Occurrence of Regulated Mycotoxins and Other Microbial Metabolites in Dried Cassava Products from Nigeria

Dried cassava products are perceived as one of the potential sources of mycotoxin ingestion in human foods. Processing either contributes to the reduction of toxins or further exposes products to contamination by microorganisms that release metabolic toxins into the products. Thus, the prevalence of microbial metabolites in 373 processed cassava products was investigated in Nigeria. With the use of liquid chromatography tandem-mass spectrometry (LC-MS/MS) for the constituent analysis, a few major mycotoxins (aflatoxin B1 and G1, fumonisin B1 and B2, and zearalenone) regulated in food crops by the Commission of the European Union were found at concentrations which are toxicologically acceptable in many other crops. Some bioactive compounds were detected at low concentrations in the cassava products. Therefore, the exposure of cassava consumers in Nigeria to regulated mycotoxins was estimated to be minimal. The results provide useful information regarding the probable safety of cassava products in Nigeria.


Introduction
The cassava root (Manihot esculenta Crantz) significantly contributes to food security, incomes, and employment opportunities in the rural areas of Sub-Saharan Africa [1], especially in Nigeria, the world's largest cassava producer [2]. Significant post-harvest deterioration of fresh cassava roots occurs because of the natural high moisture content, which accelerates microbial deterioration and undesirable biochemical changes in the products [3]. Processing is used to extend the shelf life, facilitate transport and, most importantly, detoxify the roots by removing the inherent cyanogens [4][5][6]. Hence, cassava root is processed in Nigeria into gari, tapioca, lafun, fufu, starch, and high-quality cassava flour (HQCF), among others, with all the products having different physical properties due to variations in processing methods [7][8][9]. However, these processing methods, as well as the environments and natural microflora, influence the types and concentrations of microbial metabolites in the final food products [10,11].
The various processing methods for cassava in Nigeria often result in a range of food and feed products. Cassava starch and high-quality cassava flour (HQCF) are dried, unfermented products that must be dried immediately to avoid fermentation [12,13]. Starch is produced by peeling the roots, washing, grating, pulverizing, wet-sieving, sedimentation, decanting, dewatering, drying, and milling. capita consumption of the cassava products, and the prevalence and concentrations of the different mycotoxins in the products. On the other hand, microbial specifications and permissible limits for food additives, pesticide residues, and heavy metal contaminants have all been stipulated [34]. Therefore, the objective of this study was to evaluate the prevalence of major mycotoxins and other microbial metabolites in various dried cassava products consumed in Nigeria, using a more versatile and precise mycotoxin quantitation methodology based on the proven principle of isotope dilution mass spectrometry, as previously described [23][24][25]30,31,35].
About 70% of cassava roots produced in Nigeria are processed into gari, making this the most popular cassava product in Nigeria [36]. The aflatoxin content of all types of gari samples was under the detectable limit; these results, therefore, suggest that gari is very safe from aflatoxin contamination. With averages of 1.2 µg/kg of aflatoxin B 1 (in fufu flour) and 2.9 µg/kg of aflatoxin G 1 (in HQCF), the aflatoxin levels of the dried cassava products sampled and tested were below the European Union values of 5 µg/kg tolerance level in foods [37]. The level of aflatoxin B 1 found in the HQCF of the present study was lower compared to the values (4-21 µg/kg) reported by Ediage et al. [30] for cassava flour from the Republic of Benin.
