Study On HIV-1Integrase Inhibitors Screening And Their Activity Assays

Since the first discovery of acquired immunodeficiency syndrome (AIDS) in1981，the rapid spreading of AIDS has been a serious threat to human health in theworld, causing great damage to the society. Human immunodeficiency virus-type1(HIV-1) is the etiological agent of AIDS. The active sites and unique enzymes of thereplication cycle of HIV-1provide many potential targets for chemotherapeuticintervention. One of these, Integrase (IN), which is encoded by HIV-1genome and isindispensable, catalyzes the integration process in the viral life cycle. IN helps theviral reverse transcription product cDNA to integrate into the host chromosome. Thereare no functional analogues of IN in the normal human host cells, which makes theinhibitors of IN to have lower toxicity to normal cells. At the same time, IN usessingle active site to accommodate two different DNA substrates, which will limit theresistant strains of HIV-1towards IN inhibitors to a certain extent. As such, INbecomes an important target for designing and developing the novel anti-HIV-1drugs.In vivo, integration is a multistep process which includes an assembly of the viralreverse transcription product cDNA on IN, endonucleolytic3′-processing of thecDNA and strand transfer of the cDNA to host cell DNA. Both3′-processing andstrand transfer reactions can be modeled in vitro using purified recombinant INprotein, oligonucleotide DNA substrates and divalent cationic cofactor. Applyingthese assays to screen inhibitors is a main method for IN inhibitors screening in vitro.In this thesis, recombinant IN proteins were expressed and purified, and ahigh-throughput assay for IN inhibitor screening has been developed. Compared tothe previous work, the new assay was proved to be sensitive, effective, less timeconsuming and lower cost.Drug screening method includes not only the biological experiment, computeraided virtual screening method has also played an increasingly important role. Besidesthe screening method based on biology experiment, this thesis also examined thevirtual screening methods for IN inhibitors through computer simulation and machinelearning methods. The application of the support vector machine (SVM) in inhibitorsscreening is usually realized through quantitative structure-activity relationship(QSAR) modeling. The CoMFA was applied in combination with SVM to build theprediction model, which can judge the external prediction ability. The results showthat SVM exhibits stronger modeling and predicting ability.The main content of the thesis consists of the following major aspects:(1) Expression, purification of recombinant IN protein and detection theactivity of the strand transfer reaction The recombinant expression vector pET-28a（+）/IN（F185K/C280S）wastransformed into E. coli BL21（DE3）competent cells for prokaryotic expression. Thevector contained the IN gene and F185K/C280S mutations were brought to theconstructed IN gene for the purpose of enhancing the protein solubility. Therecombinant IN protein was purified by affinity chromatography. It was validated thatthe recombinant IN protein was pure and active for screening assay development. Thehigh-throughput enzyme-linked immunosorbent assay (ELISA) was applied tomeasure the IN-catalyzed strand transfer reaction activity. This work was thecornerstone to establish the inhibitors screening assay of IN strand transfer reaction.(2) Development of a convenient and efficient inhibitor screening assay for theIN strand transfer reactionThe aim of the present study was to develop a highly effective and moreconvenient screening assay for HIV-1IN inhibitors targeting strand transfer.Biotin-labeled donor DNA and FITC-labeled target DNA were synthesized andapplied in the assay, and streptavidin-coated magnetic beads were used to capture theproduct DNA in the reaction system. Finally, the fluorescence signal was detected bya fluorescence microplate reader for the calculation of inhibition ratio of a sample.Two reported IN inhibitors S-1360and MK-0518were tested to validate the screeningassay, and the results were in accordance with previous experiment data, whichindicated that the assay could be used for the IN inhibitors screening of the strandtransfer reaction.Compared to previous assays, the assay proposed in this thesis has notableimprovements:(i) In previous work, biotin and digoxin modified donor DNA andtarget DNA were designed and composed, and streptavidin-coated magnetic beadswere involved to capture the reaction product DNA strand. The signal was readthrough the specific alkaline phosphatase chromogenic reaction. In the current assay,the fluorescence signal was detected by a fluorescence microplate reader for theevaluation of sample activity. Therefore, this method is less laborious and less timeconsuming.(ii) In previous work, alkaline phosphatase labeled antibody was used todetect the digoxin group connected with DNA. But the current assay does not needany antibody, hence its cost is much cheaper.The assay is also capable to be applied in the study of disintegration reactioncharacter and pharmacology of IN inhibitors. This work possesses significantacademic and application value.(3) CoMFA and SVM were applied to construct the screening model of theHIV-1IN pyrimidones inhibitorsThe application of SVM in inhibitors screening is usually realized through the QSAR modeling. The main task is to establish the correlation between the structureparameters and the activity of compounds. Then, the regression equation is generatedto predict the activity of new compounds. In this thesis, the CoMFA software wasused to calculate topological descriptors, polarizable descriptors, hydrophilicdescriptors and other structural parameters for68pyrimidones compounds. Thestructural parameters were selected as inputs of SVM to establish the non-linearregression model. The correlation coefficients of experimental and predicted activityvalues are0.938for the training set and0.804for the test set. It shows that the resultsobtained from the model are consistent with the experiment data. This indicates thatthe model can be used effectively in HIV-1IN inhibitors screening.This work provides valuable findings and information for the predicting model ofinhibitors by using the SVM algorithm combined with CoMFA. It provides not only amethod of virtual screening, but also biological information for the design ofanti-HIV-1drugs.