Identification of new oxospiro chromane quinoline-carboxylate antimalarials that arrest parasite growth at ring stage

Abstract Malaria still threatens half the globe population despite successful Artemisinin-based combination therapy. One of the reasons for our inability to eradicate malaria is the emergence of resistance to current antimalarials. Thus, there is a need to develop new antimalarials targeting Plasmodium proteins. The present study reported the design and synthesis of 4, 6 and 7-substituted quinoline-3-carboxylates 9(a–o) and carboxylic acids 10(a–b) for the inhibition of Plasmodium N-Myristoyltransferases (NMTs) using computational biology tools followed by chemical synthesis and functional analysis. The designed compounds exhibited a glide score of −9.241 to −6.960 kcal/mol for PvNMT and −7.538 kcal/mol for PfNMT model proteins. Development of the synthesized compounds was established via NMR, HRMS and single crystal X-ray diffraction study. The synthesized compounds were evaluated for their in vitro antimalarial efficacy against CQ-sensitive Pf3D7 and CQ-resistant PfINDO lines followed by cell toxicity evaluation. In silico results highlighted the compound ethyl 6-methyl-4-(naphthalen-2-yloxy)quinoline-3-carboxylate (9a) as a promising inhibitor with a glide score of −9.084 kcal/mol for PvNMT and −6.975 kcal/mol for PfNMT with IC50 values of 6.58 µM for Pf3D7 line. Furthermore, compounds 9n and 9o exhibited excellent anti-plasmodial activity (Pf3D7 IC50 = 3.96, 6.71 µM, and PfINDO IC50 = 6.38, 2.8 µM, respectively). The conformational stability of 9a with the active site of the target protein was analyzed through MD simulation and was found concordance with in vitro results. Thus, our study provides scaffolds for the development of potent antimalarials targeting both Plasmodium vivax and Plasmodium falciparum. Communicated by Ramaswamy H. Sarma


Introduction
Malaria still remains one of the most life-threatening vectorborne diseases in human history.It is caused by the infection of the Apicomplexa parasite from the genus Plasmodium transmitted by the female Anopheles mosquito (Pal et al., 2021).Despite well-established control strategies, around 200 million people got malaria diseases with over �400,000 deaths in the year 2020 (World Malaria Report 2020, WHO, Geneva, Switzerland).WHO African region shared the highest burden of 94% of all malaria infections and demises.With the emergence of resistance to antimalarials especially chloroquine and artemisinin-based combination therapy (ACT), it is essential to develop newer drugs as well as effective malaria vaccines to completely eradicate malaria (Cowell & Winzeler, 2019;Fairhurst & Dondorp, 2016;Madhav & Hoda, 2021;Thu et al., 2017).For this WHO called for the urgent development of new potential antimalarials, which can arrest the growth of parasites at different stages of their life cycle to control drug resistance (WHO, 2015).
Plasmodium parasites have a complex life cycle, which is tightly regulated in two hosts: human and Anopheles mosquito.Post-translational modifications (PTMs) such as phosphorylation, methylation, ubiquitination, acetylation, sumoylation and myristoylation are known to regulate critical pathways that enable the progression of malaria parasites from one stage to the next (Doerig et al., 2015).Due to the critical role(s) played by PTMs in regulating the malaria parasite life cycle, many of the enzymes/proteins that are responsible for bringing the PTMs are being considered important targets for new anti-malarial discovery (Doerig et al., 2015).
N-Myristoyltransferase (NMT, EC 2.3.1.97)is an enzyme that is responsible for the transfer of the lipid myristate (C14:0) from myristoyl coenzyme A (Myr-CoA) to specific substrate proteins (Wright et al., 2014).The vital protein, i.e. glideosome associated protein 45 and calcium-dependent protein kinase 1 (CDK1) have been shown to require myristoylation to carry out their biological functions (Rackham et al., 2013).Furthermore, in vivo genetic experiments disclosed that Plasmodium NMT has been essential for parasite (Schlott et al., 2018) and the variation at sequence level between parasite and host NMTs suggests the development of selective inhibitors as new antimalarials (Gunaratne et al., 2000).Based on these findings, Yu et al. (2012) designed and synthesized 3-methyl-4-(piperidin-4-yloxy)benzofuran derivatives as PfNMT inhibitors with IC 50 of 0.27 and 1.2 mM EC 50 for PfNMT on Pf3D7 line, respectively.These authors also developed pyridyl-based inhibitors targeting PfNMT with K i value of 0.0017 mM for PfNMT and EC 50 of 0.21 mM for Pf3D7 line (Yu et al., 2015).Further, Goncalves et al. (2012Goncalves et al. ( , 2017) ) developed 3-butyl-6-methoxy-2-methylquinoline derivative A as Plasmodium NMT inhibitors and later optimized the reported scaffold using a structure-guided approach and achieved apparent K i value of 2.92 and 2.69 mM for PvNMT and PfNMT, respectively referred it as compound B. However, these A and B molecules were not analyzed for their biological activity due to their poor solubility in the buffer.Rackham et al. (2013) explored benzo [b]thiophene derivatives as Plasmodium NMT inhibitors for the efficient treatment of malaria and reported compound C with K i value of 1.30 and 1.32 mM for PfNMT and PvNMT, respectively.The biological efficiency of a compound was reported with EC 50 value of 14.6 mM for the Pf3D7 line.Roberts et al. (2016) developed spirocyclic chromanes that inhibit all intraerythrocytic life cycle stages and reported compound D with EC 50 of 0.35 and 0.39 mM for the PfDd2 and Pf3D7 line, respectively.Together, these results advocate Plasmodium NMTs as highly promising antimalarial drug targets.The present study deals with the design and synthesis of quinoline-based dual inhibitors that bind both PvNMT and PfNMT in silico and inhibit Plasmodium falciparum growth and development in in vitro culture with an aim to develop inhibitors that can act on both P. falciparum as well as on Plasmodium vivax.

