Mycotoxins in khadi, A Traditional Non-Cereal Based Alcoholic Beverage of Botswana

Abstract

Mycotoxin contamination is a major food safety drawback towards the commercialization of food products. The commercialization of khadi, a popular fermented alcoholic beverage of Botswana necessitates the investigation of the presence of mycotoxins. Khadi brewing involves the uncontrolled and unstandardized spontaneous fermentation of sun-dried Grewia flava fruits, which could be a source of mycotoxin-producing filamentous fungi (molds). This study sought to investigate the presence of mycotoxins producing fungi and mycotoxins in 18 samples of khadi collected in Central and Northern Botswana. Periconia thailandica, Cladosporium cladosporioides, Aspergillus ochraceus, Phoma eupyrena, Setosphaeria turcica, Cladosporium sphaerospermum, Chaetomium longiciliata, and Flavodon ambrosius were identified in 10 out of 18 khadi samples. Mycotoxins were detected using the Myco-10 Randox Evidence Investigator biochip kit and confirmed using a UPLC-ESI-MS/MS. Mycotoxins such as paxilline, ochratoxin A, ergot alkaloids, aflatoxin G1/G2, and zearalenone were detected using the Myco-10 Randox Evidence Investigator biochip kit. The Myco-10 results revealed that the mycotoxins in the khadi samples were lower than the regulatory limits set by FDA or European Commission. Confirmation of results using an UPLC-ESI-MS/MS system involved confirming selected mycotoxins (AFB1, DON. ZEA, FB1, FB2, FB3, NIV, and OTA) from selected khadi samples (Palapye 1, Palapye 2, Letlhakane 2, Maun 3, Mmashoro 3, and Tonota 3). The UPLC results demonstrated that the aforementioned mycotoxins in the selected khadi samples were below the detection thresholds. The study shows that while fungal isolates were present, there is no to minimal danger/risk of exposure to toxic mycotoxins after consumption of khadi. Towards commercialization endeavors, the production process would necessitate minimal mycotoxin monitoring and product preservation but no detoxifying steps.

Introduction

Khadi is a popular traditional and non-cereal-based alcoholic beverage from Botswana, and it is usually prepared under non-standardized conditions. The brewing process involves uncontrolled spontaneous fermentation of sun-dried Grewia flava fruits and table sugar. Such uncontrolled spontaneous fermentation conditions expose the brew to mycotoxin-producing fungi contamination. Mycotoxins are a significant barrier major challenge in traditional brewing as well as a major drawback in commercialization of alcoholic beverages. Mycotoxins are a structurally diverse category of largely tiny molecular weight substances produced primarily by the secondary metabolism of various filamentous fungus, and in significant levels, they pose serious dangers to human health.^1,2 These human health implications include but are not limited to esophageal malignancies,² kidney mycotoxicosis,³ acute aflatoxicosis,¹ carcinogenic,⁴ genotoxic,⁴ teratogenic,^5,6 nephrotoxic,⁷ and hepatotoxic⁴ and others.

Contamination by filamentous fungi is a problem at every stage of the fruit-harvesting process, from pre-harvest to post-harvest storage and processing. Some of the most common fungal genera known to produce mycotoxins are Aspergillus, Alternaria, Fusarium, and Penicillium .^8,9 These mycotoxins often accumulate in the final alcoholic beverage. Concerning mycotoxins include aflatoxins, deoxynivalenol (DON), fumonisins, ochratoxins, citrinin, ergot alkaloids, patulin, and zearalenone (ZEN).^10-13 Mycotoxicity of cereal based traditional alcoholic beverages made in beverages produced from substrates is well documented in the South African market.^10,11,14-16 Despite the fact that mycotoxins have been documented in African traditional alcoholic beverages, the mycotoxin data is lacking. Fatalities after drinking khadi have been reported¹⁷ and evidence shows death of 3 after consuming this alcoholic beverage.¹⁸ Unfortunately, the total of deaths per year for the past decade linked to khadi consumption are not available on formal record. Information on mycotoxins in khadi is crucial for commercialization as it is the first step to allow development of methods to prevent contamination. Despite the importance of keeping records on mycotoxins in foods processed at the household level due to the link between food safety and health, there is currently no law in Botswana governing mycotoxin regulation in traditional beverages. However, some products, including ready-to-eat food (10 ppb), peanuts (15 ppb), and milk (0.05 ppb), are regulated for their levels of aflatoxins.¹⁹ As such, this study set out to conduct mycotoxin analyses on a popular beverage, khadi, as a critical step toward commercialization.

