Seasonal and Geographical Impact on the Mycotoxigenicity of Aspergillus and Fusarium Species Isolated from Smallholder Dairy Cattle Feeds and Feedstuffs in Free State and Limpopo Provinces of South Africa

Abstract This study evaluated the impact of seasonal and geographical variations on the toxigenicity of Aspergillus and Fusarium strains previously isolated from smallholder dairy cattle feeds and feedstuffs sampled during summer and winter in the Free State and Limpopo provinces of South Africa (SA). In total, 112 potential toxigenic fungal species were obtained and determined for their capability to produce mycotoxins on solid Czapek Yeast Extract Agar (CYA); followed by liquid chromatography-mass spectrometry (LC-MS/MS) analysis. Our result revealed that 41.96% of the fungal species produced their respective mycotoxins, including aflatoxin B1 (AFB1), aflatoxin B2 (AFB2), and zearalenone (ZEN), with higher levels of AFB1 (0.22 to 1045.80 µg/kg) and AFB2 (0.11 to 3.44 µg/kg) produced by fungal species isolated from summer samples than those in winter [(0.69 to 14.44 µg/kg) and (0.21 to 2.26 µg/kg), respectively]. The same pattern was also observed for AFB1 and AFB2 in Limpopo (0.43 to 1045.80 µg/kg and 0.13 to 3.44 µg/kg) and Free State (0.22 to 576.14 µg/kg and 0.11 to 2.82 µg/kg), respectively. More so, ZEN concentrations in summer (7.75 to 97.18 µg/kg) were higher than in winter (5.20 to 15.90 µg/kg). A similar observation was also noted for ZEN in Limpopo (7.80 to 97.18 µg/kg) and Free State (5.20 to 15.90 µg/kg). These findings were confirmed via Welch and Brown-Forsythe tests with significantly (p ≤ 0.05) higher mycotoxin levels produced by fungal strains obtained in samples during summer than those in winter. In contrast, the concentrations of mycotoxins produced by the fungal species from both provinces were not significantly (p > 0.05) different.


Introduction
Feed refers to any unprocessed or processed substance consumed by livestock to meet their nutritional demands, whereas feedstuff is the raw material from vegetable or animal sources required to formulate compound feeds for animal consumption. The proportion of essential components in livestock feeds differs based on the species, physiological status, age, and sex of the animal [1]. Dairy cattle are given the required diets to supply them with the essential nutrients mainly for milk production and dietary requirements vary significantly based on the gestation and lactation stages [2]. Another important aspect of dairy feed is its safety. In general, feed safety is a prerequisite for the safety of food of animal origin, particularly dairy products. Unfortunately, a majority of dairy cattle farmers in sub-Saharan African (SSA) nations, including South Africa (SA), lack the basic knowledge of feed safety, good farm management practices, and animal nutrition [3,4]. Dairy farmers in this region primarily focus on increasing productivity, with no concern for milk and milk product safety, often compromised when dairy cows are fed contaminated feeds. Among dairy cattle feed contaminants, mycotoxins are a severe threat to the dairy industry in SA [5,6].
Mycotoxins are toxic substances formed by toxigenic fungi which infect agricultural products on the farm or during storage under certain conditions, including improper storage, high relative humidity, high temperature, and insect and pest infestation [7,8]. These contaminants when found in dairy feeds and feed ingredients, especially at high contamination levels, might pose a serious threat to milking cows. Firstly, they may have a detrimental impact on dairy cattle health, such as decreased feeding efficiency, reduced milk yield, immunosuppression, weight loss, laminitis, low fertility, and abortion [2,9,10]. Secondly, they may compromise the food supply chain as a result of their transfer from feed to milk and milk products [11][12][13] owing to their tolerance to high temperatures [14], making milk production operations inadequate and ineffective for their complete elimination during milk production and processing [15][16][17]. As a result, humans are exposed to these toxins through the consumption of contaminated meat, milk, and milk products.
Human exposure to mycotoxins causes various health effects as they are known to be nephrotoxic, hepatotoxic, teratogenic, carcinogenic, and immunosuppressive [9,[18][19][20]. It may result in death in dire cases, such as prolonged chronic toxicity or high acute intoxication [18]. In addition to their negative impacts on animal and human health, mycotoxins also cause significant economic losses for many nations, particularly those in SSA, because of the costs directed toward food safety [21].
There is growing evidence to suggest that seasonal and geographical differences influence mycotoxin formation in food and feed [34,35]. Unfortunately, there are limited studies in SA to demonstrate this, particularly in dairy feeds and feed components. For the most part, past studies conducted in the country only revealed the prevalence of mycotoxins in dairy feeds and feedstuffs [33,36,37], while others only recognized fungi as the causative agents for the contamination [28,38]. Therefore, this study aims to screen and quantify the extent of mycotoxin production by the toxigenic fungal species invading smallholder dairy feeds in SA, based on seasonal (summer and winter) and geographical (Limpopo and Free State) variations. The findings from this study could be a significant reference for mitigating and preventing mycotoxin contamination in dairy feeds and, subsequently, in human diets derived from dairy cattle in SA.

