Phenotypic Differentiation of Two Morphologically Similar Aflatoxin-Producing Fungi from West Africa

Aflatoxins (AF) are hepatocarcinogenic metabolites produced by several Aspergillus species. Crop infection by these species results in aflatoxin contamination of cereals, nuts, and spices. Etiology of aflatoxin contamination is complicated by mixed infections of multiple species with similar morphology and aflatoxin profiles. The current study investigates variation in aflatoxin production between two morphologically similar species that co-exist in West Africa, A. aflatoxiformans and A. minisclerotigenes. Consistent distinctions in aflatoxin production during liquid fermentation were discovered between these species. The two species produced similar concentrations of AFB1 in defined media with either urea or ammonium as the sole nitrogen source. However, production of both AFB1 and AFG1 were inhibited (p < 0.001) for A. aflatoxiformans in a yeast extract medium with sucrose. Although production of AFG1 by both species was similar in urea, A. minisclerotigenes produced greater concentrations of AFG1 in ammonium (p = 0.039). Based on these differences, a reliable and convenient assay for differentiating the two species was designed. This assay will be useful for identifying specific etiologic agents of aflatoxin contamination episodes in West Africa and other regions where the two species are sympatric, especially when phylogenetic analyses based on multiple gene segments are not practical.


Introduction
Aflatoxins (AF) are potent carcinogenic metabolites produced by several Aspergillus species. Aflatoxin-producing aspergilli may infect a wide range of crops including cereals, groundnuts, cottonseed, tree nuts, and spices, and these infections frequently result in aflatoxin contamination of foods and feeds [1][2][3][4][5][6]. Aflatoxins are a serious health and economic threat worldwide [7]. Although Aspergilli produce four major aflatoxins, aflatoxin B 1 , B 2 , G 1 , and G 2 , aflatoxin B 1 is the only mycotoxin listed as a human carcinogen by the International Agency for Research on Cancer [8]. Chronic exposure to aflatoxins results in reduced immunity, growth impairment, and hepatocellular carcinoma in humans and animals [9][10][11]. Consumption of food contaminated with high concentrations of aflatoxins has resulted in severe liver damage and rapid death [12]. Epidemics of acute lethal A. minisclerotigenes. Differences in aflatoxin production on common media were consistent with phylogenetic distinctions. These observations were refined into a simple liquid fermentation assay that can provide reliable differentiation of the two species when access to DNA-based methods for species identification is lacking. The results also suggest differences between the two taxa in the conditions under which aflatoxin contamination may occur.

Aspergillus aflatoxiformans and A. minisclerotigenes Differ in Aflatoxin Production in Yeast Extract Sucrose (YES) Medium
Aflatoxin production by isolates of A. aflatoxiformans (n = 45) and A. minisclerotigenes (n = 44) was initially assessed by growth in YES medium (pH = 6.5) at 31 • C with agitation for seven days. This medium supported production of low concentrations of aflatoxins by A. aflatoxiformans (mean = 1.9 µg total aflatoxin g -1 mycelia; range = 0.02 to 5.65 µg/g), compared to A. minisclerotigenes (mean = 89 µg/g; range = 32 to 460 µg/g). Fourteen isolates from each species were re-evaluated in Adye and Mateles (A&M) medium supplemented with urea as the sole nitrogen source. In contrast to the above results, fungal isolates from both species produced high and similar concentrations of total aflatoxins (p = 0.64) in A&M medium with urea (Mean = 507 µg/g). Taken together, these results indicate that aflatoxin production by A. aflatoxiformans is much more inhibited than by A. minisclerotigenes in YES medium. This is a phenotypic distinction that differentiates these two morphologically similar species.