Design strategy
Inspired by the structural features of the different active Plasmodium NMT inhibitors, the present idea was conceived to design the new hybrid inhibitors to inhibit both PvNMT and PfNMT.The structure-guided drug design and molecular hybridization approaches were utilized to design the molecules by incorporating the important bioactive structural features of the reported Plasmodium NMT inhibitors.Various medicinal chemistry techniques, such as ring enlargement, hetero atom replacement, aromatic ring substitution, bioisosteric replacement and alkyl chain substitution, were used to design the final compounds as depicted in Figure 1.
The sulfur atom of the core moiety in the compound C was replaced with a nitrogen atom followed by ring enlargement to make it more effective toward H-bond formation due to the more nucleophilic and basic nature of nitrogen, resulting in the quinoline ring (Bagno et al., 1994).Furthermore, the presence of nitrogen also enhances the p-electron density of the quinoline ring and specifically the electron charge of the ring.The enhancement in p-electron density and electron charge is due to the 2p-2p overlap of the nitrogen atom with the aromatic ring of the quinoline moiety.Spirocyclic groups have recently demonstrated their impressive antimalarial potency (Brindisi et al., 2015).
Spiroindolone NITD609, a new antimalarial that entered clinical trials, shows remarkable potency through the inhibition of PfATP4 (Turner, 2016).Similar to this, the strength of the spirocyclic chromanes as antimalarial agents was also emphasized (Iyamu et al., 2022;Roberts et al., 2016).Therefore, the substituted piperidine ring of C was replaced with substituted spiro[chromane-2,4 0 -piperidin]-4-one to achieve p-p stacking and other types of interactions.Furthermore, the other aromatic and non-aromatic bioactive groups were also utilized to design new hybrid molecules to achieve significant Plamsodium NMT inhibitory activity.In distinct compounds, the quinoline and naphthalene groups exhibit antimalarial characteristics (Adigun et al., 2022;Benjamin et al., 2022), and the tetrahydrofurfurylamine group proved crucial in turning the quinazolines into effective antimalarials (Gilson et al., 2017).Similarly, 3-(trifluoromethyl)-5,6,7,8-tetrahydro-[1,2,4]triazolo [4,3-a]pyrazine is an important structural component of the antidiabetic drug sitagliptin phosphate (Ye et al., 2021).Therefore, the substitutions with these groups were made the compounds to best fit in the groove of the active site of the target protein.Further, the carboxylate group remained unchanged at the 3rd position of the core moiety.Bioactivity is greatly influenced by changes made to molecules' structural characteristics (Patel & Kumari, 2022).Therefore, to investigate the role of aliphatic groups on antimalarial activity, methyl and methoxy groups substitution was performed at the 6 and 7 positions of the core quinoline ring.The structure-guided drug design approach was followed to achieve the promising inhibition of the PvNMT and PfNMT for the treatment of two most lethal form of malaria with minimized cytotoxicity.

