Purpose: This study tracked changes in gluten detection in traditionally brewed soy sauce samples obtained at different stages of production using five commercial gluten ELISA kits.
Methods: Traditionally brewed tamari soy sauces formulated without wheat were produced in a pilot plant and spiked with 0 μg/g, 7.2 μg/g and 72 μg/g gluten at the moromi mash stage of production. Traditionally brewed regular soy sauce formulated with wheat was produced in a pilot plant. Tamari and regular soy sauce samples were collected during various stages of production and analyzed in triplicate using five commercial gluten ELISA kits.
Results: Gluten concentrations in the tamari moromi mash spiked with 72 μg/g gluten were 19 (4% CV) – 31 (3% CV) μg/g, as measured with the Ridascreen sandwich, Ridascreen competitive, AgraQuant G12 sandwich, Aller-Tek sandwich, and Morinaga sandwich ELISA kits. The high sensitivity of the Morinaga ELISA (LOQ, 0.27 μg/g) enabled the lowest detection of gluten in the tamari moromi mash spiked with 7.2 μg/g gluten, while other kits were unable to detect gluten in this sample. Gluten concentrations in tamari sauces after 1-6 months of fermentation were <LOQ values for all ELISA kits. Analyses of regular soy sauce samples indicated the presence of 5,600 (5% CV) – 32,000 (6% CV) μg/g gluten in koji, and 2,100 (0.1% CV) - 18,000 (4% CV) μg/g in the moromi mash. Gluten levels in regular soy sauce after 1-6 months of fermentation were <LOQ values of the ELISA kits.
Significance: This study demonstrates that reliable detection and accurate quantification of gluten, subjected to the fermentation process used in the production of soy sauce, is not possible using available ELISA technology.