was observed between control and mutant strains (****P < 0.0001, One-way ANOVA using Dunnett’s multiple comparisons test). The data for each time point represent the mean value ± standard deviation.

Cell counts of E. coli K-12 MG1655 WT and mutant strains using flow cytometry at late stationary phase.

Cells were diluted 100-fold into 1 mL of 1X PBS. n=4. The data for each time point represent the mean value ± standard deviation.