In silico studies
Due to the unavailability of an experimentally determined structure, the PfNMT protein full-length model has been predicted through different modeling approaches.PvNMT shares more than 80% of sequence identity with PfNMT (Yu et al., 2012).The only two residues, i.e.Y212 and Y334 of PvNMT, situated within 5 Å of the ligand differ and are replaced by F212 and F334 in PfNMT (Figure S1).Further, the BLASTP search with the PDB protein sequences revealed that the amino acid sequences of the other PDB PvNMT structures have 100% identity with that of the PvNMT template (4BBH) (Figure S3).Hence, a single template based on the crystal structure of the PvNMT was used for PfNMT modeling.Different models were generated using (1) homology modeling using Modeller, (2) Robetta Structure Prediction and (3) I-TASSER server and also using AlphaFold.Out of them, the Robetta server predicted the best 3D structure (Robetta Model-1) with a confidence score of 0.81.The best model was selected based on model evaluation tools.Template model superimposition was done by UCSF Chimera v1.11.2 software, showed a root mean square deviation value of 0.995 Å (Figure S4), and proved a significant level of similarity.The accuracy of the models is determined by ERRAT, PROCHECK Ramachandran plot and Verify3D.ERRAT server identifies incorrect regions of protein structures in random distributions of atoms, which can be differentiated from correct distributions.It presented a score of 93.28% for the protein model (Robetta Model-1).PROCHECK server has assessed the stereochemical quality of protein structures considering residue by residue and overall structural geometry.It showed 90.1% and an additional 9.1% in the most favored regions (Figure S5).Validation for the druggability of the obtained PfNMT protein model was performed using the CASTp server and CavityPlus (Figure S6 and Table S1-S3).The amino acid residues available in pocket 1 are also listed in the supplementary information.Molecular docking studies are used for studying protein-ligand interactions that predict several orientations of the ligand with the protein.The best orientation of a molecule bonded to the protein is calculated using the scoring function, called docking scores or binding free energy (Ram� ırez & Caballero, 2018).The docking score is an aggregate of various energies like free energy owing to protein-ligand interactions, conformational changes, internal rotations, solvent effect, association energy of ligand and receptor to form a single complex, and free energy due to vibrational mode changes (Torres et al., 2019).The lowest binding energy/docking scores indicate a stable protein-ligand complex and are considered crucial in structure-based drug design.Therefore, docking studies were used to understand different binding interactions of the designed ligands with Plasmodium NMTs, i.e.PvNMT and PfNMT.As the crystal structure of the PvNMT is available, the ligand-bound crystal structure of the PvNMT (PDB ID: 4BBH) and the homology model of PfNMT were used for docking studies.The crystalbound inhibitor, E (Figure S11) of protein PvNMT (PDB: 4BBH) was taken as control with the experimental ligands in the docking study.Moreover, the redocking with control ligand E with both proteins was performed to evaluate our docking protocol and the results were obtained identical with RMSD value of 0.000.Therefore, the designed ligands were docked with PvNMT and the homology model of the PfNMT and respective docking and glide scores are provided in Table 1.Hydrophobic interactions (light green), polar interactions (sky blue) and p-p stacking (dotted dark green) are shown within the 5 Å of the active site of the enzyme in all docking poses.All compounds fitted well in the groove of both PvNMT and PfNMT and displayed strong interactions with key residues of the active site of the enzyme (Figures 2-5 and S7-S10).
Docking results revealed that Phe105, Tyr211 and Tyr334 of PvNMT shows strong p-p stacking while Tyr334 and Ser319 were involved in making H-bonds with the different ligands.The glide scores of all the docked compounds are listed in Table 1.Among these compounds, 9a, 9b, 9g, 10a and 10b strongly interact with the active site residues of The obtained docking scores were validated by plotting the graph between the predicted and experimental IC 50 value of the compounds for the Pf3D7 line (Figure S13).However, the calculated coefficient of determination, R 2 (R-squared) value was 0.7646 which indicates that the proportion of the variation in the IC 50 values and docking scores is predictable.
PvNMT and show the glide score of À 9.084, À 7.890, À 8.624, À 8.296 and À 9.241 kcal/mol, respectively.Furthermore, Phe105, Tyr211, His213 and Phe226, of modeled protein PfNMT are involved in making strong p-p stacking, while Ser319 and Ser387 showed H-bond interaction with the ligands.The compounds 9a, 9d and 9j showed the highest glide score of À 6.975, À 7.039 and À 7.538 kcal/mol for PfNMT protein, respectively.Moreover, the docking studies of compounds 9n and 9o revealed that both compounds are welloccupied in the groove of PvNMT and PfNMT (Figures 4 and  5) with alike glide scores.Thus, the docking results indicated that the designed compounds are best fitted into the groove of both Plasmodium NMTs and show strong interactions with the key residues of the target protein(s).The similarity between the docking poses of control inhibitor E and experimental ligands was measured by calculating root-meansquare deviation (RMSD) and superimposed images were shown in Figure S12.These results thus supported the chemical development of the designed ligands for PvNMT and PfNMT.NMR, and Mass spectra and reported in Table 2.The crystal structures were further characterized by X-ray crystallography.TLC and spectral analysis supported the purity of the compounds.

NMR and mass spectral studies
The purity and preparation of these compounds 9(a-o) and 10(a-b) were determined from their proton ( 1 H) and carbon ( 13 C) NMR spectral studies (Figures S14-S47).The aromatic hydrogen protons appear in the range of 9.32 À 7.02 ppm whereas the aliphatic protons range from 4.03 to 0.67 ppm (Rackham et al., 2014).The 13 C NMR spectra showed typical aromatic carbon atoms peaks in the range of 170 À 110 ppm whereas aliphatic methyl carbons peak within 40 À 15 ppm.
The peaks observed at �61 to 62 ppm and in 13 C NMR correspond to a carbon atom that is next to carbonyl carbon attached to oxygen atoms of all the compounds (Figures S31-S47) (Kumar et al., 2016).All the newly formulated compounds 9(a-o) and 10(a-b) compositions were investigated by HRMS spectrometry (Figures S48-S64).

Crystallography study
The molecular structures of the compounds were further ascertained from X-ray crystallography (Figure 6, Table 3) and evaluated as described by Umar et al. (2019).Compounds 9a and 9b both crystallized in the monoclinic having space group P 1 21/n 1 while 9d also crystallized in the monoclinic having space group C 1 2/c 1, whereas 9c crystallized in the orthorhombic having space group P n c 2 with water as a solvent molecule.The unit of the crystal lattice of 9a, 9b, 9c and 9d has one molecule in the crystallographic asymmetric unit.The unit cell packing diagram of 9b and 9d showed chemically significant short contact interactions between the hydrogen atoms of the methoxy group with the carbonyl oxygen and the nitrogen atom (Figures S65-S68).The crystal data of all new small molecules has been submitted to the Cambridge Crystallographic Data Centre (CCDC) with submission ID of 2192668, 2192669, 2192761 and 2192670 for compounds 9a, 9b, 9c and 9d, respectively.

