Enhanced Sampling Techniques Reveal Key Information about Membrane Active Peptides

Membrane active peptides (MAPs) show promise in terms of future drug development. Whether it be adapting macrocycles for use in targeting protein-protein interactions or adopting cell-penetrating peptides (CPPs) to cause lysis in target cells, the field is burgeoning with possibilities for designer drugs. The following work encompasses the exploration of three such peptides. The pH-Low Insertion Peptide (pHLIP) is a membrane-active peptide that spontaneously folds into a transmembrane $\alpha$-helix upon acidification. This activity enables pHLIP to potentially act as a vector for drugs related to diseases characterized by acidosis such as cancer or heart ischemia. First, we explored the conformational space sampled by pHLIP while in bulk solution via constant pH molecular dynamics (MD) simulations. It was determined that pHLIP\u27s acidic residues are similar to single-residue-in-solution values and the P20G maintains a higher helicity in solution than wt-pHLIP. The next study was on the 17 N-terminal residues of the huntingtin (htt) protein (Nt17). Nt17 is essential for the patheoogenesis of Huntington\u27s Disease through its role in both htt aggregation and membrane association. We investigated Nt17 and its association with three model membranes with a consistent headgroup and tails with varying degrees of unsaturation and length. We found no correlation between the effect of lipid vesicles on aggregation and the degree of htt-lipid complexes formed, supporting that the properties of the membrane have direct influence on the aggregation mechanism. We also determined that Nt17-membrane association is regulated by complimentarily-sized hydrophobic residues in Nt17 and defects in the lipid bilayer. Finally, we developed a high-throughput assay for determining the permeability cyclic peptides. Targeting protein-protein interactions with traditional small-molecule drugs can be challenging when the binding pocket is too large. However, cyclic peptides are the key to targeting these interactions: passive membrane permeability and more consistent structure to prevent off targeting. Using a library of peptides with known permeability, we performed Gaussian accelerated molecular dynamics (GaMD) simulations on roughly 200 peptides of the library in octanol and water to estimate their permeability. Initially, we did not directly replicate the permeability from the experimental results, however, we did replicate the trend that correlates N-methylation and permeability. After using PCA to determine the peptides with biggest difference in populations between octanol and water, we more successfully reproduced the permeability data from experiment for those peptides