2D and 3D QSAR analysis of some valproic acid metabolites and analogues as anticonvulsant agents

To investigate the structural features, responsible for the variations in anticonvulsant activity of a series of twenty six valproic acid (VPA) metabolites and analogues.

Different approaches for quantitative structure--activity relationship analysis (QSAR) as conventional 2D QSAR analysis and comparative molecular field analysis (3D QSAR) were used. The 2D QSAR was performed with more than twenty structure descriptors as the partition and distribution coefficients, topological, geometrical and electronic descriptors, and indicator variables. The electronic descriptors were calculated for the energetically most stable conformers. For the need of 3D QSAR steric and electrostatic potential maps were generated. Partial least squares (PLS) analysis has been carried out for the statistical evaluation of the models and weighted least squares (WLS) analysis was used for the visualization of the results.

It was established that the two approaches--2D and 3D QSAR, prove the importance of the lipophilicity of the compounds for anticonvulsant activity. The results from both the approaches suggest that a substitution at alpha-position is essential for a higher activity.

3D QSAR is useful for describing the steric and electrostatic fields, important for the activity. For predicting the activity of new compounds 2D QSAR tools were proposed.