Research projects
The research interests of our group are in theoretical approaches to problems in biological and chemical physics. Our recent efforts have focused on a) atomistic simulations of protein folding, b) computational studies of protein aggregation, c) theoretical models of chaperon-assisted folding, d) interactions of proteins with surfaces and f) models of folding kinetics

Simulations of protein folding in explicit solvent

Protein folding remains one of the major unsolved problems in molecular biology. Recent progress in computational studies of protein folding is associated with the following two developments:

  • advances in efficient sampling algorithms for computer simulations
  • improvements of atomistic force fields aided by recent experiments on small polypeptides
Map
We use replica exchange protocol to characterize the folding thermodynamics of a small b-hairpin peptide at the atomic resolution in explicit solvent. Our simulations provide important new insights into the folding mechanism of this peptide.

Structure of Alzheimer's disease peptides

A strong causative link exists between Alzheimer's disease (AD) and amyloid b peptides, Ab. Experiments on humans and animals show that macromolecular deposits of these peptides, amyloid plaques, are toxic to the brain cells, although the exact mechanism of their toxicity remains unclear. An understanding of how Ab's aggregate to form amyloid must begin with an understanding of how these molecules behave as unaggregated monomers. We use atomistic molecular dynamics simulations in explicit solvent to study the folding of Ab peptides under normal physiological conditions. Our theoretical results are in excellent agreement with the experimental results for those few peptides that have been studied experimentally.
Solvent Ab
 

Protein aggregation

Protein aggregation refers to the process by which proteins initially dissolved in a medium spontaneously self-assemble into larger supramolecular complexes called aggregates. One particular type of protein aggregates, amyloid fibrils, serve specific biological functions in certain organisms, and constitute a promising new material in biomedical applications and nanotechnology. Amyloid fibrils are also involved in a number of neurodegenarative diseases.
Aggregation
Our major goal is to develop an efficient computational model that would allow fibril assembly to be simulated on a computer. We begin by studying initial stages of the aggregation process at which oligomers composed of a small number of proteins are formed. We carry out atomic-resolution simulations that serve dual purpose. First, oligomers are believed to be the main pathogenic agent in amyloid-related diseases such as Alzheimer's, and thus elucidating their structure is essential for the development of therapeutics for these diseases. Second, detailed atomic-level simulations of small oligomers, such as dimers, allow us to construct reliable and efficient simplified protein models for simulations of larger aggregates such as amyloid fibrils. Our strategy is applied to fragments of amyloid b peptides.

Chaperon-assisted folding

A large number of newly-synthesized proteins in living cells are unable to fold spontaneously and require external help to reach their native states. It is the functional role of chaperons, molecular mini-machines such as GroEL/GroES chaperonin complex of e. coli, to help such proteins fold. A number of mechanisms have been proposed that explain how chaperons act. Using methods of molecular modeling, we considered the effects of confining proteins inside chaperonin cavity, as one possible action mechanism. Some of our results were recently confirmed in experiments.
GroelChaperon Groes
 

Folding kinetics

Kinetics are one of the most important aspects of the folding problem. It matters that proteins not only reach their biologically functional native states but that they do so quickly to avoid aggregation. Theoretical models of folding kinetics provide a convenient framework for the general understanding of the factors that define how fast proteins fold. Computer simulations allow unbiased tests of folding models. We considered a model that represents folding as a Brownian motion on 1D free energy surface. We found the model to be adequate for proteins of varying native state topology.

   Kinetics