Parallel synthesis and pharmacological screening of nonpeptide ligands of the neuropeptide Y receptor subtype Y5.

Several series of low-molecular-mass ligands of the neuropeptide receptor subtype Y5 were prepared using a mixed strategy of synthesis on solid phase and in solution. Collections of single compounds were obtained by an automated parallel procedure which allowed quick variation and investigation of the central spacer moiety, as well as of the aromatic substituents on each side. The strategy of parallel synthesis and screening of partially purified analogs helped to select rapidly potent and selective leads which displayed comparable antagonistic potency against neuropeptide Y activity on the Y5 receptor and better receptor selectivity than the original reference compounds.

From peptide libraries to optimized nonpeptide ligands in the search for S-farnesyltransferase inhibitors.

A complete 331,776-member library of tetrapeptides made of 24 amino acid building blocks was synthesized robotically on solid phase and subjected to a deconvolution based on the inhibitory potency of the sublibraries in a HPLC assay of the S-farnesyltransferase activity in vitro. One of the non-natural peptide and noncysteine-containing leads Nip-Trp-Phe-His (Nip=p-nitrophenyl-L-alanine) was optimized chemically to give a proteolytically stable pseudopeptide with a 200-fold potency compared with the original lead. The final compound was converted to the C-terminal ethyl ester: p-F-C6H4-CO(CH2)2-CO-Bta-D-Phepsi[CH2NH]His-OEt (Bta = benzothienyl-L-alanine) and shown to behave as a prodrug which was hydrolyzed back to the C-terminal acid following cell penetration. The method confirmed that several structurally original leads can be discovered in large libraries when deconvolution relies upon a highly specific assay and that these leads can be optimized by chemical modification to impart the final compound the desired pharmacological and pharmacokinetic properties.