Cyclic peptides can achieve exquisite biochemical potency and specificity against challenging targets such as protein-protein interactions (PPIs). Although the size and polarity of most cyclic peptides fail to meet Lipinski's Rule of 5 for predicting drug-likeness, a growing number of cyclic peptides have been described that exhibit the ADME properties of small molecule drugs, including high passive cell permeability and oral bioavailability. These exceptional cases, which include natural products such as cyclosporine A (CSA) as well as a variety of model systems developed by our group, have generated enthusiasm for the idea that macrocycles may provide a ?middle way? between small molecules and biologics in the pursuit of challenging intracellular targets. However, achieving drug-like permeability in cyclic peptides is far from straightforward. Simply removing the C- and N-termini alone is rarely sufficient for achieving therapeutically relevant cell permeability. Other factors combine to determine the properties of cyclic peptides, and my group has led the effort to elucidate principles that govern those properties. This proposal aims to 1) identify novel, lariat and stapled peptides that exhibit high passive membrane permeability; 2) develop selection strategies for filtering DNA-encoded libraries of cyclic peptides based on the net polarity of the pendant macrocycles; and 3) use NMR and computational methods to study the detailed mechanisms of permeability across model lipid bilayers.
In Aim 1, we will synthesize mass-encoded libraries based on lariat and stapled peptide designs, and use methods developed in my group to evaluate their permeabilities en masse. The results will provide insights into structure-permeability relationships in this chemical space, as well as providing raw materials for the synthesis of libraries aimed at biochemical target-based screening.
In Aim 2, we will begin by synthesizing a series of DNA-tagged cyclic peptide test systems in which the permeabilities of the pendant macrocycles, which differ only by stereochemistry at 2 positions, span nearly two log units. We will test a variety of separation schemes, some of which are known to separate nucleic acids based on the polarity of covalently attached small molecules. We will then synthesize a diverse library of ~108 lariat peptides and fractionate the library based on the intrinsic polarity of the attached peptides. Deep sequencing of the pre- and post-selection libraries will illuminate the ?permeability landscape? in cyclic peptides in the 7-mer to 11-mer size range with unprecedented scope and breadth.
In Aim 3, we will use 1H and 19F 2-D NMR techniques, combined with advanced molecular dynamics simulations performed by our collaborator, Prof. Sereina Riniker (ETH), to study the detailed mechanisms underlying passive membrane permeability in cyclic peptides. We will synthesize fluorinated derivatives of CSA and other model systems and study their transport kinetics across synthetic liposomes, and compare observations with current theoretical models.
We have designed a class of molecules, inspired by natural products, that maximize molecular size and complexity while maintaining ?drug-like? cell permeability, and even oral bioavailability. While we have uncovered important general principles, such as the importance of intramolecular hydrogen bonding, the detailed mechanisms and the interplay of factors that influence cell permeability in large cyclic peptides remain poorly understood. This proposal seeks to refine our ability to predict and control cell permeability in cyclic peptides toward the design of novel chemical matter that defies conventional wisdom based on classical predictors of ?drug-likeness?.