We show that carbohydrates constitute an attractive source of readily available, stereochemically defined scaffolds for the facile attachment of side chains contained in genetically encoded and other amino acids. beta-D- and beta-L-glucose, L-mannose, and the 6-deoxy-6-N-analogue of beta-D-glucose have been employed to synthesize peptidomimetics that bind the SRIF receptors on AtT-20 mouse pituitary cells, five cloned human receptor subtypes (hSSTRs), and the NK-1 receptor. The affinity profile of various sugar-based ligands at the hSSTRs is compared with that of SRIF. Compound 19 bound hSSTR4 with a Ki of 100 nM. Subtle structural changes affect affinities. Evidence is presented that suggests that one compound (8) binds both the AtT-20 cell receptors and the five hSSTRs via a unique mode. The SARs of the glycosides at SRIF receptors differ markedly from those at the NK-1 receptor. For example a 4-benzyl substituent is important for SRIF receptor binding, but the 4-desbenzyl analogue 27 was highly potent (IC50 of 27 nM) at the NK-1 receptor. A new, nonbasic method for the synthesis of base-sensitive ethers from primary and secondary alcohols is also described.

The endogenous peptides somatostatin (SRIF) and substance P comprise very different structures. Although both bind G-protein-coupled receptors, the SRIF receptors (SSTR 1-5) recognize SRIF and related peptides which retain its beta-turn such as the potent cyclic hexapeptide SRIF agonist L-363,301 (6a), but not substance P. Conversely the NK-1 receptor binds substance P but not the above ligands. In contrast, the beta-D-glucosides 1 and 2, designed to mimic the beta-turn of 6a, bind both receptors. This observation led us to attempt the conversion of 6a into the first potent, selective cyclic hexapeptide ligand for the NK-1 receptor. To this end, we combined design with a minilibrary approach. The goal was accomplished with surprising ease, leading to the NK-1 receptor antagonist 9 (IC50 2.0 +/- 0.4 nM). This demonstrates that peptidomimetics, incorporating in this case the promiscuous beta-D-glucose scaffold, can provide valuable clues about receptor similarities not revealed by their endogenous ligands. In addition, this work suggests that the use of libraries and rational design need not be mutually exclusive approaches to lead discovery.

No abstract

No abstract