Introduction

Mycotoxins are secondary metabolites naturally produced by fungi or moulds (WHO, 2018). They are small, highly toxic chemical products that can contaminate foods or feedstuffs at any stage along the food chain (WHO, 2018). In addition, mycotoxins are considered chemical hazards in foods (Cinar and Onbasi et al., 2019) since they are known to have adverse health effects on humans and animals after ingestion, referred to as mycotoxicosis (WHO, 2018). Equally, the danger of mycotoxins should not be undermined since many mycotoxins have been reported to be carcinogenic (Sengül et al., 2016).

Mycotoxin contamination of agricultural plant products has become a significant issue around the world (Luo et al., 2021) due to post-harvest infection by mycotoxin- producing fungi such as Aspergillus spp., Alternaria spp., Penicillium spp. and Fusarium spp. (Barkai-Golan and Paster, 2008). Nevertheless, most past studies were mainly focused on cereals, grains, maise, bread, fruits, nuts, juices, sausages, infant cereals, dried fruits, legumes and dried spices (EU, 2006; de Medeiros et al., 2012; Guo et al., 2016; Eskola et al., 2019; Cinar and Onbasi et al., 2019; Kifer et al., 2019). However, vegetables, which are considered essential commodities in the diet of people (Liu et al., 2020), have also been found to be contaminated by mycotoxins (Gupta et al., 2009). The latter included aflatoxin (AF), alternariol (AOH), citrinin (CIT), deoxynivalenol (DON), fumonisin (FUM), ochratoxin (OT), T-2/HT-2 toxin (T-2/HT-2) and zearalenone (ZEN) (Adegoke and Letuma, 2013).

However, since mycotoxins are heat stable, they cannot be destroyed easily during thermal processing or cooking (Liu et al., 2020) and therefore represent a food safety risk. In Mauritius, vegetable crops which are widely produced and consumed include potato, tomato, onion, pumpkin and garlic. Till now, very few studies have recently been conducted on the detection or quantification of mycotoxin and its relative concentration in vegetables, for instance, potatoes (Youssef and Sabra, 2022), tomatoes (Ji et al., 2023), garlics (Jeswel and Kumar, 2016), pumpkins (Sahar et al., 2009) and onions (Gherbawy et al., 2015). However, to our knowledge, no studies have been reported on detecting FUM in pumpkins. The objective of this study was, therefore, to determine the level of mycotoxin contamination on selected vegetable crops commercially available in the Mauritian market.

Materials and Methods

Sample collection and fungal isolation

Samples of diseased garlic (Allium sativum), onion (Allium cepa), potato (Solanum tuberosum), pumpkin (Cucurbita moschata) and tomato (Solanum lycopersicum) suspected to be infected by pathogenic fungi were collected at the postharvest stage from open-fields, packhouses and markets of Mauritius from February 2019 to September 2021. A total of 287 samples of vegetables were collected: garlic (n=20), onion (n=40), potato (n=125), pumpkin (n=10) and tomato (n=92). To identify the fungal agent, the diseased samples were placed in moist chambers for 24 h to favour the development of fungal mycelia and subsequently examined by microscopy using a bright-field microscope.

The following day, eight squares with dimensions of 1 cm x 1 cm were cut to include a margin of the diseased and healthy tissue from the different samples. The cut pieces were gently flushed with continuous water for at least 2 h and then surface-disinfected for 30 s with sodium hypochlorite solution (1:3 sodium hypochlorite to water) and left to dry on sterile tissue paper. The samples were then transferred on Potato Dextrose Agar (PDA) amended with chloramphenicol. The inoculated plates were incubated at room temperature for 7 days in the dark. After incubation, observation of the macroscopic morphology of the colonies and microscopic examination of the fungi were conducted. If the culture obtained corresponded to the suspected fungus, it was sub-cultured using the hyphal tip method on a new PDA plate to get a pure culture (Takooree et al., 2021).

