Computer-Aided Design of Anti-HIV Drugs

描述(由申请人提供)：这项工作的目的是开发、应用和传播计算机技术，以优化先导化合物和设计选择性地阻止艾滋病毒感染和复制的新化学实体。这些进展包括开发计算工具(A)预测蛋白质-配体结合亲和力和蛋白质-配体复合体的结构，(B)评估所提议的分子的性质和药物相似性，以及(C)创建和评估潜在抑制剂的虚拟文库。计算方法的应用提高了对支配蛋白质-配体结合亲和力变化的结构和能量因素的认识，以及蛋白质突变对药物活性的不同影响的起源。这一知识为合理设计对广泛病毒突变保持有效的药物提供了基础。建议的应用集中在改进的HIV-1逆转录酶非核苷抑制剂(NNRTI)的开发进展上。GenMol用于在蛋白质的结合部位内构建单个有机分子或组合库。开发必要的评分功能需要应用于具有野生型(WT)的多系列NNRTI和所有主要的NNRTI诱导的突变体。通过在蒙特卡罗(MC)模拟中包含溶剂和构型采样的额外细节，可以更准确地表征活性的变化。因此，还对具有WT和突变体RT的一系列NNRTI进行了MC/ELR(扩展线性响应)研究，目的是开发可靠的计算筛查来评估所建议的NNRTI的潜力。在最高计算水平上，MC/FEP(自由能微扰)计算正被用于阐明泛类耐药突变(如K103N和Y188L)对奈韦拉平、Eefavirenz、DPC083和TMC125等特定药物活性的差异影响的原子水平来源。随后，将使用类似的研究通过破坏gp41-gp120复合体的融合形式的组装来设计有效的艾滋病毒进入抑制剂；特定的靶点是存在于由gp41的N末端片段形成的卷曲的脊部中的疏水口袋。

DESCRIPTION (provided by applicant): The aim of the work is to develop, apply, and disseminate computational technology for the optimization of lead compounds and for the design of new chemical entities that selectively block HIV infection and replication. The advances include development of computational tools (a) to predict protein-ligand binding affinities and the structures of protein-ligand complexes, (b) to evaluate properties and drug-likeness of proposed molecules, and (c) to create and evaluate virtual libraries of potential inhibitors. Applications of the computational methods yield enhanced knowledge of the structural and energetic factors that govern variations in protein-ligand binding affinities and of the origins of differential effects of protein mutations on drug activities. This knowledge provides a basis for the rational design of drugs that remain effective against a broad range of viral mutants. The proposed applications focus on advances for the development of improved non-nucleoside inhibitors of HIV-1 reverse transcriptase (NNRTIs). GenMol is used to construct individual organic molecules or combinatorial libraries inside a protein's binding site. Development of the necessary scoring functions requires application to multiple series of NNRTIs with wild type (WT) and all principal NNRTI-induced mutants. The added detail from inclusion of the solvent and configurational sampling in Monte Carlo (MC) simulations permits more accurate characterization of variations in activity. Thus, MC/ELR (extended linear response) studies are also performed for series of NNRTIs with WT and mutant RT with the goal of developing a reliable computational screen to evaluate the potential of proposed NNRTIs. At the highest computational level, MC/FEP (free energy perturbation) calculations are being used to elucidate the atomic level origins of the differential effects of pan-class resistance mutations such as K103N and Y188L on the activity of specific drugs including nevirapine, efavirenz, DPC083, and TMC125. Subsequent, analogous studies will be used to design potent inhibitors of HIV viral entry through disruption of assembly of the fusogenic form of the gp41-gp120 complex; the specific target is a hydrophobic pocket that exists in ridges of a coiled coil formed by N-terminal segments of gp41.