A Study on the Spatial Distribution of Zearalenone and Deoxynivalenol in Oat Bran

TL;DR

This study examines how two mycotoxins, zearalenone and deoxynivalenol, are distributed in stored oat bran and how air exposure affects fungal communities and toxin levels.

Contribution

The study reveals spatial toxin distribution patterns and fungal associations in oat bran, offering insights for preventing mold contamination during storage.

Findings

Zearalenone and deoxynivalenol levels were significantly lower in oat bran with higher air exposure.

Fusarium species were strongly correlated with toxin presence, while certain fungi showed synergistic or antagonistic relationships.

Toxin distribution in oat bran was heterogeneous and influenced by air exposure and fungal community composition.

Abstract

Zearalenone (ZEN) and Deoxynivalenol (DON) are common Fusarium toxins that are found worldwide in contaminated wheat, corn, oats, and other foods. This study investigated the spatial distribution of ZEN and DON within bagged oat bran and the relationships among fungal taxa. A total of 168 oat bran bags arranged in a three-dimensional space (X = 4, Y = 6, Z = 7) were tested for ZEN and DON concentrations via Enzyme-linked Immunosorbent Assay (ELISA) and fungal communities were analyzed by Internal Transcribed Spacer (ITS) sequencing. Samples were grouped by air-exposed surfaces: G0 (no exposure, n = 48), G1 (one exposed surface, n = 80), G2 (two or three exposed surfaces, n = 40). Results showed strong positive correlations between ZEN and DON spatial distributions (r = 0.691~0.930), with G2 having significantly lower toxin levels than G0 and G1 (p < 0.05). Fusarium spp. (e.g., F. aethiopicum, F. pseudonygamai, and F. fujikuroi) were positively correlated with ZEN and DON (p < 0.05), indicating that they are the primary producers of these mycotoxins. Talaromyces (T. funiculosus and T. stollii) and Sarocladium (S. kiliense and S. strictum) were positively correlated with ZEN, DON, and Fusarium spp., while the yeasts D. hungarica, V. victoriae, and H. sinensis exhibited a negative association with those (p < 0.05). Overall, the distribution of ZEN and DON in bagged oat bran was heterogeneous in three-dimensional space, and the distribution pattern was related to air exposure. The extent of air exposure influenced the composition of the fungal community, and the taxa correlating with Fusarium spp. showed potential synergistic or antagonistic associations, collectively influencing the accumulation of mycotoxins. This study provides a reference basis for the prevention of mold contamination during the stacked bag storage of feedstuffs.

Figures

Full text

1. Introduction

Zearalenone (ZEN) and deoxynivalenol (DON) are toxic secondary metabolites produced by Fusarium species [1]. They pose a potential risk to food and feed safety due to their pervasive contamination of staple grains such as maize, wheat, barley, oats, and sorghum [2]. While F. graminearum and F. culmorum are the predominant producers [3], recent studies have identified F. fujikuroi [4], F. aethiopicum [5], and F. asiaticum [6] as also capable of producing ZEN and DON. ZEN is an estrogenic mycotoxin causing reproductive issues across multiple species, including humans [7,8] and livestock [9,10,11,12]. DON is an enterotoxin primarily associated with inducing emesis, diarrhea, disruptions in intestinal microbiota, and intestinal lesions in animals [13,14]. Notably, the synergistic toxicity of ZEN and DON has emerged as a critical research focus due to their frequent co-occurrence in food and feed. This combined toxicity exacerbates adverse impacts on the intestine [15], reproductive system [9], liver, and kidneys [16], while also increasing immunotoxicity and cytotoxicity [17].

The stochastic nature of fungal infection in the field leads to a heterogeneous distribution of mycotoxins within agricultural commodities, a factor that is commonly overlooked during sampling [18]. Consistent with this, sampling constitutes the greatest source of error in mycotoxin testing, prompting recommendations to increase the number of samples to reduce variability [19]. Furthermore, geostatistical methods have been employed to characterize the two-dimensional spatial distribution of DON in bulk storage, revealing a heterogeneous spatial pattern that varied significantly with depth [20,21]. Similarly, Qi et al. found that storage depth within rice grains significantly influences the spatial distribution of both fungi and mycotoxins [22]. Although these studies have elucidated the spatial distribution of mycotoxins, there remains a lack of research examining mycotoxin contamination of materials in three dimensions.

