Design, synthesis, biological evaluation and in silico studies of novel 1,2,3-triazole linked benzoxazine-2,4-dione conjugates as potent antimicrobial, antioxidant and anti-inflammatory agents

2.1.1 General

¹H and ¹³C NMR were recorded using a Bruker spectrometer operating at 400 and 100 MHz, respectively. The chemical shifts were recorded in ppm relative to tetramethylsilane and with the solvent resonance as the internal standard. Data were reported as follows: chemical shift, multiplicity (s = singlet, m = multiplet), coupling constants (Hz), integration. ¹³C NMR data were collected with complete proton decoupling. Chemical shifts were reported in ppm with respect to TMS with the solvent resonance as internal standard. Elemental analyses were performed on a Perkin Elmer 2400 Series II Elemental CHNS analyzer. Column chromatography (cyclohexane/ethyl acetate (5:5)) was carried out on silica gel (300–400 mesh, Qingdao Marine Chemical ltd., Qingdao, China). Thin layer chromatography (TLC) was performed on TLC silica gel 60 F254 plates 0.2 mm 200x200 nm; The spots were visualized using UV light at 254 nm and 360 nm.

2.1.2 General procedure for the synthesis of compounds 4(a-i): Cu-catalyzed azide-alkyne cycloaddition

A mixture of dipolarophile 1 (0.2 mmol), copper (I) iodide (0.5 eq) and trimethylamine (1 eq) were added at room temperature. To this mixture, aryl azide 1 (0.4 mmol) was added in toluene and the reaction mixture was subjected to microwave irradiations at 250 W completed within 10–15 min. the crude mixture was extracted with EtOAc (3x25mL) and the combined organic layer was dried over sodium sulfate, concentrated under reduced pressure and purified through a column chromatography (cyclohexane/ethyl acetate (5:5)) or recrystallization (ethanol), to give pure 4a-i in 60–86% (percentage yields).

2.1.3 General procedure for the synthesis of compounds 5(a-f, j): Ru-catalyzed azide-alkyne cycloaddition

The dipolarophile 3 was reacted with various aromatic azides and 5 mol% Cp ∗ RuCl(PPh₃)₂ catalyst. All the reactions were performed under microwave conditions in toluene and completed within 10–15 min. The desired new compounds 5(a–f, j) were obtained in yields ranging from 44% to 60%.

2.2.1 Antimicrobial screening

First, the optical density of each microorganism suspensions was adjusted to 0.1 at OD600 for bacteria and 0.4 at OD540 for yeasts. Then, 500 μL of inoculums were dropped onto adequate agar plates. Sterile filter discs (diameter 6 mm, Biolife, Italy) were placed at the surface of the appropriate agar mediums and 10 mg/disc of the product dissolved in 10% of dimethyl sulfoxide was dropped onto each disc. Tetracycline (10 mg/mL; 10 μL/disc) and Ampicillin (10 mg/mL; 10 μL/disc), were used as reference drugs. After incubation for 18–24h at 37 °C, the antibacterial activities were evaluated by measuring an inhibition zone formed around the disc. Each assay was performed in triplicate. For this, the microbial inocula were prepared from 12 h broth cultures and spectrophotometrically adjusted to 107 CFU/mL. Serial twofold dilutions of the different amounts of the compounds (1000 to 0.09 μg/mL) were prepared in adequate broth. Then, 10 μL of the inocula of each reference strain were added to the plates containing the serial dilution and were incubated aerobically at 37 ◦C for 24h. MIC for minimum inhibition concentration was defined as the lowest concentration that completely inhibited visible cell growth during 24h incubation period at 37 °C. MBC for minimum bactericidal concentration and MFC for minimum fungicidal concentration values were determined by inoculating 10 μL of each well medium with no visible growth on Müller Hinton or Sabouraud Chloramphenicol agar plates and were defined as the lowest concentration at which 99% of the tested strains were killed after 24h of incubation at 37 °C.

