Structural Basis of 2-Phenylamino-4-phenoxyquinoline Derivatives as Potent HIV-1 Non-Nucleoside Reverse Transcriptase Inhibitors

New target molecules, namely, 2-phenylamino-4-phenoxyquinoline derivatives, were designed using a molecular hybridization approach, which was accomplished by fusing the pharmacophore structures of three currently available drugs: nevirapine, efavirenz, and rilpivirine. The discovery of disubstituted quinoline indicated that the pyridinylamino substituent at the 2-position of quinoline plays an important role in its inhibitory activity against HIV-1 RT. The highly potent HIV-1 RT inhibitors, namely, 4-(2′,6′-dimethyl-4′-formylphenoxy)-2-(5″-cyanopyridin-2″ylamino)quinoline (6b) and 4-(2′,6′-dimethyl-4′-cyanophenoxy)-2-(5″-cyanopyridin-2″ylamino)quinoline (6d) exhibited half-maximal inhibitory concentrations (IC50) of 1.93 and 1.22 µM, respectively, which are similar to that of nevirapine (IC50 = 1.05 µM). The molecular docking results for these two compounds showed that both compounds interacted with Lys101, His235, and Pro236 residues through hydrogen bonding and interacted with Tyr188, Trp229, and Tyr318 residues through π–π stacking in HIV-1 RT. Interestingly, 6b was highly cytotoxic against MOLT-3 (acute lymphoblastic leukemia), HeLA (cervical carcinoma), and HL-60 (promyeloblast) cells with IC50 values of 12.7 ± 1.1, 25.7 ± 0.8, and 20.5 ± 2.1 µM, respectively. However, 6b and 6d had very low and no cytotoxicity, respectively, to-ward normal embryonic lung (MRC-5) cells. Therefore, the synthesis and biological evaluation of 2-phenylamino-4-phenoxyquinoline derivatives can serve as an excellent basis for the development of highly effective anti-HIV-1 and anticancer agents in the near future.


Introduction
Quinoline derivatives are an important class of heterocycles that exist among the principal components of natural products [1,2]. Quinoline is widely used as a dominant compound to synthesize molecules with medical benefits. Quinolines are used as anti-cancer, antimycobacterial, antimicrobial, anticonvulsant, anti-inflammatory, and cardiovascular agents [3][4][5]. HIV-1 reverse transcriptase (RT) is an important enzyme involved in retroviral replication and represents an important target for the development of anti-HIV drugs. Highly active antiretroviral therapy (HAART) has provided substantial progress in the treatment of acquired immunodeficiency syndrome (AIDS). HAART relies on three inhibitors, such as reverse transcriptase (RT) and protease enzymes, to sufficiently control HIV infection [6][7][8]. Non-nucleoside reverse transcriptase inhibitors (NNRTIs) play an important role in HAART because of their unique antiviral activity, high specificity, and low
Next, the molecular properties of the ligands were calculated and applied to predict the solubility of the medicated compounds and their possibility of improving the efficiency of action. The lipophilicity of a compound dictates its partition coefficient, which is the value of the tendency of a compound to partition into a nonpolar lipid phase from an aqueous phase. The partition coefficient is an important determinant of medicinal properties and is a rapid and effective tool for assessing initial drug viability [16]. Based on the Lipinski's rule of five, the compounds can be used as drug candidates if they have a logP value of 0-3, because these values are optimal for the distribution of compounds across cell membranes in the body systems [17,18]. Moreover, the total polar surface area (TPSA) is used to determine the sum of all polar atoms on the surface of molecules, particularly oxygen, nitrogen, and the attached hydrogen atoms, and is applied to predict the optimization of the medicament's ability to permeate cells. The appropriate molecules that can be used as candidate medicaments have permeability across cell membranes and the blood-brain barrier with TPSA values of less than 140 and 90 Å 2 , respectively [19][20][21]. The calculation result is shown in Table 2.    Lipinski's rule of five states that molecular weight must be less than 500 Da, lipophilicity (logP) must be less than 5, H-bond donors must be less than 5, H-bond accepter must be less than 10. Other parameters were considered. For example, total polar surface area (TPSA) should be less than 140 Å, and the number of rotatable bonds should be less than 10. The results implied that the 2,4-disubstituted quinolines (4a-4d, 5a-5d, and 8a-8d) and biquinolines (7a and 7b) have logP values higher than 3, which may affect dissolution and absorption. The results of molecular docking analysis revealed that each set of compounds was not completely overlapped, especially 4a-4d, 8a-8d, and 7a-7b ( Figure 2, see also Supplementary Materials, Figures S1-S9). Moreover, the number of