All experiments in this study were conducted in accordance with the general safety and biohazard protocols approved by the Envixronmental Health and Safety, University of South Carolina.

Aspergillus parasiticus, SU-1 (ATCC 56775), a wild-type aflatoxin producer. The strain was grown on 100 mm Petri dishes containing potato dextrose agar for 2 weeks. Fresh spores collected from these colonies were used for all the experiments in this study that involved the use of SU-1. In these experiments the fungus was grown in aflatoxin-inducing yeast-extract-sucrose (YES); a rich growth medium (containing 2% w/v yeast extract, 6% w/v sucrose, pH 5.8), by inoculation of 10⁴ spores per mL of liquid medium and incubated in the dark (29°C; shaking at 150 rpm). A bacterial seawater isolate of V. gazogenes, having a 98% similarity of its 16S rRNA gene to ATCC 43942 (Farmer et al., 1988), was grown in Difco Marine Broth 2216 (BD Biosciences, Sparks, MD, USA) at 28°C in a shaking incubator (190 rpm). The bacterial isolate is currently being submitted to ATCC and NCBI.

All fungal growth quantifications were performed using dry weight measurements. Briefly, the mycelia were filtered out of the growth media using a miracloth (Millipore, Billerica, MA, USA) and dried at 75°C for 6 h and the final weight was recorded. All Vibrio growth measurements were performed using absorbance readings of growth media at 600 nm.

Aflatoxin B₁ was commercially obtained (Sigma). Three different doses (0.1, 0.2, or 0.3 μg/mL) of aflatoxin B₁ were added to the Vibrio growth medium at the start of the culture. In the control flask only the vehicle (70% Methanol) was added. To extract the metabolites from V. gazogenes the cells were first harvested by centrifugation and extracted with 60 mL acetone. A portion of the filtrate was concentrated by evaporation under N₂ gas. The concentrate was loaded onto a silica gel column (1.2 cm × 15 cm) and eluted with dichloromethane: methanol (80:1.5). The fractions were then purified on a silica gel column using chloroform and methanol (50:2). After purification the fractions were concentrated by evaporation under N₂ gas and re-suspended in 1 mL methanol for spectral analysis.

Comparative semi-quantitative estimations of accumulation of aflatoxin in growth medium was performed using thin layer chromatography (TLC) of the growth medium as described previously (Banerjee et al., 2014).

Isolation of total RNA from fungal cells exposed to aflatoxin response metabolites from V. gazogenes was performed using 30 h old cultures. This is a time point that corresponds to the activation of secondary metabolism (hence the expression of aflatoxin genes in A. parasiticus) under the growth conditions adopted in this study (Roze et al., 2007a). Purification of total RNA and preparation of complementary DNA was performed as described previously (Chanda et al., 2009). Transcript levels were quantified by performing quantitative real-time PCR assays using SsoAdvanced universal SYBR Green supermix (BioRad Laboratories, Hercules, CA, USA) and gene-specific forward and reverse primers (Table 1) that were designed using Primer3 online software¹. Reactions were performed in a CFX96 thermal cycler (Bio-Rad Laboratories, Hercules, CA, USA). As described for previous gene expression studies in A. parasiticus (Roze et al., 2007a; Chanda et al., 2009), expression value of each gene was obtained from the threshold cycle values were normalized against β-tubulin (the house keeping gene) in each sample. All RT-PCRs were performed in triplicates for each gene per sample. Data analyses were performed using CFX Manager software (Bio-Rad Laboratories).

All statistical tests were performed using GraphPad Prism Software (GraphPad, La Jolla, CA, USA). Statistical analyses to determine for statistical significance of differences between control versus experimental groups were determined using one-way ANOVA (with sample size 3). An unpaired t-test was used to determine the gene expression effects of ARMs on A. parasiticus compared to the untreated samples. Significance was set at p < 0.05.

As a first step in understanding how V. gazogenes, responds to aflatoxin B₁, we investigated the effect of three different doses of aflatoxin B₁ on the growth of V. gazogenes. The doses, 10, 30, and 50 ppb were either below, approximately equal to or fivefold higher than the highest-allowed aflatoxin level (20 ppb) in food and feed (Mazumder and Sasmal, 2001; CAST, 2003; Liu and Wu, 2010). Time-course absorbance readings were recorded to compare the growth rates of V. gazogenes, in presence of aflatoxin B₁, with untreated-controls. As shown in Figure 1, none of the aflatoxin B₁ doses demonstrated any significant effect on the growth of V. gazogenes.

Next, we investigated the effect of aflatoxin B₁ exposures on the production of prodigiosins by V. gazogenes. The prodigiosin fraction was obtained from cells (either untreated control cells or cells exposed to aflatoxin B₁) using our optimized laboratory protocol (see Materials and Methods). Since, the prodigiosins exhibit an absorbance peak at 530 nm (Figure 2A), this wavelength was used to compare prodigiosin levels between experimental treatments and controls at three different time-points of growth (12, 18, and 42 h). Our results (Figure 2B) demonstrated that although cells exposed to aflatoxin B₁ showed a minor increase in absorbance values compared to the untreated samples, the difference was not statistically significant.

