Synthesis and In Silico Study of Some New bis-[1,3,4]thiadiazolimines and bis-Thiazolimines as Potential Inhibitors for SARS-CoV-2 Main Protease

A novel series of bis-[1,3,4]thiadiazolimines, and bis-thiazolimines, with alkyl linker, were synthesized through general routes from cyclization of 1,1′-(hexane-1,6-diyl)bis(3-phenylthiourea) and hydrazonoyl halides or α-haloketones, respectively. Docking studies were applied to test the binding affinity of the synthesized products against the Mpro of SARS-CoV-2. The best compound, 5h, has average binding energy (−7.50 ± 0.58 kcal/mol) better than that of the positive controls O6K and N3 (−7.36 ± 0.34 and −6.36 ± 0.31 kcal/mol). Additionally, the docking poses (H-bonds and hydrophobic contacts) of the tested compounds against the Mpro using the PLIP web server were analyzed.

bis-1,3-Thiazole derivatives were examined against different viruses (such as hepatitis B and C viruses, poliovirus, and influenza A virus) and showed promising results (up to EC50 0.56 µM) [12].
The main protease (M pro ) or 3CL Pro is a vital viral protein that is important for the SARS-CoV-2 life cycle. Its function is to process viral polyprotein upon entry and replication  The synthesis of bis- [1,3,4]thiadiazol-2-imines 5a-h was inaugurated by nucleophilic halogen displacement by the thiol group to afford intermediate 3. Intramolecular cyclization of the latter intermediate with consecutive elimination of aniline molecules furnished the isolated products in good yields (Scheme 1). The spectroscopic data (IR, 1 H-NMR, 13 C-NMR, and MS) and the elemental analyses of bis- [1,3,4]thiadiazol-2-imines 5a-h were in agreement with the assigned structures (see Experimental section).

Computational Analysis
Molecular docking combined with molecular dynamics simulation was utilized to test the binding affinity of starting material 1, bis- [1,3,4]thiadiazol-2-imines 5a-h, and bisthiazol-2-imines (8a-e and 11a,b) against the M pro of SARS-CoV-2. The solved structure used in this study is the dimeric M pro (PDB ID: 6Y2G), this is due to the fact that it is the active dimeric form of the M pro and solved with O6K inhibitor (tert-butyl (1- 13b)), so we extracted it and used it as positive control. Additionally, we used the other inhibitor N3