Neither aflatoxins B 2 , G 2 , M 1 , M 2 , P 1 , nor ochratoxin A was detected in any of the 373 dried cassava product samples. Fumonisin B 3 was detected in only one fufu flour sample (14.5 µg/kg). However, Ediage et al. [30] oberved that 8 µg/kg of aflatoxin B 2 was present in cassava flour from the Republic of Benin. Similarly, patulin and deoxynivalenol were absent in all the samples, and the range of zearalenone concentrations (0.9-90.4 µg/kg) obtained was lower than that reported by Sulyok et al. [31] for cassava samples from Rwanda (2830 µg/kg) and Tanzania (8490 µg/kg), and the values (12 µg/kg) reported by Ediage et al. [30] for cassava flour from the Republic of Benin. The results suggest that processed cassava products in Nigeria are safe with respect to the regulated mycotoxins, also considering that the regulated levels of zearalenone and fumonisins reported for maize and other cereals in African countries, such as Niger, Ghana, the United Republic of Tanzania, Uganda, and Benin, range between 50 µg/kg and 1000 µg/kg, and 1000 µg/kg to 3000 µg/kg, respectively [37]. Implicitly, a better understanding of the impact of processing practices adopted in Rwanda and Tanzania on the relatively higher levels of zearalenone and fumonisin in samples from the two countries may be helpful in efforts towards setting future regulatory levels for these mycotoxins in cassava products. Calculation of means was based on positive samples. R: apparent recovery; LOD: limit of detection; +: represents a positive analyte but that was detected at a concentration < LOD. Figures in parentheses are a number of samples in which an analyte was detected at > LOD. Table 2. Number of non-regulated microbial metabolites detected at or above limit of detection in at least 5% of the total number of processed cassava samples.  These 16 metabolites could be the most common in dried cassava products in Nigeria. Of these, the most common metabolites associated with Aspergillus spp. were kojic acid, asperphenamate, Nbenzoyl-phenylalanine, emodin, and asperglaucide. The predominant metabolites of Alternaria spp. were cyclo (L-Pro-L-Tyr), cyclo (L-Pro-L-Val), and alternariol methyl ether. Furthermore, tryptophol was the most common metabolite associated with Fusarium spp. and brevianamid F, fellutanine A, and rugulusovi , associated with Pennicillum spp., were the most commonly identified metabolites of that species (see Table 2). These results agree with those obtained in previous studies [11,[38][39][40] of various staple foods from some countries, including cassava products.
Kojic acid, asperphenanate, N-benzoyl-phenylalanine, emodin, and asperglaucide are all metabolites associated with Aspergillus spp. (see Table 3). The concentrations of kojic acid, a 5hydroxy-2-hydroxymethyl-4-pyranone, in the dried product samples ranged from 8.35 to 1754.80 µg/kg; lafun samples had the highest, and cassava starch the lowest concentration. The kojic acid content of lafun in this study was lower than the maximum values (650,000 µg/kg and 93,700 µg/kg) recorded in cassava samples from Tanzania and Rwanda, respectively [31]. However, the processing methods for these cassava products were not indicated. Kojic acid can also be produced from various carbohydrate sources in an aerobic condition by a variety of microorganisms [41]. The fermentation process to produce lafun may be more favorable for the production of kojic acid [42] than that used for gari or fufu. Poisoning from the consumption of oriental fermented foods containing kojic acid, where its presence is common, has not been reported in humans, although there are still inconsistent and controversial results on kojic acid toxicity [43]. Additionally, Nohynek et al. [44] reported that the existing literature on the toxicity of kojic acid is somewhat inconclusive, even though it has been stated from the genotoxicity and human health risk of topical use of kojic acid that consumer exposure to fermented foods does not pose a significant risk to human health.
These 16 metabolites could be the most common in dried cassava products in Nigeria. Of these, the most common metabolites associated with Aspergillus spp. were kojic acid, asperphenamate, N-benzoyl-phenylalanine, emodin, and asperglaucide. The predominant metabolites of Alternaria spp. were cyclo (L-Pro-L-Tyr), cyclo (L-Pro-L-Val), and alternariol methyl ether. Furthermore, tryptophol was the most common metabolite associated with Fusarium spp. and brevianamid F, fellutanine A, and rugulusovi , associated with Pennicillum spp., were the most commonly identified metabolites of that species (see Table 2). These results agree with those obtained in previous studies [11,[38][39][40] of various staple foods from some countries, including cassava products.
