Mycotoxin Contamination of Wild Plants in Agricultural Landscapes: Seasonal Dynamics and the Underestimated Role of Woody Species

1. Introduction

Among the most relevant mycotoxins are aflatoxin B₁ (AFB1), deoxynivalenol (DON), and zearalenone (ZEN). These toxins, produced mainly by Aspergillus and Fusarium species, are widely detected in cereals and other agricultural products and constitute significant toxic risks, including carcinogenic, immunotoxic, and endocrine-disrupting effects [1,2,3,4,5]. Compared to cultivated crops, quantitative data on mycotoxin occurrence in wild and semi-natural plant species remain limited, particularly for vegetative organs and seasonal dynamics. Yet, these plants may serve as dietary components for wildlife and livestock and may act as reservoirs for toxigenic fungi. Nichea et al. [6] reported detectable levels of ZEN, T-2, and HT-2 toxins in wild grasses from Argentine peatlands. Kolackova et al. [7] showed that species-rich grassland communities display greater fungal colonization and a higher risk of mycotoxin contamination than monocultures. Several studies have shown that weeds and wild plants can serve as alternative hosts for toxigenic fungi. Weeds and non-crop vegetation can act as alternative hosts for toxigenic fungi, including Fusarium species capable of producing DON and ZEN, therefore serving as potential inoculum sources for adjacent crops [8,9,10]. Spores originating from infected vegetation may reach crops via wind, rain splash, soil, insects, or vertebrates, consequently causing crop infection and mycotoxin contamination [9,11]. Although mycotoxin concentrations in wild plants are often lower than in cultivated hosts, Fusarium species isolated from non-crop vegetation can produce substantial amounts of DON and ZEN under favorable conditions [8,9]. This evidence indicates that semi-natural vegetation can serve as a reservoir of toxigenic fungi causing mycotoxin contamination in agricultural systems. Semi-natural vegetation at field margins maintains varied plant, fungal, and animal communities. These habitats often contain herbaceous and woody plant species consumed by large herbivores such as roe deer (Capreolus capreolus) and fallow deer (Dama dama), as well as by small mammals including voles (Microtus spp.) and field mice (Apodemus spp.) [12,13,14,15]. These animals may ingest contaminated plant material, thereby exposing them to mycotoxins. Small mammals in field margins are important prey for avian and mammalian predators like owls, raptors, and mesocarnivores. This may facilitate the transfer of mycotoxins across trophic levels [16,17]. Lakatos et al. [18] reported the presence of mycotoxins in fallow deer tissues, denoting potential exposure of wildlife to environmental mycotoxins. Recent studies also report that multiple mycotoxins accumulate in predator tissues, supporting their transfer through trophic levels [19].

The impact of climate change underscores the importance of studying mycotoxins in natural vegetation, particularly their distribution across plant tissues and organs. Changes in temperature, precipitation, and drought can reshape fungal communities and encourage the spread of toxigenic species [20,21,22]. These dynamics are likely to be most pronounced in the Carpathian Basin, where warming and drought surpass the European average, increasing the risk of mycotoxin production in agriculture and semi-natural ecosystems [23,24]. Similar climate-driven shifts in Fusarium species composition and associated mycotoxin risks have already been predicted for Northern Europe [25]. While mycotoxins in crops and livestock are increasingly understood, patterns in wild plants—particularly in vegetative organs, reproductive tissues, and seasonal dynamics—are still underexplored. This study intends to characterize the occurrence and distribution of three major mycotoxins—Aflatoxin B₁ (AFB1), deoxynivalenol (DON), and zearalenone (ZEN)—in wild plant species growing near agricultural fields. We examined seasonal variation in contamination, assessed the co-occurrence of multiple mycotoxins, compared organ-specific (vegetative versus reproductive) accumulation and growth-form differences, focusing on woody versus monocotyledonous and dicotyledonous herbaceous species (“grasses” and “herbs”).

We hypothesized that

(i) A given plant species or plant organ can be contaminated by more than one type of mycotoxin simultaneously.

(ii) mycotoxin contamination shows seasonal variation in wild plants growing adjacent to agricultural fields;

(iii) mycotoxin contamination differs among plant organs, with reproductive and vegetative tissues manifesting distinct contamination profiles;

(iv) mycotoxin contamination varies among plant growth-form with woody and herbaceous species showing contrasting contamination profiles.

2. Materials and Methods

2.1. Field Sampling

Our fieldwork aimed to collect plant samples that are likely to serve as food for wildlife, especially European roe deer (Capreolus capreolus) and brown hares (Lepus europaeus) in the vicinity of agricultural areas. Sampling took place in an agricultural area on the Hungarian Plain in the Jászság region (Figure1). The site is a Natura 2000 nature conservation area characterized by saline soil. The climate is continental. The yearly average precipitation is 500–650 mm with unequal dispersion; drought occurs frequently, especially in summer. The average annual temperature varies from 9.5 to 11.5 °C, but summer is hot. The vegetation showed a typical character for the region, dominated by a mosaic of arable lands and grasslands with bushes at the edges and small poplar (Populus spp) and black locust (Robinia pseudoacacia) plantations. The sampling was carried out three times in the vegetation period on 16-18 July, 27-28 August, and 28-29 October 2024 along a 1.7-kilometer-long transect (coordinates N 47.559967°, E 19.933996° and N 47.561963°, E 19.911739°). Plant parts (stems, leaves, flowers, fruit) were collected from several individuals of the given plant species from different patches. The samples were desiccated at 95 oC for 24 hours, ground, and stored at -18 oC until laboratory analysis.

