Simulations of antimicrobial peptides.

Antibiotic resistance is emerging as a serious threat to public health. Because of the relative ease with which bacteria develop resistance to traditional, receptor-specific antibiotics, molecules with different mechanisms of action are sought. Antimicrobial peptides (AMPs) have been studied for over twenty years as possible alternatives to traditional antibiotics due to their quick, strong, and non-specific antimicrobial action; however, their interactions with bacteria are still not fully understood, which prohibits the design of novel sequences for human therapy. We have used computer simulations to study one family of AMPs, protegrins. We have used detergent micelles and lipid bilayers as membrane mimics, as well as simulations in water, to elucidate regions of the sequences that are vital to activity and toxicity of these peptides. Through studies of protegrin-1 and its mutants in dodecylphosphocholine (DPC) and sodium dodecylsulfate (SDS) micelles as mammalian and bacterial membrane mimics, respectively, we examined which elements of the sequence are responsible for activity and toxicity. Results from SDS micelle simulations proved too unreliable for accurate prediction of activity. Results from DPC micelle simulations were more promising and resulted in the design of new sequences with low toxicity. Simulations of the presumed biologically relevant oligomeric structure of Protegrin-1 were carried out in POPE:POPG lipid bilayers to examine at the atomistic level, the mechanism by which protegrin-1 is able to kill bacteria. In the simulation we observe the rates at which ions and water are able to pass through the pore. The model proposed from the simulations allows for quick movement of ions through the pore, which results in destabilization of the transmembrane potential across the bacterial cell membrane. An additional method for predicting toxicity of protegrin mutants was developed based on a quantitative structure-activity relationship using properties from short simulations of protegrins in water was developed. When tested against new sequences, the model was able to predict toxicity with high accuracy. In the methods employed, the level of detail from the simulations enabled us to hone in on specific amino acids that play important roles in activity and toxicity and will allow for the design of new peptides.