Synthesis and Molecular Docking Study of Novel Pyrimidine Derivatives against COVID-19

A novel series of pyrido[2,3-d]pyrimidines; pyrido[3,2-e][1,3,4]triazolo; and tetrazolo[1,5-c]pyrimidines were synthesized via different chemical transformations starting from pyrazolo[3,4-b]pyridin-6-yl)-N,N-dimethylcarbamimidic chloride 3b (prepared from the reaction of o-aminonitrile 1b and phosogen iminiumchloride). The structures of the newly synthesized compounds were elucidated based on spectroscopic data and elemental analyses. Designated compounds are subjected for molecular docking by using Auto Dock Vina software in order to evaluate the antiviral potency for the synthesized compounds against SARS-CoV-2 (2019-nCoV) main protease M pro. The antiviral activity against SARS-CoV-2 showed that tested compounds 7c, 7d, and 7e had the most promising antiviral activity with lower IC50 values compared to Lopinavir, “the commonly used protease inhibitor”. Both in silico and in vitro results are in agreement.


Introduction
Human health is one of the most important factors impacting economic development in the world. So, studying diseases and their potential treatment is an essential research area. The coronavirus disease 2019 (COVID-19) has imposed a great threat globally due its rapid spreading and mutation. Virus-encoded proteases are observed as potential drug targets. The main protease (Mpro) of SRAS-CoV-2 plays a crucial role in the viral maturation.
Studies have reported that the inhibition of Mpro can prevent the virus from replication [1][2][3][4][5]. At present, some medications and vaccines have been approved by the FDA for the treatment of COVID-19 but they could not completely eradicate the virus. The search for safe and effective antivirals today is urgent.

Chemistry
The requisite starting material 6-amino-pyrazolo [3,4-b]pyridine-5-carbonitrile 1 was used in this study. It was synthesized according to the known procedure [33] from pyrazolone, arylaldehyde and malononitrile (Scheme 1). Phosgene iminium chloride (PIC = dichloromethylene-dimethyliminium chloride) is an efficient reagent in synthetic chemistry and was used to introduce the amide chloride group into activated substrates. So, stirring of compound 1a,b with phosgenimonium chloride in 1,2-dichloroethane at room temperature provided pyridopyrimidine derivative 2a,b and Pyrazolopyridinyl derivative 4b, respectively, through the pathway outlined in Scheme 1. The structures of 2a,b and 4b were confirmed by elemental analysis and spectral (MS, IR and 1 H NMR) data (Supplementary Materials) (c.f. experimental). Their 1 H NMR spectrum in DMSO revealed in each case a characteristic signal in the region of δ 3.10-3.27 assignable to the -N(CH 3 ) 2 proton. Their IR spectra showed the absence of the characteristic band for nitrile group for compounds 2a,b which is present for compound 4b at 2217 cm −1 . Additionally, a new peak at ν 735 cm −1 for the new C-Cl appeared. Stirring of the compounds 2a,b and 4b with hydrazine hydrate in ethanol afforded the corresponding 5-imino-pyrazolo [4 ,3 :5,6]pyrido [2,3-d]pyrimidine derivatives 3a,b and 5-hydrazineylpyrazolo [4 ,3 :5,6]pyrido [2,3-d]pyrimidine derivative 6b, respectively (Scheme 1). Both spectroscopic data and elemental analyses were consistent with the assigned structures (c.f., experimental). IR spectra of compounds 6b as an example indicated the disappearance the absorption band characteristic for CN and C-Cl groups with the appearance of peaks corresponding to -NH 2 and NH groups. In the 1 H NMR spectrum (Supplementary Materials) and in the presence of D 2 O, the signals due to the NH and NH 2 groups in each compound disappeared and the signals related to methyl, aryl, and N(CH 3 ) 2 protons appeared at the expected values.
Refluxing of hydrazine derivative 3a with aldehydes and/or monosaccharaides in ethanol, with adding a few drops of acetic acid as a catalyst, gave the corresponding hydrazones derivatives 7c-g, respectively (Scheme 2). The structures of the latter products were confirmed by elemental and spectral analyses. The IR spectrum of the subsequent sugar hydrazones 7f,g as an example indicated the characteristic bands for the hydroxyl groups at (3415-3419) cm −1 in the sugar chain. The 1 H NMR (c.f. experimental) spectra of compound 7g,h demonstrated the signs of H-1 methine proton showing up as doublet at δ 7.17 and 7.20 ppm, respectively. Their 13 C NMR spectra (Supplementary Materials) of compound 7f revealed five resonances for sugar carbon atoms at δ 61.63, 70.59, 73.75, 75.48, 77.22 ppm. Treatment of each of the hydrazones 7c,d with iron (III) chloride in ethanol or with potassium iodide in DMSO gave, in each case, a single product as evidenced by TLC analysis. Elemental analyses and mass spectra revealed that each of such isolated products has two hydrogens less than the respective hydrazone. This finding was confirmed by the 1 H NMR spectra, which indicated the absence of the -N=CH-and hydrazone-NH-N=C protons. Based on this finding, the isolated products were assigned the structure 8c,d (Scheme 2). The conversion of 7 into 8 is reminiscent of other related oxidative cyclization of aldehyde N-heteroarylhydrazones with iron(III) chloride, which have been reported to proceed via generation of the respective nitrilimines, which undergo in situ 1,5-electrocyclization to give the respective fused heterocycles [43,44]. Similarly, refluxing of amino derivative 6b with 4-chlorobenzaldehyde and 4-bromobenzaldehyde and under the same reaction conditions afforded the products 10c,d via 9c,d (Scheme 3).  When the pyrazolo[4,3-e] [1,2,4]triazolo [4,3-c]pyrimidine derivative 6b was heated in ethanol in the presence of sodium acetate, it gave a product which proved identical in all respects (mp, mixed mp, IR and 1 H NMR spectra) with 3b. It has been found that compound 6b isomerized to the thermodynamically more stable pyrazolo [4,3-e] [1,2,4] [1,5-c] pyrimidine derivative 3b through tandem ring opening and ring closure reactions (Scheme 4). This rearrangement is consistent with those reported in some earlier reports [45,46]. Compound 3a was utilized for synthesis of several new compounds. So, product 11 was obtained by stirring 3a with sodium nitrite in acetic acid. Compound 3a reacted with carbon disulfide to yield triazolopyrimidinethione 12. Compound 3a was converted into compound 13 by reaction with triethyl orthoformate. Dimethyldichloromethylenirninium chloride condensed with 3a to yield 14 (Scheme 5). Structural assignments of all compounds 11-14 were based on their elemental analyses and spectroscopic data (c.f., experimental).
Based on anti-viral assay, we would endorse compound 7c for further detailed investigation, as it showed the best selectivity index (130), which is even better than the Lopinavir drug.

