Advancing the Development of Macrocyclic Peptide Therapeutics Using High-Throughput Mass Spectrometry Approaches

Abstract

Cyclic peptides are poised to revolutionize the therapeutic market as designed protein modulators. These peptides are characterized by a circular sequence of amino acids, which creates unique structural properties. Their potential as drugs is supported by existing natural products and their synthetic analogues, like vancomycin and cyclosporine2–4. The proven success of these therapeutics has stimulated interest in the discovery of new cyclic peptide drugs using advanced methods. Historically, the discovery of cyclic peptide therapeutics has relied on exploring natural product analogues and random high throughput screening2,5,6. However, these methods can be improved to test a more diverse set of cyclic peptide structures and protein targets. Recent advancements in computational protein prediction and design have made in silico screening a powerful tool for drug discovery, boosted by improvements in both physics-based and deep learning approaches. Integrating computational tools with enhanced drug discovery techniques allows for the identification of novel cyclic peptide therapeutics7,8. This strategy represents a shift from random screening to a focused approach that can uncover new structures and interactions.High-throughput screening techniques offer an efficient way to evaluate thousands of compounds for target interaction simultaneously. To match the speed of computationally identified peptide candidates, high-throughput screening is essential for cyclic peptides. One such technique, affinity selection mass spectrometry (ASMS), leverages the affinity between protein and peptide to isolate the peptide from a pool of binding candidates9. The isolated peptides are identified by characteristic fragmentation inside the mass spectrometer. However, for cyclic peptides, this characteristic fragmentation can be complex and difficult to interpret. To improve interpretation, we expanded mass spectrometry data collection with multistage MSn fragmentation and developed computational sequencing to read cyclic peptide fragmentation. This approach enables the computational assignment of unknown cyclic peptide peptide sequences, unlocking the ability to test cyclic peptides for target interaction in high throughput. Furthermore, we adapted ASMS to screen against protein targets in complex environments, including membrane-bound proteins like G-protein coupled receptors (GPCRs). Using virus-like particles to anchor GPCRs in their membranes helped preserve their native environment and allowed affinity selection against properly folded proteins. Adapting ASMS in this way opens new possibilities for discovering cyclic peptide drugs against membrane-bound proteins—a longstanding barrier in the field. Beyond discovery, high-throughput mass spectrometry approaches like HDX-MS (Hydrogen-Deuterium Exchange Mass Spectrometry) allow for the study of structural dynamics in cyclic peptides. HDX-MS is a benchtop labeling experiment that, when paired with mass spectrometry, provides insights on conformational dynamics and solvent accessibility. Applying HDX-MS to cyclic peptide pools enabled us to gain structural insights into dozens of peptides simultaneously. The data generated revealed relationships between conformational flexibility and permeability, helping identify peptide scaffolds for designing membrane-traversing peptides. Overall, these high throughput approaches significantly enhance our capability to identify and assess cyclic peptides as therapeutic candidates.