Molecular characterisation of the etiological agent of diseased samples

After 7 days of incubation, fungal isolates presumptively identified on PDA were selected for molecular identification. Briefly, mycelial growth was scraped from the plate, weighed, and ground with liquid nitrogen (Takooree et al., 2022). DNA was extracted by the Cetyl Trimethyl Ammonium Bromide (CTAB) method (Madarbokus and Ranghoo-Sanmukhiya, 2012) and subsequently subjected to PCR analysis. ITS regions of rDNA were amplified using the ITS primers, ITS5 (5’-GGAAGTAAAAGTC GTAACAAGG-3’) and ITS4 (5’-TCCTCGCTTATTGAT ATGC-3’) (Ristaino et al., 1998). The thermal cycling parameters comprised of initial denaturation at 94°C for 3 min followed by 30 cycles consisting of denaturation at 94°C for 1 min, annealing at 52°C for 1 min, and extension at 72°C for 2 min followed by a final extension at 72°C for 10 min. DNA sequencing reactions were done using a Big Dye Terminator v. 3.1 Cycle Sequencing Kit (Applied Biosystems) following the protocol outlined by the manufacturers. Sequencing reaction products were purified by the ExoSAP method. They were directly sequenced in both directions using an automated sequencer (ABI 3500 DNA sequencer (Applied Biosystems) at Inqaba Biotechnical Industries (Pty) Ltd, South Africa, using the same primers for amplification. Forward and reverse sequences were assembled and edited using CLC Main Workbench Version 22.0 (https://www.qiagenbioinformatics.com/). Consensus sequences were computed using the ClustalW (Thompson et al., 1994), integrated in MEGA11 software (Tamura et al., 2013), and deposited in GenBank (http://www.ncbi.nlm.nih.gov). All generated sequences were compared by calculating nucleotide (nt) similarities. Moreover, they were compared with previously deposited Fusarium spp. and Alternaria spp. isolates available in GenBank (Table S1), using the similarity search tool BLAST. The phylogenetic tree was made using PAUP Version 4.0 beta for the Parsimony Method, and the reliability was confirmed by bootstrapping using 1000 random replicates. Fusarium nygamai and Alternaria gypsophila were designated as the fungal outgroup to validate the obtained results.

Assessment of mycotoxin level in samples

On each sampling occasion, infected vegetable samples were examined, and sections of the diseased parts of the samples were aseptically cut with a knife and examined microscopically. It was observed that the different vegetables were found to be susceptible to infection by other mycotoxin-producing fungal agents. Table 1 thus indicates the corresponding mycotoxins for which the various vegetables were analysed. For all vegetables except for pumpkin, five cut sections of the same vegetable kind were pooled to form a composite sample, which was subsequently investigated. Pumpkin sections were analysed individually due to the limited availability of diseased samples. Each composite sample was blended to produce a slurry, which was later analysed by Enzyme-Linked Immuno-Sorbent Assays (ELISA) in duplicate. The quantitative analysis of AOH, AF, CIT, DON, FUM, OT, T-2/HT-2 and ZEN was carried out using ELISA; Beacon Analytical Inc. Alternariol Plate Kit (Cat. # 20-0288), Ridascreen® TOTAL Aflatoxin (Art. No. R4701), Ridascreen® Fast Citrinin (Art. No. R6302), Ridascreen® DON (Art. No. R5906), Ridascreen® Fumonisin (Art. No. R3401) or Ridascreen® Fast Fumonisin (Art. No. R5602), Ridascreen® Ochratoxin A 30/15 (Art. No. 1312), Rida-screen® T-2/HT-2 Toxin (Art. No. R3805) and Ridascreen® Zearalenon (Art. No. 1401) test kits (R-Biopharm, Germany). All mycotoxin extraction and analyses were conducted according to the manufacturer’s instructions. Results were obtained by reading the absorbances of standards or samples at 450 nm using an ELISA microplate reader (Rida® Absorbance 96, Germany). Calcula-tions for the concentration of each mycotoxin level were done according to the instructions. The limit of detection of the assays of AOH, AF, CIT, DON, FUM, OT, T-2/HT-2 and ZEN were: not available, 1.75 μg/kg, 15 μg/kg, 3.7-18.5 μg/kg, 25 μg/kg or 222 μg/kg, 0.4-1.6 μg/kg, 12-33 μg/kg and 0.00005- 0.00175 μg/kg respectively. In addition, to validate the ELISA assay, a standard sample suspected to be contaminated with FUM was additionally tested by both ELISA and LC-MS/MS, and the results were compared.

Data analysis

One-way analysis of Variables (ANOVA) and Tukey’s multiple comparisons test were done to determine if there was a statistically significant difference (P < 0.05) in the prevalence of mycotoxins among the samples chosen for the study. The Tukey test was the post-hoc test of choice as it is more conservative than other tests.