Computational Analysis
Molecular docking combined with molecular dynamics simulation was utilized to test the binding affinity of starting material 1, bis- [1,3,4]thiadiazol-2-imines 5a-h, and bis-thiazol-2-imines (8a-e and 11a,b) against the M pro of SARS-CoV-2. The solved structure used in this study is the dimeric M pro (PDB ID: 6Y2G), this is due to the fact that it is the active dimeric form of the M pro and solved with O6K inhibitor (tert-butyl (1- 3-yl)carbamate (alpha-ketoamide 13b)), so we extracted it and used it as positive control. Additionally, we used the other inhibitor N3 (N-[(5-methylisoxazol- 3yl) 3yl]methyl}but-2-enyl)-L-leucinamide) found in the solved structure 6LU7 to compare its affinity to M pro to that of bis- [1,3,4]thiadiazolimines and bis-thiazolimines against the active site of the M pro .
The docking was performed after equilibrating the structure for an 80 ns MDS run. The reason for performing MDS before docking is that we remove ligands and want the protein to be relaxed before making the docking calculations. We also equilibrate the system and cluster the trajectories to test different possible protein conformations against the compounds. Figure 1A reflects the system equilibration as the root-mean-square deviation (RMSD) curve (blue) is flattened around 2 Å. Additionally, the radius of the gyration (RoG) curve (red) is flat during the simulation period, averaging about 26 Å. Figure 1B shows the five representative models for M pro after clustering the trajectories depicted in colored cartoons. Additionally, the per-residue root-mean-square fluctuations (RMSF) in Å are plotted. The active site residues (H41 and C145) are marked on the RMSF curve. In addition to the protein termini, only one region (S46-M49) shows moderate fluctuations (RMSF < 1.5 Å). This region is colored in red in the structures. The active site residues are away from this region and show low fluctuations (RMSF > 0.7 Å).
The bar graph of Figure 2A shows the average (out of five values) binding energy (in kcal/mol) against the M pro of different SARS-CoV-2 conformations after MDS trajectory clustering. The error bars represent the standard deviation which was calculated by the Microsoft Excel formula ∑(x−x) 2 n−1 . Before the docking of the compounds, we tested the docking system by performing a redocking of O6K to the solved structure and we obtained a root-mean-square deviation of 0.92 Å (1806 fitted atoms). The positive controls O6K and N3 (putative inhibitors to M pro found in the solved structures 6Y2G and 6LU7) are shown in red and orange, respectively, while the compounds 1, 5a-h, 8a-e, and 11a,b are in blue ( Figure 2A). The best compound, 5h, is shown in green. The compound 5h has average binding energy (−7.50 ± 0.58 kcal/mol) better than that of the positive controls O6K and N3 (−7.36 ± 0.34 and −6.36 ± 0.31 kcal/mol). Despite the fact that the binding affinity value of 5h compound is not significantly different compared to the positive control, it is still a potential anti-SARS-CoV-2 M pro compound that deserves further investigation. Figure 2B shows the detailed interaction pattern between 5h and the M pro . Five hydrophobic contacts are formed between 5h and the residues F140, E166(2), P168, and Q189 of M pro . Additionally, the other tested compounds (except compound 1) present enhanced binding energies against M pro (−6.38 ± 0.52 down to −7.18 ± 0.29 kcal/mol for 11b and 8e, respectively) compared to the positive control N3. This reflects their potential to be tightly bound to and might inhibit the M pro of SARS-CoV-2.
Furthermore, we analyzed the docking poses of the compounds against the M pro using the PLIP web server, and a detailed list is tabulated ( Table 1). The ligand that has the nearest binding affinity to the average value is represented. These modes resemble five different protein conformations after an 80 ns MDS run. For the compounds 5a, 5b, 5c, 5d, 5e, 5f, 5g, 11a, and 11b, the most common interactions are the H-bonds (three) and the hydrophobic contacts (three). On the other hand, only hydrophobic contacts (four) are reported in compounds 5h, 8a, 8b, 8c, 8d, and 8e, while compound 1 formed only two H-bonds with the M pro . Compounds 5c, 5d, and 5g form a halogen bond with R188 or T190 residues of the M pro . The highest contributing residues in the formed interactions between the compounds and the Mpro are Q189, E166, N142, G143, S144, and C145, which formed 14, 11, 8, 8, 8, and 5 interactions, respectively. shows the five representative models for M pro after clustering the trajectories depicted in colored cartoons. Additionally, the per-residue root-mean-square fluctuations (RMSF) in Å are plotted. The active site residues (H41 