3D-QSAR-based design, synthesis and biological evaluation of 2,4-disubstituted quinoline derivatives as antimalarial agents

ABSTRACT 2,4-Disubstituted quinoline derivatives were designed based on a 3D-QSAR study, synthesized and evaluated for antimalarial activity. A large dataset of 178 quinoline derivatives was used to perform a 3D-QSAR study using CoMFA and CoMSIA models. PLS analysis provided statistically validated results for CoMFA (r2ncv = 0.969, q2 = 0.677, r2cv = 0.682) and CoMSIA (r2ncv = 0.962, q2 = 0.741, r2cv = 0.683) models. Two series of a total of 40 2,4-disubstituted quinoline derivatives were designed with amide (quinoline-4-carboxamide) and secondary amine (4-aminoquinoline) linkers at the -C4 position of the quinoline ring. For the purpose of selecting better compounds for synthesis with good pEC50 values, activity prediction was carried out using CoMFA and CoMSIA models. Finally, a total of 10 2,4-disubstituted quinoline derivatives were synthesized, and screened for their antimalarial activity based on the reduction of parasitaemia. Compound #5 with amide linker and compound #19 with secondary amine linkers at the -C4 position of the quinoline ring showed maximum reductions of 64% and 57%, respectively, in the level of parasitaemia. In vivo screening assay confirmed and validated the findings of the 3D-QSAR study for the design of quinoline derivatives.


Introduction
Malaria is a vector-borne, life-threatening infectious disease caused by Plasmodium parasites [1].In humans, Plasmodium parasites (P.falciparum, P. vivax, P. ovale and P. malariae) are transmitted by the bite of female Anopheles mosquitoes [2][3][4].Among these four species of Plasmodium parasites, P. falciparum and P. vivax are the most chronically virulent species as contributors to malaria [5] because P. falciparum binds to endothelium all over the blood stage, whereas P. vivax transmits hypnozoites in the human liver stage [6].Other Plasmodium parasite species, such as P. ovale and P. malariae, are not lifethreatening species [7].In a report in 2022, the World Health Organization (WHO) recognized malaria as the deadliest infectious disease and reported that 247 million individuals were infected by malaria in 2021.Globally, 619,000 mortalities from malaria were recorded in 2021.Children under the age of five were particularly afflicted by malaria in the WHO's African region, where this contributed up to 80% of all mortalities [8].
The parasite life cycle encompasses two stages, liver stage and the erythrocytic stage.Female Anopheles mosquitoes ingest their sporozoites into human blood during blood feeding, and these sporozoites become primary schizonts in liver cells.Primary schizonts multiplied into merozoites, followed by the release of merozoites after the rupturing of liver cells.After that, merozoites transform into gametocytes that are sucked by mosquitoes during blood feeding to continue the transmission cycle in the mosquito.These gametocytes are fertilized to produce zygotes within the mosquito.After that, the zygote transforms into an oocyst, followed by sporozoites that migrate to the mosquito's salivary gland, where they continue the transmission cycle in humans [9,10].
Antimalarial drugs available in the market are becoming resistant to Plasmodium parasites [11,12], which is reducing the number of drugs available for the effective treatment of malaria.Quinoline heterocycle is present in various antimalarial drugs, such as chloroquine, piperaquine, amodiaquine, primaquine, mefloquine, ferroquine, etc. Quinoline heterocyclecontaining drugs show antimalarial activity by often targeting the blood stage of the malaria parasite.Quinoline heterocycle is the essential moiety out of many heterocycles in antimalarial drugs for treatment of malaria [13,14].In general, quinoline ring-containing compounds have an impact on Plasmodium during the intra(endo)erythrocytic phase of the parasite's life cycle, when the parasites exhibit a marked increase in metabolic activity and utilize components of the host cell for their biosynthetic requirements.Because quinoline compounds can prevent the synthesis of hemozoin, they may be interfering with the process of haem detoxification, leaving parasites more vulnerable to oxidative stress caused by haem [14,15].
Due to the demand for novel antimalarial drugs with a distinctive mode of action, scientists from all over the world are constantly discovering new quinoline-based antimalarial compounds [13,16,17].
Three-dimensional quantitative structure-activity relationship (3D-QSAR) is a very effective drug design method that plays a vital role at the early stage of drug design.In addition, this method is also helpful for the selection of designed molecules for synthesis based on their predicted activity [18,19].3D-QSAR uses various field contributions, such as electronic, steric, hydrophobic, and hydrogen bonding, and correlates this with the activity of the molecules.In order to select better designed compounds for wet-lab experiments, contour maps from 3D-QSAR analysis are beneficial for both the selection of substitutions for favourable interactions with the target and the activity prediction of designed compounds prior to their synthesis [20].Herein, we have performed a 3D-QSAR study on a large dataset of 178 molecules for the design of 2,4-disubstituted quinoline derivatives.The antimalarial activities of the designed compounds were predicted using the best validated 3D-QSAR models.Finally, we selected the best predicted designed compounds for the synthesis and in vivo antimalarial evaluation.

