Luhua Lai

Further development and validation of empirical scoring functions for structure-based binding affinity prediction

New empirical scoring functions have been developed to estimate the binding affinity of a given protein-ligand complex with known three-dimensional structure. These scoring functions include terms accounting for van der Waals interaction, hydrogen bonding, deformation penalty, and hydrophobic effect. A special feature is that three different algorithms have been implemented to calculate the hydrophobic effect term, which results in three parallel scoring functions. All three scoring functions are calibrated through multivariate regression analysis of a set of 200 protein-ligand complexes and they reproduce the binding free energies of the entire training set with standard deviations of 2.2 kcal/mol, 2.1 kcal/mol, and 2.0 kcal/mol, respectively. These three scoring functions are further combined into a consensus scoring function, X-CSCORE. When tested on an independent set of 30 protein-ligand complexes, X-CSCORE is able to predict their binding free energies with a standard deviation of 2.2 kcal/mol. The potential application of X-CSCORE to molecular docking is also investigated. Our results show that this consensus scoring function improves the docking accuracy considerably when compared to the conventional force field computation used for molecular docking.

A New Atom-Additive Method for Calculating Partition Coefficients

A new method is presented for the calculation of partition coefficients of solutes in octanol/water. Our algorithm, XLOGP, is based on the summation of atomic contributions and includes correction factors for some intramolecular interactions. Using this method, we calculate the log P of 1831 organic compounds and analyze the derived parameters by multivariate regression to generate the final model. The correlation coefficient for fitting this training database is 0.968, and the standard deviation is 0.37. The result shows that our method for log P estimation is applicable to quantitative structure−activity relationship studies and gives better results than other more complicated atom-additive methods.

RASSE : A NEW METHOD FOR STRUCTURE-BASED DRUG DESIGN

A novel method, RASSE, has been developed to suggest reasonable structures which can fit well to the binding sites of receptors. Molecules are generated by an iterative growing procedure in which atoms are added to existing fragments. Potential ligands are then picked out by special scoring rules. This atomgrowing based method is characterized by combinatorial searching of atom types and conformations. To some extent, it is the computer simulation of combinatorial chemistry. This method has been applied to the design of inhibitors for E. coli dihydrofolate reductase and human phospholipase A2. The results demonstrate that this program is capable of generating reasonable structures, thus proving its power in drug design.

Protein ligand docking based on empirical method for binding affinity estimation.

An empirical protein-ligand binding affinity estimation method, SCORE, was incorporated into a popular docking program, DOCK4. The combined program, ScoreDock, was used to reconstruct the 200 protein-ligand complex structures and found to give good results for the complexes with high binding affinities. A quality assessment method for docking results from ScoreDock was developed based on the whole test set and tested by additionally selected complexes. The method significantly improves the docking accuracy and was shown to be reliable in docking quality assessment. As a docking tool in structural based drug design, ScoreDock can screen out final hits directly based on the predicted negative logarithms of dissociation equilibrium constants of protein-ligand complexes, and can explicitly deal with structure water molecules, as well as metal atoms.

Structural features of toxic chemicals for specific toxicity.

We have studied the structural features of toxic chemicals from the RTECS database associated with specific toxicity. The frameworks, functional groups, and structure patterns of molecules are taken into account. Potential active frameworks, groups, and structure patterns for specific toxicity are gained by computational chemistry approaches. These structural features of toxic chemicals will be helpful in understanding activities of toxic chemicals and useful in predicting the toxicity of chemicals, especially in the early stage of drug design.

Calculation of Protein Surface Loops Using Monte-Carlo Simulated Annealing Simulation

Abstract A modified program for protein loop modelling is presented with significant improvements on our previous study (Zhang et al. 1997, Biopolymers, Vol. 41, pp. 61-72), which is capable of sampling the entire conformational space and identifying the low-energy candidates by Monte-Carlo simulated annealing simulation and a cluster analysis method. Twenty flexible surface loops connecting different secondary structures were selected to test the efficiency of this program. The averaged deviations of backbone heavy atoms for four to eight-residue-loops are 0.19, 0.27, 0.46, 0.41 and 0.87Å respectively. High speed of calculation is achieved with a simplified energy function and a grid-mapping method. As a comparison of single simulation, it takes only four seconds for the simplest four-residue loop and forty-three seconds for the most complex eight-residue loop on a PII-350-Linux platform.

Predictive toxicology of chemicals and database mining

The toxic chemicals from the database Registry of Toxic Effects of Chemical Substances (RTECS) were analyzed by structural similarity comparison, which shows that the structure patterns or characteristics of toxic chemicals exist in a sufficiently large database. Then, a two-step strategy was proposed to explore noncongeneric toxic chemicals in the database: the screening of structure patterns by similarity comparison and the derivation of detailed relationship between structure and activity by using comparative molecular field analysis (CoMFA) of Quantitative Structure-Activity Relationship (QSAR) technologies. From the performance of the procedure, such a stepwise scheme is demonstrated to be feasible and effective to mine a database of toxic chemicals. It can be anticipated that database mining of toxic chemicals will be a new area for predictive toxicology of chemicals.