Aspergillus parasiticus BN9∆ku70 [13] and RH∆ku70 [14] and A. flavus CA14PTs∆pyrG [15], a ∆ku70 strain sensitive to pyrithiamine, were the recipient strains used in the msnA gene knockout experiments. A. parasiticus BN9∆ku70 and A. flavus CA14PTs∆pyrG are aflatoxigenic and produce abundant conidia and a few sclerotia. RH∆ku70 accumulates O-methylsterigmatocystin (OMST) as the end product; it produces abundant sclerotia and conidia when grown in the dark. OMST-accumulating isolates have been found to account for about 2.6% of an A. parasiticus population in a southwestern Georgia peanut field [11]. Potato Dextrose Agar (PDA; EMD Chemicals Inc., Darmstadt, Germany) was used for fungal growth and production of conidia and sclerotia for enumeration. The mixed cereal agar (MCA, 5% Gerber^® Mixed Grain Cereal, 1.5% agar) [16] was also used to promote sclerotial production.

TheEST of the A. flavus msnA gene (NAFAE55TH) was identified initially from the Aspergillus flavus Gene Index database at The Institute for Genomic Research (TIGR) based on S. cerevisiaeMSN2 and its homologue in A. nidulans (AN1652.2). The complete A. flavus msnA gene was subsequently obtained from the Aspergillus Comparative Database at Broad Institute (http://www.broadinstitute.org/annotation/genome/aspergillus_group/MultiHome.html). Restriction analysis of A. flavus msnA and flanking regions was carried out using the DNAMAN software (Lynnon Soft, Vandreuil, Quebec, Canada). The msnA disruption vector was constructed as follows: a 1.4-kb msnA 5’ and coding region and a 0.7-kb coding region near the 3’ end were amplified by PCR using primers msn5K: CTGTCTCCCGGTACCTTTGATCG and msn5P: GAGTATGCGCTGCAGCGCTGTCTC, and msn3P: GAGACAGCGCTGCAGCGCATACTC and msn3H: CGTGGGAAGCTTCATAGAGCAC, respectively. The PCR fragments after digestion were cloned sequentially into corresponding sites in pUC19. The 2.0-kb A. oryzaeptr selectable marker amplified from pPTR1 [17] was cloned into the PstI site of the above construct. This disruption vector construct, msnDV, was linearized by HindIII and KpnI to release the portion of pUC19 prior to fungal transformation.

Preparation of protoplasts, ptr-based fungal transformation and selection of mutants were performed as previously described for A. parasiticus and A. flavus [14]. Homologous recombination is the primary event in fungi with the ku70-deficient background. The msnA gene disruption was confirmed by PCR with location-specific primers based on the expected genomic patterns generated by homologous recombination in the ∆ku70 genetic background. The primers used were P1: GACACAAGGTTCGTCGGTGACT and P2: GGTACTCGCGTCGCGATTA. PCR was carried out under the following conditions in a Perkin Elmer GeneAmp PCR System 2400. Twenty-five pmol of each primer and 10 ng genomic DNA were added to 25 μL Platinum Blue PCR Supermix (Invitrogen, Carlsbad, California, USA) and subject to 30 cycles consisting of denaturation at 94 °C for 30 s, annealing at 55 °C for 30 s and extension at 72 °C for 5 min.

A genomic DNA fragment containing the full-length msnA gene including a 0.58 kb 5’UTR and a 0.24 kb 3’UTR was amplified by PCR using the Accuprime supermix (Invitrogen) with primers msn-E-pyrG, ATAGAATTCCCCGCGACTGTCCATTAGTC and msn-B-pyrG, ATAGGATCCTTTGTGAAGACCATGT. The PCR product was first cloned into the EcoR1 and BamH1 sites of pPG28 [18]. The A. nidulans autonomously-replicating sequence in the 5.2 kb HindIII fragment from pHELP1 [19] was cloned into the resulting construct. The circular plasmid was transformed into an A. flavus CA14∆msnA strain by the pyrG-based transformation protocol [15].

