Synthesis and In Silico Investigation of Isatin‐Based Schiff Bases as Potential Inhibitors for Promising Targets against SARS‐CoV‐2

Abstract

Despite the significant development in vaccines and therapeutics cocktails, there is no specific treatment available for coronavirus disease 2019 (COVID‐19), caused by the new severe acute respiratory syndrome coronavirus 2 (SARS‐CoV‐2). Targeting the main protease (M^pro) of SARS‐CoV‐2, which possesses a key role in producing the essential viral structural and functional proteins, can be considered an efficient way to control this potentially lethal infection. Recently, some of Michael acceptor‐pharmacophore containing inhibitors have been suggested as successful suppressors of the main protease. Here, we synthesized the Isatin‐based Schiff bases possessing the structural pattern of a Michael acceptor‐like portion employing synthesis procedures. In silico investigation of these compounds was not limited to the main protease. We have also evaluated their possible inhibitory activity against the other identified druggable targets using homology modeling, molecular docking, and molecular dynamics simulations. Our investigations revealed that the dimethyl biguanide carrying Schiff bases of Isatin‐derivatives have the best binding mode and interaction energy. The dimethyl biguanide moiety‐containing compounds have formed promising interactions with the key amino acid residues Cys145 and HIS41 of M^pro with a binding free energy of −7.6 kcal/mol which was lower than the positive control compound Carmofur (−6.3 kcal/mol). It also leads to the higher affinity and the much inhibitory potential against the SARS‐CoV‐2 RdRp and Spike glycoproteins, human TMPRSS2, and ACE2 receptors.

In the current study we synthesized C5‐ substituted Isatin‐based Schiff bases employing green synthesis procedures. Although the in silico investigation of these compounds was not limited to the main protease of SARS‐CoV‐2, we have found the dimethyl biguanide carrying final products shows the best binding mode and interaction energy against M^pro. Considering the structural pattern of a Michael acceptor‐like portion in these compounds to interact with the Cys–His catalytic dyad, these results were not surprising.

Introduction

Following the viral infections severe acute respiratory syndrome (SARS) and Middle Eastern respiratory syndrome (MERS), which were epidemic in 2002–2003 and 2012 respectively, today humankind are in a new tough battle against the other human Beta‐coronavirus (SARS‐CoV‐2) which causes the high contagion and potentially fatal infection COVID‐19 (coronavirus disease 2019).[ ¹ , ² ] The current COVID‐19 pneumonia pandemic, which was first described in humans in Wuhan, China, in December 2019, has affected almost everything from social and economic aspects of our life to humans’ health problems and death.

Apart from the non‐specific therapeutic cocktails, including immunosuppressive and antimicrobial agents utilized in COVID‐19 clinical management, we have also been trying to control its highly pathogenic virus through different semi‐specific inhibitory effects such as suppression of human transmembrane serine protease 2 (TMPRSS2). That enzyme is essential for the cleavage and activation of SARS‐CoV‐2 spike protein as the primary phase of the viral adhesion, ^[3] inhibition of the pH‐mediated endocytosis process and the viral genome release from early endosomes through different strategies such as increasing of intracellular pH, ^[4] prevention of the viral replication and virus exports from the infected cells via targeting the RNA‐dependent RNA polymerase (RdRp), ^[5] and inhibition of the main protease (M^pro) of SARS‐CoV‐2.

SARS‐CoV‐2 M^pro can be considered a great target for drug discovery and rational drug design research against this potentially fatal infection since inhibiting M^pro prevents a successful production process of viral mature structural and functional proteins. ^[6] Therefore, both survival of the virus and its replication will be disrupted. ^[7] The binding pocket of M^pro is divided into a series of sub‐sites (including S1, S2, S4, and S1’) possessing the catalytic pair amino acid residues Cys145 and His41 which act as the nucleophile and general acid/base or a π‐π interaction creator group, respectively. It has been shown through a molecular docking study with Ritonavir which is already in clinical trials for COVID‐19. ^[8]

The highly conserved sequence and structure of the M^pro among Beta‐coronaviruses make feasible the drug design and development progress for the newly born viral infection COVID‐19 on the earlier discovered M^pro inhibitors for SARS and MERS. Among existing SARS/MERS‐CoVs M^pro inhibitors as lead compounds, in the initial phase, non‐covalent inhibitors X77 and Baicalein were recognized as potential inhibitors for SARS‐CoV‐2 M^pro (PDB IDs: 6W63 ^[44] and 6M2N, ^[45] respectively). Furthermore, successful peptide‐like small molecules (N3 and 13b) have been suggested for the purpose (Figure 1) since they include diverse electron‐deficient reactive Warheads groups, i. e., Michael acceptors, aldehydes, epoxy ketones, and other ketones to attack Cys145 covalently.