Method Validation
Mycotoxins synthesized by Fusarium and Aspergillus strains in this research were identified and quantified by means of liquid chromatography-mass spectrometry (LC-MS/MS). The mycotoxins tested for in this study include AFB 1 , AFB 2 , AFG 1 , AFG 2 , OTA, ZEN, and DON. Standard validation parameters for each matrix were achieved based on recovery, retention time, linearity, the limit of detection (LOD), and the limit of quantification (LOQ). As seen in Table 1, the analytical technique revealed linearity with R 2 for all mycotoxin levels varying from 0.9966 to 0.9995. Furthermore, the LODs and LOQs of various mycotoxins varied from 0.01 to 4.42 and 0.04 to 13.40 µg/kg, respectively, whereas the mean recoveries varied from 71.4 to 101.9%, falling within the European Commission (EU) acceptable performance criteria [39]. Figure 1 represents the multiple reaction monitoring (MRM) chromatograms of AFs (AFB 1 and AFB 2 ) produced by A. flavus and ZEN formed by F. oxysporum in contrast to corresponding mycotoxin standards.

Method Validation
Mycotoxins synthesized by Fusarium and Aspergillus strains in this research were identified and quantified by means of liquid chromatography-mass spectrometry (LC-MS/MS). The mycotoxins tested for in this study include AFB1, AFB2, AFG1, AFG2, OTA, ZEN, and DON. Standard validation parameters for each matrix were achieved based on recovery, retention time, linearity, the limit of detection (LOD), and the limit of quantification (LOQ). As seen in Table 1, the analytical technique revealed linearity with R 2 for all mycotoxin levels varying from 0.9966 to 0.9995. Furthermore, the LODs and LOQs of various mycotoxins varied from 0.01 to 4.42 and 0.04 to 13.40 µ g/kg, respectively, whereas the mean recoveries varied from 71.4 to 101.9%, falling within the European Commission (EU) acceptable performance criteria [39]. Figure 1 represents the multiple reaction monitoring (MRM) chromatograms of AFs (AFB1 and AFB2) produced by A. flavus and ZEN formed by F. oxysporum in contrast to corresponding mycotoxin standards.

