Clinically effective inhibitors of the HIV protease are generally considered major successes in structure-based drug design. Despite these successes, a cure for AIDS has remained elusive. This is due to the rapid evolution of HIV in the host and the emergence of drug resistance. Continued target identification and drug design will be an important, ongoing effort, yet probably will not result in a definitive therapy against the HIV-1 virus. For viruses that rapidly develop resistance pathways, the rational design of multi-drug therapy becomes an increasingly pressing goal. The present work seeks to explore strategies for multi-drug therapies using a combination of bioinformatics and experimental approaches. A simple experimental model system will be created for exploring multidrug resistance in a high-throughput manner. This model system involves polymerase chain reaction (PCR) random mutagenesis and coexpression of the HIV protease, apoptotic Bax protein and beta-galactosidase in E. coli. In this system, Bax protein can be used to efficiently mimic HIV vitality testing and beta-galactosidase provides a colorimetric assay for screening protease activity in vivo. The rate of mutation can be adjusted in PCR mutagenesis and a low mutation rate will be used to provide a dense, local exploration of sequence space. With the aid of this experimental system, a bioinformatics tool will be developed to predict drug resistance from protein sequence information. This uses a linear transformation model along with a position-specific scoring system. This tool is used to create a """"""""vector space"""""""" model of drug efficacy. This model uses methods similar to principal component analysis to calculate orthogonal vectors in """"""""drug efficacy space"""""""". These vectors are used to determine proportions in drug cocktails. The hypothesis is that a sequential application of these cocktails will span the vector space and, therefore, be the best possible therapy given these available drugs. The model system will be used in an iterative manner to test the multidrug therapy strategy and to refine the parameters of the computational tools. This exploratory project marks the first step in establishing a general design strategy for multidrug therapy.
| Cheng, Ting-Jen; Brik, Ashraf; Wong, Chi-Huey et al. (2004) Model system for high-throughput screening of novel human immunodeficiency virus protease inhibitors in Escherichia coli. Antimicrob Agents Chemother 48:2437-47 |