Materials and Methods

Sample collection

Khadi samples were randomly collected from local brewers in Botswana’s towns and villages of Tonota, Palapye, Serowe, Letlhakane, Mmashoro and Maun. Brewers in the southern part of Botswana were excluded because they brew khadi using tubers instead of G. flava fruits. During the sampling process, all the khadi samples were collected at a stage when the khadi was ready for consumption that is after completion of fermentation as was deemed appropriate by each individual brewer. The samples were collected aseptically using 500 mL sterile bottles, transported to the laboratory at a temperature of 0°C to 4°C in a cooler box with ice packs and kept at −20°C until analyses.

Isolation of fungi from khadi samples

Hundred microliters of each aliquot of khadi were 10-fold serially diluted using 1 mL phosphate buffered saline (PBS; g/L: NaCl 8.0, KCl 0.2, disodium phosphate 1.44, and potassium phosphate 0.24), pH 7 and kept on ice. The dilutions (up to 10⁻⁶) were then spread plated on YPD agar (2% glucose, 1% yeast extract, 2% peptone, and 1.5% agar) supplemented with 100 µg/mL of streptomycin, penicillin, and ampicillin. The plates were incubated at 30°C for 5 days. The fungal colonies on the YPD plates were then cut and transferred to malt extract agar (MEA) (Oxoid, England) and potato dextrose agar plates (PDA) (Difco, Detroit, MI) to get pure cultures which were then incubated at 30°C for 5 days. For further study, the purified cultures were stored in 25% glycerol at −80°C.

Identification of fungi from khadi samples

For identification of fungal isolates, we first extracted gDNA using the microwave method.²⁰ In brief, the fungal mat was rinsed with 1 mL TE buffer, centrifuged, and lysed using 100 µL TE buffer and 50 µL 10% SDS. The lysate was incubated for 30 minutes at 65°C and centrifuged to remove the supernatant. The cell pellet was placed in a microwave oven and heated twice for 1 minute at 900 W. The pellet was then re-suspended in 200 µL TE buffer with one volume of phenol:chloroform:isoamyl alcohol (25:24:1) for 15 minutes. The aqueous phase was recovered by centrifugation; DNA was precipitated with 95% ethanol and centrifuged at 12 000 rpm for 20 minutes. DNA was then rinsed with 1 mL 70% ethanol and re-suspended in 200 µL deionized water.

Two microliters of the suspension was then used as a template for PCR amplification of a 560 bp DNA fragment of the D1/D2 region of the 26S rDNA using universal PCR primers, NL1 (5′-GCATATCAATAAGCGGAGGAAAAG-3″) and NL4 (5′-GGTCCGTGTTTCAAGACGG-3″) as reported elsewhere.^21,22 PCR amplification was performed using ProFlex PCR Systems (Applied Biosystems, USA) in a 20 µL reaction volume containing 2.5 µL 10× PCR buffer, 2.0 µM of each primer, 0.2 mM dNTPs, and 1.25 U Taq DNA polymerase (Takara Bio Inc., Japan) using the following cycling conditions: initial denaturation at 98°C for 3 minutes; 38 cycles of denaturation (98°C for 15 seconds), annealing (54°C for 1 minute), and extension (72°C for 1 minute), a final extension step at 72°C for 7 minutes and held at 4°C until required. Sterile water was used instead of template DNA for negative controls that were run alongside test samples. All amplicons were purified using a QIAquick PCR product purification kit (Qiagen, GmBH, Germany) according to manufacturer’s instructions. The amplicons were sequenced by Inqaba Biotech (Pretoria, South Africa). SnapGene^® Viewer software ver. 4.2.11 (GSL Biotech) sequence editing tool was used to generate contiguous sequences (http://www.snapgene.com). Species identification was done by using a BLAST nucleotide sequence analysis tool (https://blast.ncbi.nlm.nih.gov/Blast.cgi?PROGRAM=blastn&PAGE_TYPE=BlastSearch&LINK_LOC=blasthome) as well as using pairwise identification tool available on the Westerdijk Fungal Biodiversity Institute (CBS-KNAW) strain database (http://www.westerdijkinstitute.nl/).