2.2.2 Antioxidant assays

DPPH^• radical ABTS^•+ radical scavenging assays have been achieved based on the same protocol done by with slight modification (Abdelhamid et al., 2018). The samples, at different concentrations (0.078, 0.15, 0.31, 0.62, 1.25 and 2.50 mg/mL), were pipetted in separate test tubes and then mixed with DPPH^• radical or ABTS^•+ radical solutions. After 30 min of incubation in the darkness and at a temperature of 25 °C. The absorbance of the resulting solution was measured at 520 nm with a spectrophotometer.

Inhibition of free radicals in percent (PI%) was calculated by following equation:

• PI% = 100 × (A_Control – A_Sample)/A_Control.

2.2.3 Anti-inflammatory test

2.3.1 Molecular docking study

The 3D-coordinates of S. aureus tyrosyl-tRNA synthetase, C. albicans N-Myristoyltransferase, human COX-2 enzyme, and human Peroxiredoxin 5 crystals were retrieved from the Protein Data Bank by selecting 1JIJ, 1IYL, 1PXX, and 1HD2 entries, respectively. For the preparation of protein crystal structures, the Protein Preparation Wizard (PPW) tool was used, which involves a three-step procedure: preprocessing, optimization, and protein minimization. Initially, the proteins were preprocessed by introducing hydrogen atoms to the structure and removing crystallographic water molecules beyond 5 Å. The H-bond network was adjusted in PPW's optimize tab in order to fix overlapping hydrogens and the most likely positions of thiol and hydroxyl hydrogen atoms. Finally, using the force field Optimized potentials for liquid simulation 3e (OPLS3e), restrained minimization was attempted until the mean root mean square deviation (RMSD) of the non-hydrogen atoms converged at 0.30 Å (Zrieq et al., 2021). Based on the promising biological activity compounds, 4a,4b, and 4d were selected for docking, which were treated by using the LigPrep tool to explore the chirality, ionization states, ring conformations, and tautomers of each input structure. At the site of a co-crystallized ligand, a grid was generated utilizing the Receptor grid generation tool, which describes the targeted protein properties as well as the shape used to provide more precise ligand pose scoring. A standard precision docking methodology was used with a default force field, and a detailed investigation of ligand binding affinities was done by computing docking score and binding free energies (ΔG) between protein and ligands.

2.3.2 Molecular dynamics (MD) simulation study

The MD simulation studies were carried out using the Desmond tool of Schrödinger molecular modelling, which allows us to comprehend the binding of a ligand–protein complex in simulated physiological conditions (Lee et al., 2021, Ayipo et al., 2022, Acar Çevik et al., 2022). A molecular dynamic simulation study was performed for the promising compound 4b in a complex with the human COX-2 enzyme (PDB ID: 1PXX). The ligand–protein docked (4b-1PXX) complex system was developed using a simple point charged (SPC) solvent system with an orthorhombic box of 10 ˚A on each side. To maintain the system eclectically neutral, the built system charge was neutralized with appropriate Na⁺ and Cl^– ions (Ghosh et al., 2021, Ahmad et al., 2022). After building the solvated system, the 4b-1PXX complex was minimize and relaxed using Desmond's default protocol, which relaxed the system into a local energy minimum. The model system is minimized using a hybrid approach of the steepest decent and the limited-memory Broyden-Fletcher-Goldfarb-Shanno (LBFGS) algorithms (Abdelhamid et al., 2018). The simulation was run in NPT ensemble mode with a 300 K temperature and 1 bar of pressure for 100 ns. For isothermal-isobaric conditions, the 'Nose-Hoover chain thermostat' and 'Martyna-Tobias-Klein barostat' algorithms were ensembled at 100 and 200 ps, respectively, using the Ensemble panel (Patel et al., 2018, 2022). Throughout the simulation, the recording interval trajectory was taken every 100 ps and 1000 frames. The Simulation Interactions Diagram Panel was used, which graphically displayed information regarding protein and ligand behavior and interactions throughout a simulation.

2.3.3 ADME study

Pharmacokinetic properties of the titled compounds were predicted using ADME (absorption, distribution, metabolism and excretion) descriptors by a SwissADME online server (https://www.swissadme.ch/).

