Mixtures of Mycotoxins, Phytoestrogens, and Other Secondary Metabolites in Whole-Plant Corn Silages and Total Mixed Rations of Dairy Farms in Central and Northern Mexico

Mycotoxins and endocrine disruptors such as phytoestrogens can affect cattle health, reproduction, and productivity. Most studies of mycotoxins in dairy feeds in Mexico and worldwide have been focused on a few (regulated) mycotoxins. In contrast, less known fungal toxins, phytoestrogens, and other metabolites have been neglected and underestimated. This study analyzed a broad spectrum (>800) of mycotoxins, phytoestrogens, and fungal, plant, and unspecific secondary metabolites in whole-plant corn silages (WPCSs) and total mixed rations (TMRs) collected from 19 Mexican dairy farms. A validated multi-metabolite liquid chromatography/electrospray ionization–tandem mass spectrometric (LC/ESI–MS/MS) method was used. Our results revealed 125 of >800 tested (potentially toxic) secondary metabolites. WPCSs/TMRs in Mexico presented ubiquitous contamination with mycotoxins, phytoestrogens, and other metabolites. The average number of mycotoxins per TMR was 24, ranging from 9 to 31. Fusarium-derived secondary metabolites showed the highest frequencies, concentrations, and diversity among the detected fungal compounds. The most frequently detected mycotoxins in TMRs were zearalenone (ZEN) (100%), fumonisin B1 (FB1) (84%), and deoxynivalenol (84%). Aflatoxin B1 (AFB1) and ochratoxin A (OTA), previously reported in Mexico, were not detected. All TMR samples tested positive for phytoestrogens. Among the investigated dietary ingredients, corn stover, sorghum silage, and concentrate proportions were the most correlated with levels of total mycotoxins, fumonisins (Fs), and ergot alkaloids, respectively.

Studies on the broad spectrum of mycotoxins and other endocrine disruptors of natural origin in the feedstuffs and diets of dairy cows and other food-producing animals are essential. However, they are still very limited [2,4]. Therefore, this study aimed to determine co-occurrences and concentrations of mycotoxins and other fungal metabolites (derived from the genus Alternaria, Aspergillus, Fusarium, Penicillium, other fungi, and ergot alkaloids) as well as secondary plant metabolites (such as phytoestrogens and others) in TMRs and WPCSs from large dairy cattle farms located in northern and central Mexico. The analysis was achieved using a validated multi-metabolite analysis. The possible associations of the main dietary ingredients to the levels of mycotoxins and other secondary metabolites contained in the TMRs were also assessed.
Regarding the proportion of inclusion (dietary content), commercial high-energy density concentrate was the dietary ingredient most abundant in the evaluated TMRs, with an average inclusion of 45.5% on a dry matter (DM) basis, ranging from 28.3% to 60%. WPCSs averaged 38.9% DM of the rations, varying from 27.5% to 53%. The diets that included rolled corn, corn meal, and protein-rich concentrate presented an average of 25.5%, 24.2%, and 22.6% DM, respectively. The remaining ingredients were included in the TMR formulations with an average of inclusion (proportion) less than 15% DM (Table 1).

Occurrence and Concentrations of the Detected Metabolites 2.2.1. General Overview
The 125 identified biological compounds in representative samples of WPCSs (114) and TMRs (118) were grouped based on their reported primary producers. These compounds consisted of 94 fungal metabolites, derived from the genera Alternaria (number of detected metabolites: 8), Aspergillus (12), Fusarium (38), Penicillium (16), or other fungi (17) and ergotderived alkaloids (3). Thirteen compounds were plant-derived metabolites (including 9 phytoestrogens), 18 were unspecific metabolites (multi-kingdom-derived, i.e., derived from fungi, bacteria, and/or plants), and 1 was of bacterial origin ( Figure 1). Figure 1 and Table 2 illustrate the occurrence and concentrations of the groups of metabolites. Additionally, Table 2 shows the significance level of a paired comparison between the WPCSs and TMRs per respective farm.  1. Occurrence and distribution of concentrations (µ g/kg on DM basis, log 10 scale) of groups of secondary metabolites detected in whole-plant corn silages (in yellow) and total mixed rations (in gray) on dairy farms in Mexico. The total number of secondary metabolites detected per group is shown in parentheses. Asterisks (*) show significant differences (p-value < 0.05) between the concentrations of the respective groups in whole-plant corn silages and total mixed rations according to the Wilcoxon matched-pairs signed rank test (p-values in Table 2). Means are shown as "+". Occurrence and distribution of concentrations (µg/kg on DM basis, log 10 scale) of groups of secondary metabolites detected in whole-plant corn silages (in yellow) and total mixed rations (in gray) on dairy farms in Mexico. The total number of secondary metabolites detected per group is shown in parentheses. Asterisks (*) show significant differences (p-value < 0.05) between the concentrations of the respective groups in whole-plant corn silages and total mixed rations according to the Wilcoxon matched-pairs signed rank test (p-values in Table 2). Means are shown as "+".

