Development of a direct competitive microcystin immunoassay of broad specificity.

produced by several cyanobacterial species.1 The toxicological mechanism is mainly based on a strong inhibition of protein phosphatases, which may lead to severe—sometimes lethal— liver damage.2,3 In addition, microcystins are considered to be tumor promotors. Hence, the control of microcystins is critical for water quality. Recent publications have stated that about 90 microcystins and nodularins have already been characterized.4,5 Only a few microcystins and one nodularin are available as standard compounds. There is some concern that analytical methods for the detection of hepatotoxic peptides may overlook toxins and perhaps cause a significant number of false negative samples. This applies to instrumental techniques (e.g., HPLC, MS) as well as for immunochemical assays. Only the mouse bioassay is considered to give a relatively definite answer,6 however, this test for acute toxicity is relatively insensitive, and criticism has been expressed against these crude animal tests. Several immunoassays have been developed for the analysis of microcystins.7–16 Unfortunately, none of them have gained widespread application, which may be caused by insufficient sensitivity, unfavorable cross-reactivities and difficult availability. Most desirable would be an immunoassay of broad reactivity towards all microcystins,6 allowing a measurement of the sum concentration of these compounds, which would be a useful index for regulatory purposes. A known monoclonal antibody9 against microcystin-LR (MC-LR) was used in this study. The antibody (clone M8H5) had been generated by immunization with a BSA-microcystin-LR conjugate, prepared by coupling the carboxylic groups of microcystin-LR to the protein with a water-soluble carbodiimide (conjugation density MC-LR/BSA 1.3 mol/mol). Because MC-LR has two carboxylic groups available for conjugation (see Fig. 1), the immunogen is not structurally unambiguous. Most characteristics of M8H5 support the notion that the effective conjugation site was at amino acid (3), not at (6). If (6) would have been the conjugation site, one should expect a larger influence of the arginine at position (4). In addition, the loss of a methyl group at (3) seems to be of minor importance. Other clones isolated in parallel to M8H5 showed relatively similar cross-reactivity patterns.9 Although the antibody M8H5 showed high affinity to the antigen (k = 1.3 × 1010 L/mol), the crossreactivity pattern of M8H5 was not broad enough in the previously tested assay formats. The generation of new antibodies17 may solve such selectivity problems. As shown in many papers, the selectivity of antibodies can be considerably influenced by the orientation of the hapten during immunization. If broad reactivity is desired, the common substructure must be placed at the “outer” side, which is located opposite to the conjugation site to the protein carrier. This approach probably works better with large antigens/haptens, with extended common epitopes. In other cases, all parts of the analyte influence the binding by the antibody, and consequently the cross-reactivity pattern is uneven. However, it should be 1445 ANALYTICAL SCIENCES DECEMBER 2001, VOL. 17 2001 © The Japan Society for Analytical Chemistry