Fate of aflatoxins M1 and B1 within the period of production and storage of Tarkhineh: A traditional Persian fermented food

Abstract The objective of the study was to assess the amount of aflatoxin M1 (AFM1) and aflatoxin B1 (AFB1) during fermentation, drying, and storage of Tarkhineh—a traditional Persian fermented food—over four months. Tarkhineh samples were produced based on a traditional method. Various concentrations of AFB1 (2.5, 5, 7.5, and 10 µg/kg) and AFM1, stood at 0.25, 0.5, 0.75, and 1 µg/kg, were added to Iranian yogurt drink, called doogh, samples. Tarkhineh samples were evaluated for AFB1 and AFM1 on days 0, 2, 6, and 8 and also after drying and four months of storage. In cases of repeatability, recovery, and reproducibility, the high‐performance liquid chromatography through fluorescence detector (HPLC‐FD) method was successfully done to demonstrate aflatoxins (AFs) in Tarkhineh samples. The fermentation process had a considerable consequence on the reduction in AFM1 and AFB1 as compared to the control group, evidenced by 65.10%–81.20% and 55.80%–74.10%, respectively, after eight days of fermentation (p < .05). The highest reduction in AFB1 existed in samples containing 2.5 µg/kg toxin, followed by 5, 7.5, and 10 µg/kg, respectively. A similar trend was found for AFM1, as the highest concentration was found in samples containing 0.25 µg/kg toxin, followed by 0.5, 0.75, and 1 µg/kg, respectively.

. Among eighteen aflatoxin compounds, aflatoxin B 1 (AFB 1 ) is the highest frequent mycotoxin that existed in foods for both humans and animals (Babakhanian et al., 2015;Iqbal et al., 2013). Due to the highly cytotoxic, mutagenic, immunosuppressive, teratogenic, and carcinogenic effects of AFB 1 , being recognized as the first group of human carcinogen, defined in International Agency for Research on Cancer (IARC). To protect consumers' health, the maximum authorized amount of AFB 1 in foodstuffs, as well as cereals and breakfast cereals, was considered to be 5 and 2 µg/kg among the European Commission, respectively (European Commission, 2006).
The cytochrome P450 enzyme system metabolizes AFB 1 into aflatoxin M 1 (AFM 1 ) in the lactating animals' liver, feeding on AFB 1contaminated diet (Fallah et al., 2011). According to IARC, AFM 1 has been regarded as the second group of human carcinogens and is observed as a vexingly complicated issue, considering food hygiene (IARC, 2002). AFM 1 is frequently found in pasteurized, raw, and ultrahigh-temperature (UHT) milk and dairy outputs, including butter, yogurt, ice cream, cream, and cheese in many countries around the world (Fallah, 2010;Fallah et al., 2009;Hassan & Kassaify, 2014;Iqbal et al., 2017;Škrbić et al., 2015;Zheng et al., 2013;Zinedine et al., 2007). Because of the health-related problems of AFM 1 , over sixty countries have set up the highest authorized amount for AFM 1 distributed among dairy productions based on their economic conditions and development (Kos et al., 2014;Nabizadeh et al., 2018). The countries of the European Commission, Turkey, Argentina, Honduras, and also Iran have set up the level permitted with 50 ng/kg in infant, raw, and pasteurized milk (Codex Alimentarius Commission, 2001). In the past decades, several approaches such as fermentation, heating, and irradiation and also the addition of chlorinating, oxidizing, hydrolytic compounds, and medicinal plants have been used to decontaminate food products from AFM 1 (Sarlak et al., 2017).
Tarkhineh, Tarkhowana or Doowina in Kurdish language, is one of the locally made foods in the western part of Iran (Kermanshah, Ilam, Kurdistan, and Lorestan), produced throughout heating process and fermenting wheat meal, medicinal plants (mint, pennyroyal, and ziziphora), and doogh fermented for approximately a week and then proceeded by sun-drying for 3-4 days (Bahrami et al., 2016). The doogh is a traditional Persian dairy drink that prepared from churning yogurt in a vessel comprising the potable water, salt, and spices, and definitely, it is one of the most popular beverages in Iran and some other Middle East countries (Bahrami et al., 2016;Sarlak et al., 2017). In a lately represented study done by Bahrami et al. (2016), the medium level concentration of AFM 1 in samples of Tarkhineh from Kermanshah, Kurdistan, Ilam, and Lorestan provinces was recorded to be 11.0 ± 1.2 ng/kg. The contamination of doogh, wheat, and spices such as mint, pennyroyal, and ziziphora is the potential reason for the existence of AFM 1 and AFB 1 in Tarkhineh. In another study, Sarlak et al. (2017) indicated that incorporating 9 log CFU/ml probiotic Lactobacillus acidophilus into the fermented doogh significantly reduced its AFM content (0.5 µg/kg) compared with the control group. A similar AFM 1 reduction in fermented milk during production and storage was reported (Arab et al., 2012). Based on this study, no permissible limit has been established for AFM 1 in Tarkhineh in Iran, given that Tarkhineh is commonly consumed in winter and autumn in different cities of Iran and is produced by households in rural areas under unsuitable hygienic conditions. According to our knowledge, there is no information on the fate of AFM 1 and AFB 1 within the period of production and storage of Tarkhineh. Therefore, it is highly important to analyze the fate of AFM 1 and AFB 1 within producing and storing Tarkhineh. Thus, the study presented was conducted with the aim of determination of the occurrence of AFM 1 and AFB 1 during fermentation, drying, and storage of Tarkhineh over four months.