Mycotoxins and Other Fungal Secondary Metabolites
Of the 94 fungal-derived metabolites detected, 58 have been previously reported as mycotoxins (Table 3). Among the major mycotoxins, ZEN, deoxynivalenol (DON), FA1, FA2, FB1, FB2, FB3, and FB4 were detected in TMRs as well as in WPCSs. The carcinogenic mycotoxins AFB1 and OTA were not detected in the assessed feed samples. Toxin metabolites related to the parent major mycotoxins such as nivalenol (NIV), DON-3-glucoside, and hydrolyzed FB1 were found in silages and in rations. 15-acetyl-deoxynivalenol was seen only in WPCSs (11%). ZEN was detected in all the TMR samples and in 68% of the WPCSs. The dietary levels of ZEN had an average of 38.7 µg/kg and a maximum concentration of 246 µg/kg. DON was also found with a higher frequency in TMRs (84%) than in WPCSs (53%). The average concentration of DON in TMRs was 615 µg/kg, and the maximum level was 1660 µg/kg. FB1 and FB2 were the most detected and with the highest levels among the fumonisins (Fs) in TMRs, with occurrences of 84% and 64% and an average of 218 µg/kg and 103 µg/kg, respectively. NIV was detected more frequently in TMRs (68%) than in WPCSs (42%). The levels of NIV were significantly higher (p-value = 0.0061) in samples of TMRs (mean: 872 µg/kg; max: 2600 µg/kg) than in WPCSs (mean: 269 µg/kg; max: 614 µg/kg) ( Table 3). The totals of Fs (sum of FA1, FA2, FB1, FB2, FB3, FB4 hydrolyzed B1) and a total of type B trichothecenes (sum of DON, 15-acetyl-DON, DON-3-glucoside, and NIV) were superior in WPCSs than in TMRs; however, the differences were not significant. The maximum amount of total Fs in silages was 4410 µg/kg, and it was 1670 µg/kg in mixed rations. The maximum concentration of type B trichothecenes in WPCSs was 4230 µg/kg, whereas the highest concentration in the analyzed TMR samples was 5510 µg/kg. Concerning other less studied mycotoxins derived from Fusarium spp. such as beauvericin, beauvericin A, bikaverin, and enniatins (ENNs) (A, A1, A2, B, B1, and B2), epiquisetin, equisetin, fusaric acid, culmorin, moniliformin, and siccanol, among others, were also detected. Beauvericin, bikaverin, and moniliformin were detected in all the analyzed TMRs. Siccanol was the Fusarium-derived metabolite with the highest concentration, averaging 2510 µg/kg and with a maximum level of 6130 µg/kg. The second most produced fusarial metabolite was 15-hydroxyculmorin (average: 1270 µg/kg; max: 1510 µg/kg). The total content of ENNs was significantly superior (p-value = 0.0144) in TMRs than in WPCSs, with a respective average of 11.2 µg/kg and 5.96 µg/kg. Among mycotoxins produced by Penicillium spp., mycophenolic acid was detected in TMRs, with an occurrence of 42%. Citrinin, primarily Penicillium-derived but also produced by some Aspergilli, was seen in only 1 TMR sample (5%). Concerning the metabolites derived from Alternaria spp., tentoxin and tenuazonic acid showed the highest occurrence in TMRs, being detected in 79% and 53% of the samples, respectively. Tenuazonic acid presented the highest concentration in TMRs among this group of metabolites (average: 49.4 µg/kg, range: 30.3 µg/kg-82.8 µg/kg).
Tenuazonic acid presented the highest concentration in TMRs among this group of metabolites (average: 49.4 µg/kg, range: 30.3 µg/kg-82.8 µg/kg). In the analyzed silages, the most frequently occurring were alternariolmethylether (47%) and tentoxin (42%). The Alternaria-derived metabolites with the highest levels were infectopyrone (average: 97 µg/kg; range: 23.9 µg/kg-176 µg/kg) and tenuazonic acid (average: 40.2 µg/kg; range: 30.1 µg/kg-60.4 µg/kg). Other metabolites produced by Alternaria spp. such as altenuisol, alternariol, altersetin, and macrosporin were also detected in both studied matrices. Among the metabolites produced by fungi of the genus Aspergillus, flavoglaucin was detected in all the TMR samples, followed by phenopyrrozin, which was found in 79% of the samples. Phenopyrrozin also showed the highest occurrence of Aspergilli-derived compounds in WPCSs. Kojic acid presented the highest levels among the Aspergillus-derived metabolites in silages and TMRs, although its occurrence was low. Other metabolites/mycotoxins such as averufin, fumigaclavine, fumiquinazolin D, and seco-Sterigmatocystin were detected in both WPCSs and TMRs. Among the ergot alkaloids, dihydroergosine occurred the most. Chanoclavine was detected in low frequency (5%) in TMRs and WPCSs. Festuclavine was detected only in WPCSs. The concentrations of the individual ergot alkaloids were low (≤12.5 µg/kg). Concerning the compounds produced by other fungi species, the most frequently detected in TMR samples were ilicicolin B (89%), cercosporin (79%), monocerin (74%), and sporidesmolide II (74%). The metabolites derived from other fungal species that were most frequently detected in the analyzed WPCS samples were monocerin (89%), sporidesmolide II (84%), and ilicicolin B (79%). Monocerin was the compound derived from other fungi with the highest concentration in WPCSs as well as in TMRs, with a respective average of 115 µg/kg and 85.9 µg/kg. The levels of monocerin were significantly superior (p-value = 0.0024) in the silages. Samples of silage and TMRs showed content of ascochlorin, ascofuranone, bassianolide, beauveriolide III, ilicicolin A, and myriocin. Cytochalasin J, destruxin B, and mycousnine were identified only in TMRs, whereas phomalone and sporidesmolide III were identified only in WPCSs.

