In silico evaluation of quinoline-4-carboxamides as potential antimalarial agents

2.1. Computational methods

3.1. Analysis of frontier molecular orbitals

The FMO analysis of quinoline-4-carboxamide derivatives (1-43) was carried out to assess their electronic properties, including chemical stability and reactivity. The energy gap (Egap) between the HOMO and the LUMO is a crucial indicator of these properties. A smaller Egap is associated with higher reactivity and lower stability (“soft” molecules), whereas a larger gap suggests lower reactivity and increased stability (“hard” molecules).

All quantum chemical descriptors were computed at the B3LYP/6-31G(d,p) level of theory. The evaluated parameters include IP, EA, chemical potential (μ), global hardness (η), global softness (σ), electronegativity (χ), electrophilicity index (ω), binding energy (H), energy change (ΔE), and dipole moment. These have been summarized in Table S1, Figure 3, Supplementary Tables S2-S47, and Figures S1-S45.

Close

Figure 3.

(a) FMOs of ligand 24, showing isodensity surfaces at 0.02 electrons Bohr⁻³ (red indicates electron-rich regions; blue indicates electron-deficient regions) for the HOMO and LUMO; (b) MEP map of ligand 24 at the same isodensity surface (0.02 electrons Bohr⁻³), highlighting electron-rich (red), electron-deficient (blue), and neutral (green/yellow) regions; (c) Density of states (DOS) plot and corresponding HOMO-LUMO energy gap for ligand 24.

Export to PPT

Among the tested ligands, compound 24 displayed a HOMO-LUMO energy gap of 2.99 eV, with corresponding hardness and softness values of 1.49 and 0.665 eV, respectively. Ligands 37 and 38 exhibited slightly larger energy gaps (3.763 eV and 3.961 eV), indicating greater stability. The reference drug quinine (D3) showed a comparable gap of 2.973 eV, implying potentially higher biological activity. The Egap values ranged from 2.332 eV (compound 7) to 8.873 eV (compound 41), reflecting a broad spectrum of electronic behaviors. These differences underscore the variability in chemical reactivity and potential pharmacological performance across the series.

MEP maps for ligand 24 have been depicted in Figure 3, with additional visualizations provided in the Supplementary Information (Figures S1-S45). In these maps, electrophilic regions are indicated in blue, nucleophilic regions appear in red, and areas of partial nucleophilicity are shown in yellow. Electronegative atoms, such as oxygen, exhibit negative electrostatic potential and are represented in shades of red, orange, or yellow, suggesting regions susceptible to electrophilic attack. Conversely, hydrogen atoms display positive potential, illustrated in blue, denoting favorable sites for nucleophilic interaction. Neutral areas are rendered in green, indicating regions of near-zero electrostatic potential.

3.2. In silico molecular docking

Among the selected targets, P. falciparum aminopeptidase N (PfA-M1, PDB ID: 3EBH) emerged as the most promising interaction partner for our compound series. This zinc metalloprotease is a validated antimalarial target of high biological significance. Located in the parasite’s food vacuole, PfA-M1 is essential for the final steps of hemoglobin digestion, a process that supplies the parasite with vital amino acids for its growth and proliferation. Inhibition of this enzyme disrupts this critical nutrient supply, leading to parasite death, and represents a therapeutic strategy that is distinct from conventional antimalarials, offering potential to circumvent existing resistance mechanisms. Molecular docking serves as a pivotal computational approach to evaluate the binding affinity and interaction profiles of ligands with target proteins, thereby elucidating their potential biological activity. In this study, a comprehensive docking analysis was performed for quinoline-4-carboxamide derivatives (1-43) and three reference drugs, mefloquine, hydroxychloroquine, and quinine, against six protein targets: glutamyl-tRNA amidotransferase (PDB ID: 5C5Z), 1D-myo-inositol 2-acetamido-2-deoxy-α-D-glucopyranoside deacetylase (PDB ID: 1Q7T), enoyl-[acyl-carrier-protein] reductase [NADH] (PDB ID: 5VRL), interleukin-6 receptor subunit beta (PDB ID: 8D74), aminopeptidase N (PDB ID: 3EBH), and epidermal growth factor receptor (PDB ID: 4UV7). The resulting binding affinities and interaction data have been summarized in Table 1.