| Materials
AFB 1 and AFM 1 standards were bought from Sigma-Aldrich (UK) with purity certified greater than 99%. AFLA-Test immunoaffinity column for extract clean-up was purchased from LC Tech GmbH (Dorfen, Germany) and utilized according to the guidelines of the manufacturer. All chemicals used in this investigation were of HPLC grade and obtained from Merck (Darmstadt, Germany).
The deionized water was applied in all procedures. The wheat and raw milk samples were grasped from the regional markets in Kermanshah, located in the western part of Iran. Lactobacillus delbrueckii subsp. bulgaricus and Streptococcus salivarius subsp. thermophilus were ordered from the microbial collection of Iranian Research Corporation for Science and Technology (Tehran, Iran).

| Preparation of aflatoxins M 1 and B 1
The stock solutions of AFB 1 and AFM 1 were prepared via dissolving 0.5 mg in 10 ml chloroform and 5 µg in 5 ml acetonitrile, respectively, and consequently, they were stored in −20°C in an amber flask, preventing photodegradation. The appropriate concentrations AFB 1 and AFM 1 , considered as the aqueous working solutions, were obtained from the stock solutions to an appropriate volume of methanol: water 40:10 v/v, ranging from 0.02 to 20 µg/kg for AFB 1 and from 0.125 to 2 µg/kg for AFM 1 , which were then kept at refrigerator temperature (4 ± 1°C).

| Analysis of HPLC
High-performance liquid chromatography (HPLC) utilized in this paper is performed via a KNAVER HPLC accompanied with a fluorescence detector (FD detector, model RF-20A) and LCTech postcolumn photochemical derivatization of UV system. The HPLC system consisted of a C18 analytical column with 250 4.6 mm I.D., 5 mm.
All HPLC analyses were done under isocratic conditions by the use of a mobile phase composed of a combination of ultrapure water: acetonitrile (90:10 v/v, AFB 1 ) with a flow rate of 1.5 ml/ min and acetonitrile: methanol: water (20:20: 60 v/v, AFM 1 ) with a flow rate of 1.2 ml/min. The wavelengths of the FD were set at 329 and 460 nm for the excitation and emission of AFB 1 and at 365 and 455 nm for the excitation and emission of AFM 1 , respectively. The column temperature was held at 40°C for AFB 1 and 30°C for AFM 1 . The standard and sample injection solution volumes were 20 μl.

| Preparation of Tarkhineh samples
Tarkhineh samples were manufactured based on a traditional method as published in a previous study (Mashak et al., 2014).
An amount of 1000 g wheat meal was soaked in 4000 ml sour doogh (a beverage produced by beating unflavored yogurt until it is smooth) and then fermented for 8 days. Subsequently, 20 g dried Mentha longifolia powder and 20 g salt were incorporated into the dough-like mixture. Eventually, it was exposed to sunlight in tiny parts to dry.

| Proximate composition of Tarkhineh samples
Moisture, protein, fat, and ash contents of Tarkhineh samples were measured by a standard method (AOAC, 1995). Compositional values are reported on a percent (%) basis.

| Spiking of Tarkhineh samples
The initial counts of AFB 1 and AFM 1 in the raw milk and wheat were clarified by using HPLC-FD, as described in Section 2.3. AFB 1 and AFM 1 were not detected in the raw milk and wheat samples. Various concentrations of AFB 1 , which include 2.5, 5, 7.5, and 10 µg/kg, and AFM 1 , which include 0.25, 0.5, 0.75, and 1 µg/kg, were added to the doogh samples. Tarkhineh samples were evaluated for the existence of AFB 1 and AFM 1 on days 0, 2, 6, and 8 and also after drying and four months of storage.

