Cyclic di-AMP inhibits Listeria monocytogenes thymineless death during infection

Abstract

Anti-folate antibiotics are used to treat meningitis and refractory listeriosis caused by drug-resistant Listeria monocytogenes (Lm). Their bactericidal activity is attributed to the deactivation of thymidylate synthase (ThyA), which subsequently induces bacterial cell death when thymidine is depleted-a process known as thymineless death (TLD). Despite decades of study, the mechanisms of TLD, especially during infection, remain unclear. Cyclic di-AMP (c-di-AMP), a common bacterial second messenger that regulates bacterial stress responses, is elevated in response to anti-folate antibiotics. In this study, we found that elevated c-di-AMP is required to inhibit TLD in Lm. Conversely, reducing c-di-AMP levels in the ΔthyA mutant led to increased bacterial cell death under thymidine starvation and significant reduction in intracellular growth. Furthermore, we found that ΔthyA exhibited a more pronounced growth defect during oral infection compared to intravenous infection, due to limited thymidine availability in the gallbladder, which acts as a bottleneck for ΔthyA in establishing infection. Notably, decreasing c-di-AMP levels abolished the infection capacity of ΔthyA in both infection models. Finally, we identified that the c-di-AMP-binding protein PstA contributes to bacterial cell death when c-di-AMP concentrations are low. Deletion of pstA in the ΔthyA background rescued the elevated cell death caused by c-di-AMP depletion both in vitro and during mouse infections. Our study identifies a previously unrecognized mechanism of TLD regulation mediated by c-di-AMP. This expands fundamental knowledge of TLD in the context of infection and provides insight into potential combined therapeutic strategies for listeriosis targeting both anti-folate and c-di-AMP metabolic pathways.

Figure 1.. C-di-AMP functions as a global regulator for antibiotic resistance.

(A) The heatmap indicates the log₂ ratio of the CFU recovery of ΔdacA::disA (low c-di-AMP strain) relative to WT strain exposed to various antibiotics. The ΔdacA::disA strain, in which DacA was replaced by an IPTG-inducible diadenylate cyclase, was cultured without IPTG for 16 h to halt c-di-AMP production. Blue texts indicate antibiotic classes according to their mechanisms of action, “ND” indicates that no bacteria recovery after treatment and antibiotic susceptibility was not determined. (B) and (C) Antibiotic susceptibility of Lm strains measured by disk diffusion assays. The inhibition zone of each strain was normalized to that of the WT. For all panels, mean values of biological replicates are plotted, and error bars indicate ±SD, P values were calculated using Ordinary one-way analysis. Asterisks indicate that differences are statistically significant (*, P < 0.05; ****, P < 0.0001), and ns indicates no significant difference. (D) Schematic of thymidine and c-di-AMP metabolism in bacteria. ThyA, thymidylate synthase; NupC, nucleoside permease; DacA, diadenylate cyclase; PgpH, phosphodiesterase; PdeA, phosphodiesterase; MdrMTAC, multidrug transporters.

Figure 2.. C-di-AMP is crucial for maintaining cell viability during thymidine deprivation in vitro.

(A) Intracellular c-di-AMP concentration of Lm strains quantified using the CDA-Luc assay. The Lm strains were grown overnight statically at 30 °C in BHI broth supplemented with different concentrations of thymidine. (B) Growth curves of Lm strains with varying concentrations of thymidine supplementation in BHI broth incubated statically at 30 °C. (C) Composite phase-contrast and red fluorescence images of Lm strains stained with propidium iodide. ΔthyA and ΔthyA::pdeA strains were grown on LB agar without thymidine and supplemented with propidium iodide during the 11 hours of imaging, which penetrates only the cell envelopes of dead cells. (D) Relative CFU recovery of bacteria incubated in LB broth without thymidine. The percentage of survival at each time point was calculated by normalizing the CFU count to the value at time 0. PdeA, c-di-AMP phosphodiesterase, PdeA_84–657, enzymatic domain of PdeA. PdeA_DHH-AAA, enzymatic dead mutant of PdeA. For all panels, mean values of biological duplicates are plotted, and error bars indicate ±SD, P values were calculated using either Two-way ANOVA (A) or Ordinary one-way ANOVA (B and D) analysis. Asterisks indicate that differences are statistically significant (*, P < 0.05; **, P < 0.01; ***, P < 0.001, ****, P < 0.0001), and ns indicates no significant difference.