Toxigenicity of Aspergillus and Fusarium Species
Data on the mycotoxigenicity of the fungal isolates are summarized in Tables 2 and 3 as well as in Supplementary Tables S1 and S2. About 41.96% of the total 112 fungal species were positive, producing their respective mycotoxins (AFB 1 , AFB 2 , and ZEN). Furthermore, 51.47 and 27.27% of the fungal isolates recovered during summer and winter, as well as 37.5 and 47.92% of those recovered from Free State and Limpopo feeds tested positive for the analyzed mycotoxins. More so, our results revealed that A. flavus was the sole AF producer among the 83 Aspergillus species isolated from the feed samples, producing only AF B-types (AFB 1 and AFB 2 ). Of the 29 A. flavus strains examined in this work, 82.76 and 37.93% were positive for AFB 1 and AFB 2 , respectively. However, 90 and 45% of summer A. flavus strains produced AFB 1 and AFB 2 . Similarly, 66.67 and 22.22% of winter A. flavus strains produced the same mycotoxins. It is worth noting that 75 and 31.25% of A. flavus from Free State samples, as well as 92.31 and 46.15% of A. flavus from Limpopo, were positive for AFB 1 and AFB 2 , respectively. Interestingly, none of the strains of A. ochracheus and A. niger, the well-known producers of OTA, produced mycotoxin. Also, no strain of A. fumigatus in this study produced any toxin.
In addition, ZEN, the sole Fusarium mycotoxin found in this work, was primarily formed by strains of F. oxysporum and F. equiseti, with prevalence rates of 63.64 and 71.43%, respectively. It must be emphasized that 3/3 (100%) and 2/4 (50%) of the summer and winter F. equiseti produced ZEN, whereas 2/2 (100%) and 3/5 (60%) of the same isolate from Free State and Limpopo feed samples were toxigenic with respect to this toxin. Among the F. oxysporum strains examined, 5/8 (62.5%) and 2/3 (66.67%) from summer and winter samples, and 5/8 (62.5%) and 2/3 (66.67%) of those from Free State and Limpopo tested positive for ZEN. Meanwhile, all Fusarium spp. failed to produce DON.
As shown in Tables 2 and 3 and Figure 2, the concentrations of AFB 1 produced by A. flavus in summer (range: 0.22-1045.80; mean: 127.01 µg/kg) were higher than in winter (range: 0.69-14.44; mean: 3.85 µg/kg). AFB 2 showed the same trend in summer (range: 0.11-3.44; mean: 1.40 µg/kg) and winter (range: 0.21-0.26; mean: 0.24 µg/kg). Additionally, ZEN concentrations formed by both F. oxysporum and F. equiseti strains were higher in summer (range: 7.75-97.18; mean: 22.55 µg/kg) than in winter (range: 5.20-15.90; mean: 9.46 µg/kg). Furthermore, AFB 1 levels produced by A. flavus from Free State (range: 0.22-576.14; mean: 70.47 µg/kg) were lower than those recorded in A. flavus from Limpopo (range: 0.43-1045.80; mean: 122.13 µg/kg). A similar pattern was also seen in AFB 2 levels formed by A. flavus from Free State (range: 0.11-2.82; mean: 1.14) and Limpopo (range: 0.13-3.44; mean: 1.22 µg/kg). ZEN concentrations produced by F. oxysporum and F. equiseti strains from Free State samples (range: 5. 20-15.90; mean: 9.83 µg/kg) were lower than the concentrations of ZEN produced by both fungal strains from Limpopo samples (range: 7.80-97.18; mean: 29.88 µg/kg).     Tables 4 and 5 showed the Welch and Brown-Forsythe robust test for the equality of means for AFB1 and AFB2 biosynthesis by A. flavus, as well as ZEN formation by F. oxysporum and F. equiseti in SA smallholder dairy cattle feeds and feed ingredients with re- Figure 2. Minimum, mean, and maximum mycotoxin levels (µg/kg) produced by fungal isolates recovered from SA dairy cow diets during the summer and winter seasons; A1 and A2: AFB 1 produced by A. flavus during summer and winter in Free State and Limpopo; B1 and B2: AFB 2 produced by A. flavus during summer and winter and Free State and Limpopo; C1 and C2: ZEN produced by F. oxysporum and F. equiseti during summer and winter in Free State and Limpopo. Tables 4 and 5 showed the Welch and Brown-Forsythe robust test for the equality of means for AFB 1 and AFB 2 biosynthesis by A. flavus, as well as ZEN formation by F. oxysporum and F. equiseti in SA smallholder dairy cattle feeds and feed ingredients with respect to seasons (summer and winter) and locations (Free State and Limpopo). The results showed that season had a significant effect (p ≤ 0.05) on all the mycotoxins (AFB 1 , AFB 2, and ZEN) formed by the toxigenic fungal strains, while the potential of the fungal isolates to produce mycotoxins in both locations was not statistically different (p > 0.05).

