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

Bacterial strains and growth conditions

All bacterial strains and plasmids used in this study are listed in Table 1. Escherichia coli was cultured at 37 °C on Lysogeny Broth (LB) agar (Difco Laboratories, Detroit, MI, USA). P. chlororaphis strains were routinely cultured on LB or in M9 minimal salts medium supplemented with 0.2% glucose and 1mM magnesium sulphate (M9-glc) at 28 °C. Media was supplemented with the following antibiotics: ampicillin (Amp; 100 µg/ml), gentamicin (Gm; 15 µg/ml) for E. coli, and piperacillin (Pip; 40 µg/ml), Gm (20 µg/ml), tetracycline (Tc; 15 µg/ml) for PA23. All antibiotics were obtained from Research Products International Corp. (Mt. Prospect, IL, USA).

Acanthamoeba strain and culture conditions

Acanthamoeba castellanii (ATCC 30234) was grown axenically without shaking in 20 ml of PYG medium (proteose peptone 10 g/L, yeast extract 5 g/L, glucose 10 g/L and the additives: 4 mM MgSO₄•7H₂O, 0.4 mM CaCl₂, 0.05 mM Fe(NH₄)2(SO₄)2•6H₂O, 2.5 mM Na₂HPO₄•7H₂O, 2.5 mM KH₂PO₄, 0.1 M glucose and 3.4 mM sodium citrate•2H₂O) in T75 tissue culture flasks (Sarstedt, Saint-Leonard, QC, Canada) in a humidified incubator at 25 °C. Before the experiment, cultures were washed three times with Ac buffer (PYG medium lacking proteose peptone, yeast extract and glucose) (Moffat & Tompkins, 1992) to remove non-adherent cells or any existed cysts. Amoeba trophozoite density was measured using a Neubauer cell counting chamber. To obtain amoeba cell-free supernatant, cells were grown in Ac buffer for three days at 25 °C and then filtered using 0.22 µm filters (Sarstedt). Supernatants were stored at −80 °C.

P. chlororaphis PA23—Ac co-culture assays

To study amoeba-bacterial interactions, Ac trophozoites were washed three times with Ac buffer. Amoebae were adjusted to 10⁶ cells/ml, and 1-ml aliquots were transferred into wells of a 24-well plate and incubated at 28 °C for 1 h to allow adherence. Wells were washed three times with Ac buffer to remove non-adherent amoebae. Next, bacterial suspensions grown in M9-glc were washed twice with M9-glc and adjusted to 10⁸ CFU/ml. A 1-ml volume of bacteria was added to each well and allowed to incubate at 28 °C for 15 days. Growth and viability of amoeba trophozoites was determined by microscopic visualization of eosin-stained cells (dead cells stain red; live cells remain unstained; cysts are morphologically distinct) following the method of Neilson, Ivey & Bulmer (1978). Briefly, a 10-µl volume of a 0.5% (v/v) basic eosin solution was added to a 10-µl volume of Ac trophozoites grown in the presence and absence of bacteria. Ac cells were examined and enumerated with a Neubauer cell counting chamber. The number of extracellular bacteria residing in the co-culture samples was determined through viable plate counting. Amoebae grown in the absence of bacteria were used as a negative control. Experiments were repeated three times.

Effect of secondary metabolites on Ac trophozoite viability

Cultures of PA23 strains were incubated for 3 days at 28 °C, and then cells were pelleted, and the supernatant passed through a 0.2-µm filter to remove all bacterial cells. The experiment setup was the same as described above, except that bacterial supernatant was added to the wells. The morphological changes of amoeba were monitored using an inverted microscope (Zeiss Observer Z1 inverted microscope; Carl Zeiss Microscopy GmbH, Göttingen, Germany). To determine the effect of purified compounds on Ac, amoebae were adjusted to 10⁶ cells/ml in Ac buffer containing commercially purified PRN (Sigma, St. Louis, MO, USA) at the following concentrations: 0 µg/ml (control), 0.1, 0.5, 1, 5 and 10 µg/ml, or KCN (Sigma) at the following concentrations: 0 µg/ml (control), 50, 100, 200, 400 and 800 µg/ml. For PHZ analysis, a 15-ml overnight culture of PA23-PRN⁻ grown in LB medium was used to extract PHZ following the method of Selin et al. (2010). In brief, PHZ extractions were quantified with UV-visible spectroscopy (Biochrom Ltd. Cambridge, England), and the absorption maxima for PCA and 2-OH-PHZ were measured at 367 and 490 nm, respectively. Amoeba cells were incubated with extracted PHZ at the following concentrations: 0, 10, 20, 35, and 50 µg/ml. Ac subcultures containing various concentrations of PRN, PHZ and KCN were incubated in 24-well plates at 28 °C, and amoeba viability was monitored at 1, 6, 12, 18, 24, 48 and 72 h. Five replicates were included per assay, and the experiment was repeated three times.

