Peptide and nonpeptide lead discovery using robotically synthesized soluble libraries.

The method of combinatorial synthesis of peptide and nonpeptide libraries on solid phase is analyzed and the automation of the mix and divide key step described. A set of amino acids leading to a high molecular diversity is proposed as well as a number of scaffolds for the preparation of variable polyamide libraries. Adequacy of the resin bead quantities to library size and to the ratio of the synthesized peptide types is discussed. Examples of the use of capillary electrophoresis and of spectroscopic methods (MS, MS/MS, and NMR) for the analysis of the library content are given. The iterative deconvolution SURF (synthetic unrandomization of randomized fragments) is compared with positional scanning and the success of coupling of mixtures evaluated. It is concluded that extension of the original mix and divide method and the SURF deconvolution (as proposed by Houghten et al. Nature (London), 354: 84-86 1991) to nonpeptide libraries affords new leads that can be optimized towards useful therapeutics.