3.2.1 Antimicrobial activity

The antimicrobial potential of the novel synthesized compounds was evaluated towards six different pathogenic species including 2 g-positive bacteria (S. aureus ATCC 25923, M. luteus NCIMB 81660), 2 g-negative bacteria (E. coli ATCC 25922, P. aeruginosa ATCC 27853), and 2 yeasts (C. albicans ATCC 90028, C. krusei ATCC 6258). As shown in Table 2, all synthesized analogues were found to be more active against highly resistant S. aureus MRSA with 4a (MIC 0.25 mM), 4b (MIC 0.23 mM) and 4d (MIC 0.21 mM) had MIC values less than the standard drug, tetracycline (MIC 0.28 mM). Towards M. luteus and P. aeruginosa, a potent activity was ascribed for all compounds in respect to the standard, tetracycline, however significant to moderate activity was achieved with E. coli. Interestingly, all tested compounds exhibited excellent antifungal properties against the fungal strains, C. albicans and C. krusei with MIC values lower than the standard drug, ampicillin (MIC 0.36 mM). The remaining compounds exhibited good to moderate antimicrobial properties.

3.2.2 Antioxidant activity

The antiradical potential of our synthesized compounds was evaluated spectrophotometrically as the capacity to scavenge DPPH^• and ABTS^•+.

• DPPH^• radical scavenging analysis

The capacity of the tested compounds and BHT to scavenge DPPH^• is outlined in Table 3. From the results, it can be observed that out of the synthesized series 1,4-disubstituted 1,2,3-triazole (4a-i), compounds 4a (IC₅₀ 1.88 ± 0.07 µM) followed by 4b (IC₅₀ 2.19 ± 0.09 µM), 4d (IC₅₀ 2.30 ± 0.01 µM), and 4f (IC₅₀ 2.98 ± 0.02 µM), respectively, exhibited excellent DPPH^• radical scavenging properties with lower IC₅₀ values than BHT (IC₅₀ 3.52 ± 0.08 µM), used as standard antioxidant. Meanwhile, compounds 4c, 4e, 4g-i showed the weaker DPPH^• radical scavenging ability than BHT. In regards to 1,5-disubstituted 1,2,3-triazole analogues, their DPPH^• radical scavenging ability is weaker than those of 1,4-disubstituted 1,2,3-triazole with only 5c (IC₅₀ 2,59 ± 0.34 µM) had potent activity than the standard, BHT, however the remaining compounds had lesser activity.

• ABTS^•+ radical scavenging analysis

The antioxidant results related to the capture of ABTS^•+ radical in comparison to reference compounds are summarized in Table 3. As shown, data revealed that out of the synthesized compounds, 4a (IC₅₀ 2.17 ± 0.02 µM), 4b (IC₅₀ 2.38 ± 0.43 µM), 4f (IC₅₀ 3.80 ± 0.01 µM) and 4d (IC₅₀ 4.07 ± 0.57 µM) displayed relatively high ABTS^•+ scavenging activities, in relation to BHT (IC₅₀ 4.64 ± 0.11 µM). The rest of compounds appeared to possess good to moderate scavenging potential. The same trend has been observed for compounds 4e, 4g-i, 5b, 5f and 5g displaying the weakest activities.

Our results generally indicate that the scavenging ability of our compounds seems to be even more potent in DPPH assay than ABTS test suggesting that the positive contribution of electron-donating substituents is more pronounced with DPPH test.

