Effect of Temperature, Water Activity and Incubation Time on Trichothecene Production by Fusarium cerealis Isolated from Durum Wheat Grains

Fusarium cerealis is a causal agent of Fusarium Head Blight in wheat, and it produces both deoxynivalenol (DON) and nivalenol (NIV). Nevertheless, the effect of environmental factors on the growth and mycotoxin production of this species has not been studied so far. The objective of this study was to investigate the impact of environmental factors on the growth and mycotoxin production of F. cerealis strains. All strains were able to grow in a wide range of water activity (aW) and temperatures, but their mycotoxin production was influenced by strain and environmental factors. NIV was produced at high aW and temperatures, while optimal conditions for DON production were observed at low aW. Interestingly, some strains were able to simultaneously produce both toxins, which could pose a more significant risk for grain contamination.


Introduction
Fusarium head blight (FHB) is an important fungal disease affecting wheat production worldwide. Economic losses occur due to reductions in grain yield and quality, mainly from the degradation of gluten proteins by the fungus resulting in lower bakery properties [1]. In addition, grains could be contaminated with mycotoxins that, in some cases, are regulated at the maximum permitted levels, making cereal trading more difficult. In addition, these mycotoxins are harmful for human and animal health [2]. The main mycotoxins produced are trichothecenes which are sesquiterpenoides that have in common a tricyclic 12,13-epoxytrichothec-9-ene (EPT) core structure [3]. According to the substitution groups on the EPT core, they are divided into four classes (types A-D). Deoxynivalenol (DON) and nivalenol (NIV) are included within the type B trichothecenes. These mycotoxins affect protein synthesis and, in consequence, DNA biosynthesis by inhibiting the peptidyl transferase activity of 60S ribosomes [4]. In humans and animals, exposure to mycotoxins causes acute and chronic effects such as abdominal pain, feed refusal, anorexia, diarrhea, immunological problems, vomiting, skin dermatitis, and hemorrhagic lesions [5,6]. Trichothecenes can act as immunostimulatory or immunosuppressive, depending on the dose, exposure frequency, and type of immune function assay [7]. In plants, they can cause dwarfism, chlorosis, and inhibition of root and shoot elongation and act as a virulence factor in FHB [8]. DON is the main trichothecene found in cereals and cereal-based products [9]. However, it has been reported that NIV is more cytotoxic than DON [10][11][12]. In addition, several reports showed additive and/or synergetic effects when DON is combined with NIV; however, these interactions are dose-dependent [13][14][15]. For instance, Alassane-Kpembi et al. [16] showed that below a cytotoxicity level of 50%, the combination of DON and NIV causes Fusarium cerealis strains RCFG 6029, RCFG 6046, and RCFG 6076 were isolated from FHB symptomatic durum wheat grains from the main growing area in Argentina. The identity of the RCFG 6029 strain was confirmed by sequencing analysis of EF-1α gene (Accession N • KX359404), while the identity of the remaining strains was confirmed by using species-specific primers and inter-simple sequence repeats (ISSR) analysis [32,41]. Moreover, a phenotypic and molecular characterization of their mycotoxin production profile was previously carried out. The three strains produced DON and NIV on durum wheat grain cultures: the RCFG 6029 strain produced 16 µg/kg of NIV and 160 µg/kg of DON, the RCFG 6046 strain produced 193 µg/kg of NIV and 270 µg/kg of DON, and the RCFG 6076 strain produced 9284 µg/kg of NIV and 170 µg/kg of DON [27]. Cultures were purified by subculturing single macroconidia and preserved in sterile 15% glycerol at −80 • C [42]. Preserved cultures are maintained in the culture collection of the Research Institute on Mycology and Mycotoxicology (IMICO), CONICET-UNRC.

Medium
Durum wheat was finely milled by using a Romer mill (Romer Labs Inc., Union, MO, USA). Mixtures of 2% (w/v) of milled wheat in water were prepared, and 2% (w/v) agar (technical agar N • 2, Oxoid, Waltham, MA, USA) was added. The a W of the basic medium was adjusted to 0.99, 0.97, 0.95, 0.93, and 0.90 by addition of different amounts of glycerol [43]. The media were autoclaved at 120 • C for 20 min. Flasks of molten media were thoroughly shaken, prior to being poured into 9 cm sterile Petri dishes. The a W of representative samples (2 of each treatment) of media was checked with an Aqualab Series 3 (Decagon Devices, Inc., Pullman, WA, USA). Additional, uninoculated control plates were prepared and measured at the end of the experiment in order to detect any significant deviation of the a W .