In vitro anti-plasmodial activity
NMT plays a vital role in the life cycle of the plasmodium parasite and its inhibition leads to the death of the parasite (Wright et al., 2014).In the present study, quinoline-based molecules were identified by molecular docking and screened for their antiplasmodial activity against both the CQ-sensitive Pf3D7 line and the CQ-resistant PfINDO line in vitro.The corresponding IC 50 values of the screened compounds are reported in Table 4.Most of the compounds exhibited micromolar to submicromolar range antiplasmodial activity within the range of 3.96 À 66.23 mM.Evaluation of the antiplasmodial activity with the structural features of the screened compounds disclosed exciting structure-activity relationships (SARs), which would be useful in developing the new scaffold for malarial treatment.The bioactivity was strongly related to the substitution of core quinoline moiety at different positions as well as with the substitution in the attached aromatic system at position 3. SAR with in silico results revealed that the compound 9b, with methoxy group at 6th position, carboxylate group as an aliphatic system at 3rd position, and naphthalene ring as an aromatic system were the best fit into the groove of PvNMT and also have good interaction with the model protein PfNMT with the IC 50 value of 10.92 mM for Pf3D7 line.To analyze the relation of substituents at different positions of the quinoline ring with antiplasmodial potency, the aliphatic substituents with two atom lengths at the 6th position were replaced with one atom length methyl group.The resultant compound 9a showed increased potency against the P. falciparum parasite with an IC 50 value of 6.58 mM in in vitro culture and a higher glide score for PvNMT in silico.The compounds 9a and 9b thus appear to have inhibition activities for PvNMT and PfNMT with moderate antiplasmodial activity against the P. falciparum parasite.Further, the methyl group of 9a was replaced with the fluoride group to remove the substituted aliphatic chain at position 3 and the resultant compound 9e exhibited the bioactivity profile against the parasite with an IC 50 of 37.38 mM.The results indicated that the reduction in length and size of the substituted aliphatic group or positional replacement of the methoxy group is not tolerable at this position of core quinoline moiety, hence a vast change in IC 50 value has been observed.
The second observation was done at position 3 of the quinoline ring with the carboxylate group.The carboxylate group of 9b and 9a was converted into the carboxylic acid, resulting in compounds 10b and 10a, respectively.This substitution resulted in a compound with decreased aliphatic chain length, imparting the acidic character and increasing the polarity, which played an important role in determining the bioactivity profile of the compounds (Ertl et al., 2020).Thus, the molecular interactions with PvNMT were increased but the antiplasmodial activity of the compounds 10b got reduced with an IC 50 value of 66.23 mM for the Pf3D7 line while 10a lost its activity.It is possible that these compounds may be beneficial for P. vivax but were not effective against P. falciparum malaria.The results indicated that the change into the carboxylate group at position 3 of the core quinoline ring was not favorable to achieving the potential antiplasmodial activity of the molecules.Finally, the effect of substituted aromatic or non-aromatic systems with different linkers at the 4th position of core moiety was examined by replacing the naphthalene ring with a quinoline ring, fivemembered tetrahydrofuran ring with one carbon amine linker, spiro[chromane-2,4 0 -piperidin]-4-one and 3-(trifluoromethyl)-5,6,7,8-tetrahydro-[1,2,4]triazolo[4,3-a]pyrazine groups.Antiplasmodial results indicated that the compound 9n, with methoxy group at the 7th position, carboxylate group as an aliphatic system at the 3rd position, and tetrahydrofurfurylamine at the 4th position was found potent against CQ-sensitive Pf3D7 line with an IC 50 value of 3.96 mM while compound 9o, substituted with spiro[chromane-2,4 0 -piperidin]-4-one at 4th position was most active against CQ-resistant PfINDO line with an IC 50 value of 2.8 mM.According to the literature, both chemical groups are crucial in adjusting the compounds' antimalarial potency (Gilson et al., 2017;Iyamu et al., 2022).Both compounds, 9n and 9o also showed good glide scores against PvNMT as À 7.639 and À 7.876 kcal/mol, respectively, superior to the glide score against model protein PfNMT.These results indicated that these compounds may also be equipotent against P. vivax parasite.Positional replacement of the substituent methoxy group from the 7th to 6th position of the quinoline ring in the corresponding compounds resulted in compounds 9j and 9k, leading to a decrease in the activity to 12.38 and 9.40 mM, respectively for Pf3D7 line.Substitution of the methyl group at the 6th position resulted as compounds 9g and 9h remarkably affected the bioactivity profile and the potency against the parasite, decreasing up to 18.9 mM.
It was observed that the substitution at the 4th position of the core quinoline moiety with different aromatic and non-aromatic systems was proven an important factor to improve the antimalarial potency of the compounds.Observation revealed that the replacement of tetrahydrofurfurylamine with spiro[chromane-2,4 0 -piperidin]-4-one at the 4th position in compound 9n decreases the activity against the CQ-sensitive line while improving the activity against CQresistant line.Moreover, the substitution of b-naphthol and 3-(trifluoromethyl)-5,6,7,8-tetrahydro-[1,2,4]triazolo[4,3-a]pyrazine made compounds almost inactive against the parasite.
ORTEP crystal structures of compounds 9a and 9d indicated that the lowest energy state and orientation of the compounds were different which may affect the interaction of the compounds with the target protein and may not effectively inhibit NMT, resulting in a less biologically active molecule.This phenomenon was also supported by the docking images of the screened compounds with the target protein.The SAR indicated that the optimum length of the substituted aliphatic groups at both, the 3rd and 6th position, appropriate positional substitution of the methoxy group of a core moiety, and selective substitution at the 4th position by taking aromatic and non-aromatic groups were necessary to achieve the promising antiplasmodial activity.The results were also in good agreement with the hemolysis assay, hence the compound 9b was marked as a dual inhibitor of PvNMT and PfNMT with moderate antiplasmodial activity and minimum toxicity while 9n and 9o were marked as the best compounds of the study against P. falciparum CQsensitive and CQ-resistant line, respectively, which showed promising anti-plasmodial activity.Therefore, further studies are required to explore the insight and to increase the antimalarial potency from submicromolar to nanomolar rage.