Figure 1. Location of the sampling field. Background imagery from Google Earth (© Google).

Figure 1. Location of the sampling field. Background imagery from Google Earth (© Google).

2.2. Mycotoxin Analysis

The concentration of zearalenone (ZEN) was determined using a competitive enzyme immunoassay with the Europroxima Zearalenone ELISA kit (catalog number: 5121ZON). A sensitive solid-phase extraction (SPE) method was used for sample preparation and cleanup. Briefly, 1 g of finely ground plant sample was mixed with 4 ml of 84% aqueous acetonitrile and homogenized head-over-head for 60 min. After allowing the extracts to settle, 1 ml of the supernatant was evaporated to dryness under a gentle stream of nitrogen at 50 °C. The resulting residue was dissolved in a mixture of 400 µL of 100% methanol and 1600 µL of distilled water. A C18 solid-phase extraction (SPE) column was activated with 5 ml of 100% methanol, followed by 5 ml of distilled water. Subsequently, 2 ml of the dissolved sample was transferred onto the activated column at a flow rate of 1 ml/min. The column was washed with 3 ml of 20% aqueous acetone and 3 ml of 30% aqueous methanol. Zearalenone was then eluted using 2 × 1 ml of methanol. The eluate was evaporated to dryness under a mild nitrogen stream at 50 °C, and the residue was reconstituted in 500 µL of dilution buffer. From this solution, 50 µl was used for ELISA measurements. During the ELISA procedure, 50 µl of standard solutions (ranging from 0.125 to 10.0 ng/ml) and 50 µl of the prepared samples were pipetted into the microtiter plate wells in duplicate. Next, 25 µL of zearalenone-HRP conjugate and 25 µL of antibody solution were added to each well, excluding the blank wells. The plate was sealed, shaken for 1 minute, and incubated in the dark at 37 °C for 1 hour. Following incubation, the wells were washed three times with the provided rinsing buffer. Then, 100 µl of substrate solution was added to each well, and the plate was incubated for 30 minutes at 20–25 °C in the dark. The reaction was terminated by the addition of 100 µL of stop solution, and the optical density (O.D.) was promptly measured at 450 nm. For data analysis, the mean O.D. of the blank was subtracted, and the absorbance values of the standards and samples were expressed as a percentage relative to the maximal absorbance of the zero standard. The ZEN equivalent concentrations of the samples were determined from the calibration curve and multiplied by a dilution factor of 2.5, corresponding to the SPE procedure. Data were collected and analyzed using a Thermo MultiskanTM FC microplate reader (Waltham, MA, USA) with SkanIt RE software (version 6.1.1.7). Absorbance was measured at 450 nm, with a reference at 630 nm.

The concentration of Aflatoxin B1 (AFB1) was quantified using a competitive enzyme immunoassay, specifically the Europroxima Aflatoxin B1 ELISA kit (catalog number: 5121 AFB). Absorbance was measured photometrically at 450 nm using a microtiter plate reader. Approximately 10-15 g g of the sample was ground into a fine, homogeneous compound. A 1 g portion of the ground sample was weighed, and 9 ml of 80% methanol was added. The mixture was thoroughly shaken for 10 min at 20–25 °C. Subsequently, the sample was centrifuged for 10 min at 2000 × g. An aliquot of 50 µL of the resulting supernatant was diluted with 150 µL of dilution buffer to achieve a solution containing 20% methanol. From this diluted supernatant, 50 µl was used for ELISA measurements. Before the assay, all reagents were equilibrated to room temperature. We performed the measurements according to the manufacturer’s instructions. Fifty microliters of standard solutions (from 0.0157 to 0.5 ng/ml) and 50 µL of sample solutions were put into the microtiter plate wells twice. Then, 25 µL of Aflatoxin-HRP conjugate and 25 µL of antibody solution were added to each well, except the blank ones. The plate was sealed, shaken briefly, and kept in the dark at 37 °C for 1 hour. After this, the liquid was removed, and the wells were washed three times with washing buffer. Next, 100 µL of substrate solution was added to each well, and the plate was kept in the dark for 30 minutes at 20–25 °C. The reaction was stopped by adding 100 µL of stop solution, and the absorbance was measured right away at 450 nm. For data analysis, the mean optical density (O.D.) of the blank was subtracted, and the absorbance values of the standards and samples were shown as percentages of the maximum absorbance (zero standard). The AFB1 equivalent concentrations in the samples were found using a semi-logarithmic calibration curve. Finally, the calculated values were multiplied by a dilution factor of 16 to obtain the final AFB1 content in plant samples.

