Synthesis, characterization, and in silico studies of 1,8-naphthyridine derivatives as potential anti-Parkinson’s agents

Abstract 1,8-Naphthyridine scaffold is a nitrogen-containing heterocyclic compound known for its versatile biological activities. The structure-activity relationship (SAR) has shown that modification at the 3rd position of the nucleus with various secondary amines enhances the binding efficiency and potency towards the Adenosine receptor (A2A type). In this paper, we have reported some newly synthesized derivatives of 1,8- Naphthyridine, and the prepared compounds were assessed for their potential to constrain A2A receptors through molecular docking. Based on the SAR studies, modifications were done at the 3rd position of the nucleus by incorporating secondary amines. The synthesized compounds were characterized by FT-IR, 1H and 13C NMR. All the synthesized compounds 10a-f and 13a-e showed good binding efficiency towards the A2A receptors and might act as an A2A receptor antagonist, as predicted by in-silico studies. 1-Ethyl-7-methyl-3-(pyrrolidine-1-carbonyl)-1,8-naphthyridine-4(1H)-one (10c) in first series showed the highest docking score of −8.407 and binding energy (MMGBSA dG bind) of –56.60 kcal/mol and N-(4-2-diethylaminoethoxyphenyl)-1-ethyl-7-methyl-4-oxo-1, 4, 4a, 8a- tetrahydro-1,8-naphthyridine-3-carboxamide (13b) showed the highest docking score of –8.562 and free binding energy (MMGBSA dG bind) score of –64.13 kcal/mol which was comparable to the bound ligand. MD simulations study also suggested that compounds 10c and 13b would form stable complex human A2A receptor. These findings need to be validated by further in vitro assays. Communicated by Ramaswamy H. Sarma


Introduction
Recent report from World health organization (WHO) revealed that stressful lifestyle affects the quality and psychosocial well-being of people. Approximately 56.8 million people are affected by neurodegenerative diseases and represent a major threat to human health. It has been estimated that this number will be increased 4-fold by the year 2050 (Kannappan et al., 2011;Vizza et al., 2019;Winiarska-Mieczan et al., 2020). Parkinson's disease (PD) that has been characterized by features which arises as a result of degeneration of dopaminergic and nondopaminergic neurons (Pringsheim et al., 2014;Rizek et al., 2016) is the most frequent neurodegenerative disorder after Alzheimer's disease. It is deliberated that oxidative stress in the substantia nigra (SN) plays a noteworthy function in the loss of neurons which produces dopamine (DA). This results in the distinctive deficit of dopamine in SN which has been observed in the post-mortem brain of PD patients (Tysnes & Storstein, 2017).
The characteristic features associated with PD are the immense shortage of dopaminergic neurons in substantia nigra and accrual of Lewy bodies in the central nervous system (CNS). Despite that, the main etiology of Parkinson's is yet not clear (Fox et al., 2018;Moustafa et al., 2016). In present days, only few effective pharmacologically active drugs like levodopa (1), benserazide (2), selegiline (3), rasagiline (4), entacapone (5), tolcapone (6), ropinirole (7) and pramipexole (8) are used in the management of PD, but permanent cure is still questionable (Diaz & Waters, 2009;Ellis & Fell, 2017;McFarthing et al., 2020;Munchau & Bhatia, 2000;Schapira, 2009). The chemical structures of these drugs are shown in Figure 1. All the above-mentioned drugs directly help to improve the level of dopamine, but the long-term use of these drugs is associated with a high risk of dyskinesia. Research is going on in the field of the non-dopaminergic drugs to avoid the major side effects associated with dopaminergic-based drugs. Adenosine receptor (A2A type) takes special attention in the management of neurodegenerative disorders like Parkinson's. Blockade of the A2A subtype of adenosine receptor directly affects the D2 receptor and enhances the level of dopamine in substantia nigra. Nowadays, there is a likelihood of treating acute or chronic neurological disorders like Parkinson's disease by A2A antagonists as a new pharmacological tool (Azam et al., 2009;Richardson et al., 2006). 1,8-Naphthyridine (nitrogencontaining heterocyclic compound) is known for its multipurpose biological activities. Versatile biological activities of 1,8-naphthyridine scaffold has received special attention by the researchers to develop some novel molecules (Ojha et al., 2021). Recent researches showed that 1,8-naphthyridine derivatives possess neuroleptic properties in addition to cholinesterase inhibition, calcium ion regulators, antimicrobial, anti-cancer, anti-allergic, antioxidant, EGFR inhibition, inhibition of platelet aggregation, protein kinase inhibition, ionotropic agent, adenosine receptor agonist, MDR modulator, anti-secretory, bronchodilator, anti-convulsant, adrenoceptor antagonist, anti-malarial, antiinflammatory, anti-mycobaterial, and antihypertensive activity (Gurjar & Pal, 2019;Madaan et al., 2015). It is believed to be one of the therapeutic agents which have a great application for the management of neurodegenerative disorders. Nalidixic acid (1ethyl-7-methyl-4-oxo-[1,8]-naphthyridine-3-carboxylic acid) possess structure resemblance to 1,8-naphthyridine. Therefore, we further proposed to synthesize some nalidixic acid-based derivatives to present a new family of molecules for the treatment of neurodegenerative diseases (Lin et al., 2008).
The structure activity relationship of nalidixic acid shows that modification at the 3rd position of the nucleus with various secondary amines ( Figure 2) enhances the binding efficiency and potency towards the A2A receptor. In this paper, we have reported some newly synthesized derivatives of 1,8naphthyridine, and the prepared compounds were assessed for their potential to constrain A2A receptors through molecular docking (Ahmed & Kelley, 2017;Ball, 2000;Pinna et al., 2005).