H-bond donors was 0 for 4a-4d, and the TPSA values of 4a-4d, 8a-8d, 7a, and 7b were lower than the previous reported values of 5a-5d. Compounds 4a-4d, 8a-8d, 7a, and 7b were not selected for future development because of the evidence mentioned above. At this point, we concluded that these compounds have low hydrophilicity; the logP values were higher than 3 for 4a-4d, 5a-5d, 8a-8d, 7a, and 7b, and the TPSA values for 4a-4d, 7a, and 7b were lower than 70 Å 2 . Moreover, 5b, 5c, and 8b had moderate activity against HIV-1 RT that was not similar or equal to that of currently available drugs ( Figure 1). Next, 6a-6d were designed, and 2-amino-5-cyanopyridine was applied as the substituent at the 2-position instead of 4-aminobenzonitrile in order to increase the number of N atoms in the structure. From the previous study, the molecular docking and interaction results between 5a-5d and HIV-1 RT demonstrated that these compounds aligned in an identical location and bound to HIV-1 RT on Lys101 and His235 residues using hydrogen bonding and on TYR318 residues using π-π stacking in an allosteric site. A study in the literature reported that Lys101 is an amino acid residue with a key importance for the inhibitory activity against HIV-1 RT [22]. The binding energy, the number of conformations, and the interaction between the designed and synthesized compounds and amino acid residues on HIV-1 RT from the molecular docking results (Table 1) revealed similar results for 6a-6d to those for 5a-5d. The calculation results of the pharmacokinetic parameters (Table 2) show that 6a-6d demonstrated logP values in the range of 0-3, lower than those of 4a-4d, 5a-5d, 8a-8d, 7a, and 7b, and exhibited TPSA values higher than those of 4a-4d, 5a-5d, 8a-8d, 7a, and 7b, which implied that these compounds may tend to perform well in permeating in cell membranes and penetrating in the blood-brain barrier. Moreover, the overlaying of 6a-6d exhibited that these compounds completely overlapped and aligned in the cavity pocket of HIV-1 RT, and the binding area was similar to those of 4a-4d, 5a-5d, and 8a-8d as shown in Figure 2. Then, the 2-pyridinylamino-4-phenyloxyquinoline derivatives 6a-6d were synthesized according to the method for 5a-5d, except that 2-amino-5-cyanopyridine was used for coupling instead of 4-aminobenzonitrile. Compounds 6a-6d were evaluated for their inhibitory activity (%) against HIV-1 RT. Compounds 6b and 6d exhibited higher inhibitory activities than 5b and 5d when compared with the same substituent function. Compounds 6b and 6d exhibited inhibitory activities against HIV-1 RT with inhibition rates of 44.5 ± 2.8 and 45.1 ± 1.5, respectively, at 1 µM concentration; these values were similar to that of NVP (53.4 ± 2.7). Table 1 shows that 6b and 6c interacted in the pocket of HIV-1 RT via the LYS101, HIS235, and PRO236 residues using hydrogen bonding and via the TYR188, TRP229, and TYR318 residues using π-π stacking. The numbers of interactions of 6b and 6d with HIV-1 RT were higher than those of 4a-4d, 5a-5d, 7a, 7b, and 8a. However, 8a-8d demonstrated moderate efficacy in inhibiting HIV-1 RT and had low productivity in the synthesis process. The IC50 values of 6b and 6d were then deter-mined. The results of inhibitory analysis of 6b and 6d compared with the currently available drugs are presented in Table 3. Compounds 6b and 6d exhibited IC50 values of 1.93 and 1.22 µM, respectively, similar to that of NVP (IC50 = 1.05 µM). RPV is an NNRTI that is usually used for HAART in patients with HIV infection. RPV has a high specificity against HIV-1 RT with low cytotoxicity in humans [23]. In our research, RPV was one of the three structures used in the molecular hybridization approach to evaluate novel compounds as potent NNRTIs. The cyanovinyl substituent, which was the same as the side chain in the structure of RPV, was designed and applied to 5b, 6b, and 8b. Then, the synthesized compounds 9b, 10b, and 11b were obtained from 5b, 6b, and 8b, respectively, using diethyl cyanomethyl phosphonate in basic conditions, as demonstrated in Scheme 3. The products were obtained in 56%-64% yield from 5b, 6b, and 8b and consisted of trans-and cis-forms in a ratio of 7:3. However, the mixture of 11b could not be separated into each form. Compounds 9b, 10b, and 11b showed lower inhibitory activities against HIV-1 RT compared with 5b, 6b, and 8b. Moreover, the inhibitory activities of the trans-and cis-forms were not much different. In this studies, the binding energy values between designed compounds and HIV-1 RT as shown in Table 1 demonstrated that the calculated binding energy of NVP is found to be higher than that of RPV because the size of the NPV core structure is larger than the binding pocket of RPV in the HIV-1 RT enzyme. This could lead to the