Allergen analysis and authenticity control: Development of mass spectrometry‐based methods for animal‐derived food products

Abstract

Food allergy and food fraud involving animal‐derived products are two of the most significant issues in food markets. On one hand, immunoglobulin E (IgE)‐mediated allergic reactions after ingestion of fish, crustaceans, eggs, or milk are among the most prevalent and can happen even after ingestion of trace amounts. On the other hand, new rules regarding product commercialization (e.g., novel food regulation) are more and more created while fraudulent species substitution in fishery products is very common. Sensitive and accurate analytical methods for allergen quantification and species identification in commercial food products are therefore urgently required whether to help food industries inform allergic consumers, to ensure the food compliance with new regulations or to combat food fraud.In the past few years, bottom‐up proteomic techniques, which rely on the detection of peptide biomarkers resulting from a tryptic digest of food proteins using liquid chromatography coupled to tandem mass spectrometry (LC‐MS/MS), have been emerging in this field. The selection of reliable allergen‐specific or species‐specific peptide biomarkers is one of the most crucial steps when developing such methods whether for qualitative protein detection (i.e., screening analysis) or protein absolute quantification.The first part of this dissertation relates therefore to the selection of allergen peptide biomarkers for fish, invertebrates, eggs, and milk in an experimental way using a single chaotropic urea extraction buffer. The allergenic proteins responsible for those severe reactions are mainly parvalbumin, tropomyosin, ovalbumin, and caseins.The protein extraction was first assessed via sodium dodecyl sulfate‐polyacrylamide gel electrophoresis (SDS‐PAGE) and the major allergens were well extracted. After that, data‐dependent MS/MS spectra which were obtained using digest samples of authentic animals or pure allergenic ingredients were processed against a matching protein database, and identified peptides were filtered according to several criteria such as the sequence length, amino acid composition, specificity, reproducibility, and sensitivity. Myosin proteins were also a target of choice for fish detection due to their high prevalence and sequence homology even if they are non‐allergenic. Contrary to fish and invertebrate databases which included entries for only one specific protein (i.e., parvalbumin, myosin, or tropomyosin), egg and milk databases contained all known allergenic proteins or even additional proteins for egg yolk. Two extra selection steps were achieved for egg and milk markers based on more stringent criteria regarding their sensitivity after targeting them in the corresponding allergenic ingredients. At this point, 17, 13, 10, 10 and 12 potential markers were respectively selected for fish, invertebrates, egg white, egg yolk and milk. An alignment algorithm was used for all those markers to get an idea about their biological specificity. The biological specificity was verified experimentally for fish and invertebrate potential markers by targeted analysis in digest samples of animal species that are relevant in the food industry (vertebrates and invertebrates). This verification was not done for egg and milk markers as it was not regarded as significant due to reported cross‐reactivity among avian eggs and among mammalian milk. Two potential parvalbumin markers were detected in other nonfish vertebrates, while all potential tropomyosin markers except one were specific to at least an invertebrate class belonging to the same phylum. Parvalbumin and myosin markers as well as tropomyosin markers were exclusively found in vertebrates or invertebrates. Marker detectability was checked by analyzing processed fish products as well as cooked fish for potential parvalbumin and myosin markers, while commercial insect‐based food products such as cereals bars or pasta were studied for potential tropomyosin markers. All expected fish and invertebrate markers were detectable in those complex food products. Detectability of egg and milk markers was assessed by analysis of bread, cookies, and chocolate samples contaminated at different stages of the sample preparation with trace amounts (100 pg/g) of eggs and milk (i.e, fortified, spiked, and incurred digest samples). Two egg white markers and seven milk markers were detected in all those samples. The most suitable markers in terms of sensitivity and specificity were finally chosen for each allergenic product. Thus, besides two myosin fish global markers, five parvalbumin markers were retained including at least one of the investigated fish species. In addition, five tropomyosin markers were chosen, their specificity allowing us to distinguish crustacean tropomyosin from that of insects/arachnids, or mollusks. At last, two ovalbumin markers and three casein markers were confirmed to be the most suitable allergen markers respectively for egg white and milk. All those retained markers could be compiled in a single multiplex method. The automation of the sample preparation could also be a promising improvement whether for qualitative or quantitative analysis.