Discussion
Animal feed contamination by microorganisms, especially toxigenic fungal species is a worldwide concern owing to the mycotoxins they produce in agricultural products, which can be harmful to human and animal health [40] with a serious impact on the economy of any nation [21]. Mycotoxin contamination greatly contributes to agricultural product loss in SSA [41]. The prevalence of mycotoxin-producing fungi, particularly, Fusarium and Aspergillus, have previously been documented in feeds destined for smallholder dairy cows in both Free State and Limpopo provinces of SA [30], implying the presence of the mycotoxins recovered and reported herein.
In the present study, a total of 112 Fusarium and Aspergillus strains obtained from smallholder dairy cow diets across different seasons (summer and winter) from Free State and Limpopo provinces of SA were investigated for their mycotoxin-producing potentials. The study also investigated the impact of variation in seasonal and geographical locations on the mycotoxin production capacity of the fungal strains. A. flavus strains analyzed in this present work produced only B-type AFs (AFB 1 and AFB 2 ) but none of the G-types (AFG 1 and AFG 2 ). These findings are in line with those reported by Chilaka et al. [38], Lasrem et al. [42], and Yogendrarajah et al. [43], based on the formation of only AFB 1 and AFB 2 by strains of A. flavus recovered from various agricultural commodities from SA, Sri Lanka, and Tunisia, respectively. It must be emphasized that the levels of AFs produced by the A. flavus strains in this study ranged from 0.22 to 1045.80 µg/kg, wherein 9/24 (38%) of the fungal species produced AF levels above the SA maximum tolerable levels (10 µg/kg) for dairy cattle feeds [5]. We also found that all the A. fumigatus, A. niger, and A. ochraceus strains reported in this work failed to produce any AF analog.
A. niger has been documented as one of the major OTA producers of agricultural products [44,45]. However, OTA production by Aspergillus spp. occurred under different environmental conditions, including temperature, water activity (a w ), and pH [46], as well as the carbon source [47]. Wei et al. [47] confirmed that arabinose was the optimum carbon source for OTA production in A. niger CBS 513.88, followed by glucose and sucrose. Nevertheless, none of the 25 strains of A. niger examined in this study produced OTA. This disparity might be because the A. niger strains lack the genetic potential to synthesize OTA. Susca et al. [48] investigated several OTA-producing and non-producing A. niger strains recovered from various agricultural commodities such as maize, pistachio, walnuts, and cashew nuts from several countries (Turkey, Italy, USA, Brazil, Argentina, China, Greece, Indian, Iran, and Spain), and discovered that the non-OTA-producing strains lacked OTA biosynthetic cluster genes (ota1, ota2, ota3, and ota5), with further genome sequence analysis data revealing OTA gene deletion in the non-OTA-producing A. niger isolates. More so, the culture substrate (CYA) utilized in this study could be linked to the A. niger strains' non-ochratoxigenic nature, as Esteban et al. [49] observed that mean OTA levels produced by A. carbonarius and A. niger are substrate dependent. Our findings regarding OTA production by A. niger are in agreement with that observed by Munitz et al. [50], who found that none of the 19 A. niger strains obtained from Argentinian blueberry produced OTA, similar to the study conducted in Africa by Njobeh et al. [51] where over 17 strains of A. niger isolated from Cameroonian food commodities were non-OTA producers. Our results were also in line with the study performed by Sultan and Magan [52], wherein none of the A. niger strains recovered from peanuts sourced from different geographical regions of Egypt tested positive for mycotoxin. Chilaka et al. [38] also conducted similar research in SA and found that the 10 strains of A. niger isolated from maize, an essential dairy cattle feed ingredient in the country, could not produce OTA. However, this result is unsurprising as earlier reports on OTA contamination of dairy feed in SA revealed that the mycotoxin was not detected in the feed samples [36,53]. A. ochraceus, another notable OTA-producing fungus [38,54], isolated from the feed samples, was not ochratoxigenic, and as conjectured, neither the experimental conditions, including the culture medium and laboratory conditions (temperature, a w , and pH) were ideal for OTA biosynthesis by the fungal species or the isolates lacked the required genes for OTA production. According to Chilaka et al. [38] and Wang et al. [54], the fungal species metabolic profiling relies on the laboratory conditions and the growing medium, while O'Callaghan et al. [55] confirmed that the expression of the biosynthetic genes (polyketide synthase) involved in OTA biosynthesis by A. ochraceus is highly regulated by nutritional cues and pH. Moreover, our findings on OTA production by both A. ochraceus and A. niger contradict those of Rosa et al. [56], who reported that the two fungi isolated from dairy cow feed in the Brazilian state of Rio de Janeiro produced OTA. Adekoya et al. [57] also demonstrated that 87% of A. niger isolates from Nigerian and SA fermented food products tested positive for OTA.
Among the Fusarium spp. tested in this work, only cultures of F. oxysporum and F. equiseti were positive for one of the tested mycotoxins (ZEN). This result concurs with the research conducted on mosquitoes in SA by Phoku et al. [58]. Barros et al. [59] also confirmed ZEN formation by F. equiseti in Argentinean soybean. However, ZEN concentrations reported in this present work were low, ranging from 5.20 to 97.18µg/kg, a range far below the SA acceptable limit (500 µg/kg) in dairy cattle feeds [5]. Nevertheless, earlier reports have indicated that ZEN is more prevalent in Western Europe, North America, and Eastern Europe than in African countries [60]. However, our results disagree with the report of Chilaka et al. [38] wherein none of the F. oxysporum isolated from SA maize tested positive for ZEN. Another important Fusarium toxin, DON, was detected in this study as none of the tested fungal isolates could produce the toxin. According to Neme and Mohammed [61], DON is primarily formed by Fusarium spp., in particular, F. culmorum and F. graminearum. Also, fumonisin B 1 (FB 1 ) and fumonisin B 2 (FB 2 ) could not be confirmed in this study due to the lack of reference materials.
Data on the impact of seasonal and geographical variations on the toxigenic potentials of fungal species invading dairy cattle diets and feed raw materials is scarce worldwide, including in SA. In terms of seasonal variation, significant (p ≤ 0.05) effects were noted on the biosynthesis of AFB 1 and AFB 2 by A. flavus and ZEN by F. oxysporum and F. equiseti, in which higher levels of the mycotoxins were formed by the summer fungal species in comparison to the winter fungal isolates. These findings are in excellent agreement with what Mohale et al. [62] reported on mycotoxin production by toxigenic fungal species associated with stored maize (essential dairy feed ingredients) from various parts of Lesotho, with findings revealing mean concentrations of AFB 1 (73, 338 µg/kg) and AFB 2 (3, 117 µg/kg) produced by A. flavus in 2009/10 cropping season being significantly higher than AFB 1 (821 µg/kg) and AFB 2 (0 µg/kg) produced during 2010/11 cropping season. Interestingly, our investigation, on the contrary, revealed that location had no significant (p > 0.05) effect on the toxigenicity of the fungal isolates. Our results on the effect of location on the fungal mycotoxin profiling agree with those of Yazid et al. [44], who found that agro-ecological location has no influence on AFB 1 formation by A. flavus invading grain corn farms (Kampong Dadong and Rhu Tapai) in Terengganu, Malaysia. The authors emphasized that the mycotoxigenicity of the fungal isolates is more dependent on fungal strain specificity than geographical locations.
The ability of the isolated fungal species to produce different mycotoxins is a concern for the welfare of dairy cattle in the areas, as earlier mycotoxin analysis of similar feeds from the two provinces revealed the presence of all the attendant mycotoxins (AFB 1 , AFB 2 , and ZEN) reported in this study, in addition to OTA, DON, AFG 1, and AGF 2 , which were not produced by the toxigenic microbiota [6]. However, the study confirmed a low incidence of OTA, AFG 1, and AFG 2 [6]. As motioned previously, the failure of the tested fungal isolates to produce these toxins may be linked to the culture medium [38], strain specificity [44], or the absence of mycotoxin cluster genes in the fungal species [55].
The high level of mycotoxins formed by both Fusarium and Aspergillus strains in summer feeds in this work implies that fungal species produce high levels of mycotoxin in summer than in winter. This is likely attributed to climatic differences since toxigenic fungal colonization of crops and their ability to produce mycotoxin is dependent on climatic conditions [7,63]. Temperature, rainfall, and relative humidity are the primary climatic parameters promoting fungal growth and mycotoxin production in food and feed [63]. Even though fungal proliferation and mycotoxin biosynthesis are linked, the ideal relative humidity and temperature for mycotoxin biosynthesis differ based on the fungus and its associated mycotoxins [64]. Although A. flavus spp. can produce AFs under a wide range of temperature conditions, an ideal range for their maximum production is between 25 to 35 • C [65]. At high temperatures, more AFB is synthesized than AFG, but the biosynthesis of both toxins is surmised to be similar at low temperatures [66]. Limpopo and Free State areas of SA are usually hot and humid during summer but cold and mild during winter. Free State summer temperatures range from 15 to 31 • C, and winter temperatures range from 1 to 18 • C, whereas summer temperatures in Limpopo range from 18 to 38 • C and a winter temperature range from 12 to 24 • C. The higher temperature conditions observed during summer than in winter in the two provinces must have been the reason for the increased AF production by A. flavus strains recovered from the summer samples.
It has been demonstrated that relative humidity levels ranging from 88 to 95% promote fungal proliferation and subsequent toxin production in agricultural products [67]. According to Muga et al. [68], 90% relative humidity greatly enhances AF formation in SA-stored maize kernels. These conditions are identical to the ambient climatic parameters often encountered during summer in both Limpopo and Free State areas of SA and hence account for the elevated AFs and ZEN concentrations produced by summer fungal isolates that did not differ significantly (p > 0.05) between the two provinces. The higher mycotoxin concentrations produced by the toxigenic fungal species during summer may also be attributed to post-harvest conditions, such as improper feed management and storage practices [69].