repulsive interactions inside the binding pocket and eventually increasing the binding energy. In the case of the de-signed compounds, their quinoline-based core structures are slightly bigger than that of RPV but they still fit in very well with the binding pocket. Besides, the core structure and the side chains of the designed compounds can also interact very well with amino acid residues inside the HIV-1 RT binding pocket. As a result, their binding energy with the enzyme is lower or comparable to that of RPV. Although EFV displays a similar size of core structure to the synthesized compounds, they have different substituents which could lead to different interactions and binding energy. As shown, EFV only exerts one H-bond with the amino acid residue in the HIV-1 RT pocket. Therefore, its binding energy is found to be higher than that of the synthesized compounds. It is worth mentioning that the obtained binding energy is solely predicted based on the calculation via a molecular docking technique. Further examination of the biological activity and the effect of 2,4-disubstituents of the synthesized compounds are still needed.  Additionally, 4-8(a-d), (9)(10)(11)b, and the three common drugs (NVP, EFV, and RPV) were evaluated for their cytotoxicity against various cancer cell lines, as shown in Tables 4 and 5. In the literature, NNRTIs, namely, NVP, EFV, and RPV, have been report-ed as being toxic against a wide range of cancer cells in vitro [24][25][26] but only have minor toxicity against normal tissue cells [24]. The toxicity of NNRTIs against cancer cells promoted the idea that these drugs can be used to prevent or even treat HIV-1 infection and cancer. All of the synthesized compounds that showed activity against HIV-1 RT were cytotoxic to cancer cell lines. The results of the cytotoxicity assay revealed that 6b, EFV, and RPV had strong activities against acute lymphoblastic leukemia (MOLT-3) cells (IC50 = 12.7 ± 1.1, 24.6 ± 1.2, and 4.3 ± 0.5 µM, respectively), cervical carcinoma (HeLA) cells (IC50 = 25.7 ± 0.8, 46.2 ± 1.3, and 11.3 ± 1.2 µM, respectively), and promyeloblast (HL-60) cells (IC50 = 20.5 ± 4.1, 33.7 ± 1.1, and 11.5 ± 0.8 µM, respectively). Compounds 6b, EFV, and RPV had substantial cytotoxicity against the cancer cell lines, whereas 6b and RPV had very low cytotoxicity against normal embryonic lung (MRC-5) cells. However, 6d had high inhibitory activity with a very low IC50 value against HIV-1 RT and had no cytotoxicity against normal embryonic lung (MRC-5) cells compared with EFV and RPV. All of the evidence brought us to the conclusion that the substituents of quinoline at the 4-position, namely, 2 ,6 -dimethyl-4 -cyanophenoxy and 2 ,6dimethyl-4 -formylphenoxy, and the substituent at the 2-position, namely, 5 -cyanopyridin-2 ylamino, had inhibitory activities against HIV-1 RT. Thus, 6b and 6d are candidates for the development of HIV treatment in the near future.

Molecular Docking Studies
The crystal structure of HIV-1 (4G1Q) [27] was obtained from the Protein Data Bank. 4G1Q is a crystal structure of HIV-1 RT in a complex with RPV, which is a commercial NNRTI drug. The protein was prepared by removing the water molecules, ligand, and other unnecessary small molecules from the crystal structure of the ligand-HIV-1 RT complex (PDB code: 4G1Q) before molecular docking. The geometry of quinoline deriva- tives, 4-7(a-d) and 8-10(b), was fully optimized using the density functional theory at B3LYP/6-31G (d, p) level implemented in Gaussian 09 [28]. The binding inter-actions of 2,4-diphenoxyquinoline (4[a-d]), 2-phenylamino-4-phenoxy-quinoline (5-6[a-d]), 4,4 -di-(4 -formylphenoxy)-2,2 -biquinoline (7a), 4,4 -di-(2 ,6 -dimethyl-4 -formylphenoxy)-2,2biquinoline (7b), and 2-phenoxy-4-phenylamine-quinoline (8 [a-d], 9-11[b]) with HIV-1 RT were simulated by molecular docking using AutoDock 4.2 [29]. Default AutoDock settings were used; the population size for the Lamarckian genetic algorithm was set to 150 individuals, and the number of genetic algorithm was set to 200. The maximum number of evaluations was 2,500,000, and the maximum number of generations was 27,000. Blind docking was carried out using a grid box with a size of 80 × 80 × 80 A • along the x, y, and z axes, respectively, to observe the binding sites of inhibitors in HIV-1 RT. The grid center for HIV-RT was located at x = 49.082, y = −28.29, and z = 37.541, and the spacing was 0.375 A • . Discovery Studio 4.0 software was applied to visualize the lowest energy conformation. The identification of ligand binding modes was concluded by iteratively evaluating a number of ligand conformations and estimating the energy of their inter-actions with the target. All experiments were accomplished in triplicate, and the replicates showed similar results.