Figure 1.

X77 and Baicalein as non‐covalent inhibitors of the main protease versus the peptide‐ like M^pro inhibitors N3 and 13b.

SARS‐Cov‐2 M^pro structure is deposited in PDB (PDB codes: 6LU7 ^[23] and 6LYF ^[46] ) with peptide‐like irreversible inhibitors N3 and 13b respectively as the co‐crystallized ligands. In the active site, Michael acceptor group‐ containing inhibitor N3 creates covalent modification of Cys145 as an electrophile in its extended conformation similar to the α‐ketoamide 13b. The formation of the covalent bond between the M^pro and the co‐crystallized inhibitor N3 as a structural analogue of broad‐spectrum viral protease inhibitor Rupintrivir happens through the Michael addition of the vinyl group; and the α‐ketoamide functional group forms a hemithioacetal with Cys145. ^[9]

A high‐throughput virtual screening (HTVS) among approved drugs and/or drug candidates have suggested six small molecules as efficient inhibitors of SARS‐Cov‐2 M^pro including the synthetic organoselenium compound Ebselen, organic disulfide Disulfiram, thiadiazolidine compound Tideglusib, plant‐derived naphthoquinone compound Shikonin, imidazole containing disulfide compound PX‐12, and antineoplastic agent Carmofur which is an organohalogen pyrimidine derivative (Figure 2). Furthermore, the experimental researches have confirmed the effectiveness of these compounds on COVID‐19 in some of different aspects. ^[10]

Figure 2.

Small molecule M^pro inhibitors antineoplastic drug Carmofur, the strongest screened small molecule Ebselen, Disulfiram, Tideglusib, Shikonin, and PX‐12.

Although the molecular mechanism of the M^pro inhibition by small molecules such as Carmofur is not crystal clear and seems different from N3, the X‐ray crystallography of M^pro‐Carmofur complex reveals generally similar pattern with M^pro‐N3 complex. According to the molecular simulation and mass spectroscopy, carbonyl group of Carmofur covalently binds to the essential catalytic amino acid residue Cys145. Different from N3 which occupies four subsites S1, S2, S4 and S1,[ ¹¹ , ¹² ] the crystal structure of SARS‐CoV‐2 M^pro revealed the transfer of the hexylurea side chain from Carmofur to Cys145 in M^pro, the fatty acid tail of the drug as a residual hexylcarbamothioate appears to interact with His41 and other existing residues into the hydrophobic S2 subsite in the binding pocket of this cysteine protease (PDB: 7BUY).

Therefore, it seems that for further drug design studies against this highly infectious viral disease through the targeting of the main protease, small molecules carrying an electron deficient warhead can be considered as a clever step forward.

Successful developed mRNA and protein‐based vaccines are currently available against SARS‐CoV‐2 that have lighted a flame of hope in the world‘s dark moments in the recent global pandemic. Apart from the efficacy of these vaccines for preventing disease or severe disease that results via achieving “herd immunity”, and the symptomatic treatment using immunotherapeutics (Dexamethasone, Tocilizumab, Mavrilimumab, Baricitinib, Bamlanivimab, Etesevimab, and interferons), anti‐viral therapeutic (Remdesivir, Ritonavir/Lopinavir, Favipiravir), antimicrobial and antimalarial medications (Azithromycin, Doxycycline, Ivermectin, and Hydroxychloroquine), Camostat mesylate, Heparin, and etc. as well as convalescent plasma, there is a lack of safe, effective, and specific therapeutic drugs against this potentially lethal viral infection.[ ¹³ , ¹⁴ ] Therefore, we still are not free from the need of a confirmed and absolute medication for COVID‐19 patients. Thus, achieving the small‐molecule candidate drugs such as Paxlovid targeting the key proteins in the life cycle of SARS‐CoV‐2 seems to be a critical goal that must follow.

Since several derivatives of Isatin which is a well‐known structural heterocyclic part of many naturally occurring compounds demonstrated excellent antiviral activity against a vast range of pathogenic viruses, numerous studies have been conducted to develop Isatin derivatives as potential SARS‐CoV main protease and RdRp inhibitors[ ¹⁵ , ¹⁶ ] which plays a viable role in the viral replication cycle. According to the literature review, nonvoluminous electron‐pulling groups at C‐5 of the Isatin ring as well as the appropriate N‐1 substitution are linked to better activity in most of the reported antiviral Isatins. So we synthesized C‐5‐ substituted analogs of Isatin based on molecular hybridization strategy and have found them effective in our in silico investigations.