Plant Secondary Metabolites (Phytoestrogens and Others)
Among the plant-derived metabolites, nine phytoestrogens in WPCSs and in TMRs were found. The isoflavones daidzein, daidzin, genistein, genistin, and glycitin and the isoflavone glucoside onionin were identified in all the TMR samples. Biochanin, coumestrol, and glycitein also presented a high occurrence in TMRs (≥79%). The most frequently detected phytoestrogens in WCPSs were daidzin (68%), genistin (63%), and genistein (58%). The other phytoestrogens were identified in <40% of the silage samples. The concentrations of the all the found phytoestrogens were significantly higher (p-value < 0.5) in the TMRs than in the WPCSs. Genistin was the (plant) metabolite with the highest concentration detected in both feed matrices, averaging 1000 µg/kg and 118,150 µg/kg in WPCSs and TMRs, respectively. Concerning the compounds cataloged in the group of other plant metabolites, abscisic acid was predominant in occurrence and concentration; for instance, its occurrences in WPCSs and TMRs were 42% and 100%, respectively. The levels of abscisic acid were significantly superior (p-value = 0.0012) in TMRs than in the silages. Three tropane alkaloids (anisodamine, atropine, and hyoscine) were identified in both analyzed matrices. The 3 mentioned tropane alkaloids occurred in 16% of the WPCSs, whereas in TMRs anisodamine was found at a frequency of 16% and atropine as well as hyoscine in 11% of the samples. These alkaloids were detected in concentrations lower than 1000 µg/kg; the levels were higher in WPCSs, but without significance (i.e., p-value > 0.05) ( Table 3).

Unspecific (Multi-Kingdom) and Bacterial Metabolites
Multiple metabolites that can be produced by unrelated organisms belonging to diverse kingdoms such as Plantae, Fungi, and/or Eubacteria were detected in both feed commodities. Four compounds belonging to this category, asperglaucide, cyclo (L-Pro-L-Tyr), cyclo (L-Pro-L-Val), neoechinulin A, and rugulusovine, were identified in all the assessed TMR samples. Asperphenamate, brevianamid F, chrysophanol, emodin, skyrin, fellutanine A, and iso-rhodoptilometrin occurred in TMRs at a rate superior to 50%. Other unspecific metabolites such as 3-nitropropionic acid, chrysophanol, citreorosein, N-benzoylphenylalanine, and tryptophol were detected in frequencies between 20% and 40%. Regarding the occurrence in WPCSs, all the assessed samples contained cyclo (L-Pro-L-Tyr), cyclo (L-Pro-L-Val), and rugulusovine, with frequencies of over 50% for brevianamid F, citreorosein, emodin, fellutanine A, iso-rhodoptilometrin, and skyrin. The highest concentrations in the category of unspecific metabolites corresponded to the bioactive cyclic dipeptides cyclo (L-Pro-L-Val) and cyclo (L-Pro-L-Tyr) in silages as well as in TMRs, with an average concentration above 2100 µg/kg. The content of both cyclic dipeptides was significantly higher (p-value < 0.001) in WCPSs than in the complete rations. The other compounds of this group presented average concentrations lower than 400 µg/kg in WPCSs and TMRs (Table 3).  Table S1 presents the exact values of the co-contamination levels (average ± SD, median, minimum, maximum) of the groups of metabolites. Additionally, the significance (p-values) of the comparison via the Wilcoxon matched-pairs signed rank test between co-contamination of the diverse groups' metabolites in samples of WPCSs and TMRs are also presented in Table S1. All the samples were co-contaminated with cocktails of toxins/metabolites. In total, the assessed WPCSs showed, on average, 29 metabolites per sample (range: 13-39 metabolites per sample), and TMRs showed 55 metabolites per sample (range: 31-66 metabolites per sample). Silages presented, on average, 17 mycotoxins per sample, varying from 6 to 27 mycotoxins per sample. The assessed TMR samples showed a mean of 24 mycotoxins per sample, ranging from 9 to 31 mycotoxins per sample. The mycotoxin co-contamination level was significantly higher (p < 0.0001) in TMRs. Fusarium-derived metabolites was the category among the fungal metabolites with the highest diversity of detected compounds, showing, on average, 14 metabolites per sample (max: 20 metabolites per sample) in WPCSs and 18 metabolites per sample (max: 24 metabolites per sample) in TMRs. The co-contamination grade with compounds derived from Aspergillus spp., Fusarium spp., Penicillium spp., and other fungi, in addition to the total fungal metabolites, phytoestrogens, plant metabolites, unspecific metabolites, and total metabolites, was significantly higher (p-value < 0.05) in the complete rations than in the silages.