Molecular Docking
To pick up the mode of action of the tested compounds, a molecular-docking study was utilized to determine the binding modes against SARS-CoV-2 main protease "M pro ", which are significant targets to create anti-SARS-CoV2 agents. These targets were chosen dependent on their key roles in viral protein formation; therefore, targeting these proteins give potential advantages in killing the virus. The co-crystalized ligand carbamate" was re-docked to guarantee the validity of the docking parameters and methods using Auto Dock vina to represent the position and orientation of the ligand recognized in the crystal structure. The distinction of RMSD value between co-crystalized ligands to the original co-crystal ligand was <2 Å, which affirmed the accuracy of the docking protocols and parameters. As reference ligand, the co-crystalized ligand interacted with the active site of the M pro protein via 5 hydrogen bonds, with the active amino acid residues (THR26, GLU166, CSY145, CYS145, SER144) with bond distance (3.139, 3.299, 2.890, 2.959, 3.023, respectively), and the value of the free binding energy was -7.5 Kcal/mol (Tables 2 and 3) (Figure 3). Comparing our tested compounds based on their binding to M pro protein, 6 compounds (7c, 7d, 7e, 7f, 10c, and 10d) showed preferable binding affinity to the M pro protein more than the co-crystalized ligand did, which was evidenced by the lower values of their free binding energy (−8.4, −8.3, −8.5, −8.0, −8.1, and −8.1 Kcal/mol, respectively), and by the hydrogen bond interaction with the key amino acid residues in the active site of the M pro protein (Tables 2 and 3) (Figure 4). Based on the docking results, we have selected the most promising tested compounds, and ordered them according to their activities (7e > 7c > 7d > 7f >10c = 10d). Therefore, these compounds may have interesting applications against the Alpha variant of SARS-CoV-2. It is encouraging to expand this series via the synthesis of more analogues in an optically pure form, and to test them all experimentally.

Structure Activity Relationships
The structure-based activity analysis (SAR) of synthesized compounds revealed the compounds having an electron-withdrawing group bonded to the phenyl ring as shown in 7c (IC 50 = 1.2), 7d (IC 50 = 2.34), and an electron-donating group bonded to the phenyl ring as observed in 7e (IC 50 = 2.3), displaying promising antiviral activity against SARS-CoV-2 with wider selectivity indices (SI = 71-130). The presence of a sugar moiety in compounds 7f, 7g, 10f, and 10g did not result in increased antiviral activity. Compound 10c showed moderate activity.

