Limitations of the enzyme‐linked immunosorbent assay for routine determination of legume inoculant quality

Abstract

Peat inoculants containing strains of either Rhizobium or Bradyrhizobium spp. were used to determine correlations between cell numbers and A405 values obtained with double antibody sandwich enzyme‐linked immunosorbent assay (DAS ELISA) and indirect ELISA. ELISA values of inoculants containing strains of Rhizobium were weak and non‐specific; with Bradyrhizobium spp. strains, readings were higher and cross‐reactions negligible when heated inoculant suspensions were allowed to stand for 3 h before ELISA determinations were made. With soybean inoculant, correlation coefficients of r = 0.93 and 0.83 were obtained with DAS and indirect ELISA, respectively. A linear curve relating log cell numbers to A405 values was used to determine the reliability of DAS ELISA values obtained over 2 years with tests on commercially produced soybean inoculants. In the range 5 times 108‐ca 3 times 109 cells/g inoculant, DAS ELISA estimates closely followed plate counts but no significant correlation was found when inoculants contained >ca 3 times 109 cells/g. With a minimum requirement of 1 times 109 cells/g inoculant, discrepancies between DAS ELISA estimates and plate counts obtained with inoculants produced with gamma‐irradiated peat would have resulted in the erroneous rejection or acceptance of 14.5% of all inoculants tested, based on DAS ELISA estimates. With inoculants produced with steam‐sterilized peat, which was unfavourable for survival of strain WB1, 70.0% of the inoculants rejected because of low plate counts would have been acceptable on the basis of DAS ELISA estimates.