Co-occurrence of Mycotoxins and Phytoestrogens
Co-occurrence analyses (frequency of detection of combinations, in %) between mycotoxins/metabolites evidenced in WPCSs and TMRs are presented in Figures 3 and 4, respectively. In the silages, the most frequent combinations of mycotoxins detected were among ZEN and fusarial emerging mycotoxins such as aurofusarin, beauvericin, beauvericin A, bikaverin, fusaric acid, and siccanol, which presented occurrences of over 50%. The co-occurrence of DON and ZEN was detected in 53% of the WPCS samples. The Aspergillus-produced metabolite phenopyrrozin and the Penicillium-derived questiomycin derivate showed co-occurrences of over 50% with ZEN and the emerging fusarial mycotoxins such as aurofusarin, beauvericin, beauvericin A, bikaverin, fusaric acid, and siccanol ( Figure 3). Regarding the most recurrent combinations of mycotoxins/metabolites in TMRs, the co-occurrence among aurofusarin, beauvericin, beauvericin A, bikaverin, fusaric acid, and siccanol were higher than 75%. The mycoestrogen ZEN presented a high degree of co-occurrence with mycotoxins DON (84%), FB1 (84%), FB2 (68%), and NIV (68%). Remarkably, more than half of the TMR samples presented co-occurrence of the Alternaria-derived mycotoxins alternariolmethylether and tenuazonic acid and the Aspergillus-produced flavoglaucin and phenopyrrozin; questiomycin and its derivate also co-occurred with the fusarial mycotoxins DON, ZEN, FB1 aurofusarin, beauvericin, beauvericin A, bikaverin, and enniatin B1 (Figure 4).
The co-occurrence rates of estrogenic compounds (for instance, phytoestrogens and mycoestrogens) and other detected plant secondary metabolites (abscisic acid, anisodamine, atropine, and hyoscine) in TMRs are shown in Figure 5. All tested samples presented cooccurrence among the phytoestrogens daidzein, daidzin, genistein, genistin, glycitin, and onionin and the fusarial mycoestrogen ZEN. The occurrence of these mentioned estrogenic compounds with the Alternaria mycoestrogens alternariol and its monomethylether corresponded to 11% and 42%, respectively. The co-occurrence rates of estrogenic compounds (for instance, phytoestrogens and mycoestrogens) and other detected plant secondary metabolites (abscisic acid, anisodamine, atropine, and hyoscine) in TMRs are shown in Figure 5. All tested samples presented co-occurrence among the phytoestrogens daidzein, daidzin, genistein, genistin, glycitin, and onionin and the fusarial mycoestrogen ZEN. The occurrence of these mentioned estrogenic compounds with the Alternaria mycoestrogens alternariol and its monomethylether corresponded to 11% and 42%, respectively.  Table S1).

Figure 2.
Scatter plots illustrate the grade of co-contamination (number of metabolites/sample) by group, whole-plant corn silages (in yellow) or total mixed rations (in gray) from Mexico. Asterisks (*) confirm significant differences (p-value < 0.05) between the number of metabolites per sample in the respective group, whole-plant corn silages or total mixed rations, according to the Wilcoxon matched-pairs signed rank test (p-values in Table S1).

Relationship between Concentrations of Mycotoxin/Metabolite Groups and the Dietary Ingredients
Spearman's correlation coefficients (rho(ρ)) among groups of metabolites detected in total mixed rations with the main ingredients of the TMRs are shown in Figure 6

Relationship between Concentrations of Mycotoxin/Metabolite Groups and the Dietary Ingredients
Spearman's correlation coefficients (rho(ρ)) among groups of metabolites detected in total mixed rations with the main ingredients of the TMRs are shown in Figure 6. The respective p-values of the correlation coefficients are presented in the supplementary Table  S2. Correlations with individual mycotoxin, phytoestrogen, and tropane alkaloid levels were also assessed (data not shown); strong and moderate correlation coefficients and their respective p-values are presented in the text. A moderate positive correlation was observed between the total proportion of concentrate with total ergot alkaloids (ρ = 0.63, p-value = 0.038), the ergot alkaloid dihydroergosine (ρ = 0.63, p-value = 0.006), and the penicillium-derived metabolite pestalotin (ρ = 0.59, p-value = 0.008). The dietary content of corn stover presented a moderate positive correlation with levels of Fusarium-derived metabolites (ρ = 0.63, p-value = 0.0173), beauvericin (ρ = 0.60, p-value = 0.007), total mycotoxins (ρ = 0.52, p-value = 0.0220), and total fungal metabolites (ρ = 0.50, p-value = 0.0305). Sorghum silage showed a moderate positive correlation with the total Fs levels (ρ = 0.   Table S2.

Discussion
This investigation describes for the first time the occurrence of mixtures of mycotoxins, phytoestrogens, and other secondary metabolites in the WPCSs and TMRs of dairy farms in Mexico. The presented results confirmed the ubiquitous presence of mycotoxin mixtures in feeds and complete rations of dairy cows, as indicated in previous studies [2,22,[39][40][41]. The multi-mycotoxin approach used in this study showed that previous reports on mycotoxin contamination are underestimations, as demonstrated by the mixtures fluctuating from 9 to 31 different mycotoxins (13 to 43 total fungal metabolites and 31 to 66 total secondary metabolites) per ration, evidencing the realistic scenario of simultaneous dietary exposition of dairy cattle to multiple mycotoxins and endocrine disruptors. Fusarium-derived mycotoxins/metabolites represented the most relevant fungal metabolites considering the high co-occurrence rates and levels in both WPCSs and TMRs. Our outcome confirms the importance of Fusarium spp. as a primary contributor to contamination with mycotoxins (such as ZEN, DON, and Fs), emerging mycotoxins (such as beauvericin), and other less studied metabolites in dairy cattle feeds, which have also been described in other regions such as South America [42], Europe [2,5,43], and Asia [22].  Table S2.