and C145) are marked on the RMSF curve. In addition to the protein termini, only one region (S46-M49) shows moderate fluctuations (RMSF > 1.5 Å). This region is colored in red in the structures. The active site residues are away from this region and show low fluctuations (RMSF < 0.7 Å). tion. Figure 2B shows the detailed interaction pattern between 5h and the M pro . Five hydrophobic contacts are formed between 5h and the residues F140, E166(2), P168, and Q189 of M pro . Additionally, the other tested compounds (except compound 1) present enhanced binding energies against M pro (-6.38 ± 0.52 down to −7.18 ± 0.29 kcal/mol for 11b and 8e, respectively) compared to the positive control N3. This reflects their potential to be tightly bound to and might inhibit the M pro of SARS-CoV-2.  Furthermore, we analyzed the docking poses of the compounds against the M pro using the PLIP web server, and a detailed list is tabulated ( Table 1). The ligand that has the nearest binding affinity to the average value is represented. These modes resemble five different protein conformations after an 80 ns MDS run. For the compounds 5a, 5b, 5c, 5d, 5e, 5f, 5g, 11a, and 11b, the most common interactions are the H-bonds (three) and the hydrophobic contacts (three). On the other hand, only hydrophobic contacts (four) are reported in compounds 5h, 8a, 8b, 8c, 8d, and 8e, while compound 1 formed only two H-  The binding affinities are listed among the number of H-bonds, halogen bonds, and hydrophobic contacts and the residues that take part in their formation in Table 1. Bold residues are the common residues interacting with the ligands, while the active site dyads are underlined. These residues are essential in the protease function as they lie within the active site pocket. When some small molecules block this pocket, it may interfere with the protease function ( Figure 3A), which is yet to be verified experimentally. The best compound 5h (magenta sticks) is fitted in the active site pocket similar to the positive control O6K (green sticks) ( Figure 3B). This indicates its possible usefulness as a SARS-CoV-2 M pro inhibitor. According to the SwissADME web tool (http://www.swissadme.ch/index.php) (accessed on 6 June 2016), compound 5h has poor solubility, has low gastrointestinal absorption, has a 674.84 g/mol molecular weight, 0.17 bioavailability score, and does not have adverse pharmacokinetics properties.
protease function (Figure 3A), which is yet to be verified experimentally. The best compound 5h (magenta sticks) is fitted in the active site pocket similar to the positive control O6K (green sticks) ( Figure 3B). This indicates its possible usefulness as a SARS-CoV-2 M pro inhibitor. According to the SwissADME web tool (http://www.swissadme.ch/index.php) (accessed on 6 June 2016), compound 5h has poor solubility, has low gastrointestinal absorption, has a 674.84 g/mol molecular weight, 0.17 bioavailability score, and does not have adverse pharmacokinetics properties.  Table 1). (B) The superposition of the M pro (blue cartoon) is complexed with O6K (green sticks) and 5h (magenta sticks) for comparison. The active site pocket residues are labelled with its 1-letter codes.
We trust our simulation as the RMSD of our trajectory coincides with that of the large simulation trajectory found at the following link: (https://figshare.com/search?q=10.6084%2Fm9.figshare.12009789) (accessed on 23 March 2020). Figure 4A shows the RMSD of our simulation trajectory (80 ns) in blue and the 1 Figure 3. (A) The most reported interacting residues of SARS-CoV-2 Mpro (green cartoon) with the ligands are represented using the solved structure (PDB ID: 6LU7). The active site dyads (H41 and C145) are represented by red sticks, while the yellow sticks represent the most reported interacting residues, N142, G143, S144, E166, and Q189 (underlined in Table 1). (B) The superposition of the M pro (blue cartoon) is complexed with O6K (green sticks) and 5h (magenta sticks) for comparison. The active site pocket residues are labelled with its 1-letter codes.
We trust our simulation as the RMSD of our trajectory coincides with that of the large simulation trajectory found at the following link: (https://figshare.com/search?q=10.608 4%2Fm9.figshare.12009789) (accessed on 23 March 2020). Figure 4A shows the RMSD of our simulation trajectory (80 ns) in blue and the 1 microsecond trajectory in red. Additionally, we performed blind docking of the compounds utilizing AutoDock Vina implemented in PyRx software against both systems (after trajectory clustering) and found comparable results. The compounds show comparable results to positive controls, and 5h was the best based on the average binding affinities in both systems.