General
All the chemicals used for the synthesis purpose were of analytical grade. Silica coated TLC (thin layer chromatography) was used to monitor the reaction and visualized by ultraviolet radiations. The percentage yields reported were based on pure products and were not optimized. 1 H and 13 C NMR spectra were recorded on a Bruker Avance II 400 MHz spectrometer in CDCl 3 or DMSO-d6 solvents at 400 and 100 MHz, respectively. All chemical shifts (d) were expressed in parts per million (ppm) relative to the standard TMS, and the peak patterns were indicated as (s) singlet, (d) doublet, (t) triplet and (m) multiplet.

General method for the synthesis of compound 10a-f
To a stirred solution of nalidixic acid (1) (500 mg, 2.15 mmol; 1 eq) in DCM (10 mL), added amines (dimethylamine, diethylamine, pyrrolidine, piperidine, morpholine and methylpiperazine) (2 eq, 2 M in THF) and DIPEA (1 mL, 6.46 mmol; 3  eq) at 0 C followed by T3P (2 eq, 50% sol in Ethyl acetate) under Nitrogen atmosphere and stirred at room temperature (RT) for 50 mins. After completion of reaction (monitored by TLC), the reaction mixture was evaporated under reduced pressure, diluted with sat. NH 4 Cl (20 mL) and extracted with ethyl acetate (50 mL Â 2). The combined organic layer was dried over Na 2 SO 4 and evaporated under reduced pressure to get crude residue. The crude residue was triturated with diethyl ether (10 mL Â 2) and filtered over sintered funnel to get the desired products (10a-e).

General method for the synthesis of compounds 13a-e
Intermediate (12; 1 eq) was dissolved in 40 mL of ethyl methyl ketone and potassium carbonate (2eq) was added with stirring to absorb the moisture from the reaction mixture. After 1 h of stirring, when a slurry was formed then the chlorinated amines (3 eq) was added with reflux at 80-100 C for 15-20 h. After the completion of reaction (monitored with TLC using ethyl acetate as a mobile phase) excess of solvent was removed using rotary evaporator. The semisolid product so obtained was washed with diethyl ether (3 times) to obtain the solid product (13a-e). The product was further dried completely using rotary evaporator and kept in desiccator for further use as such.

Molecular modelling studies
Molecular docking studies and drug-likeness analysis was carried out using Maestro (Version 11.6), which is a commercial software by Schrodinger, LLC on  Sastry et al., 2013). The protein preparation is a three-step process which involves import and refine, review and modify and minimization using OPLS3e force field.

Ligand preparation
The structures of the synthesized molecules were drawn by employing 2D sketcher tool of the Maestro suite and thus generated ligands were put for ligand preparation with the help of LigPrep tool of Maestro (Sankhe et al., 2021). In this step the minimized energy state of the ligands in 3D were generated at pH 7.0 ± 2.0 by using OPLS3e force field.

Ligand docking
Molecular docking studies in extra precision (XP) mode were carried out by employing GLIDE (Grid-based Ligand Docking with Energetics) module of Maestro . First a receptor grid was generated by using the co-crystallized ligand using the receptor grid generation tool in Glide. A grid is used to generate progressively more accurate scoring of the ligand poses. XP docking is computationally more demanding than standard precision (SP) and high-throughput virtual screening (HTVS) mode. XP mode is less forgiving than SP mode and it can weed out false positives (Friesner et al., 2006).