Dataset
3D-QSAR models were generated using a dataset of 178 quinoline analogues (Table S1 under supplementary materials), consisting of 7-(2-phenoxyethoxy)-4(1H)-quinolones (28 compounds), 3-substituted-2-methyl-4(1H)-quinolones (51 compounds), and 1,2,3,4-tetrahydroacridine-9(10 H)-ones (99 compounds), from the same research group and laboratory [21][22][23] using SYBYL X 1.2 software.In the reported literature, Cross et al. evaluated all 178 compounds against the multidrug-resistant malarial strain W2 (chloroquine and pyrimethamine resistant) and calculated antimalarial activities as EC 50 values.Reported EC 50 values (nM) were converted into pEC 50 (-log(1/EC 50 )), and used as a dependent variable for the generation of 3D-QSAR models.The whole dataset of 178 compounds was divided into a training set of 125 molecules and a test set of 53 molecules.The selection of molecules for the training and test sets was made with the knowledge that the spectrum of activity represented by the test set molecules is similar to that of the training set.As a result, the test set served as the training set's actual representation.This resulted from the arbitrary selection of 53 molecules as a test set with evenly spaced biological data.

3D QSAR methodology
The Sybyl X 1.2 software's SKETCH tool was used to draw chemical sketches of all the compounds in the training and test sets.Using Gasteiger-Marsili charges, the Tripos force field, and the Powell method, energy reduction was carried out with a gradient difference of no less than 0.01 kcal/mol.The maximal common substructure (MCS) as described by the Distill alignment module was used.Training and test set compounds were aligned on the MCS using compound Tr1 (Table S1 under supplementary materials) as a template molecule.

Comparative molecular field analysis (CoMFA) and comparative molecular similarity indices analysis (CoMSIA) modelling
The CoMFA steric and electrostatic fields were estimated for each compound in the training and test sets using the sp 3 -hybridized carbon as the probe atom, using the Coulomb and Lennard-Jones potentials, respectively.Column filtering was set to 2.0 kcal/mol with a distance-dependent dielectric constant, and the default energy cut-off values were kept at 30.00 kcal/mol for steric and electrostatic fields.For the CoMSIA study, we employed the same default parameters as the CoMFA computation of similarity index descriptors.Compared with the Lennard-Jones and Coulomb potentials, the Gaussian function has a distinct shape.As a result, similarity indices were calculated for the molecules both inside and outside of their molecular surfaces [24].Statistical parameters of 3D-QSAR models were calculated and validated using partial least squares (PLS) regression analysis [25].

Activity prediction studies
Validated CoMFA and CoMSIA models were used for the prediction of the activity of the designed compounds.The sketch function of the Sybyl X software was used to draw the 3D structures of the designed compounds.Utilizing the Tripos force field and Gasteiger-Huckel charges, energy minimization of drawn structures was performed.

