Offprints of several author's journal articles inserted / Includes bibliography / xx, 285 leaves : ill ; 30 cm. / Title page, contents and abstract only. The complete thesis in print form is available from the University Library. / Thesis (Ph.D.)--University of Adelaide, Dept. of Paediatrics, 1985

Diss. (sammanfattning) Uppsala : Univ., 2002. / Härtill 4 uppsatser.

Diss. (sammanfattning) Uppsala : Uppsala universitet, 2007. / Härtill 4 uppsatser.

Thesis (M.Sc. (Chemistry)) - Rhodes University, 2008.

Thesis (Ph. D.)--University of Washington, 2004. / Vita. Includes bibliographical references (leaves 127-141).

Thesis (M.Sc. (Chemistry)) - Rhodes University, 2009.

Thesis (Ph.D.)--University of Western Australia, 2008.

Thesis (Ph. D.)--School of Pharmacy and Dept. of Chemistry. University of Missouri--Kansas City, 2004. / "A dissertation in pharmaceutical science and chemistry." Advisor: Ashim K. Mitra. Typescript. Vita. Description based on contents viewed Feb. 27, 2006; title from "catalog record" of the print edition. Includes bibliographical references (leaves 175-199). Online version of the print edition.

HIV/AIDS is a global pandemic that poses a great threat especially in Sub-Saharan Africa where the highest population of those infected with the virus is found. It has far reaching medical, socio-economic and scientific implications. The HIV-1 protease enzyme is a prime therapeutic target that has been exploited in an effort to reduce morbidity and mortality. However problems arise from drug toxicity and drug-resistant mutations of the protease which is a motivation for research for new, safer and effective therapies. Evidence exists to show that there are significant genomic differences in Subtype B and C that have a negative effect on the intrinsic binding of inhibitors. It is imperative to look at all perspectives from epidemiological, molecular to the pharmacological ones so as to achieve rational design of therapeutic agents. This study involved the use of in silico structural analysis of the effects of mutations in the active site. The data was provided by the National Institute of Communicable Diseases consisting of HIV-1 Subtype C protease sequences of 29 infants exhibiting drug-resistance to ritonavir and lopinavir. The major active site mutations causing drug resistance identified in this study were M46I, I54V and V82A using the Stanford HIV database tool. Homology modeling without extra restraints produced models with improved quality in comparison to those with restraints. MetaMQAPII results differed when models were visualized as dimers giving erroneous modeled regions in comparison to monomers. A broader study with a larger dataset of HIV-1 subtype C protease sequences is required to increase statistical confidence and in order to identify the pattern of drug resistant mutations. Homology modeling without extra restraints is preferred for calculating homology models for the HIV-1 subtype C. Further investigations needs to be done to ascertain the accuracy of validation results for dimers from MetaMQAPII as it is designed for evaluation of monomers.

Cheminformatic approaches have been employed to optimize the bis-coumarin scaffold identified by Onywera et al. (2012) as a potential hit against the protease HIV-1 protein. The Open Babel library of commands was used to access functions that were incorporated into a markov chain recursive program that generated 17750 analogues of the bis-coumarin scaffold. The Morita-Baylis-Hillman accessible heterocycles were used to introduce structural diversity within the virtual library. In silico high through-put virtual screening using AutoDock Vina was used to rapidly screen the virtual library ligand set against 61 protease models built by Onywera et al. (2012). CheS-Mapper computed a principle component analysis of the compounds based on 13 selected chemical descriptors. The compounds were plotted against the principle component analysis within a 3 dimensional chemical space in order to inspect the diversity of the virtual library. The physicochemical properties and binding affinities were used to identify the top 3 performing ligands. ACPYPE was used to inspect the constitutional properties and eliminated virtual compounds that possessed open valences. Chromene based ligand 805 and ligand 6610 were selected as the lead candidates from the high-throughput virtual screening procedure we employed. Molecular dynamic simulations of the lead candidates performed for 5 ns allowed the stability of the ligand protein complexes with protease model 305152. The free energy of binding of the leads with protease model 305152 was computed over the first 50 ps of simulation using the molecular mechanics Poisson-Boltzmann method. Analysis structural features and energy profiles from molecular dynamic simulations of the protein–ligand complexes indicated that although ligand 805 had a weaker binding affinity in terms of docking, it outperformed ligand 6610 in terms of complex stability and free energy of binding. Medicinal chemistry approaches will be used to optimize the lead candidates before their analogues will be synthesized and assayed for in vivo protease activity.