Effect of temperature and water activity on the production of fumonisins by Aspergillus niger and different Fusarium species

Background Fumonisins are economically important mycotoxins which until recently were considered to originate from only a few Fusarium species. However recently a putative fumonisin gene cluster was discovered in two different Aspergillus niger strains followed by detection of an actual fumonisin B2 (FB2) production in four strains of this biotechnologically important workhorse. Results In the present study, a screening of 5 A. niger strains and 25 assumed fumonisin producing Fusarium strains from 6 species, showed that all 5 A. niger strains produced FB2 and 23 of 25 Fusarium produced fumonisin B1 and other isoforms (fumonisin B2 and B3). Five A. niger and five Fusarium spp. were incubated at six different temperatures from 15-42°C on Czapek Yeast Agar +5% salt or Potato Dextrose Agar. A. niger had the highest production of FB2 at 25-30°C whereas Fusarium spp. had the maximal production of FB1 and FB2 at 20-25°C. Addition of 2.5-5% NaCl, or 10-20% sucrose increased the FB2 production of A. niger, whereas addition of glycerol reduced FB2 production. All three water activity lowering solutes reduced the fumonisin production of the Fusarium species. Conclusion The present study shows that the regulation of fumonisin production is very different in A. niger and Fusarium, and that food and feeds preserved by addition of sugar or salts may be good substrates for fumonisin B2 production by A. niger.


Background
The fumonisins were discovered in 1988 and are divided in four series A, B, C, P [1][2][3]with the B 1 (FB 1 ), B 2 (FB 2 ) and B 3 (FB 3 ) as the most abundant naturally occurring homologues [4,5]. They were first isolated from Fusarium verticillioides (= F. moniliforme pro parte [6]) strain MRC 826 by Gelderblom et al. [7]. FB 1 is mainly produced by F. verticillioides and F. proliferatum [8]. However, production of type B fumonisins by other Fusarium spp. has also been reported, e.g. from F. dlaminii, F. napiforme, F. nygamai and F. oxysporum [8][9][10]. Fumonisins are important mycotoxins because they are suspected to cause human and animal toxicoses by the consumption of contaminated cornbased food and feeds [11]. Fumonisins have been shown to induce outbreaks of equine leukoencephalomalacia in horses and pulmonary edema and hydrothorax in pigs [5,12]. The fumonisins are structurally similar to sphingolipids and have shown to inhibit the sphingolipid biosynthesis via the ceramide synthase pathway [13,14]. To avoid possible health risks, the U.S. Food and Drug Administration recommends that corn products should not be used for human consumption when contaminated with more than 2-4 mg/kg total fumonisins (depending on the product) [15]. whereas EEC has a regulatory limit of 0.2-2 mg/kg (depending on the product) [16].
Recently putative homologues to the F. verticillioides fumonisin gene cluster were found in two different Aspergillus niger genomes [32,33], and it was subsequently shown that three full genome sequenced strains and the ex type strain of A. niger actually can produce FB 2 in comparable amounts to Fusarium strains [34] when grown on agar substrates with high amounts of sugar, glycerol or NaCl.
This was followed by the discovery of additional FB 4 production (~20% the amount of FB 2 ) by A. niger [35] in agar cultures and naturally A. niger contaminated Thai coffee beans [35].
The objectives of the present work were to i) screen A. niger and Fusarium strains, for production of FB 1 , FB 2 and FB 3 on three different agar substrates, ii) investigate the effect of incubation temperature on the production of fumonisins and iii) study the effect of the solutes glycerol, NaCl and sucrose on the production of fumonisins. The current work is performed on agar media instead of natural substrates in order to more easily asses the water activity.

