Mycobiota and aflatoxigenic profile of Portuguese almonds and chestnuts from production to commercialisation

Aflatoxin (AF) contamination of nuts is an increasing concern to the consumer’s health. Portugal is a big producer of almonds and chestnuts, but there is no scientific knowledge on the safety of those nuts. AFs B1, B2, G1 and G2 are produced mainly by some species of Aspergillus belonging to section Flavi, which is composed of a large number of very closely related species. While these species are difficult to differentiate morphologically and even genetically, they differ in a characteristic that is of paramount importance for food safety, as only some are responsible for the production of the highly toxigenic AFs. Taxonomy and species identification are therefore subject of great interest, and the establishment of schemes for species and for aflatoxigenic strains identification that are simultaneously accurate, sensitive, robust and expedite is mandatory. This work had three major goals: the first was to provide knowledge on the general mycobiota, aflatoxigenic fungi and AF contamination of Portuguese almonds and chestnuts, and its evolution throughout the various stages of production (field, storage and processing). For this matter, 45 chestnut samples were collected from orchards from Trás-os-Montes. Forty-seven almond samples were collected in Trás-os-Montes at different stages of production: field, storage and processing. All fungi belonging to genus Aspergillus were isolated and identified to the section level, and all isolates belonging to section Flavi were further tested for their aflatoxigenic ability. Fungi representative of other genera were identified to the genus level. Almond samples were tested for AF contamination. The mycobiota of almonds and chestnut was found to vary in terms of both matrix and stage of production. Chestnuts were mainly contaminated with the genera Fusarium, Cladosporium, Alternaria and Penicillium, and the genus Aspergillus was only rarely found, whereas almonds were more contaminated with Aspergillus. No Aspergillus section Flavi were isolated from chestnuts. In almonds, Fusarium, Cladosporium, Alternaria and Penicillium decreased from field to the end of processing, whereas Aspergillus increased significantly, including those from section Flavi. In total, 352 fungi belonging to section Flavi were isolated from Portuguese almonds, of which 231 isolates (66%) were aflatoxigenic. Even so, only one sample from storage was found to be contaminated with AFs (4.97 µg/kg) at a level below the maximum levels recently imposed by the Commission Regulation (EU) No 165/2010. The second goal of this work was to characterise and identify the isolates of Aspergillus section Flavi by applying a polyphasic approach including classic phenotypic and molecular methods as well as the innovative technology protein spectral analysis Matrix-Assisted Laser Desorption/Ionisation-Time of Flight Intact-Cell Mass Spectrometry (MALDI-TOF ICMS), and to devise accurate and sensitive schemes for species identification. For the morphological analysis, fungi were cultured on different media and were characterised for several macro and micro morphological features. Morphological analysis was complemented with biochemical analyses, which consisted of determining the extrolite profiles relative to AFs and cyclopiazonic acid. A group of selected isolates was identified molecularly based on the sequencing of the ITS region and partial calmodulin gene. Spectral analysis was made by MALDI-TOF ICMS to obtain spectra of protein masses. Dendrograms of relatedness were obtained for each set of data and used to compare sensitivity and accurateness of the different approaches. From the preliminary morphological analysis, three morphotypes were identified: as “A. flavus morphotype” (36.4% of the isolates), “A. parasiticus morphotype” (55.4%), and “A. tamarii morphotype” (8.2%). The 3 morphotypes were then divided into 9 phenotypes based on their extrolite profile. Genotypic and spectral analyses clustered the selected isolates into the same 3 groups created by morphological analysis. Furthermore, all sets of data, including the morphological complemented with extrolite profile, were able to further resolve the isolates into more restrictive clusters. They all positioned two of the 9 phenotypes in two unidentified terminal clades closely related to A. parasiticus. The third goal was to test a molecular method based on multiplex PCR and RT-PCR for the ability to differentiate aflatoxigenic and non-aflatoxigenic isolates. Two genes of the AF biosynthetic pathway, aflD (= nor1) and aflQ (= ord1= ordA), were tested for presence and expression (by PCR and RT-PCR, respectively). The presence of both genes did not correlate with aflatoxigenicity. In terms of gene expression, aflD was not considered a good marker for differentiating aflatoxigenic from non-aflatoxigenic isolates, but aflQ showed a good correlation between expression and AF-production ability. In conclusion, Portuguese almonds and chestnuts seem to be generally safe in terms of AF contamination. Nevertheless, the majority of the isolates of Aspergillus section Flavi obtained from Portuguese almonds was found to be aflatoxigenic, which may constitute a problem in terms of food safety if storage and processing conditions are not effectively controlled. At present, these conditions seem to be guaranteed, since only one almond sample was found to be contaminated. At the species identification level, good agreement was obtained between the 3 methods of analysis since they all generated similar dendrograms with concordant strain clustering. Morphological analysis has shown sensitive and reliable as a preliminary method for species identification only when complemented with the extrolite profile. The calmodulin gene showed to be more robust and reliable as genomic marker for this group of fungi than the ITS region, providing good DNA barcoding potential. MALDI-TOF ICMS results confirmed that this technique is highly reliable for fungal identification, and is faster and less expensive in terms of labour and consumables when compared with other biological techniques, which is essential whenever there is a paucity of characters for defining many fungal species and when high numbers of isolates are involved. Expression analysis of the aflQ gene seems to be a good method for the differentiation of aflatoxigenic and non-aflatoxigenic isolates