Head-to-tail peptide cyclization: new directions and application to urotensin II and Nrf2

Abstract

Peptides have recently re-gained interest as therapeutic candidates but their development remains confronted with several limitations including low bioavailability. Backbone head-to-tail cyclization is one effective strategy of peptide-based drug design to stabilize the conformation of bioactive peptides while preserving peptide properties in terms of low toxicity, binding affinity, target selectivity and preventing enzymatic degradation. However, very little is known about the sequence-structure relationship requirements of designing linkers for peptide cyclization in a rational manner. Recently, we have shown that large scale data-mining of available protein structures can lead to the precise identification of protein loop conformations, even from remote structural classes. Here, we transpose this approach to head-to-tail peptide cyclization. Firstly we show that given a linker sequence and the conformation of the linear peptide, it is possible to accurately predict the cyclized peptide conformation improving by over 1 A over pre-existing protocols. Secondly, and more importantly, we show that is is possible to elaborate on the information inferred from protein structures to propose effective candidate linker sequences constrained by length and amino acid composition, providing the first framework for the rational peptide head-to-tail cyclization. As functional validation, we apply it to the design of a head-to-tail cyclized derivative of urotensin II, an 11-residue long peptide which exerts a broad array of biologic activities, making its cognate receptor a valuable and innovative therapeutic or diagnostic target. We propose a three amino acid candidate linker, leading to the first synthesized 14-residue long cyclic UII analogue with excellent retention of in vitro activity.

Highfold: accurately predicting cyclic peptide monomers and complexes with AlphaFold

“…Traditional experiments can deliver excellent observation accuracy, but it often demands extensive time and effort [4][5][6] . Although there have been recent advances in approaches of Computer-Aided Drug Design (CADD) methodologies [7][8][9][10][11][12][13][14] , the computational prediction of cyclic peptide monomer structure still is mildly unsatisfactory. Those methods typically utilize physics-based energy functions and Machine Learning (ML) algorithms such as Monte Carlo sampling methods, indicating that they require an immense amount of time to sample for finding the lowest energy structures.…”