Optimization of extraction
The efficiency of five different extraction solvents to extract FB 2 from A. niger (NRRL 567) varied significantly, with methanol:water (3:1) being most efficient, followed by acetonitrile:water (3:1) with a 20% lower efficiency and methanol:dichloromethane:ethyl acetate (1:2:3) (30% lower efficiency). The use of water (25°C) and hot water (100°C) was not suitable for extraction of FB 2 from A. niger NRRL 567 with a relative efficiency of <1%, when compared to methanol:water. The most effective solvents concerning F. verticillioides IBT 9400 were acetonitrile:water (3:1) as the best followed by methanol:water (3:1)(98%) and water at room temperature with an efficiency of 93%. The use of hot water as extraction solvent was less efficient (76%), compared to methanol:water. The acidic methanol:dichloromethane:ethyl acetate (1:2:3), had the lowest extraction efficiency with 26%. The difference between each of the two replicates for all incidences was in the interval 1-40% with an average of 10%.

Validation of methanol-water extraction from A. niger
The recovery of FB 2 from two spiked non-FB 2 producing strains showed a recovery of 75% ± 10% (IBT 20381) and 85% ± 10% (IBT 19345). The calibration curves from standards and spiked samples, used to calculate recovery from all had R 2 better than 0.995. The relative standard deviation (RSD) of the extracted amounts of FB 2 from the 8 isolates (n = 5) varied within 4-50%, with an average RSD of 20%. LOD were found to be 0.1 μg/cm 2 fungal culture.

Screening of strains for fumonisin production
The results from the screening experiment are shown in table 1. The A. niger strains were able to produce FB 2 on all three substrates, with the highest production on RC and CYAS. None of the A. niger strains produced detectable amounts of neither FB 1 nor FB 3 . LC-MS/MS analyses have shown that FB 2 is produced along with FB 4 , although the amount of FB 4 normally lies in the range 5-20% of the FB 2 amounts [35].
Of the A. niger strains, NRRL 567 had the highest production of FB 2 on RC and CYAS. But on PDA the three strains, NRRL 567, NRRL 2001 and IBT 24631, produced very similar amounts of fumonisins, although lower than RC and CYAS. The FB 2 production on RC and CYAS of A. niger NRRL 2001, IBT 24631 and 24634 differed only slightly, whereas FB 2 production by the other two, NRRL 567 and NRRL 3 were clearly favored by growth on RC. All Fusarium spp. with the exception of the two F. dlaminii strains IBT 2937 and IBT 2938 produced fumonisins under these conditions. Six Fusarium strains, F. napiforme IBT 2932, F. proliferatum IBT 9109 and IBT 9337, F. verticillioides IBT 9496 and F. nygamai IBT 9395 produced amounts close to the detection limit of FB 1 on RC; in addition IBT 2932 also produced fumonisins on PDA. F. nygamai IBT 2934, IBT 8554 and IBT 8557 showed a higher production of FB 2 than FB 1 on PDA (data not shown), and F. proliferatum IBT 9397 had the highest concentration of total fumonisin measured. Only seven strains had a total production of fumonisin above 1 μg/ml extract on RC compared to 14 strains on PDA.
Seven Fusarium strains did not have any measurable production of FB 1 , FB 2 nor FB 3 on PDA compared to two nonproducers on RC. On the other hand PDA supported production of higher amounts of fumonisins, strains which had a barely detectable amount of FB 1 on RC, also showed production of either FB 2 or both FB 2 and FB 3 on PDA.
Strains for the next experiments were selected on the basis of the above mentioned experiments. Besides the five A.
niger strains five Fusarium strains were selected, with both good and poor producers at 25°C, two strains with a high FB production, F. verticillioides IBT 9400 and F. oxysporum IBT 9514, one with an average FB production F. proliferatum IBT 41107, one with a low FB 1 production F. napiforme IBT 2932 and a strain with a higher production of FB 2 than FB 1 F. nygamai IBT 8554 was selected.