General
All melting points were measured on a Gallenkamp Melting point apparatus and are uncorrected. The IR spectra were recorded on a Shimadzu FT-IR 8101 PC infrared spectrophotometer (Shimadzu, Tokyo, Japan) using KBr disks. The NMR spectra were preserved on a Varian Mercury VX-400 NMR spectrometer (Varian, Palo Alto, CA, USA). 1 H NMR spectra were run at 400 MHz and 13 C NMR spectra were run at 75.46 MHz in deuterated chloroform (CDCl 3 ) or dimethyl sulfoxide (DMSO-d 6 ) as specified in individual compound characterizations. Chemical shifts are given in parts per million and were referenced to those of the solvents. Mass spectra were recorded on a Shimadzu GCMS-QP 1000 EX mass spectrometer at 70 eV. Elemental analyses were registered on an Elementar-Vario EL (Germany) automatic analyzer.

General Procedure for the Synthesis of 7c-g and 9c,d
To a mixture of 1 (0.6 g, 2.5 mmol) and the appropriate aldose 2a-d (2.5 mmol) in ethanol (15 ml), a catalytic amount of glacial acetic acid (0.1 ml) was added To a mixture of derivatives 3a or 6b (10 mmol) and the appropriate aldehyde (10 mmol) or respective monosaccharides (10 mmol) in ethanol (30 mL), a catalytic amount of glacial acetic acid (1.0 mL) was added. The reaction mixture was refluxed for eight hours. After cooling at room temperature, the precipitated solid was collected by filtration and recrystallised from the proper solvent. Method B: 2 M solution of iron (III) chloride in ethanol (2 mL) was added dropwise to a boiling solution of (7c,d) or (9c,d) (10 mmol) in ethanol (50 mL). Heating was continued for 20 min and the mixture was then kept overnight at room temperature. The reaction was allowed to cool and the precipitate isolated by filtration and washed with a little methanol and recrystallized from dioxane.

Cytotoxic Concentration 50 (CC 50 ) and Viral Inhibitory Concentration 50 (IC 50 ) Calculation
The assay was performed according to the procedure that was previously described (Feoktistova, M., Geserick, P., Leverkus, M. 2016. Crystal Violet Assay for Determining Viability of Cultured Cells. In Cold Spring Harb Protoc, pdb.prot087379) with minor modifications. Vero E6 cells were seeded into 96-well plates in 100 µL of high glucose Dulbecco's modified Eagle's medium (DMEM) containing medium 10% fetal bovine serum (FBS), 100 units/mL penicillin, and 100 µg/mL streptomycin at 37 • C in 5% CO 2 . After 24 h (90-100% confluence monolayer of Vero E6), each compound was diluted into varying concentrations in a separate U-shape 96-well plate (with a range of concentration from 10 µg/mL to 1 ng/mL) using DMEM containing 2% FBS (maintenance medium). A volume of 100 µL of each dilution was transferred into a new U-shape 96-well plate and supplemented with 100 TCID 50 in 100 µL maintenance medium. In parallel, the wells dedicated for CC 50 calculation were supplemented with 100 µL maintenance medium without virus. Aliquots of 100 µL of infection media containing 100 TCID 50 were used as virus control. After 1 h of incubation, 100 µL of each well was transferred to the corresponding wells into the 96-well plates containing Vero E6 cultures. The plates were incubated for 72 h, the cell monolayers were washed with PBS and subjected to cell fixation using 100 µL of 10% formalin for 1 h. Subsequently, the plates are washed well 3 times with 1 × PBS and dried well before staining with 50 µL (0.5%) crystal violet to each well ((0.5 g crystal violet powder (Sigma-Aldrich), 80 mL distilled H 2 O and 20 mL methanol)) for 30 min. The plates were then washed well with rinsed water and air-dried at room temperature for 2 to 24 h. To distain crystal violet, 200 µL methanol was added to each well, and the plate was incubated with its lid on a bench rocker (20 oscillations/minute) for 20 min at room temperature. Finally, the optical density of each well at λ 590 nm (OD590) was measured with a plate reader. The average OD of each dilution without or with virus was compared to control cells and control virus wells to calculate CC 50 and IC 50 values using nonlinear regression in GraphPad Prism 5.01.

Molecular Docking Study
The structures of all tested compounds were modeled using the Chemsketch software (http://www.acdlabs.com/resources/freeware/, accessed on 15 December 2021). The structures were optimized and energy minimized using the VEGAZZ software [49]. The optimized compounds were used to perform molecular docking. The three-dimensional structures of the molecular target were obtained from Protein Data Bank (PDB) (www.rcsb. org, accessed on 15 December 2021): SARS-CoV-2 (2019-nCoV) main protease M pro (PDB: 6Y2F, https://www.rcsb.org/structure/6Y2F, accessed on 15 December 2021). The steps for receptor preparation included the removal of heteroatoms (water and ions), the addition of polar hydrogen, and the assignment of charge. The active sites were defined using grid boxes of appropriate sizes around the bound cocrystal ligands. The docking study was performed using Autodock vina [50] and Chimera for visualization [51].