Simulation study of antimicrobial agents : theories, properties, and interactions with biomembrane

The competition between the evolution of drug-resistant bacteria and the development of new antibiotics has never stopped since the first application of antimicrobial agent. Considering the long development cycle for a new agent and the invalidation of more and more old agents due to resistance, sustained investments are necessary to keep drug design leading in this competition. MD simulation as a useful tool for rational drug design reveals the drug-target interactions at an atom level, offering great assistance in understanding the killing mechanisms of antibiotics. Using MD simulations in combination with enhanced sampling methods, two novel antimicrobial agents and their self-assembly, binding modes with the targets, and conformational changes in solution were studied in this thesis. Furthermore, for a better application of one of the enhanced sampling methods, PT-WTE, we also came up with an algorithm to generate the temperature sequences and corresponding bias factors. The first study case, CSM5-K5, which is synthesized by grafting polylysine chains onto a chitosan backbone, was experimentally proved to self-assemble in aqueous medium and function as a membrane active antibiotic. According to our MetaD simulation results, the –NHC(=O)•••–NHC(=O) H-bonds between chitosan chains are strong enough to counteract the electrostatic repulsion introduced by the polylysine, and play a critical role in maintaining the dimerization of CSM5-K5. And as proved by the later conventional MD simulation, this dimerization also accelerates the lipid rearrangement process upon binding to the bacterial membrane, which will eventually cause membrane defects and lead to the cell death. CSM5-K5 presents a successful example of drug design by employing the polylysine chains to ensure the selectivity of bacteria over the mammalian cell and the antimicrobial activity, while the chitosan chains to supply the aggregation mechanism to further enhance the effectiveness. In the second case, the binding phase space of a recently discovered antibiotic, TXB, and its target, LII, is investigated. Totally four binding modes were identified from this highly flexible binary complex ensemble generated by PTMetaD-WTE. Three of them involve the ring motif of TXB and have relatively higher binding affinities, indicating the importance of the ring motif of TXB in LII recognition. These three binding modes where all the four NH groups on the ring form H-bonds with the negatively charged residues of LII, were also demonstrated indirectly by a crystallographic structure of a TXB analogue binding to a chloride anion. Among the four amino acids on the ring motif, the chiral D-Thr residue is critical for the antimicrobial activity. We then performed MetaD and PT-WTE simulations to study the necessity of this chiral D-Thr in the binding with LII. Our results suggest that different chirality leads to different NH orientation of Thr with respect to the ring plane. Only in the ring motif with D-Thr, a favored binding cavity is achievable with all the four NH groups pointing to the same side of the ring plane. To improve the application of PT-WTE, we came up with a scheme to generate the temperature sequence and the corresponding bias factor optimally. The number of replicas for a complete coverage of a specific temperature range is adjustable in this scheme while keeping the AAP between neighboring replica-pairs unchanged. Two series of PT-WTE simulations were tested with the number of replicas as 16 and 8, and the corresponding bias factors as 37.9 and 174.2, respectively. However, the bias factor of 174.2 is oversized for a better application of PT-WTE.Doctor of Philosoph