The effect of temperature on growth and production of fumonisins by A. niger and Fusarium spp
Only one of the A. niger strains was able to grow at 15°C, although very slowly (extensive data shown in [Additional file 1]). The growth increased at higher temperature  A: FB 2 production by Aspergillus niger after 7 days of growth at different temperatures (15-42°C) on CYAS and peaked at 30-37°C, followed by a slight reduction of the growth at 42°C. All Fusarium strains were able to grow at 15°C followed by an increased growth at higher temperatures and peaked at 25-30°C, above this temperature the growth decreased and no growth was observed at 42°C. The effect of temperature on the production of FB 2 by the five A. niger strains is shown in figure 1A. None of the isolates produced detectable amounts of FB 2 at 42°C, even though all strains grew well. The only A. niger strain, NRRL 2001, that was able to grow at 15°C, did not have any detectable production of FB 2 .
At 37°C, the optimal temperature for growth by A. niger, the FB 2 production was very limited. NRRL 567 had the highest production of the five strains at 25°C and 30°C; however at 20°C NRRL 2001 showed the highest production. This strain showed also the highest growth rate at this temperature. Four strains, NRRL 567, NRRL 3, NRRL 2001 and IBT 24634 had the highest FB 2 production at 25°C, followed by 30°C, 20°C, and 37°C ( Figure 1A). One strain, IBT 24631, differed from the other four, since it had a maximal production at 30°C, followed by 25°C, 20°C and 37°C.
The concentration of FB 1 , FB 2 and FB 3 produced by Fusarium spp. is shown in figure 1B,

The effect of glycerol, NaCl and sucrose on the growth and production of fumonisins by A. niger and Fusarium spp
All strains of A. niger and Fusarium spp. were able to grow at all glycerol concentration (0-255 g/l) [see Additional file 2]. The growth of A. niger was only slightly reduced at a w 0.99; below this there was a continuous decrease in the growth. For the Fusarium spp. the growth was reduced at a w 0.99 and below, which was the same as the A. niger. All A. niger strains were able to produce FB 2 at all glycerol concentrations tested ( Figure 2). The effect of glycerol on the quantitative FB 2 production for A. niger was very strain dependent, but in general the average FB 2 production was reduced 8.2% per 0.01 a w unit (R 2 = 0.97). Three of the five Fusarium strains were able to produce fumonisin at all glycerol concentrations: F. nygamai IBT 8554, F. oxysporum IBT 9514 and F. verticillioides IBT 9400 ( Figure 2). For two of these, IBT 8554 and IBT 9400, fumonisin production was increased up to 20% when glycerol was added. The other two strains F. napiforme IBT 2932 and F. nygamai IBT 8554 did not have a measurable production of fumonisins at a w 0.99 and 0.98. The average total fumonisin production was reduced 18% per 0.01 a w unit (linear regression R 2 = 0.91).
For the five A. niger growth in the presence of NaCl was partially inhibited below a w 0.98 and all strains were able to grow at all a w tested. The Fusarium species had the highest growth rates at the highest a w and the growth declined until a w 0.93 where they were unable to grow.
The FB 2 production of all A. niger strains was stimulated when NaCl was added with FB 2 being detected at all NaCl concentrations ( Figure 2). The production peaked at a w 0.985-0.97 depending on the strain. A reduction of the FB 2 production started at a w 0.97-0.94 and a comparison by a regression analysis to the cultures with no NaCl added the average production of FB 2 decreased 13% per 0.01 a w unit (R 2 = 0.98). The addition of NaCl clearly reduces the production of fumonisins by Fusarium ( Figure  2), and was not detected in any culture below a w 0.955. One strain, F. napiforme IBT 2932, did not even have a measurable production of fumonisin when NaCl was added. But no correlations in the average decrease of fumonisin production was observed.
The A. niger and Fusarium spp. strains grew better on sucrose, compared to NaCl and glycerol. The addition of sucrose increased the production of FB 2 for all the A. niger strains ( Figure 2). The increase in the FB 2 production was followed by either a reduction or stagnation in the production level at higher sucrose concentrations. The average decrease in the total fumonisin production were 8.3% per 0.01 a w unit (R 2 = 0.87) Only three of the five Fusarium strains were able to produce FB 1 , FB 2 and FB 3 when cultured on different sucrose concentrations (Figure 2). Because a commercial potato extract without sucrose was not obtainable, the base potato extract was prepared on home-made boiled potatoes. The three Fusarium strains with a detectable fumonisin production were F. oxysporum IBT 9514, F. proliferatum IBT 41107 and F. verticillioides IBT 9400. Fusarium oxysporum IBT 9514 had a maximal production at a w 0.995, while F. proliferatum IBT 41107 peaked at a w 0.999 and F. verticillioides IBT 9400 peaked at a w 0.999. A reduction of the fumonisin production was observed at higher sucrose concentrations. There were no correlations in the average decrease of fumonisin production Effect of glycerol, NaCl and sucrose on the FB 2 production of Aspergillus niger and the total production of FB 1 , FB 2 and FB 3 by F. verticillioides, F. proliferatum, F. napiforme, F. nygamai and F. oxysporum