The three-dimensional spatial heterogeneity of mycotoxins in stored bulk materials arises from the complex interplay of multiple factors: initial stochastic fungal contamination in the field, subsequent cross-contamination between adjacent materials, and uneven storage conditions (e.g., gradients in water activity and temperature). Warm and humid environments facilitate the proliferation of toxigenic fungi, including Fusarium species, which directly results in increased levels of mycotoxins such as DON [23]. Katsurayama et al. reported a higher incidence of fungal contamination, primarily involving Aspergillus and Fusarium species, in rice cultivated in wetlands compared to that grown in dryland environments [24]. In addition, mycotoxin production can be modulated by interactions within the fungal community. For instance, A. flavus is capable of inhibiting the production of mycotoxins by both F. graminearum and F. verticillioides [25]. Therefore, understanding the interactions among mycotoxins, environmental factors, and fungal communities from a three-dimensional spatial perspective provides critical guidance for preventing mycotoxin contamination in stored materials.

Recent studies on the spatial distribution of mycotoxins have predominantly focused on two dimensions or specific depths, resulting in a limited understanding of the full three-dimensional spatial heterogeneity of mycotoxins in stored bulk materials. This study introduces a novel three-dimensional spatial analysis of the correlation and distribution of ZEN and DON contamination in bagged oat bran. Additionally, ITS-based fungal profiling was employed to examine the relationship between mycotoxin contamination and fungal infection in oat bran, providing a scientifically informed reference for material stacking and sample collection.

2. Materials and Methods

2.1. Sample Collection

Oat bran samples were obtained from a local feed warehouse. The material, packaged in commercial bags, was stored on four pallets placed centrally in the warehouse for approximately 12 months. The schematic representation of the placement is depicted in Figure 1a. Each pallet held 42 bags (6 bags per layer × 7 layers), resulting in a total of 168 bags. A composite sample from each bag was obtained using a feed sampler. To account for potential vertical stratification within the bags, a diagonal sampling method was employed to cover the top, middle, and bottom positions. Three subsamples (~100 g) were then thoroughly mixed to form a single 300 g composite sample per bag. All samples were immediately stored at −80 °C until subsequent mycotoxin and microbial analysis.

2.2. ZEN and DON Measurement

The concentrations of ZEN, DON, and AFB_1_ were measured by enzyme-linked immunosorbent assay (ELISA) kits. Sample preparation and analysis were performed using ZEN Rapid Detection Kit (HEMI796), DON Rapid Detection Kit (HEM1896), and AFB_1_ Rapid Detection Kit (HEM0196) (Huaan Magnech Bio-Technology Co., Ltd., Beijing, China). The limit of detections (LODs) was 1.05 μg/kg for AFB_1_, 20 μg/kg for ZEN, and 250 μg/kg for DON, respectively. For all samples whose concentrations exceeded the linear range of the standard curve, appropriate gradient dilution was performed during reanalysis to ensure that the final measured values fell within the linear range of the kit, thereby guaranteeing data accuracy.

2.3. ITS High-Throughput Sequencing

These 168 samples were classified into three groups (G0, G1, and G2) based on the number of air-exposed surfaces: G0 (no exposure, n = 48), G1 (one exposed surface, n = 80), and G2 (two or three exposed surfaces, n = 40), of which 160 samples (G0:45, G1:75, G2:40) were subsequently used for Internal Transcribed Spacer (ITS) sequencing. Regarding the reason for including both two- and three-air-exposed surfaces in the G2 group: Only four samples with three exposed surfaces were obtained. To ensure the statistical power and reliability of the results, the samples with two exposed surfaces (n = 36) and those with three exposed surfaces (n = 4) were combined into a single group, designated as G2 (n = 40).

Total genomic DNA was extracted from oat bran samples using the TGuide S96 Magnetic Soil/Stool DNA Kit (Tiangen Biotech Co., Ltd., Beijing, China) according to the manufacturer’s instructions. The internal transcribed spacer (ITS) region of the 5.8S, 18S, and 28S rRNA genes was amplified, and the PCR products were visualized on an agarose gel and purified using the Omega DNA purification kit (Omega Inc., Norcross, GA, USA). The purified PCR products were collected, and the paired ends (2 × 250 bp) were sequenced on the Illumina Novaseq 6000 platform.