Kojic acid, asperphenanate, N-benzoyl-phenylalanine, emodin, and asperglaucide are all metabolites associated with Aspergillus spp. (see Table 3). The concentrations of kojic acid, a 5-hydroxy-2-hydroxymethyl-4-pyranone, in the dried product samples ranged from 8.35 to 1754.80 µg/kg; lafun samples had the highest, and cassava starch the lowest concentration. The kojic acid content of lafun in this study was lower than the maximum values (650,000 µg/kg and 93,700 µg/kg) recorded in cassava samples from Tanzania and Rwanda, respectively [31]. However, the processing methods for these cassava products were not indicated. Kojic acid can also be produced from various carbohydrate sources in an aerobic condition by a variety of microorganisms [41]. The fermentation process to produce lafun may be more favorable for the production of kojic acid [42] than that used for gari or fufu. Poisoning from the consumption of oriental fermented foods containing kojic acid, where its presence is common, has not been reported in humans, although there are still inconsistent and controversial results on kojic acid toxicity [43]. Additionally, Nohynek et al. [44] reported that the existing literature on the toxicity of kojic acid is somewhat inconclusive, even though it has been stated from the genotoxicity and human health risk of topical use of kojic acid that consumer exposure to fermented foods does not pose a significant risk to human health.   Unlike kojic acid, asperphenamate is an unusual ester of N-benzoyl-phenylalanine and N-benzoyl-phenylalaninol produced by Aspergillus spp., Penicillium spp., and plants [45,46]. The concentration of this metabolite was highest in yellow kpo-kpo gari (270.2 µg/kg), and lowest in fine yellow gari (6.8 µg/kg). Similarly, the concentration of N-benzoyl-phenylalanine, which has the same biogenetic pathway as asperphenamate, was also highest in yellow kpo-kpo gari (141.1 µg/kg) and lowest in fine yellow gari (1.0 µg/kg). Figure 2 shows the overlaid ESI (-) MRM-chromatogram (sum of all XICs) of asperphenamaten, equisetin and epi-equisetin in a representative sample, which also contained natural toxins in cassava (linamarin and lotaustralin).
Emodin (1,3,8-trihydroxy-6-methylanthracene-9,10-dion, a natural compound belonging to the anthraquinone family, was prevalent (41.8%) in the dried cassava product samples. It occurs naturally either in a free state or combined with sugar in a glucoside and in rhubarb, cascara sagrada, aloe, and other plants. It has been found to have many health benefits, including antitumor effects on human cells [47]. Thus, emodin content in foods may not necessarily be of fungal origin [48]. The emodin concentrations in the dried product samples (quantified in free form) ranged from 0.17 to 31.17 µg/kg, with fufu flour having the highest, and cassava starch the lowest, concentrations.
The asperglaucide content of the samples was highest in lafun (385.8 µg/kg) and lowest in fine white gari (25.4 µg/kg). Asperglaucide is reported to have an anti-inflammatory effect and the ability to inhibit cysteine peptidase. Table 4 reveals the prevalence and concentrations of Alternaria, Fusarium, and Penicillium spp. metabolites in samples of various types of cassava products from Nigeria. Cyclo (L-Pro-L-Tyr), or maculosin, is a diketopiperazine formed by the fusion of tyrosine and proline that has been reported as a secondary metabolite of various fungi and bacteria on knapweed as reported by Stierle et al. [49]. These researchers also identified Cyclo (L-Pro-L-Tyr) as a host-specific phytotoxin produced by Alternaria alternata on spotted knapweed [49]. In the samples, the concentrations of this metabolite ranged from 22.4 µg/kg to 199.9 µg/kg; fufu flour exhibited the lowest and yellow kpo-kpo gari the highest concentration. Related to cyclo (L-Pro-L-Tyr) is another diketopiperazine known as cyclo (L-Pro-L-Val), which is formed by the fusion of valine and proline [50]. This was found in higher concentrations in fine white gari (625.3 µg/kg) than in yellow kpo-kpo gari (57.2 µg/kg). Alternariol monomethyl ether, which is produced by different species of Alternaria spp., has been reported to have low acute toxicity [51,52]. This metabolite has frequently been detected in apples and their products, apple juice concentrates, mandarins, olives, pepper, tomatoes and their products, oilseed rape meal, sunflower seeds, sorghum, wheat [53], and in edible oils (olive oil, rapeseed oil, sesame oil, sunflower oil), among others [54]. The alternariol methyl ether content of the dried cassava product samples ranged from 0.02 µg/kg to 1.49 µg/kg. Fufu flour had the lowest content, and fine yellow gari, the highest. Samples of cassava starch and tapioca did not contain alternariol methyl ether at detectable levels.
