Quantification of the allergen soy (Glycine max) in food using digital droplet PCR (ddPCR)

Abstract

To meet the increasing need for quantification of allergens and to have an alternative to commercially available ELISA and PCR systems, the Austrian Agency for Health and Food Safety started establishing in-house PCR systems. To obtain low limits of detection (LOD) and quantification (LOQ), target sequences are preferably sought in multicopy genomes like mitochondrial- or chloroplast DNA. These molecules are of high but varying abundance even among tissues of the same organism. Beyond that, DNA might be degraded by processes of food manufacturing which additionally affects their quantification. Therefore, a reliable correlation of the allergen portion in a sample and its chloroplast-DNA concentration cannot be preassumed. This incoherence is not further considered (e.g., by a matrix-related reference material), and therefore, our quantitative results can only be understood as the mass of soy which maintained its biochemical activity, related to the soy content of the reference material used. To convert absolute results expressed in copies per microliter (Cp/µL) as obtained by digital droplet PCR (ddPCR) into a unit of mass fraction (e.g., milligram per kilogram), a conversion function is generated by the measurement of a reference material in the same run. For the specific detection and quantification of the allergenic ingredient soy (Glycine max) in food a primer/probe system has been developed which amplifies a 140 bp product of the ndhH gene of the chloroplast DNA. It is specific for soy and does not react with even closely related plant species. Digital droplet PCR (ddPCR) was selected for quantification for its particular advantages and the method has been validated in-house. It was found to be applicable to various matrices including meat products, flour, milk, and fatty creams, with recovery rates between 60 and 100%. The limit of detection and the limit of quantification (LOQ) are 0.16 mg/kg and 0.60 mg/kg, respectively. Repeated analysis of analyte-free food matrices spiked with reference material provided acceptable values for precision: The relative standard deviation (RSDoverall) of the whole method (including DNA extraction) is below 25%. The recovery of pure soy material (pulverized beans) was between 112.5 and 135.0%. The presented method is shown to be reliable and accurate, provided that samples and reference material are extracted and amplified in the same way.