Predictive mycology and use of natural antifungals to prevent the mycotoxin food hazard

Abstract

Filamentous moulds may cause spoilage in raw materials, foods and feeds. Some of them synthesize mycotoxins which are a risk for human and animal health. For this reason, from the food safety point of view, only mycotoxins, as chemical hazards, are important. Nevertheless, despite the absence of direct correlation between mould growth and mycotoxins production, the prevention of fungal growth in raw materials and foods leads invariably to the prevention of mycotoxins presence. Due to the fact that moulds can contaminate foods from raw materials till end products, different strategies could be used at the different steps in the food chain. Preharvest strategies include the use of resistant varieties, crop rotation, soil preparation, optimal irrigation, fertilizer, herbicides, insecticides and chemical and biological agents application. Post-harvest strategies include improved drying and storage conditions, together with the use of natural and chemical agents. Predictive models may be used as a strategy to predict and prevent mycotoxigenjc fungal growth and mycotoxins accumulation. The present PhD work focused in two main strategies: a) The use of antifungals of natural origin to prevent from mycotoxigenic fungi and mycotoxins Equisetum arvense and Stevia rebaudiana extracts were analized as possible natural agents to inhibit growth and mycotoxin accumulation in in vitro and in vivo experiments. Both extracts were effective against mycotoxigenic moulds and the mycotoxigenic Aspergillus and Fusarium isolates studied were completely inhibited by a 3% of E. arvense. However, the effect decreased in the in vivo test. In the last case, Equisetum was effective against Aspergillus section Flavi and Fusarium section Liseola growth at high water activity levels and with high infection levels, but mycotoxins levels were not significantly affected. b) The assessment of the usefulness of predictive models to manage the mycotoxin problem In an initial experiment, four particular points which deserved in depth study to assess the viability of predictive microbiology in the moulds field were identified: 1) models should be developed for longer time periods; 2) food and raw materials prone to mycotoxin contamination are usually stored under marginal conditions for mould growth, thus performance of models should be checked under such conditions; 3) the impact of the inoculum size in the performance of the models; and 4) the impact of the potential intraspecies variability among isolates in prediction performance. Prediction of time to growth by kinetic models was clearly linked to inoculum size. On the other hand, the performance of predictive models may be compromised under marginal conditions for fungal growth, the higher variability of results under these conditions results in the need for a higher number of replicates required, specifically for kinetic models. For last, a high intraspecific variability on growth and mycotoxin levels has proven to be wider for the both isolates studies: A. carbonarius and P. expansum. For this reason, a greater number of strains should be included to develop models under non optimal condition for both, growth and for mycotoxin production. A matrix was built from which the number of strains and replicates to be planned for new experiments can be assessed for a reliable estimation of growth parameters and we conclude that increasing the number of strains in an experiment increases the explained variability much more than including further replicates. Finally, a first attempt was done to model aflatoxins production as a function of growth parameters and time. Aflatoxins accumulation was shown to be better correlated to colony area than either colony diameter or fungal biomass. Luedeking-Piret model was used for this purpose, and reasonable percentages of variability were explained. To conclude, probability models applied either to mould growth or mycotoxin production might be a valuable tool in food safety management through the food chain.