Chemistry
Except as otherwise specified, all chemicals were used as received.We purchased aniline, morpholine, dimethylformamide, palladium acetate (Pd(OAc) 2 ), xantphos, sodium tertbutoxide (Naot-Bu) and hexafluorophosphate azabenzotriazole tetramethyl uranium (HATU) from Sigma Aldrich, and o-toluidine, p-toluidine, o-anisidine, p-anisidine, and N, N-diisopropylethylamine from CDH Fine Chemical.Merck Keiselgel aluminium-backed plates coated with silica were used for analytical thin layer chromatography (TLC), and (120 micron) silica was used for column chromatography.Components were seen using potassium permanganate and ultraviolet light following TLC.The melting temperatures of the final synthesized target compounds and intermediates were calculated using an electrothermal melting point (MP) apparatus.By using the KBr dispersion procedure, Fourier-transfrom infrared spectroscopy (FTIR) spectra were captured on the Jasco FTIR instrument.A Bruker Avance II 400 spectrometer was used to record 1 HNMR spectra at 400 MHz and 13 C NMR spectra at 100 MHz in the designated solvent while utilizing the internal standard, residual protic solvent dimethyl sulphoxide-d6 (DMSO-d6).The chemical shifts were measured in ppm.Measurements of all coupling constants were made in hertz (Hz) and multiplicities are reported as follows: s (singlet), d (doublet), t (triplet), q (quartet), dd (doublet of doublets) and m (multiplet).On an Agilent mass spectrometer, mass spectra (MS) were captured in electron spray ionization (ESI) mode.On a Perkin Elmer elemental analyser, the elemental analysis was carried out for CHN elements in the synthesized compounds.

Synthesis of 2-hydroxy-(N-substituted quinoline)-4-carboxamide (3a-d)
In a clean, dry round-bottom flask (RBF), 2-hydroxyquinoline-4-carboxylic acid (1) (0.010 mol) was dissolved in DMF at 0°C under nitrogen gas.To this, HATU (0.0150 mol) and DIPEA (0.030 mol) were added and stirred for 1 h.Substituted aromatic amines (2a-d) (1 eq.mol) were added to the reaction mixtures and allowed to stir for 1-2 h at 0-5°C, then taken to room temperature and allowed to stir for 12-30 h, and reactions were monitored using TLC.Upon completion of the reactions, the reaction mixtures were poured into the ice-water mixture.Precipitates were poured out and filtered to get crude solid products (3a-d).The crude products (3a-d) were purified using column chromatography over a silica column, with a mixture of hexane: ethyl acetate (90:10) as the mobile phase [26].

Synthesis of 2-chloro-(N-substituted quinoline)-4-carboxamide (4a-d)
In a clean and dry RBF, 2-hydroxy-N-substituted quinoline-4-carboxamides (3a-d) (0.01 mol) were refluxed with phosphorus oxychloride (POCl 3 ) at 90-100°C for 8-10 h.TLC was performed to observe the reaction's progression.After the completion of reactions, POCl 3 was removed and reaction mixtures were neutralized to pH 7 using a solution of sodium bicarbonate at 0-5°C, then reaction mixtures were dissolved in ethyl acetate.The resulting mixtures were washed with water, and the organic layers were separated and dried over anhydrous Na 2 SO 4 .These solutions were filtered out to get crude solid products, and the solvent was removed under reduced pressure.The crude products (4a-d) were purified using column chromatography over a silica column, with a mixture of hexane: ethyl acetate (90:10) as the mobile phase [27].

Synthesis of 4-substituted thiazol-2-amines (15 and 17)
In a clean and dry RBF, substituted 2-bromoethanone (13 and 16) (0.038 mol) was dissolved in DMF and a mixture of K 2 CO 3 (0.046 mol) and EtOH under nitrogen gas.The reaction mixture was refluxed for 2 h with thiourea (14).TLC was performed to observe the reaction's progression, and after the completion of the reaction, mixtures were poured into an ice-water mixture.Precipitates were poured out and filtered through a Buchner funnel to get crude solid products (15,17).The crude products (15,17) were purified using column chromatography over a silica column, with a mixture of dichloromethane: methanol (90:10) as the mobile phase [29].

Calculation of molecular properties of dataset and synthesized compounds
For the comparison of molecular properties of the dataset (training and test) and synthesized compounds, the online workstation SwissADME, available at http://www.swissadme.ch/index.php#,was used.Physicochemical properties including molecular weight (MW), lipophilicity (log Po/w), water solubility (log S), number of H-bond donors (HBDs), number of H-bond acceptors (HBAs), topological polar surface area (TPSA), and molar refractivity (MR) were calculated.