Results and Discussion

Morphological identification

Vegetables such as garlic, onion, potato, pumpkin and tomato were found to be infected by Alternaria spp., Aspergillus spp., Fusarium spp. or Penicillium spp. from different regions of Mauritius. Microscopic analysis at the site of infection and from PDA cultures was conducted to identify the causative agent (Figure 1) presumptively. The suspected fungal isolates infecting each crop were thus confirmed based on their morphological characteristics, which included mainly the colour and shape of colonies grown on PDA, shape, colour and branching pattern of the spores or conidia or conidiophores and the presence or absence of septa (Meena et al., 2017). Each isolate of the Alternaria spp., Aspergillus spp., Fusarium spp. or Penicillium spp. demonstrated variations in culture morphology as shown in Figure 1. It can be observed that the colour of isolated Alternaria species varies, from olive-brown green or pale brown [a(i)] to white colony [b(i)], as reported by Meena et al. (2017). As for Fusarium isolates, they also vary in colony morphology, such as white to pink cottony, with purple tinge [c(i)]; orange [c(ii)]; and pale yellow and white cottony [c(iii)], similar to past studies (Nirmaladevi et al., 2016; Swamy et al., 2020)

Figure 1. Fungal isolates from infected crops showing colony morphology, spores or conidia, conidiophores, hyphae and reproductive structures stained with cotton blue lactophenol dye using light microscopy, 40X: a(i) & a(ii) Alternaria alternata; b(i) & b(ii) Alternaria solani; c(i) Fusarium oxysporum; c(ii) Fusarium graminearum; c(iii) Fusarium equiseti; c(iv) macroconidia of Fusarium spp.; d(i) & d(ii) Aspergillus spp.; e(i) & e(ii) Penicillium spp.

Moreover, microscopic examination revealed the presence of conidia of Alternaria spp. appearing brown and oval in shape with septa [a(ii) & b(ii)], macroconidia of Fusarium spp. demonstrating three to five septa together with cells appearing pointed on both ends and having a foot-like shape [c(iv)], conidiophores of Penicillium spp. with some oval small spores formed in chains [d(ii)] and round conidiophores of Aspergillus spp. with numerous oval-shaped spores [e(ii)] (Figure 1). Those microscopic observations are congruent with microscopic features observed by other authors for Alternaria spp. (Troncoso-Rojas and Tiznado-Hernández, 2014), Fusarium spp. (Thrane, 1999; Nirmaladevi et al., 2016; Swamy et al., 2020), Penicillium spp. (Palou, 2014) and Aspergillus spp. (Plascencia-Jatomea et al., 2014).

Molecular characterisation of the etiological agent of diseased samples

After sequencing the PCR products and obtaining the consensus sequences for each isolate, the BLAST analysis revealed the fungal species to be Fusarium oxysporum, Fusarium equiseti, Fusarium graminearum, Alternaria alternata and Alternaria solani, which showed 100% identity with GenBank sequences. Submission of the sequences to NCBI GenBank was carried out, and accession numbers were obtained, as shown in Table S2.

A phylogenetic tree was constructed to illustrate the possible genetic relatedness among Fusarium and Alternaria spp. isolates from onion, potato, pumpkin and tomato samples, and several mycotoxigenic fungi already published in the literature (Figure 2). The tree revealed four main clades, with two outgroups (Fusarium nygamai and Alternaria gypsophilae) occurring in separate clades. Among the clade of Fusarium species, the sequences of F. oxysporum, F. equiseti and F. graminearum were observed to form separate subclusters. A. alternata and A. solani formed the other clade. Indeed, it can be observed from the resulting phylogenetic tree that all the isolates of Fusarium spp. and Alternaria spp. were clustered together with 100% bootstrap value, supporting their correct identification.

Figure 2. Phylogenetic tree of Fusarium spp. and Alternaria spp. isolates from mycotoxin-contaminated vegetable samples compared with reference sequences obtained from NCBI.