Conclusions
In conclusion, the present study revealed variations in mycotoxin levels produced by toxigenic fungal isolates previously recovered from contaminated smallholder dairy cows' feeds and feed raw materials sourced during two different seasons (summer and winter) and geographical regions (Limpopo and Free State) of SA. The high formation of AFB 1 and AFB 2 by A. flavus strains, as well as ZEN by F. oxysporum and F. equiseti in this study, demonstrate that similar feeds are likely contaminated by attendant mycotoxins. It is equally important to necessitate feed safety awareness-raising programs specifically on fungal and mycotoxin contamination among various smallholder dairy cattle farmers in SA. More so, dairy cattle farmers in both regions under study must be informed of eco-physiological conditions that enhance fungal proliferation and mycotoxin formation in dairy cow feeds. Also, farmers and consumers must be educated on the risk of carry-over of these toxins from feeds to meat, milk, and other dairy products, as well as the associated health implications in animals and humans.
It thus, follows that data generated in this study should lay the foundation for monitoring the mycotoxins in food and feed that can necessitate particular attention, especially for feed during the summer. However, the present results might not accurately reflect the general scenario of the mycoflora toxigenicity in the diary feeds because of the small sample size and a small proportion of evaluated toxigenic fungal isolates. To fully comprehend the impact of seasonal and geographic variations on the toxigenic potential of mycobiota associated with SA smallholder dairy cattle feeds, a broader survey of mycotoxin-producing fungi with a large sample size that includes year-to-year climatic variations and diverse agro-ecological zones are required.

