Abstract A multianalyte immunosensor array can be im- plemented by immobilization of different haptens in dis- tinct areas of a single cavity or flow cell. In this case a mixture of different antibodies for different analytes is used in an indirect ELISA-format. The selection of the right hapten structures is very important to build up an ar- ray successfully. A system of independent hapten/anti- body combinations is needed, with one immobilized hap- ten (coating antigen) reacting only with one antibody. If more than one antibody binds to a coating antigen no ideal calibration curves are obtained. This phenomenon is known as shared-reactivity and can lead to double- sigmoidal curves. To use monoclonal antibodies to 2,4,6- trinitrotoluene (TNT) and 2,4-dichlorophenoxyacetic acid (2,4-D), two different haptens had to be found, one only reacting with the TNT-antibody, the other only binding to the 2,4-D-antibody. 2,4-Dichlorophenoxybutyric acid was used for the 2,4-D antibody and 2,4,6-trinitrophenyl-8- aminooctanoic acid for the TNT antibody. Although 4-ni- trotoluene, 2,4-dinitrotoluene and 4-amino-2,6-dinitro- toluene showed only very low cross-reactivities to the 2,4-D antibody the corresponding haptens 4-nitrophenyl- acetic acid, 2,4-dinitrophenyl-6-aminohexanonic acid, and 4-amino-2,6-dinitrotoluyl-(N)-glutarate are useful coating antigens for this antibody. The structure of the coating antigens had no significant influence on the mid- points (IC50) of the test for 2,4-D and even haptens with very low cross-reactivities could be used. With all haptens a test midpoint of about 0.2 μg/L for 2,4-D was achieved. For the direct assay format with immobilized antibodies the same test midpoint of 0.2 μg/L for 2,4-D was ob- tained. As a conclusion, the selectivity of a monoclonal antibody should not be influenced by the used tracer or coating antigen as well. It could be shown that the affinity constants of an antibody to the analytes are the main sen- sitivity and selectivity determining parameters for com- petitive immunoassays. A two-dimensional microtiter plate array was used to determine the analytes 2,4-D and TNT in parallel with a mixture of antibodies. 1 Introduction Immunoassays, for instance the enzyme-linked im- munosorbent assay (ELISA), are very widely used in medical and pharmaceutical analysis. In the environmen- tal field their use is mostly restricted to a screening tool [1]. Immunoassays have several advantages compared with conventional analysis like the low costs per sample, the high sensitivity and high sample throughput. On the other hand, their use is limited by the high development costs and the fact that the result of the measurement is only a single value of a substance equivalent [2]. Due to the cross-reactivities of antibodies, quantitative analysis is not possible without prior information about the composi- tion of the “unknown” sample [3]. Many attempts have been made to overcome the cross-reactivity problem. One way is to design antibodies by protein modeling [4] that only should bind one single compound. This may be pos- sible with substances, where only very few cross-reacting structures exist, but will probably never be achieved with substance classes like triazines, where many similar com- pounds occur. A more promising way is to build up an ar- ray structure with many different antibodies to make use of the specific cross-reactivity pattern of each antibody [5]. By this way other disadvantages of immunoassays can be overcome, too. With different antibodies for differ- ent analytes in an array structure, different analytes can be measured in parallel at the same time. Most known ap- proaches make multianalyte detection possible by com- bining two or more individual sensors. This kind of paral- lelization reaches its physical and economical limits with the increasing number of analytes or antibodies [6]. In this paper an approach for the multianalyte detection by enzyme immunoassay, based on an indirect ELISA- Andreas J. Schuetz · Michael Winklmair · Michael G. Weller · Reinhard Niessner Selection of hapten structures for indirect immunosensor arrays Fresenius J Anal Chem (1999) 363 : 625–631 © Springer-Verlag 1999 Received: 29 July 1998 / Revised: 21 October 1998 / Accepted: 10 November 1998 ORIGINAL PAPER Presented at the Analytica Conference, Munich, 21.–24.04.1998 A. J. Schuetz · M. Winklmair · M. G. Weller () · R. Niessner Institute of Hydrochemistry, Technical University of Munich, Marchioninistrasse 17, D-81377 München, Germany