Among the eight Fusarium species, two F. oxysporum strains from potatoes (P2 and P198) clustered well with the reference sequence (Accession No. KC478622). The latter strain has been reported to be highly virulent, causing severe Fusarium wilt in tomato plants in India (Murugan et al., 2020), thus pointing to the potentially high virulence of the two local potato strains. The remaining three strains of F. oxysporum isolated from potato (P124), tomato (T1) and onion (G11) were found to cluster with the reference sequence (Accession No. KP719138). Stefańczyk et al. (2016) reported that the latter Fusarium species were more frequently observed than other species, such as F. avenaceum, F. solani and F. sambucinum. However, according to the author, F. oxysporum was not found to be pathogenic in the laboratory, causing dry rot symptoms in potato tubers. Nevertheless, Stefańczyk et al. (2016) indicated that the agent can still present a disease risk in field or storage areas.

In addition, the other two isolates of F. equiseti (BO3 and FV1) clustered with the reference sequence (Accession No. MK290391). Since BO3 and FV1 strains both originated from the region of Moka (Supplementary Table 2), their geographical proximity could explain their genetic relatedness. It is worth mentioning that the reference strain was previously associated with causing a relatively high disease incidence in maize in South India (Swamy et al., 2020). Equally, the other reference isolate (Accession No. KP264661) was found to be a producer of trichothecene and ZEN (Stefańczyk et al., 2016). Finally, one isolate of F. graminearum from pumpkin (GF1) clustered well with the reference sequence (Accession No. KP292811). The latter strain was reported to be pathogenic in potatoes in Poland and also can produce trichothecene and ZEN (Stefańczyk et al., 2016). Hence, the isolates in this study are indeed a potential threat to the crops in Mauritius in terms of both their occurrence and ability to produce high levels of different mycotoxins.

The four recovered species of Alternaria, three A. alternata (P140, P170 and ME11) and one A. solani (G4), were observed to be clustering with the reference sequence (Accession No. KX118413). This strain was previously reported to be highly pathogenic and elaborated a high level of mycotoxins. Meena et al. (2017) said that this strain was highly toxigenic and elaborated all three mycotoxins (tenuazonic acid, AOH and alternariol mono-methyl ether) produced by Alternaria spp. Hence, this could adversely affect the tomato plant and reduce the quality and safety of the edible fruits (Meena et al., 2017). In addition, the isolate G4 (A. solani) was observed to be clustering with a very pathogenic and mycotoxigenic isolate from potato leaf (reference sequence, Accession No. KX139151), with significantly high level of mycotoxins (tenuazonic acid: 14.98 ± 1.43 μg/ml, AOH: 26.43 ± 4.52 μg/ml and alternariol monomethyl ether: 46.18 ± 4.12 μg/ml) (Meena et al., 2017). Thus, this study indicates that the fungal isolates from vegetable crops of Mauritius also have the potential to elaborate a high level of mycotoxins.

Level of mycotoxin in the vegetable samples

Most infected vegetable samples tested for different mycotoxin levels had mean concentrations within safe limits (Table 2). However, garlic and onions had a mean level of CIT (5,448.55 µg/kg) and OT (9.25 µg/kg) higher than their respective safe limits. One garlic sample had a CIT level as high as 37,179.60 µg/kg. In a previous study in Egypt, some infected onion samples were found to have a mean concentration of 30,000 µg/kg of CIT, thereby also exceeding the tolerable level (Zohri et al., 1992). Moreover, in this study, among the onion samples tested for OT, one onion sample had a higher value of 40 µg/kg compared to the safe limit of 5 µg/kg. Nevertheless, some previous studies reported that OT was below the limit of detection in infected onion samples collected from Belgium, Brazil, Egypt, India (van der Perre et al., 2013) and Saudi Arabia (Gherbawy et al., 2015).

In addition, in this study, the mean FUM level detected in potato samples was 740.14 µg/kg. Nevertheless, a relatively high level of FUM (5,531 µg/kg) was seen in a potato sample when compared with a safe limit of 4,000 µg/kg. In a past study conducted by El-Hassan et al. (2007) in Egypt, a concentration of 98 µg/kg of FUM was recorded in infected potato tubers collected. This was an acceptable level for human consumption. However, in another study, the level of FUM in sweet potatoes, another starchy tuber, was above the maximum tolerable limit (267.86 µg/kg) (CE, 2010; Amri and Lenoi, 2016). It is worth mentioning that there was no statistically significant difference in the FUM level for samples tested by ELISA and LC/MS-MS.