In vivo antimalarial activity
The Institutional Animal Ethics Committee of Nirma University, Ahmedabad, India, approved the protocol for animal experiments (IP/PCEM/MPH/24/2019/005). Swiss Albino mice were kept in the central animal facility of Nirma University in Ahmedabad with a standard diet and water ad libitum.One week of acclimatization time was given to the animals before the commencement of the studies.Using a conventional technique, a blood-stage infection was developed in 6-8-week-old Swiss Albino mice using a reference protocol [31].On day 0 of the study, 1 × 10 7 parasitized erythrocytes were injected intraperitoneally into each mouse.From day 3 after the first infection, all the synthesized compounds were given twice daily at a dose concentration of 6.5 mg/kg body weight/12 h for the next 3.5 days.Mice in the control group received only vehicle.Based on the standard methodology, dose concentration and dose frequency were determined [32].Giemsa staining of a thin blood smear, which was prepared from blood drawn by tail-nick, was used to determine the level of parasitaemia.A minimum of 5-10 microscopic fields were used to count the infected red blood cells and calculate the % parasitaemia.Mice with parasitaemia levels that were undetectable or considerably lower than those in the control group were considered cured.After the study was completed, the animals were euthanized.

3D-QSAR study
The CoMFA and CoMSIA models were developed using a dataset of 178 quinoline analogues for the design of 2,4-disubstituted quinoline derivatives.Statistical parameters such as r 2 cv , q 2 , r 2 ncv , SEE, r 2 pred , and F were calculated for validation of 3D-QSAR models and found statistically significant (Table 1) based on the Distil alignment (Figure S1 under the supplementary materials).For a total number of 24 components, CoMFA and CoMSIA models were internally validated with q 2 values 0.677 and 0.741, non-cross validated (r 2 ncv ) values of 0.969 and 0.962, F-value 194 and 114 and standard error of estimation (SEE) value 0171 and 0.198, respectively.Based on the test set compounds, CoMFA and CoMSIA models were externally validated with predicted r 2 (r 2 pred ) values of 0.698 and 0.686, respectively.The experimental and predicted pEC 50 of the training and test set compounds are tabulated in Table S2 and plotted in Figure S2 under the supplementary materials.

Contour map analysis
Contour maps were generated as a final outcome of 3D-QSAR analysis.According to the chemical structures of a compound, the contour maps suggested the favourable and unfavourable locations of various fields, such as electrostatic, steric, hydrophobic, HBA and HBD in the three-dimensional (3D) space.Therefore, favourable substitutions that could increase antimalarial activity were taken into account for the design of the novel molecules.Figure 1 depicts the steric and electrostatic contour maps derived from the CoMFA analysis.The CoMFA steric map (Figure 1a) shows the sterically favourable (contributed 80%) and unfavourable (20%) areas of substitutions in the template compound (Tr1) (Table S1 under the supplementary materials) as green and yellow contours, respectively.Sterically favourable green contours were observed near the -C4 and extended -C2 position on the quinoline ring, demonstrating that the activity of the designed compounds may be increased by the substitutions of bulky groups in these areas.Two sterically unfavourable yellow contours were observed: one near the substitution (-R 1 ) at -C3 position (-CO 2 Me) of quinoline ring and the other was found above the -C5 position, implying that the bulky substitutions on -C3 and -C5 positions of the quinoline ring may lead to a decrease in the antimalarial activity of the designed compounds.CoMFA electrostatic map is shown in Figure 1b, where blue and red-coloured contours indicated electrostatic favoured (contributed 80%) and disfavoured (contributed 20%) regions.A large blue contour covered the -C5, -C6 and -C7 positions of the quinoline ring, which recommends that positively charged groups at these places might increase the inhibitory activity of designed compounds.A red-coloured contour covered the -N atom and -C8 position of the quinoline ring and suggested that presence of any electronegative groups in these regions may be beneficial for the better activity of designed compounds.
Hydrophobic, HBD and HBA fields are additionally calculated by CoMSIA analysis.The CoMFA and CoMSIA steric and electrostatic contour maps were positioned in a similar manner.In the CoMSIA hydrophobic contour map (Figure 2a) the yellow contour (contributed 80%) indicates a hydrophobic favourable region and grey colour (contributed 20%) indicates hydrophobic disfavourable regions.A yellow contour covered -C3, -C4 positions of quinoline ring, which indicated that hydrophobic groups will be conducive for the activity.The presence of grey contour at the -C3 position of the quinoline ring suggested that a hydrophobic substitution at this position would decrease the activity of designed compounds.In the HBA contour map (Figure 2b), the magenta contour (contributed 80%) shows a favoured region and red contour (contributed 20%) shows a disfavoured region.The magenta contour was observed out of the molecular area.Red contour covered substitution on the -N1 position of quinoline ring, which suggested an acceptor group at this position will not provide any contribution to biological activity.In the HBD contour map (Figure 2c) cyan and purple contours indicate H-bond donor  favoured (contributed 80%) and disfavoured (contributed 20%) regions, respectively.The presence of cyan contour above the -C4 position of the quinoline ring suggested the substitution of a donor group in designed compounds.The HBD disfavoured purple contour was observed near -C7 and -C8 positions.