The concentration of deoxynivalenol (DON) was quantified using a competitive enzyme-linked immunosorbent assay (ELISA) kit (Catalog No: E-TO-E003, Elabscience). Absorbance readings were obtained using a microplate reader equipped with a 450 nm filter and a reference wavelength of 630 nm. For sample preparation, a specific extraction protocol designed for plant parts with high water absorption was followed. Briefly, 1 g of the homogenized plant sample was placed in a 50 mL centrifuge tube, and 10 mL of deionized water was added to it. The mixture was vortexed for 5 min and centrifuged at 4000 rpm for 10 min at room temperature. After centrifugation, 0.1 mL of the supernatant was transferred to a new tube and thoroughly mixed with 0.9 mL of the Reconstitution Buffer. From this final mixture, 50 µL was used for ELISA measurements. During the assay procedure, all reagents and samples were brought to room temperature (25 °C). Then, 50 µL of the standard solutions (ranging from 0 to 243 ppb) and 50 µL of the prepared samples were pipetted into the microtiter plate wells in duplicate, immediately followed by the addition of 50 µL of Antibody Working Solution to each well. The plate was sealed, gently oscillated for 5 s, and incubated in the dark at 37 °C for 30 min. After incubation, the liquid was removed, and the wells were washed five times with 300 µL of wash buffer, allowing 30-second intervals between washes. Next, 100 µL of HRP Conjugate was added to each well, and the plate was incubated in the dark at 37 °C for 30 min. The five-step washing procedure was repeated thrice. For color development, 50 µL of Substrate Reagent A and 50 µL of Substrate Reagent B were added to each well, mixed gently for 5 s, and incubated in the dark at 37 °C for 15 min. The reaction was stopped by adding 50 µL of Stop Solution, and the optical density (OD) was measured within 10 min. For data evaluation, the absorbance percentages of each standard and sample were calculated relative to the zero standard. The final DON concentration of the samples was determined from a semi-logarithmic calibration curve and multiplied by the corresponding dilution factor of 100 to obtain the concentration in the original plant material. The analytical limits of detection (LOD) were 6 ppb for zearalenone (ZEN), 0.5 ppb for aflatoxin B1 (AFB1), and 150 ppb for deoxynivalenol (DON).

2.3. Statistical Analyses

The collected plant samples were classified according to three categorical variables: month of sampling (July, August, or October), growth-form (woody, grasses, herbs), and plant part (leaf, stem, fruit – fruits and seeds –, flower). We examined whether there were significant differences in the concentrations of the measured mycotoxins (ZEN, AFB1, DON) between months, taxa, and plant parts. In our measurement results, the zero values reported for mycotoxin concentrations do not represent actual zeros but technical zeros, corresponding to concentrations below the limit of detection (LOD). We determined the co-occurrence frequencies of different toxins in plant species, growth form, organs, and months. In total, eight combinations (cases) were identified, depending on whether the mycotoxins were present (+) or absent (–) in each sample. We counted the frequency of all possible occurrence combinations. Concentrations below the LOD were interpreted as the absence of mycotoxin.

All analyses were performed using IBM SPSS Statistics 30.0.0.0 running on Windows 11. We calculated descriptive statistics. The data series for mycotoxin concentrations did not meet the assumptions of normality or homogeneity of variance, and the available sample sizes were too small to reliably apply multivariate statistical models capable of controlling for confounding among the three independent variables (growth-form, plant organ, and month of sampling). Due to these methodological limitations, we conducted a non-parametric test (Mann-Whitney test) to evaluate the effects of each independent variable separately. We examined the individual group pairs using the Mann-Whitney U test, with 3 (month, growth-form) and 6 (plant part) group pairs per category variable [26,27].

2.4. Data Visualization and Distribution Analysis

Violin plots were used to visualize the distribution of toxin concentrations across groups. The violin plot combines the features of a box plot and a density function, allowing the simultaneous interpretation of the distribution shape of measured values within groups as well as robust measures of central tendency and dispersion.

Each group (e.g., sampling month or taxonomic category) is represented by a separate violin. The shape of each violin is based on kernel density estimation calculated from the observed values; the width of the violin at a given position on the y-axis is proportional to the relative frequency expected within that concentration range. Accordingly, wider sections of the violin indicate higher data density (i.e., a greater number of observations), whereas narrower sections reflect lower density. The vertical extent of the violin represents the observed range of concentrations and allows the visualization of distributional properties such as asymmetry (skewness), peakedness (kurtosis), and potential multimodality. A narrow box plot was overlaid on each violin plot. The lower and upper edges of the box correspond to the first (Q1) and third (Q3) quartiles, respectively, with the box height representing the interquartile range (IQR = Q3 − Q1). The horizontal line within the box indicates the median. This combined visualization enables direct comparison of robust descriptive statistics of central tendency and dispersion alongside the overall distribution shape across groups.

To improve visual interpretability, only extreme outliers were excluded from the violin and box plot rendering. Extreme outliers were identified separately for each group using Tukey’s rule, with observations falling outside the range [Q1 − 3×IQR, Q3 + 3×IQR] considered extreme. Importantly, this filtering was applied exclusively for graphical display purposes; all statistical analyses were conducted on the complete dataset with outliers retained, ensuring that statistical inferences were not influenced by visual trimming. Statistically significant comparisons are indicated on the figures, with significance levels denoted by asterisks (* p < 0.05; ** p < 0.01; *** p < 0.001). The violin plots were created using the R environment (v. 4.3.3, R Core Team 2024) with the following packages: ‘dplyr’ [28], ‘ggplot2’ [29], ‘RColorBrewer’ [30], ‘ggsignif’ [31], ‘purrr’ [32].