Aflatoxin Production by A. aflatoxiformans and A. minisclerotigenes in Liquid Fermentations
To further assess differential influences of medium composition on aflatoxin production by A. aflatoxiformans and A. minisclerotigenes, four isolates representative of each species from different geographic regions of Nigeria were assayed in three media, (i) YES, (ii) A&M supplemented with urea, and (iii) A&M containing ammonium with agitation. Each of these media support aflatoxin production by fungi with L-or S-morphology within section Flavi [29,44,45]. The two A&M media with different nitrogen sources have been utilized previously to assess aflatoxigenicity of isolates from soils and crops and to study the etiology of aflatoxin contamination [33,[49][50][51]. Aspergillus aflatoxiformans and A. minisclerotigenes produced similar concentrations of aflatoxins B 1 and G 1 in the A&M medium with urea (p = 0.71 for AFB 1 and p = 0.42 for AFG 1 ), and individual isolates did not differ (Table 1; p > 0.05). Production of aflatoxins B 1 and G 1 in A&M medium containing ammonium differed among isolates (Table 1; p < 0.001). Aspergillus minisclerotigenes produced significantly greater concentrations of aflatoxin G 1 than A. aflatoxiformans (p = 0.039) in A&M with ammonium, while the two species produced similar concentrations of aflatoxin B 1 (Table 1; p = 0.71). Both species produced greater concentrations of aflatoxins in A&M with urea compared to A&M with ammonium. Aspergillus aflatoxiformans produced 7.5 times more aflatoxin B 1 and 31 times more aflatoxin G 1 on average in the medium with urea compared to that with ammonium (Table 1). Aspergillus minisclerotigenes produced 3.7 times more aflatoxin B 1 and nearly 8 times more aflatoxin G 1 in medium supplemented with urea versus ammonium (Table 1). These results are consistent with previous studies that reported increased production of aflatoxins B 1 and G 1 by African S BG isolates in A&M medium containing urea compared to the medium with ammonium as the sole nitrogen source [29]. Production of both aflatoxins B 1 and G 1 by A. aflatoxiformans was significantly lower in YES medium (pH = 4.75) compared to that of A. minisclerotigenes (Table 1; p < 0.001), as also observed during the initial evaluation of these fungi in YES medium (pH = 6.5). Isolates of A. minisclerotigenes produced at least 50 times more aflatoxin B 1 and 25 times more aflatoxin G 1 in YES medium compared to isolates of A. aflatoxiformans (Table 1). Since YES medium contained higher concentrations of sucrose (15%) compared to either of A&M medium (5%), the effect of sucrose concentration on aflatoxin production was tested in YES medium under shaking and stationary conditions. Total aflatoxin production was inhibited by A. aflatoxiformans in YES medium irrespective of sucrose concentration from 5-20% under shaking and stationary conditions at 31 • C (Table S1). Aflatoxin production, pH, and fungal biomass were higher when cultures were stationary versus shaking (Table S1).

pH Modification by A. aflatoxiformans and A. minisclerotigenes
All eight fungal isolates modified the pH of each medium during growth. Medium composition influenced the extent to which pH changed. At the end of fermentation, pH was in order of YES > A&M with urea > A&M with ammonium. Although the final pH of A&M medium with urea differed significantly among individual isolates of both species (Table 1; p < 0.001), the average final pH for A. aflatoxiformans did not differ from that of A. minisclerotigenes (Table 1; p = 0.83). The final pH of YES medium differed both among individual isolates (Table 1; p < 0.001) and between the two species (Table 1; p < 0.002). However, differences were minor, and the two species overlapped making final culture pH not useful for distinguishing the two species. It is noteworthy that aflatoxin production did not depend on the initial pH of YES medium because aflatoxin production in YES fermentations at either pH = 6.5 (initial aflatoxin screen assay) or at pH = 4.75 (Table 1) was much lower for A. aflatoxiformans than A. minisclerotigenes in YES irrespective of the pH. A&M medium with ammonium was the most acidic at the end of the fermentation, and the final pH did not differ among isolates (p = 0.498) or between species (p = 0.41). Similar pH modification was reported for the A&M media in previous studies [29,44]. The A&M medium containing ammonium was more acidic by the end of the fermentation compared to either A&M with urea or YES medium. Previously, influences of fungal growth on culture medium pH have been used to group aflatoxin producing fungi [23,31]. However, these groupings have not consistently reflected DNA-based phylogenetic relationships.

Growth of A. aflatoxiformans and A. minisclerotigenes in Liquid Fermentations
All isolates, irrespective of species, produced the highest biomass in YES medium, and mycelial mass did not differ among isolates (Table 1; p = 0.89) or between species (p = 0.33). Differences were detected in biomass production in A&M medium with ammonium among isolates (p < 0.01) but not between species (p = 0.28). Fungal growth was significantly different among isolates and between species in A&M medium containing urea (p < 0.001); A. minisclerotigenes produced greater biomass in this medium (Table 1; p < 0.001). Concentrations of aflatoxins B 1 and G 1 produced in each medium was independent of fungal growth and biomass production. Although A. aflatoxiformans produced the least concentration of aflatoxins in YES medium, it produced the greatest mycelial mass in this medium. Its mycelial mass was comparable to that of A. minisclerotigenes indicating that decreased production of aflatoxins by A. aflatoxiformans was not due to influences on growth (Table 1). Nutrient influences on aflatoxin biosynthesis [30] may have resulted in the observed differences.

