Home |  People |  Publications |  Research |  Teaching |  Links |  Contact | 
Prediction and design of protein structures using the physics of protein folding
How are amino acids in a protein arranged and orchestrated into a unique 3-D structure so that it can function? What are the basic physical principles behind the folding of a protein? Protein folding is one of the toughest problems in science and designated as a grand challenge. Some proteins require as little as microseconds to fold into their native structures, yet it is still a challenge to fold a protein in computers.

With the genome sequencing efforts, there have been complete genetic blueprints for hundreds of organisms. The problem is to assimilate that data into a utilitarian form. A common approach to facilitate this task is to determine the 3-D structures of these proteins by experimental methods such as X-ray crystallography and nuclear resonance (NMR) spectroscopy. However many protein sequences are not easily accessible to structure determination by experiment. In the last two years, the number of sequences in the comprehensive databases, such as SwissProt/TrEMBL or GenPept, has increase more than 100 percent, yet the number of experimentally determined structures deposited in Protein Data Bank has only increased by 40 percent.

We have developed a purely physics-based, ab-initio structure prediction method called Zipping and Assembly Method (ZAM). ZAM uses replica exchange molecular dynamics (REMD) and the AMBER ff96 force field with Generalized Born implicit solvation. The core of ZAM is " zipping and assembly" (ZA) search strategy by which we believe proteins physically fold up. ZA principle suggests that local structuring occurs along the chain at early times, followed by either substructure growth with the addition of neighbouring amino-acids to form new contacts (zipping) or coalescence with other substructures (assembly). Thus, using ZAM we not only can predict high resolution structure of a protein but its physical folding routes as well.

Our aim is to bridge the gap between the sequences and structures using ZAM. Particularly we are interested in (i) designing new protein structures and interactions (ii) improving homology modeling, (iii) exploring the impact of mutation on foldable sequences and (iv) understanding the physical folding routes of proteins with different topologies.

Protein-protein, protein-ligand interaction using flexible docking

Protein-protein, protein-ligand interactions are a vital part of regulation and tranmission of information among cells. Our long-term goal is to have a better understanding of the physical principles of how biomolecules recognize each other and how these interactions build up macromolecular machines and networks.
The big payoff in protein structure modeling using ZAM will be in understanding ligand binding, protein-protein interactions, drug discovery, and predicting protein function. Current docking methods attempt to predict the bound ligand by keeping the protein fixed while moving the target ligand around the binding site along with energy minimization. The major problems with this approach are: (i) proteins are dynamic, flexible, and deformable, so keeping them fixed often misses the correct binding modes, (ii) relying on pure energy minimization is not sufficient to differentiate between ligands that have affinity of binding to the target from those that don't

We are developing a multiscale computational model that will explore the conformational dynamics of protein and ligand at low resolution and use these conformations to calculate binding free energies in all-atom based models. The main goal of this research project is to understand the physical principles of biological networks from atomic interactions to whole cellular systems.

Center for Biological Physics Arizona State University Bateman Physical Sciences Building F-Wing, Room 359 Tempe, AZ 85287-1504