In the current research, we present an efficient synthesis procedure for the development of SARS‐CoV‐2 RdRp and M^pro inhibitors which are structurally based on Isatin scaffold. In continuation of our earlier research, ^[17] in this work, the proposed Isatin‐based Schiff bases are including a Thio/semicarbazide and/or N, N‐dimethylbiguanide structural portion which is multi‐targeted functional scaffolds in medicinal chemistry possessing confirmed activity against SARS‐CoV‐2 via diverse mechanisms of action.[ ¹⁸ , ¹⁹ , ²⁰ , ²¹ ] These compounds were synthesized utilizing Glycerin (Glycerol) as recently proposed valuable green solvent for organic syntheses ^[22] (Scheme 1) and investigated as potential inhibitors of the SARS‐CoV‐2 main protease using molecular simulations (PDB IDs: 6LU7). ^[23]

Scheme 1.

General structures of the proposed Isatin‐derived M^pro inhibitors in this research.

Aside from the main protease and RNA‐dependent RNA polymerase (RdRp) (PDB ID: 6M71) ^[24] that can be considered as the previously confirmed drugable therapeutic targets for Isatin derivatives against COVID‐19, we have also evaluated the possible inhibitory potential of these compounds against the other druggable targets such as Spike receptor‐binding domain bound with Angiotensin‐converting enzyme 2 (ACE2) (PDB ID: 6M0J), ^[25] and TMPRSS2 using homology modeling, molecular docking, and molecular dynamics simulations hoping to find notable results serendipitously.

Experimental

Materials and Methods

Similar to the synthetic processes of our previously published study that the discussed compounds in the current research were synthesized as functional intermediates using a novel magnetic nanocatalyst for the first time, ^[17] both chemicals and reagents were obtained from Sigma Aldrich. Thin layer chromatography (TLC) was used to determine the progress of the reactions processes and purity of the intermediate compounds and the final products. Chemical characterization was utilized by following equipments such as IR spectra recording at 4000–400 cm⁻¹ on KBr pellets using a Shimadzu 8201 pc. ¹H and ¹³CNMR spectra were recorded on Bruker DRX instrument at 300 MHz. Elemental Analyzer (Varian EL III) was utilized to analyze, elements C, H, N present in the synthetic compounds. ^[17] In this research we used the molecular simulations study as the early phase of the drug design approaches to investigate and predict the potential interactions of synthesized Isatin‐based Schiff bases with the main identified druggable targets against COVID‐19 including the SARS‐CoV‐2 main protease (PDB ID: 6LU7), ^[23] RNA‐dependent RNA polymerase (RdRp) (PDB ID: 6M71) ^[24] and the Spike receptor‐binding domain (PDB ID: 6M0J), ^[25] Angiotensin‐converting enzyme 2 (ACE2) bound with the Spike receptor‐binding domain (PDB ID: 6M0J), ^[25] and Transmembrane serine protease 2 (TMPRSS2) using homology modeling, molecular docking, and molecular dynamics simulations.

Chemistry

General synthetic procedure for compounds CO₁‐CO₁₂

The (substituted) Isatin‐3‐thio/semicarbazones (CO₁ ‐CO₈ ) and N, N dimethyl biguanidin imino isatin derivatives (CO₉ ‐CO₁₂ ) can be produced in the Glycerol as the non‐costly, green, and safe solvent. The prepared mixture of 1 mmol Isatin and 1 mmol amino group‐containing organic compound (such as Thio/semicarbazide and/or N, N‐dimethylbiguanide (Metformin)) was stirred under mild conditions for an adequate amount of time (Table 1). The progress of the reaction was monitored by TLC (Chloroform: Methanol (9 : 1). After completion of the reaction, H₂O (3–5 mL) was added to the reaction mixture and the obtained crude compounds were collected by filtration and washed with water/methanol. Then, the final products were separated by recrystallization from ethanol to afford (CO₁ ‐CO₁₂ ) in an excellent yield. These compounds were synthesized by condensation of Isatin with Thio/semicarbazide and Metformin (N, N‐dimethylbiguanide) as summarized in Scheme 2.

Table 1.

The optimization reactions of Schiff bases production in the presence of Glycerol.

Entry	Amount of Glycerol (mol %)	Derivatives with	Time (min)	Yield (%)a
1	–	O‐ containing tails	300	0
2	–	S‐ containing tails	300	0
3	10	O	30	94
4	15	S	36	92
5	5	O	150	94
6	10	S	165	94
7	15	O	120	92
8	5	S	140	92
9	10	Metformin tail	120	90
10	15	Metformin tail	120	89
11	5	Metformin tail	80	90

Scheme 2.