Mycotoxins detection from khadi using the Randox Evidence Investigator

Simultaneous semi-quantitative screening of mycotoxins in feed was carried out using the Myco 7 biochip array kit (EV4065; Randox Food Diagnostics, Crumlin, UK). The kit contains multianalyte biochips, assay diluent, conjugate diluent, multianalyte conjugate, a set of multianalyte calibrators (spanning the range of each assay), multi-analyte controls, signal reagent, washing buffer, calibration disk, and barcodes. The biochips were supplied in carriers (3 × 3 biochips per carrier), and a carrier handling tray was provided with the system that allows the simultaneous handling of 6 carriers (54 biochips). Data were generated and processed with the semi-automated benchtop biochip analyzer Evidence Investigator (EV3602; Randox Food Diagnostics, Crumlin, UK).

Extraction of mycotoxins from khadi samples (Acetonitrile method)

To extract mycotoxins from the samples, a total of 2.5 mL of each khadi sample was mixed with 20 mL pure acetonitrile (Sigma-Aldrich, USA) in a 50 mL centrifuge tube. Sodium chloride (2 g) and magnesium sulfate (0.25 g) were added to the mixture.²³ The resultant mixture was vortexed for 3 minutes followed by centrifugation at 4000 rpm for 10 minutes at 4°C. The supernatant was collected into a new 2 mL centrifuge tube and the pellet was discarded.

Randox Evidence Investigator Myco-10 calibration

The Randox Evidence Investigator was calibrated based on the nine-point calibration curve using Randox Evidence Investigator Myco-10 calibrator (Randox Food Diagnostics) (https://www.randox.com/evidence-investigator-multiple-industries/), which covered the calibration range of all assays. In brief, calibration data was used to calculate the linearity, limit of detection (LOD) and limit of quantification (LOQ) of the calibration process. The LOD is the detection limit of an analytical procedure and is the lowest concentration of analyte in a sample that can be detected while the LOQ is the smallest concentration of an analyte that can be reliably measured by an analytical procedure.

To cover worldwide guidance limits, the overall sample dilution factor could be selected at lower and higher levels (dilution factor 20 for sensitive detection and 250 for monitory level, respectively). The multi-analyte calibrators included all the analytes used for the standardization of each of 7 competitive immunoassays on the biochip at 9 different concentration levels, with the following typical calibration ranges for sensitive detection and monitory level, respectively: 0 to 3.7 and 0 to 46.25 ppb expressed as aflatoxin B1 equivalence for the aflatoxin B1 assay, 0 to 20 and 0 to 250 ppb expressed as aflatoxin G1 equivalence for the aflatoxin G1 assay, 0 to 2000 and 0 to 25 000 ppb expressed as DON equivalence for the DON assay, 0 to 80 and 0 to 1000 ppb expressed as fumonisin B1 equivalence for the fumonisins assay, 0 to 16 and 0 to 200 ppb expressed as ochratoxin A equivalence for the ochratoxin A assay, 0 to 80 and 0 to 1000 ppb expressed as T-2 toxin equivalence for T-2 toxin assay, and 0 to 40 and 0 to 500 ppb expressed as zearalenone equivalence for the zearalenone assay.

Mounting of samples onto Myco-10 biochip

To begin the sample mounting process, 150 µL of diluted wash buffer was pipetted into each of the 9 biochip wells (Myco-10 biochip, Randox Diagnostics, UK). Fifty microliters of each of the extracted samples was added to the appropriate biochip well. Mixing of the samples and the was buffer was done by gently tapping the edges of the handling tray. The trays were placed on a thermal shaker for 30 minutes at 370 rpm and 25°C. After incubation, a working strength solution conjugate (provided in the Myco-10 kit) of 100 µL was pipetted into each biochip well. The biochip was then incubated at 25°C for 60 minutes at a shaking speed of 370 rpm. The biochip was then washed 4 times by pouring 375 µL of wash buffer to each biochip well. Another wash cycle was carried out 4 times and the biochips were finally soaked in wash buffer for 2 minutes. After the final wash, the biochips were tapped onto a lint free paper towel to remove any residual wash buffer. After tapping, the biochip wells were filled with wash buffer to avoid drying out and left to soak until directly prior to imaging.

Myco-10 biochip imaging

Each biochip carrier was imaged individually and biochips awaiting imaging were protected from light by covering them with a lint free paper towel. Each biochip well was tapped onto a lint free paper towel to remove the wash buffer. A signal reagent (provided in the Myco-10 kit) volume of 250 µL was added to each biochip well and covered to protect from light for 2 minutes. The biochip carrier was placed into the Evidence Investigator (EV3602; Randox Food Diagnostics, Crumlin, UK), images from each well were captured and mycotoxins per well quantified. In the Evidence Investigator, the degree of binding is determined using a chemiluminescent light source and then quantified using a charge coupled device (CCD) camera and imaging system. The Randox Evidence Investigator detects and quantifies 10 mycotoxins per sample (paxilline, fumonisin, ochratoxin A, aflatoxin G1/G2, ergot alkaloids, diacetoxyscirpenol, deoxynivalenol, aflatoxin B1/B2, and zearalenone). The signal from sample chips was compared to the signal from the control (calibrators) to obtain a semi-quantitative result.