Susceptibility of Maize to Aflatoxin Contamination by A. aflatoxiformans and A. minisclerotigenes
In order to assess the aflatoxin-producing potential of A. aflatoxiformans and A. minisclerotigenes under different environmental conditions, aflatoxin production by the two species was further evaluated on maize at 25 • C, 30 • C, 35 • C, and 40 • C ( Table 2). Maize is an important staple in Nigeria, and average annual production exceeds 10 million metric tons per year [52]. It is estimated that Nigerians may be exposed to~5.0 mg aflatoxin person -1 year -1 through maize consumption [53]. Temperatures were chosen based on climate data from maize growing regions in Nigeria [54,55] and temperatures typically encountered during storage. Overall, aflatoxin production by members of both species was high at 25 • C, 30 • C and 35 • C, and A. aflatoxiformans produced significantly greater concentrations of total aflatoxins compared to A. minisclerotigenes at each of these temperatures (Table 2; p < 0.001). Furthermore, isolates differed in aflatoxin-producing potential at 25 • C, 30 • C and 35 • C (p < 0.001), and some A. minisclerotigenes isolates produced concentrations of aflatoxins similar to those of A. aflatoxiformans ( Table 2). The greatest concentrations of aflatoxins were observed at 30 • C followed by 35 • C and 25 • C for each species. Notably, although fungal isolates produced the least aflatoxin at 25 • C (Range: 13.0-154 µg/g), the observed concentrations are still unacceptable and dangerous for human and animal consumption. Neither A. aflatoxiformans nor A. minisclerotigenes produced detectable concentrations of aflatoxins at 40 • C (LOD = 0.42 µg/g of maize grain). However, A. aflatoxiformans produced on average 5.6 and 11.7 fold more aflatoxins than A. minisclerotigenes at 30 • C and 35 • C, respectively. These temperatures are typical of maize production areas in Nigeria. Due to the ability of A. aflatoxiformans to produce greater concentrations of aflatoxins in maize compared to A. minisclerotigenes, even if the two species are present at similar frequencies, A. aflatoxiformans may be considered the more important causal agent of aflatoxin contamination in Nigeria. Nevertheless, the aflatoxin-producing potential of A. minisclerotigenes is sufficient to render it potentially dangerous in terms of crop contamination. The results on YES medium (

Assay to Differentiate A. aflatoxiformans and A. minisclerotigenes
We tested the utility of YES medium to reliably differentiate A. aflatoxiformans from A. minisclerotigenes by evaluating total aflatoxin production of 11 representative isolates of each species grown in YES medium for 3 days at 31 • C with and without agitation. Different sets of isolates were used in this assay from those used in the previous experiment to validate the ability of the observed aflatoxin production phenotypes to differentiate the two species. Reference isolates of both species were included (NRRL A-11612 of A. aflatoxiformans and NRRL A-11611 of A. minisclerotigenes). To extend the utility of the assay to laboratories where incubators with shakers may not be available, cultures were grown under stationary conditions as well as with shaking incubation. Results were similar to those previously observed, with isolates of A. minisclerotigenes producing greater concentrations of total aflatoxins in YES medium under either shaking (17.0 to 300 µg/g) or stationary (25.4 to 1,052 µg/g) conditions (Table 3). These same isolates were further tested for aflatoxin production in A&M medium with urea as the sole nitrogen source, and the ratio of total aflatoxins in the urea medium to total aflatoxins in YES medium was calculated for each isolate. Overall, aflatoxin concentrations were greater for all isolates when grown under stationary versus shaking conditions irrespective of species and medium ( Table 3). Isolates of A. aflatoxiformans produced at least 122 and 124 times more total aflatoxins in A&M medium with urea versus YES medium, with and without agitation, respectively. However, ratios of aflatoxin production by A. minisclerotigenes in A&M medium with urea versus YES medium were in the range of 0.86-13.3 with agitation and 0.23-11.9 when stationary. Based on these results, fungal isolates can be identified as A. aflatoxiformans when ratios of aflatoxin concentrations in A&M medium with urea versus YES medium are greater than 80, and as A. minisclerotigenes when ratios are less than 80, with or without agitation (Tables 1 and 3). This ratio cutoff point was selected to allow for any potential outlier isolates from each species based on results in Tables 1 and 3. Table 3. Ratios of total aflatoxins produced in the A&M medium with urea to that in YES medium with agitation and under stationary conditions for Aspergillus minisclerotigenes (AM) and A. aflatoxiformans (AA). Total aflatoxins are expressed as µg/g mycelial weight. The mycelia were captured on #4 Whatman filter paper during filtration of the acetone extracted culture. Dry weights were determined after drying in a forced air oven for 48 h at 40 • C.