The ∆msnA and the parental strains were point inoculated on PDA plates (100 × 15 mm) in triplicate and grown for five days at 30 °C for the determination of vegetative growth based on the colony diameter. For enumeration of conidia, two culture plugs were cored with Transfertube^® (Spectrum, Houston, Texas, USA) from each of the three seven-day-old replicate PDA plates. The plugs were placed in a microfuge tube containing 0.5 mL ethanol to kill conidia and then vortexed 2 min with a Disruption Genie apparatus (ZYMO RESEARCH, Orange County, California, USA). Each sample was diluted 100-fold with a 0.01% Triton solution, and conidia were counted two to four times using a hemocytometer. The calculated numbers were the total conidia from the two agar plugs. Strains were cultured on PDA and MCA plates at 30 °C for a week for sclerotial formation.

The ∆msnA strains produced an unknown diffusible pigment on MCA plates not observed on PDA plates. The water-soluble orange pigment was isolated from 25 MCA plates inoculated with A. parasiticus BN9∆msnA. The frozen and thawed cultures were filtered through No. 4 paper (Whatman, Piscataway, New Jersey, USA) and the filtrate passed through a 33mm Millex GP 0.22 µ syringe filter (Millipore, Billerica, Massachusetts, USA). The filtrate was stirred overnight with Amberlite XAD-2 resin (Rohm and Haas, Philadelphia, Pennsylvania, USA). The pigment was eluted from the XAD-2 resin with MeOH, and the solvent was removed under reduced pressure. The remaining solids were dissolved in water and applied to a Sephadex G-25 column (GE Healthcare, Piscataway, New Jersey, USA). The column was eluted with water, and the orange band was collected and lyophilized yielding 44 mg of the pigment. The absorbance spectrum between 350 and 700 nm for the pigment dissolved in water was recorded on an Agilent 8453 photodiode array UV-Vis spectrophotometer (Agilent, Santa Clara, California, USA). Fourier Transform Infrared (FTIR) analysis, which provides information about the chemical bonding and the molecular structure of a compound, was performed. The infrared spectrum of the mixture of 0.5% of the orange pigment in KBr was obtained by diffuse reflectance on a Nicolet 6700 FTIR (Thermo Scientific, Waltham, Massachusetts, USA).

The amounts of aflatoxins and kojic acid produced on MCA plates in triplicate by four- and eight-day-old cultures of wild type and ∆msn strains of A. parasiticus BN9 (only kojic acid was determined for RH), and A. flavus CA14 were determined. For each sample, the entire contents of the Petri dish (100 × 15 mm, 25 mL per plate) were added to a Waring MC-2 blender cup and blended for 30 s with 20 mL hot water. The mixture was filtered through Whatman No. 4 paper and a 33 mm Millipore Millex GP 0.22μm syringe filter. The filtered extracts were analyzed for aflatoxins and kojic acid on a HPLC system (Agilent) consisting of a quaternary pump, autosampler, photodiode array detector, and fluorescence detector. The analyses were performed on a column of Intersil ODS-3, 5 μ, 4.6 × 250 mm (GL Sciences, Torrance, California, USA) at a flow rate of 1.0 mL/min. The injection volume was 20 μL. The mobile phase for aflatoxins was H₂O:CH₃CN:MeOH (45:25:30) and for kojic acid MeOH:0.1% H₃PO₄ (25:75), isocratic. Detection of aflatoxins was based on fluorescence at 365 nm excitation and 455 nm emission with “PHRED” postcolumn photochemical derivatization (Aura Industries, New York, New York, USA). Detection of kojic acid was based on UV absorption at 265 nm. The retention times of the standards for aflatoxins B₁, B₂, G₁ and G₂ were 10.6 min, 9.4 min, 8.7 min and 7.8 min, respectively, and for kojic acid the retention time was 5.4 min.

An aliquot of spore suspension was added to each Petri dish plate (100 × 15 mm) containing 30 mL potato dextrose broth (PDB, Merck, Darmstadt, Germany) to give a final concentration of 3 × 10⁵ spores per milliliter. Four dishes were used for each strain. Stationary cultures were incubated at 30 °C in the dark for 4 days. Harvested mycelium were pooled and rinsed with sterilized distilled water and pulverized to a fine powder with a mortar and pestle in the presence of liquid nitrogen. Total RNA was prepared using TRIzol^® reagent (Invitrogen) and treated with amplification grade DNase I followed by purification with an RNeasy Plant Mini kit (Qiagen, Valencia, California, USA).