The only Fusarium spp. metabolite which was present in more than 75% of the cassava product samples was tryptophol. This is an aromatic alcohol that induces sleep in humans and is produced by many microbial species [55]. It is also produced by the trypanosoma parasite in wine as a secondary product of alcoholic fermentation [55]. Tryptophol may also be formed from an amino acid (tryptophan) during fermentation [31]. Lafun had the highest (1121.9 µg/kg), and fine yellow gari the lowest (121.3 µg/kg), tryptophol content (see Table 4). Brevianamid F, fellutanine A, and rugulusovin metabolites associated with Penicillium spp. were prevalent (>75%) in the cassava product samples (see Table 4). Brevianamid F is a cyclic dipeptide produced by many species of Penicillium and an intermediate in the production of many other fungal metabolites [31]. The brevianamid F content of the dried cassava product samples was highest in fine white gari (44.0 µg/kg) and the lowest in fufu flour (7.1 µg/kg). Fellutanine A is one of the bio-active diketopiperazine alkaloids often produced by Penicillium fellutanum and Penicillium simplicissimum [56], which is also a non-annulated analogue of cyclo (L-Trp-L-Trp). This implied that fellutanine can also be produced from the amino acid tryptophan during fermentation [56]. The concentration range of this metabolite in the samples was 0.02 µg/kg to 4.14 µg/kg, with fine white gari having the highest, and tapioca the lowest, concentrations. The rugulusovin content of the dried product samples ranged between 0.06 µg/kg and 2.05 µg/kg. The values were highest in tapioca and lowest in fine yellow gari. Rugulusovin was not detected in high-quality cassava flour, possibly because of the absence of fermentation in the processing method (Tables 2 and 4).
As shown in Tables 2-4, some emerging mycotoxins, namely, beauvericin, moniliformin, emodin, alternariol methyl ether, and tenuazonic acid occurred in more than 5% of the total number of cassava products studied. Sterigmatocystin and O-Methylsterigmatocystin occurred in more than one sample, but less than 5% of the samples (see Supplemental Table S1). While there are no significant differences in the concentration of Emodin among the cassava products, Yellow kpo-kpo gari had significantly higher (p < 0.05) concentrations of alternariol methyl ether than any of the other cassava products, suggesting possible role of processing method or presence of carotenoids (antioxidants that are present in yellow cassava roots), in the formation of alternariol methyl ether. In addition, the fermented cassava products (Lafun, fufu flour, fine yellow gari, fine white gari and yellow kpo-kpo gari) exhibited consistent, significantly higher(p < 0.05) concentrations of microbial metabolites than non-fermented products (cassava starch, HQCF, and tapioca) (see Tables 3 and 4). Lafun contained significantly higher (p < 0.05) concentrations of kojic acid (1755 ± 7196 µg/kg) than all the other products. Yellow kpo-kpo gari contained significantly higher (p < 0.05) concentrations of asperphenamate (270± 655 µg/kg), N-benzoyl-phenylalanine (141± 384 µg/kg), cyclo (L-Pro-L-Tyr) (200 ± 18 µg/kg) and cyclo (L-Pro-L-Val) (57± 56 µg/kg) than all the other products, while fine white gari contained significantly higher (p < 0.05) concentration of brevianamid F (44 ± 52 µg/kg). Available knowledge suggests that food processing causes the masking of some mycotoxins through oxidation, reduction, or conjugation phenomenon [24]. From the preceding, there is an indication that the existing diverse traditional cassava processing practices in different countries, most of which involve fermentation by different chance-microorganisms, could alter the metabolites found in cassava products. In the light of this, further understanding of the diversity and concentration of emerging mycotoxins in cassava products would be required with regard to the effect of different processing practices and the presence of beta carotenoids in some cassava varieties.

Conclusions
The results of this study showed that regulated mycotoxins, based on European regulations, were not prevalent in any dried cassava product sample from Nigeria. The results, therefore, indicate that consumers of dried cassava products made in Nigeria are not exposed to high levels of regulated mycotoxins. Nevertheless, the study recommends further studies on the role of different processing practices in the alteration of the contents of emerging mycotoxins in cassava products. Additionally, precautions in the form of establishing hygiene and industrial standards for raw materials combined with other operational protocols in cassava processing companies are needed to prevent accidental exposures of consumers to high concentrations of toxins in improperly processed products. Establishing protocols for manufacturing practices will be in line with the global practice of establishing permissible limits, preventing food toxins, and reviewing the limits and practices from time-to-time, taking account of new advances in scientific and technical knowledge on the toxins and any new variants of the associated microorganisms.

