Pseudomonas chlororaphis PA23 metabolites protect against protozoan grazing by the predator Acanthamoeba castellanii

Abstract

Background: Pseudomonas chlororaphis strain PA23 is a biocontrol agent that is able to protect canola against the pathogenic fungus Sclerotinia sclerotiorum. This bacterium secretes a number of metabolites that contribute to fungal antagonism, including pyrrolnitrin (PRN), phenazine (PHZ), hydrogen cyanide (HCN) and degradative enzymes. In order to be successful, a biocontrol agent must be able to persist in the environment and avoid the threat of grazing predators. The focus of the current study was to investigate whether PA23 is able to resist grazing by the protozoan predator Acanthamoeba castellanii (Ac) and to define the role of bacterial metabolites in the PA23-Ac interaction.

Methods: Ac was co-cultured with PA23 WT and a panel of derivative strains for a period of 15 days, and bacteria and amoebae were enumerated on days 1, 5, 10 and 15. Ac was subsequently incubated in the presence of purified PRN, PHZ, and KCN and viability was assessed at 24, 48 and 72 h. Chemotactic assays were conducted to assess whether PA23 compounds exhibit repellent or attractant properties towards Ac. Finally, PA23 grown in the presence and absence of amoebae was subject to phenotypic characterization and gene expression analyses.

Results: PRN, PHZ and HCN were found to contribute to PA23 toxicity towards Ac trophozoites, either by killing or inducing cyst formation. This is the first report of PHZ-mediated toxicity towards amoebae. In chemotaxis assays, amoebae preferentially migrated towards regulatory mutants devoid of extracellular metabolite production as well as a PRN mutant, indicating this antibiotic has repellent properties. Co-culturing of bacteria with amoebae led to elevated expression of the PA23 phzI/phzR quorum-sensing (QS) genes and phzA and prnA, which are under QS control. PHZ and PRN levels were similarly increased in Ac co-cultures, suggesting that PA23 can respond to predator cues and upregulate expression of toxins accordingly.

Conclusions: PA23 compounds including PRN, PHZ and HCN exhibited both toxic and repellent effects on Ac. Co-culturing of bacteria and amoebae lead to changes in bacterial gene expression and secondary metabolite production, suggesting that PA23 can sense the presence of these would-be predators and adjust its physiology in response.

Keywords: Amoebae; Antibiotic; Biocontrol; Cyst; Gene expression; Predator-prey interaction; Pseudomonad; Trophozoite.

Figure 1. Growth of Acanthamoeba. trophozoites on PA23 and derivative strains in M9-glc.

Amoeba growth and viability were monitored for 15 days. Asterisks indicate significant difference from the PA23 WT as determined by two-way ANOVA (*, P < 0.001; **, P < 0.0001). Note: PHZ⁻, PRN⁻ and HCN⁻ mutants are statistically significant at day 15 only, whereas the gacS⁻, AI-deficient, and phzR⁻ strains are statistically significant at days 1, 5, 10 and 15. Experiments were performed three times; one representative data set is shown.

Figure 2. Effect of Acanthamoeba castellani. trophozoites on the growth of PA23 and derivative strains in M9-glc.

Bacteria and amoeba were co-cultured for 15 days, and bacteria were enumerated on days 1, 5, 10 and 15. By day 15 there were no viable bacteria remaining. Asterisks indicate statistical significance of difference using two-way ANOVA (*, P < 0.01; **, P < 0.001; ***, P < 0.0001). Experiments were performed three times; one representative data set is shown.

Figure 3. Incubation of Acanthamoeba castellani. trophozoites with bacterial cells and cell-free supernatant.

PA23 WT cell-free supernatant (A), gacS-cell-free supernatant (B), GFP-tagged gacS-cells containing WT cell-free supernatant (C), and trophozoites in Ac buffer (D). Red arrows highlight swollen Ac trophozoites, and black arrows indicate Ac cell lysis. Images were captured using a Zeiss Observer Z1 inverted microscope under 40 × magnification. Scale bar = 10 µm.

Figure 4. Acanthamoeba castellani. trophozoites challenged with PRN (0–10 µg/ml) (A), PHZ (0–50 µg/ml) (B) and KCN (0–800 µg/ml) (C).

Asterisks indicate statistical significance of difference using two-way ANOVA (*, P < 0.01; **, P < 0.001). Three replicates were used per trial, and the experiment was repeated three times. One representative data set is shown.

Figure 5. Chemotactic response of Acanthamoeba castellani towards PA23 WT and derivative strains.

(A) Schematic diagram illustrating Petri plate set up. Active amoebae were placed in the center well; the test bacterium was placed in the test well, and PA23 WT, the gacS mutant or saline was added to the control well. Chemotactic preference assays were carried out against saline control (B), PA23 WT (C), and the gacS mutant (D). The chemotactic response was determined as follows: the number of amoebae migrating towards the test well/the number of amoebae migrating towards the control well. Values > 0 indicated attraction; values < 0 indicated repellent activity. Assays were performed in triplicate and the experiment was repeated three times. Error bars indicate ± SD; columns labelled with the same letter do not differ significantly by the Tukey test (P > 0.05).

Figure 6. The impact of Acanthamoeba castellani cells and cell free supernatant on prnA, phzA, phzI, phzR, gacS and rpoS expression in Pseudomonas chlororaphis PA23.

Co-cultures with Ac trophozoites (▴), Ac cell-free supernatant (■) and bacteria alone (•) were analyzed for β-galactosidase activity (Miller units) at 24, 48 and 72 h. Asterisks indicate statistical significance of difference using two-way ANOVA (*, P < 0.01). Experiments were performed three times; one representative data set is shown.