General synthetic procedure for compounds CO₁‐CO₁₂.

Molecular simulation investigations

Preparation of protein and ligands

The Structure of the main protease (M^pro) of SARS‐CoV‐2 (PDB ID: 6LU7), ^[23] SARS‐Cov‐2 RNA‐dependent RNA polymerase (RdRp) with PDB ID: 6M71, ^[24] and Spike receptor‐binding domain bound with Angiotensin‐converting enzyme 2 (ACE2) (PDB ID :6M0J) ^[25] have been downloaded from the Protein Data Bank (http://www.pdb.org). In order to docking study, specific chain (Chain A) of the M^pro and RdRp proteins (6LU7 ^[23] and 6M71 ^[24] respectively), have been selected for the preparation of receptor protein input file.

As the three‐dimensional (3D) structure of TMPRSS2 is not available in the Protein Data Bank, the sequence of TMPRSS2 (O15393) is retrieved from Uniprot and was fetched onto the SWISS‐MODEL server to create homology models of the protein. The top ranked template structures for the modeling were selected to build TMPRSS2 models. The Hepsin structure (PDB ID: 5CE1) ^[26] was subjected as the template to protein preparation and the verified homology model of TMPRSS2 with good quality was further prepared for molecular docking studies. The 3D structures of ligands were drawn in ChemDraw Ultra 8.0 and after geometry optimization of ligands using the semi‐empirical AM1 Hamiltonian, the structures were saved as pdb files. The general molecular structures of the studied synthetic ligands in this research are shown in supporting information (Figure S1). The input pdbqt files of the proteins and ligands for the docking simulation were generated using Auto Dock Tools.

Drug likeness properties

Prediction of the overall pharmacokinetic behavior of the potentially effective synthetic compounds was tested using Swiss ADME which is based on Lipinski's rule of five. Different molecular parameters such as molecular weight, numbers of hydrogen bond donors, number of hydrogen bond acceptors, and LogP values were analyzed as per Lipinski's rule of five. Further, other significant properties such as total polar surface area (TPSA) and number of rotatable bonds are calculated using Swiss ADME program. ^[27]

Docking study

Here, molecular docking study was accomplished using Auto Dock Vina package 1.1.2 and the binding energy between the receiver and the ligands is calculated. The value is stored in a score using a grid method which searches the available conformational space for the protein and ligand which gives an effective assessment of the binding energy between conformations. The grid box parameters for 6M71 ^[24] were set to size 30 Å×30 Å×30 Å and center [118 Å, 119 Å, 140 Å]; For 6LU7 ^[23] the grid box parameters were set to size 30 Å×30 Å×30 Å and center [−19 Å, 12.599 Å, 70 Å]; for 6 M0 J ^[25] the grid box is set at the junction of Spike‐RDB and ACE2 with size 20 Å×20 Å×20 Å and center [−32.483, 26.077, 7.923] for Spike‐RDB and size 30 Å×30 Å×30 Å and center [−25.100, 19.090, 3.074] for ACE2 respectively. The grid box was generated at the catalytic site of TMPRSS2 with size 20 Å×20 Å×20 Å and center [10, −6, 26] covering the important amino acid catalytic residues His296, Asp345, and Ser441 of TMPRSS2. Exhaustiveness was set on 100 and the spacing between the grid points was 1.0 Å. Plip web server was applied for the analysis and visualization of docking results and intermolecular interactions between ligand and the receptor molecules. Interactions shown between amino acids and different ligands, including hydrophobic contacts and Van der Waals interactions and hydrogen bonds, can provide insights into recognition of molecular mechanisms of these potentially active compounds. ^[28]

Molecular Dynamics Simulations and MM‐GBSA Calculations

Molecular dynamics (MD) simulations on top two docked complexes (CO₁₂ in complex with SARS‐CoV‐2 RdRp (PDB : 6M71) ^[24] and M^pro (PDB : 6LU7) ^[23] were conducted using the GROMACS version 2019.1, ^[29] 2019. The calculations were performed on Ubuntu operating system (version 18.04) with AMBER03. all‐atom force field with 75 ns time duration. All the MD simulations systems were solvated using SPC water model. The number of atoms in simulated systems were 110411 and 48683 for SARS‐CoV‐2 RdRp and M^pro respectively. Then net charge of the system was neutralized by addition of 6 Na+/Cl− ions for SARS‐CoV‐2 RdRp and 4 Na+/Cl− for M^pro. Energy minimization of 5,000 steps followed by equilibration for 50 ps. Berendsen thermostat algorithm ^[30] was used for maintaining the system at constant volume (100 ps) and at a constant temperature (310 K) in NVT equilibration. The Particle‐ Mesh Ewald (PME) ^[31] method was used to calculate the electrostatic interactions. The Coulomb radius and Fourier grid spacing were set at 1.2 and 0.16 nm, respectively while the van der Waals interactions were limited to 1.2 nm. In order to energy stabilization the MD trajectories were saved at every 10 ps which obtained 50000 frames from each production simulation. Root mean square deviation (RMSD) and root mean square fluctuation (RMSF) were calculated.

