Three types of immunochemical tests were compared for assaying the storage globulins, glycinin and β-conglycinin, in 15 differently processed soyabean products. They included immunoenzymatic tests carried out either on microtitre plates (Elisa) or on nitrocellulose membranes (dot blot), and the passive haemagglutination inhibition test (PHI). These assays also differed in their detection systems: enzymo-colorimetric activity measured spectrophotometrically at 405 nm (Elisa) or detected visually on membranes (dot blot), and red blood cell agglutination determined spectrophotometrically at 540 nm (PHI). The hyperimmune sera used in these assays were produced in rabbits against native or SDS-denatured glycinin and native β-conglycinin. The concentrations of antigens were determined in borate buffer extracts of the soyabean products, and were expressed as mg/g of crude protein. The treatments applied to the soyabean products included defatting, steam-heating with or without previous water extraction of oligosaccharides, hot aqueous ethanol denaturation, and proteolysis. Nine of them were introduced into milk replacers which were given to preruminant calves at a level of intake of 58 to 60 g dry matter per kg of metabolic weight per day, for approximately 100 days. The soyabean products provided 59 to 72% of the dietary nitrogen, the remainder coming from skim milk powder or sweet whey supplemented with synthetic amino acids. The apparent digestibility of nitrogen, the live weight gain over the entire experimental periods and the plasma anti-soya antibody titres were measured. The data obtained by the different tests for each globulin were highly and positively correlated to one another, even though the dot blot tests for the native globulins provided significantly higher results (P < 0.05). The apparent digestibility of nitrogen, measured for the diet or calculated for the soya products, was highly and positively correlated with the concentrations of immunoreactive globulins, for the vast majority of the compared immunochemical assays. Conversely, the plasma anti-soya antibody titres and the levels of immunoreactive globulins varied in the same way, irrespective of the assay format. In contrast, the relationships between live weight gain and the levels of immunoreactive globulins in the products were rarely significant. Finally, the relationships between in vivo and in vitro parameters, when significant, were the strongest with β-conglycinin. It is concluded from these results that all these assays employing hyperimmune sera raised against the native soyabean globulins for determining the levels of antigenic proteins in soyabean products intended to be incorporated into milk replacers are useful because they all could predict the apparent digestibility of dietary or soyabean nitrogen, and the immunogenicity of soyabean in the preruminant calf. Elisa remains the method of choice in the laboratory conditions, while the PHI and the dot blot assays could be used when the technical environment is not appropriate for Elisa.
