Evaluation and comparison of three enzyme-linked immunosorbent assay formats for the detection of ricin in milk and serum

When applied to the detection of a specific protein toxin in food or biological fluids in the incidence of a potential contamination, it is crucial that the assay be both sensitive and specific. In order to identify an immunoassay which is sensitive, simple, and accurate for the detection of ricin in milk and serum, three formats of sandwich enzyme-linked immunosorbent assay (ELISA) were compared utilizing the same pair of antibodies. The ELISA using a biotinylated primary detection antibody and streptavidin-linked horseradish peroxidase (HRP) system was shown to be the most sensitive assay with limits of detection (LOD) of 25 pg/mL in phosphate buffered saline (PBS), 50 pg/mL in non-fat milk, mouse serum, and 100 pg/mL in whole milk. The second best was the ELISA using a streptavidinylated primary detection antibody and biotin–HRP system, the LOD for ricin was 100 pg/mL in PBS and milk, and 1 ng/mL in serum. The ELISA using a non-tagged primary detection antibody and HRP-labeled secondary antibody performed the least sensitive among all and the LOD was 1 ng/mL in all matrices tested. Compared with the direct ELISAs (without using the capture antibody), the sandwich ELISAs were 50–500-fold more sensitive in PBS buffer. Estimation of the accuracy of these immunoassays using the Coefficient of Variability (CV) showed that the most sensitive ELISA format also had the lowest inter- (4.28%) and intra-assay CV (2.15%) although the inter- and intra-assay CV for the other two ELISAs were less than 10% and 6%, respectively, well below the maximum acceptable level. To conclude, the ELISA using a biotinylated primary detection antibody and streptavidin–HRP system is the best assay for detection of ricin in PBS, milk and serum among three ELISA formats tested and the application of this assay will be valuable to food safety research and clinical diagnosis.