To provide an ecological and toxicological context, measured mycotoxin concentrations were compared with European Union guidance values established for animal feed materials relevant to young ruminants [33,34].

3. Results

3.1. Co-Occurrence Patterns of Multiple Mycotoxins in Wild Plant Samples

Only 13% of the 134 samples contained none of the investigated mycotoxins (Table S1). The lowest frequency (4%) was found for samples in which only AFB1 was present, whereas the highest frequency (20%) was observed in samples in which all three toxins co-occurred. Among samples with only a single mycotoxin detected, ZEN was the most frequent (16%). The pairwise co-occurrence of ZEN and DON was the most frequent (15%).

Examining the dataset by month, we observed that in July, the most prevalent combination was the presence of all three toxins simultaneously, accounting for 25% of the samples (Table S2). In contrast, August marked a peak incidence of samples with ZEN detected alone, accounting for 26% (Table S3). In October, the highest proportions were observed in samples containing both DON and ZEN mycotoxins, as well as in those containing only DON, with both combinations accounting for 29% of the dataset (Table S4). Notably, in July, the occurrence of samples showing only AFB1 mycotoxin was the lowest at 4%. In August, there were no instances of DON and AFB1 co-occurring in the absence of ZEN. Furthermore, in October, AFB1 was completely absent from the samples, whether alone or in combination with ZEN. DON occurrence increased, but ZEN and AFB1 decreased in October.

3.2. Co-Occurrence of Mycotoxins in Plant Organs

Stems were the most mycotoxin-free (24%), followed by the flowers (15%), fruits (9%), and leaves (3%; Tables S5-S8). The most contaminated by all mycotoxins were the leaves (28%), followed by the stems (16%), the flowers (15%), and the fruits (9%). The most prevalent contamination in leaves was the co-occurrence of all three mycotoxins, ZEN alone in stems (32%) and flowers (31%), and ZEN and AFB1 together in fruits (22%). Notably, AFB1 was not detected without the presence of other mycotoxins in the stem samples. Furthermore, the flower samples did not reveal any cases of AFB1 and ZEN present together.

3.3. Co-Occurrence of Mycotoxins by Growth-Form Groups

Co-occurrence of all three mycotoxins represented the highest proportion in the case of woody taxa at 28%, followed by the grasses and herbs (17% and 10%, respectively, Tables S9-S11). None of the mycotoxins were detected in 17% of the grasses, 12% of the herbs, and 10% of the woods. The most frequently observed mycotoxin presence was ZEN alone in grasses (46% of the cases). Among herbs taxa, the most common scenario was the co-occurrence of ZEN and AFB1, making up 24% of the samples. The co-occurrence frequencies were rather evenly distributed in the woods. The least-observed combinations for each growth-form were as follows: for woody plants, the sole presence of AFB1 was the rarest, occurring in only 3% of cases. In grasses, several combinations were completely undetected, including the sole presence of AFB1, the sole presence of DON, and the joint occurrence of AFB1 and DON. For herbs species, the co-occurrence of AFB1 and DON was the least common combination, accounting for just 2% of the samples.

3.4. Temporal Variation in Mycotoxin Concentration of Plant Species

AFB1 concentration distributions did not differ significantly among months, as determined by pairwise Mann–Whitney U tests (Figure 2A; Tables S12 and S13). The minimum value and even the lower-quartile and the median – except July – were “0”. However, the violin plots indicated a gradual decrease in distribution ranges and maximums from July to October, accompanied by a right-skewed distribution in August due to several high-concentration observations. These extreme values were detected primarily in fruits and vegetative tissues of woody species, including Rosa canina and Rosa rubiginosa, where AFB1 concentrations exceeded the regulatory limit and reached up to 55 µg/kg in leaves and 71.96 µg/kg in fruits (Tables S18–S22).

We found seasonal differences in mycotoxin concentrations for ZEN and DON. (Figure 2B and C). ZEN concentrations differed significantly between summer and autumn months (Table S13). July and August did not differ significantly, although August samples exhibited a broader, more evenly distributed concentration range (Table S12). In contrast, October samples showed significantly lower ZEN concentrations and a more compact distribution. This seasonal pattern coincided with the occurrence of extreme ZEN values in August, including a high concentration measured in leaf samples of Verbascum phlomoides (89.78 µg/kg) (Figure 2C; Tables S18-22).

DON concentrations showed a distinct seasonal pattern. We observed a significant difference between August and October (Mann–Whitney U test, p < 0.05), whereas July did not differ significantly from either month (Table S13). The higher median and upper quartile in July were compensated by the extremely high maximum in August, measured in vegetative tissues of Salix cinerea (1,541.72 µg/kg), which represents a concentration close to the European Union guidance value for feed materials relevant to young ruminants. The violin plots indicated that this August–October contrast reflected a general shift toward higher DON concentrations in October (Figure 2B; Tables S18-S22).