General procedure for the preparation of 2,4-diphenoxyquinoline (4a-4d). A mixture of 1 (10 mmol) and hydroxyl benzene (21 mmol) in DMF (30 mL) with anhydrous Cs 2 CO 3 (20 mmol) was heated in a sealed tube, stirred at 120 • C for 8-16 h, and cooled. Afterward, the mixture was poured into ice-water and extracted thrice with ethyl acetate. The combined organic layers were washed with saturated NaCl and dried over Na 2 SO 4 . The crude product was purified on a silica gel column (eluent: hexane/ethyl acetate) to obtain 4a-4d with 55-65% yield.
General procedure for the preparation of 9b, 10b, and 11b. Potassium tert-butoxide (1.50 mmol) was added to an ice-cooled solution of diethyl cyanomethyl phosphonate (1.50 mmol) in THF (20 mL). The mixture was stirred at 0 • C for 30 min and then at room temperature for another 30 min. A solution of 4-(4 -formylphenoxy)-2-arylaminoquinoline (5b and 6b) or 2-(4 -formylphenoxy)-4-phenylamino-quinoline (8b, 1 mmol) in THF (13 mL) was added dropwise to the reaction mixture. The solution was continued for 8-10 h. After the reaction was completed, the corresponding solution was added with water and extracted with ethyl acetate. The organic layer was dried over Na 2 SO 4 and concentrated under reduced pressure. The residue was purified on a silica gel column (eluent: hexane/ethyl acetate) to obtain 9b, 10b, and 11b with 56%-64% yield. The E:Z isomer ratio for 9b and 10b was 7:3, whereas 11b was inseparable.  buffer. Finally, the peroxide substrate (ABST) was added to the MTP. A colored reaction product emerged during the cleavage of the substrate catalyzed by peroxide enzyme. The absorbance of the sample was measured at the optical density (OD) at 490 nm using the MTP (enzyme-linked immunosorbent assay [ELISA]) reader. The percentage inhibitory activities of the RT inhibitors were calculated by comparing with that of the sample without an inhibitor. The resulting color intensity was directly proportional to the RT activity. Percentage inhibitory values were calculated using the following formula: Inhibition (%) = [1-(OD value with RT and inhibitor−OD value without RT and inhibitors)/(OD value without inhibitors with RT−OD value without RT and inhibitors)] × 100. Half maximal inhibitory concentration (IC50) was determined as the concentrations of the compounds of interest and the three control drugs at 50% cell growth inhibition using the plotting of the sigmoid curve between the log of the concentration on the x axis and the inhibition rate on the y axis.

Cytotoxic Activity
The cell lines were seeded in a 96-well microplate (Costar No. 3599) at a density of 5 × 10 3 − 2 × 10 4 cells/well (100 µL/well). Background control wells contained the same volume as the complete culture medium. The microplate was incubated for 24 h at 37 • C with 5% CO 2 and 95% humidity (Shellab). Samples at various concentrations were added to the microplate and incubated for another 48 h. Cell viability was determined by 3(4,5-dimethylthiazol-2-yl)-2-5-diphenyl tetrazolium bromide (MTT) assay (Sig-ma-Aldrich) [36,37]. The reagent was dissolved in phosphate-buffered saline at 5 mg/mL and filtered to sterilize and remove the small amount of insoluble residue present in some batches of MTT. MTT solution (10 µL/100 µL medium) was added to all wells of each assay, and the plates were incubated at 37 • C with 5% CO 2 and 95% humidity for 2-4 h. Subsequently, dimethyl sulfoxide (100 µL; Merck, Germany) was added to dissolve the resulting formazan by sonication. The plates were read on a microplate reader (Molecular Devices, CA, USA) using a test wavelength of 550 nm and a reference wavelength of 650 nm. XTT assay for suspension cells was used for MOLT-3 cells [36]. The plates were incubated for 4 h after the addition of a 50-µL mixture of 1 mg/mL (5 mL) and 0.383 mg/mL (100 µL) phenazine methosulfate. The absorbance of the orange formazan compounds was measured at the wavelengths of 492 and 690 nm. IC50 values were determined as the drug and sample concentrations at 50% cell growth inhibition.