Discussion
We found that A. niger was able to produce FB 2 , in agreement with Frisvad et al. [34] who showed that A. niger produced FB 2 on the agar substrates RC and CYAS. On the other hand they did not measure any production of FB 2 on PDA, whereas all five A. niger strains in the present study had a detectable, albeit low, production of FB 2 on this substrate. Besides F. verticillioides, also F. napiforme, F. nygamai, F. proliferatum and F. oxysporum were found to produce fumonisins on laboratory agar substrates. The ability to produce fumonisins by these species correlates with findings of Nelson et al. [8] and Kpodo et al. [9]. However Nelson et al. [8] described the production of FB 1 by F. dlaminii, but this was not supported in our study. The findings of strains, capable of producing more FB 2 than FB 1 was also described by Musser & Plattner [36] as well as Leslie et al. [37]. Apart from the four A. niger strains shown to produce FB 2 by Frisvad et al. [34], additional 12 strains did produce this mycotoxin. Among the 18 A. niger strains investigated until now, only two have been unable to produce fumonisins in detectable amounts on the media investigated.
Astoreca et al. [38] found the optimal temperature for growth of A. niger to be 30°C, the highest investigated temperature in their study. In correlation to this study Palacios-Cabrera et al. [39] also found that A. niger grew optimally at temperatures of 30°C, which was also the optimal temperature for linear growth in our study. Marin et al. [40] found that the growth of both F. verticillioides and F. proliferatum was best at 25 to 30°C, which is in agreement of our results. According to Marin et al. [40]F. verticillioides is more tolerant to temperature above 30°C than other Fusarium spp., however this was not obvious in our study.
We found that the optimal production of FB 2 by A. niger was at 25°C and in one case 30°C. Since there has been only one report of the fumonisin production by A. niger we have compared our results to the production of ochratoxin A. Our results correlates with an investigation of Esteban et al. [41], who showed that the optimal temperature for production of ochratoxin A by A. niger was at 20-25°C. In contrast to this, other authors found that the optimal ochratoxin A production in a synthetic grape juice medium was significantly better at 15°C compared to both 25 and 35°C [42]. Findings in our study showed that at 20°C a significant decrease in the FB 2 production occurred compared to 30°C. Earlier studies show that the optimal temperature for production of fumonisins by F. proliferatum is at 15-20°C where F. verticillioides prefers the higher temperatures of 30°C [27]. This partly correlates with the results from our study, where both isolates showed the highest production at 20°C, but also produced fumonisins at 30°C. However fumonisin produc-tion by F. verticillioides was less inhibited than F. proliferatum by the higher temperature. Dilkin et al. [26] and Alberts et al. [25] found the optimal temperature for fumonisin production to be 25°C, followed by 20 and then 30°C. These results deviate from our results, because four isolates had the best production at 20°C and one at 25°C. Marin et al. [27] described the production of FB 1 by F. verticillioides at 37°C, which was also observed in our study. From our results a general pattern in the fumonisin production for both genera was observed, namely the maximal production of fumonisins being 5°C below the optimal growth.
Comparing A. niger with the Fusarium spp., all A. niger strains grew better at all tested a W values. Leong et al. [42] found the optimal a w for ochratoxin A production by A. niger to be a w 0.95, whereas Esteban et al [43] found it to be in the range of 0.96-0.99, and that it was very strain dependent. These values are lower than those observed in our study for fumonisins where four of the A. niger strains had the highest production of FB 2 at a W 0.99, with one isolate produced most FB 2 at a w 0.98. Earlier studies have shown that the optimal a W for fumonisin production by Fusarium is in the interval 0.97-0.98 [27,44]. The optimal a w value from our study was a W 0.995, which is a bit higher than the above mentioned a w values. Frisvad et al. [34] also found that the addition of 5% NaCl (a w = 0.97) or 20% sucrose (a w = 0.99) increased the production of FB 2 by A. niger. The present study showed that not all the used strains had the same pattern. Only four of the five strains had an increase in FB 2 production at a w 0.97 compared to the zero sample when grown on NaCl. The last A.niger NRRL 3 had only an increase at a w 0.985 followed by a decrease at higher NaCl concentrations. The same was observed with sucrose where three strains had an increase in the FB 2 production compared to a sucrose concentration of 3% in standard medium. The last strains had a decrease in the production at the previous mentioned sucrose concentrations. In conclusion it is clear that there is very large strain variability in fumonisin production at different water activities between the A. niger strains used in these experiments.
Further studies on the effect of physiological variables on fumonisin production by A. niger is needed on large numbers of strains, due to large strain differences, and should further be backed up by studies of commodities where A. niger is common. This will lead us to a better understanding of how large a food safety problem fumonisin production by A. niger is. Only one report of fumonisins from A. niger in food (green coffee beans) has been reported, however the amounts were well below the regulatory limit [35]. Fumonisin contamination in a series of foods, often contaminated with A. niger, may show to be an overlooked mycotoxin problem, since surveys have only mostly targeted corn and rice, based on the belief that it was only Fusarium spp. that could produce these important mycotoxins.