Inoculation, Incubation, and Growth Assessment
Petri plates were inoculated with a 4 mm diameter agar disk that was taken from the margin of a 7-day-old colony of each isolate grown on synthetic nutrient agar [42] at 25 • C and transferred face down to the center of each plate. Inoculated plates of the same a W were sealed in polyethylene bags and incubated at 15, 20, 25, and 30 • C for 28 days. At 7, 14, 21, and 28 days after incubation, some plates were taken in triplicate for mycotoxin analysis. A full factorial design was used, where the factors were a W , temperature, and strain, and the response was growth (total number of plates: 5 a W × 4 temperatures × 4 incubation periods × 3 strains × 3 replicates).
Assessment of growth was made every day during the incubation period, and two diameters of the growing colonies were measured at right angles to each other until the colony reached the edge of the plate. Colonies' radii were plotted against time, and linear regression was applied in order to obtain the growth rate (mm/day) as the slope of the line. After the incubation period, uninoculated controls and treatments were frozen for later mycotoxin determination.

Nivalenol and Deoxynivalenol Analysis
For mycotoxin analysis, three replicates per treatment were destructively sampled after 7, 14, 21, and 28 days. DON and NIV extraction was completed according to Li et al. [44], with some modifications. Briefly, 8 g of culture medium were shaken with 20 mL of acetonitrile in a vortexing machine (Vortexer) for 5 min and then on an orbital shaker (150 rpm) for 1 h. Subsequently, the sample was centrifuged at 10,000× g rpm for 5 min (Sorvall™ ST 16, Thermo Fisher Scientific, Waltham, MA, USA). Next, 8 mL of the supernatant were transferred into a 15 mL centrifuge tube containing 1.5 of NaCl. The tube was shaken for 2 min by using a vortexing machine. Then, it was centrifuged at 3500 rpm for 5 min, and 5 mL of the supernatant were transferred into a vial. The acetonitrile was blown to dryness under nitrogen flow. The dried residue was redissolved in 1 mL (acetonitrile/water = 1:9, v/v) and homogenized with a vortexing machine, and then a nylon filter (0.22 µm) was applied to remove impurities. Samples were stored at 4 • C until analysis by HPLC. The HPLC system consisted of a Waters e2695 separations module (Milford, MA, USA) with a stainless steel C18 reversed-phase column (150 × 4.6 mm, 5 µm particle size; LunaPhenomenex) connected to pre-column (20 × 4.6 mm, 5 µm particle size, Luna-Phenomenex). Mycotoxins were detected by UV at 220 nm using a Waters 2998 diode array detector. The mobile phase was water:methanol (88:12, v/v) at a flow rate of 1.3 mL/min. The limit of detection (LOD) was 15 µg/kg, and the limit of quantification (LOQ) was 40 µg/kg. The retention times of NIV and DON were 5.2 min and 10.7 min, respectively.

Statistical Analysis
The growth rate and mycotoxin concentration data were transformed to normal scores, and they were evaluated by analysis of variance (ANOVA) using InfoStat for Windows version 2018 [45]. For normality and homogeneity of variances, data were tested using Shapiro-Wilk test and Levene test, respectively. Statistical significance was judged at the level of p ≤ 0.0001. When the analysis was statistically significant, the post hoc Fisher's least significant difference (LSD) test was used for the separation of means (significance judged at p ≤ 0.05).

Effect of a W and Temperature on Lag Phase
The ANOVA analysis (Table 1) shows that all the single variables and two-and threeway interactions had a significant impact on the lag phase of the three strains, except the a W × strain interaction. The Table 1 reveals that the a W influenced the lag phase (F = 860.95) the most, followed by temperature and strain. Table 1. Analysis of variance (ANOVA) on the effects of strain (S), temperature (T • ), water activity (a W ), and their interactions on lag phase and growth rate of Fusarium cerealis in a wheat-based medium.  Figure 1A-C show that the shortest lag phases were observed at 25 • C, 0.99 a W. They corresponded to the F. cerealis RCFG 6076 (23.47 h) and F. cerealis RCFG 6046 (24.86 h) strains. In general, the lag phases increased at a more stressful a W (0.93-0.95) and low temperatures (15-20 • C). However, the behavior of the RCFG 6076 strain showed some differences, since at 15 • C the lag phase was relatively short and did not show significant differences at a W ≥ 0.97. These values were similar to those obtained at 30 • C, 0.95 and 0.97 a W . The largest lag phase (107.74 h) was observed for the RCFG 6029 strain at 15 • C, 0.93 a W ( Figure 1A). Table 1 shows that F. cerealis growth rate was significantly influenced by all the single factors evaluated in the study (strain, a W , and temperature) and by their interactions. The a W had the most important effect on the growth rate. Figure 1D-F show the growth rates of all strains according to the different a W and temperatures studied. For the three F. cerealis strains, the maximum growth was observed at 25-30 • C, 0.99 a W , while the minimal growth was observed at 15 • C, 0.93 a W . The growth of all strains decreased as the water availability of the media was reduced. No growth was observed at 0.90 a W during the incubation period, regardless of the temperature, for any tested strain.