Efficiency and accuracy of the extraction methods

The extraction method was confirmed for accuracy and efficiency by spiking 3 khadi samples from 3 different sampling locations with known concentrations of AFB1 (0.1 ppb) and FB1 (1 ppb). The Myco-10 extraction method was also used as instructed by the manufacturer (Randox Diagnostics, UK) and compared with the above extraction method (Section 2.4.1). The Myco-10 extraction method involved taking 5 mL of the khadi sample and adding 25 mL of acetonitrile: methanol: water (50:40:10). The mixture was then vortexed for 60 seconds and hand rolled for 10 minutes. After rolling, the mixture was centrifuged for 10 minutes at 3000 rpm to complete the extraction. The mycotoxin quantification (Sections 2.4.3 and 2.4.4) and spikings were done as stated above.

Mycotoxins detection from khadi using the UPLC-ESI-MS/MS

Sample preparation

To prepare the samples for mycotoxins analysis using the UPLC, all the samples were diluted in duplicates with 75% water: 25% methanol solvent into respective 2 mL microcentrifuge tubes and centrifuged for 5 minutes at 13 000 rpm. The supernatant was transferred to a 2 mL microcentrifuge tube and the pellet discarded. One milliliter of diluted sample was then aliquoted into an analysis vial.

Standard solutions

The following mycotoxin standards and metabolites were purchased from Biopure (Tulln, Austria): fumonisin B1, fumonisin B2, fumonisin B3, ochratoxin A, aflatoxin B1, nivalenol, diacetoxyscirpenol, and zearalenone. Acetonitrile (ACN), water and methanol (MeOH) were purchased from Biosolve (Valkenswaard, Netherlands) and formic acid (FA) from Merck (Amsterdam, Netherlands). The working solutions were prepared as follows: fumonisins solution containing FB1, B2, and B3 (each at 5 μg/mL) was diluted in acetonitrile/water (50/50 v/v) and stored 4°C; aflatoxin B1 (at 1 μg/mL), ochratoxin A solution (1 μg/mL), and the other mycotoxins solution containing nivalenol, deoxynivalenol, zearalenone (each at 10 μg/mL) were diluted in acetonitrile and were stored −20°C.

Mycotoxin quantification

A Waters Acquity ultra performance liquid chromatography (UPLC) was coupled to a Xevo triple quadrupole tandem mass spectrometer (MS/MS) (Waters, Milford, MA, USA) and used for high-resolution UPLC-MS/MS analysis. Multiple mycotoxins were separated in multiple reaction monitoring (MRM) using electrospray ionization in positive mode. The operating parameters used were as follow: capillary voltage, 3.5 V; cone voltage range, 15 to 50 V; collision energy range, 10 to 40 eV; source temperature, 140°C; desolvation temperature, 400°C; desolvation gas, 800 L/h and cone gas, 50 L/h. Separation was achieved with an Aquity BEH C18 (2.1 ×100 mm, 1.7 μ particle size) column at 50°C and a flow rate of 0.35 mL/min. An injection volume of 2 µL was used and the mobile phase consisted of water acidified with 0.1% formic acid (A) and acetonitrile acidified with 0.1% formic acid (B). The following gradient was used: Initial condition was 98% A, held for 0 to 0.5 minutes. followed by a gradual gradient change of solvent A to 60% from 0.5 to 7 minutes. This was followed by a further gradient change of solvent A to 30% from 7 to 10 minutes. A quick gradient change followed for solvent A to 5% from 10 to 11 minutes. Another gradient change of solvent A followed to 0% from 11 to 12 minutes after which the gradient quickly changed to initial condition of 98% solvent A from 12 to 12.1 minutes. Initial condition of 98% solvent A was kept for 12.1 to 14 minutes.

Matrix-matched linear calibration, LOD, and LOQ

To assess the linearity and detection sensitivity, blank matrices previously tested with undetectable target analytes were used. The blank matrices were extracted and evaporated the same manner as the samples. The resultant residues were dissolved using a series of mycotoxin standard solutions prepared in 20% methanol:80% water. The matrix-matched standard solutions were filtered through the 0.22-µm membrane before being analyzed with a UPLC-ESI-MS/MS method. The obtained peak area of each analyte was correlated with their concentration to examine the linearity. The LOD (S/N = 3) and LOQ (S/N = 10) of the devised method were calculated using the S/N from the lowest concentration of matrix-matched calibration standard solution.

