(220f) A Biased Review about Computer Simulations of Protein Folding and Aggregation

Amyloid fibrils are ordered aggregates of misfolded peptides or proteins associated with type 2 diabetes mellitus, Alzheimerâ??s disease, Parkinsonâ??s disease, etc. Amyloid diseases are apparently unconnected; however, intermolecular secondary structure (mostly β-sheets) is present in all amyloid aggregates.

Assumptions about the dominance of a particular type of force are common in computer simulations of protein aggregation â??most of the times these simplifications are based on experimental observations. However, sometimes the simplifications are rooted on the knowledge of other systems or simply on accepted rules of thumb. Two common simplifications are to overemphasize electrostatic interactions (at one end) or hydrophobic interactions (at the other end). Even though the promoters of each approach make good points in defending them, they (for the most part) ignore the role that solvent may have in stabilizing the structure. Most simulations (are performed with very few peptides, a single protein, a single monomer added to an existing fibril. The fact that aggregation happens at multiple time scales adds complexity to the theoretical work. Prof. Hallâ??s group has used a simplified model for peptides consisting of a few united atoms bearing a square-well potential that represents either hydrogen bonding or hydrophobic forces to study the self-assembly of peptides of alanine and glutamic acid. It is fascinating that such a simple model predicts the correct assembly, even after incorporating the solvent into an oversimplified potential of mean force. Another mechanism that has been explored is molecular crowding; this approach allows for the use of a peptide at infinite dilution in a crowding media. These studies support the idea of incorporating interfaces --we have already shown that interfaces act as an enrichment site in the study of peptide aggregation.

The few reports devoted to study the effect of surfaces on amyloid deposits formation are mostly dedicated to finding a link between the interaction of the deposits with cell surfaces and their toxicity. Concentration and misfolding of the peptides at the solid/liquid interface support the hypothesis that the surface serves as a catalyst. Recently, we have studied the fibrillation of a series of diblock peptides in the presence of poly(styrene) latex with various surface chemistries and of liposomes with different phospholipid compositions. It was concluded that fibrillation depends on the composition of the diblock and on the presence or absence of interfaces but no adsorption was observed. Interestingly, the replacing of an isoleucine block by a leucine one in diblock peptides of the same length inhibits fibrillation. The consequences of these findings on current modeling efforts as well as the development of a new model are discussed.