Study area

The present study was performed with a descriptive, longitudinal and non-experimental design. A dairy production unit was selected for convenience, using the non-probabilistic method, and was followed for 24 mo. The DPU was located in the Central High Plateau of Mexico (21°48'N, 102°03' W; 1986-2008 m asl), with a dry temperate and a semi-warm semi-dry climate, with summer rains, average temperature of 18.4 º C, an annual rainfall of 518.4 mm, and a maximum altitude of 2,300 m asl.

During the observation period, the DPU had an average of 2,749 Holstein cows housed in confinement within open-air pens enclosed with metal fences, shaded areas and freely accessible feeders. The cows were milked with automated equipment, obtaining an average daily production of 28.1 L per cow. The milk production is exported to agro-industrial plants in the region. Diets were developed as a total mixed ration (TMR) in a feed mill, where the corn silage and the concentrate were blended in mixer wagons. The TMR release is formulated to meet the nutritional requirements of dairy cows. The corn for silage is obtained directly from the agricultural areas of the UPL, while the protein-energy concentrate was purchased from the Local Association of Dairy Farmers of Aguascalientes, using as main ingredients canola, soybean, corn, rolled, sorghum, soybean meal, corn distilled dry grains, cottonseed, alfalfa and oats hay, as well as premixed vitamins and minerals.

Sample collection and handling

A total of 288 samples (1.0 kg) were obtained: total mixed ration (TMR), concentrated feed and corn silage were dried up in a forced air circulation oven (OF-22G-TECH, JEOI Lab Companion, Corea). The samples were pulverized (500-800 µm) in a continuously operating universal mill (MF series 10 Basic, IKA®-Werke, Germany) and stored in cooling system (4-5 °C) until processing (<7 d). During the (morning and evening) milking, a total of 288 samples of raw milk were obtained directly from the collector tank (500 ml) for each production batch (high, medium and low). The samples were transported in refrigeration and were preserved in freezing (-20 °C) until processing (<7 d).

Five agricultural plots were sampled with the help of a soil punch, choosing four sampling points on the surface of each plot of land and taking five 100 g sub-samples at each point, at a depth of 3-30 cm. The five sub-samples were gathered in a zip lock bag. Finally, they were screened (500-800 µ) and kept in refrigeration until processing.

Quantification of aflatoxins

Food samples were analyzed according to the official method 990.33 of the AOAC²¹, using solid phase columns (SPE; SupelcleanTM LC-18 SPE tube, Sigma-Aldrich, USA). The eluate extracted from the samples derivatized with trifluoroacetic acid was analyzed by HPLC (detection limit 2 µg/kg) with a fluorescence detector (binary Pump Vary Pro Star; FP 2020 detector, Varian Associates Inc., Victoria, Australia), a C18 column, and a column guard (LC-18 LC-18; Thermo Fisher Scientific, Massachusetts, USA). The quantification data were obtained using the Galaxie software (version 1.9.302.530), and AF concentrations were calculated using standard curves of purified AFs (Sigma-Aldrich, St. Louis, MO, USA). The AFM1 was quantified in raw milk with the competitive ELISA technique using a commercial kit (Ridascreen fast® aflatoxin M1 R-1121, R-Biopharm, Germany; detection range 0.005-0.08 µg/kg). The samples were homogenized and centrifuged according to the manufacturer's instructions. The absorbance was measured at 450 nm in an ELISA microplate reader (BioTek Instruments, Inc., USA). The results were interpreted upon the basis of the calibration curve made with purified 1 AFM (Sigma-Aldrich, St. Louis, MO, USA).