Effect of a W and Temperature on Growth Rate
The highest growth rate was observed for the RCFG 6076 and RCFG 6046 strains at 25 • C, 0.99 a W (8.27 and 8.14 mm/day, respectively), and the values were statistically significantly different compared to the rest of the obtained growth rates. Within some a W , temperature showed little or no effect on the growth rates. At 25-30 • C, there were not any statistically significant differences among the growth rates at 0.93, 0.95, and 0.97 a W for the RCFG 6076 strain ( Figure 1E), at 0.95 and 0.97 for the RCFG 6029 strain ( Figure 1D), and at 0.95 for the RCFG 6046 strain ( Figure 1F).  On the other hand, the lowest growth rate was observed for the RCFG 6046 strain at 15 • C, 0.93 a W (0.74 mm/day).
In order to identify the optimum and marginal conditions that allowed the growth of F. cerealis strains, three contour maps were performed. They show the isopleths for different growth rates according to a W and temperature. Figure 2 shows that F. cerealis can grow in a wide range of a W (0.99 to 0.93) and temperatures (15-30 • C).

Effect of aW, Temperature, and Incubation Time on Nivalenol and Deoxynivalenol Production
All the factors under study (aW, temperature, incubation time, and strain) and their interactions influenced NIV and DON production by the three F. cerealis strains on a wheat-