Effect of compounds on the maturation of Pf3D7
Since developed Plasmodium NMT inhibitors show promising anti-plasmodial activity on both, CQ-sensitive Pf3D7 and CQresistant PfINDO line, we analyzed the time course effect of the compounds on the maturation of Pf3D7 parasites.Pf3D7 culture was treated with IC 90 concentration of compounds 9a (59.1 mM), 9h (88.9 mM), 9k (84.5 mM), 9n (35.5 mM), 9o (60.3 mM) and solvent 0.4% DMSO as a control.The microscopic observation of the morphological changes in Figure 7 indicated that control-treated parasites grew normally and matured into functional schizonts.
Further, microscopic observation deduced that the test compounds prevented the maturation of Pf3D7 at their IC 90 .It was revealed that compounds 9a, 9h and 9k delayed the  maturation of trophozoites and prevented them to form functional schizonts.Compounds 9n and 9o stopped the development of the parasites at the ring stage and prohibited trophozoite formation.Therefore, it can be concluded that the compounds developed in the study not only led to the death of parasites, but they also had the ability to inhibit the maturation of parasites.

Hemolysis assay and ADME prediction
Human RBCs are one of the best cells for evaluating the toxicity of compounds.Hence, the in vitro hemolytic assay was performed to study the toxicity of test compounds on the human host.After incubating the RBCs in the presence of test compounds (9a-9o and 10a-10b), it was observed that the lysis of erythrocytes increased.CQ was used as positive control and it showed very less toxicity.Among all tested compounds, 9b showed less toxicity.At 25 mM concentration, 9b showed only 0.38% lysis while at the highest concentration, i.e. 200 mM, 9b showed nearly 13% cell lysis.Also, 9c showed the highest toxicity on RBCs among all screened compounds.At 25 mM, nearly 9% lysis was observed while at 200 mM, 9c showed approx.50% cell lysis.The best compounds 9n and 9o of the study showed lysis of 5.64% and 7.51%, respectively which was below 10% and not toxic at the test concentration.Figure 8 showed the hemolysis data of the best five compounds of the study.The hemolysis data for the remaining other compounds could be accessed from the supplementary information (Figure S69).
Different pharmacokinetics parameters, i.e. absorption, distribution, metabolism and excretion (ADME) and drug-likeness were predicted by the QikProp module of the Schr€ odinger 2021-2, and the results were discussed in Table 5.To examine the biomolecular similarities and differences between the synthesized compounds and the marketed antimalarial medications, such as primaquine (PQ), chloroquine (CQ), artemether (AR) and artesunate (AE), different characteristics of these pharmaceuticals were also included.The data of Table 5 shows that ADME parameters of most of the synthesized molecules fall within the range recommended by Schr€ odinger based on 95% of known drugs in its database.
The QikProp calculated total solvent accessible surface area (SASA) which represent to surface of the ligand where solvent molecule can be contact (Konstantinidis et al., 2021).The Schr€ odinger suggested identical range for SASA 300.0 À 1000.0 square Å.Similarly, hydrophobic component of the SASA (FOSA, 0.0 À 750.0), hydrophilic component of the SASA (FISA, 7.0 À 330.0) exposed due to other atoms were calculated (Silva et al., 2019).QPlogS (-6.5 to 0.5 mol/dm 3 ) and QPlogK hsa (-1.5 to 1.5) are prediction of aqueous solubility and binding to human serum albumin were calculated to predict the same.The prediction of percent human oral absorption (%HOA) >80% is considered high, and <25% is considered poor.The drug likeness of the compounds were predicted using rule of five and rule of three.The rule of five is defined as molecular weight < 500, QPlogPo/w < 5, donorHB � 5, accptHB � 10.Those The MW, HBD, HBA and QPlogPo/w were calculated using the QikProp module of the Schr€ odinger suit.MW, molecular weight; HBD, hydrogen-bond donors; HBA, hydrogen-bond acceptors; QPlogPo/w, predicted octanol/water partition coefficient (-2.0 to 6.5).
compounds who follow rule of five are considered to be more drug-like (Benet et al., 2016).Similarly, rule of three defined as QPlogS > À 5.7 mol/dm À 3 , QPPCaco > 22 nm/s (predicted apparent Caco-2 cell permeability), and #primary metabolites (#metab) < 7.Those compounds who follow rule of three are considered to be more orally available (Lionta et al., 2014).

Molecular dynamic simulation
A 100 ns MD simulation study was performed to examine the conformational stability of PvNMT alone and PvNMT after binding with both, the test ligand 9a and control ligand E with the active site of the enzyme.To describe the structural stability of protein, we have shown the time-dependent graph of RMSD, R g , RMSF and SASA, shown in Figure 9.

Root mean square deviation (RMSD)
The structural and conformational changes of the protein during the simulation are calculated by RMSD, shown in Figure 9(A).It also describes the equilibration and stability of the protein.Here we have calculated the C a atoms RMSD for all three systems.For the PvNMT system, the protein attains the equilibrium state after 15 ns, and after that, it shows the constant RMSD value of around 1.5 Å.For complex PvNMT-9a, the RMSD value increases initially and reaches the equilibration after 70 ns.RMSD value for the PvNMT-E system shows constant throughout the simulation after 5 ns.PvNMT-  9a system shows the maximum changes compared to other systems, and the PvNMT-E system shows the least changes.