Clean reads were then conducted on feature classification to output ASVs (amplicon sequence variants). Taxonomy annotation of the ASVs was performed based on the Naive Bayes classifier in QIIME2 using the SILVA database (release 138.1) with a confidence threshold of 70%. Alpha diversity was performed to identify the complexity of species diversity of each sample, utilizing QIIME2 software. Beta diversity calculations were analyzed by Non-metric Multidimensional Scaling (NMDS) to assess the diversity in samples for species complexity. One-way analysis of variance was used to compare bacterial abundance and diversity. Redundancy Analysis (RDA) and Relevance Network based on the Pearson correlation coefficient were used to explore the relationship of environmental factors and fungal taxa. Linear Discriminant Analysis (LDA) and effect size (LEfSe) coupled with random forest analysis were applied to evaluate the differentially abundant taxa. The online platform BMKCloud (https://www.biocloud.net) (accessed on 22 May 2025) was used to analyze the sequencing data.

2.4. Statistical Analysis

A one–way analysis of variance (ANOVA) for mycotoxin concentrations in each stratum was performed using IBM SPSS v19.0 (SPSS, Inc., Chicago, IL, USA), Tukey’s HSD multiple analysis of variance was used to compare the means among groups, p < 0.05 was considered statistically significant, 0.05 ≤ p < 0.10 considered a trend of statistical significance, and the relationship between ZEN and DON was expressed using the Pearson correlation coefficient. Regarding the multiple correlation analysis, the false discovery rate (FDR) correction (Benjamini–Hochberg method) was applied to the p-values, and only correlations that remained significant after FDR correction have been retained. Heatmaps of the spatial distribution of ZEN and DON were drafted using MATLAB R2022b. Data are shown as “mean ± standard deviation”.

3. Results

3.1. Descriptive Statistics of ZEN and DON Concentrations in 168 Bags of Oat Bran

The distribution of both ZEN and DON across the 168 bags of oat bran was highly heterogeneous, as indicated by high coefficients of variation (38.74% for ZEN and 54.88% for DON; Table 1). The analysis of the normal distribution showed that ZEN exhibited a non-normal distribution (p < 0.05), whereas DON demonstrated a normal distribution (p > 0.05) (Figure S1). As shown in Table 1, the concentration of ZEN ranged from 456.40 to 61,013.70 μg/kg, with a higher mean of 33,752.03 μg/kg. In comparison, the concentration of DON ranged from 1465.07 to 42,196.98 μg/kg, with a lower mean of 16,207.78 μg/kg. The 1st quartile and 3rd quartile values of ZEN were 28,571.02 and 42,452.30 μg/kg, respectively, while the DON values were 9342.35 and 22,291.15 μg/kg, respectively. The median and standard deviation values of ZEN were 33,545.00 and 13,077.00 μg/kg, respectively, while those for DON were 15,619.81 and 8894.73 μg/kg, respectively.

3.2. Correlation of ZEN and DON Concentration

The concentrations of DON and ZEN were represented using three-dimensional heatmaps based on spatial coordinates. These visualizations revealed analogous distribution patterns for both mycotoxins across all layers, with spatial coincidence of the peaks and valleys of their concentrations (Figure 1). Consistent with this observation, the spatial distributions of ZEN and DON exhibited a strong positive correlation, with consistently high Pearson correlation coefficients across all three dimensions (X: 0.760–0.901; Y: 0.731–0.891; Z: 0.691–0.930; Table 2, Table 3 and Table 4).