The mean values for the level of various mycotoxins in the different crops are presented in Table 3. It is to be noted that due to natural variability in the occurrence of mycotoxins in other samples, the values fell in a broad range, thus masking any statistically significant differences (Gibbs, 2013). Additionally, the average AOH and FUM concentrations indicated no significant differences (P > 0.05) among the different crops analysed (potato, pumpkin and tomato). It is also worth mentioning that pumpkin samples tested for these two mycotoxins had a relatively lower level than the other two crops. This could be related to the fact that pumpkin crops have a very short supply chain in Mauritius and are not stored for an extended period. As a result, they are a low-risk commodity for mycotoxin contamination. Equally, pumpkin is the only crop which is not imported, and Mauritius is self- sufficient in the production of pumpkins (AMB, 2021).

Moreover, concerning T-2/HT-2, levels in potato samples were statistically higher than those of tomatoes (P < 0.05). As for DON, there was no significant difference (P > 0.05) in the level between potatoes and onions. The high level of DON and T-2/HT-2 in potatoes produced by Fusarium spp. could be related to the fact that potatoes are stored for an extended period in storage areas, which can be damp and humid. Moreover, since they are seasonal crops, potatoes and onions are imported from several countries to meet the local demand (FAREI, 2021; AMB, 2022). Similarly, in a past study, potato was reported to be more susceptible to infection by Fusarium spp. than other vegetables tested, such as chilli, pointed gourd, garlic and onion (Gupta et al., 2009). ZEN concentration in onion, potato and tomato samples showed a significant difference, with ZEN level in onions being statistically higher than in potatoes or tomatoes. In fact, similar to potato crops, onions are also stored longer in packhouses (FAREI, 2021; AMB, 2022), presenting ample opportunities for fungal growth and mycotoxin contamination.

High levels of mycotoxin in crops could be related to several factors, with climate and storage conditions during postharvest being the most important ones (Bryden, 2012). High temperatures and moisture are contributory factors for the growth of mycotoxigenic fungi (Lacetera et al., 2019). Mauritius, being a tropical island, unfortunately, faces highly humid conditions, which can enhance the growth of mycotoxigenic fungi on crops, thereby compromising the safety and quality of produce. (Fernández-Cruz et al., 2010). It is likely that climate change, characterised by higher air temperatures and erratic rainfall events in tropical regions, could exacerbate growth and mycotoxin production by certain species of pathogenic and mycotoxigenic fungi.

Moreover, mycotoxin contamination of crops can occur at any stage of production, including cultivation, harvesting, storage, processing, transport or retail (Rahmani et al., 2009; Dombrink-Kurtzman, 2008; Yang et al., 2014; Darwish et al., 2014; Afolabi et al., 2019). Hence, ensuring the safety of agricultural produce is more complex as it requires the concerted effort of all actors in the value chain.

Unlike other food hazards, mycotoxin contamination of food is also a global “one-health” concern (Alshannaq and Yu, 2017) that can simultaneously impact human, animal and environmental health. In this study, many of the vegetable samples collected had minor signs of fungal spoilage. They were either sold at reduced prices as vegetables of inferior quality to street food vendors or were processed and transformed into other products. Since mycotoxins are generally heat stable, they cannot be destroyed easily during thermal processing or cooking (Liu et al., 2020). In addition, chronic exposure to low levels of mycotoxins, albeit frequently undetectable, is responsible for chronic carcinogenicity in humans (Alshannaq and Yu, 2017).

As for vegetables which are unfit for sale due to fungal spoilage or other, they are often discarded or used as animal feed. Studies have documented that mycotoxins accumulated in food animals are not only a health threat for animals but also represent a food safety and public health issue through ingestion of potentially contaminated animal-derived food products (Alshannaq and Yu, 2017; Duchenne et al., 2021).

Conclusions

The presence of mycotoxins in our strategic crops (garlic, onion, potato, pumpkin and tomato) represents a significant threat to the food safety and security of Mauritius. The relatively high concentrations of citrinin, fumonisin and ochratoxin detected in some vegetable samples tested urge for awareness and continued surveillance of these products in the future. Moreover, our findings indicated that extended storage of produce such as potatoes and onions can increase the risk of harbouring a high level of mycotoxin. Therefore, this research highlights the paramount importance of carrying out further mycotoxin testing in the fresh produce chain to ensure the safety and health of consumers. Future work will shed light on temperature and humidity conditions prevailing in storage houses conducive to mycotoxin contamination of vegetables.