MD simulation study
To better understand the interaction between top docked compounds and the human A2A receptor, MD simulation study was carried out of 100 ns. The simulation was performed as apo and complexed with compounds using GROMACS 2018.3 on LINUX system with NVIDIA GPU support (Berendsen et al., 1995). For force field AMBER ff99SB-ILDN was used with a water model tip3p. The Antechamber tool of AMBERTOOLS was used to create ligand topology. A cubic unit cell of 1 nm edge was used to fill the water for solvation. Further respective ions (Naþ/ClÀ) were added to neutralize the system. The steepest descent algorithm and conjugate gradient methods were used successively to perform the energy minimization proceeded with equilibration keeping the ligand restrained with a force constant of 10.0 kJ/mol along with protein. Further two-phase equilibration was performed, each for 100 ps at the time steps of 2 fs. First, NVT ensemble to stabilize the temperature at 300 K and NPT ensemble to stabilize the pressure at 1 bar. Now well equilibrated system at desired temperature and pressure was released for position restrain to run the MD for 100 ns with a time step of 2 fs. For holonomic constraints Linc algorithm was selected and for neighbor searching Varlet cutoffscheme was used. For electrostatics calculations vdW and PME were used. The GROMACS in-built tools were used to perform analysis and Xmgrace for plotting the data.

Drug likeness analysis
Drug-likeness of the synthesized compounds computed using QikProp module of Maestro. QikProp predicts physically significant descriptors of the ligands under investigation and employs that for ADMET calculations (Kumar et al., 2020). The pharmacokinetic properties of the synthesized compounds were determined by an overall ADME-compliance score denoted by #stars. Several other properties like molecular weight, hydrogen-bond donors, hydrogen-bond acceptors, partition coefficient in oil/water, percent absorption, violation of rule of five etc. were also computed.
3 Results and discussion

Chemical synthesis
After in depth literature study, we designed and synthesized some novel compounds by modifying the carboxylic group (3rd position) of nalidixic acid (9) with secondary amines. The 1,8-naphthyridine derivatives (Scheme I) were synthesized by replacing the -OH group with secondary amines like diethylamine, dimethylamine, piperidine, morpholine, pyrrolidine and 1-methylpiperazine. To synthesize desired 1,8naphthyridine derivatives, a concise synthetic route was adopted. The hydroxyl group of nalidixic acid was replaced with amines and the formation of amide bond produced the desired novel products (10a-e).

Molecular docking
Most of the compounds showed comparable docking score and interactions with reference to the co-crystallized ligand with the Human A2A Adenosine Receptor (PDB ID 3EML). The co-crystallized ligand showed a docking score of -9.785 with a binding energy (MMGBSA dG bind) of -72.65 kcal/mol as shown in Table 1 and Figure 3(f). The bound ligand showed H-bonding interactions with GLU169 and ASN253 amino acid residues. It also showed hydrophobic interactions with TYR9, ALA63, ILE66, ALA81, VAL84, LEU85, PHE168, MET174, MET177, TRP246 LEU249, MET270, TYR271 and ILE274 residues. Polar interactions were observed with SER67, THR88, ASN253 and HIE250 residues. It also showed charged negative interactions with GLU169 and p-p stacking with PHE168, HIE250 residues. The docking protocol was also validated by superimposing the generated docked pose of the bound ligand over its co-crystallized pose in the X-ray crystal structure and the RMSD (root mean square deviation) value was 0.461.
(continued) was the highest among both series of synthesized compounds. It also showed the crucial H-bond interaction with ASN253 residue and another H-bond interaction with SER67 residue. Its hydrophobic interaction pattern was similar to the bound ligand as it showed possible hydrophobic interactions with ILE66, VAL84, LEU85, LEU167, PHE168, MET174, MET177, TRP246, LEU249, LEU267, MET270, TYR271 and ILE274 residues. Its polar and charged negative interactions were also similar to the bound ligand. The core interaction of p-p stacking was also observed with PHE168 residues. For the rest of the compounds of this series the 3D-ligand interaction diagrams have been depicted in Figure 3(a-e). N-(4-2-dimethylaminoethoxyphenyl)-1-ethyl-7-methyl-4-oxo-1,4,4a,8a-tetrahydro-1,8-naphthyridine-3carboxamide showed the lowest docking score of À5.896 and didn't show any H-bond interactions. Another notable observation was that the compounds of this series seem to be larger than the pocket and hence some solvent exposure was observed.