3.3. Spatial Distribution Characteristics of ZEN and DON Concentrations

According to Table 2, Table 3 and Table 4, there is no significant difference in the concentration of ZEN across layers of the X dimension (p > 0.05). However, in the Y and Z dimensions, the concentration of ZEN across layers showed a significant difference (p < 0.05). Similarly, no significant difference in DON concentration was observed across layers in the X-dimension (p > 0.05), while it tended to be significantly different in the Y and Z dimensions (0.05 ≤ p < 0.10). Furthermore, based on a preliminary consideration of ventilation effects, the concentrations of ZEN and DON at air-exposed surfaces (X1, X4, Y1, Y6, Z7) were relatively low (Table 2, Table 3 and Table 4), and this result was also supported by the heatmap (Figure 1). Specifically, in the Y dimension, compared with the Y4, the concentration of ZEN in the Y1 was significantly lower (p < 0.05); In the Z dimension, compared to Z3, the concentration of ZEN in the Z2 significant lower (p < 0.05).

To further explore the relationship between the number of air-exposed surfaces and the distribution of ZEN and DON, three groups were classified: G0 (no exposure), G1 (one exposed surface), and G2 (two or three exposed surfaces) (Figure 2a). As shown in Figure 2b,c, the concentrations of both ZEN and DON decreased significantly with an increasing number of air-exposed surfaces (G0 > G1 > G2), with G2 levels significantly lower than those of G0 and G1 (p < 0.05). This pattern confirms that the number of air-exposed surfaces is a key determinant of the heterogeneous three-dimensional distribution of ZEN and DON.

3.4. High-Throughput ITS Sequencing Analysis

3.4.1. Community Structures of Oat Bran Fungi

To elucidate the air-exposed distribution pattern of ZEN and DON, ITS sequencing was performed on oat bran samples from groups G0 (n = 45), G1 (n = 75), and G2 (n = 40). The ACE, Simpson, Shannon, and Chao1 indices showed no significant differences in α-diversity among the groups (p > 0.05) (Figure 3a). NMDS analysis revealed a significant separation in fungal community composition and structure among the G0, G1, and G2 (stress < 0.2) (Figure 3b). The Venn diagram shows the shared and unique ASVs among the G0, G1, and G2 (Figure 3c). Fungal communities were dominated by Ascomycota (85.95%) and Basidiomycota (3.28%) at the phylum level (Figure 3d). At the genus level, the most abundant taxa were Fusarium (19.01%), Xeromyces (14.16%), Sarocladium (13.12%), and Alternaria (12.41%) (Figure 3e). At the species level, dominant species included X. bisporus (14.39%), A. hordeicola (12.06%), S. kiliense (11.78%), and F. aethiopicum (11.29%) (Figure 3f).

3.4.2. Correlation Analysis of ZEN and DON with Fungi

To further explore the effects of environmental factors—ZEN, DON, and Air (defined by the number of air-exposed surfaces)—on fungal communities, their correlations with the abundance of fungal phyla, genera, and species were analyzed (Figure 4a–d). The analysis revealed a negative correlation between Air and the concentrations of ZEN and DON (Figure 4a,b,d), supporting the observed order of ZEN and DON concentrations as G0 > G1 > G2 (Figure 2b,c). This further indicated that the distribution of both mycotoxins is related to the extent of air exposure. Besides, Air had a stronger effect than ZEN and DON on the fungal taxa at the phylum level, and the dominant phyla Ascomycota and Basidiomycota had no significant correlation with DON and ZEN (p > 0.05) (Figure 4a,c). ZEN and DON had a stronger effect than Air on the fungal taxa at the genus and species level, and the dominant genera and species Fusarium (including F. aethiopicum) and Sarocladium (including S. kiliense) showed a positive correlation with ZEN and DON, while Xeromyces (including X. bisporus) and Alternaria (including A. hordeicola) showed a negative correlation with ZEN and DON (p < 0.05) (Figure 4b–d). Since Fusarium is the known producer of ZEN and DON, correlation analysis between the concentration of both mycotoxins and the abundance of Fusarium species was performed (Figure 4e,f). The results showed that F. aethiopicum, F. pseudonygamai, and F. fujikuroi were dominant Fusarium species and exhibited a significantly positive correlation with ZEN and DON (p < 0.05), suggesting that they may likely be the main sources of ZEN and DON contamination in oat bran.