Sampling of Dried Cassava Products Traded in Nigeria
Three hundred and seventy-three samples of dried cassava products were taken from processors and vendors located in the humid forest (92), derived savannah (267), and Southern Guinea Savannah (14) zones. The distribution was as follows: tapioca: 36 samples, white kpokpo gari: 52, yellow kpokpo gari: 12, fine yellow gari: 50, fine white gari: 113, fufu flour: 36, lafun: 30, starch: 15, and high-quality cassava flour (HQCF) 29. All of the products were properly sampled by quartering before sending to the laboratory for analyses. Each cassava product (200 g) collected was a representative of the sampling frame, which was based on the relative quantities of the products processed from fresh cassava and traded in each agroecological zone. All of the samples were collected during the rainy season. Samples were kept in polypropylene bags and transported to the Center for Analytical Chemistry Laboratory in the Department of Agrobiotechnology, University of Natural Resources and Life Sciences, Vienna, for analysis.

Determination of Mycotoxins and Other Microbial Metabolites
The mycotoxin metabolites were determined using the method reported in our previous study by Malachova et al. [57]. The accuracy of the method was verified on a routine basis by inter-laboratory comparison trials organized by BIPEA (Gennevilliers, France). Z-scores are in the acceptable range of −2 < z < 2 for ca. 90% of the submitted results, with other results, mainly deriving from matrices, that have not been previously validated [58].
The samples (5 g) were weighed into a 50-mL polypropylene tube (Sarstedt, Nümbrecht, Germany), and the extraction solvent (acetonitrile/water/acetic acid 79:20:1, v/v/v) was added at a ratio of 5 mL per gram of sample. Samples were extracted for 90 min on a GFL 3017 rotary shaker (GFL, Burgwedel, Germany) and diluted with the same volume of dilution solvent (acetonitrile/water/acetic acid 79:20:1, v/v/v), and the diluted extracts injected [58]. Centrifugation was not necessary due to sufficient sedimentation by gravity. Apparent recoveries of the analytes were taken from the analysis of 15 individual spiked samples.
A QTrap 5500 LC-MS/MS System (Applied Biosystems, Foster City, CA, USA) equipped with a TurboIonSpray ® electrospray ionization (ESI) source and a 1290 Series HPLC System (Agilent, Waldbronn, Germany) was used for the liquid chromatography-mass spectrometry/mass spectrometry (LC-MS/MS) screening of target microbial metabolites [59]. Chromatographic separation was performed at 25 • C on a Gemini ® C18 column (Phenomenex, Torrance, CA, USA ), of 150 × 4.6 mm ID, with 5 µm particle size, equipped with a C18 4 × 3 mm ID SecurityGuard™ cartridge (Phenomenex, Torrance, CA, USA). ESI-MS/MS was performed in the time-scheduled multiple reaction monitoring (MRM) mode, both in positive and negative polarities, in two separate chromatographic runs per sample by scanning two fragmentation reactions per analyte. The MRM detection window of each analyte was set to its expected retention times of ±27 and ±48 s in the positive and the negative modes, respectively. Confirmation of positive analyte identification was obtained by conducting two MRM assays per analyte (except in the case of moniliformin, which exhibited only one fragment ion). This yielded 4.0 identification points according to European Union Commission Decision 2002/657 [60]. Additionally, the LC retention time and the intensity ratio of the two MRM transitions agreed with the related values of an authentic standard within 0.1 min and 30% rel., respectively. The MRM transitions for all the major toxins and the metabolites identified in this work have previously been detailed in Malachova et al. [53]. Quantitation was performed using external calibration in connection with apparent recoveries previously determined for cassava [28]. Isolation and identification of microorganisms were not included in this study.

Statistical Analysis
Analysis of variance (ANOVA) and separation of the mean values (using Duncan's multiple range test at p < 0.05) were done using Statistical Package for the Social Sciences (SPSS) software, version 21.0 (SPSS, Inc., Chicago, IL, USA).

Supplementary Materials:
The following are available online at www.mdpi.com/2072-6651/9/7/207/s1, Table S1: List of all the 91 microbial metabolites detected in more than one cassava product from Nigeria.