Fungal Isolation and characterization

The fungi were isolated using the pour plate technique with four serial dilutions in sterile peptone water²²; the samples were seeded in a potato, dextrose and agar (PDA), malt extract with rose Bengal and Czapeck medium, and incubated (28 °C) in the darkness during 7 days. The colonies were isolated and purified, and the microscopic morphological characteristics consistent with the description of the genus were identified¹³. Isolates consistent with A. flavus^8,13 were submitted to the process of fixating with glutaraldehyde (2%), gradual alcohol dehydration, and conditioning with a critical point dryer (Samdri Tousimis Research 795, Rockville, Meryland) and a metalizer (Desk II, Denton Vacuum, USA) were digitized in a scanning electron microscope (JEOL JSM-5900 LV,JEOL, USA) and conducted 10 measurements of each structure (stipe, gallbladder, spore, esclerocio and fialide), and using a JEOL software (Scanning Electron Microscope). The morphological structures of the isolates identified were compared against known strains of A. flavus, known as AF-36, AF-Cuatitlan (AF-C) and AF-Tamaulipas, (AF-T)^(14.23). The strains obtained in this study were recorded at the NCBI (National Center for Biotechnology Information) with the respective access codes.

The aflatoxigenic capacity of the isolates was characterized by thin-layer chromatography (TLC)²⁴; silica gel plates without fluorescence indicator were utilized (Z265829, Sigma-Aldrich, USA) activated in a high temperature oven (OF-22G-TECH, JEOI Lab Companion, Corea). The plates with purified standards of AFs (6636-50MG, Sigma Aldrich, USA) were placed inside of a chromatographic chamber with mobile phase chloroform - acetone - isopropanol (85:10:5, v/v/v) for an hour and a half. Eventually, the dry plates were visualized in a transilluminator. The technique ammonium vapors in coconut agar technique was also used according to the methodology described above²⁵. Monosporic isolates were inoculated in sterile coconut agar and were left to incubate in the dark (30 °C, 5-7 d). Subsequently, ammonium hydroxide (200 µl; J.T. Baker, Mexico) at 25 % was added to the cover of the Petri dishes, and the distribution and the intensity of the color was observed. The presence of color was taken as indicative of AF production.

Molecular analysis

The genomic DNA of monosporic cultures of Aspergillus spp. was extracted according to previously standardized methods²⁶. The quality of the obtained DNA was displayed with electrophoresis (45 min to 85 volts) in agarose gel (1%) with a TAE 1X damper and was quantified by comparing against known concentrations of DNA of phage λ (Thermo Fisher Scientific, MA USA). The DNA visualization was performed using a device for photoanalysis (Bio-Rad Gel Image®- Doctm XR, CA USA) with the Quantity One software version 4.6.7.

An amplification of a fragment corresponding to the region of internal spacers of the transcript (ITS; ITS1-5.8S-ITS2 RNAr) and ITS4 (5´-TCCTCCGCTTATTGATATG -3´) was carried out in accordance with the previously described protocols^27,28; the gene of calmodulin (CaM) was amplified with the primers CMDA7F (5'-GCCAAAATCTTCATCCGTAG-3') and CDMA8R (5'-ATTTCGTTCAGAATGCCAGG-3') and the regulatory gene of the aflatoxin biosynthetic pathway (aflR) was amplified using the primers aflR-F (5'-GGGATAGCTGTACGAGTTGTGCCAG aflR-3') and aflR-R (5'-TGGKGCCGACTCGAGGAAYGGGT-3') of Eurofins Genomics, Louisville KY, USA. The polymerase enzyme Go-Taq (Promega, Madison, WI USA) and a model 9700 thermal cycler (Applied Biosystems) were utilized in the amplification. The PCR products obtained (ITS, calmodulin and aflR) were separated by electrophoresis in (1%) agarose gel and they were observed using as intercalating agents SYBR® Gold and Orange DNA Loading Dye (Thermo Fisher Scientific, MA USA); the resulting bands were observed in an image analysis device (BIO-RAD IMAGE® GEL Molecular- DOCTM XR CA, USA) with the Quantity One software (version 4.6.7). Molecular weight marker ladders (Axygen Biosciences, CA, USA) were included. PCR products were purified with the reagent ExoSAP-IT® PCR Product Cleanup (Afflymetrix, Thermo Fisher Scientific Inc. Santa Clara, California, USA). The purified PCR products were sequenced in forward and reverse chains with the dideoxy method²⁹. The samples were injected into a sequencer (ABI 3730XL Genetic Analyzers), and the resulting sequences were recorded in an electropherogram. The electropherograms were visualized with the Chromas Lite software and were compared with the records of the NCBI using BLAST (Basic Local Alignment Search Tool).