3.5. Mycotoxin Contamination of Plant Organs

Although the data dispersion was large and the distributions were concentrated around lower values, as with other investigated parameters, our results revealed differences in mycotoxin concentrations among plant organs (Figure 3A–C, Tables S14 and S15). The minimum values and lower quartiles were below the LOD for each organ. Stem and flower medians were 0 µg/kg, but the maximums were similar to the other organs (Table S14). Median AFB1 concentrations were significantly higher in leaves than in stems and fruits (Table S15). Despite impressive differences between leaves and fruits, with higher values in fruits, the test did not reach statistical significance. The median (5.32 µg/kg) and the maximum (71.96 µg/kg) AFB1 concentrations were the highest in fruits and the lowest in leaves (2.985 µg/kg and 55 µg/kg, respectively). The AFB1 concentrations exceeded the limit in leaves of Populus alba and Rosa rubiginosa, in stems of Salix alba and Prunus spinosa, in fruits of Rosa canina, Gleditsia triacanthos, Rosa rubiginosa, Lathyrus tuberosus, Epilobium parviflorum, and in the flowers of Chenopodium album (Tables S18-S22).

Deoxynivalenol (DON) exhibited the clearest organ-specific differences. The minimum values and lower quartiles were below the LOD for each organ. Stem and fruit medians were also “0” µg/kg, and the maximums were low except for leaves (Table S14). Leaves were the most contaminated and differed significantly from all other organs, whereas stems, fruits, and flowers did not differ significantly from each other (Table S15). This pattern reflects a clear association of DON contamination with leaf tissues. Notably, extreme DON concentrations were detected in vegetative tissues, including 1,541.72 µg/kg as the highest in Salix cinerea, indicating that a small number of highly contaminated leaf samples contributed substantially to the observed differences (Tables S18-S22).

In contrast, zearalenone (ZEN) showed more moderate organ-specific variation (Table S14). Although distributions largely overlapped across organs, leaf samples showed a broader concentration range than stems and fruits. Pairwise tests indicated a significant difference only between leaves and fruits, suggesting a weaker organ-specific pattern compared to AFB1 and DON (Table S15). We detected low contamination in general, but the lowest was in fruits. None of the values were above the limit. Overall, organ-specific differences in mycotoxin contamination were most evident for DON and AFB1, while ZEN exhibited comparatively homogeneous distributions. Importantly, these differences were primarily driven by a limited number of samples with elevated concentrations rather than consistent shifts in median values across all plant organs.

3.6. Mycotoxin Contamination of Major Plant Growth Forms

Violin plots grouped by major plant growth forms showed distinct patterns in mycotoxin concentrations among woody species, monocotyledonous (grasses) plants, and dicotyledonous herbaceous (herb) species; however, the magnitude and consistency of these differences varied among the investigated mycotoxins (Figure 4A–C, Table S16). AFB1 concentration distributions differed significantly among plant growth forms. The contamination in grasses was significantly lower than in woody species or in herbs (Table S17). No significant difference was found between woody plants and herbs, although the median of the contamination in woody plants was higher than in herbs (3.52 and 0 µg/kg), but it was opposite in the case of maximums (73.3 and 81.08 µg/kg). Consistent with this, co-occurrence data showed that AFB1 alone was rare in woody plants (3%), whereas monocotyledonous taxa exhibited multiple combinations where AFB₁ was absent or occurred only in mixtures.

DON exhibited the most pronounced growth form–specific differences. Woody species showed substantially higher and more variable DON concentrations compared to both grasses and herbs, while the latter two groups did not differ significantly from each other (Tables S16 and S17). This pattern is supported by co-occurrence data, where DON-containing combinations (e.g., DON alone: 10%, ZEN + DON: 13%, and all three toxins: 28%) were markedly more frequent in woody species. In contrast, there were no occurrences of DON alone or in combination with AFB₁, showing a markedly lower contribution to DON contamination. DON occurred at intermediate frequencies (e.g., DON alone: 7%, ZEN + DON: 17%), but without the extreme values observed in woody taxa. The results identify woody species as a major contributor to elevated DON levels in semi-natural vegetation adjacent to agricultural fields.

In contrast, zearalenone (ZEN) showed weaker differentiation among plant growth forms. Pairwise comparisons showed a significantly higher ZEN concentration in grasses than in woods, whereas comparisons among the other groups were not significant, even though the herb values, except for the maximum, were apparently lower than those of grasses (Tables S16 and S17). ZEN was the most frequently detected single toxin in herbaceous taxa, accounting for 46% of samples in grasses and 17% in herbs. In woody species, ZEN occurred both alone (7%) and in combination with other toxins, particularly in triple co-occurrence (28%).

The highest concentrations exceeding the limit values were found for AFB1 in woody plants and next in herbs. DON contamination exceeded the limit in only one case involving a woody species. ZEN concentrations in grasses were high but did not exceed the limit (Tables S18-S22).

Overall, differences among plant growth forms were most pronounced for DON, intermediate for AFB1₁, and least evident for ZEN. As observed in seasonal and organ-specific analyses, statistically significant differences were mainly driven by shifts in distributional variability and the presence of highly contaminated samples, rather than consistent differences in central tendency across all groups.