| Solid-phase extraction
In order to perform the clean-up procedure, 25 g of each sample was combined with 0.5 g NaCl. After adding 50 ml ultrapure water: methanol (40:10 v/v), 14 ml of the resultant solution was combined with 86 ml phosphate-buffered saline. Then, extracts of the sample were carefully moved through C18 SPE column. This trend was proceeded with ultrapure water: methanol (40:10 v/v).
Following this, being washed with 10 ml water, the column was dried by N 2 gas. The resulting residue was dissolved in 2 ml methanol, transmitted to an HPLC vial, and analyzed by HPLC-FD (Sarlak et al., 2017).

| Validation of HPLC method
Considering the validation procedure for aflatoxin residues distributed among animal-related products, described by the European Communities (EC), the method validation was implemented (European Communities, 2002). Linearity, repeatability, limit of detection (LOD), and limit of quantification (LOQ) were examined in this study. Accordingly, a signal-to-noise ratio (S/N) of 3 and 10, LOD, and LOQ were ascertained, where the samples were spiked with different concentrations of AFs (Bahrami et al., 2016). The linearity was determined by injecting different concentrations of AFM 1 and AFB 1 , respectively, at 0.125-2 and 0.02-20 µg/kg. The percent of recovery was also calculated to confirm the accuracy of the method. Withinday (run-to-run) precision of the HPLC-FD method, considered as RSD%, was calculated by extracting and analyzing AFs in one sample three times in the same condition.

| Survival of Lactobacillus bulgaricus and Streptococcus thermophilus in Tarkhineh samples
Enumerations of L. bulgaricus and S. thermophilus in the stored Tarkhineh samples were conducted using MRS agar. Then, the plates were incubated within the period of 48 hr, at 37 ± 1°C for under microaerophilic conditions (Jay et al., 2005).

| pH determination of Tarkhineh samples
For pH measurement, 5 g of the samples was homogenized with double-distilled water. 5 min later, pH was examined via a pH meter (Metrohm digital pH meter model 632; Switzerland) at room temperature. For evaluation of each toxin, the two-way repeated-measures analysis of variance including a within-subject factor (six levels of storage time) and a between-subject factor (four different treatments) was applied.

| Exact composition of Tarkhineh samples
Fat, moisture, protein, and ash contents of Tarkhineh samples were measured to be 1.23%, 3.12%, 14.57%, and 7.45%, respectively. The obtained outputs are in agreement with what was done in the previous studies (Mashak et al., 2014;Tabatabaei-Yazdi et al., 2013).

| Validation study
The HPLC chromatograms of AFM 1 and AFB 1 standards, and their samples as well, are depicted in Figures 1 and 2, respectively.
According to the performance of the HPLC-FD method, the current method in this study was evaluated based on the recovery percentage, LOD, LOQ, r 2 , RSD, and linearity. With regard to our results, the curve of calibration for AFM 1 and AFB 1 was linear at disparate concentrations between 0.02 and 20 µg/kg for AFB 1 and ranging from 0.125 to 2 µg/kg for AFM 1 . The linear regression equations of AFB 1 and AFM 1 were y = 39.359 x + 14.592 and y = 47.805 x + 4.6708, respectively. These revealed a linear connection between the peak area and the corresponding concentrations of AFB 1 and AFM 1 . There was a significant correlation between results and the concentration of AFB 1 injected, in which the calculated coefficient of determination (r 2 ) was 0.996. Moreover, LOD, 0.005 µg/kg, and LOQ, 0.015 µg/kg, were found.
The r 2 , LOD, and LOQ for the injected AFM 1 were 0.9937, 0.02, and 0.045 µg/kg, respectively. The comparison of the obtained results with the performance standard introduced in the Commission Regulation (EC) No 401/2006 exhibited the method that met the performance criteria in terms of repeatability, recovery, and reproducibility for AFs in Tarkhineh samples.

