P01-12

Kinetic Analysis of Membrane Permeation Process of Cyclic Peptides Using Markov State Models with Molecular Dynamics Simulations

Kei TERAKURA *1Masatake SUGITA1, 2Keisuke YANAGISAWA1, 2Yutaka AKIYAMA1, 2

1Department of Computer Science, School of Computing, Institute of Science Tokyo
2Middle-Molecule IT-based Drug Discovery Laboratory (MIDL), Institute of Science Tokyo
( * E-mail: terakura@bi.c.titech.ac.jp )

Cyclic peptides are gaining attention as novel pharmaceuticals because they can target intracellular protein-protein interactions, which have been difficult to address with small molecules or antibody drugs. However, the poor membrane permeability of cyclic peptides is a significant bottleneck in drug development. To design membrane-traversing cyclic peptides, understanding the membrane permeation mechanisms of cyclic peptides, especially permeation by passive diffusion, is crucial for their design. Understanding the mechanisms is challenging due to the diversity of cyclic peptide conformations and the long timescale of membrane permeation. To address these issues, research has been conducted using molecular dynamics simulations in addition to experimental studies. These simulation-based studies have yielded a certain degree of predictive accuracy and insights. Nevertheless, the discussions have only been based on one-dimensional or two-dimensional reaction coordinates, such as the overall peptide position and angle relative to the membrane, and local changes in three-dimensional structures during the membrane permeation process have not been considered.
In this study, we analyzed the membrane permeation process of cyclic peptides using a Markov State Model (MSM) based on multidimensional reaction coordinates, including position, orientation, and detailed conformation of the peptide. MSM is a powerful tool that quantifies long-timescale behavior using multiple short simulations and suggests key features of functional dynamics. We applied MSM to understanding the membrane permeation mechanism of cyclic peptides. We conducted approximately 1,000 unbiased simulations of 100 ns each across more than 15 peptides. The initial conformations were selected from previous research data [1]. The eigenvectors obtained from the MSM indicated several motions that represent potential bottlenecks in the membrane permeation process. However, some challenges still remain, such as converging timescales.

[1] Sugita, M.; Fujie, T.; Yanagisawa, K.; Ohue, M.; Akiyama, Y. Lipid Composition Is Critical for Accurate Membrane Permeability Prediction of Cyclic Peptides by Molecular Dynamics Simulations. J. Chem. Inf. Model. 2022, 62, 4549–4560, DOI: 10.1021/acs.jcim.2c00931