To further analyze the differential fungal genera and species associated with air exposure, LefSe analysis showed that g_Ramaria, g_Geoglossum, g_Sebacinales, g_Thozetella, g_Diaporthe, and g_Scleroderma were significantly enriched in G2 (LDA score > 2.0), while g_Hannaella, g_Clavispora, g_Cornuvesica, g_Geotrichum, g_Gibellulopsis, and g_Talaromyces were significantly enriched in G0 (LDA score > 2.0) (Figure 5a). Among them, the abundance of g_Geotrichum, g_Gibellulopsis, and g_Talaromyces was positively correlated with ZEN, DON, and the abundance of g_Fusarium, F. aethiopicum, F. pseudonygamai, and F. fujikuroi (p < 0.05). In contrast, the abundance of g_Hannaella was negatively correlated with ZEN, DON, and the abundance of F. pseudonygamai and F. fujikuroi (p < 0.05) (Figure 5c). Further analysis of the species level of g_Hannaella showed that H. sinensis had the highest abundance and was negatively correlated with ZEN, DON, and the abundance of F. pseudonygamai and F. fujikuroi (p < 0.05) (Figure 5d,e). Random forest analysis identified F. aethiopicum, G. simoni, C. acuminata, S. strictum, T. funiculosus, T. stollii, and S. kiliense as key species exerting a significant influence on the group differences. These species exhibited positive correlations with ZEN, DON, and the abundance of g_Fusarium, F. aethiopicum, F. pseudonygamai, and F. fujikuroi, whereas D. hungarica demonstrated a negative correlation with the same variables (p < 0.05) (Figure 5b,f).

In addition, the fungal network correlation analysis showed that Xermyces (including X. bisporus) and Aspergillus (including A. penicillioides and A. vitricola) had higher abundance in the G2 and G1 groups, respectively, and were negatively correlated with other genera and species (p < 0.05) (Figure 3e,f and Figure S2). The abundance of Vishniacozyma (including V. victoriae) was negatively correlated with ZEN, DON, and the abundance of Fusarium (F. aethiopicum, F. pseudonygamai, and F. fujikuroi) (p < 0.05) (Figure 4b–d and Figure S2).

4. Discussion

This study analyzed neatly stacked and bagged oat bran obtained from a local feed mill. Preliminary screening for common mycotoxins (including AFB_1_, ZEN, and DON) revealed that ZEN and DON were the predominant contaminants at high concentrations. Consequently, these two mycotoxins were selected for detailed spatial analysis.

The frequent co-occurrence of ZEN and DON in contaminated agricultural commodities is well-documented. This is substantiated by regional studies, such as the analysis conducted on 1381 rice samples, which identified ZEN and DON as the most prevalent mycotoxin combination [4]. Additionally, a global analysis of 74,821 feed samples confirmed a positive correlation between ZEN and DON concentrations in major grains, including maize and wheat [26]. These findings are consistent with the present study, which demonstrated a strong correlation between the three-dimensional spatial distributions of ZEN and DON in oat bran. The spatial heatmap and correlation analysis revealed analogous distribution patterns across all layers for ZEN and DON, with their peaks and valleys largely overlapping. This finding is in accordance with expectations, as both ZEN and DON are produced by the Fusarium spp., resulting in the frequent co-occurrence of these mycotoxins in similar locations [2,3]. Furthermore, Casado et al. conducted an analysis using 150 samples obtained from bulk maize beds within a three-dimensional space (X = 10, Y = 5, Z = 3) to examine the correlation between FB_1_ and FB_2_. The findings indicated that the distribution of FB_1_ and FB_2_ across the layers exhibited a similar trend [27]. Conversely, a separate study found no correlation between the presence of DON and ochratoxin A (OTA) in wheat within a two-dimensional space (X = 10, Y = 5) [20]. This discrepancy can be attributed to the production of distinct mycotoxins by various toxigenic fungi. Specifically, FB_1_ and FB_2_ are produced by F. verticillioides, while DON and OTA are produced by Fusarium spp. and Aspergillus spp., respectively [28].