Radius of gyration (R g )
To examine the extent of folding of the protein, R g is measured, which evaluates the effective size of the protein.R g of the protein backbone atoms for all the systems for comparing the folding nature is shown in Figure 9(B).The PvNMT-9a shows the highest R g value, which means it is in an extended state compared to other systems.PvNMT and PvNMT-E systems show almost similar R g values around 21.3 Å.In the presence of ligand 9a, PvNMT shows slightly extended conformation compared to the compact structure given by ligand E maintained throughout the simulation period.

Solvent accessible surface area (SASA)
The surface area of the solute protein over which the contact between the protein and water molecules takes place.Here we have shown the SASA of the protein in Figure 9(C) and observed that SASA, on average for the PvNMT system, is showing the value �192.53 nm 2 .For PvNMT-9a and PvNMT-E systems, the SASA value is 190.72 and 189.82 nm 2 , respectively.PvNMT system is showing a slightly larger value as compared to other systems due to the presence of ligands.
The larger value of SASA means larger interaction of the protein with the solvent molecules.

Root means square fluctuation (RMSF)
The flexibility of the protein residues in PvNMT is affected by the presence of ligands 9a and E. We calculate the flexibility  by computing the RMSF as shown in Figure 9(D).It is observed that the residues of the PvNMT-9a system show the maximum fluctuation, which is well correlated with the higher RMSD of this system.The PvNMT is stable in the presence of ligand E and shows almost the same flexibility of residues as that of the PvNMT system.

Conclusion
A structure-based drug design coupled with molecular hybridization strategy was utilized to design new quinoline scaffold-based dual inhibitors of PvNMT and PfNMT enzymes for the treatment of P. vivax and P. falciparum malaria.The designed molecules were analyzed in silico through docking studies with PvNMT and the homology model of PfNMT, followed by chemical synthesis and biological screening against CQ-sensitive Pf3D7 and CQ-resistant PfINDO strain in vitro.
The docking study revealed that the compounds completely fit into the groove of PvNMT and PfNMT enzymes and shows strong interactions with the amino-acid residues of the ligand binding pockets.The compounds 9a shown strong interactions with the amino-acid residues of PvNMT with the glide scores of À 9.084 kcal/mol while À 6.975 kcal/mol for PfNMT.The in vitro antimalarial screening indicated that most of the compounds possessed submicromolar IC 50 values and compound 9n, having tetrahydrofurfurylamine as a substituted non-aromatic group at position 4 of core quinoline moiety and carboxylate group as the substituted lipophilic chain was found most potent against Pf3D7 with the IC 50 value of 3.96 and 6.38 mM against PfINDO line.The compound 9o showed promising activity against the CQ-resistant PfINDO line with the IC 50 value of 2.8 and 6.71 mM for the Pf3D7 line.The time course effect of the compounds on the maturation of Pf3D7 parasites concluded that the compounds developed in the study not only led to the death of parasites, but they also had the ability to inhibit the maturation of parasites.The hemolysis assay revealed that compound 9b shows 0.38% lysis at 25 mM concentration while 9n and 9o were also safe at 25 mM concentration with 5.64% and 7.51% cell lysis.Further, an MD simulation study was performed to investigate the conformational stability of PvNMT after binding of compound 9a with the active site of the enzyme.Among all the synthesized compounds, four compounds 9a-9d were successfully crystalized and analyzed by single-crystal X-ray crystallography.The results supported that the synthesized compounds may act as dual inhibitors of PvNMT and PfNMT for the treatment of P. vivax and P. falciparum malaria.The reported compounds may be considered for the further development of new antimalarials with a novel mechanism of action as well as non-toxic for a safe therapy.

Method and materials
All the chemicals and starting materials were purchased from commercial sources and were used as received or synthesized via literature procedures.The reactions were monitored and R f values were determined using analytical thin layer chromatography (TLC) with Merck silica gel 60-120 F 254 precoated plates (0.25 mm) thickness.Spots on the TLC plates were visualized using ultraviolet light both at short (254 nm) and long wave (365 nm) UV light.RMSF of residues.In all panels, the color code of the drug complex is PvNMT (black), the complex of PvNMT with 9a (magenta) and the complex of PvNMT with control ligand E (green).

General method for the preparation of compound 3(a-d)
The intermediates 3(a-d) were synthesized using standard organic chemistry protocols as reported in our previous article (Kumar et al., 2016).Diethyl (ethoxymethylene) malonate 1 (25 mmol) and corresponding substituted aniline 2(a-d) (25 mmol) were mixed in equivalent proportions at ambient temperature, giving rise to an exotherm of 18 � C. Benzene (6.5 ml) was added, and the resulting solution was heated under reflux (83 � C) for 2h.The solution was concentrated in vacuo to obtain an oil, which was crystallized on standing.
The crude product was slurred in hexane, filtered off and air dried to give the compound 3(a-d) (m.p. 47-50 � C).

General method for the preparation of compound 4(a-d)
Quinolines 4(a-d) were synthesized using standard organic chemistry protocols as reported in our previous article (Kumar et al., 2016).A solution of 3(a-d) (85 mmol) in POCl 3 (1.34mol) was refluxed at 100 � C for 8h.The cooled solution was concentrated in a vacuum and the resulting brown oil partitioned between CH 2 Cl 2 (500 ml) and water (250 ml).Organic extracts were dried through Na 2 SO 4 and concentrated in vacuo to give a brown oil which was chromatographed on silica gel eluting with 17% EtOAc:Hexane to give 4(a-d) as an off-white to light yellow solid.