Conclusion
The present study shows that the regulation of fumonisin production is very different in A. niger compared to Fusarium, with the latter preferring high a w (>0.99) and low temperature (20-25°C) and with A. niger preferring lower a w and higher temperature (25-30°C). Fumonisin produced by A. niger may be an overlooked health risk in foods not previously associated with fumonisins, especially because A. niger is known as a common food spoilage organism on a wide variety of foods [45].

Methods
Unless otherwise is stated all solvents were HPLC grade, chemicals were analytical grade and water was purified on a Milli-Q system (Millipore, Bedford, MA). Media were prepared in 9 cm Petri dishes, each with 20 ml medium, and strains were inoculated by single point and dishes incubated in micro perforated plastic bags at 25°C for 7 days in darkness. The colony diameter measured is an average of the smallest and broadest diameter. All samples were as a minimum made in biological duplicates on two individually plates.

Fungal strains and media
All strains (Table 2) were from the IBT culture collection at Center for Microbial Biotechnology, DTU. The Fusarium strains were selected from species known to produce fumonisin. Five A. niger strains were selected, for the physiologically study, three of these have been used in industry according to collection databases (NRRL 3, NRRL 567 and NRRL 2001), whereas the last two were isolated from black pepper (IBT 24634, IBT 24631). The last 9 strains were only used for validation. All A. niger strains were characterized and identified using a polyphasic approach, as in our previous articles [46,47], in order to ensure that they were A. niger sensu stricto. Besides this the ITS region and parts of the β-tubulin and calmodulin genes were amplified and sequenced as described previously [48][49][50]. Results shown in Table 2 (public database numbers of the strains).