The heterogeneous spatial distribution of mycotoxins within bulk commodities presents a significant challenge for conducting accurate risk assessments. Cheng et al. compared the spatial clustering of aflatoxins, the sampling strategies (random and systematic), and the number of samples [29]. They found that the accuracy of the results was primarily affected by the clustering of aflatoxins in three-dimensional space. Casado et al. analyzed a 26-ton truckload of wheat and revealed that DON exhibited a markedly heterogeneous distribution in two-dimensional space [20]. Johansson et al. studied AFB_1_ and FB_1_ levels in damaged corn, whole kernels, and other materials, discovering that both mycotoxins mainly accumulated in damaged maize kernels and settled at the bottom of the commodities [30]. The results of the present study align with previous findings, demonstrating that ZEN and DON contamination in oat bran exhibits a heterogeneous distribution in three-dimensional space. This suggests that when sampling materials, it is crucial to consider not only the quantity of samples but also their spatial distribution to ensure the collection of the most representative samples. However, the heterogeneous distribution of mycotoxins within a material stack is primarily driven by gradients in water activity and temperature, which are directly influenced by differential contact with air [31,32]. A study on wheat in bottom-ventilated cylinders found that DON was significantly enriched in the upper peripheral regions, where poor ventilation led to higher humidity [33]. Qi et al. conducted a study on the variations in fungal communities during rice storage and demonstrated that the lower layers, which were better ventilated, exhibited reduced contamination by toxigenic fungi [22]. The present study determined that the concentrations of ZEN and DON followed the order (G0 > G1 > G2) and exhibited a negative correlation with the number of air-exposed surfaces. Therefore, improved ventilation can effectively mitigate the accumulation of mycotoxins during material storage. A notable limitation of this study is the lack of direct measurement of environmental parameters, such as relative humidity, temperature, and airflow, during the storage period. Fungal growth and mycotoxin distribution may be substantially influenced by these unmonitored variables.

Mycotoxin contamination is highly associated with the species and abundance of infecting toxigenic fungi. Generally, ZEN and DON have been reported to be produced mainly by F. graminearum and F. culmorum [2,3]. Notably, recent studies have highlighted the significant pathogenicity of other species: F. pseudonygamai, a pathogen of pearl millet (Pennisetum glaucum) [34], and F. fujikuroi, a major pathogen of rice [4,35], both causing substantial yield losses. In the present study, F. graminearum and F. culmorum were not identified in oat bran samples, whereas the F. aethiopicum, F. pseudonygamai, and F. fujikuroi were identified with higher relative abundance. This suggests that these three Fusarium species may play a decisive role in the production of ZEN and DON in moldy oat bran. Furthermore, the co-occurrence of multiple mycotoxins, which have distinct structures and molecular targets, may result in additive or synergistic toxicological effects [31]. Numerous studies have identified the Ascomycota genera g_Fusarium, g_Sarocladium, g_Alternaria, and g_Aspergillus as the predominant fungal taxa in stored materials [36,37,38]. Among these, specific strains of g_Sarocladium (e.g., S. strictum, S. zeae) can synthesize emerging mycotoxins such as beauvericin (BEA) and enniatin B_1_ (ENN B_1_), while g_Alternaria can produce potentially carcinogenic mycotoxins, such as alternariol (AOH) and tenuazonic acid (TeA) [39,40]. The F. fujikuroi species complex (FFSC), known causal agents of post-flowering stalk rot (PFSR) in maize, recent studies have identified Sarocladium species (including S. kiliense and S. zeae) as pathogens of PFSR [41,42,43]. Plant-pathogenic fungi release enzymes like β-glucosidase, xylanase, and pectinase to degrade the plant cell wall’s cellulose, hemicellulose, and pectin, promoting infection [34,43,44]. Both g_Sarocladium and g_Fusarium possess high cellulase and β-glucosidase activities, thereby exacerbating their coordinated infection [45]. Combined with the findings of this study, these results suggest that Sarocladium (including S. kiliense and S. strictum) may share a a compatible ecological niche and cooperative colonization with Fusarium spp. However, a negative correlation was observed between g_Fusarium and g_Alternaria, which is consistent with previous reports [25]. This pattern can be attributed to their high niche overlap in key environmental parameters such as water activity, temperature, and C-source utilization. A study on the interaction between g_Fusarium, g_Alternaria, and their toxins demonstrated that TeA suppressed the growth of F. graminearum and F. culmorum, resulting in decreased DON. Conversely, both Fusarium strains were capable of degrading AOH. Despite consistent water activity and temperature, the findings indicate complex interactions between phytopathogenic fungi and their mycotoxins [46].