Experimental
Melting points were measured on an Electrothermal IA 9000 series digital melting point apparatus. IR spectra were recorded on Pye Unicam SP 3300 and Shimadzu FTIR 8101 PC infrared spectrophotometers. NMR spectra were recorded on a Varian Mercury VX-300 NMR spectrometer operating at 300 MHz ( 1 H-NMR) and 75 MHz ( 13 C-NMR) and run in deuterated dimethylsulfoxide (DMSO-d6). Chemical shifts were related to that of the solvent. Mass spectra were recorded on a Shimadzu GCeMS-QP1000 EX mass spectrometer at 70 eV. Elemental analyses were measured by using a German made Elementar vario LIII CHNS analyzer.

Experimental
Melting points were measured on an Electrothermal IA 9000 series digital melting point apparatus. IR spectra were recorded on Pye Unicam SP 3300 and Shimadzu FTIR 8101 PC infrared spectrophotometers. NMR spectra were recorded on a Varian Mercury VX-300 NMR spectrometer operating at 300 MHz ( 1 H-NMR) and 75 MHz ( 13 C-NMR) and run in deuterated dimethylsulfoxide ( . Chemical shifts were related to that of the solvent. Mass spectra were recorded on a Shimadzu GCeMS-QP1000 EX mass spectrometer at 70 eV. Elemental analyses were measured by using a German made Elementar vario LIII CHNS analyzer.

N,N -(Hexane-1,6-diyl)bis[3-phenyl-4-(p-tolyl)thiazol-2(3H)-imine] (8b).
Yellow solid (79%); m.p. 169 and C145) protocol [39,40]. SARS-CoV-2 main protease M pro dimer structure (PDB ID: 6Y2G) was downloaded from the protein data bank (https://www.rcsb.org/) (accessed on 20 March 2020) [27]. O6K, α-ketoamide inhibitor, is the positive control molecule that was solved in the M pro structure and was used to examine the affinity of the compounds 1, 5a-h, 8a-e, and 11a,b to the M pro active site. O6K is a covalently bound ligand, but we redock it to the solved structure in a non-bonded fashion, and it gives a root-mean-square difference of 0.966 Å. Additionally, the peptidyl Michael acceptor, N3, found in the solved structure of Mpro (PDB ID: 6LU7) is also used as a positive control for comparison [30]. The M pro dimer was subjected to molecular dynamics simulation (MDS) which ran for 80 nanoseconds, as reported [41] before the docking study. The MDS was conducted on the WEBGRO macromolecular simulation utilizing GROningen MAchine for Chemical Simulations (GROMACS) software and CHARMM27 force field [42,43]. A minimization for 10,000 steps of the conjugate gradient is performed before the MDS run. TIP4P water model was used with a constant number of atoms, volume, and temperature (NVT) ensemble in cubic periodic boundary conditions. Na + and Cl − were added to the system for the salt concentration of 154 mM, while the temperature and pressure were adjusted to be 310 K and 1 atm, respectively, to resemble physiological conditions. Clustering of the trajectories was performed using the UCSF Chimera 1.14 software [44]. A representative structure from each cluster was used when testing the binding affinity using the AutoDock Vina software [39,45].
The search box was adjusted to cover the active site dyad (H41 and C145) with dimensions of 30 Å × 30 Å × 30 Å centered at (25.2, 47.3, 38.4) Å. Protein-ligand interaction profiler (PLIP) webserver (https://plip-tool.biotec.tu-dresden.de/plip-web/plip/index) (accessed on 5 May 2021) was utilized to check the binding modes and the data are tabulated and then represented using PyMOL software 2.0.4 in the results section [46][47][48]. In the current study, we used an exhaustiveness value of 100. This is due to the many rotatable bonds we have in some ligands.

Conclusions
The present study disclosed the preparation of 1,1 -(hexane-1,6-diyl)bis( 3-phenylthiourea) which was employed as a key intermediate for the synthesis of a new series of bis- [1,3,4] thiadiazolimines, and bis-thiazolimines, with an alkyl linker, through its reaction with various hydrazonoyl halides or α-haloketones, respectively. The newly synthesized derivatives' structures were confirmed by elemental analysis and spectral data. Docking studies were applied to test the binding affinity of the synthesized products against the M pro of SARS-CoV-2.
The study results showed that compound 5h is the best one as it has average binding energy (−7.50 ± 0.58 kcal/mol) better than that of the positive controls, O6K and N3 (−7.36 ± 0.34 and −6.36 ± 0.31 kcal/mol). Additionally, the docking poses (H-bonds and hydrophobic contacts) of the tested compounds against the M pro using the PLIP web server were analyzed. This work paves the way for the design and synthesis of bis-thiadiazoles and bis-thiazoles-based libraries, which could lead to the innovation of efficient treatment against SARS-CoV-2 main protease (M pro ).

Data Availability Statement:
The data presented in this study are available on request from corresponding author.

Conflicts of Interest:
The authors declare no conflict of interest.