Moreover as a widely‐accepted methods the Poisson‐Boltzmann MM/PBSA was used for computing the protein‐inhibitor affinity.[ ³² , ³³ , ³⁴ ] For calculating the binding free energy of M^pro and RdRp docked complex with CO₁₂. The ‘g_mmpbsa’ tool ^[35] with default parameters was used for molecular mechanics potential energy (electrostatic + Van der Waals interactions) and solvation free energy (polar + non‐polar solvation energies) calculations. The last stable 20 ns (200 frames) trajectories assess by the RMSD plot from each docked complex were used to estimate binding free energy.

Results and discussion

Chemistry

In this aim, the synthetic method involves the use of Glycerol as a catalytic solvent. ^[36] Our previously reported protocol ^[17] proceeds via a novel magnetic nanocatalyst. Both of these methods followed for the synthesis of Isatin‐based Schiff bases in this research involve a green pathway. The final products (CO_1‐CO₁₂) are obtained within a short time and in excellent yields. The general pathway of this procedure is presented in Scheme 1. The optimization conditions are given in Table 1. The chemical structures of the produced Schiff bases are presented in Scheme 2 as well as the first row of Table 1. Proton nuclear magnetic resonance (¹HNMR) spectroscopy, ¹³CNMR spectroscopy, and mass spectrometry techniques were used to characterize the compounds. The structural assignment of all the obtained products can be probed in three categories 1) Isatin‐3‐semicarbazones (CO₁, CO₃, CO₅, CO₇), 2) Isatin‐3‐thiosemicarbazones (C₂, C₄, C₆, C₈) and 3), and N, N dimethyl biguanidin imino isatin derivatives (C₉, C₁₀, C₁₁, C₁₂).

The analytical and spectroscopic data for the final products have been prepared and presented in supporting information (Figures S5–S40). As an example, 2‐(2‐oxoindolin‐3‐ylidene) hydrazine‐1‐carboxamide (CO₁) from the first category. The elemental analysis and mass spectrometry of compound CO₁ have the gross formula C₉H₈N₄O₂. The ¹HNMR spectrum of CO₁ exhibited a singlet at 7.13 corresponding to the NH₂ moiety in Semicarbazide‐ derived portion, two singlet signals at 11.10 (corresponding to the Semicarbazide‐ derived NH group) and 11.73 (corresponding to the −NH unit in Isatin). Four aromatic protons give rise to characteristic signals in the aromatic region of the spectrum (6.91–7.60). Signals at 141.94, 155.45, 163.17 ppm were observed in the ¹³CNMR spectral profile. The signals were attributed to the C=N, N−C=O, and Isatin‐C=O carbons respectively.

Molecular docking

At the first phase, docking was performed with reference molecules of respective proteins to validate the docking protocol. The molecular fit of the Carmofur and Ribavirin triphosphate was used as reference inhibitors in order to validate the results of molecular docking studies against SARS‐CoV‐2 M^pro and RdRp respectively. Carmafor as the reference ligand was docked in the active site of the main protease (PDB: 6LU7) ^[23] and Ribavirin triphosphate was studied as the potential reference drug hit against RdRp (PDB : 6 M71). ^[24]

Estimated binding energies and the possible interactions of the studied ligands (CO₁‐CO₁₂) docked into the binding site of the 6LU7 and 6M71 enzymes are summarized in Table 2, 3. After successful docking of all the ligands with these receptors, the docking results showed the efficiency of Isatin‐based Schiff bases (CO₁‐CO₁₂) as inhibitors against SARS‐CoV‐2 M^pro and RdRp. As can be seen from Table 2 ligands CO₉‐CO₁₂ show the best binding values and the rest studied ligands have reasonable but higher binding energies compared with Carmafor. The binding energy of CO₁₂ (−7.0 kcal/mol) is just 0.2 kcal/mol less than Ribavirin triphosphate and the other studied ligands have close binding energy to this reference ligand which shows their inhibitory potential. The current investigation shows that the compound CO₁₂ has formed better interactions within the active site of the main protease in comparison to RdRp (Table 2).

