PROBING THE INTERACTIONS OF MP1 PEPTIDE WITH MODEL MEMBRANES USING MULTISCALE MOLECULAR DYNAMICS

Bioactive peptides play a crucial role in several biological processes, including host defense against microorganisms. Their mechanism of action typically involves interactions with cell membranes, disturbing lipid packing and inducing the formation of pores or defects, leading to leakages or cell death. Polybia-MP1 (MP1) exhibits broad-spectrum bactericidal activity against both Gram-positive and Gram-negative bacteria, while remaining non-hemolytic and non-cytotoxic to eukaryotic cells. The MP1 peptide stands out for its selective action against bacteria and tumor cells, targeting anionic phospholipids like phosphatidylglycerol (PG) and phosphatidylserine (PS). Although its macroscopic effects are known, the molecular details of its collective mechanism, specifically regarding pore formation and the impact of membrane complexity, remain to be fully elucidated. In this work, we employ a multiscale computational strategy to investigate the collective behavior of MP1 and its analogues. Preliminary All-Atom (AA) molecular dynamics simulations characterize the initial adsorption of individual peptides, identifying local perturbations such as changes in lipid order parameters and bilayer curvature. To overcome the temporal and spatial limitations of the atomic scale, we transition to the Coarse-Grained (CG) framework. This approach allows us to simulate systems with multiple peptides to observe concentration-dependent events, including peptide aggregation and the possible formation of transmembrane pores. In the CG approach, we first benchmark the model by investigating the interactions of MP1 with simple lipid bilayer compositions, such as pure phosphatidylcholine (PC) and phosphatidylethanolamine (PE) bilayers. These systems were selected to allow comparison with existing experimental data on MP1–membrane interactions, such as the subtle shifts in the phase transition temperature (Tm) of PE bilayers. By integrating AA and CG perspectives, including backmapping strategies, we aim to identify the structural features that optimize lytic action and selectivity.

 This work was supported by CNPq (404205/2024-0, 174388/2024-1 and 201531/2025-9) and FAPESP (2025/23025-6 and 2022/07231-7). Computational resources were provided by National Laboratory for Scientific Computing (LNCC) through the Santos Dumont supercomputer. We also acknowledge the support of the PSMN (Pôle Scientifique de Modélisation Numérique) and Centre Blaise Pascal’s IT test platform at ENS de Lyon (Lyon, France) for the computer facilities. The platform operates the SIDUS solution developed by Emmanuel Quemener.