Chemotaxis assays

Bacteria were grown in M9-glc medium at 28 °C for 24 h. Axenically grown Ac cultures were prepared as described above. Petri dishes (60 × 15 mm) containing 5 ml of 1.5% water agar had three wells created 20 mm apart. An aliquot of 50 µl of Ac (10⁶ cells/ml) was transferred into the centre well. One of the two outside wells contained 50 µl of “test” bacterial inoculum, while the second well contained a 50 µl suspension of PA23 WT, the gacS mutant, or saline as the “control” sample. A hand-crafted grid coverslip was placed underneath the plate for counting amoebae that had migrated from the centre well towards the outside wells moving under the agar. The chemotactic index was calculated based on the formula: number of amoebae migrating towards the test sample/number of amoebae moving towards the control sample. Experiments wherein the test and control wells contained the same sample, namely saline and PA23 WT were used to demonstrate equal attraction as a means of calibrating the test. Experiments were repeated three times.

Analysis of transcriptional fusions in the presence and absence of Ac

The activity of prnA-, phzA-, phzI-, phzR-, rpoS-, and gacS-lacZ transcriptional fusions was determined in PA23 cultured in the presence and absence of Ac trophozoites. Active amoebae were adjusted to 10⁶ cells/ml in Ac buffer, as described above. Overnight bacterial cultures grown in M9-glc were adjusted to an optical density of 0.1 (2 × 10⁸ CFU/ml) prior to co-culture with Ac or amoeba-free supernatant. Samples were grown for 24, 48, and 72 h in M9-glc at 28 °C. The effect of Ac cells or cell-free supernatant on PA23 gene activity was determined by β-galactosidase assays (Miller, 1972) and experiments were repeated three times.

Antifungal activity

To assess the ability of PA23 and derivative strains to inhibit the growth of S. sclerotiorum in vitro, radial diffusion assays were performed as described by Poritsanos et al. (2006). Bacteria were grown in M9-glc in the presence and absence of trophozoites for 72 h at 28 °C. Five replicates were analyzed for each strain and experiments were repeated three times.

Autoinducer detection assay

AHL analysis was conducted by spotting a 5-µL aliquot of cultures onto LB agar seeded with C. violaceum CV026. Strain CV026 is able to detect exogenous AHLs with carbon chain lengths ranging from C4-C8, resulting in a deep purple halo surrounding the bacterial colony (Latifi et al., 1995). The diameter of purple pigment surrounding the bacterial colonies was measured (Poritsanos et al., 2006). Five replicates were analysed for each strain, and the experiment was repeated three times.

Protease analysis

Extracellular protease production was determined qualitatively by inoculating a 5-µL volume of bacterial culture onto a 1.5% agar plate containing 2% skim milk (Difco). Protease activity was indicated by a zone of lysis surrounding the colony after 36–48 h growth at 28 °C (Poritsanos et al., 2006). Zones of clearing were measured for each strain. Data represent the average of five replicates, and the experiment was repeated three times.

Motility analysis

To assess the impact of amoebae on PA23 flagellar (swimming) motility, bacteria were grown overnight in M9-glc in the presence and absence of trophozoites. A 5-µl volume of culture containing bacteria or bacteria plus Ac were inoculated below the surface of a 0.3% LB agar plate (Poritsanos et al., 2006). The plates were incubated at 28 °C and the swim zone diameter was measured at 24, 48 and 72 h. For the assays, five replicates were analyzed, and the experiment was repeated three times.