Effect of a W , Temperature, and Incubation Time on Nivalenol and Deoxynivalenol Production
All the factors under study (a W , temperature, incubation time, and strain) and their interactions influenced NIV and DON production by the three F. cerealis strains on a wheatbased medium ( Table 2). The analysis of variance showed that the most important factor that influenced the NIV production was strain, while, for DON production, the incubation time and a W were the most important. Figures 3 and 4 show the NIV and DON production by F. cerealis strains at 15, 20, 25, and 30 • C; at four different a W ; and during 7, 14, 21, and 28 days of incubation. In general, F. cerealis was able to produce DON and NIV at all the conditions tested; however, the amount and the type of toxin was strain-dependent, since they showed different trichothecene production profiles. The toxin levels observed for the RCFG 6076 strain were higher than those observed for the other strains under almost all the temperatures and a W that were tested. The highest NIV level (23,918 µg/kg) was produced by the RCFG 6076 strain at 30 • C, 0.99 a W , after 14 days of incubation. The maximum NIV production of the other strains was 13,432 µg/kg (30 • C, 0.99 a W , 14 days) and 3,527 µg/kg (25 • C, 0.99 a W , 7 days) for RCFG 6046 and RCFG 6029, respectively. The lowest level detected (41 µg/kg) was observed at 20 • C, 0.93 a W , after 28 days of incubation for the RCFG 6076 strain. Before this period, the toxin was not detected in samples at low a W and temperatures (LOD: 15 µg/kg, LOQ: 40 µg/kg) (Figure 3).
In general, the optimal condition for NIV production was 25-30 • C, 0.99 a W ; as a W was reduced, NIV production declined. For the RCFG 6076 and RCFG 6046 strains, NIV production was higher at 30 • C, but it decreased at 25 • C and 20 • C; while, for the RCFG 6029 strain, higher production was observed at 25 • C, but it decreased at 30 • C. At 15 • C, NIV was only produced by the RCFG 6076 strain at all a W that were tested after 28 days of incubation.
The three F. cerealis strains were capable of producing DON; however, the production profile was different for each strain (Figure 4). DON was produced by all the studied strains only after 28 days of incubation. The maximum production (2954 µg/kg) was observed at 30 • C, 0.93 a W (RCFG 6076 strain), while the minimum value obtained was 660 µg/kg at 15 • C, 0.93 a W (RCFG 6046 strain). DON production was mostly observed at a W ≤ 0.97 at all temperatures, although the maximum levels were mainly observed at 0.93 a W . At 20 • C, all strains produced DON but at different a W . At 15 • C, only the RCFG 6046 strain produced the toxin at 0.93 a W . The RCFG 6029 strain only produced DON at 20 • C, 0.99-0.97 a W .   Maximum NIV levels were detected at 14 days of incubation, while, for DON, they were detected after 28 days of incubation. Considering this, we analyzed the optimum conditions for temperature and a W , where the F. cerealis strains produced the maximum levels of NIV and DON, in order to identify those conditions with the highest toxicological risks for each strain ( Figure 5). Higher NIV levels were detected at 0.99 a W for the three strains. The RCFG 6029 strain produced the maximum NIV levels at 25 • C; the RCFG 6076 strain at the range of 25-30 • C, and the RCFG 6046 strain at 30 • C. The maximum DON levels produced by RCFG 6029 were observed at 20 • C, and 0.97-0.99 a W . For RCFG 6076 strain, the highest levels were observed at 25-30 • C and 0.93-0.95 a W . While, for RCFG 6046 the highest production was detected at 25 • C, 0.95 a W . However, appreciable levels were also obtained at 15 • C, 0.93 a W ( Figure 5). Maximum NIV levels were detected at 14 days of incubation, while, for DON, they were detected after 28 days of incubation. Considering this, we analyzed the optimum conditions for temperature and aW, where the F. cerealis strains produced the maximum levels of NIV and DON, in order to identify those conditions with the highest toxicological risks for each strain ( Figure 5). Higher NIV levels were detected at 0.99 aW for the three strains. The RCFG 6029 strain produced the maximum NIV levels at 25 °C; the RCFG 6076 strain at the range of 25-30 °C, and the RCFG 6046 strain at 30 °C. The maximum DON levels produced by RCFG 6029 were observed at 20 °C, and 0.97-0.99 aW. For RCFG 6076 strain, the highest levels were observed at 25-30 °C and 0.93-0.95 aW. While, for RCFG 6046 the highest production was detected at 25 °C, 0.95 aW. However, appreciable levels were also obtained at 15 °C, 0.93 aW ( Figure 5). There were a few conditions where the strains co-produce both toxins ( Figure 6). Only strains RCFG 6076 and RCFG 6046 were able to co-produced NIV and DON after 28 days of incubation. The RCFG 6046 strain only produced both toxins at 0.95 aW and 25 °C, There were a few conditions where the strains co-produce both toxins ( Figure 6). Only strains RCFG 6076 and RCFG 6046 were able to co-produced NIV and DON after 28 days of incubation. The RCFG 6046 strain only produced both toxins at 0.95 a W and 25 • C, while the RCFG 6076 strain co-produced DON and NIV at 0.95 a W and temperatures ≥ 20 • C and at 0.97 a W and 20 • C. In most cases, NIV production was higher than DON production. while the RCFG 6076 strain co-produced DON and NIV at 0.95 aW and temperatures ≥ 20 °C and at 0.97 aW and 20 °C. In most cases, NIV production was higher than DON production.

