Pyruvate-depleting conditions induce biofilm dispersion and enhance the efficacy of antibiotics in killing biofilms in vitro and in vivo

Abstract

The formation of biofilms is a developmental process initiated by planktonic cells transitioning to the surface, which comes full circle when cells disperse from the biofilm and transition to the planktonic mode of growth. Considering that pyruvate has been previously demonstrated to be required for the formation of P. aeruginosa biofilms, we asked whether pyruvate likewise contributes to the maintenance of the biofilm structure, with depletion of pyruvate resulting in dispersion. Here, we demonstrate that the enzymatic depletion of pyruvate coincided with the dispersion of established biofilms by S. aureus and laboratory and clinical P. aeruginosa isolates. The dispersion response was dependent on pyruvate fermentation pathway components but independent of proteins previously described to contribute to P. aeruginosa biofilm dispersion. Using porcine second-degree burn wounds infected with P. aeruginosa biofilm cells, we furthermore demonstrated that pyruvate depletion resulted in a reduction of biofilm biomass in vivo. Pyruvate-depleting conditions enhanced the efficacy of tobramycin killing of the resident wound biofilms by up to 5-logs. Our findings strongly suggest the management of pyruvate availability to be a promising strategy to combat biofilm-related infections by two principal pathogens associated with wound and cystic fibrosis lung infections.

Figure 1

Exposure of P. aeruginosa biofilms to active pyruvate dehydrogenase (PDH) coincides with a reduction in the biofilm biomass in a manner independent of biofilm age. Biofilms were grown for 4 days in 24-well polystyrene plates in five-fold diluted LB. (A) Remaining biofilm biomass following exposure to 5, 10, and 20 mU PDH, as determined using CV staining. Inset, CV-stained biofilms prior to and post treatment with PDH. (B) Absorbance of biofilm supernatant following exposure to PDH or heat-inactivated PDH. (C) Brightfield images of biofilms grown for 3, 6, and 7 days prior to and post treatment with 10 mU PDH. (D) Biofilm biomass following exposure to heat-inactivated PDH. Remaining biofilm biomass following exposure to (E) 10 mM lactate or cofactors and products of the PDH catalyzed reaction, namely ß-NAD⁺ and acetyl-CoA, ß-NADH, and 10 mM nitrate, or to (F) increasing concentration of pyruvate (0, 1, 10, and 100 mM) in the presence and absence of PDH. Untreated biofilms were used as controls. PDH treatment was done in the presence of CoA, ß-NAD⁺, TPP, and MgSO_4. *Significantly different (p < 0.05) from untreated biofilms. All experiments were carried out in triplicate. Error bars denote standard deviation.

Figure 2

Pyruvate depletion coincides with P. aeruginosa biofilms showing signs of dispersion. (A) Representative confocal images of biofilms grown for 4 days in 24-well polystyrene plates that were left untreated (control), or were exposed to 10 mU PDH or 10 mU heat-inactivated PDH (HK_PDH). Red arrow indicates central hollowing of microcolony architecture. White size bar = 100 µm. (B) Number of microcolonies as percent of the total colonies counted per treatment group, showing void formation indicative of dispersion. (C) Average diameter of microcolonies that appeared to be dispersed or not dispersed. Inset, representative images of microcolonies that were considered in the analysis as “not dispersed” or “dispersed”. Error bars represent standard deviation of biological replicates. *Statistically different (p < 0.05) from untreated biofilms.

Figure 3

Pyruvate-depletion induced dispersion is dependent on MifR and functional ldhA. (A,B) Biofilms by indicated P. aeruginosa PAO1 mutant strains grown for 4 days in 24-well polystyrene plates that were left untreated or exposure to 10 mU PDH and cofactors for 16 h. (A) Representative confocal images and (B) quantitative analysis of the biofilm biomass by P. aeruginosa PAO1 and indicated mutant strains. (C) Remaining biofilm biomass of P. aeruginosa PA14 and indicated mutant strains that were either left untreated or exposure to 10 mU PDH and cofactors for 16 h, as determined using CV staining. White size bar = 100 µm. *Statistically different from untreated biofilms (p < 0.05). All experiments were carried out at least in triplicate. Error bars denote standard deviation.