Design of 2,4-disubstituted quinoline derivatives
We designed a total of 40 (1d-40d) (Table S3 under supplementary data) 2,4-disubstituted quinoline derivatives using the contour map information from the 3D-QSAR study.We were interested in quinoline derivatives, and in consequence of that, a 3D-QSAR study was performed on a large dataset of 178 substituted quinoline derivatives.We have attempted to confirm the conclusion drawn from the contour map analysis described in the previous section for the design of 2,4-disubstituted quinoline derivatives.The main strategy of molecular modification for the design of quinoline derivatives is the substitution of functional groups, which were found to be important with the contour map analysis.The quinoline ring was maintained in the designed molecules as a core moiety because it is known for its antimalarial activity.We designed two series of 2,4-disubstituted quinoline derivatives with amide (quinoline-4-carboxamide) and secondary amine (4-aminoquinoline) linkers at the -C4 position of the quinoline ring.These linkers served the purpose of hydrogen-bonding interactions, as per the HBD contour map analysis.In the first series of 4-carboxamidequinoline, sterically bulky and hydrophobic substituents such as tolyl, trifluoromethylphenyl, methoxyphenyl, ethoxyphenyl, and trifluoromethoxyphenyl groups were used for the further extension of the -C4 amide group as a requirement of the steric contour map.In the second series of 4-aminoquinoline compounds, we extended the -C4-amino group by substituting 4-nitrophenyl thiazole, 4-chlorophenyl thiazole, 2,4-dichlorophenyl thiazole, 2,4-difluorophenyl thiazole, 2-trifluoromethylphenyl thiazole and the biphenyl thiazole moiety.In the first series of compounds, at the -C2 position of the quinoline moiety, morpholine and aniline rings were substituted as a requirement of steric bulkiness.In the second series of compounds, the unsubstituted phenyl, pyridine and pyrimidine rings were maintained to serve the purpose of steric bulk.We kept other positions such as -C3, -C5, -C6, and -C7 unsubstituted as a requirement of electrostatic and hydrophobic contour maps.This strategy enabled the formation of newly designed 2,4-disubstituted quinoline derivatives with all structural requirements as per the contour map analysis (Figure 3).

Prediction of activity of designed compounds
The antimalarial activity of the designed compound was predicted using both 3D-QSAR models (CoMFA and CoMSIA), and the best predicted compounds were selected for synthesis and biological evaluation.The predicted pEC 50 values of the designed compounds are given in Table S3 under supplementary data.The results of the predicted activity of the designed compounds validated the selection of bulky substituents at the -C4 and -C2 positions of the quinoline ring.In terms of preference for the linkers, amide linkers containing compounds were slightly better predicted as compared with secondary amine linkers.In the 1 st series, 2-methylphenyl group (1d), 2-methoxyphenyl (3d) groups at the -R 1 position and aniline at -R 2 of the designed compounds exhibited pEC 50 values of more than 9, and similarly designed compounds with 4-methoxyphenyl (7d) and 2-methoxyphenyl (4d) groups at the -R 1 position and morpholino ring at the -R 2 position demonstrated pEC 50 values of more than 8.8, which is more than most of the training set compounds.In the 2 nd series of compounds biphenyl ring was predicted with pEC 50 values of more than 9.5.The designed compound with 4-nitrophenyl ring at the -R 1 position and phenyl ring at the -R 2 position (24d) was predicted with pEC 50 values of more than 8.7.Finally, the designed compounds with the best predicted activity were chosen for synthesis and biological evaluation.