The extent of air exposure influenced the composition of the fungal community, likely due to the drier conditions prevalent in areas with ventilation. It is noteworthy that Xeromyces (X. bisporus) [47] and Aspergillus (A. vitricola and A. penicilioides) [48,49], both characterized by extreme xerophilicity, were observed in higher abundance in G2 and G1, respectively. These three fungal species showed a negative correlation with the majority of other fungal species, a phenomenon that can be attributed to the more favorable ventilation and drier conditions present in G2 and G1. This is in contrast to the preference of most other fungi for warmer and more humid environments [33]. The synthesis and accumulation of mycotoxins in stored commodities are intricately regulated by complex interactions within the fungal community. The g_Gibellulopsis and g_Fusarium co-occur in the soil of American ginseng with decayed roots [50]; T. funculosus can cause maize ear rot [51], while T. stollii has been reported to be able to cause pineapple fruitlet core rot disease with Fusarium ananatum [52]. The above research findings, together with the results of the present study, suggest that toxigenic fungi may accelerate mycotoxin production through synergistic interactions, though the exact mechanisms of these interactions require further investigation. In addition to cooperative interactions, complex fungal network environments also exhibit competitive interactions among fungal taxa. Both studies have demonstrated that V. victoriae functions as an effective biocontrol agent against the invasion of Botrytis cinerea through the utilization of organic acids and energy derived from the kiwifruit [53] and pears [54]. Zhu et al. have shown that the g_Vishniacozyma plays a significant role in inhibiting and eliminating other fungal genera during the grape ripening process [55]. Similarly, g_Dioszegia (D. hansenii and D. hungarica), as a potential antagonist yeast, reduced strawberry gray mold incidence by 82% to 86% [56]. In addition, g_Hannaella (especially H. sinensis) has been reported to produce volatile organic compounds that can inhibit the growth of A. flavus, thereby reducing aflatoxin concentrations in maize and cereals [57,58]. Similar studies have found that H. sinensis is also effective in controlling rot in rice and apples [59,60,61]. Janakiev et al. studied Fusarium control in pear and found that the autochthonous antagonistic yeasts Hannaella spp. had a significant effect on the growth of Fusarium spp. from 42.3% to 70.2% [62]. The findings presented above, in conjunction with the results of the current study, indicate that H. sinensis, V. victoriae and D. hungarica may possess the capacity to inhibit the production of ZEN and DON by Fusarium spp. This competitive interactions might be a powerful biological control agent of the biology itself against Fusarium spp. and other mold infections.

Based on the research findings and limitations previously discussed, several promising directions for future investigation deserve exploration. Firstly, to accurately elucidate the environmental drivers underlying the observed three-dimensional gradients of mycotoxins, real-time monitoring systems should be integrated into storage trials to capture dynamic fluctuations in temperature, humidity, and airflow. Concurrently, optimized, spatially-informed sampling protocols should be developed and tested based on the identified contamination gradients to enhance the accuracy and representativeness of risk assessment. Finally, given the complex synergistic and antagonistic interactions among fungal taxa suggested by this study (e.g., Fusarium, Sarocladium, Hannaella), controlled co-culture experiments coupled with metabolomic analysis are necessary to clarify the underlying chemical and ecological mechanisms.

5. Conclusions

The three-dimensional spatial distribution of ZEN and DON in oat bran was heterogeneous, and a spatial correlation was observed between these two mycotoxins. Furthermore, the distribution patterns of ZEN and DON were associated with the number of air-exposed surfaces, with mycotoxin concentrations being lower in areas with superior ventilation. In moldy oat bran, ZEN and DON were strongly associated with Fusarium spp. (F. aethiopicum, F. pseudonygamai, and F. fujikuroi). Correlation patterns indicated frequent co-occurrence of Talaromyces (T. funiculosus and T. stollii) and Sarocladium (S. kiliense and S. strictum) with Fusarium spp., while several yeasts, including H. sinensis, D. hungarica, and V. victoriae correlated with reduced mycotoxin accumulation. In conclusion, this study elucidates the three-dimensional spatial distribution of mycotoxins within materials and provides the scientific basis for the optimum sampling method and rational stacking storage of materials.