MD simulations study
The dynamic behavior and stability of Human A2A Adenosine Receptor model and its complex with ligands in aqueous media were studied by performing MD simulation for 100 ns. The information collected by simulation were about conformational changes in protein residues, ligands and change in their interaction patterns, their stability and strength of binding during the entire run. Further RMSD, PCA, RMSF, SASA, H-bonding and radius of gyration were  Figure 5). It has an equilibrium phase i.e., time required to reach the equilibrium and productive phase where average deviations can be analyzed. In the current plot we can clearly observe the apo-protein (black) reached the equilibrium at 10 ns and remained stable during entire run with the average  deviations between 0.5 and 1 nm. Recptor-10c complex attained the equilibrium at around 5 ns and remained stable with intermittent deviations of 0.5-1.5 nm, whereas receptor-13b complex was found to be comparatively more stable which attained the equilibrium at around 3 ns and average standard deviation of 0.5-1.0 nm.

Root mean square fluctuation (RMSF) analysis.
RMSF measures the rigidity of polypeptide chain. The conformational changes in residues and its flexibility to each other can be analyzed by RMSF. The Human A2A Adenosine Receptor structure is barrel of a-helices connected by small loops. In RMSF plot we can clearly observe that the   corresponding fluctuations of apo and complexes are almost similar as flexibility of residues of the same protein would be same. In RMSF plot (Figure 6(a)) both the ends showing the highest flexibility due to free helices attached by a loop 18 to 56 and 451 to 472 (Figure 6(b), green segment). There is one moderate fluctuations between residues number 56 to 91 and one major fluctuation from 120 to 181 due to presence of a small helix between long loops showing variant flexibility. From 181 to 216 there is a trapped helix showing small flexibility but there is an intermittent fluctuation from 220 to 380 as a group of helices are free to float attached with long loop at the ends (Figure 4, yellow). There is a moderate fluctuation from 420 to 450 due to small helix able to fluctuate as bonded by a loop.
3.2.2.3 Solvent accessible surface area (SASA) and intermolecular hydrogen bonding. SASA defines the change in accessibility of protein to the solvent. It helps to analyze the change in folding pattern of inner hydrophobic core to outer hydrophilic surface of protein structure. The SASA range of apo as well as bound complexes lies between 230-260 mm 2 throughout the simulation period (Figure 7(a)). There are no sudden major changes in SASA of apo as well as complexed structure which shows it remained stable throughout the simulation and there was no unfolding in protein structure. Intermolecular hydrogen boding attributes to binding strength of ligands to its protein. The number and distribution of H-bond during the simulation is represented in Figure 7(b). In both the complexes few conformations were showed maximum of 2 hydrogen bonds formed between protein and ligands. However, the density of forming one hydrogen bond was consistent during the simulation. Thus, it shows that ligands can maintain the binding with the site of interaction of receptor during the simulation.

Drug-likeness analysis
As shown in Table 2, in silico ADMET analysis suggested that the synthesized compounds would be drug like molecule as for all the synthesized compounds #stars value was 0. A lower #star value indicates that all the property descriptors calculated by QikProp for these compounds fell within the range of 95% of known drugs. Furthermore, there was no violation of Lipinski's rule of five and hence it also suggests that these compounds followed the general rule of bioavailability for oral drugs. For most of the compounds the percent of oral absorption was within the acceptable range. Compound 10f from the 1st series showed the least absorption of 82.799%. Rest all compounds showed more than 90% oral absorption. Some compounds like 10b, 10c, 10d, and 13d showed the probability of 100% oral absorption.

Conclusion
The newly synthesized 1,8-naphthyridine derivatives possesses anti-Parkinson's activity. All the synthesized compounds 10a-f and 13a-e showed good binding efficiency towards the A2A receptors and act as an A2A receptor antagonist, confirmed by the help of in-silico studies. Compound (10c) in first series showed the highest docking score of -8.407 and binding energy (MMGBSA dG bind) of -56.60 kcal/mol and compound (13b) showed the highest docking score of -8.562 and free binding energy (MMGBSA dG bind) score of -64.13 kcal/mol which was comparable to the bound ligand. MD simulations study also suggest that compounds 10c and 13b would form stable complex human A2A receptor. These findings need to be validated by further in vitro assays.