Table 2.

Molecular docking analysis of studied compounds against M^pro (6LU7) and RdRp (6M71).

Ligands	6LU7			6M71
	Free Energy of Binding (kcal/mol)	H bond	Hydrophobic	Free Energy of Binding (kcal/mol)	H bond	Hydrophobic
CO1	−6.2	LEU141, GLY143, SER144, CYS145	THR25, LEU27	−6.3	ILE548, ALA840, ASP845, ARG858	ILE548, ALA840
CO2	−5.4	GLY143	MET165, GLN189, HIS41	−5.5	ILE548, PHE843, ARG858	ILE548, ARG836, ALA840
CO3	−6.2	GLY143, SER144, GLU166	MET165	−6.4	ALA406, GLN408, ILE450, MET542, GLY670	ALA406, GLN408
CO4	−5.6	GLY143	MET165, GLN189	−5.7	TYR546, ILE548, ALA840, PHE843, ARG858	ALA547, ILE548
CO5	−6.3	HIS41, GLY143	MET165, GLN189	−6.1	ASP452, ARG553, THR556, ARG624,	LYS621
CO6	−5.8	GLY143	MET165, GLN189	−5.7	TYR546, ILE548, PHE843, ARG858	ALA547, ILE548
CO7	−6.3	GLY143, SER144, CYS145, GLU166	MET165, GLN189	−6.3	THR409, LEU544	LYS411
CO8	−5.9	GLY143	MET49, MET165, GLN189	−5.9	TYR546, ILE548, ALA840, PHE843, ARG858	ILE548, ARG836
CO9	−7.2	HIS41, GLY143, SER144, CYS145	MET165, GLN189	−6.8	GLN408, THR409, LYS411	TYR546
CO10	−7.5	HIS41, GLY143, SER144, CYS145	MET165, GLN189	−6.8	GLN408, THR409, LYS411	TYR546
CO11	−7.5	HIS41, GLY143, SER144, CYS145	MET165, GLN189	−6.9	GLN408, THR409, LYS411	TYR546
CO12	−7.6	CYS145, SER144, GLY143 a, HIS41	MET165, GLN189	−7.0	GLN408, THR409, LYS411	TYR546, ILE847
Carmofur	‐6.3	GLY143, SER144, CYS145 a	–	–	–	–
Ribavirin triphosphate	–	–	–	−7.2	ASP452, ARG553, THR556, ASP623, SER682, ASN691, SER759	–

[a] The identical critical amino acids (in the reference molecule‐target complex as well as the best‐evaluated ligand‐target complex) are shown bolded.

Table 3.

Molecular docking analysis of studied compounds against ACE‐2 and SPIKE (6M0J).

Ligands	ACE2			Spike
	Free Energy of Binding (kcal/mol)	H bond	Hydrophobic	Free Energy of Binding (kcal/mol)	H bond	Hydrophobic
CO1	−5.5	ASP350, ASP382	PHE40, ASP350	−5.6	GLN498, ASN501	ARG403, TYR495, PHE497, TYR505
CO2	−5.7	ALA348, ASP350, ARG393	PHE40, TRP349	−5.2	GLN498, ASN501	TYR453, TYR495
CO3	−5.8	ASP350, ASP382	PHE40, ASP350	−5.6	ARG403, GLU406	TYR495
CO4	−4.6	TYR41, ASN330, ASP355, ARG357	–	−5.4	GLN498, ASN501,	TYR453, TYR495
CO5	−5.8	ASP350, ASP382	PHE40, ASP350	−5.9	ARG403, GLU406, TYR453	TYR495
CO6	−4.9	ASN33, HI34, GLU37, ALA387, ARG393	PRO389	−5.3	ARG403, GLU406, TYR453	TYR495
CO7	−5.6	ARG393, GLY352,	PHE390, LEU391	−6.1	ARG403, GLU406, TYR453	TYR495, PHE497, TYR505
CO8	−6.1	ALA348, ASP350, ARG393	PHE40, TRP349	−5.5	ARG403, GLU406, TYR453	TYR495, PHE497
CO9	−5.6	TYR41, GLY326, ASN330, ARG357	–	−6.3	TYR453, SER494, GLY496, ASN501	–
CO10	−5.7	ASN33, ALA386, ARG393	GLU37	−6.3	TYR453, SER494, GLY496, ASN501	–
CO11	−5.7	ASP30, ASN33, PHE390, ARG393	PRO389	−6.3	ARG403, TYR453, GLY496	‐
CO12	−5.8	ASP30, ASN33, PHE390, ARG393	PRO389	−6.4	ARG403, TYR453, GLY496	–