Discussion
In a previous study we have analyzed the mycotoxin profile of F. cerealis and we observed that it was capable of producing a wide range of mycotoxins and secondary metabolites including DON and NIV, both in vitro and in planta [27]. However, the environmental factors that influence its growth and trichothecene production have not been determined so far. The present study evaluated for the first time the influence of temperature, a W and time of incubation on F. cerealis growth and DON and NIV production, on a wheat-based medium. The three F. cerealis strains evaluated were able to grow at 0.99-0.93 a W and 15-30 • C. Maximum growth rates were observed at 0.99 a W and 25-30 • C and the optimal conditions for growth ranged from 0.99 to 0.97 a W for all tested temperatures. Minimal growth rates were observed at 15 • C, 0.93 a W . No growth was observed at 0.90 a W regardless of temperature and time of incubation tested. These growth parameters are in agreement with those reported for others FHB pathogens such as F. graminearum and F. culmorum which grew best at 0.995-25 • C and were able to grow in the range of 0.995-0.93 a W and 15-30 • C. In addition, no mycelial growth was observed at a W ≤ 0.90 [37][38][39][40].
Fusarium head blight is greatly affected by weather conditions. Infections during the flowering period are favored by extended periods (2-3 days) of >90% relative humidity and temperatures in the range of 15-30 • C [46]. These conditions are optimal for Fusarium pathogen development, as demonstrated in this study for F. cerealis.
Regarding mycotoxin analysis, the production patterns observed for the three F. cerealis strains were totally different.
Nivalenol was detected in all the conditions in which F. cerealis was able to grow. The highest levels were detected at 25-30 • C, 0.99 a W , between 14 and 21 days of incubation; after this period, the levels decreased. These temperatures and a W were the same in which the maximum growth was observed for F. cerealis, while the minimum levels detected for NIV were at 15 and 20 • C, 0.93 a W , after 28 days of incubation. Therefore, F. cerealis is capable of producing NIV at suboptimal conditions for growth, but it takes a longer period of time. These same results were observed for other Fusarium species such as F. graminearum and F. culmorum, which produced higher amounts of trichothecene at the highest a W and temperatures analyzed [38,40]. For instance, in Argentina, Ramirez et al. [39] observed that F. graminearum produced the maximum amount of DON at 0.995 a W and 30 • C on irradiated wheat grains. For those NIV-producing species such as F. poae, F. meridionale, F. culmorum, and F. asiaticum, it was reported that the optimum conditions for toxin production were between 20 and 30 • C and 0.98 and 0.995 a W [37,40,47,48].
Regarding DON production, the three F. cerealis strains were capable of producing the toxin, but only after 28 days of incubation. DON was produced at lower levels than NIV. The maximum levels were also obtained at 25-30 • C; however, contrary to NIV production, at 0.95 and 0.93 a W . This means that DON production was favored by the stressful conditions related to a W . This is in accordance with Schmidt-Heydt et al. [49], who developed a model to predicted DON production by F. culmorum, relating the specific gene expression of the key biosynthetic genes with a W levels. This model suggests that high DON levels would occur under water stress. Moreover, Marin et al. [50] observed an induction in the Tri5 gene expression (implicated in DON biosynthesis) of F. graminearum under water stress. Taking this into account, there could be a risk of grain contamination with DON during storage, since, according to Magan et al. [51], fungal spoilage and contamination of grains with mycotoxins may continue during storage if the moisture, temperature, and aeration are suitable for fungal growth and toxin production. They observed that DON contamination on stored wheat inoculated with F. graminearum exceeded the maximum permitted level (<1250 ppb) when it was stored at 30 • C and 0.95 and 0.93 a W. In addition, Garcia Celá et al. [52] analyzed mycotoxins (included DON and NIV) in naturally contaminated stored wheat and wheat inoculated with F. graminearum and concluded that the optimum production of these compounds may occur at 20-25 • C and 0.95 a W . They observed that the DON and NIV contamination levels were significantly affected by a W .
In addition, the present study shows that it is necessary to include more than one strain in ecophysiological studies to obtain more reliable results, due to the intraspecific variability that exists in a species, mainly within Fusarium [53]. For instance, in our study, we observed that, regarding growth conditions, the three strains shared the same patterns. However, when we analyzed the mycotoxin production, they had different production profiles. The RCFG 6076 strain produced the highest mycotoxin levels in all conditions. This is in accordance with our previous work, where we observed that this strain produced the maximum NIV level on durum wheat grain cultures [27]. Moreover, we observed that the RCFG 6029 strain produced DON in completely different conditions compared to the other strains. We could not have observed these results using only one strain.
The co-production of both toxins was observed in a few conditions, most of them for the RCFG 6076 strain, at 0.95 a W and between 20 and 30 • C, after 28 days of incubation. This is in agreement with our previous report of DON and NIV co-production by F. cerealis strains on durum wheat grain cultures and in planta. However, when F. cerealis strains grew on autoclaved durum wheat grains at 25 • C for 28 days, many of them produced higher amounts of DON than NIV. However, if we consider the two strains under study that co-produce both toxins, RCFG 6076 produced higher amounts of NIV, while RCFG 6046 produced higher amounts of DON [27]. This is consistent with the present study.

Conclusions
F. cerealis strains were capable of producing NIV in a wide range of temperatures and a W . High amounts of this toxin were produced at 0.99 a W and 25-30 • C, which are the same environmental conditions for optimal growth. These conditions are needed for Fusarium head blight development. So, in years with conducive conditions for the disease, grains could be contaminated with high NIV levels if they are infected by F. cerealis. Although DON production was not detected under these conditions, the risk of grain contamination with this mycotoxin could occur under storage conditions, since the optimal production seems to be at low a W . Moreover, the toxin was produced after 28 days of incubation by all strains. Why DON is produced after this period of time is a question that we still have to answer. This is, to our knowledge, the first ecophysiological study carried out about F. cerealis, so we do not have other studies to compare our results with. In order to answer this question, we are carrying out more studies, especially gene expression studies.
In addition, two strains were able to simultaneously produce both toxins. Therefore, grains could end up contaminated with both DON and NIV if F. cerealis is present, posing a more significant risk for human and animal consumption.

Conflicts of Interest:
The authors declare no conflict of interest.