Discussion
This investigation describes for the first time the occurrence of mixtures of mycotoxins, phytoestrogens, and other secondary metabolites in the WPCSs and TMRs of dairy farms in Mexico. The presented results confirmed the ubiquitous presence of mycotoxin mixtures in feeds and complete rations of dairy cows, as indicated in previous studies [2,22,[39][40][41]. The multi-mycotoxin approach used in this study showed that previous reports on mycotoxin contamination are underestimations, as demonstrated by the mixtures fluctuating from 9 to 31 different mycotoxins (13 to 43 total fungal metabolites and 31 to 66 total secondary metabolites) per ration, evidencing the realistic scenario of simultaneous dietary exposition of dairy cattle to multiple mycotoxins and endocrine disruptors. Fusarium-derived mycotoxins/metabolites represented the most relevant fungal metabolites considering the high co-occurrence rates and levels in both WPCSs and TMRs. Our outcome confirms the importance of Fusarium spp. as a primary contributor to contamination with mycotoxins (such as ZEN, DON, and Fs), emerging mycotoxins (such as beauvericin), and other less studied metabolites in dairy cattle feeds, which have also been described in other regions such as South America [42], Europe [2,5,43], and Asia [22].
Mexican regulations establish maximum limits only for AFs in cereals and cereal products. The maximal levels for AFs for humans are 20 µg/kg and for cattle consumption 100-300 µg/kg [44,45]. No limits or guidance levels are set for other mycotoxins. Because Mexican regulations on mycotoxins in animal feed are outdated and not strict enough, they should be updated accordingly [46]. As reference values for this discussion, the advisory limits/guidance values established by the FDA and EU Commission [19,47,48] will be considered. On an 88% DM basis, the FDA sets for complete rations of dairy animals levels of 5000 µg DON/kg, 30,000 µg Fs (FB1 + FB2 + FB3)/kg, and 20 µg AFs/kg. For OTA (and others ochratoxins) as well as ZEN, the FDA has no regulatory limits [47,48]. The EU Commission recommends 500 µg/kg for ZEN and 5000 µg/kg for DON [19]. Previous studies on mycotoxins in Mexico cattle feeds focused mainly on classic mycotoxins such as AFs, Fs, OTs, ZEN, and DON [11,49]. In the case of the individual mycotoxin levels such as DON (max: 1670 µg/kg), ZEN (max: 246 µg/kg), and total Fs (max: 1670 µg/kg), no sample presented contamination levels higher than the FDA or EU Commission's regulatory levels. However, the sum of related toxic metabolites can be higher than such regulatory limits, for example, the sum type B trichothecenes (amount of DON, 15-acetyl-DON, DON-3glucoside, and NIV). The highest concentration of total type B trichothecenes detected in TMR samples was 5510 µg/kg (equivalent to 6405 µg/kg on an 88% DM), which is above the levels of the single parent mycotoxin included in the legislation (DON). This evidence shows that although the analysis of individual analytes can be below the guidance levels, the total content of related metabolites such as modified mycotoxins, in this case type B trichothecenes, can be above the guidance value, representing a risk. Our results confirm that mycotoxin regulations target the tip of the iceberg if we consider the multiple mycotoxins co-occurrence (i.e., the real-world scenario), as suggested previously [2,4] [55].
ZEN also presented higher occurrence and median in Mexican (100%; 15.2 µg/kg) than in European (67.7%; µg/kg) WPSCs. However, the ZEN maximum level detected in a WPCS sample from Europe (1670 µg/kg) was around 6 times higher than that reported in the present study (278 µg/kg). Although our investigation did not detect AFB1 and OTA, as initially expected, it is essential to clarify that these fungal compounds with carcinogenic properties have been widely reported in dairy cattle feeds (such as cereals) and dairy products (AFM1) in Mexico, representing a real and latent veterinary and public health risk in this country [12,46,49,[57][58][59][60]. Averufin, sterigmatocystin, and versicolorin C, considered to be possible precursors of AFs [61][62][63][64][65], were detected. Sterigmatocystin was previously reported in Mexican maize [59]. Like AFs, sterigmatocystin is known to be a carcinogenic compound with immunotoxin and immunomodulatory activity. Data on the exposure of dairy cows and other animals to sterigmatocystin and the related toxicological implications are limited [66][67][68].
Our results highlight teanuazonic acid as one of the most abundant Alternaria mycotoxins in TMRs. However, animal epidemiological and toxicological information on Alternaria-produced toxins (e.g., alternariol, alternariolmethylether, and teanuazonic acid) is still required. Health risks associated with Alternaria toxins in feeds must be investigated and clarified [69]. We also detected the Penicillium-derived compound mycophenolic acid, mainly related to post-harvest contamination during the ensiling process [8,40,70]. Previous studies showed average levels of mycophenolic acid of 54 µg/kg and 47.5 µg/kg in TMRs from the Netherlands and Austria, respectively [2,71]. The mean of the evaluated TMR samples (32 µg/kg) was lower than the cited European reports. Additionally, kojic acid, produced primarily by Aspergillus spp. but also by some Penicillium spp. [72], has demonstrated antibacterial and immunomodulatory activity [73][74][75]. Citrinin, which is primarily Penicillium-derived [76] but also produced by some Aspergillus spp. [77], was also found. Moreover, several less known metabolites produced by other fungi were detected in TMRs. Some of them, such as cercosporin, the illicicolins, and cytochalasins, have antibacterial activity [78][79][80][81]. The diversity of mycotoxins and fungal secondary metabolites detected in TMRs is due to their multi-commodity composition (Table 1).
Concerning the risk associated with toxicological interactions of mycotoxins [4,82], this study demonstrated a high occurrence of a wide variety of mycotoxins (most of them not considered in legislation at the international level) and other fungal secondary metabolites in the TMRs of dairy cattle. In addition, our findings also showed that phytoestrogens constituent a class of metabolites ubiquitously contained in dairy cow rations. The concern in veterinary medicine and public health related to phytoestrogens is due to their endocrine-disrupting activity. These estrogenic compounds are found primarily in Leguminosae plants, such as clovers (Trifolium spp.), alfalfa (Medicago sativa), and soybeans (Glycine max), and they can act as endocrine disruptors, impairing the reproductive performance of livestock [25][26][27][28][29][30][31][32][33][34][35][36][37]83,84]. In TMR samples, the phytoestrogens that most occurred and the highest concentrations presented were isoflavones such as genistin, daidzein, glycitin, and daidzein (Table 3). However, coumestrol, which is reported to be more potent in estrogenic activity than isoflavones, presented concentrations below the reported critical range (18-180 mg/kg) [85]. The interaction of phytoestrogens with other estrogenic xenobiotics (such as mycoestrogens) is currently the focus of interest [86][87][88]. In this study the co-occurrence of these estrogenic compounds with mycoestrogens such as ZEN, alternariol, and alternariolmethyether was corroborated in TMRs of dairy cattle ( Figure 5), matching previous results of a similar survey carried out in Austria [2]. Along with the mentioned phytoestrogens, other plant-derived compounds detected in TMRs of dairy cows were the phytohormone abscisic acid [89] and the tropane alkaloids anisodamine, atropine, and hyoscine [90]. These alkaloids can have a wide range of biological activity (e.g., anticholinergic effects) and are mostly detected in high concentrations in plants belonging to the Solanaceae and Erythroxylaceae families [91]. These tropane alkaloids were previously detected in cattle feed from Tunisia and Spain in lower concentrations [92] than those presented here. However, according to a scientific opinion of the panel on contaminants of the European Food Safety Authority (EFSA), toxicosis due to tropane alkaloids in livestock is relatively rare [93]. We consider the presence of these alkaloids in TMRs to be a consequence of the existence of native Solanaceae weeds in the feed crops of Mexican dairy cattle. Due to the detected occurrences (≤16%) and low concentrations (<300 µg/kg) in TMR samples, these alkaloids seem not be a risk for the fed cattle.
Our results revealed corn stover as the most correlated ingredient with the content of total mycotoxins, Fusarium-derived metabolites, and fungal metabolites ( Figure 6). Corn stover is the stalks, leaves, and husks that remain in the field after corn harvest [94]. It has been reported as a source of abundant exposure to Fusarium mycotoxins such as Fs, ZEN, and DON [95,96]. The content of ergot alkaloids correlated to the proportion of concentrate in the diet, confirming previous reports that related cereal grains with ergot alkaloids [97]. The proportion of sorghum silage in the rations presented the highest correlation with total content of Fs but also of FA2, FB2, FB3, FB4 hydrolyzed FB1, and citrinin. A previous study performed in the state of Nuevo León, Mexico, evidenced a contamination rate by Fs of 62% [98]. In Uruguay, it was found that 40% of the freshly harvested samples of sorghum presented contamination with Fs [99]. In Brazil, the occurrence of FB1 in sorghum was 74% [100]. These reports demonstrated that Fs contamination is common in this crop. In contrast to prior investigations in other regions such as Europe and South America [2,26,43,71], our results do not suggest WPCSs as one of the most contributing feedstuffs to mycotoxin/metabolite contamination. Concerning the correlations between the dietary ingredients and the levels of mycotoxins/metabolites, it is crucial to consider that more consistent association and relationship assessments require higher sample sizes and additional studies.
The complex mixtures of different mycotoxins, phytoestrogens, and other metabolites evidenced in the WPCSs and rations of dairy cattle in Mexico indicate, along with previous reports/studies, that unexplored and unpredictable toxicological interactions, such as synergistic as well as antagonistic toxic effects, are happening. Extensive studies using a multi-metabolite approach should be performed in other Mexican regions and other Latin American countries on dairy feed and other animal feed but also food for human consumption, including animal-derived products such as dairy products. More governmental interest and research are essential to ensure the safety of animal feed and derived foods, which will support animal health and the productive potential of herds, as well as the delivery of safe products to consumers.