Figure 4

Pyruvate-depletion induced dispersion is independent of BdlA, DipA, and RbdA previously described to play a role in nutrient-induced dispersion. (A) CV-stained biomass and (B) representative brightfield images of biofilms by P. aeruginosa PAO1 and indicated mutant strains that either left untreated or exposure to 10 mU PDH and cofactors for 16 h. White size bar = 100 µm. *Statistically different from untreated biofilms (p < 0.05). All experiments were carried out at least in triplicate. Error bars denote standard deviation.

Figure 5

Effect of enzymatic depletion of pyruvate on biofilms by P. aeruginosa clinical isolates and Staphylococcus aureus. (A) Biofilm biomass of biofilms by P. aeruginosa strains PAO1, PA14, and clinical strains isolated from the cystic fibrosis lung, chronic wounds, and the urinary tract, left untreated or exposed to 10 mU PDH, as determined using CV staining. (B) Biofilm biomass of 4-day old S. aureus biofilms left untreated (control), and exposed for 16 h to 10 mU PDH or heat-inactivated PDH (HK_PDH), as determined using CV staining. (C) Representative confocal images of S. aureus biofilms left untreated (control), and exposed to 10 mU PDH or heat-inactivated PDH (HK_PDH). (D) COMSTAT analysis of the S. aureus biofilm biomass. (E–G) S. aureus biofilms were grown for 4 days in the absence (control) or continued presence of heat-inactivated PDH (HK PDH) or 10 mU PDH. (E) CV staining of remaining biofilm biomass following 4 days of growth. (F) Representative confocal images of S. aureus biofilms left untreated or continuously exposed to 10 mU PDH. (G) COMSTAT analysis of confocal of S. aureus remaining biofilm biomass. Size bars = 100 µm. All experiments were carried out at least in triplicate. Error bars represent standard deviation. *Significantly different from untreated biofilms (p < 0.05).

Figure 6

Pyruvate-depletion induced dispersion renders P aeruginosa biofilm cells more susceptible to the antibiotic tobramycin. (A) Viability (CFU) of P. aeruginosa PAO1 biofilms left untreated and exposed to tobramycin (150 µg/ml) or 10 mU PDH plus 150 µg/ml tobramycin for 1 h. (B) Viability (CFU) of P. aeruginosa PAO1 planktonic cells left untreated or exposed to 10 mU PDH for 1 h. Experiments were carried out at least in triplicate. Error bars represent standard deviation. *Significantly different (p < 0.05) from untreated biofilms.

Figure 7

Pyruvate-depletion induced dispersion of P. aeruginosa biofilms within second-degree porcine burn wounds reduces the bacterial burden and enhances the efficacy of tobramycin in killing biofilm cells. Burn wounds were infected with P. aeruginosa ATCC® 27312™ and left untreated for 24 h. Wounds were subsequently exposed to increasing concentrations of PDH, tobramycin, or silver sulfadiazine. At days 1, 3 and 6 post infections, adherent and non-adherent bacterial cells were removed using flush & scrub, and the number of viable cells (CFU) per wound determined using viability counts. (A) Number of adherent (biofilm) and non-adherent (planktonic) cells present in wounds. (B) Effect of increasing concentrations of PDH on the P. aeruginosa biofilm population present in burn wounds. Efficacy of PDH (100 and 200 mU), tobramycin (100 µg/ml), or co-treatment of PDH and tobramycin on the (C) P. aeruginosa biofilm population and (D) planktonic population present in burn wounds. Log reduction was determined relative to untreated biofilms post 3 and 6 days infection. (E) Number of adherent (biofilm) and non-adherent (planktonic) cells present in wounds post exposure to silver sulfadiazine (SSD) 3 and 6 days post infection. Prior to SSD treatment, infected wounds were left untreated for 24 h post infection. With the exception of data shown in (B), all experiments were done in in triplicate, with each biological replicate being comprised of 3 wounds per treatment group (n = 9). Experiments shown in (B) are representative data obtained using only 3 wounds per treatment group (n = 3). Error bars represent standard deviation. Error bars represent standard deviation. *Significantly different (p < 0.05) from untreated wounds. **Significantly different (p < 0.05) from wounds treated only with PDH or tobramycin.