Results and Discussion

Khadi contaminated with mycotoxin-producing fungi

To determine if there was filamentous fungal contamination, khadi samples were grown on filamentous fungi isolation media. The findings revealed that a majority of the samples were contaminated with a variety of filamentous fungal species of phylogenetically diverse background (Supplemental Figure S1). Aspergillus ochraceus, Cladosporium sphaerospermum, Cladosporium cladosporioides, Phoma eupyrena, and Setosphaeria turcica, to name a few, were among the 14 filamentous fungal isolates found (Table 1 and Supplemental Figure S1). The predominant isolates were C. sphaerospermum (Serowe and Maun), A. ochraceus (Palapye and Tonota), and Flavodon ambrosius (Mmashoro) (Table 1), accounting for 21.4% of the total isolates.

Table 1.

Identity of filamentous fungi isolates from the khadi samples based on D1/D2 domains of 26S rDNA.

Sample collection location	Isolate name	Accession number	Identity (%)
Serowe 1	Cladosporium sphaerospermum	KY873376.1	99
Serowe 3	Periconia thailandica	KY753888.1	99
Palapye 2	Cladosporium cladosporioides	KM246047.1	99
Palapye 3	Aspergillus ochraceus	KX958037.1	99
Palapye 3	Phoma eupyrena	JF766684.1	99
Maun 1	Setosphaeria turcica	LT715640.1	99
Maun 1	Cladosporium sphaerospermum (2)	KY873376.1	99
Maun 2	Chaetomium longiciliata	KP336823.1	99
Mmashoro 2	Flavodon ambrosius (3)	NG_060421.1	99
Tonota 3	Aspergillus ochraceus (2)	KX958037.1	99

The number in the bracket shows strains with the same accession number from the same location.

The introduction of filamentous fungi to khadi might be through the fruits or as contamination from the brewers or brewing equipment. Khadi that has visible filamentous fungal growth on it may have a shorter shelf life or become unfit for human consumption. However, the presence of filamentous fungi hardly never indicates the presence of mycotoxins. The presence of these filamentous fungal isolates on the khadi samples may indicate that the food item possesses the requisite ideal conditions, such as high water activity, the presence of sugars, and the presence of organic acids capable of decreasing pH 8. Notably, the presence of filamentous fungal isolates in beverages such as Aspergillus ochraceus can signal the potential existence of mycotoxins. The presence of Aspergillus ochraceus therefore, is of extremely concerning because it has been shown to produce aflatoxins and ochratoxins which are harmful to mammalian cells and can cause cancer, nephrotoxicity, hepatotoxicity, and teratogenic effects.^9,24 Fusariotoxins/trichothecenes and aflatoxins, which are produced by fungus belonging to the Fusarium and Aspergillus genera, respectively, have also drawn particular attention because they are the most dangerous for humans and animals and have significant economic implications. Fusariotoxins have a terpenoid biosynthesis route as one of their characteristics (eg, trichothecenes like toxin T2 and deoxynivalenol, the main precursors of which are mevalonic acid and farnesyl pyrophosphate). Other isolates of interest that were identified were C. sphaerospermum and C. cladosporioides, which are filamentous fungi previously isolated from flowers of edible fruit plants as well as being isolated from honey, dairy products and grapes prior to harvesting.^8,25-27 The Cladosporium species have been described in the past as mycotoxin-producing and possibly pathogenic fungi that are frequently found in outdoor settings.²⁸ The subsequent mycotoxin studies were necessary because the filamentous fungi were deemed to be a potential signal of mycotoxin presence in the khadi samples.

Mycotoxins detection from khadi using the Randox Evidence Investigator

Randox Evidence Investigator Calibration

The presence of fungal isolates (Table 1 and Supplemental Figure S1) in the khadi samples prompted the examination of the beverage for mycotoxins. Quantification of mycotoxin was done with the Randox Evidence Investigator. To optimize the sensitivity of the method and asses its accuracy, the Randox Evidence Investigator was calibrated using the supplied calibrators of a known concentration. The acquired data was then used to provide information on the limits of detection (LODs) and limits of quantification (LOQs). LOD and LOQ are terms used to describe the smallest concentration of an analyte that can be reliably measured by an analytical procedure with an acceptable degree of accuracy and precision. The LOD and LOQ were calculated based on 3- and 10-fold signal to noise (S/N) ratios, respectively (Table 2). The LOD ranged from 0.13 to 101.33 ppb while the LOQ ranged from 0.43 to 337.78 ppb with DON having the highest LOD and LOQ. The linearity was established by the calibration curve based on the expected values and actual values from calibrators. The coefficient correlations R² ranged from 0.97 to 0.99 thus showing good linearity (Table 2) hence validating results from the Evidence Investigator (Supplemental Figure S2).