General method for the preparation of compound 8
Spiro[chroman-2,4 0 -piperidin]-4-one 8, was synthesized by following the reported protocol with slight modifications (Ghatpande et al., 2022).A mixture of 1-boc-4-piperidone 5 (4.0 mmol) and pyrrolidine (6.0 mmol) with o-hydroxyacetophenone 6 (4.0 mmol) was dissolved in anhydrous MeOH at r.t. and stirred at 80 � C for 6 to 8 h.After the completion, the reaction mixture was cooled, MeOH was removed under reduced pressure followed by extraction with ethyl acetate and dried over anhydrous sodium sulfate.The resulting crude was purified by silica gel column chromatography to afford the corresponding product 7. Further, compound 7 (9.452mmol) was dissolved in anhydrous DCM followed by the addition of TFA (78.407 mmol) and stirred at r.t. for 2 h.After the completion, the reaction mixture was neutralized with 10 N NaOH and extracted with DCM and water.The organic phase was dried over anhydrous sodium sulfate and purified by silica gel column chromatography to afford the corresponding product 8 as a light-yellow solid.

General method for the preparation of the target compounds 9(a-o)
The target compounds 9(a-o) were synthesized using a nucleophilic aromatic substitution (SNAr) synthetic approach by following standard organic chemistry protocols as reported by Singh et al. (2020) with slight modifications.

Protein modeling and molecular docking analysis
In the present work, we aim to explore the possible biological target of the synthesized small molecules to examine the possibilities of an existing relationship between the experimentally observed biological results of compounds using molecular docking results.The three-dimensional structure of PfNMT protein was not available in the Protein Data Bank, therefore protein modeling based on the ligand-bound structure of PvNMT (4BBH) was used as a template to perform homology modeling to obtain the PfNMT three-dimensional structure (Hema et al., 2021).The PvNMT (4BBH) amino acid sequence shares more than 80% sequence identity and remarkably high coverage with that of PfNMT.The BLASTP search with the PDB protein sequences revealed that the amino acid sequences of the other PDB PvNMT structures have 100% identity with that of the PvNMT template (4BBH) (Figure S62, S3).Hence, a single template was used for PfNMT modeling.Protein modeling was carried out using, 1. Homology modeling using Modeller, 2. Robetta Structure Prediction and 3. I-TASSER server for protein 3D structure.The models were compared to the AlphaFold predicted structure of the same protein.The generated models were then validated using ERRAT, Verify3D and Procheck.The best model was further evaluated for drugability using CASTp server and CavityPlus.
The crystallographic structure of PvNMT (PDB ID: 4BBH, resolution: 1.63 Å), retrieved from RCSB Protein Data Bank (http://www.rcsb.org)and homology model of PfNMT, both were used for molecular docking studies.We used the glide module (Friesner et al., 2006) of Schr€ odinger suite ver.2021-2 for molecular docking installed on Microsoft Windows 10based system.The computer is equipped with an 8-core Intel i5 processor, 16 GB of LPDDR4x Ram and an 8 GB Intel Iris Xe graphic card.All ligands were minimized using workspace operations of Maestro 12.8, optimized with the MOPAC2016 module, aligned and prepared using the LigPrep module of the Schr€ odinger suite.The receptor grid was generated using the receptor grid generation in the Glide application by specifying the bound ligand, which was identified by the SiteMap tool (Halgren, 2009).The centroid coordinates of grid box for PvNMT were X ¼ 24.052, Y ¼ 43.048 and Z ¼ 64.906 and dimension was 25 � 25 � 25Å: Similarly, centroid coordinates of grid box for PfNMT model protein were X ¼ 10.890, Y ¼ 17.140 and Z ¼ À 23.170 and dimension was 30 � 30 � 30Å: The ligands were docked in the binding site of proteins using the XP mode of Glide and the docked conformers were evaluated using the Glide (G) Score, and the best-docked pose with the lowest Glide score value was recorded.The most favorable ligand orientation with the lowest free energy (binding affinity) was selected for further structure analysis of the protein-ligand complex, which included calculations of hydrogen bonds, hydrophobic interactions, bond lengths etc., using Maestro of Schr€ odinger suite (Friesner et al., 2006).The docking results were validated by calculating the coefficient of determination R 2 (R-squared) between the experimental IC 50 values and the predicted IC 50 value.The predicted IC 50 value was calculated by putting the docking scores in the general equation of straight line obtained from the graph plotted between docking scores vs experimental IC 50 of the test compounds, marked as training set (Figure S13).

Molecular dynamic simulation
Gromacs 5.1.4(Hess et al., 2008) version installed in CentOS Linux release 7.9.2009(Core) was used to perform the MD simulations of each system.The system hardware was equipped with Intel Skylake Intel(R) Xeon(R) Gold 6130 CPU @ 2.10 GHz), 32 Cores per node, 96 GB of RAM and MGA-G200 graphics chi.The parameters are taken from the CHARMM all-atom force field (Vanommeslaeghe et al., 2009).CHARMM-GUI (Kim et al., 2017) tool is used to generate the parameters of the ligands NHW, 9b and control ligand E. All atoms of PvNMT, including complexes PvNMT-9a and PvNMT-E, are equilibrated in a cubic box, having a size of approximately 112.43 � 112.43 � 112.43 Å.The water model TIP3P (Jorgensen et al., 1983) is used to solvate the system and the details have been provided in Table S4.Each system is energetically minimized with the steepest descent followed by the conjugate gradient algorithm.Long-range electrostatic interactions are calculated with the particle-mesh Ewald (PME) method (Essmann et al., 1995), and for van der Waals (vdW) interactions, a cutoff of 14 Å is applied.The temperature is maintained at 300 K using Berendsen's algorithm (Berendsen et al., 1984) with a coupling constant of 0.5 ps.The Leapfrog algorithm is used for integrating Newton's equations of motion over a two fs time step.Bond lengths of all the bonds are constrained with the LINCS algorithm (Hess, 2008).Water molecule bond lengths are constrained by employing the SETTLE algorithm (Miyamoto & Kollman, 1992).Position restrained simulations by varying on all the heavy atoms of protein are done to equilibrate the solvent around the protein for 500 ps.We have performed MD simulation on each system of PvNMT, PvNMT-9a and PvNMT-E complexes for 100 ns.