Quantitative analysis of PRN and PHZ

Production of PRN was quantified by HPLC as described by Selin et al. (2010) with the following modifications. PA23 cultures started at an initial concentration of 10⁸ CFU/ml were grown in the presence and absence of the amoeba (10⁶ cells/ml) at 28 °C in 30 ml M9-glc. PRN was extracted and quantified after 96 h. Toluene was added to the culture supernatants as an internal control. Peaks corresponding to the toluene and PRN were analyzed by UV absorption at 225 nm using a Varian 335 diode array detector. For analysis of PHZ production, PA23 cultures (10⁸ CFU/ml), were grown in the presence and absence of Ac (10⁶ cells/ml) in 30 ml M9-glc at 28 °C for 72 h. PHZ was quantified according to the method of Selin et al. (2010) as described above. PRN and PHZ analysis was performed in triplicate and experiments were repeated twice.

Statistical analysis

An unpaired Student’s t test was used for statistical analysis of PRN, PHZ, AHL production, swimming motility, AF activity and protease production. The Tukey test was applied to determine the chemotactic preference of amoebae for each of the bacterial strains. The two-way ANOVA test was applied for amoeba-bacterial co-culture assays and gene expression analysis.

PA23 affects Ac trophozoite viability

To determine the impact of PA23 on Ac trophozoite viability and cyst formation, the PA23 WT and derivative strains, including regulatory (rpoS⁻, gacS⁻, phzR⁻ and AI-deficient) and biosynthetic mutants (PRN⁻, PHZ⁻, and HCN⁻) were offered as prey. Previous phenotypic analysis showed that the gacS⁻, phzR⁻ and AI-deficient strains produce little to no antibiotics and degradative enzymes (Poritsanos et al., 2006; Selin, Fernando & De Kievit, 2012). As illustrated in Fig. 1, in the presence of the gacS⁻, phzR⁻, and AI-deficient strains, Ac trophozoite numbers increased on days 1 and 5, after which the amoebae remained active but the population declined slowly. When co-incubated with PA23 WT, PHZ⁻, PRN⁻, HCN⁻ and rpoS⁻cells, the number of trophozoites steadily decreased over time (Fig. 1). Fig. S1 depicts the proportion of viable, encysted and dead Ac at each time point. At day 15, there were less trophozoites present in co-cultures with the PHZ⁻ strain, compared to PA23 WT. We have previously demonstrated that this bacterium and the rpoS mutant secrete increased levels of PRN relative to WT (Manuel et al., 2012; Selin et al., 2010). Conversely, trophozoite numbers were significantly elevated when grown on the PRN⁻ and HCN⁻ strains compared to those grown on the PHZ- strain at day 15 (Fig. 1). Collectively, these results indicate that PA23 exoproducts play a role in the inhibition of Ac growth.

Bacterial persistence upon co-culturing with Ac trophozoites

To investigate whether bacterial growth was affected by the presence of amoebae, bacteria were co-cultured with Ac and viability was assessed over time. The number of PA23 WT, rpoS⁻ and PHZ⁻ cells increased from 10⁸ CFU/ml on day 0 to between 9.7 × 10⁸ and 9.8 × 10⁸ CFU/ml on day 1 (Fig. 2). The HCN⁻ and PRN⁻ strains also increased from 10⁸ CFU/ml to 7.5 × 10⁸ and 7.6 × 10⁸ CFU/ml, respectively. The QS-deficient phzR⁻ and AI⁻ derivatives showed smaller increases in population size, whereas the gacS numbers declined to 3.1 × 10⁷ CFU/ml. On day 5, the PA23 WT and the PRN over-producing PHZ⁻ and rpoS⁻ strains continued to increase in abundance. The PRN⁻ and HCN⁻ populations also increased but to a lesser degree. For the gacS⁻ mutant, there were no viable bacteria detected, while the number of QS-deficient cells was dramatically reduced. Bacteria populations continued to decrease by day 10, with the largest number of cells remaining for the PA23 WT, PHZ⁻ and rpoS⁻ strains (Fig. 2). There were no viable cells recovered on day 15 (data not shown). In the absence of Ac, there were no observable differences in bacterial viability between strains over time (Fig. S2).