While absolute binding energies provide a convenient computational metric, their biological significance emerges only when translated into physiologically meaningful scales. In the present study, a 1 kcal mol⁻¹ improvement in predicted binding free energy corresponds to an approximately five-fold increase in binding affinity at 310 K, as derived from the relationship ΔΔG = –RT ln(K₁/K₂). Consequently, the observed 2.0 kcal mol⁻¹ advantage of ligand 24 over chloroquine (–10.3 vs. –8.3 kcal mol⁻¹) equates to an ∼11-fold higher predicted affinity for the P. falciparum aminopeptidase N active site. This magnitude is consistent with shifts observed between sensitive and resistant parasite strains when key point mutations reduce chloroquine affinity by 1.5-2.5 kcal mol⁻¹ [19].

To ensure the reliability of docking results, the structural quality of aminopeptidase N (3EBH) was rigorously validated. Ramachandran plot analysis showed that 93% of residues resided in the most favored regions (Figure 4a). Additionally, ProSA-web yielded a z-score of −7.53 (Figure 4b), indicating excellent structural integrity when compared with high-resolution crystal and nuclear magnetic resonance (NMR) structures. These assessments were complemented by quality factor and Verify3D evaluations (Figure S46 in the SI), confirming the suitability of 3EBH for molecular docking.

Close

Figure 4.

(a) Ramachandran plot of protein 3EBH generated using PROCHECK; (b) ProSA-web z-score validation of 3EBH structure.

Export to PPT

The docking results were further interpreted to clarify the biological significance of the observed score differences. Overall, most quinoline derivatives demonstrated superior binding affinity for aminopeptidase N (3EBH) compared to the other evaluated targets. Among these, ligand 24 exhibited the strongest interaction with 3EBH (−10.3 kcal/mol), surpassing both other ligands and the reference drug quinine (−8.3 kcal/mol) (Table 1), with a meaningful difference of approximately 2.0 kcal/mol. In addition, ligand 24 displayed favorable binding toward other targets, including 5C5Z (−6.1 kcal/mol), 1Q7T (−8.2 kcal/mol), 5VRL (−7.8 kcal/mol), 8D74 (−9.8 kcal/mol), and 4UV7 (−10.2 kcal/mol). Ligands 37 (−10.0 kcal/mol) and 38 (−9.5 kcal/mol) also demonstrated strong binding to 3EBH, further supporting their potential as promising computational leads. It is important to note, however, that docking scores should be interpreted with caution. Minor differences (e.g., −8.5 vs. −8.3 kcal/mol) may fall within the error range of docking algorithms and are therefore unlikely to reflect meaningful biological variation. In contrast, larger variations (≥1.0-1.5 kcal/mol) are generally considered biologically relevant, as they may correspond to an order-of-magnitude difference in binding affinity. For this reason, emphasis was placed on the stronger binding observed for ligand 24 relative to quinine, while smaller differences close to reference values were not overinterpreted.

Further interaction analysis (Table 2, Figures 5 and S47-S49 in the SI) indicated that ligand 24, which showed the strongest binding affinity, engages in multiple stabilizing interactions within the P. falciparum aminopeptidase N (3EBH) binding pocket. These include hydrogen bonds with ALA461 and GLU497, halogen bonding with GLU572, and extensive hydrophobic contacts such as π–π stacking with HIS496 and TYR580, and π–π-alkyl interactions with VAL459, VAL523, TYR575, and MET1034. In contrast, chloroquine interacts primarily through hydrogen bonding with ALA461, GLY460, LYS1044, TYR1077, TYR574, and GLU497, along with weaker hydrophobic interactions such as π–σ and π–π T-shaped contacts. This difference in interaction profiles explains the superior binding affinity of ligand 24, which benefits from stronger and more diverse stabilizing contacts with key residues in the active site. Ligands 37 and 38 demonstrated multiple hydrogen bonding and hydrophobic contacts, supporting their observed binding affinities.