ADME prediction
The different pharmacokinetic parameters were predicted by using the QikProp module of the Schr€ odinger suit.All the ligands were minimized and optimized using the MOPAC2016 module before prediction.The QikProp job to predict the ADME parameters was started in the normal mode once the ligands were optimized.

Plasmodium falciparum culture and inhibition assay
A well-described protocol was followed to culture the Pf3D7 and PfINDO lines as described previously (Singh et al., 2020).In this method, human erythrocytes (4% hematocrit) in RPMI media (Invitrogen) were supplemented with 0.5% albumax and 4% hematocrit and were synchronized by repeated sorbitol treatment.The three sets of the growth inhibition assay were performed simultaneously that was repeated twice.The 4 mL of the test compound was added to each well of 96 well flat bottom plate so as to reach the desired concentration (0-100 mM).To this parasite culture (96 mL) at 1% parasitemia (3-10 h p.i) and 2% hematocrit was added and the plate was incubated for the next 48 h.0.4% DMSO and 4 mM chloroquine were used as controls.After 48 h, DNA fluorescent dye-binding assay using SYBR green (Invitrogen) was performed to evaluate the growth of the parasite in each well as described by Smilkstein et al. (2004).

Effect of compounds on the maturation of Pf3D7
Pf3D7 culture at 1% parasitemia and 2% hematocrit (0-6 h, p.i) was treated for different time intervals by IC 90 of test compounds.At each time interval, thin giemsa-stained smears were prepared and morphological changes induced by drug treatment were observed by microscopy.

Hemolysis assay
The toxicity of compounds to host cells was determined by the in vitro hemolytic assay as described earlier (Ahmedi et al., 2022;Amin & Dannenfelser, 2006).RBCs from a healthy individual were collected in tubes containing EDTA, an anticoagulant.The erythrocytes were harvested by centrifugation for 10 min at 2000 rpm at 20 � C and washed three times in PBS.To the pellet, PBS was added to yield a 10% (v/v) erythrocytes/PBS suspension.The 10% suspension was then diluted 1/10 in PBS.From each suspension, 100 ml was taken in triplicate and added to Eppendorf tubes.Complete hemolysis was achieved with 1% Triton X-100.Tubes were incubated for 1 h at 37 � C and then centrifuged for 10 min at 2000 rpm and 20 � C. Supernatant absorbance was measured spectrophotometrically at 450 nm.The hemolysis percentage was calculated by the following equation: % Hemolysis ¼ A 450 of test compound treated sample À A 450 of buffer treated sample A 450 of 1% TritonX À 100 treated sample À A 450 of buffer treated sample � 100

Figure 2 .
Figure 2. Molecular interactions and orientations of the best two docked compounds, 9a and 9b with PvNMT.Ligands are shown as sticks.

Figure 3 .
Figure 3. Molecular interactions and orientations of the best two docked compounds, 9a and 9b with a homology model of PfNMT.Ligands are shown as sticks.

Figure 4 .
Figure 4. Molecular interactions and orientations of the best bioactive compounds 9n, and 9o with PvNMT.Ligands are shown as sticks.

Figure 5 .
Figure 5. Molecular interactions and orientations of the best bioactive compounds 9n, and 9o with homology model of PfNMT.Ligands are shown as sticks.

Figure 7 .
Figure 7. Time course effect of compounds at IC 90 (mM) on Pf3D7 maturation: Parasites were treated with varying time intervals.Morphological changes induced were studied by staining smears of each time point with giemsa.1500 cells were observed to select the best representative image.

Figure 8 .
Figure8.Hemolysis caused by CQ and best five compounds.Hemolysis was determined by recording absorption at 450 nm and comparing it to hemolysis achieved with 1% Triton X-100 (reference for 100% hemolysis).This data is a mean of triplicate experiments.The student's t-test was used to verify statistical significance (p < 0.05).

Figure 9 .
Figure 9. Molecular dynamics simulations of drug-protein complexes during 100 ns.(A) RMSD of the C a backbone, (B) radius of gyration, R g , (C) SASA and (D)RMSF of residues.In all panels, the color code of the drug complex is PvNMT (black), the complex of PvNMT with 9a (magenta) and the complex of PvNMT with control ligand E (green).

Table 1 .
Calculated binding free energies (kcal/mol) of the synthesized compounds using molecular docking through Schr€ odinger.

Table 2 .
Details of different substituents of the final compounds and their physicochemical properties.

Table 3 .
Crystallographic Parameters for the Compounds

Table 4 .
Anti-plasmodial activity of the synthesized compounds.

Table 5 .
Predicted ADME properties and drug-likeness of the synthesized molecules.