Statistical analysis

The quantitative data of the total mixed ration, milk production, AF concentration in feed and milk, and of the measurements of the structure of each isolate were subjected to a one-way variance analysis (ANOVA), considering as separate factors the time of year and the level of milk production (high, medium or low) in which the cows were classified in batches. The differences between the means and the 95 % confidence intervals were determined using the honestly significant difference (HSD) Tukey’s test; P<0.05 was considered significant for all statistical analyses.

Frequency of Aspergillus spp

It was identified a total of 283 fungal isolates; 56.8 % came from samples of feed for dairy cows, and the rest, from the agricultural land (Table 1). A total of 88 isolates (31.1 %) exhibited morphological features corresponding to those described for the genus Aspergillus. Figure 1 shows the conidiophores and conidia, and the characteristic metulae. Isolates were also found with a morphology compatible with the genera Penicillium, Fusarium, Rhizopus, Mucor, Cladosporium, Richoderma, Alternaria, Curvularia and Bipolaris. The proportions of each gender are shown in Table 1.

Panels: A) Conidiophore: Cc = conidial head, Ee = Stipe, Pe = foot cell, Mo = mycelium. B) conidial head uniseriate portions: Co = Conidium or spore; Fe = Phyalid, Va = Vesicle C) Biseriate conidial head: Co = Conidium, Fe = Phyalid , Va = Vesicle, i = Metula.

Figure 1 Aspergillus morphological structures 

Only six isolates exhibited a morphology consistent with A. flavus and came from the feed concentrate (AC1, AC2 and AC3) and corn silage (EM1, EM2, EM3); when cultivated in PDA, they exhibited an olive green hue with a whitish periphery. Most of the isolates (5/6) exhibited the presence of sclerotia of dark brown color; only one did not manifest these structures. At a microscopic level, conidiophores of A. flavus exhibited irradiate and uniseriate conidial heads, the stipe with rough walls, the spherical vesicle, globe-shaped conidia with irregularities in their surface, and a whitish, septate mycelium (Figure 2).

Panels: A) Cc = conidial head, Ee = stipe. B) Va = vesicle. C) Conidiophore. Fe = phyalids, Co = conidium. D) Cc = conidium, Fe = phyalid. E) Mo = mycelium, So = septum. F) Eo = Sclerotium.

Figure 2: Morphological structures of Aspergillus flavus 

The six isolates of A. flavus from feed exhibited significant differences in the morphometry of their structures when compared against the control strains (AF-C and AF-T; Table 2). These strains presented conidial heads, with rough radiated stipe and spherical vesicles, while AF-C only exhibited (uniseriate) phyalids and (biseriate) AF-T phyalids metulae; the conidia were globe-shaped with irregularities on the surface. Four strains isolated were classified as L (long sclerotium, >400 µm), one as S (short sclerotium, <400 µm), and one as without sclerotium.

The isolates identified as A. flavus were analyzed using TLC and ammonium vapors in coconut agar techniques in order to identify the presence of aflatoxin (Figure 3). Five isolates (AF-AC1, AF-AC2, AF-AC3, AF-EM2 and AF-M3) had a positive reaction to the presence of AF, while the other (AF-EM1) produced no aflatoxins, with both techniques. These results coincided with the amplified aflR gene.