| Fate of aflatoxins M 1 and B 1 during Tarkhineh fermentation and storage
The results of the fate of AFB 1 and AFM 1 during Tarkhineh fermentation are presented in Tables 1 and 2, respectively. Based on our findings, the fermentation process had an immense consequence on the reduction in AFM 1 and AFB 1 compared with the control group, evidenced by 65.10%-81.20% and 55.80%-74.10%, respectively, after eight days of fermentation (p < .05). Moreover, the reduction percentage of toxins significantly varied with toxin concentrations in Tarkhineh samples (p < .05). In detail, the highest reduction in AFB 1 existed in samples containing 2.5 µg/kg toxin, followed by 5, 7.5, and 10 µg/kg, respectively. A similar trend was also found for AFM 1 , as the highest concentration was examined in samples containing 0.25 µg/kg, which was followed by 0.5, 0.75, and 1 µg/kg, respectively. This is probably because of the practice of S. thermophilus and L. bulgaricus, which can be mainly related to weak noncovalent bound interactions, like pertaining to hydrophobic pockets on the bacterial surface (Campagnollo et al., 2016;Iqbal et al., 2015). El-Khoury et al. (2011) evaluated the capability of some strains of lactic acid bacteria, especially S. thermophilus and L. bulgaricus, to remove AFM 1 within the period of yogurt production. Their findings indicated that S. thermophiles and L. bulgaricus bound 70% and 87.6% of AFM 1 , respectively. Further, reported 77%-99% AFM 1 was eliminated by Lactobacillus spp. in an extensive and medium amount (Turbic et al., 2002). Our findings are also in agreement with those reported for Feta cheese (Motawee & McMahon, 2009), acidophilus milk , yogurt (Govaris et al., 2002; Montaseri F I G U R E 1 HPLC chromatograms of AFM 1 standard (2 µg/kg). Inset is the chromatogram of the sample F I G U R E 2 HPLC chromatograms of AFB1 standard (20 µg/kg). Inset is the chromatogram of the sample et al., Sevim et al., 2019), doogh (Sarlak et al., 2017), and buttermilk and kefir (Wiseman & Marth, 1983). The reduction in AFB 1 in nondairy products such as pistachio nuts (Rastegar et al., 2017), peanut and olive oils (Fan et al., 2013), and corn kernels (Hojnik et al., 2021) was also reported.
The viability of S. thermophilus and L. bulgaricus during the production of Tarkhineh sample was also evaluated. According to our findings, the total viable counts of L. bulgaricus and S. thermophilus were 7.88 ± 0.04-9 ± 0.02 log CFU/ml, 7.45 ± 0.05 log CFU/g, and 6.45 ± 0.07 log CFU/g during fermentation, drying, and storage of Tarkhineh samples, respectively (Table 3). The reduction in S. thermophilus and L. bulgaricus can be attributed to the acidic condition of the samples during fermentation (Tsakalidou & Papadimitriou, 2011). The survival of bacterial cells during the drying and storage of samples can be originated from the adhesion forces between food matrices and the strains and finally overcome the limitations, including the physicochemical and osmotic conditions (Jay et al., 2005). Our outcomes are in agreement with those reported for Feta cheese (Mohammadi et al., 2020), doogh (Sarlak et al., 2017), milk (Abdelmotilib et al., 2018;Serrano-Niño et al., 2013), and peanut grains (Silva et al., 2015).
Furthermore, contributors, such as low pH and the formation of organic acids and other fermentation by-products, result in the reduction in AFM 1 (Bahrami et al., 2016). The consequences of this study exhibited that the pH values of Tarkhineh samples were declined within fermentation and drying and storage steps (Table 4). In the current study, a definite link was observed between pH reduction and AFM 1 and AFB 1 decomposition. The previous studies also reported that AFM 1 and AFB 1 reduced with a decrease in pH (Hassanin, 1994;Sarimehmetoğlu & Küplülü, 2004), being complied with our findings. In stark contrast, though, Blanco et al. (1993) and Wiseman and Marth (1983) demonstrated that aflatoxins B 1 , B 2 , G 1 , G 2 , M 1 , and M 2 did not alter during yogurt, buttermilk, and kefir fermentation. Van Egmond et al. (1977) and Munksgaard et al. (1987) found that the concentration of AFM 1 was increased after the fermentation process, which is in contrast with our findings (Munksgaard et al., 1987;Van Egmond et al., 1977). As indicated in the previous studies, the production of organic acids and other fermentation by-products can be considered a noteworthy method for detoxification of AFM 1 and AFB 1 in food products compared with other chemical approaches.

TA B L E 1 Fate of aflatoxin B 1 during Tarkhineh production and storage
As presented in Tables 1 and 2 was approximately 4-to 4.5-fold more than in the milk (Manetta et al., 2009). Moreover, the results of the presented study demonstrated that the concentrations of AFB 1 and AFM 1 were noticeably constant after four months of sample storage at room temperature.

| CON CLUS ION
With regard to this study's results, the fermentation of doogh during Tarkhineh production can significantly reduce the concentrations of AFM 1 and AFB 1 (p < .05). Our findings indicated that the most probable reason for toxin reduction was the low pH of doogh during fermentation and also the presence of starter culture microorganisms, including L. bulgaricus and S. thermophilus as well as native probiotics in the Tarkhineh to bind the mycotoxins. Accordingly, the fermentation process could remarkably reduce AFB 1 and AFM 1 concentrations during fermentation while the concentrations of AFB1 and AFM1 could be constant after four months of sample storage.
The obtained results were in an agreement with those reported for milk and dairy outputs.

ACK N OWLED G M ENT
It must be noted that efforts of the Department of Food Science

CO N FLI C T O F I NTE R E S T S
There is no conflict of interest found in this study.

DATA AVA I L A B I L I T Y S TAT E M E N T
The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.