HIV Integrase Modeling and Computer-Aided Inhibitor Development

该项目的主要目的是阐明HIV-1整合酶蛋白与DNA和/或抑制剂的络合结构，利用所获得的结构知识设计更好的该酶抑制剂，以开发新的抗HIV药物，并应用任何其他计算机辅助药物设计方法，以帮助识别新的、有前途的HIV-1整合酶抑制剂。HIV整合酶(IN)是病毒编码的酶，负责将逆转录病毒DNA整合到宿主基因组中。艾滋病毒生命周期中的这一步对病毒复制至关重要。抑制整合被认为是抗艾滋病治疗的一个有吸引力的靶点，因为目前还不知道IN的细胞同源物，从而增加了开发有效的、低毒的抗IN药物的希望。在给予蛋白酶和逆转录酶(RT)抑制剂的鸡尾酒疗法期间出现多药耐药病毒表型，进一步突出了替代治疗方法的必要性。In是一种32 kDa的蛋白质，是病毒颗粒中包含的Gag-Poll融合蛋白前体的产物。在RT完成前病毒DNA合成后，IN从病毒DNA末端各切割两个核苷酸(“3‘-加工”)。在随后迁移到宿主细胞核后，IN催化将在3‘-处理步骤中产生的凹陷的3’末端插入宿主DNA的一条链中。这种反应称为3‘末端连接(也称为整合或链转移)，在病毒DNA的两端同时发生。随后的缝隙连接被认为是由细胞DNA修复酶执行的，以产生完全整合的前病毒DNA。以前的工作，主要是基于NCI数据库中的3D药效团搜索，已经产生了一些IN的抑制剂。随着HIV-1 IN的更多、更好的实验结构(通过X射线结晶学和核磁共振)以及ASV整合酶等密切相关的酶的出现，已经有可能对更大的结构进行建模，包括全长蛋白质的多聚体模型，而对这些结构的实验结构尚不可用。我们已经利用所有可用的实验证据，通过分子建模技术生成了这样的结构。正如Karki等人，2004年所描述的，特别强调在3‘-加工之后但在链转移之前，获得与病毒DNA相对的酶的活性部位的模型。该模型可用于基于结构的、在体内保持活性的抑制剂的设计。我们利用这些结构模型研究了各种二酮基酸HIV-1 IN抑制剂的潜在结合模式，这些模式还没有实验的复杂结构可用。结果表明，二酮酸类IN抑制剂可能与催化部位的金属离子发生络合作用，并可防止病毒DNA 3‘端暴露于人DNA。这些模型被成功地完全用于抑制剂的开发，利用了包括我们的数据库项目中描述的资源，特别是通过对包含2600多万个可购买筛选样本的数据库进行电子筛选。目前的努力一直集中在基于配体的抑制剂设计上，利用来自少数已进入后期临床试验或被批准为抗HIV药物的分子的结构信息。基于这些结构基序，设计了一系列IN相关专利未涵盖的新化合物，并通过数据库项目中新实施的半定制在线合成请求系统(SCSORS)提交报价进行合成。从全球10多家不同供应商引用的8000多种化合物中，已经选出了近100种化合物并提交购买。到目前为止，从这组化合物中已经得到了大约25个化合物，其中一些化合物表现出一定的抗IN活性。预计很快会有另外几十种化合物，一旦收到就会进行化验。

The principal objective of this project is to elucidate the structure of the HIV-1 integrase protein, complexed with DNA and/or inhibitors, to use the structural knowledge thus obtained to design better inhibitors of this enzyme with the goal of developing new anti-HIV drugs, and to apply any other computer-aided drug design method that may be helpful in identifying new, promising HIV-1 integrase inhibitors. HIV integrase (IN) is the virally encoded enzyme responsible for integration of the retroviral DNA into the host genome. This step in the life cycle of HIV is essential for viral replication. Inhibition of integration is seen as an attractive target in the development of anti-AIDS therapies because no cellular homologue to IN is known, thus raising the hope that effective anti-IN based drugs with low-toxicity can be developed. The emergence of multidrug-resistant virus phenotypes during administration of cocktails of protease and reverse transcriptase (RT) inhibitors has further highlighted the need for alternative therapeutic approaches. IN is a 32kDa protein that is a product of the gag-pol fusion protein precursor contained in the virus particle. Upon completion of proviral DNA synthesis by RT, IN cleaves two nucleotides from each viral DNA end ("3'-processing"). After subsequent migration to the host cell's nucleus, IN catalyzes the insertion of the recessed 3'-terminus, generated during the 3'-processing step, into one strand of the host DNA. This reaction is termed 3' end joining (also known as integration or strand transfer) and occurs for both ends of the viral DNA simultaneously. The subsequent gap-joining is presumed to be performed by cellular DNA repair enzymes to yield a fully integrated proviral DNA. Previous work, mainly based on 3D-pharmacophore searches in the NCI database, had yielded a number of inhibitors of IN. With the advent of more, and better, experimental structures (by X-ray crystallography and NMR) of HIV-1 IN as well as of closely related enzymes such as ASV integrase, it has become possible to model larger structures including multimeric models of the full-length protein, for which experimental structures are not available as of yet. We have generated such structures by means of molecular modeling techniques using all available experimental evidence. Special emphasis was placed on obtaining a model of the enzyme's active site with the viral DNA apposed to it as it might be after 3'-processing but before strand transfer, as described in Karki et al., 2004. This model is useful for structure-based inhibitor design of inhibitors which retain activity in vivo. We have made use of these structural models to study the potential binding modes of various diketo-acid HIV-1 IN inhibitors for which no experimental complexed structures are available. The results indicate that the diketo-acid IN inhibitors probably chelate the metal ion in the catalytic site and also prevents the exposure of the 3'-processed end of the viral DNA to human DNA. These models were success fully used for inhibitor development, utilizing resources including those described in our database project, in particular through in silico screening of a database of more than 26 million purchasable screening samples. Current efforts have been focusing on ligand-based inhibitor design, making use of the structural information coming from those few molecules that have made it into late-phase clinical trials or been approved as anti-HIV drug. Based on these structural motifs, a series of novel compounds not covered by IN-related patents were designed and submitted for quotation for synthesis via the newly implemented Semi-Custom Online Synthesis Request System (SCSORS) mentioned in the Database project. From the more than 8,000 compounds quoted by more than 10 different suppliers world-wide, a set of nearly 100 has been chosen and submitted for purchase. About 25 compounds have been obtained so far from this set, with some anti-IN activity exhibited by a few. Several tens of additional compounds are expected soon and will undergo assaying as soon as they are received.