Analysis of mycotoxins in the food chain by enzyme immunoassay: possibilities and limitations

Abstract

Indole alkaloids have already been detected in a number of foods and beverages in widely varying concentrations. However, the data on the occurrence and detection of these compounds is still very limited. A similar picture of the data situation emerges for the mycotoxicological quality of special products such as processed cereal-based foods for infants and young children (PCF) as well as for plant-based milk alternatives (PBMAs) consumed by both adults and children. The studies conducted in this work provided new insights into the detection and occurrence of mycotoxins, especially indole diterpene alkaloids (IDTs), in foods. The first study was concerned with the development of an enzyme immunoassay (EIA) for the detection of the IDTs penitrem A (PTA) in methanolic mycelial extracts prepared from fungal cultures of mouldy food. To cover as broad a spectrum of indole-type toxins as possible, two established EIAs for the detection of paxilline (PAX) and ergoline alkaloids were included in the experiments. Analysis of the mycelium extracts by EIAs revealed that fungi derived from spoiled food produce PTA, PAX and other ergoline alkaloids. Comparative investigations using LC-MS/MS showed that in addition to PTA and PAX, their analogues (penitreme B-F; deoxy-PAX) could also be detected in mycelium extracts. This finding provided evidence that the PTA-EIA detects not only PTA but also its analogues. A further investigation of selected PTA-positive fungal isolates by sequencing the internal transcribed spacer region and the β-tubulin gene resulted in an identification of P. crustosum and P. polonicum, whereby P. polonicum was identified as a PTA former for the first time. The second study investigated the microbiological and mycotoxicological quality of PCF from the German market. A number of microbiological parameters (aerobic mesophilic plate count, moulds, Enterobacteriaceae, Cronobacter spp., presumptive Bacillus cereus) were investigated, using standard microbiological methods in each case. The mycotoxin studies included deoxynivalenol (DON), zearalenone (ZEN), alternariol (AOH), T-2/HT-2 toxin (only samples containing oats) and ergot alkaloids, with sample extraction based on established methods. Study 2 showed the detection of low concentrations of AOH, DON, T-2/HT-2 (oat-containing products only) in PCF, whereas ZEN and ergot alkaloids could not be detected. Another important finding was that a few samples contained low numbers of opportunistic pathogens, in particular Cronobacter sakazakii, Acinetobacter spp. and Pantoea spp. as well as enterotoxigenic Bacillus wiedmannii. The results obtained within the study were all within the existing European Union regulations or international guidelines. Study 3 investigated the applicability of mycotoxin-EIAs for direct analysis of PBMAs of different composition without prior sample extraction. For this purpose, the extent of matrix influence of different PBMA matrices on the EIAs for the detection of aflatoxin B1 (AFB1), sterigmatocystin (STC), Ochratoxin A (OTA), DON and T-2/HT-2 was first determined. It was found that the matrix influence turned out to be test-specific and consequently different high dilutions of the samples in the respective EIAs were necessary. After analysis of the PBMA samples from the German market, positive results were obtained in at least one test system for 43 % of the samples, with most positive results being close to the calculated detection limit. Concomitant LC-MS/MS analyses of EIA-positive samples for AFB1, STC, and OTA yielded poor quantitative agreement, which was attributed to matrix interference from sample ingredients. This investigation provided evidence that the sample dilutions used in the EIA were insufficient to overcome matrix interference from the constituents present in PBMAs, thus compromising the reliability of the EIA results. As the use of higher dilutions conflicts with sufficiently sensitive detection systems, the use of EIAs for routine mycotoxin analysis in PBMAs requires further investigation to develop a viable sample preparation method.