3.2.3 Anti-inflammatory activity

Inflammation is a physiological response of immune system to infections injuries, that involves many enzymatic and cellular processes to protect the body from all kinds of trauma. The acute anti-inflammatory potentiality of the prepared samples was assessed using xylene-induced ear edema in mice. The increase or decrease ear thickness of each group was measured after 3 h of xylene application. Results showed that the sample 4a was the most active against the inflammation. The percentages values showed a dose-dependent effect on the ear inflammation. At 100 mg/kg and after 3 h of xylene application, the ear edema was almost totally suppressed (84.41%) which had the same results as the drug reference “aspegic 300 mg/Kg” (PI = 87.53%). On the other hand, at only 25 mg/kg of 4a, the anti-inflammatory activity exhibited a potentiality exceed 50% of edema inhibition (61.03%) which is a strong activity compared to our reference (the dose is almost twelve times more; 300 mg/kg vs 25 mg/kg)). The samples 4b had a medium activity compared to 4a, but always more active than the reference. However, the inhibition percentage of the tested 4d showed an ineffective value at the highest tested dose (100 mg/kg; PI = 42.84%). As previously demonstrated, the central carbocyclic or heterocyclic ring system as found in many COX-2 inhibitors can be replaced by a central 1,2,3-triazole unit without losing COX-2 inhibition potency and selectivity. The high anti-inflammatory (via COX-2) inhibition potency of some 1,2,3-triazoles having a vicinal diaryl substitution pattern along with their ease in synthesis through versatile Ru(II)-catalyzed click chemistry make this class of compounds interesting candidates for further design and synthesis of highly selective and potent COX-2 inhibitors (Othman et al., 2021b) (Table 4).

3.4.1 Drug-likeness assessment

The most active analogues have been further predicted for their druglikeness and pharmacokinetic properties including absorption, distribution, metabolism and excretion toxicity (ADME) parameters to verify that the designed molecules are a viable drug (Table 5). The selected ligands were predicted to comply with Lipinski's rule-of-five, as well as Ghose, Veber, Egan and Muegge filters. They displayed good lipophilic property with consensus log Po/w value of 4.28 good bioavailability score (55%), high gastrointestinal absorption (GI) and were predicted to be not blood–brain-barrier (BBB) permeant as well as not P-gp substrate. Simultaneously, predictive data showed that the designed compounds are inhibitors of cytochrome P450 isoenzymes CYP 1A2 and CYP2C19, but not inhibitor of CYP2D6 and particularly CYP3A4, known as the most abundant in the liver and an active participant in the metabolism of approximately 60% of known drugs. The skin permeation ability of the selected compounds (LogK_P), was in the range −6.72 cm/s to −6.02 cm/s. As per as medicinal chemistry properties, all compounds showed none PAINS alert and the majority of them with also none brenk alert and generally good synthetic accessibility. The bioavailability radar plot in which the following properties were taken into account such as flexibility, lipophilicity, saturation, size, polarity and solubility indicate that the selected analogues exceed the pink area zone by one parameter, signifying its good predicted oral bioavailability.

Table 5 ADME properties of the most active compounds.

Entry	4a	4b	4d	Entry	4a	4b	4d
Physicochemical Properties /Lipophilicity/Druglikeness				Pharmacokinetics
Molecular weight	320.30	334.33	354.75	GI absorption	High	High	High
Num. heavy atoms	24	25	25	BBB permeant	No	No	No
Num. arom. heavy atoms	21	21	21	P-gp substrate	No	No	No
Fraction Csp3	0.06	0.11	0.06	CYP1A2 inhibitor	Yes	Yes	Yes
Num. rotatable bonds	3	3	3	CYP2C19 inhibitor	Yes	Yes	Yes
Num. H-bond acceptors	5	5	5	CYP2C9 inhibitor	No	Yes	No
Num. H-bond donors	0	0	0	CYP2D6 inhibitor	No	No	No
Molar Refractivity	87.41	92.38	92.42	CYP3A4 inhibitor	No	No	No
TPSA (Å2)	82.92	82.92	82.92	Log Kp (cm/s)	−6.72	−6.55	−6.48
Consensus Log Po/w	2.06	2.47	2.54	Radar chart
Lipinskiˈs Rule	Yes	Yes	Yes
Veber	Yes	Yes	Yes
Egan	Yes	Yes	Yes
Ghose	Yes	Yes	Yes
Muegge	Yes	Yes	Yes
Bioavailability Score	0.55	0.55	0.55
Medicinal Chemistry
PAINS	0	0	0
Brenk	0	0	0
Leadlikeness	Yes	Yes	No
Synthetic accessibility	3.36	3.42	3.35
BOILED-Egg graph

The BOILED-Egg graph (Table 5) clearly specifies that compound 4a, 4b and 4d are not a substrate for P-glycoprotein (PGP^-) which diminished the possibility of their resistance by tumor cell lines through efflux.

