De novo design of passively permeable cyclic peptides and incorporation of noncanonical amino acids to improve predicted binding affinity

Abstract

Cyclic peptides fill an intermediate niche between traditional small molecule therapeutics and larger biologics. In ideal cases they combine the passive permeability and oral availability of small molecules with the binding specificity of larger biologics. However, achieving both of these features in a single designed cyclic peptide has remained elusive, with the majority of clinically approved cyclic peptides being derived from natural products. In order to address this problem my doctoral research focuses on developing new methods for the de novo computational design of cyclic peptides to have improved permeability and binding properties. Collaborators and I designed over 70 cyclic peptides that have apparent passive membrane permeabilities > 1*10^-6 cm/s that are of a broader size and conformational range than have been previously reported. I also worked to expand computational methods for the incorporation of noncanonical amino acids at cyclic peptide binder interfaces to increase predicted binding affinity.