Synthesis and screening of an indexed motif-library containing non-proteinogenic amino acids

In an effort to increase the probability of finding novel peptides in resin-bound combinatorial libraries displaying affinity to various macromolecular targets, we increased the diversity of a solid-phase library considerably by synthesizing multiple structures on each bead - a motif-library - including 45 building blocks. The building blocks consist of L-aa, D-aa and eight hydrophobic non-proteinogenic alpha-amino acids. A library with the format O-Z0-1-O-Z0-1-O-XX-resin was synthesized giving the four motifs OOOXX, OZOOXX, OOZOXX, OZOZOXX corresponding to 364.500 different motifs (45(3) x 4 theoretical combinations). The positions O are defined amino acids while Z represents three mixtures pi, omega, phi, where pi is a mixture of polar and charged residues, omega is a mixture of aliphatic residues and phi is a mixture of aromatic residues. X represents a mixture of all 45 residues. The library was screened with the macromolecular target streptavidin which served as a model receptor. Binding peptides were sequenced by microsequencing. We included small amounts of norvaline and norleucine in the library, which served as index residues to be able to distinguish between LD-amino acids and other residues with the same retention time in the HPLC system. Beads that interact with the receptor were found, and the binding motifs that appeared had no homology to known binding motifs found in either L-aa or D-aa libraries, instead motifs with the non-proteinogenic residues L-phenylglycine, O-benzyl-L-hydroxyproline and O-benzyl-L-tyrosine dominated. The novel peptides inhibit binding of biotin to streptavidin but do not bind to avidin, and the affinity is higher than the peptides found in linear all L-aa peptide libraries.