Surpassing the results, Table 2, Isatin‐based thiosemicarbazones (CO_{2 ,} CO₄ , CO₆ , and CO₈ ) have revealed the lowest docking scores. Their interaction with binding pocket of targets shows they just have hydrogen bond interaction with GLY143. The highest energy levels are observed for Metformin carrying derivatives CO_{9 ,} CO₁₀ , CO₁₁ , and CO₁₂ . This is quite probable as they quite fit in the binding site due to the present of the N, N dimethylbiguanide atoms in their structures which leads the best affinity in the binding pocket of SARS‐Cov‐2 M^pro and RdRp which leads to increase the inhibitory potential of these compounds against studied enzymes. Figure 4 shows the interaction of CO₁₂ as the best inhibitor in the binding pocket of targets. As can been seen, it has hydrogen interactions with Cys145 and His41 critical residues of the main protease which are the most prominent and potential target for the inhibitory effect of the protein. ^[37] Moreover, molecular docking was performed to find types of interactions and the binding affinity in order to check docking features at the junction of spike‐RBD and ACE2 interface in SARS‐CoV‐2 spike receptor‐binding domain bound with ACE2 target. The results are presented in Table 2. As can be seen, the CO₁₂ showed better binding energy than the other investigated compounds as the potential inhibitor of the viral entry pathway Figure 3.

Figure 4.

3D interaction of CO₁₂ with the amino acids of ACE2 (a) and Spike (b).

Figure 3.

3D interaction of CO₁₂ with the amino acids of 6LU7 (a) and 6M71 (b).

The dimethyl biguanide group as substituent causes better binding score mainly because of its engagement in stacking interactions with amino acid side chains of the Spike and ACE2 protein. Methyl group in 5‐methyl Isatin portion provides high density of resonance electrons which results higher binding affinity. Spike protein binding with ACE2 in presence of CO₁₂ can be seen in Figure 4.

In this study, the virtual screening of CO₉ ‐CO₁₂ was performed against the serine protease TMPRSS2. As can be seen from the results, in the catalytic site, CO₉ , CO₁₀ , CO₁₁ and CO₁₂ are forming strong interactions with TMPRSS2 residues.

Here, Camostat was chosen as the reference ligand against TMPRSS2 and the comparative resulted data depicts the inhibitory potential of studied ligands. According to the binding modes presented in Table 4, the existing differences between the number and the nature of the observed interactions explain the lower binding energy of the reference standard molecule Camostat in comparison to the best‐tested ligand CO₉ (ΔG=−0.9 kcal/mol).

Table 4.

Molecular docking analysis of the best studied compounds against TMPRSS2.

Ligands	Free Energy of Binding (kcal/mol)	H bond	Salt Bridge	Hydrophobic
CO9	−6.8	GLN438, SER441 a	–	THR459
CO10	−6.3	SER441, GLY462	–	THR459
CO11	−6.7	SER441, GLY462	–	THR459
CO12	−6.7	HIS296, GLY439, SER441		GLU299, LEU302
Camostat	−7.7	TYR337, LYS342, SER436, *SER441, TRP461, GLY462, GLY464	HIS296,LYS342,ASP435	HIS296, LYS342, CYS437, *GLN438, TRP461

[a] The identical critical amino acids (in the reference molecule‐target complex as well as the best‐evaluated ligand‐target complex) are shown bolded.

Drug likeness properties

The computational prediction of pharmacokinetics properties of potential RdRp and M^pro inhibitors were evaluated by SWISS ADME online server according to Lipinski's ‘rule‐of‐five’ (Table 4<tabr4<). Lipinski's ‘rule‐of‐five’ states that a compound to exhibit drug likeness has no more than one violation of defined criteria ^[38] and as can be seen in Table 4<xtabr4<, all ligands possess the allowed values.

Molecular dynamics simulations

In this study, in order to investigate the stability, conformational changes and dynamics of docked M^pro and RdRp with CO₁₂ inhibitor complexes, MD simulations were carried out which reveals the interaction and structural stability of inhibitor complexes with protein on atomic level. Root mean square deviation (RMSD) of Cα atoms of protein backbone was monitored throughout 75 ns simulation. The RMSD value shows stability around 0.35 and 0.3 for RdRp and M^pro with inhibitors, respectively.