Conclusions
This study demonstrated the ubiquitous contamination of WPCSs and TMRs by a wide spectrum of mycotoxins/metabolites (derived from the genera Fusarium, Alternaria, Aspergillus, and Penicillium) and endocrine disruptor compounds such as phytoestrogens and other metabolites in Mexico. Overall, Fusarium-produced mycotoxins and metabolites were the dominant fungal contaminants. In the assessed TMR samples, ZEN was found with a frequency of 100%, Fs of 89%, and DON of 84%. Although the detected individual levels of the classic mycotoxins (ZEN, DON, FB1, and FB2) were below the maximum/guidance values of Mexican, EU, and FDA regulations, the fact that multiple (regulated, modified, and emerging) mycotoxins co-occurred in complex mixtures, fluctuating from 9 to 31 toxins per sample, should cause concern. Most detected mycotoxins/metabolites are not well studied; their effect as mixtures and their toxicological implications have not been determined. Long-term and subclinical effects on herds' health, production, and reproduction produced by complex mixtures of toxins and endocrine disruptors are unpredictable and require more research. Regarding the ingredients that represent more risk for mycotoxin contamination in TMRs, corn stover was the most correlated feedstuff to high total mycotoxins levels, and sorghum silage was most correlated to Fs contamination. Our results also revealed that dietary concentrate proportion had the strongest correlation to ergot alkaloid contamination in the TMRs of Mexican dairy cattle.