The effect of PA23 metabolites on Ac viability

To further explore the impact of PA23 exoproducts on Ac trophozoites, amoebae were challenged with cell-free supernatant from the PA23 WT and the gacS mutant (Fig. 3). After 1 h incubation with WT supernatant, amoeba cells started to swell and this continued until they began to burst at 2 h (Fig. 3A). Conversely, incubation with gacS⁻ supernatant did not affect amoeba morphology (Fig. 3B). Next, we investigated how feeding on nontoxic bacteria in the presence of antifungal (AF) metabolites impacts amoebae. Ac trophozoites were co-cultured with GFP-tagged gacS⁻ mutant cells resuspended in PA23 cell-free supernatant. As depicted in Fig. 3C, at 1 h, amoebae had lost their amoebic shape. After 2 h incubation, trophozoites were fluorescing green consistent with uptake of the gacS⁻ cells. Despite the fact that the trophozoites were actively feeding, they underwent the same morphological changes as when challenged with PA23 WT supernatant alone (Fig. 3A), including cell lysis (Fig. 3C). Collectively these finding indicate that secreted PA23 metabolites exert deleterious effects on Ac trophozoites.

To explore whether purified compounds would exhibit the same toxicity, trophozoites were challenged with PRN (0-10 µg/ml), PHZ (0-50 µg/ml) and KCN (0-800 µg/ml). As illustrated in Fig. 4A, when exposed to PRN at a concentration of 1 µg/ml or lower, there was no impact on amoeba viability. However at higher PRN levels, the number of Ac trophozoites declined in a dose-dependent fashion (Fig. 4A). PHZ was also found to exhibit toxic effects on the amoebae. Exposure to ≤20 µg/ml PHZ had little effect on protozoan survival, but at higher concentrations (35-50 µg/ml), amoeba viability decreased to less than 50% after 24 h (Fig. 4B). Exposure to KCN led to a reduction in the number of Ac trophozoite at concentrations of 400 µg/ml and above (Fig. 4C).

PA23 exoproducts affect the chemotactic response of Ac

Bacterial metabolites can exhibit either attractant or repellent effects, and this can ultimately impact predator grazing. To study the chemotactic response of Ac towards bacteria, binary choice assays were undertaken, as depicted in Fig. 5A. Compared to saline control, trophozoites were more attracted to the gacS⁻, QS-deficient, and PRN⁻ strains, all of which lack PRN production (Fig. 5B). Whereas PA23 WT, and the PRN hyper-producing PHZ⁻ and rpoS⁻ mutants exhibited a repellent effect. Amoebae were marginally attracted to the HCN⁻ strain (Fig. 5B). Employing PA23 as the control, amoebae preferentially migrated towards all of the strains except for the PRN overproducers (PHZ⁻ and rpoS⁻ mutants; Fig. 5C). Trophozoites clearly had a strong preference for the gacS⁻ derivative because when it was included as the control, Ac consistently migrated towards this bacterium (Fig. 5D). Once again, the PRN-producers (PHZ⁻, rpoS⁻, HCN⁻ and PA23 WT) exhibited the strongest repellent activity (Fig. 5D).

Growth in the presence of Ac affects PA23 gene expression

To determine whether bacteria can sense the presence of the predator, PA23 was grown together with amoebae cells or cell-free supernatants and monitored for changes in gene expression. For this assay, biosynthetic (prnA and phzA) and regulatory genes (phzI, phzR, rpoS, gacS) were analyzed. No changes in gene expression were observed in bacteria incubated with Ac cell-free supernatants. However, co-incubation of PA23 with trophozoites resulted in elevated expression of phzA and prnA at both 48 h and 72 h (Fig. 6). For the QS genes phzI and phzR, co-culturing resulted in a significant increase in phzI-lacZ activity at all time points tested, whereas phzR activity was elevated at only 48 h (Fig. 6). Growth with trophozoites led to an increase in the rpoS-lacZ activity at 72 h, while no change in gacS expression was observed at any of the time points (Fig. 6).

Impact of Ac on PA23 phenotypic traits

Phenotypic analysis was undertaken to determine whether changes in secondary metabolite production or other traits were brought on by growth in the presence of Ac. As outlined in Table 2, co-incubation with amoebae led to increased PRN and PHZ production, consistent with the elevated phzA and prnA gene activity. Other phenotypic traits, including fungal inhibition, protease activity, and swimming motility, were unaffected by Ac (Table 2).