It is necessary to understand the molecular basis for the action of HIV protease and its inhibitor-resistant variants in order to develop new inhibitors and new therapeutic strategies. Previously, the comparison of the crystal structures and specificities of two distinctly different retroviral proteases from HIV and Rous sarcoma virus (RSV) were used to identify important inhibitor-protease interactions and the protease residues that are critical for recognition of substrates. These key residues were predicted to occur in the inhibitor-resistant variants of HIV protease, and are the most commonly mutated residues in the known resistant variants. Variants of both HIV-1 and RSV proteases will be studied in order to model the development of resistance to inhibitors. The crystals structures of these variant proteins will be determined, their specificities for peptide substrates representing the natural polyprotein cleavage sites will be measured, and their effects on viral replication will be studied. New inhibitors will be evaluated for their ability to inhibit HIV protease variants and for their antiviral effects. Computational methods have been developed that predict the relative efficiency of peptide substrates of HIV protease. These calculations will be used to predict the effects of the resistant mutants on cleavage of the polyproteins and viral replication and the activity of different inhibitors. One major advantage is that these predictions will be applicable to any newly discovered resistant mutants of HIV protease, and will help to design new inhibitors to overcome resistance.