Classification and Design of HIV-1 Integrase Inhibitors Based on Machine Learning.

A key enzyme in human immunodeficiency virus type 1 (HIV-1) life cycle, integrase (IN) aids the integration of viral DNA into the host DNA, which has become an ideal target for the development of anti-HIV drugs. A total of 1785 potential HIV-1 IN inhibitors were collected from the databases of ChEMBL, Binding Database, DrugBank, and PubMed, as well as from 40 references. The database was divided into the training set and test set by random sampling. By exploring the correlation between molecular descriptors and inhibitory activity, it is found that the classification and specific activity data of inhibitors can be more accurately predicted by the combination of molecular descriptors and molecular fingerprints. The calculation of molecular fingerprint descriptor provides the additional substructure information to improve the prediction ability. Based on the training set, two machine learning methods, the recursive partition (RP) and naive Bayes (NB) models, were used to build the classifiers of HIV-1 IN inhibitors. Through the test set verification, the RP technique accurately predicted 82.5% inhibitors and 86.3% noninhibitors. The NB model predicted 88.3% inhibitors and 87.2% noninhibitors with correlation coefficient of 85.2%. The results show that the prediction performance of NB model is slightly better than that of RP, and the key molecular segments are also obtained. Additionally, CoMFA and CoMSIA models with good activity prediction ability both were constructed by exploring the structure-activity relationship, which is helpful for the design and optimization of HIV-1 IN inhibitors.

Study on suitable analysis method for HIV-1 non-catalytic integrase inhibitor.

BACKGROUND: Integrase (IN) is an essential protein for HIV replication that catalyzes insertion of the reverse-transcribed viral genome into the host chromosome during the early steps of viral infection. Highly active anti-retroviral therapy is a HIV/AIDS treatment method that combines three or more antiviral drugs often formulated from compounds that inhibit the activities of viral reverse transcriptase and protease enzymes. Early IN inhibitors (INIs) mainly serve as integrase strand transfer inhibitors (INSTI) that disrupt strand transfer by binding the catalytic core domain of IN. However, mutations of IN can confer resistance to INSTI. Therefore, non-catalytic integrase inhibitors (NCINI) have been developed as next-generation INIs. METHODS: In this study, we evaluated and compared the activity of INSTI and NCINI according to the analysis method. Antiviral activity was compared using p24 ELISA with MT2 cell and TZM-bl luciferase system with TZM-bl cell. Each drug was serially diluted and treated to MT2 and TZM-b1 cells, infected with HIV-1 AD8 strain and incubated for 5 and 2 days, respectively. Additionally, to analyze properties of INSTI and NCINI, transfer inhibition assay and 3'-processing inhibition assay were performed. RESULTS: During screening of INIs using the p24 ELISA and TZM-bl luciferase systems, we found an inconsistent result with INSTI and NCINI drugs. Following infection of MT2 and TZM-bl cells with T-tropic HIV-1 strain, both INSTI and NCINI treatments induced significant p24 reduction in MT2 cells. However, NCINI showed no antiviral activity in the TZM-bl luciferase system, indicating that this widely used and convenient antiretroviral assay is not suitable for screening of NCINI compounds that target the second round of HIV-1 replication. CONCLUSION: Accordingly, we recommend application of other assay procedures, such as p24 ELISA or reverse transcription activity, in lieu of the TZM-bl luciferase system for preliminary NCINI drug screening. Utilization of appropriate analytical methods based on underlying mechanisms is necessary for accurate assessment of drug efficacy.

Classification of active and weakly active ST inhibitors of HIV-1 integrase using a support vector machine.

Using a support vector machine (SVM), two computational models were built to predict whether a compound is an active or weakly active strand transfer (ST) inhibitor based on a dataset of 1257 ST inhibitors of HIV-1 integrase. The model built with MACCS fingerprints gave a prediction accuracy of 91.82% and a Matthews Correlation Coeffiient (MCC) of 0.73 on test set, and the model built with 40 MOE descriptors gave a prediction accuracy of 93.64% and an MCC of 0.79 on test set. Some molecular properties such as electrostatic properties, van der Waals surface area, hydrogen bond properties and the number of fluorine atoms are important factors influencing the interactions between the inhibitor and the integrase. Some scaffolds like ß-diketo acid and its derivatives, naphthyridine carboxamide or the isosteric of it and pyrimidionones may play crucial rule to the activity of the HIV-1 integrase inhibitors.

Novel quinolinonyl diketo acid derivatives as HIV-1 integrase inhibitors: design, synthesis, and biological activities.

Novel quinolinonyl diketo acids were designed to obtain integrase (IN) inhibitors selectively active against the strand transfer (ST) step of the HIV integration process. Those new compounds are characterized by a single aryl diketo acid (DKA) chain in comparison to 4, a bifunctional diketo acid reported by our group as an anti-IN agent highly potent against both the 3'-processing and ST steps. Compound 6d was the most potent derivative in IN enzyme assays, while 6i showed the highest potency against HIV-1 in acutely infected cells. The selective inhibition of ST suggested the newly designed monofunctional DKAs bind the IN-DNA acceptor site without affecting the DNA donor site.

Cluster analysis and three-dimensional QSAR studies of HIV-1 integrase inhibitors.

Three-dimensional quantitative structure-activity relationship (3D QSAR) and cluster analysis were applied to a variety of HIV-1 integrase inhibitors. One structure was chosen from each of 11 classes of inhibitors to represent the whole class in descriptor-based cluster analysis. The 11 classes of inhibitors were classified into two groups. The molecular field analysis (MFA) models for these two clusters had r2 values of 0.90 and 0.95 and q2 values of 0.85 and 0.91 that were noticeably enhanced from those of conventional QSAR models. The five test compounds, which were proposed to have a common binding site near the metal in HIV-1 integrase based on docking studies by Sotriffer et al., were utilized to compare the predictive capability of MFA and conventional QSAR models. Among these five compounds, only L-chicoric acid belongs to cluster 1 and the other four belong to cluster 2. MFA models give better overall predictions and more importantly the activity of these test compounds is better predicted by the MFA model derived from the cluster each test compound belongs to. The necessity of dividing the inhibitors into two groups to obtain predictive QSAR models supports the likelihood of two separate binding sites.

Inhibition of human immunodeficiency virus type I integrase by naphthamidines and 2-aminobenzimidazoles.

Retroviral integrases catalyze two of the steps of insertion of proviral DNA into the host genomic DNA. Inhibitors that target the second step, strand transfer into the host DNA, have been demonstrated to have antiviral activity in cell culture. We describe two classes of HIV-1 integrase inhibitors that block strand transfer, one based on a naphthamidine core and one on a benzimidazole core. While the naphthamidine compounds showed some propensity to interact with the DNA substrate, both classes were shown to bind directly to integrase. The naphthamidine compounds showed activity in cell culture, and a direct effect on integrase was indicated by an increase in 2-LTR products in the presence of a naphthamidine compound. These two classes of compounds represent potential starting points for the development of new classes of integrase inhibitors.