Biological Screening of a Large Combinatorial Library

Abstract

Encoding technology has allowed for the creation of libraries of 50,000 or more low-molecular-weight compounds for biological testing. The current challenge is to properly and efficiently screen among these compounds for useful biological activities. In this example, actives against two related G-protein coupled receptors were sought from a combinatorial library of 56,000 members. The library was synthesized on solid phase using the split synthesis method and photochemically released for testing. At a screening concentration of 0.5-1 /LM, 86 unique structures were identified as active against one receptor and 24 were active against the other. Due to the random nature of compound sampling, five library equivalents or 280,000 beads were screened to ensure greater than 99% representation of library members. As a result, many actives appeared multiple times in the screen, verifying the encoding process. Further confirmation was obtained by resynthesis and testing of predicted active structures. A clear bias for specific "R" groups at each point of variation in the combinatorial library with little overlap between the two receptors has produced a clear structure-activity relationship on which to base further work.