Antibacterial Peptides on Lipid Membranes: An Integrated Experimental and Computational Pipeline for Conformational Analysis and Mechanistic Determinant Identification

The growing morbidity and mortality associated with multidrug-resistant bacteria highlight the urgent need for alternatives to conventional antibiotics, whose effectiveness is increasingly compromised. Antimicrobial peptides (AMPs) represent promising therapeutic candidates due to their structural diversity and distinctive mechanisms of action. However, the development of stable and highly effective AMPs requires reliable strategies for predicting activity and understanding their molecular mechanisms. A detailed understanding of how AMPs interact with bacterial membranes and disrupt cellular integrity is essential for rational peptide design. Integrating computational approaches with experimental validation provides a powerful framework for optimizing peptide structure and predicting biological activity. Among computational methodologies, molecular dynamics (MD) simulations have become a valuable tool in drug discovery and molecular biology. These simulations capture biomolecular behavior at atomic resolution and fine temporal scales, enabling mechanistic insights that complement experimental observations. In this thesis, classical all-atom molecular dynamics (aaMD) simulations were employed as the primary computational pipeline to investigate peptide-membrane interactions and associated structural dynamics. The simulations conducted led to several key findings: (i) aaMD analysis provided a detailed atomic-level characterization of the tetra-branched antimicrobial peptide SET-M33 and its linear analogue Q33, supported by experimental evaluation of secondary structure and residue-level membrane interactions; (ii) application of the aaMD framework to a library of short synthetic antimicrobial peptides revealed residue-specific membrane binding patterns consistent with experimental data; (iii) clustering analysis and lipid interaction studies confirmed the presence of stabilizing π–π interactions, in agreement with previous NMR observations; and (iv) experimentally determined peptide-liposome binding affinities, measured using surface plasmon resonance (SPR) and quartz crystal microbalance (QCM), showed strong agreement with computational predictions obtained from MD simulations. Overall, this work establishes an integrated computational-experimental pipeline for achieving atomic-level insight into peptide-membrane interactions. The findings demonstrate the value of combining molecular simulations with biophysical techniques to validate computational measurements and guide peptide design. This approach supports the rational development of antimicrobial agents by identifying membrane-binding characteristics and structural features for antibacterial activity.