Fluorescent dye Cy3 and Cy5 labeled probes were prepared using the indirect labeling method of aminoallyl-cRNA according to the protocol provided by TIGR. A total of 6 μg of aminoallyl-cRNA were needed in each probe labeling. The aminoallyl-cRNA was synthesized and amplified using an Amino Allyl MessageAmp™ II aRNA Amplification kit (Ambion, Austin, Texas, USA). Mono-reactive dyes Cy3 and Cy5 (Amersham, Piscataway, New Jersey, USA) were coupled respectively to aminoallyl-cRNA from wild type and mutant. The unincorporated free dyes were removed using the RNeasy MinElute cleanup kit (Qiagen).

The microarray used was Aspergillus flavus NRRL3357 27.6k oligo array [20]. A direct comparison design was applied, including Mutant/Wild type. Four technical replicates were used, including two dye swaps to compensate for cyanine dye effects. Following hybridization and washing according to the TIGR protocol, the microarray slides were scanned by a Genepix 4000B (Molecular Devices, Sunnyvale, California, USA), and the images were analyzed using GENEPIX 6.0 software.

Microarray data were normalized and analyzed using GeneSpring GX 10.0 software (Silicon Genetics, Redwood City, California, USA). Two criteria were used for selecting positive spots, (Signal-Background) mean > 500 unit as expression intensity filter, and at least two of the four replicates showing positive. These filters were imposed to remove genes with very minor differential expression or genes with little evidence for expression. Data normalization was performed using local regression LOWESS (locally weighted scatterplot smoothing). Differentially expressed genes were identified by performing a one-way ANOVA on the normalized data using a T test with no assumption of equal variance. The cutoff criteria in fold change analysis were set as P < 0.05 in significant difference and fold change >2.

Quantitative real-time PCR (qPCR) was carried out in a 20-μL reaction volume with the SYBR Green Master Mix (Applied Biosystems, Foster City, California, USA) in a StepOne^TM thermal cycler (Applied Biosystems). PCR conditions were as follows: an initial step of 95 °C for 10 min and 40 cycles with each cycle consisting of 95 °C for 15 s and 60 °C for 1 min. The primers sets used were designed by the PrimerExpress 3.0 software following the guidelines specified by Applied Biosystems. The specificity of the PCR products was confirmed by the melt curve analysis. The relative expression levels of the genes examined were determined by the relative standard curve method, in which standard curves based on 5-point, 10-fold serial dilutions were constructed for the endogenous (18S) and the target gene in each experiment using genomic DNA as the template.

ROS generation was assessed using the substrate 2’,7’-dichlorofluorescein diacetate (DCFH-DA; Sigma D6883). Oxidation of non-fluorescent DCFH-DA by ROS, such as H₂O₂ and the hydroxyl radical, yields the fluorescent product dichlorofluorescin (DCF). DCF fluorescence spectra is usually measured using excitation/emission wavelength of 490/525 nm. In this study, we used the OneStepPlus qPCR instrument (Applied Biosystems) that is able to measure dye fluorescence with the program set for the detection of SYBR^®-DNA complex at excitation/emission wavelength of 488/522 nm. DCFH-DA stock solution (100×, 1 mM in dimethyl formamide) was prepared. Four agar plugs were cored from each PDA culture plate grown for 7 days at 30 °C. They were placed into a microfuge tube, and 1mL freshly made 1× PBS containing DCFH-DA at a final concentration of 1 μM was added. The reaction was allowed to continue at 37 °C for specified time periods (1, 5 or 24 hr). At the end of each time period, three replicates of 30 μL from each sample were loaded into the optical 48-well reaction plate. The program was set to hold at 37 °C and fluorescence measurements were taken for 6 cycles at 10 min intervals.

