Success in generating catalytic antibodies as enzyme mimics lies in the strategic design of the transition-state analog (TSA) for the reaction of interest, and careful development of screening processes for the selection of antibodies that are catalysts. Typically, the choice of TSA structure is straightforward, and the criterion for selection in screening is often binding of the TSA to the antibody in a microtiter-plate assay. This article emphasizes the problems of TSA design in complex reactions and the importance of selecting antibodies on the basis of catalysis as well as binding to the TSA. The target reaction is the derivatization of primary amines with naphthalene-2,3-dicarboxaldehyde (NDA) in the presence of cyanide ion. The desired outcome is selective catalysis of formation of the fluorescent derivative in preference to nonfluorescent side-products. In the study, TSA design was directed toward the reaction branch leading to the fluorescent product. Here, we describe a microtiter plate-based assay that is capable of detecting antibodies showing catalytic activity at an early stage. Of the antibodies selected, 36% showed no appreciable binding to any of the substrates tested, but did show catalytic activity in derivatizing one or more of the amino acids screened. In contrast, only two out of 77 clones that showed binding did not show catalysis. Thus, in this complex system, observation of binding is a good predictor of the presence of catalytic activity, and failure to observe binding is a poor predictor of the absence of catalytic activity.