Chemicals and Reagents
Formic acid, acetonitrile, LC-MS grade methanol, and dichloromethane used in this study were procured from Merck (Darmstadt, HE., Germany).

Fungal Strains and Inoculation
Pure fungal strains (n = 112) used in this study, including Aspergillus (83) and Fusarium (29), were recovered from smallholder dairy cow feeds collected from various dairy farms in Limpopo and Free State provinces in SA between summer 2018 and winter 2019. Information on fungal isolation and identification can be found in Adelusi et al. [30]. The fungal isolates were tested for their potential to produce mycotoxins, including AFs (AFB 1 , AFG 1 , AFB 2, and AFG 2 ), DON, OTA, and ZEN by inoculating them on solid CYA agar (Appendix A contains the procedures for agar preparation) treated with streptomycin and chloramphenicol to prevent bacterial growth. Following inoculation, the CYA plates were cultured in the dark for 21 days at 25 • C according to the method of Adekoya et al. [57].

Multi-Mycotoxin Extraction
Extraction of AFs from Aspergillus isolates was performed following the agar plus method described by Adekoya et al. [57] with slight modifications. One gram of pure colonies (together with the medium) was gently removed from the CYA plates and transferred to an amber vial containing 4 mL of dichloromethane. The content was vortexed for about 2 min, left for 60 min, and then filtered into a 1.5 mL amber vail using a Milex syringe filter (0.22 µm). ZEN was extracted from Fusarium isolates using the method reported by Adekoya et al. [57]. Briefly, 10 g of individual isolate with the medium was transferred into a 250 mL conical flask, and 50 mL of acetonitrile: water (60/40, v/v) as extraction solvent was added. The content was agitated for 60 min before being filtered via a Whatman #4 filter paper (Merck, Johannesburg, South Africa), and pH was adjusted to 6.2 ± 0.3 by adding 1 M H 2 SO 4 . After that, the filtrate was extracted thrice using 25 mL of dichloromethane in a 250 mL separating funnel. Following this, acetonitrile (25 mL) was added to the dichloromethane-extracted content and passed over a thin layer of sodium sulfate anhydrous to remove excess moisture. Both Aspergillus and Fusarium extracts were dried using a stream of nitrogen (N 2 ) gas and the content was stored at 4 • C prior to analysis.