Chemistry
The synthesis of 2,4-disubstituted quinoline derivatives was carried out by acid-amine coupling, hydroxyl group substitution, elimination, halo-amine coupling, and Hantzsch condensation reactions.Two different types of targeted compounds were synthesized using amide and secondary amine linkers, as quinoline-4-carboxamide and 4-aminoquinolines, respectively.As depicted in Scheme 1, the first series of targeted compounds (5)(6)(7)(8)(9)(10)(11)(12) was synthesized in three steps.In the first step, 2-hydroxy-(N-substituted quinoline)-4-carboxamide (3a-d) were synthesized via acid-amine coupling reaction between 2-hydroxyquinoline-4-carboxylic acid (1) and substituted aromatic amines (2a-d) using N,N-diisopropylethylamine (DIPEA) and HATU.In the next step hydroxyl group substitution reaction was performed on 3a-d to convert them in corresponding chloro compounds (4a-d) using phosphorus oxychloride (POCl 3 ).In the third step two different types of substitutions were placed on the -C2 position of the quinoline ring by halo-amine coupling reaction.Final compounds 5-8 and 9-12 were synthesized by reaction of 4a-d with aniline and morpholine, respectively, using palladium acetate (Pd(OAc) 2 ) and xantphos along with sodium tert-butoxide.Synthesized compounds' chemical structures were verified by FTIR, NMR, mass spectrometry, and elemental analytical data.Synthesized compounds demonstrated characteristic -NH stretching peaks of amide and secondary amine linkers in the range of 3500-3205 cm −1 and C=O stretching peaks near 1352-1250 cm −1 , whereas compounds showed aromatic -C-H stretching peaks near a broad range at 3156-3031 cm −1 , and aliphatic -C-H stretching peaks in the region of 3035-2825 cm −1 in FTIR spectra.The presence of various sets of hydrogen atom resonances corresponding to the protons in the quinoline, thiazole, phenyl and biphenyl ring system, and methyl and methoxy along with amide and secondary amine linkers is the key characteristic of 1 H NMR spectra.The group of protons associated with the quinoline, thiazole, phenyl, morpholine and biphenyl ring system showed up in the δ 8.3-3.619ppm range and demonstrated the expected multiplicity and integration values and correlated with -C atoms (δ 166.24-48.65) in the 13 C NMR spectra.Singlets were observed for the amide proton peaks (-NH) and were found at the typical range of δ 10.294-9.591ppm for this group, which correlated with the -C atom of the amide group (C=O) at δ 176-166 ppm in the 13 CNMR spectrum.The methyl protons in the synthesized compounds showed a peak in the range of δ 2.4-2.2ppm and correlated with -C atoms (δ 21.23-17.23 ppm) in the 13 C NMR spectrum, methyl proton of the methoxy group was observed at δ 4.112-2.911ppm (δ 57.10-56.01ppm in 13 C NMR).Persistent base M + 1 peaks (456.15-348.16)were linked to mass spectra of synthesized compounds.

Molecular properties of dataset and synthesized compounds
The molecular properties of a compound are crucial in determining its biological activity.We calculated and compared the molecular properties of all the synthesized compounds and training and test set compounds, such as log S, log P o/w , MW, HBD, TPSA, HBA, MR, and the number of rotatable bonds.The water solubility parameter (log S) significantly impacts how drugs are absorbed and distributed.Log S values were calculated for both the dataset and synthesized compounds, in which dataset compounds exhibited values in the range of −2.77 to −9.44 and synthesized compounds showed values in the range of −4.03 to −5.78.Log S greater than −4 is considered a good value for better drug absorption.All the synthesized compounds showed a log S value of more than −4.The Scheme 2. Synthetic scheme for the synthesis of 2-phenylquinoline-4-amine derivatives (19,20).Reagents and condition: (a) DMF, K 2 CO 3 and EtOH, reflux 2 h, (b) isopropyl alcohol, reflux at 90°C for 8-9 h.lipophilic parameter (log P o/w ) demonstrated a value in the range of 6.61-1.64 for the dataset compounds; however, the log P o/w value of synthesized compounds showed a range of 3.66-2.69,which is an optimal range of log P o/w for any molecule to become a drug candidate.Binding at the target site and transport across the membrane are highly dependent on the log P o/w value of a compound.Similarly, the values of other molecular properties such as TPSA, which governs transport across the cell membrane, were found in the range of 85.46-32.86 and 79.9-41.99 for the dataset and synthesized compounds, respectively, and demonstrated that there is no significant difference in the values of both the dataset and synthesized compounds.Steric parameters MR and MW (g/mol) provide information on the size of the molecules and were found in the ranges of 181.17-57.19 and 199.25-647.72,respectively, for dataset compounds.For synthesized compounds, MR and MW values were found in the range of 113.32-298.5 and 382.41-328.37,respectively.For some of the dataset compounds, MW values are greater than 600 as compared with all the synthesized compounds, which obeyed Lipinski's rule of less than 500 MW.HBA (electronic parameter) values in the range of 8-1 and 4-2 for dataset and synthesized compounds, respectively, and HBD (electronic parameter) values in the range of 2-1 for both dataset and synthesized compounds followed Lipinski's rules for HBA and HBD groups.Overall, the calculated molecular properties of the synthesized compounds were found to be in a more acceptable range as compared with the dataset compounds, as per Lipinski's rule of five (Table 2).