Table 2.

The calibration of the Randox Evidence Investigator using the control toxins provided with the kit to determine the LOD and LOQ of each toxin.

Compound	Linear equation	R2	SD	LOD	LOQ
PAX	Y = 0.9634x + 0.0252	0.99	6.48	20.18	69.39
FUM	Y = 1.0061x − 0.1004	0.97	0.68	2.03	6.77
OTA	Y = 1.0088x − 0.0131	0.98	0.27	0.81	2.70
AFG1	Y = 0.9883x − 0.0103	0.98	0.33	1.01	3.38
EA	Y = 1.0056x − 0.1973	0.97	4.49	13.41	44.70
DAS	Y = 1.0816x − 0.4864	0.97	7.33	20.33	67.76
DON	Y = 1.0907x − 2.4123	0.98	36.84	101.34	337.78
T2	Y = 1.0935x − 0.0721	0.98	1.47	4.04	13.47
AFB1	Y = 1.0769x − 0.0023	0.98	0.05	0.13	0.43
ZEA	Y = 1.0766x − 0.0459	0.99	0.72	2.02	6.73

Abbreviations: AFB1/B2, aflatoxin B1/B2; AFG1/G2, aflatoxin G1/G2; DAS, diacetoxyscirpenol; DON, deoxynivalenol; ERG, ergot alkaloids; FUM, fumonisins; OTA, ochratoxin A; PAX, paxilline; ZEA = zearalenone.

Three samples from 3 different sampling areas were then spiked with a known concentration of AFB1, and FB1 to test the efficiency of the acetonitrile method (Section 2.4.1) and the Myco-10 method (Section 2.4.5) according to the manufacturer’s instruction. The recovery of the samples ranged from 45% to 107% using the acetonitrile method where the best recovery was observed on FB1 and AFB1. Table 2 shows that the recovery challenge is not with the biochip because the quantification of the control standards read values close to the true concentrations as stated in the manual provided by the manufacturer. This is further confirmed by the R² of the sample calibrators (Supplemental Figure S2). The results in Table 3 also show that the Myco-10 extraction method was not suitable for use in this experiment. The Myco-10 extraction method was based on the use of chemicals namely acetonitrile, methanol and water as the extraction solvents and no organic salts. The acetonitrile method was more effective as it counteracted the presence of water in khadi by using organic salts to bind to the water molecules.²⁹ This step was missing in the Myco-10 extraction method resulting in the negative values in Table 3. Due to the lack of meaningful results from the Myco-10 method, it was deemed unsuitable for use in this study.

Table 3.

The product recovery of the spiked samples using the above extraction method.

Sample	FB1 (1 ppm) (%)	AFB1 (0.1 ppm) (%)
Maun 1 (ACN method)	91.40	53.10
Mmashoro 2 (ACN method)	103.70	45.80
Serowe 1 (ACN method)	107.30	61.90
Serowe 1 (Myco-10 method)	−98.2	−75

Mycotoxins from khadi

After the optimization of the acetonitrile method, the khadi samples were analyzed for presence of mycotoxins. All the samples contained fumonisin, ochratoxin A, diacetoxyscirpenol, and zearalenone (Table 4). The least detected toxins from the brews were paxilline (detected from Mmashoro 1, Mmashoro 2, Palapye 1, Palapye 3, Serowe 1, and Serowe 3), deoxynivalenol (detected from Maun 1, Maun 2, Serowe 1, Tonota 1, Tonota 2, and Tonota 3) and aflatoxin B1/B2 (detected from Mmashoro 3, Serowe 1, and Tonota 2). The Serowe 1 sample contained all 10 mycotoxins. The quantification of mycotoxin shows that DAS was the most abundant (ranged from 3.47 to 33.79 ppb) and aflatoxin B1/B2 was the least abundant (ranged from 0 to 0.049 ppb). Other toxins included PAX, which ranged from 0 to 0.16 ppb, FUM which ranged from 0.14 to 2.27 ppb, OTA which ranged from 0.33 to 3.84 ppb, AFG1/G2 which ranged from 0 to 0.12 ppb, ERG which ranged from 0 to 4.93 ppb, DON which ranged from 0 to 21.86 ppb, and ZEA which ranged from 0.004 to 0.49 ppb.