Conclusions
Aspergillus aflatoxiformans and A. minisclerotigenes have indistinguishable morphologies, overlapping secondary metabolite profiles, and ability to produce high concentrations of aflatoxins in both synthetic media and crops [29,37,41]. Aflatoxin contamination of crops in West Africa has often been attributed to A. aflatoxiformans due to the aflatoxin-producing ability of this species and its frequency of occurrence in agricultural soils and crops [33][34][35]. In contrast, only a single isolate of A. minisclerotigenes was reported from West Africa [31,41] until recently when A. minisclerotigenes was reported in Nigerian chilies at high frequencies (8% of all Aspergillus section Flavi isolates), confirming the co-occurrence of A. aflatoxiformans and A. minisclerotigenes in West Africa [37]. Aspergillus minisclerotigenes was first described a decade ago, and it was reported that only DNA sequences could separate A. minisclerotigenes and A. aflatoxiformans (referred to as A. parvisclerotigenus) [41]. Both A. minisclerotigenes and A. aflatoxiformans can contaminate widely consumed crops such as maize, groundnuts, and chilies with unacceptable concentrations of aflatoxins [37]. Furthermore, both species produced toxic concentrations of aflatoxins at 25 • C, 30 • C, and 35 • C in maize (Table 3), indicating hazardous potential at temperatures that are prevalent both pre-and post-harvest. More than 99% of the human population in several areas of West Africa suffers chronic exposure to aflatoxins due to consumption of contaminated cereals [53,56]. This creates a clear need for reliable detection and identification of etiologic agents of aflatoxin contamination of crops in this region. Attribution of etiologies is complicated by sympatric species, such as A. minisclerotigenes and A. aflatoxiformans, that share morphological characteristics. Such attribution is especially challenging where DNA-based technologies and the necessary technical expertise may not be easily available or accessible. The microbiological assay developed during the current study is based on differential aflatoxin production in YES medium. This assay utilizes simple liquid fermentation to differentiate A. aflatoxiformans from A. minisclerotigenes within 72 h.

Fungal Isolates and Inoculum Preparation
Isolates of A. aflatoxiformans (n = 45) and A. minisclerotigenes (n = 43) previously recovered from dried red chilies produced in Nigeria [37] were included in this study. The fungal inoculum was prepared for each isolate as described previously [44,57]. Briefly, conidial suspensions from water vials were centrally seeded onto 5/2 agar (5% V8 juice, 2% agar, pH = 6.0) and isolates were grown in the dark for 5-7 days at 31 • C. Conidia were swabbed with sterile cotton swabs and transferred into glass vials with Teflon septa containing 10 mL sterile ultrapure water. Conidial concentrations were estimated with a turbidity versus colony forming unit curve [58], and the final concentration of each suspension was adjusted to 10 6 conidia/mL.