While growth and prodigiosin production by V. gazogenes was not affected in presence of aflatoxin B₁, we observed that exposure to aflatoxin B₁ resulted in a distinct alteration of color in the growth medium (Figure 3A) suggesting the presence of a different metabolite compared to untreated cells. Based on the ‘blue-shift’ in color of the growth medium (bright red to orange) upon addition of aflatoxin B₁, we hypothesized that the bacterium synthesizes an additional metabolite fraction under these conditions with a corresponding absorbance lower than that of the prodigiosin fraction. To test this, we performed UV-Vis spectral analysis on the metabolite fractions of aflatoxin B₁-treated samples. The Vibrio metabolite fractions obtained from aflatoxin B₁ treated samples demonstrated a new absorbance peak at 240 nm (Figure 3B). This suggested that aflatoxin B₁ exposure affects the cellular metabolism of V. gazogenes resulting in a different metabolite profile, compared to the untreated control. Here, we denote this additional metabolite fraction in response to aflatoxin B₁ exposure as ‘aflatoxin response metabolites (ARMs).’

Next we proceeded to investigate whether ARMs affect the aflatoxin synthesis in the model aflatoxin B₁ producer, A. parasticus. The activation of ARM production by Vibrio occurred upon addition of aflatoxin B₁ to their growth medium. Therefore, we envisioned this alteration of metabolite profiles as a defensive response from Vibrio cells. We hypothesized that ARMs will have a specific inhibitory effect on aflatoxin synthesis in the producer cells. To test this we studied the growth and aflatoxin production by A. parasiticus in presence of two different doses of the ARMs metabolite fraction (1 and 2 μg per mL of growth medium); the doses were chosen arbitrarily. To compare the levels of aflatoxin biosynthesis in A. parasiticus exposed to ARMs exposed with the untreated cells, we adopted a semi-quantitative approach in which we compared the intensities of aflatoxin ₁ and aflatoxin B₂ bands on the TLC plates (see Materials and Methods). As predicted, our TLC results generated from 40 h cultures of A. parasiticus, demonstrated that ARMs applied at the concentration of 2 μg per mL of growth medium inhibited both aflatoxin B₁ and aflatoxin B₂ by approximately twofold (Figure 4A). Since, the drop in aflatoxin synthesis could also have resulted from the inhibition of A. parasiticus growth, we next compared the dry-weights of the A. parasiticus mycelia exposed to 1 and 2 μg per mL of ARMs extract with the untreated control mycelia. As shown in Figure 4B, addition of ARMs to the growth medium did not result any significant change in A. parasiticus dry weight, suggesting that inhibition of aflatoxin synthesis in A. parasiticus by ARMs was a direct effect and not a growth dependent effect.

Our UV-Vis spectral analysis suggested that ARMs were synthesized by V. gazogenes upon exposure to aflatoxin. We then proceeded to confirm that this fraction (peak absorbance at 240 nm) was composed of metabolites of molecular masses that are different from the Vibrio’s prodigiosin fraction (peak absorbance at 530 nm). As shown in Figure 5, HPLC traces showed that the prodigiosin fraction predominantly demonstrated the expected molecular weight of 324 D, corresponding to the known prodigiosin. The HPLC trace of ARMs on the contrary was clearly different, with a predominantly displayed molecular mass 232 D, which demonstrate that the metabolite fraction of ARMs was chemically different from the Vibrio’s prodigiosin fraction. These results suggest that the differential metabolite profile in response to aflatoxin exposure can occur either due to synthesis of new metabolites by V. gazogenes or due to breakdown of prodigiosins resulting in novel smaller molecules with aflatoxin synthesis inhibitory activity.

The fungal growth and aflatoxin results then prompted us to investigate whether aflatoxin biosynthesis was inhibited at the level of transcript accumulation of aflatoxin genes. To conduct this analysis we performed a quantitative comparison of transcript accumulation of two genes nor-1, and ver-1 that encode two enzymes, Nor-1, Vbs, and Ver-1, respectively, involved in the aflatoxin biosynthetic pathway (Chanda et al., 2009). Activation of these genes in A. parasiticus occurs at 24 h, and transcripts of all aflatoxin enzymes accumulate by 30 h, when the fungus is grown in YES growth medium (Roze et al., 2007a). Hence, we chose to examine the effects of ARMs extract on A. parasiticus at three different time-points, 24, 30 and 40 h, a time-point when aflatoxin is synthesized by the fungus at peak levels (Roze et al., 2007a). In addition to these genes, we also compared the transcript accumulation of the aflatoxin pathway regulator, aflR, at the same time points. As shown in Figure 6, nor-1, ver-1 as well as the aflR genes transcript levels demonstrated ≥5 fold reduction in presence of ARMs extract compared to the vehicle control by 30 h. Hence, our results suggest that ARMs extract reduces aflatoxin synthesis at the level of transcript accumulation.

Since the regulatory network of the aflatoxin biosynthesis pathway is integral to the global network of secondary metabolism in A. parasiticus (Chanda et al., 2009), we also proceeded to investigate whether, ARMs target the global regulation of secondary metabolism. One key global regulatory complex of fungal secondary metabolism is the Velvet complex which consists of proteins VeA, LaeA and other regulators (Bayram et al., 2008). Central to this complex is the cross-talk between the two global regulatory proteins, LaeA, a methyltransferase that is key to the epigenetic regulation of aflatoxin biosynthetic pathway (Bok and Keller, 2004), and VeA, a light responsive regulatory protein that migrates from cytoplasm to the nucleus in absence of light to particpate in the VeA complex (Bayram et al., 2008). In this study we investigated whether ARMs affect transcript accumulation of either laeA or veA genes. To our surprise we found that, while no significant changes occurred in veA transcripts, the laeA transcript accumulation was reduced by ∼2 fold by 30 h and ∼4 fold by 40 h (Figure 7), suggesting that ARMs inhibit aflatoxin biosynthesis at least in part, through inhibition of laeA gene activation.