Roterman I. (1), Skowronek M. (1), Konieczny L. (2), Stopa B. (2), Rybarska J. (2), Piekarska B. (2), Górny M. (2)
1 Department of Biostatistics and Medical Informatics
2 Institute of Medical Biochemistry
Collegium Medicum - Jagiellonian University
INTRODUCTION
Mathematical modeling of ligand-protein complex creation has been done for years. To develop new drug one employs simulation of ligand docking to active sites of enzymes. Two mathematical methods are basic to this procedure: energy minimization and molecular dynamics simulation. So far only low-mass molecules attachable to the binding site have been treated as ligands to modify the activity of enzymes (1).
There are a few well known programs which solve problems in computer-aided drug design. Some of them are only in 'numerical' form (Amber(2), ECEPP), and they can be run on supercomputers without any additional equipment. There are also a few with elaborate visualization tools. It is easy to run them on work-stations.Large memory and high speed calculations are required to calculate the energy minimum protein conformation and simulate ligand docking.
A new family of ligands has been discovered in our group (3). Some proteins are able to bind not only a single separate molecule but also a micelle which incorporates even ten or more molecules in the form of a single supramolecular ligand. Binding to protein does not disturb the coherence of the ligand.
The main computational problem concerns the relation of the protein structure to the supramolecular organization of the ligand and its location in the protein molecule. The criteria for complementarity can be defined based on the results of computer simulation of the micelle docked to the protein.
All calculations presented in this work were done on HP/Convex Exemplar SPP in Academic Computer Center CYFRONET in Krakow.
RESULTS
A mono-molecular micelles containing eight molecules of different bis-azo dyes were mathematically constructed and evaluated in terms of their stability and structural flexibility. The set of bis-azo dyes was selected to represent different degrees of flexibility, different symmetries and different relations to the hydrophobic and hydrophilic groups present in the molecule. A few orientations were selected to evaluate the influence of the starting structure for each molecule. Molecular organizations representing low-energy structures were selected to evaluate the stability of the system using molecular dynamics simulation.The energetically stable conformations were examined as potential ligands able to be bound to the polypeptide chain.
Analysis and comparison of the final structures yielded criteria for compatibility with the polypeptide chain structures present in the protein molecule. Some of the dyes created regular ribbon-like structures tending propagate to infinitely long chains; others resembled a propeller-like organization with finite numbers of units in the micelle. One dye tended to create doubles as the optimal form for its intermolecular organization. Comparison of the energy components elucidated the influence of functional groups on the structure of the micelle. Strong internal interaction prevailed over intermolecular interaction, resulting in a low tendency toward supramolecular organization.
CONCLUSIONS
There are several different organizations of supramolecular systems created by bis-azo dyes. Those that are stable in ribbon-like form can be easily incorporated to the Beta-form of the polypeptide chain and can bind as a single specific ligand. The globe-like forms of supramolecular organization found for particular bis-azo dyes exclude the possibility of interaction with protein.
The structures of micelles found with mathematical modeling yielded criteria for potential supramolecular ligand-protein interaction. The molecule should be rather rigid, symmetrical and bi-polar with a hydrophobic core to create a micelle structurally compatible with the polypeptide chain forms represented in proteins. This binding, which is very often outside the center of activity, can enhance the activity of the enzyme.
REFERENCES
1. Goodsell D.S. et al. (1990) Proteins Struct Func Gen. 8, 195-202
2. Kollman P.A. et al. (1991) AMBER, UCSF
3. Roterman I. et al.(1998) Computers & Chemistry - in press