Figure 3.. Elevated c-di-AMP production is required for the intracellular growth of ΔthyA.

(A) Schematic of macrophage infection with Lm. The iBMDMs were infected with Lm with an MOI of 1 in the presence or absence of extracellular thymidine supplementation. Gentamicin was supplemented at 0.5 hpi to kill the extracellular bacteria. (B) and (C) Intracellular growth curve of Lm in iBMDMs as describe in (A). (D) Fluorescence microscopy images of iBMDMs infected with Lm at an MOI of 10 at 6 hpi. Nuclei were stained with DAPI (blue), F-actin with Alexa Fluor 568 phalloidin (red), and Lm with anti-Lm antibody (green). (E) Cell length quantification of intracellular Lm as indicated in (D). Mean values of cell length of 15 bacterial cells from three different macrophages were plotted. (F) and (G) Plaque size of L2 fibroblast with Lm infection. The monolayer of fibroblasts was infected with Lm for 2 days with gentamicin and various concentrations of thymidine supplementation and then stained with neutral red. The plots represent the median of 20 plaques measured from three different wells, with plaque diameters normalized to the control. (H) (I) and (J) Intracellular growth curve of Lm in iBMDMs with or without SXT treatment. The iBMDMs were infected with Lm with an MOI of 1 in the presence or absence of IPTG supplementation. Gentamicin and SXT was supplemented at 0.5 hpi. For panel (B), (C), (H), and (I), mean values of biological duplicates are plotted. For all panels, error bars indicate ±SD, P values were calculated using Ordinary one-way ANOVA (B, C, F, G) or Two-way ANOVA (E) analysis. Asterisks indicate that differences are statistically significant (*, P < 0.05; **, P < 0.01; ****, P < 0.0001), and ns indicates no significant difference.

Figure 4.. Elevated c-di-AMP production is required for the growth of ΔthyA during mouse infection.

(A) Schematic of retro-orbital (R.O.) Lm infection in mice. C57BL/6 mice were infected with 10⁵ CFU/mouse via the retro-orbital route. Livers and spleens were harvested for CFU enumeration at 72 hpi. (B) Weight loss of infected mice as indicated in panel (A). (C) and (D) CFU recovery of Lm from spleens and livers as describe in (A). (E) Schematic of oral Lm infection in mice. C57BL/6 mice were given 5 mg/mL streptomycin in drinking water for 48 hours, fasted overnight, and then orally inoculated with 10⁸ Lm. Livers and spleens were harvested for CFU enumeration at 72 hpi. (F) to (L) CFU recovery of Lm from mouse tissue or feces samples, as described in (E). For panel B, mean values of five biological replicates are plotted, and error bars indicate ±SD, P values were calculated using Ordinary one-way ANOVA analysis. For all CFU recovery in mouse infections, biological replicates are plotted. Horizontal black bars indicate the median of the data. P values were calculated using Mann-Whitney analysis. Asterisks indicate that differences are statistically significant (*, P < 0.05; **, P < 0.01; ***, P < 0.001, ****, P < 0.0001), and ns indicates no significant difference. LoD denotes the limit of detection.

Figure 5.. PstA promotes thymineless death in the absence of c-di-AMP.

(A) Optical density (600 nm) of Lm strains grown in BHI with 1.25 μg/ml thymidine for 6 hours. (B) Bacterial survival in the absence of thymidine. Survival rate at each time point was calculated by normalizing CFU counts to the value at time 0. (C) and (D) CFU recovery of Lm from spleens and livers. C57BL/6 mice were infected with 10⁵ CFU Lm per mouse via the retro-orbital route. Livers and spleens were harvested for CFU enumeration at 72 hpi. For panels (A) and (B), mean values of biological replicates are plotted, and error bars indicate ±SD, P values were calculated using Ordinary one-way ANOVA analysis. For all mouse infection panels, biological replicates are plotted. Horizontal black bars indicate the median of the data. P values were calculated using Mann-Whitney analysis. Asterisks indicate that differences are statistically significant (*, P < 0.05; **, P < 0.01; ***, P < 0.001, ****, P < 0.0001), and ns indicates no significant difference.