4. Discussion

4.1. Conceptual Synthesis of the Main Findings

The present study indicates that mycotoxin contamination in wild plants is primarily shaped by seasonal dynamics and growth-form affiliation, while plant organ plays a comparatively minor and less consistent role. The contamination in vegetative organs, especially leaves, was unexpectedly high, exceeding that in reproductive organs. Co-occurrence of multiple mycotoxins was common, indicating that contamination is driven by shared ecological and environmental factors rather than isolated processes. The results reflect the timing of the mold infection and mycotoxin production, as well as the dominant roles of abiotic and biotic factors in natural or semi-natural habitats. Seasonal variation resulted not only in changes in toxin concentrations but also in shifts in mycotoxin profiles, from multi-toxin occurrence toward more toxin-specific patterns across time. Environmental factors vary not only over time but also in space, resulting in high variability in mycotoxin content among plant species and within the same species. Notably, woody species emerged as important contributors to AFB1 and DON contamination, highlighting their previously underestimated role as reservoirs of mycotoxigenic fungi in semi-natural vegetation. Together, these findings contest crop-centered outlooks, emphasize the ecological relevance of wild plant communities to mycotoxin dynamics, and draw attention to the risks of mycotoxicosis for wild animals and, in some cases, to human food risks, e.g., the consumption of wild rose fruits.

4.2. Co-Occurrence of Mycotoxins

The present study demonstrates that wild plant species in semi-natural vegetation adjacent to agricultural fields can harbor detectable levels of mycotoxins, including ZEN, AFB₁, and DON. This confirms that mycotoxin contamination occurs not only in cultivated crops but also in natural vegetation, supporting previous observations from grasslands, meadows, and wild plant communities [6,7]. Building on these findings, mycotoxin co-occurrence was common in wild plants near agricultural fields, with ZEN, AFB₁, and DON co-occurring in approximately 20% of samples. This rate is similar to wheat-based systems, where 16.4%, 6.8%, and 6.8% of samples contained two, three, and four mycotoxins, respectively [35]. Notably, recurring mycotoxin combinations included enniatins with DON, OTA, or fumonisins, as well as more complex mixtures such as AFB₂ + AFG₂ + OTA + ZEN [35]. Substantially higher co-occurrence rates have been observed in processed cereal-based products; for instance, 90% of couscous samples contained at least two mycotoxins—most commonly tenuazonic acid, tentoxin, DON, and OTA [36]. Earlier surveys of cereals found co-occurrence rates of approximately 40–60% [37,38]. Similar multi-mycotoxin patterns have also been documented in wheat using UPLC-MS/MS analyses, where complex mixtures of Fusarium-, Alternaria-, and Aspergillus-derived toxins were frequently detected within the same samples [39]. Although the overall frequency in wild plants is lower, the presence of multiple mycotoxins suggests that co-occurrence results from shared ecological drivers rather than solely from agricultural practices or post-harvest mixing. Mechanistically, overlapping mycotoxigenic fungal communities and environmental conditions, such as temperature and moisture, underlie these patterns. Similar relationships between toxin profiles and specific Fusarium communities have been demonstrated in Swedish wheat, where DON, ZEN, NIV, and other Fusarium toxins were closely associated with the occurrence of particular Fusarium species [40]. Mycotoxigenic fungi, particularly Fusarium spp., frequently co-colonize host plants and exhibit broad host ranges, including non-crop species [10,41]. Recent studies further demonstrated that individual weed species may harbor multiple Fusarium strains capable of producing distinct mycotoxin profiles, including DON, 3-ADON, NIV, enniatins, and ZEN under field conditions [10]. From a toxicological perspective, co-occurrence is especially relevant because mycotoxins may act additively or synergistically, interactions not captured by single-compound risk assessments [42,43]. These findings highlight the need to consider mycotoxin mixtures in ecological and exposure assessments. Recent reviews on herbs, spices, and plant-derived supplements similarly emphasize the frequent co-occurrence of multiple mycotoxins in non-cereal plant matrices, including combinations of aflatoxins, ochratoxin A, fumonisins, trichothecenes, and zearalenone [44]. Similar patterns of widespread contamination have been reported in fresh leafy vegetables. For example, 72.5% of spinach samples contained at least one mycotoxin, with zearalenone (ZEN) detected in 50% of cases [45]. These findings support the view that non-cereal plant tissues, particularly leaves, can serve as important matrices for mycotoxin occurrence and co-occurrence under field conditions. Importantly, these results expand current knowledge by demonstrating that co-occurrence of multiple mycotoxins can already be observed in wild vegetation, highlighting its potential as an environmental reservoir for complex mycotoxin mixtures. Consequently, the noted frequencies support the need to consider mycotoxin mixtures in ecological and exposure assessments [42,43]. Recent risk prioritisation frameworks further emphasize that combined exposure to co-occurring mycotoxins represents a major challenge for food safety assessment because mixture effects are insufficiently captured by single-compound approaches [46].