Table 4.

The quantification of mycotoxins from khadi samples.

Sample	PAX	FUM	OTA	AFG1/G2	ERG	DAS	DON	T2	AFB1/B2	ZEA
Control	1.51	0.27	0.057	0.066	1.10	1.30	8.89	0.25	0.0081	0.14
Maun 1	0	1.30	2.53	0.092	4.57	27.06	5.57	0.24	0	0.16
Maun 2	0	1.48	2.26	0.1116	4.86	25.13	10.02	0.30	0	0.094
Maun 3	0	0.61	0.87	0.067	2.43	14.53	0	0.06	0	0.091
Mmashoro 1	0.23	1.14	1.97	0.12	4.16	21.12	0	0.20	0	0.11
Mmashoro 2	0.51	1.46	2.26	0.088	4.45	24.74	0	0.15	0	0.12
Mmashoro 3	0	1.90	2.79	0.096	4.10	29.23	0	0.079	0.0006	0.13
Palapye 1	0.27	1.55	2.22	0.060	3.60	25.21	0	0.13	0	0.19
Palapye 2	0	1.81	2.33	0.079	3.26	27.30	0	0.13	0	0.12
Palapye 3	0.33	2.27	3.52	0.072	3.78	29.91	0	0.093	0	0.068
Serowe 1	1.32	1.14	2.13	0.12	3.22	26.06	21.86	0.33	0.0049	0.49
Serowe 2	0	1.45	1.89	0.058	3.80	23.11	0	0.091	0	0.13
Serowe 3	0.16	1.15	1.76	0.074	3.35	22.19	0	0.12	0	0.11
Tonota 1	0	1.72	3.84	0.11	4.61	33.79	8.46	0.24	0	0.18
Tonota 2	0	1.38	2.82	0.12	4.93	29.54	10.95	0.26	0.0013	0.19
Tonota 3	0	1.26	2.58	0.078	3.84	26.51	7.09	0.27	0	0.13
Letlhakane 1	0	1.02	1.60	0.07	3.56	19.63	0	0.027	0	0.042
Letlhakane 2	0	0.18	0.44	0	0	4.34	0	0	0	0.009
Letlhakane 3	0	0.14	0.33	0	0	3.47	0	0	0	0.004

Abbreviations: AFB1/B2, aflatoxin B1/B2; AFG1/G2, aflatoxin G1/G2; DAS, diacetoxyscirpenol; DON, deoxynivalenol; ERG, ergot alkaloids; FUM, fumonisins; OTA, ochratoxin A; PAX, paxilline; ZEA = zearalenone.

Mycotoxins such as zearalenone have been detected from traditional opaque beers of Botswana¹³ while others such as ochratoxin A, deoxynivalenol, fumonisin B1, and fumonisin B2 have been detected in various European beers.³⁰ In another investigation on the South African maize-based alcoholic beverage umqombothi, the authors detected mycotoxins such as T2, DON, fumonisin, zearalenone, and aflatoxins with evidence pointing at the raw material or substrate as the source of the mycotoxins.^15,16 Both dried and fresh fruits tend to have mycotoxins.^31,32 Mycotoxins accumulate during the harvesting, drying and storage due to conditions that favor mycotoxin accumulation.^33,34 The A. ochraceus produces ochratoxin A, which was detected in all the khadi samples in this work. The detection of other mycotoxins such as aflatoxins is characteristic of other fungi such as A. flavus, A. parasiticus, A. bombycis, A. ochraceus, A. nomius, and A. pseudotamar.^1,9,24,35-37 In this work, only A. ochraceus was isolated from khadi and thus A. ochraceus could be the only possible source of aflatoxins detected in khadi.

Different countries have different mycotoxins limitations.³⁸ The Food and Drug Administration (FDA) allows aflatoxins to be at a maximum level of 4 to 20 ppb while European Commission (EC) has a rigorous restriction of 0.5 to 10 ppb per food item⁵ yet all the khadi samples were below the permissible range. The FDA also has limits on fumonisins (0.2-4 ppb), aflatoxins (20 ppb), ZEA (20-200 ppb), and T2 (80-100 ppb).⁵ The khadi samples were all below these limits, implying that they may not cause any major health risks when consumed (Table 4). These findings imply that there is no risk of mycotoxin poisoning after consumption of khadi.