Liquid Fermentation Assays and Assessment of Aflatoxin Production
Three different liquid media were used to evaluate aflatoxin production by fungal isolates: Yeast extract and sucrose medium [45] and Adye and Mateles (A&M) medium [43] amended with either 22.5 mM ammonium sulfate ((NH 4 ) 2 SO 4 ) or 22.5 mM urea as the sole nitrogen source. The exact compositions of liquid media were previously reported [44]. Urea was filter sterilized and added aseptically to autoclaved medium, while ammonium sulfate was added prior to autoclaving the medium [44].
Aflatoxin production by A. aflatoxiformans (n = 45) and A. minisclerotigenes (n = 43) was initially evaluated in YES medium (pH = 6.5). Erlenmeyer flasks containing 70 mL of the medium were aseptically inoculated with conidial suspensions (10 6 conidia/mL of the suspension; 100µL/flask) and incubated in the dark for 7 days at 31 • C with agitation. At the end of the fermentation period, aflatoxins were extracted by the addition of 70 mL acetone to each 70 mL fermentation. After acetone addition, cultures were allowed to sit in the dark for an hour with periodic swirling to increase mixing. Acetone extracts were separated from the fungal mycelia by filtering the contents of the fermentation flasks through Whatman no.4 filter paper using vacuum filtration. Mycelia were dried in a forced air oven (40 • C, 48 h) and weighed to determine the total biomass (dry weight). Acetone extracts (4 µL) were directly spotted onto thin layer chromatography (TLC) plates (Silica gel 60, EMD, Darmstadt, Germany) and separated with ether:methanol:water (96:3:1) adjacent to 4 µL of aflatoxin standard containing 1µg B 1 , 1µg G 1 , 0.3 µg B 2 , and 0.3 µg G 2 per mL of benzene:acetonitrile (98:2) (Aflatoxin Mix Kit-M Supelco, Bellefonte, PA). Aflatoxins were measured on TLC plates by scanning fluorescence densitometry under 365 nm UV light (TLC Scanner 3, Camag Scientific Inc., Wilmington, NC, USA). If aflatoxin was not detected, 12 µL of the extract was spotted onto TLC plates and analyzed as described above. Aflatoxin quantities (total µg) on TLC plates were determined by comparing the area under the peaks generated by each sample to the area under the peaks for the corresponding aflatoxin standard (aflatoxin B 1 , B 2 , G 1 , and G 2 ) generated by the TLC scanner. Total µg of aflatoxins detected on the TLC plate were divided by the proportion of the total volume of the original extract spotted on the TLC plate to calculate the µg aflatoxin per fermentation. Total aflatoxin per fermentation was divided by the dry weight of the mycelia to determine the µg aflatoxin produced per µg mycelia in each fermentation. If aflatoxin was not detected from 12 µL, the extract was partitioned twice with dichloromethane and concentrated prior to quantification as previously described [33]. Total µL of aflatoxins were estimated as mentioned above from a concentrated extract volume of 4 mL.
Based on results from initial screening of isolates in YES medium, four representative isolates each of A. aflatoxiformans and A. minisclerotigenes were chosen for the evaluation of aflatoxin production in YES, A&M with ammonium sulfate, and A&M with urea media. Fungal isolates NRRL A-11612 and NRRL A-11611 from Nigerian groundnuts [31] were used as reference isolates of A. aflatoxiformans and A. minisclerotigenes, respectively. The remaining six isolates were chosen to represent isolates recovered from chilies sampled from different locations in Nigeria. Inoculum for each of the eight isolates was prepared as described above. Isolates were inoculated aseptically into each of the three media to a final concentration of 10 6 conidia/mL. Treatments were replicated four times. All media were adjusted to pH = 4.75 before autoclaving. Fermentations were carried out for 7 days in the dark at 31 • C after which pH was measured, and the experiment was terminated by addition of 70 mL acetone (50% acetone vol/vol). Aflatoxins were extracted, concentrated, and quantified as described above.
In order to test the effect of sucrose concentration in YES medium on aflatoxin production by members of A. aflatoxiformans and A. minisclerotigenes, isolates were inoculated into YES medium (pH = 4.75) containing 5%, 10%, 15%, and 20% sucrose. Isolates were replicated three times and incubated with and without agitation for 3 days at 31 • C in the dark. Total aflatoxins, mycelial mass, and pH were measured after incubation as described above.
A microbiological assay utilizing YES and A&M with urea media was designed to differentiate A. aflatoxiformans and A. minisclerotigenes.
Eleven isolates each of A. aflatoxiformans and A. minisclerotigenes were evaluated for aflatoxin production under shaking and stationary conditions for 3 days at 31 • C. Fungal isolates were selected such that isolates were representative of location and year of sampling. Aflatoxin concentrations, biomass production, and pH were estimated at the end of the incubation period. Ratios of aflatoxin concentrations produced in A&M medium with urea to that in YES were calculated.

Aflatoxin Production in Maize Grain
Fungal isolates evaluated for aflatoxin production in liquid fermentations (Table 1) were also assessed for aflatoxin production in maize (Zea mays L.) (Table 2). Healthy, undamaged maize kernels adjusted to 25% moisture were autoclaved in Erlenmeyer flasks (10 g per flask) for 20 min at 121 • C. Aflatoxin production on autoclaved maize is a good predictor of aflatoxin production in viable maize kernels [44]. Maize was inoculated with 100 µl of conidial suspensions (10 6 conidia/mL), adjusted to 30% moisture, and incubated for 7 days at 25 • C, 30 • C, 35 • C and 40 • C in the dark. Each treatment was replicated three times. At the end of the incubation period, maize-fungal cultures were ground in 50 mL, 85% acetone, in a Blender (Seven-Speed laboratory blender, Waring Laboratory, Torrington, CT, USA) at full speed for 30 s. Aflatoxins were extracted and quantified as previously reported [59].

Data Analysis
Aflatoxin concentrations and fungal biomass were expressed in µg/g and g of dried mycelial weight, respectively. Aflatoxins produced by individual isolates, pH at the end of incubation, and fungal biomass were analyzed using Analysis of Variance as implemented in JMP 11.1.1 (SAS Institute, Cary, NC, USA, 2013). Means were separated using Tukey's HSD test (p = 0.05). Differences in mean aflatoxin concentrations, pH, and mycelial mass between species were compared using Student's t-test (p = 0.05). Data were tested for normality and, if required, log transformed before analysis. True means are presented for clarity. All experiments were conducted with a randomized complete block design.