4.3. Seasonal Variation

Seasonal variation appeared as a key factor influencing both mycotoxin concentrations and simultaneous occurrence patterns. ZEN concentrations were significantly higher in August than in July or October, while DON showed a shift toward higher values in October, and AFB₁ was no longer detected in late-season samples. These results show that seasonal variability is not limited to changes in toxin levels but also involves restructuring of mycotoxin profiles over time. Early summer samples showed a higher frequency of multi-toxin co-occurrence, whereas late summer was dominated by ZEN, followed by a transition toward DON-dominated profiles in autumn. Such patterns correspond to the known ecology of mycotoxigenic fungi. Temperature and moisture strongly influence fungal growth and toxin production [20,47]. Experimental studies show that fungi produce distinct toxins under different environmental conditions. This leads to temporal shifts in toxin profiles as climate changes [22,48]. Seasonal dynamics likely reflect shifts in fungal community composition and activity. Distinct fungal taxa occupy distinct ecological niches. Climate-driven shifts in Fusarium species composition have also been predicted for Northern Europe, with an increasing prevalence of F. graminearum and associated DON contamination under warmer, more humid conditions [25]. These succession patterns determine which mycotoxins dominate in plant tissues [48,49]. The absence of AFB₁ in October samples matches the ecological preferences of Aspergillus species. These fungi are typically associated with warmer and drier conditions. This illustrates the strong climatic control over toxin-specific occurrence patterns [22,49]. In contrast, the increased relative importance of DON in autumn may reflect the persistence or late-season activity of Fusarium species. These fungi thrive in cooler, more humid conditions. Late summer (August) represented a period of increased variability. This time showed wider concentration ranges and more extreme values. Such a pattern may indicate a transitional phase in fungal community dynamics. Multiple ecological drivers may promote toxin production during this period. In addition, plant senescence and the accumulation of plant debris may increase substrate availability for fungal colonization and secondary metabolite production [9,47]. Overall, these findings show that seasonal drivers of mycotoxin occurrence act through direct effects on fungal growth. They also operate via shifts in community composition and substrate availability. This highlights the crucial role of considering temporal dynamics when assessing mycotoxin contamination in both cultivated and semi-natural vegetation. This is especially important under evolving climatic conditions. Seasonal fluctuations in dominant mycotoxins have also been observed in pasture ecosystems, where zearalenone-related metabolites and ochratoxin concentrations varied markedly across years and between wet and dry seasons [50].

4.4. Plant Organ-Specific Patterns

Results partially support that mycotoxin contamination varies among plant organs, with recorded patterns being toxin-specific and mainly influenced by extreme rather than consistent differences. Significant differences in DON and AFB₁ were driven by a few highly contaminated samples. DON contamination is closely tied to leaf tissues. The strong association with leaf tissues suggests that vegetative plant parts may provide favorable conditions for Fusarium colonization and toxin production. Leaves represent a large, continuously exposed surface to environmental inoculum, which may facilitate fungal establishment [20,49]. In contrast, stems and fruits showed lower or less variable DON levels, showing a more limited role in DON accumulation. AFB₁ exhibits a distinct pattern of distribution. AFB₁ displayed elevated concentrations detected in both fruits and leaves. This may reflect the ecological preferences of Aspergillus species, which are often associated with exposed plant tissues under warm conditions [22,49]. The presence of high AFB₁ levels in fruits suggests that reproductive organs may act as localized centers of contamination, although this pattern was not consistent across all samples. ZEN shows limited organ-specificity in its distribution. The ZEN pattern shows only weak organ-specific differentiation, unlike both DON and AFB₁, indicating that its distribution is less dependent on plant organ type. This may suggest wider ecological tolerance or more uniform colonization patterns of ZEN-producing fungi across plant tissues [48]. A small number of high-concentration samples largely influenced organ-level differences, implying that stochastic or localized infection events may drive mycotoxin distributions. Similar high variability and occasional extremes have been reported in both crop and non-crop systems [1,2]. Ecologically, these data indicate that plant organs contribute variably to environmental mycotoxin exposure. Rather than a uniform pattern, contamination is driven by localized fungal function and environmental conditions. This demonstrates the importance of accounting for both organ type and variability when assessing mycotoxin occurrence in semi-natural vegetation.

4.5. Growth-Form Specific Differences

Growth-form affiliation was a strong determinant of mycotoxin concentrations. Significant differences were detected among woody species, grasses (monocotyledonous herbaceous plants), and herbs (dicotyledonous herbaceous species) for all examined toxins, with particularly pronounced differences for DON. Woody plants showed markedly higher DON concentrations compared to herbs, a result that contrasts with the common assumption that woody taxa are less relevant hosts for toxigenic fungi. This finding suggests that woody plants in field margins and shelter belts may act as underestimated reservoirs for mycotoxin-producing fungi. These results align with the concept of host specificity among fungi, which can range from monophagous to polyphagous strategies. While many Fusarium species preferentially infect monocotyledonous hosts, polyphagous taxa can colonize a wide range of plant species, including woody plants under suitable environmental conditions. Overall, these findings support the view that mycotoxin contamination in natural vegetation is strongly species-dependent, reflecting differences in host suitability for distinct toxigenic fungal communities [10]. For example, bermudagrass and limpograss exhibited markedly higher zearalenone-4-sulfate concentrations than bahiagrass, highlighting that vegetation structure and species composition may strongly influence mycotoxin accumulation patterns [50].

4.6. Plant Species Comparison

As previously noted, mycotoxin contamination of naturally growing species received little attention in the past. Most of the data comes from studies focusing on medicinal herbs, herbal teas, spices, and plant-derived products [44,51,52]. Previous studies reported low levels of aflatoxin contamination in commercial rosehip products [52], whereas in our study, several Rosa species showed substantially higher AFB1 concentrations. These species are often used in food (syrups, tea, etc.) and medicinal products. However, Burkin & Kononenko [53] collected data about wild-growing gramineous plants. This study mainly focused on mycotoxin contamination in meadows and hayfields in European Russia. They collected data from false oatgrass (Arrhenatherum elatius), different bromes species (Bromus sps.), and couch grass (Elymus repens). In all these plants, ZEN and DON mycotoxins were detectable. Our results show similarity with this. Oat grass and smooth brome were contaminated with ZEN and DON; however, couch grass only showed ZEN contamination.