LC-MS/MS Apparatus and Condition
The dried extracts were reconstituted in 1500 µL of LCMS-grade methanol (MeOH). Each aliquot (750 µL) was transferred into an amber vial, diluted with an equal volume of dilution solvent that consisted of MeOH and acetonitrile (ACN) (50:50, v/v), and 5 µL was injected into an LC-MS/MS [70]. The detection and quantification of AFs and ZEN produced by the fungal isolates were achieved on a Shimadzu LC-MS/MS 8040 instrument (Shimadzu Corporation, Tokyo, Japan), which comprised an LC-30AD Nexera chromatograph attached to a CTO-20 AC Prominence Column Oven and SIL-30 AC Nexera autosampler. The analytes were separated chromatographically by passing extracts through a RaptorTM ARC-18 (2.7 UM, 2.1 × 100 mm) column (Restek Corporation, Bellefonte, PA, USA). Solvents A (0.1% formic acid in deionized water) and B [0.1% formic acid in ACN and MeOH (50:50, v/v)] were used as mobile phases.
Mobile phases A and B were delivered at 0.2 mL/min constant flow rate and 400 bar maximum pressure. The gradient elution operation started with 0.1 min at 10% solvent B, ramped to 95% B in 8.4 min, and maintained for 3 min. The column was re-equilibrated for 1 min using 10% solvent B before the start of the next run, which lasted for 4.5 min, bringing the entire run time to 17 min. After chromatographic separation, all toxins were identified and quantified utilizing a Shimadzu 8040 triple-quadrupole MS 8040 (Shimadzu Corporation, Kyoto, Japan) operated in a positive ionization mode using an electron spray ionization (ESI+) source. The desolvation line (DL) temperature was set at 250 • C, the interface nebulizing gas flow rate was 3 L/min, the heat block temperature was set at 400 • C, and the drying gas flow rate was set at 15 L/min. Finally, data were acquired using the MRM method operated using optimized MS conditions for the analytes ( Table 6). The obtained data were processed and visualized by means of Shimadzu LabSolutions software.

Method Validation
Several European Commission criteria [39] were evaluated to determine the method performance in this study. A multi-mycotoxin analytical technique used was validated using blank CYA media. The evaluated parameters included matrix effect, linearity, recovery, LOD, and LOQ. Both matrix-matched and neat standard calibration curves were constructed to investigate matrix effects on the analyzed samples. The levels of mycotoxins produced by the toxigenic fungal isolates were quantified using matrix-matched calibration (MMC) curves. Furthermore, the linearity was determined by spiking a blank CYA medium at seven (7) concentrations using the matrix-matched calibration curve. The calibration curves were generated by plotting the analyte peak areas (y) against the corresponding concentrations (x). The calibration curve was fitted using linear regression.
Retention times (RT) and coefficient of determination (R 2 ) were obtained for individual mycotoxins. Both LOD and LOQ for each mycotoxin were also estimated using MMC. The LODs were calculated as 3.3 times the ratio of the residual standard deviation to the slope (Equation (1)), and LOQs were calculated as the concentration equivalent to 10 times the ratio of the residual standard deviation to the slope (Equation (2)) according to Shrivastava and Gupta [71]. All detected mycotoxins (AFB 1 , AFB 2, and ZEN) were quantified by comparing their peak area on the calibration curve to the corresponding mycotoxin standards. In addition, the apparent recovery for individual fungal toxins was determined by spiking blank (CYA) media at 50 µg/kg (low) and 100 µg/kg (high) and comparing the observed concentrations and spiked concentrations following extraction as

Data Analysis
The IBM Statistical Package for SPSS version 27 (IBM Corp., Armonk, NY, USA, 2021) was employed to analyze the data. A one-way analysis of variance (ANOVA) was used to evaluate the possible differences among the mycotoxin levels produced based on season and location. The Levene test was used to assess the equality of variances among the mycotoxin concentrations produced across the two seasons and locations. Lastly, Welch and Brown-robust Forsythe's test for equality of means was performed wherein the equality of variances could not be established. Differences were considered significant if p was ≤0.05.