Hamsa Priya Mohana Sundaram

Water structure in aqueous osmolyte solutions, deduced from the slight alteration in the water-water radial distribution function, the decrease in water-water hydrogen bonding, and tetrahedral ordering based only on the orientation of nearest water molecules derived from the molecular dynamics simulations, appears to have been perturbed. A careful analysis, however, reveals that the hydrogen bonding and the tetrahedral ordering around a water molecule in binary solutions remain intact as in neat water when the contribution of osmolyte-water interactions is appropriately incorporated. Furthermore, the distribution of the water binding energies and the water excess chemical potential of solvation in solutions are also pretty much the same as in neat water. Osmolytes are, therefore, well integrated into the hydrogen-bond network of water. Indeed, osmolytes tend to preferentially hydrogen bond with water molecules and their interaction energies are strongly correlated to their hydrogen-bonding capability. The graph network analysis, further, illustrates that osmolytes act as hubs in the percolated hydrogen-bond network of solutions. The degree of hydrogen bonding of osmolytes predominantly determines all of the network properties. Osmolytes like ethanol that form fewer hydrogen bonds than a water molecule disrupt the water hydrogen-bond network, while other osmolytes that form more hydrogen bonds effectively increase the connectivity among water molecules. Our observation of minimal variation in the local structure and the vitality of osmolyte-water hydrogen bonds on the solution network properties clearly imply that the direct interaction between protein and osmolytes is solely responsible for the protein stability. Further, the relevance of hydrogen bonds on solution properties suggests that the hydrogen-bonding interaction among protein, water, and osmolyte could be the key determinant of the protein conformation in solutions.

Short surfactant-like amphiphilic peptide, A3K, resembling a surfactant with a hydrophobic tail (A3) and a polar headgroup (K), is experimentally determined to form a membrane. Although the peptides are known to exist as β-strands, the exact packing architecture stabilizing the membrane is unknown. Earlier simulation studies have reported successful packing configurations through trial and error. In this work, we present a systematic protocol to identify the best peptide configurations for different packing patterns. The influence of stacking peptides in square and hexagonal packing geometry with the neighboring peptides in parallel and antiparallel orientations was explored. The best peptide configurations were determined from the free energy of bringing 2-4 peptides together as a bundle that can be stacked into a membrane. The stability of the assembled bilayer membrane was further investigated through molecular dynamics simulation. The role of peptide tilting, interpeptide distance, the nature and the extent of interactions, and the conformational degrees of freedom on the stability of the membrane is discussed. The consistency with the experimental findings suggests hexagonal antiparallel as the most relevant molecular architecture.

We find, through atomistic molecular dynamics simulation of native cyclodextrins (CDs) in water, that although the outer surface of a CD appears like a truncated cone, the inner cavity resembles a conical hourglass because of the inward protrusion of the glycosidic oxygens. Furthermore, the conformations of the constituent α-glucose molecules are found to differ significantly from a free monomeric α-glucose molecule. This is the first computational study that maps the conformational change to the preferential hydrogen bond donating capacity of one of the secondary hydroxyl groups of CD, in consensus with an NMR experiment. We have developed a simple and novel geometry-based technique to identify water molecules occupying the nonspherical CD cavity, and the computed water occupancies are in close agreement with the experimental and density functional theory studies. Our analysis reveals that a water molecule in CD cavity loses out about two hydrogen bonds and remains energetically frustrated but possesses higher orientational degree of freedom compared to bulk water. In the context of CD-drug complexation, these imply a nonclassical, that is, enthalpically driven hydrophobic association of a drug in CD cavity.

Alanine-rich tetrapeptides like A3K dominantly exist as polyproline II helices in dilute aqueous solutions. However, during self-assembly, based on the free energy calculation in implicit solvent for various peptide conformations, only the peptides in the β-strand conformation can be packed closely. This necessitates the conformational transition to the β-strand commonly observed during peptide self-assembly such as in amyloid fibril formation. In fact, the closest interpeptide distance of 4.8 Å is consistent with the interstrand distance determined from the X-ray diffraction pattern of many amyloid fibrils. The position of free energy minimum obtained from implicit solvent calculation matches exactly with the explicit solvent simulation through umbrella sampling when the peptide conformations are restrained, demonstrating the applicability of the former for rapid screening of peptide configurations favorable for self-assembly. The barrier in the free energy profile in the presence of water arises out of the entropic restriction on the interstitial water molecules while satisfying the hydrogen bonding of both the peptides by forming water mediated hydrogen bond bridge. Further, the high energy barrier observed for the β-strand suggests that peptides initially tend to self-assemble in the polyproline II structure to mitigate the desolvation energy cost; the transition to the β-strand would happen only in the later stage after crossing the barrier. The umbrella sampling simulations with peptides allowed to change conformations, relative to each other, confirm the dynamic conformational transition during the course of the self-assembly supporting the “dock and lock” mechanism suggested for amyloid fibrillar growth.