Disruption of msnA in A. parasiticus BN9 and RH, and A. flavus CA14 was confirmed by PCR based on expected genomic patterns (Figure S1). Colonies of the ∆msnA strains of A. parasiticus and A. flavus exhibited a restricted, densely-packed appearance on PDA plates. The radial colony growth of the ∆msnA strains was 2 to 3-fold less than that of the parental strains (Figure 1A). Compared to the parental strains the ∆msnA strains produced more conidia as estimated from an identical area (Figure 1B). The increase in conidiation of A. parasiticus BN9∆msnA (36%) was less than that of A. parasiticus RH∆msnA (>600%) and A. flavus CA14∆msnA (84%). Although A. parasiticus RH is a strain that produces many sclerotia, RH∆msnA was unable to produce any on PDA; BN9∆msnA and CA14∆msnA were also unable to produce sclerotia on the same medium.

To further examine whether sclerotial formation of the three ∆msnA strains was abolished, we grew them on mixed cereal agar plates. All strains did not produce sclerotia on this medium either. They, however, produced a diffusible orange-colored substance that increased in intensity after prolonged incubation (Figure 2). In contrast, the orange-colored pigment was barely produced by the wild-type strains after the same 7-day period of growth.

Based on the color we speculated that the orange pigment might be a complex derived from kojic acid, a metabolite commonly produced by some strains of A. flavus and closely related A. oryzae. The UV-Vis spectrum of the pigment purified from the A. parasiticus BN9∆msnA culture matched with the published spectrum of the complex of 3 kojic acid moieties per Fe(III) [21] and had a peak absorbance at 398 nm and a broad shoulder at 460 nm (Figure S2). The kojic acid in the complex was confirmed by FTIR analysis using kojic acid as the reference. The separate FTIR spectra of kojic acid and the pigment are shown in Figure 3A. The overlaid image indicates that the two spectra are nearly identical (Figure 3B); two extra peaks in the carbonyl region (1,500 to 1,700 cm⁻¹) appear in the pigment and another observed difference is at about 2,550 cm⁻¹.

To determine how production of aflatoxins and kojic acid was affected by msnA disruption, we carried out a time-course quantitative determination. The normalized data (to growth area, Table 1) showed that the production of total aflatoxins and kojic acid increased from day 4 to day 8 in the parasiticus BN9∆msnA strain. The ∆msnA strain accumulated 50% more aflatoxins and 20-fold more kojic acid than the parental strain at day 8. The A. parasiticus strain RH is a blocked strain that accumulates O-methylsterigmatocystin as the end product. Although it does not produce aflatoxins, the derived RH∆msnA strain produced 10-fold more kojic acid. The A. flavus CA14∆msnA strain produced only small amounts of aflatoxins, but it produced four-fold more kojic acid than the parental strain.

Microarray assays identified many differentially expressed genes in the ∆msnA strains of A. parasiticus BN9 and A. flavus CA14 (Table S1 and Table S2). For A. parasiticus, approximately 85% of the genes (0.005% saturation and 500 cutoff) were down-regulated; only 12 genes were up-regulated (°2-fold), and they included those oxidative stress defense genes encoding superoxide dismutase, catalase, and cytochrome c peroxidase. The up-regulation of these genes was confirmed by qPCR (Table 2). For A. flavus, approximately two thirds of the genes profiled were down-regulated; the genes up-regulated were diverse including many of those encoding hypothetical proteins, and only one up-regulated catalase A gene for the A. flavus ∆msnA strain was found.

The recent generation of the A. flavus CA14-derived double mutant makes it possible to retransform a knockout strain, such as the ∆msnA strain, using a second selectable marker along with an intact gene to confirm gene function [15]. By this approach, the defects in colony growth were remediated in the A. flavus CA14∆msnA strain after the intact msnA genomic DNA was reintroduced (Figure S3).

ROS production by the wild-type strains, ∆msnA strains of A. parasiticus and A. flavus, and by the A. flavus msnA addback strain was determined to see whether ROS levels correlated with observed changes in growth, development and secondary metabolite production. ROS production was found to fluctuate in the first 5 hr incubation assays (data not shown). Only after 24 hr incubation a more distinct pattern emerged; an increase in ROS production was seen in all ∆msnA strains (Figure 4). ROS production by the A. flavus msnA addback strains was decreased to levels comparable to that of the wild-type A. flavus.