Sampling and Sample Preparation
Representative samples of TMRs and WPCSs were collected from 19 dairy farms in 5 states in northern and central Mexico-for instance, Coahuila (5), Guanajuato (3), Hidalgo (1), Jalisco (7), and Querétaro (3) (Figure 7). The average herd size of the participating farms was 1512 (SD ± 986) lactating cows, varying from 100 to 3500 lactating cows. The main cattle breed of the farms was Holstein-Friesian. Each representative sample of TMRs and WPCSs consisted of at least of 30 incremental samples. Data on the TMR formulation (most important ingredients and their respective proportions) were collected via personal interview (questionnaire-guided).

Sampling and Sample Preparation
Representative samples of TMRs and WPCSs were collected from 19 dairy 5 states in northern and central Mexico-for instance, Coahuila (5), Guanajuato dalgo (1), Jalisco (7), and Querétaro (3) (Figure 7). The average herd size of the p ing farms was 1512 (SD± 986) lactating cows, varying from 100 to 3500 lactating c main cattle breed of the farms was Holstein-Friesian. Each representative sample and WPCSs consisted of at least of 30 incremental samples. Data on the TMR form (most important ingredients and their respective proportions) were collected via interview (questionnaire-guided). The samples were manually collected with gloves from the feed bunk dire the serving (TMR) (according to Penagos-Tabares et al., 2022 [2]) and from opened and "ready to be fed" WPCS bunker silos (according to McElhinney et [101]). The amount of composited samples (> 30 incremental samples) was 1 -1.5 lected samples were homogenized (properly manually mixed), vacuum-pack stored in the dark at − 20 °C until sample preparation. Sampling was carried ou the period of July-August of 2022. For the sample preparation, TMR and WPCS were air-dried (at 65 °C for 48 h) and the whole samples were subsequently m final particle size < 0.5 mm using a mill (Hamilton Beach Model 80335R, Hamilt Brands Inc., China). Finally, aliquots of 5 grams (± 0.01 g) of each homogenize sentative sample were designed for analysis. The samples were placed into 50 m propylene conical tubes (Sarstedt, Nümbrecht, Germany) and sent to Tulln an de Austria, for multi-metabolite analysis. The sample preparation was carried out a boratory of Animal Nutrition of Facultad de Estudios Superiores Cuautitlán, M Veterinaria y Zootecnia (UNAM), located in Cuautitlán Izcalli, México.

Multi-Mycotoxin Analysis (LC-ESI-MS/MS)
The validated multi-metabolite (>800) liquid chromatography/electrospra tion-tandem mass spectrometric (LC/ESI-MS/MS) method was carried out at the The samples were manually collected with gloves from the feed bunk directly after the serving (TMR) (according to Penagos-Tabares et al., 2022 [2]) and from already-opened and "ready to be fed" WPCS bunker silos (according to McElhinney et al., 2016 [101]). The amount of composited samples (>30 incremental samples) was 1-1.5 kg. Collected samples were homogenized (properly manually mixed), vacuum-packed, and stored in the dark at −20 • C until sample preparation. Sampling was carried out during the period of July-August of 2022. For the sample preparation, TMR and WPCS samples were air-dried (at 65 • C for 48 h) and the whole samples were subsequently milled to a final particle size < 0.5 mm using a mill (Hamilton Beach Model 80335R, Hamilton Beach Brands Inc., China). Finally, aliquots of 5 grams (±0.01 g) of each homogenized representative sample were designed for analysis. The samples were placed into 50 mL polypropylene conical tubes (Sarstedt, Nümbrecht, Germany) and sent to Tulln an der Donau, Austria, for multimetabolite analysis. The sample preparation was carried out at the Laboratory of Animal Nutrition of Facultad de Estudios Superiores Cuautitlán, Medicina Veterinaria y Zootecnia (UNAM), located in Cuautitlán Izcalli, México.

