We are designing small molecule mimetics of insulin, based on knowledge of the important residues of insulin identified through site-directed mutagenesis and interspecies differences. The structure of the insulin receptor is not known, so we cannot use the most direct method to design a ligand, i.e. using the receptor surface as a template. Instead we must infer the receptor surface by examining a variety of insulin mutants whose affinity for the receptor is known. We have used molecular modeling with graphics to design insulin-like peptides. The peptides have been synthesized their biological activity assayed and their structures analysed using NMR. So far, we have tested three structures, all of which are a mixture of D- and L- amino acids. None of the peptides has been shown to have biological activity. The most recent design has 31 residues and evidence from NMR shows that the peptide has a persistent structure in solution. This project is significant for both drug design, and, on a more fundamental level, our ability to design proteins.
Showing the most recent 10 out of 508 publications
