Novel computational chemistry models for exploring the structure-activity relationships of HIV-1 protease substrates and inhibitors (Immune deficiency)

This thesis presents a detailed study of the substrates and inhibitors of Human Immunodeficiency type 1 (HIV-1) protease using computational chemistry methods. Successful inhibition of (HIV-1) protease halts the replication of the Acquired Immunodeficiency Syndrome (AIDS) virus. The primary objective of these studies was to move towards an improved strategy for the discovery of promising active inhibitor candidates to be deployed against (HIV-1) protease.; The data sets consisted of three congeneric series of inhibitors with experimentally determined inhibitory activity against (HIV-1) protease and a set of substrates. Each of the studies involved: (1) computer modeling of the compounds and optimization of their three-dimensional structure; (2) analysis of the physical properties of these compounds using several computational tools; and (3) statistical analysis of the results to determine the predict power of the model. Comparison between the computed binding energies and experimental activity data (pIC₅₀) of 38 protease inhibitors that are structurally similar to the AIDS drug Indinavir® found a high degree of correlation (r2 = 0.82). Three-dimensional Quantitative Structure Activity Models (3D-QSAR) models using Comparative Molecular Field Analysis (CoMFA) yielded predicted activities that were in excellent agreement with the corresponding experimentally determined values for all three data sets of protease inhibitors. Comparative Molecular Similarity Indices (CoMSIA) models not only displayed high predictive power but also had the added advantage of producing high quality visual images that can provide useful insights into structure-activity relationships.; The inclusion of the calculated enzyme-inhibitor binding energy as an additional descriptor in the CoMFA model, yielded a significant improvement in the internal predictive ability of our model for the third data set ( q2 = 0.45 to q2 = 0.69). Inclusion of the calculated non-polar buried surface led to a further improvement to q2 = 0.71. This novel approach of combining the Structure-Based and Ligand-Based approaches resulted in models with exceptional predictive power.