Confirmation of presence of mycotoxins in khadi using the UPLC-ESI-MS/MS

To confirm the identity of the mycotoxins obtained using the rapid method of mycotoxin quantification (Myco-10 kit), the samples were analyzed using a UPLC-ESI-MS/MS system. The MS spectra of the mycotoxins in selected khadi samples in comparison to the standards suggest that there were no mycotoxins detected (Table 6 and Supplemental Figures S3-S10). As shown in Table 5 and Supplemental Figures S3-S10, there was an excellent linearity for all 8 calibration curves of the mycotoxins, with a coefficient of greater than 0.98. Our results suggest that the sensitivity of the detection method was very high as the LOD was as low as 1 to 10 ppb (Table 5). The results in summation showed that all the samples were free of any concerning levels of mycotoxins.

Table 6.

Mycotoxin detection through UPLC-ESI-MS/MS from the khadi samples with the FDA and EC limits.

Mycotoxin	Sample	Toxin quantification	FDA limit (ppb)	EC limit (ppb)
AFB1	Palapye 1, Palapye 2, Letlhakane 2, Mmashoro 3, Maun 3, and Tonota 3	ND	<20	2
NIV	Palapye 1, Palapye 2, Letlhakane 2, Mmashoro 3, Maun 3, and Tonota 3	ND
DON	Palapye 1, Palapye 2, Letlhakane 2, Mmashoro 3, Maun 3, and Tonota 3	ND	1000	1 ppb BW/day
ZEA	Palapye 1, Palapye 2, Letlhakane 2, Mmashoro 3, Maun 3, and Tonota 3	ND	0.25 ppb BW/day	20-200
OTA	Palapye 1, Palapye 2, Letlhakane 2, Mmashoro 3, Maun 3, and Tonota 3	ND	Not set	2000
FB1	Palapye 1, Palapye 2, Letlhakane 2, Mmashoro 3, Maun 3, and Tonota 3	ND	2000-4000	200-4000
FB2	Palapye 1, Palapye 2, Letlhakane 2, Mmashoro 3, Maun 3, and Tonota 3	ND	2000-4000	200-4000
FB3	Palapye 1, Palapye 2, Letlhakane 2, Mmashoro 3, Maun 3, and Tonota 3	ND	2000-4000	200-4000

AFB1, aflatoxin B1; BW/day, body weight per day; DON, deoxynivalenol; FB1, fumonisin B1; FB2, fumonisin B2; FB3, fumonisin B3; NIV, nivalenol; OTA = ochratoxin A; ZEA = zearalenone.

Table 5.

Limit of detection for the analytes in this method.

Analyte	Linear equation	R2	LOD (ppb)
Aflatoxin B1	Y = 147489x – 139.572	0.9938	5
DON	Y = 20322.9x – 12.406	0.9938	10
Zearalenone	Y = 1629.66x – 22.1094	0.9985	10
Nivalenol	Y = 214.41x – 4.0601	0.9970	10
Ochratoxin A	Y = 106682x – 84.2093	0.9897	10
Fumonisin B1	Y = 216963x – 183.785	0.9981	1
Fumonisin B2	Y = 448027x – 385.787	0.9980	1
Fumonisin B3	Y = 410526x – 256.431	0.9943	1

The mycotoxins levels in the khadi samples are below the FDA and EC regulation standards. The results from the Myco-10 biochip detected mycotoxins from all the khadi samples (Table 6) but the UPLC-ESI-MS/MS showed that the mycotoxins in the samples were below limit of detection (Table 6). These findings indicate that khadi can be consumed with little danger of mycotoxin poisoning. If mycotoxins are present at quantities exceeding set regulatory limits, they can have major health consequences.^1,2 Although the khadi samples were contaminated with filamentous fungi (Table 1), the absence of mycotoxins is not surprising given that literature implies that conditions conducive to fungal growth and development are not conducive to mycotoxin production.⁸ Similarly, the absence of filamentous fungi does not always imply the absence of mycotoxins, as mycotoxins might persist after the disappearance of fungi.³⁹

Conclusion

Mycotoxin contamination research is crucial step toward commercialization of innovative food products. Khadi, like most spontaneously fermented beverages, was discovered to harbor fungal isolates. Although some of isolates are known to produce mycotoxin, our findings suggest that there are minimal risks of mycotoxins toxicity associated with khadi. Using 2 distinct procedures, all of the mycotoxins were either below regulatory levels or undetectable. This results from this study thus show that khadi mycotoxin levels are within safe limits for human consumption.