Vibrio gazogenes, ATCC 43942, was used in this study and was grown in Difco Marine Broth 2216 (BD Biosciences, Sparks, MD, United States) at 28°C in a shaking incubator (190 rpm) in the dark for 24 h before harvesting the cells for interaction experiments. This condition rendered ~1.6e⁷ Vg cells after harvest that we used for all treatments in this study (unless other concentrations are mentioned under the specific experiments). Other bacterial strains used included Serratia marcescens (ATCC 39006), Escherichia coli (ATCC 25922), Staphylococcus aureus (ATCC 25923), Vibrio harveyi (ATCC 14126), and Vibrio fischeri (ATCC 7744). Vibrio strains were grown in Marine Broth at 28°C in the dark in a shaking incubator (190 rpm) for 24 h, while the other strains were grown in Nutrient Broth in the dark at 37°C at 250 rpm for 12 h before harvesting. Under these conditions, all bacterial strains produced equal concentrations of cells as determined for Vg (~1.6e⁷ cells total). Fungal strains used in this study were stored as glycerol stocks at −80°C until they were needed. For the fungal experiments, the A. flavus wild-type strains NRRL 3357 or SRRC1529 (also known as AF70) strain were used. YES (6% sucrose, 2% yeast extract, pH 5.8) medium was used for experiments with liquid culture, and potato dextrose agar (PDA) was used for experiments with solid culture. All fungal strains were grown on glucose minimal medium (GMM) for spore production at 29°C. In some experiments using the pyrG auxotrophic strain, TJES 19.1, 0.56 g/L uracil, and 1.26 g/L uridine were supplemented, and for the argB auxotrophic strain, TJES 20.1, 1 g/L arginine was supplemented in the growth medium. In some cases, all three supplements were added, as described below. Heterokaryon analysis was assessed on GMM + 0.25% Triton X-100 medium.

Unless otherwise specified, all Vg treatments consisted of 24 h marine broth cultures of Vg that were pelleted by centrifugation at 4000 g for 15 min at room temperature (25°C), then resuspended in 1 mL YES media (~1.6e⁷ cells). This suspension was added to 100 mL YES growth medium inoculated with 10⁷ spores/mL of A. flavus and incubated in the dark at 29°C and 150 rpm for 40 h. For generating heat-inactivated and non-viable Vg and other bacteria (E. coli, S. aureus, V. harveyi, V. fischeri cells, and S. marcescens), the cultures were resuspended in 1 mL YES broth and incubated at 100°C for 10 min. Non-viability was confirmed by plating the heat-treated cells on bacterial growth medium (Marine agar at 29°C in the dark for Vibrio strains and Nutrient Agar at 37°C in the dark for other bacterial strains) and ensuring that no growth occurred over 24 h. Vg treatment of A. flavus on agar plates was conducted by center-point inoculating 10 μL (total of ~10³ fungal spores) on the plate previously spread with ~1.6e⁷ cells/mL of the bacteria. Studies on viable corn seed (inbred maize line Va35) were conducted by pooling three sets of 10 surface-sterilized seeds per experimental group and treating as follows: (i) uninoculated and untreated, (ii) A. flavus inoculated but untreated, and (iii) inoculated and treated with Vg. For the treatment group, 1.6 × 10⁷ Vg cells suspended in 1 mL YES media were applied to the seeds and semi-dried prior to the fungal spore inoculation. All groups were incubated at 30°C in the dark for 5 days prior to aflatoxin analysis.

For light imaging experiments, A. flavus spores (~10² spores) were inoculated in 20 μL of YES growth medium in a glass-bottom dish with or without Vg added to the growth medium. A widefield microscope (Finite Optical System, Fischer Scientific) was used for bright-field microscopy. A. flavus colonies growing on PDA with or without Vg treatment were visualized by stereomicroscopy (EZ4 W, Leica microsystems). Mycelial pellets harvested at 40 h post-inoculation were used for transmission electron microscopy (TEM) studies. For TEM, the pellets were fixed using 3% glutaraldehyde and 2% osmium tetroxide, dehydrated in an ethyl alcohol series, and made into resin blocks. The resin blocks were trimmed and cross-sectioned into approx. 80 nm thick sections using a diamond knife (Micro Star Technology Inc., Huntsville, Texas) on a Sorvall Porter-Blum MT2-B Ultra-Microtome. Sections were loaded onto copper grids and later stained with lead citrate and 5% uranyl acetate. The stained grids were then imaged using a Hitachi H8000 TEM with pressure maintained at 10⁻⁷ Torr and an accelerating voltage maintained at 200 kV.