3.4.2 Molecular docking: Receptor-ligands interactions

After identifying the activity and the pharmacokinetics proprieties of the most active compounds, they also were chosen for a virtual screening by molecular docking and dynamic simulation for predicting their interactions mode in order to elucidate their possible mechanism of action. The different targets were selected as follows: S. aureus tyrosyl-tRNA synthetase enzyme was selected as an emerging and attractive enzyme for finding new antibacterial agents. The monomeric enzyme, N-Myristoyltransferase (NMT) that catalysis the transfer of the myristoyl group of myristoyl-CoA to the N-terminal glycine of various eukaryotic cellular proteins have been demonstrated for its viability of pathogenic fungi, such as C. albicans and C. neoformans. Thus, NMT would be a potential target for the discovery of novel fungicidal drugs. On the other hand, cyclooxygenase (COX-2) is widely known for its crucial role in converting arachidonic acid to prostaglandins (PG), that play an important role in the process of inflammation. Therefore, COX-2 has been validated as a molecular target for treating inflammatory diseases. Concerning the enzyme Human Peroxiredoxin 5, known as a novel thioredoxin peroxidase, which was widely expressed in mammalian tissues suggesting its important role as antioxidant in organelles that are major sources of ROS, has been selected as a potent antioxidant target. The molecular docking results (Table 6) on the bacterial target S. aureus tyrosyl-tRNA synthetase suggested that compound 4a had been shown to have the highest binding energy at −5.9 kcal/mol; from which conventional hydrogen bonding interactions of the amino acid residues Asp40, and Gln190, of the target with the ligand 4a. Furthermore, hydrophobic interactions were also observed between His50 of the bacterial target and the same compound.

Another potent compound, 4b, interacted with significant hydrogen and hydrophobic interactions, which accounts for its binding affinity towards both the targets, C. albicans N-Myristoyltransferase and human COX-2 enzymes; the docking scores were −7.98 kcal/mol and −9.142 Kcal/mol with 1IYL and 1PXX, respectively.

The central triazole ring of compound 4b and the phenyl groups of Tyr225 and Tyr354 residues generate three hydrophobic interactions in the C. albicans N-Myristoyltransferase protein; another hydrophobic interaction exists between the 2H-benzo[d][1,3]oxazine-2,4(1H)-dione and the phenyl group of Tyr354. In compound 4b, the carbonyl oxygen in the 2H-benzo[d][1,3]oxazine-2,4(1H)-dione scaffold is implicated in hydrogen bonding with Tyr335 (Fig. 4). Hydrophobic interactions with Tyr225 is also seen in the co-crystallized ligand, which supports the potency of compound 4b towards the C. albicans N-Myristoyltransferase protein. With a binding affinity of −9.14 kcal/mol (ΔG = -64.17 kcal/mol), compound 4b has the second-highest binding affinity with the human COX-2 enzyme. The docking scores and ΔG of the same molecule with human Peroxiredoxin 5 are −3.67 and −48.03Kcal/mol, respectively. In the human COX-2 enzyme, compound 4b made one hydrophobic interaction with the 2H-benzo[d][1,3]oxazine-2,4(1H)-dione ring and the phenyl ring of residue Trp387 (Fig. 5). Compound 4d, which comprises a 2-chlorophenyl group in the 1,2,3-triazole structure, has a binding affinity of −5.85 kcal/mol with the bacterial S. aureus tyrosyl-tRNA synthetase target, and −7.95, −8.94, and −3.76 kcal/mol with the fungal and human targets 1IYL, 1PXX, and 1HD2, respectively.

Close

Fig. 4

Binding of compound 4b into the cavity of C. albicans N-Myristoyltransferase (PDB ID: 1IYL).

Export to PPT

Close

Fig. 5

Binding of compound 4b to into the cavity of Human COX-2 enzyme (PDB ID: 1PXX).