Multi-Mycotoxin Analysis (LC-ESI-MS/MS)
The validated multi-metabolite (>800) liquid chromatography/electrospray ionizationtandem mass spectrometric (LC/ESI-MS/MS) method was carried out at the Institute of Bioanalytics and Agro-Metabolomics of the University of Natural Resources and Life Sciences, Vienna, located in Tull an der Donau, Austria, according to previous descriptions. Water purification was completed using a Purelab Ultra system (ELGA LabWater, Celle, Germany). Glacial acetic acid (p.a.) and ammonium acetate (LC-MS grade) were bought from Sigma-Aldrich (Vienna, Austria). HiPerSolv Chromanorm HPLC gradient grade acetonitrile was purchased from VWR Chemicals (Vienna, Austria), and LC-MS Chromasolv grade methanol was acquired from Honeywell (Seelze, Germany). Standards of >800 fungal, plant, and unspecific secondary metabolites were supplied by several research institutions or commercial providers and are listed in Supplementary Table S3. For simultaneous quantification of multiple metabolites, 5 grams (±0.01 g) of each TMR and WPCS sample was extracted in 20 mL of the extraction solvent (acetonitrile/water/acetic acid 79:20:1, v/v/v) following the procedures reported by Steiner et al. (2020) [102]. These volumes were placed into the QTrap 5500 LC-MS/MS system (Applied Biosystems, Foster City, CA, USA) equipped with a TurboV electrospray ionization (ESI) source coupled to a 1290 series UHPLC system (Agilent Technologies, Waldbronn, Germany). Subsequently, quantification from external calibration by serial dilutions of a stock solution of analyzed compounds was accomplished. Finally, the outcomes were adjusted for apparent recoveries defined through spiking experiments, according to Steiner et al. (2020) [102]. This analytical methodology has been validated [96] and used to study the occurrence of multiple metabolites in complex feedstuff matrices such as silages, pastures, concentrates, and TMRs [2,5,22,39,56]. The method accuracy has been verified on a routine basis by proficiency testing organized by BIPEA (Genneviliers, France). Satisfactory z-scores between −2 and 2 have been achieved for >95% of >1800 results submitted so far. Supplementary Table S4 presents performance values of LC/ESI-MS/MS analysis for mycotoxins, phytoestrogens, and other fungal, plant, and unspecific metabolites detected in WPCSs and TMRs.

Data Analysis
Concentrations of metabolites were presented in µg/kg on a DM basis. Descriptive statistics (i.e., occurrences and the average, median, and range of the concentrations) were processed considering only the positive values (x ≥ limit of detection (LOD)) using Microsoft ® Excel ® . Values lower than the limit of quantification (LOQ) were calculated as LOQ/2. The normality assessment of the data was completed via the D'Agostino and Pearson test, Anderson-Darling test, Shapiro-Wilk test, and Kolmogorov-Smirnov test. All the tests indicated the non-normal distribution of the handled data. Considering the dependence, the differences between concentrations of metabolites in TMRs and WPCSs of each respective farm were assessed via the (nonparametric) Wilcoxon matched-pairs signed rank test, and statistical differences were considered significant at p-value < 0.05. The co-occurrence analyses of mycotoxins and plant metabolites were performed separately using Microsoft Excel, generating matrices plotted in heatmaps. Moreover, a two-tailed Spearman's correlation test was conducted to explore possible relations among dietary ingredients and levels of metabolites. Spearman's correlation coefficients were considered significant at a p-value < 0.05. Accordingly, the correlation coefficients were interpreted according to Hinkle et al. 2003 [103]: "very high" (0.90 up to 1.00), "high" (0.70 up to 0.90), "moderate" (0.50 up to 0.70), "low" (0.30 up to 0.50), and "negligible" (<0.30). Low and negligible correlations were not considered for the interpretation of the results. The statistical analyses and graphs were completed using GraphPad Prism version 9.5 (GraphPad Software, San Diego, CA, USA).

Supplementary Materials:
The following supporting information can be downloaded at: https:// www.mdpi.com/article/10.3390/toxins15020153/s1, Figure S1: Distribution of concentration (µg/kg DM, linear scale) of (a) groups of fungal metabolites and mycotoxins and (b) total phytoestrogens and plant metabolites detected in in whole-plant corn silages (yellow) and total mix rations (gray) in dairy farms in Mexico. Asterisks (*) show significant differences (p-value < 0.05) between the concentrations of the respective groups in whole-plant corn silages and total mixed rations according to the Wilcoxon matched-pairs signed rank test (p-values in Table 2). Means are shown as "+"; Table S1: Description of the co-contamination level of the diverse groups of analysis detected in whole-plant corn silages and total mixed rations of Mexican dairy farms; Table S2: p-values of the Spearman's correlation coefficients (ρ) among groups of metabolites detected in total mixed rations with the main dietary ingredients. Significantly different (p-value < 0.05) presented in black cells; Table S3: List of 863 targeted metabolites to analyze whole-plant corn silages and total mixed rations from Mexican dairy farms via a validated multi-metabolite liquid chromatography/electrospray ionization-tandem mass spectrometry (LC/ESI-MS/MS); Table S4: Performance values of liquid chromatography/electrospray ionization-tandem mass spectrometry (LC/ESI-MS/MS) analysis for mycotoxins, phytoestrogens, and other fungal, plant, and unspecific metabolites detected in whole-plant corn silage and total mixed rations of dairy cattle in Mexico.