A total of 250 mL of YES agar was divided into five 50 mL aliquots and autoclaved. Pure prodigiosin (Cayman Chemical Co., Ann Arbor, MI) was added from 1 mg/mL stock to every 50 mL of tempered YES agar medium to yield final concentrations of 0, 0.5, 1.0, 2.0, and 4.0 μg/mL prodigiosin. Ten milliliter aliquots of the prodigiosin supplemented YES agar medium was added to five 60 × 15 mm Petri plates for each concentration of prodigiosin used. The dishes were center-point inoculated with 2 μL of a 1e⁷ solution of AF70 spores and placed at 30°C for 3 days in the dark. After the incubation period, colony diameter was measured and then agar was excised and extracted for determination of aflatoxin levels.

The influence of Vg treatment on the hyphal fusion process was investigated using an established assay described previously (Zhao et al., 2017). Briefly, equal concentrations of the A. flavus pyrG auxotrophic strain (TJES 19.1) and argB auxotrophic strain (TJES 20.1) were mixed and spotted onto GMM amended with arginine (1 g/L) and uracil (5 mM)/uridine (5 mM) with or without 1.6e⁷ Vg. After incubation at 29°C for 5 days, the newly formed conidia were collected in Phosphate Buffered Saline, and 10⁵ conidia were spread onto the surface of GMM + 0.25% Triton X-100 agar plates (the 0.25% Triton X-100 restricts colony diameter to help with precise colony counts). Heterokaryotic colonies were counted after incubation for 3 days at 29°C. A cross of the hyphal anastomosis mutant, ΔhamI -TXZ9.16 with TJES 20.1, was used as a negative control.

Colony diameters were measured by assessing the mean distance from the outer edge of the growing colony to the center of the colony following growth on PDA for 5 days at 30°C in the dark. For comparing the extent of hyphal branching between the Vg-treated and untreated colonies, the colonies grown on PDA were observed under a stereomicroscope (EZ 4w, Leica Microsystems) and a scoring method used to quantify the visible differences in hyphal branching under the microscope. Twenty investigators were asked to assess a control (untreated) A. flavus colony and a Vg-treated colony. In a blinded fashion, they were asked to assign a score of 0–10, with 0 representing no branching and 10 representing the extent of branching observed in control.

Two polyene endocytosis inhibitors, filipin and natamycin (Sigma Chem. Co, St. Louis), were used to study their effect on Vg-PG uptake and aflatoxin production. Stock solutions (1 mmol/L) of the inhibitors were made using DMSO. We used filipin at 3 μmol/L and natamycin at 6 μmol/L for treatments. These doses corresponded to the minimal inhibitory concentrations (MICs) previously observed for the disruption of Aspergillus endosomal development (Van Leeuwen et al., 2009).

Since mycelial uptake of Vg-PG correlates with the transition from normally non-pigmented fungal hyphae to a reddish color upon complete uptake at 40 h, the effect of the inhibitors on Vg-PG uptake was assessed by visual monitoring of the mycelia and by comparing its color to the control mycelia. A less intense red color corresponded to less uptake. To determine whether the inhibition of Vg-prodigiosin uptake was statistically significant, we developed a blinded visual scoring method to conduct a quantitative comparison of Vg-prodigiosin uptake in the presence and absence of endocytosis inhibitors. A total of 20 separate examiners, who had no prior knowledge of the samples, were asked to provide a color intensity score to 40 h mycelial pellets representing untreated control (displaying noninhibited Vg-prodigiosin uptake) and the natamycin and filipin-treated pellets (displaying inhibited Vg-prodigiosin uptake). Examiners were asked to provide a color score on a scale of 0 to 10, where 0 indicated no red color observed in mycelial pellets, as seen in 40 h untreated pellets (negative control), and 10 indicated entirely red mycelial pellets, corresponding to uninhibited Vg-prodigiosin uptake (positive control). The choice of 40 h old pellets for color scoring was based on the rationale that mycelia show the highest color intensity (corresponding to an almost complete Vg-PG uptake) at this time point under our culture conditions. A higher red color score, therefore, represented higher Vg-PG uptake.

All statistical tests were performed using GraphPad Prism Software (GraphPad, La Jolla, CA, USA). Statistical analysis was carried out using one-way ANOVA followed by Tukey’s multiple comparison test. Statistical significance of difference was set at p < 0.05.

Fungal dry weight was measured as described previously (Kenne et al., 2018) by collecting mycelia from the growth medium and then drying the mycelial mass in an oven at 80°C for 6 h. Weight difference before and after drying was recorded as the dry weight.

Aflatoxin was extracted from liquid and solid cultures using a standard chloroform-methanol extraction protocol described previously (Kenne et al., 2018). In addition, as described previously (Gummadidala et al., 2019), aflatoxin was quantified in the extracts using ELISA (with limit of detection at 2.5 ppb) and HPLC (with limit of detection at 0.1 ppb).