Panels: A) AF-36, non-aflatoxigenic (AF-). B) AF-Cuatitlán, and C) AF-Tamaulipas, aflatoxigenic (AF+). D) AF-EM1, AF- isolate from corn silage. E) AF-AC2, AF+ isolate from concentrated feed. F) AF-EM3, AF+ isolate from corn silage.

Figure 3 Intensity of color in a coconut agar medium in contact with ammonia vapors of Aspergillus flavus isolates after 4 d of growth 

Molecular analysis

Of the 88 isolates morphologically identified as Aspergillus spp., 49 % were amplified for the calmodulin gene, and 31 % were amplified for ITS (Figure 4). The analysis of the sequences obtained showed that the species were A. oryzae (45.5 %), A. niger (10.2 %), A. ochraceus (3.4 %), A. pseudodeflectus (4.5 %), A. ustus (10.2 %), A. flavus (6.8 %), A. versicolor (5.6 %), A. nidulans (5.6 %), A. sublatus (4.5 %) and A. Sydowii (3.4 %).

Rails: MWM) Molecular weight marker (100pb). Controls: Non-aflatoxigenic A. flavus (AF-36) and aflatoxigenic fungi (Cuautitlan, AF-C; Tamaulipas, AF-T). A. flavus isolates of corn silage (EM1, EM2 and EM3) and feed concentrate (AC1, AC2 and AC3) intended for dairy cows

Figure 4 Electrophoresis in an agarose gel at 1% of the PCR products amplified from the calmodulin gene (Panel A), internal transcribed spacer of the rDNA region 5.8S-ITS2 (Panel B), and regulatory gene of the aflatoxin biosynthetic pathway, aflR (Panel C) 

The five local isolates identified as aflatoxigenic A. flavus amplified a fragment of 796 pb for the aflR gene. The analysis of the sequences of the control strains (AF-36, AF-C and AF-T) and the six isolates of A. flavus (AF-AC1, AF-AC2, AF-AC3, AF-EM1, AF-EM2 and AF-EM3) have been shown to have an identity percentage of over 90% with the isolates of A. flavus registered in the database of the NCBI (Table 3).

Aflatoxins detected in feed and in milk

Of the 288 samples obtained from food, 286 (99.3 %) exhibited detectable levels of AFs (18.5 ± 3.7 µg/kg); of which 10.4 % exceeded the maximum limits allowed by the Mexican law (20 µg/kg). The presence of AFM₁ (0.021 µg/kg) was detected in 39.9% of the 183 samples of raw milk, and 12.0% exceeded the maximum permissible limit used as a standard by the agro-industry which acquired the raw milk (0,050 µg/kg). The highest incidence of AF in feed occurred in summer and autumn (P<0.01), compared to winter and spring (17.4^a ± 3.2 and 14.8^ab ± 1.6 vs 8.1 ^bc ± 0.5 and 5.9^c ± 0.5 µg/kg, respectively). In turn, the concentration of AFM₁ was directly correlated with the concentration of AFs in the TMR in a double squared model (P<0.01; R²= 30.5 %). Also the presence of AFM₁ in milk was significantly correlated with the level of milk production and the consumption of TMR; on average, the cows consumed daily 621 µg of aflatoxins in 42 kg of TMR; had a daily production of 26.2 L of milk with a total load of 0.603 µg of AFM₁ ―which represented a great capacity of dairy cows to metabolize the AF―, and eliminated only a fraction (0.09 %; Table 4); in general, cows with a high production were exposed to the ingestion of larger amounts of AFs (658 µg/cow/day), and the presence and elimination of AFM₁ in the milk of highly productive cows was greater than in cows with a medium and low production (0.22 µg/cow/d).

Frequency of Aspergillus spp. and A. flavus

The fungal genus Aspergilus had the highest occurrence at the DPU (31 %), followed by Penicillium (13.8 %) and Fusarium (12.7 %). These genera have already been identified in Mexico in stored yellow corn⁹ and in corn hybrids intended for human and animal consumption³⁰. It has been suggested that the frequent presence of micotoxigenic fungi and their toxins in corn is due to improper handling of the crop and to ineffective implementation of strategies for the control of infections^5,31.

