Montserrat Barbany | PubMed Health

PubMed2004

Towards a MIP-based alignment and docking in computer-aided drug design.

Structural alignment of ligands in their biological conformation is a crucial step in the building of pharmacophoric models in structure-based drug design. In addition, docking algorithms are limited in some cases by the quality of the scoring functions and the limited flexibility of the environment that the different programs allow. On the other hand, GRID molecular interaction potentials (MIPs) have been used for a long time in 3D-QSAR studies. However, in most of these studies the alignment of the molecules is performed on the basis of geometrical or physico-chemical criteria that differ from the MIPs used in the partial least squares statistical analysis. We have previously developed a method to use the same scoring function for the molecular alignment and for 3D-QSAR studies. This methodology, based on the use of GRID potentials, consists in the weighted averaging of similarities of the relevant MIPs of the molecules to be aligned. Here we present a method to obtain the weights for the different GRID probes in the average based on the structural information on protein-ligand complexes for relevant systems. The method, implemented in MIPSIM, is shown to yield good accuracy in the prediction of the alignments for two systems: a set of three inhibitors of dihydrofolate reductase and a set of fifteen non-nucleoside HIV-1 reverse transcriptase inhibitors (NNRTIs). The smooth GRID potentials are shown to capture the flexible character of the active site, as opposed to traditional docking scoring energy functions.

Binding Sites↗

PubMed2003

On the generation of catalytic antibodies by transition state analogues.

The effective design of catalytic antibodies represents a major conceptual and practical challenge. It is implicitly assumed that a proper transition state analogue (TSA) can elicit a catalytic antibody (CA) that will catalyze the given reaction in a similar way to an enzyme that would evolve (or was evolved) to catalyze this reaction. However, in most cases it was found that the TSA used produced CAs with relatively low rate enhancement as compared to the corresponding enzymes, when these exist. The present work explores the origin of this problem, by developing two approaches that examine the similarity of the TSA and the corresponding transition state (TS). These analyses are used to assess the proficiency of the CA generated by the given TSA. Both approaches focus on electrostatic effects that have been found to play a major role in enzymatic reactions. The first method uses molecular interaction potentials to look for the similarity between the TSA and the TS and, in principle, to help in designing new haptens by using 3D quantitative structure-activity relationships. The second and more quantitative approach generates a grid of Langevin dipoles, which are polarized by the TSA, and then uses the grid to bind the TS. Comparison of the resulting binding energy with the binding energy of the TS to the grid that was polarized by the TS provides an estimate of the proficiency of the given CA. Our methods are used in examining the origin of the difference between the catalytic power of the 1F7 CA and chorismate mutase. It is demonstrated that the relatively small changes in charge and structure between the TS and TSA are sufficient to account for the difference in proficiency between the CA and the enzyme. Apparently the environment that was preorganized to stabilize the TSA charge distribution does not provide a sufficient stabilization to the TS. The general implications of our findings and the difficulties in designing a perfect TSA are discussed. Finally, the possible use of our approach in screening for an optimal TSA is pointed out.

Antibodies, Catalytic↗