Two doses of Vg were used to treat A. flavus growing in liquid and solid growth media. Results showed that the administration of 1.6e⁷ bacterial cells resulted in an almost complete inhibition (>99%) of aflatoxin production both in liquid (Figure 1A) and solid (Figure 1B) growth media. The lower dose (0.4e⁷ cells) resulted in less inhibition (~50% reduction) of aflatoxin production. Due to the more significant observed reduction in aflatoxin production, 1.6e⁷ Vg cells were used for all subsequent experiments. No significant effects on total fungal biomass as measured by dry weight was observed upon Vg treatments (data not shown). Studies were also performed to determine if viable Vg was essential for Vg-dependent aflatoxin reduction. As shown in Supplementary Figure S1, treatment using heat-treated, non-viable Vg resulted in a similar reduction (>95%) of aflatoxin production as treatment with viable Vg. In a separate experiment, Vg, when applied to A. flavus infected corn seeds, resulted in a similar reduction (>90%) in aflatoxin (Supplementary Figure S2), which supports the idea that Vg could restrict aflatoxin contamination in crops in a manner similar to our batch experiments using artificial media.

During Vg treatments, we consistently observed that by the 40-h time-point, the fungal mycelia turned red (a color similar to Vg-prodigiosin that we denote here as Vg-PG; see Figure 2A). Bright-field microscopy also indicated the presence of Vg-PG at discrete locations along the hyphae (see hyphal image after Vg treatment in Figure 2A), suggesting a hyphal uptake of Vg or a Vg-PG enriched fraction. TEM images of hyphae of Vg-treated and untreated A. flavus cultures showed no trace of intact rod-shaped Vibrio cells inside Vg- treated hyphae (Supplementary Figure S3), confirming that intact uptake of Vg did not occur. During the Vg treatment we also observed a steady decrease of Vg as reflected by the OD600 readings of the treatment broth until it reached non-detectable levels by 40 h (Supplementary Figure S4), suggesting a possibility of bacteria lysis in the growth medium during co-culture, that could explain why no signs of intact Vg cells were observed in TEM images.

Next, we proceeded to examine the influence of prodigiosin on aflatoxin production by the fungus. First, we treated A. flavus with different bacteria (including Vg) to compare the effects of the treatments on aflatoxin production. The non-Vg bacteria included V. harveyi and V. fischeri (two Vibrio spp. that do not produce prodigiosin), S. marcescens (a non-Vibrio prodigiosin producer), E. coli (as a representative Gram-negative bacterium) and S. aureus (a representative Gram-positive bacterium). As shown in Supplementary Figure S5, significant aflatoxin reduction compared to untreated control was observed only upon treatment with the prodigiosin producers, Vg and S. marcescens. To determine whether prodigiosin was the reason for the specificity of aflatoxin inhibition we then tested the ability of pure prodigiosin to inhibit aflatoxin production. As shown in Figure 2B, aflatoxin was significantly inhibited at all levels of prodigiosin with a ~ 88–97% decrease corresponding to 0.5–4.0 ug/mL pure prodigiosin. In addition, some inhibition of fungal growth as measured by colony diameter was observed at all levels but not enough to account for the observed decreases in aflatoxin production. Collectively, the data imply that prodigiosin is one critical bioactive component linked with Vg-dependent aflatoxin inhibition.

Based on the observations of Vg-PG localized at discrete locations within fungal hyphae during Vg treatment (in Figure 2A), further investigation was conducted to test if Vg treatment influenced hyphal fusion. Hyphal fusion is a fundamental biological process in filamentous fungi coordinated by fungal endocytic machinery and endosomal transport (Riquelme et al., 2018; Fischer and Glass, 2019). Hence we reasoned that if Vg-PG internalized in endosomes and thereby disrupted endosomal functions, hyphal fusion would be disrupted in a Vg-dependent manner. Results from the hyphal fusion assay (Figure 3) indicated that the process of heterokaryon formation from crosses between TJES 19.1 and TJES 20.1 was almost completely inhibited (by >95%) in the presence of Vg. This observation was similar to the negative control (crosses between the hyphal anastomosis mutant, ΔhamI, and TJES 20.1). In addition, Vg treatment’s influence on hyphal branching within the colony (an endosome-dependent process) was examined. As shown in Supplementary Figure S6, hyphal branching were significantly reduced upon Vg treatment.

Next, we tested if inhibiting endocytosis uptake of Vg-PG could increase aflatoxin production. Two established endocytosis inhibitors, filipin, and natamycin, were used for this study. As shown in Figure 4, the visual observations corresponded to the color scores. Mycelial color intensity corresponding to Vg-PG uptake dropped in the presence of the inhibitors. The drop in the color score for both inhibitor treatments was statistically significant. Aflatoxin levels correlated well with the levels of color scores. While both inhibitor treatments significantly increased aflatoxin production compared to the untreated control (A. flavus + Vg), natamycin appeared to cause a higher increase in aflatoxin levels compared to the filipin-treated samples.