Export to PPT

3.4.3 Molecular dynamic (MD) simulation

MD is a comprehensive analysis tool for determining the structural stability and dynamics of proteins caused by ligand/inhibitor binding. In this study, we have conducted a 100 ns MD simulation of the best scored dock complex (4b-Human COX-2 enzyme). The MD simulation trajectory was used to assess key parameters, including the protein Cα atoms RMSD, RMSF, ligand atoms RMSD, and protein ligand interactions.

RMSD is defined as a difference in the structure of a protein or protein–ligand complex from a reference structure, typically the beginning frame. Assessing the protein's RMSD over the simulation can provide details on its structural conformation, indicating the protein's stability and whether the simulation has stabilized. Ligand RMSD can reveal the ligand's stability with respect to the protein, as well as the variation of its intrinsic conformation. The RMSD is a key quantitative measurement in which small deviations suggest stable conformation and vice versa (Acar Çevik et al., 2022, Abdelhamid et al., 2018, Ahmad et al., 2021, 2022, Boulaamane et al., 2022, D. E. Shaw Research, 2021, Girase et al., 2022, Ghosh et al., 2021, Malani et al., 2021, Patel et al., 2018, 2022, Pawara et al., 2021a, 2021b). Fig. 6A shows the RMSD of Cα atoms of 1PXX (protein) in teal color and compound 4b (ligand) RMSD in a brown line. Throughout the simulation time, no major changes in RMSD values of protein and ligand were noticed as compared with the initial frame. The average RMSD values for 1PXX alone and for the ligand in relation to protein were 2.64 Å and 2.89 Å, respectively. The formation of a stable 4b-1PXX complex is expected because the variance in RMSD values of the protein and ligand with respect to protein was substantially lower than the acceptable limit of 3.0 Å.

Close

Fig. 6

MD simulation analysis of compound 4b in complex with Human COX-2 enzyme (PDB ID: 1PXX) (A) RMSD (Protein RMSD is shown in teal while RMSD of compound 4b are shown in brown) (B) Protein RMSF (C) 2d Interaction diagram and (D) Protein–ligand contact analysis of MD trajectory.

Export to PPT

During simulation, the stability of the protein–ligand complex is influenced by the individual amino acid residues of the protein. The RMSF parameter can be used to investigate the residue fluctuations during the simulation. The RMSF value was derived from the MD simulation trajectory and is presented in Fig. 6B. With the exception of the loop region and the C-terminal Overall, the amino acids were stable during the simulation window with an RMSF range of between 0.60 and 3.50 Å. The amino acids that have made contact with the ligand during the simulation trajectory include: Met113, Val116, Leu117, Arg120, Ile345, Tyr348, Val349, Leu352, Ser353, Tyr355, Leu359, Leu365, Tyr385, Trp387, Arg513, Phe518, Val523, Glu524, Ala527, Leu531, Leu534, and Met535. The amino acids that fluctuate more are Asp53 with an RMSF of 3.49 Å, and Trp139 with an RMSF of 3.074 Å and Ile274 with an RMSF of 2.144 Å. It's believed that such high fluctuations in RMSF are acceptable since there are no significant molecular interactions with bound ligand in that region, and the structure is highly flexible. As a result, the stability of the protein–ligand complex was certainly explained by low RMSF values at ligand-contact residues.

Fig. 6C and 6D show the ligand 2D interaction and the histogram depicting the interactions type of compound 4b with 1PXX throughout the simulation, respectively. Carbonyl oxygen of 2H-benzo [d][1,3]oxazine-2,4 (1H)-dione is involved in amino acid mediated (Arg120 and Glu524) hydrogen bonding. On the other hand, charged positive residue Arg120, makes an π-cation contact with the core triazole ring for 23% of the simulation time. Major hydrophobic interactions are observed with the amino acids Met113, Leu117, Ile345, Val349, Leu352, Trp387, Phe518, Val523, Leu531, and Leu534. Three hydrogen bonds are observed between the Arg120, Tyr355, and Tyr385 residues in the 4b-1PXX complex. Overall, the simulation revealed more hydrophobic interactions, and amino-acid mediated water bridges stabilized the compound 4b in the cavity of the 1PXX protein.