Fumonisin analysis
The fumonisin were extracted using the method previously described by Frisvad et al. [34]. Six plugs (D = 6 mm) were cut out of the colony from the center and in a radius towards the edge of the colony and transferred to a clean 2-ml vial, 800 μl of methanol:water (3:1) was added, and extracted by ultrasonication for one hour. All extracts were filtered through a 13 mm PTFE 0.45 μm syringe filter (National Scientific, Rockwood, Tennessee) into a new vial and used directly for analysis.
The LC-MS analysis was performed on a LC/MSD VL single quadrupole (Agilent, Santa Clara, California). The separation of 3 μl extracts were done at 40°C on a 50 × 2 mm, i.d. 3 μm size, Luna C-18 (II) column (Phenomenex, Torrance, California), fitted with a security guard column, and using a water:acetonitrile (both containing 20 mM formic acid) gradient at a flow rate of 0.3 ml/min. The gradient started at 30% acetonitrile, and increased to 60% acetonitrile over 5 minutes. During further 1 minute it was increased to 100% acetonitrile and maintained here for 2 minutes before the gradient in 1 minute was returned to starting conditions and kept there for 5 minutes. The mass spectrometer (MS) was operated in positive electrospray ionization mode and was automatically calibrated on the instrument ESI tuning mix. The MS was used in selected ion monitoring (SIM) mode for measuring [M+H] + : FB 1 (m/z 722), FB 2 and FB 3 (m/z 706). The capillary voltage was held at 3000 V, the fragmentor voltage was at 70 V and the nebulizer pressure was at 2.5 bar. The drying gas flow was 12 l/min with a temperature of 350°C. The detection limit was measured to 0.01 μg/ml from dilutions of a FB 1 and FB 2 certified standard (Biopure, Tulin, Austria), with concentrations of 50.2 μg/ml and 51.0 μg/ml, respectively. The fumonisin concentrations of the extracts were calculated from a standard curve created from dilutions of the FB 1 and FB 2 standard mixture. FB 2 presence was further confirmed in selected extracts by LC-MS/MS [35] and LC-HRMS [34].

Efficiency of extraction solvents
The five different extraction solvents tested on A. niger NRRL 567 and F. verticillioides IBT 9400 were: i) methanol:water (3:1), ii) acetonitrile:water (3:1), iii) water at room temperature (25°C), iv) water at 100°C and v) methanol:dichloromethane:ethyl acetate (1:2:3) with addition of 1% (v/v) formic acid. The extraction process for [i-iv] was the same as mentioned above. For the fifth extraction solvent there were a few extra steps: After ultrasonication the extract was transferred to a new vial and the organic phase was evaporated in vacuo. The residue was redissolved by ultrasonication in 500 μl methanol for 20 minutes. All extracts were filtered through a PTFE 0.45 μm syringe filter before analysis.
Validation of methanol:water extraction FB 2 extraction was validated by spiking 5 plugs of two non-fumonisin producing strains of Aspergillus niger (IBT 19345 and IBT 20381) with 100 μl FB 2 standard containing 5000, 2500, 1000, 500 and 100 ng FB 2 . After spiking, the plugs were left for 2 hours, and extracted as described above. Recovery was determined by comparing the slope of the spiked curve to slope of the curve from diluted samples in methanol-water (3:1). Five blank samples of the 2 strains were also analysed.
Reproducibility of extraction efficiency was further determined from 7 A. niger strains (IBT 4983, IBT 18741, IBT 19558, IBT 26774, IBT 28086, IBT 28104, NRRL 567) selected from high, medium and low fumonisin producing strains. Five replicate plates were made from each strain and extracted as described above.

The effect of temperature on the growth and production of fumonisin by A. niger and Fusarium spp
To assess the effect of temperature on the production of fumonisins, A. niger strains were inoculated on CYAS and Fusarium strains were inoculated on PDA. The plates were incubated in darkness at 15,20,25,30,37 or 42°C respectively for 7 days. The fungi used were the five A. niger strains listed in Table 2, and the following Fusarium species: F. napiforme IBT 2932, F. nygamai IBT 8554, F. verticillioides IBT 9400, F. oxysporum IBT 9514 and F. proliferatum IBT 41107.

The effect of glycerol, NaCl and sucrose on the growth and production of fumonisin by A. niger and Fusarium spp
For investigation of the effect of glycerol, NaCl and sucrose on the production of fumonisin the following experiments were performed: A. niger was inoculated on CYA and Fusarium on PDA with different concentration of glycerol (0-255 g/l), salt (0-125 g/l) and sucrose (0-530 g/ l). For the Fusarium-sucrose experiment, the PDA mixture was made from boiled potatoes instead of a commercial blend. In short terms, 200 g potatoes was peeled and diced and autoclaved at 121°C with 1 liter of water, 15 g agar and 0-530 g sucrose. The values are means of the triplicates plus/minus the standard deviation.