4.7. Organ-Specific Patterns and Monitoring

Contrary to the crop-oriented focus of most mycotoxin monitoring programs, plant parts were a less decisive factor in explaining mycotoxin variability. Significant differences among plant organs were detected only for DON, mainly between leaves and stems, and between leaves and fruits. This finding supports growing evidence that vegetative tissues can accumulate substantial mycotoxin concentrations, sometimes exceeding those found in reproductive organs [54]. The limited number of significant differences among plant parts suggests that restricting monitoring efforts to seeds or fruits may underestimate overall mycotoxin presence in natural vegetation. Similar accumulation of DON and ZEN in vegetative tissues has been reported previously, indicating that mycotoxin presence is not restricted to reproductive organs but may extend to leaves and stems under field conditions [54,55]. Plant metabolism may further influence these accumulation patterns because several mycotoxins can occur in conjugated or “masked” forms within vegetative tissues, affecting their distribution and detectability in plants [56].

4.8. Ecological and Agronomic Implications

Wild plants growing near agricultural fields may influence mycotoxin dynamics at multiple levels. They can serve as alternative hosts for toxigenic fungi, facilitate spore dispersal to crops via wind or insect vectors, and contribute to indirect exposure pathways for wildlife and livestock. Herbivorous mammals such as roe deer and brown hare consume a wide range of wild plant species and plant organs [57,58,59], potentially ingesting mycotoxins that not only can harm their survival and reproduction [18,60,61,62], but moreover may accumulate in tissues and enter the human food chain through game meat consumption. Moreover, small mammals feeding on contaminated vegetation may act as vectors of mycotoxins to higher trophic levels, including predators [19]. Semi-natural field margins are widely recognized as ecological hotspots that support complex plant–fungus–animal interactions and may facilitate indirect exposure pathways for environmental contaminants [63,64]. Recent research underscores that the assessment and regulation of mycotoxin contamination in food are frequently insufficient, potentially leading to an underestimation of environmental contamination [45]. Moreover, the current mandatory limits set by the European Union do not include fresh and minimally processed vegetables. This situation can be analogously applied to semi-natural vegetation. If adequate regulations are absent even for commercially available leafy greens, the monitoring of toxins entering the food chain through wild plants (e.g., via game meat consumption) represents an even more significant oversight in existing risk assessment frameworks. Although the concentrations measured in this study were generally lower than the extreme values reported for forage or medicinal plants in other regions [51,65], several samples exceeded or approached European Union guidance values established for animal feed materials relevant to young ruminants, particularly in the case of AFB1. Although DON concentrations generally remained below the official EU guidance values, some individual samples exceeded ecological and toxicological relevance.

4.9. Limitations and Future Directions

The present study has several limitations. Sample sizes were uneven across months, taxa, and plant organs, which may have influenced the robustness of statistical comparisons. In addition, the study focused on mycotoxin concentration patterns rather than fungal community composition, which could provide important mechanistic insights into the observed differences. Future studies, including mycobiome analyses, would help to better understand the ecological drivers of mycotoxin occurrence.

Another limitation concerns the analytical approach. Although immunoanalytical methods are suitable for screening, they may be affected by cross-reactivity and have lower specificity than chromatographic techniques used for multi-mycotoxin detection, such as HPLC or LC-MS/MS [66,67]. Therefore, future studies applying chromatographic confirmation methods would further strengthen the analytical robustness of the present findings.

Subsequent studies should also integrate fungal community characterization, multi-year sampling, and experimental techniques to better understand the role of wild vegetation as a reservoir of toxigenic fungi and to assess the ecological and food safety implications under changing climate conditions.

5. Conclusions

This study demonstrates that wild plant species growing in semi-natural vegetation adjacent to agricultural areas can harbor detectable concentrations of multiple mycotoxins, including ZEN, AFB1, and DON. The results indicate that mycotoxin occurrence in vegetation is primarily shaped by seasonal conditions and growth-form affiliation, while plant organ plays a comparatively minor role. Notably, woody plant species exhibited unexpectedly high DON concentrations, and contamination in vegetative organs, especially leaves, was higher than in reproductive organs, suggesting that field margins and shelter belts may function as underestimated reservoirs for mycotoxin-producing fungi.

Seasonal differences, particularly the elevated ZEN concentrations observed in late summer, highlight the strong influence of temperature and moisture conditions on mycotoxin dynamics, underscoring the potential sensitivity of these processes to ongoing climate change. The limited organ-specific patterns further suggest that focusing monitoring efforts exclusively on generative plant parts may underestimate overall mycotoxin presence in natural vegetation.

Taken together, these findings challenge traditional crop-centered perspectives on mycotoxin contamination and emphasize the ecological relevance of wild plants in agricultural landscapes. While the study is constrained by uneven sampling and the absence of fungal community data, it provides a basis for future integrative research combining mycotoxin analysis with fungal ecology to better assess environmental reservoirs and exposure pathways under changing environmental conditions.