As can be seen from Figure 5 at the first 50 ns of the dynamic simulation, RMSD values of the nsp12 protein increases slowly from 0.2 to 0.35 nm and remain stable between 0.35 and 0.4 nm for the next 25 ns while the analysis of RMSD for CO₁₂‐M^pro complex showed that this value increase from 0.15 to 0.35 nm and then begin to relax from 50 ns. Moreover, (RMSF) values of backbones C_α atoms for both complexes were calculated and plotted by averaging over all the conformations sampled during 75 ns simulation. As shown in Figure 6 in the case of RdRp, the RMSF of residues (260–431) and (850–920) fluctuates significantly during protein‐ligands interaction. While the highest peaks in the region between residues 268–280 was observed in the M^pro complex. Furthermore, binding affinities of the investigated molecules were calculated using MM‐PBSA methods as reflected in Table>5tabr5>, while due to complexity, the entropy part was not calculated. The MM‐PBSA binding energy offers a good compromise between accuracy and computational cost.[ ³⁹ , ⁴⁰ , ⁴¹ , ⁴² , ⁴³ ] The total binding energy was −293.347±22.034 kcal mol⁻¹ for RdRp and −612.172±54.849 for M^pro complex.

Figure 5.

2D interaction of Studied CO₁₂ with SARS‐CoV‐2 spike and human ACE2 receptor in their interacting site.

Figure 6.

Molecular dynamic simulations result: (a) RMSD plot of RdRp in complex with CO₁₂; (b) RMSF plot of RdRp in complex with CO₁₂; (c) RMSD plot of M^pro in complex with CO₁₂; (d) RMSF plot of M^pro in complex with CO₁₂.

The SARS‐CoV‐2 main protease is different from human proteases, which means it allows us to design specific drugs with selective toxicity. A considerable number of peptidomimetic and small‐molecule inhibitors of the SARS‐CoV‐2 main protease were developed using reactive electron‐deficient warhead groups, including Michael acceptors, aldehydes, and epoxy ketones, electrophilic ketones such as halomethyl ketones, trifluoromethyl ketones, and keto‐ amides.[ ⁴⁴ , ⁴⁵ , ⁴⁶ , ⁴⁷ , ⁴⁸ ] In the current research, the dimethyl biguanide carrying Schiff bases (compounds CO₉ , CO₁₀ , CO₁₁ , and specially CO₁₂ ) have revealed the best energies of interaction with the SARS‐CoV‐2 M^pro and also RdRp. Considering our calculations for CO₁₂ binding affinity against M^pro (−7.6 kcal/mol) Table 2 reveals that its affinity score is better than that for two original M^pro ligands: N3 (−7.1 kcal/mol) ^[48] and Carmafor (−6.3 kcal/mol). However, in this study we consider Carmafor as reference mainly because of its structural similarity to our studied ligands. It can be found that the dimethyl biguanide moiety in the best investigated ligands leads to their higher affinity to the binding pockets and subsequently the much inhibitory potential against the selected targets. Considering the binding site of M^pro, which is highly conserved among all coronaviruses,[ ⁴⁹ , ⁵⁰ ] we expect an effective compound against one of them may be useful against a broader spectrum of these viruses. ^[51]

Conclusion

In the current research, compound 1‐N,N‐dimethylimidamide‐3‐(5‐methyl‐2‐oxoindolin‐3‐ylidene)guanidine (CO₁₂ ) has formed promising interactions with the key amino acid residues Cys145 and HIS41 of M^pro with the lower binding free energy than the non‐peptidic small‐molecule Carmofur as the standard reference. We also targeted human TMPRSS2 and ACE2 receptors and the SARS‐CoV‐2 RdRp as well as the structural glycoproteins Spike to find other possible antiviral effects of the prepared Schiff bases. Considering the preformed docking studies which are confirmed by the results of MD simulations, the Metformin containing Isatin‐derived Schiff bases (CO₉ ‐CO₁₂ ) have revealed the lower binding energy values against the selected targets correlated with their dimethyl biguanide moiety that provides multiple hydrogen bonds. These observations should be followed by experimental studies to evaluate the antiviral efficacy and toxicity of this successful molecular hybridization pattern.

Supporting Information Summary

Electronic supplementary information is available: The analytical and spectroscopic data for the final products (CO₁‐CO₁₂), and the molecular docking related information (validation of the docking studies and the 35 tested ligands interactions with the binding residues of the investigated targets).

Conflict of interest

The authors declare no conflict of interest.

Supporting information

As a service to our authors and readers, this journal provides supporting information supplied by the authors. Such materials are peer reviewed and may be re‐organized for online delivery, but are not copy‐edited or typeset. Technical support issues arising from supporting information (other than missing files) should be addressed to the authors.

Supporting Information

Z. Esam, M. Akhavan, M. Lotfi, A. Bekhradnia, ChemistrySelect 2022, 7, e202201983.

Data Availability Statement

The data that support the findings of this study are available in the supplementary material of this article.