In vivo antimalarial activity
All the synthesized compounds were evaluated for in vivo antimalarial activity.A reduction in the level of parasitaemia in infected mice was measured, and the relative inhibitory activity of these compounds was determined.The extent of the reduction in infected mice was compared with the control group.Compounds 5 (Figure S14 under supplementary materials), 8 and 19 significantly decreased the level of parasitaemia, and compounds 6, 9 and 12 were found to be less effective, as shown in Figure 4(a).Other synthesized compounds (7, 10, 11 and 20) did not show a reduction in the level of parasitaemia.Overall, the extent of reduction of the level of parasitaemia was less than chloroquine (standard), which demonstrated maximum suppression as depicted in Figure 4(a).Compound 5 from the series 1 (amide linker) of synthesized compounds proved to be the most active compound in this in vivo antimalarial activity, and exhibited a 64% reduction in the level of parasitaemia.Compound 5 is substituted with a 2-methyl group at the -R position which is attached with amide linker at the -C4 position of quinoline ring and aniline group at the -C2 position.Another compound 8 with a 4-OCH 3 group substituted at the -R position showed 53% reduction in parasitaemia.In the case of the compound 6 where the 2-CH 3 group was replaced with a 2-OCH 3 group at the -R position, less reduction in the level of parasitaemia was observed (23%).In the case of other synthesized compounds (9)(10)(11)(12) with morpholine ring at the -C2 position, less inhibitory activity was observed from compounds 9 (23%) and 12 (22%).Compound 19 showed 57% reduction in the parasitaemia and was found to be the most active compound from the secondary amine linker series of designed compounds with biphenylthiazol-2-amine group at the -C4 position of the quinoline ring.

Conclusion
In the field of antimalarial drug research, the design and discovery of novel antimalarial agents are required to address the problem of drug resistance and also for the continuing supply of novel compounds in the pipeline.In an attempt to identify novel antimalarial agents, we performed a 3D-QSAR study on a large dataset of 178 quinoline analogues.PLS analysis resulted in satisfactory statistical parameters for both CoMFA (r 2 ncv = 0.969, q 2 = 0.677) and CoMSIA (r 2 ncv = 0.962, q 2 = 0.741) models.Both the models demonstrated good predictive ability based on external validation

Figure 3 .
Figure 3. Strategy for the design of substituted quinolines using CoMFA and CoMSIA contour maps generated using 3D-QSAR study.

Figure 4 .
Figure 4. P. berghei-infected mice in vivo efficacy data of synthesized compounds.Mice were infected with P. berghei on day 0. Mice were dosed twice daily for 3.5 days starting 3 days post infection.All the compounds were administered through the intraperitoneal route with dose concentration of 6.5 mg/ kg/12 h, and as a control chloroquine was administered at the same dose using the intraperitoneal route.(a) Reduction in %parasitaemia vs. day post inoculation using synthesized compounds.Results represent the mean for four animals per dose ± S.D. (b) Kaplan-Meyer survival analysis of synthesized compounds administered in mice after blood stage infection.

Table 1 .
Statistical parameters using PLS analysis for CoMFA and CoMSIA models.