In this study, A. flavus was isolated from the feed for dairy cows at a ratio of 6.8% of the total identified Aspergillus spp. (6/88). This species was not identified in agricultural soil intended for the alternating production of fodder maize and oats. In Mexico, A. flavus has been identified in commercial corn kernels³² from soil for crops of grain corn¹⁰. This suggests that fungal populations of this species are distributed across the agricultural areas of Mexico.

In this study, the morphometric analysis showed that one-third of the isolates of A. flavus contained abundant short sclerotia (< 400 µm) and were therefore classified as S strains, while the rest were classified as L strains, as they had few but larger sclerotia (>400 µm). The importance of this morphology and its association with the aflatoxigenicity of S strains has already been pointed out; on the other hand, L strains contain both toxigenic and atoxigenic isolates^33,34.

Molecular analysis

All of the 25 sequences obtained from the isolates identified morphologically as A. flavus exhibited a large percentage of identity (> 90%) with registered sequences of A. flavus strains. Amplified fragments were also obtained for the 468 pb CaM gene, for the 796 pb aflP gene, and for the ITS with a range of 600-800 pb (ITS1-5.8S-ITS2). It has been pointed out that the rDNA region of the internal transcribed spacers is the official DNA bar code for fungi because it is the most frequently sequenced marker and is a useful tool for the description of A. flavus species⁸. It has also been said that the CaM gene is able to distinguish between almost all species of Aspergillus⁸. On the other hand, the aflR gene is necessary for the transcription of most genes involved in the activation of enzymatic reactions that are necessary for the formation of aflatoxins^35-37. Based on the above, this work integrates molecular identification to the morphological identification of six isolates of A. flavus obtained from the dairy production chain in the Central High Plateau region of Mexico.

Aflatoxigenic capacity

The six isolates identified as A. flavus were evaluated with TLC and vapors of ammonium in coconut agar in order to identify their capacity to produce aflatoxins; five were classified as aflatoxigenic fungi, and one was classified as a non-producer of aflatoxins. These results evidenced a behavior comparable to those obtained in other studies conducted in Mexico¹⁰ in which A. flavus isolates were identified as genetically related to A. flavus AF-36³². This non-aflatoxigenic strain has been employed in the development of biological control strategies to reduce aflatoxin production of aflatoxigenic strains in cotton-producing fields¹⁵. The above suggests that the main aflatoxigenic A. flavus populations are distributed worldwide across agricultural ecosystems, and that they coexist with strains that have no capacity to produce aflatoxins.

Aflatoxins detected in feed and in milk

A total of 99.3 % (286/288) of the samples of feed for dairy cows had detectable levels of AFs (18.5 ± 3.7 µg/kg), of which 10.4% exceeded the maximum limits allowed by Mexican law (20 µg/kg); furthermore, 39.9 % (73/183) of the raw milk samples exhibited the presence of AFM₁. Comparable proportions of AF contamination have been detected in balanced feeds produced in Mexico⁵, in the total mixed rations of stables in Mexico³⁸, and in samples of feed of Asian origin³⁹. The incidence of AFM₁ has been reported in raw bovine milk in Mexico^6,7; also, incidence of AFM₁ has been identified in milk for human consumption in various countries^3,40-42. This suggests that the aflatoxin contamination of food for human and animal consumption is a global public health issue.

This study showed an association between the level of milk production of cows with the amount and concentration of AFM₁ eliminated in the milk, in such a way that highly productive cows consumed a larger amount of feed and were exposed to larger amounts of aflatoxins in their feed; therefore, the total amount was larger, and the concentration of AFM₁ in raw milk was higher (P<0.05). It has been pointed out that the AFM₁ elimination rate in milk is influenced by the amount of feed consumed per day and the amount of milk produced per day by each animal, as well as by the stage of lactation⁴³.