Affinity ligand selection from a library of small molecules: assay development, screening, and application

A facile and cost-effective process for screening synthetic libraries for an affinity ligand is described. A high throughput 96-well plate filtration method was designed to screen both discrete compounds and mixtures of compounds attached to a solid support. Human serum albumin (HSA) was used as a target protein to demonstrate the proof of concept. Detection and quantitation by fluorescence was accomplished with the use of fluorescamine to conjugate the protein in the filtrate. It is found that mixtures demonstrating low average binding reflect an overall lower hit rate of the components, whereas deconvolution of mixtures with high protein binding consistently provides a high hit rate. This differs from many of the previous experiences screening solid-phase mixtures in which high false positive rates are noted to occur. A total of 100K compounds were tested: 25K as discrete samples and 75K as mixtures. An overall hit rate of 8% was observed. Secondary screening of compounds measured specificity, recovery, and dynamic binding capacity. The effectiveness of the method is illustrated using an affinity column made with a representative lead compound. A similar purity was achieved in a single-step purification of HSA from serum as compared to that obtained by two steps of ion-exchange chromatography. The process for primary screening of a large number of compounds is simple, inexpensive, and applicable to any soluble target protein of known or unknown function from crude mixtures and may have additional utility as a generic chemical affinity tool for the functional characterization of novel proteins emerging from proteomics work.