Close

Figure 5.

2D ligand-protein interaction diagrams for ligand 24 and quinine with 3EBH.

Export to PPT

Active site prediction using the CASTp server (Figure 6a) highlighted key residues involved in ligand binding. Figures 6(b-d) further illustrate non-covalent interaction distributions and residue contact maps for ligands 24, 37, 38, and quinine within the 3EBH binding cavity.

Close

Figure 6.

(a) CASTp-based active site prediction for 3EBH; (b) Residue interaction mapping; (c) Non-covalent interaction distribution; (d) Residue-ligand interaction map for ligands 24, 37, 38, and quinine.

Export to PPT

Figure 7 emphasizes the superior binding profile of ligand 24 with 3EBH, highlighting its potential as a drug candidate. The compound showed a well-defined fit within the binding pocket, forming stable hydrogen bonds and hydrophobic interactions, outperforming known antimalarial drugs in affinity and interaction specificity. These findings support ligand 24 as a promising candidate for further development as an antimalarial or anticancer agent targeting key proteins such as 3EBH and 4UV7.

Close

Figure 7.

(a) Protein–ligand binding cavity for 3EBH; (b) Ligand 24 docked in the active site; (c) Active site surface view; (d) Hydrogen bond visualization between ligand 24 and 3EBH.

Export to PPT

3.3. Analysis of physicochemical and pharmacokinetic properties

The evaluation of physicochemical and pharmacokinetic properties is essential in early drug development to assess the suitability of compounds as potential therapeutic agents. In this study, the quinoline-4-carboxamide derivatives (1-43) were analyzed based on Lipinski’s “Rule of Five” and Veber’s criteria to predict oral bioavailability and drug-likeness. According to Lipinski’s rule, an ideal orally active drug candidate should possess a molecular weight (MW) ≤ 500 g/mol, a log P ≤ 5, no more than five hydrogen bond donors (HBDs), no more than 10 hydrogen bond acceptors (HBAs), and a topological polar surface area (TPSA) ≤ 140 Ų [33,34]. Veber’s rule further stipulates fewer than 10 rotatable bonds and a TPSA ≤ 140 Ų as essential for optimal oral bioavailability.

Using SwissADME, ligands 1-43 were analyzed for compliance with established drug-likeness rules. All compounds met Lipinski’s rule of five and Veber’s rule, suggesting favorable oral bioavailability potential. Most ligands displayed molecular weights below 500 g/mol and exhibited balanced hydrogen bonding capacity and lipophilicity. Bioavailability scores further supported their suitability for advancement in drug development. The ADMET analysis was also carefully re-examined, confirming that none of the 43 derivatives triggered PAINS, hepatotoxicity, or Ames mutagenicity alerts. This strengthens confidence that predicted toxicity risks do not undermine the favorable docking outcomes. Additionally, all ligands were screened for PAINS, with no alerts detected, reinforcing their potential as clean, drug-like candidates. Systematic evaluation against multiple drug-likeness criteria revealed that, although a subset of compounds presented a single violation of Lipinski’s parameters (compounds 9, 10, 16, 19, 25-28, 32-37, 40-41, and 43), the majority fully complied with the thresholds. Importantly, all derivatives satisfied Veber’s rule, indicating strong potential for oral absorption. Assessments with the Ghose, Egan, and Muegge filters consistently classified the entire series as drug-like, with no PAINS alerts observed across the dataset (Table S48). These findings demonstrate that the designed derivatives possess favorable safety predictions and robust drug-likeness properties, complementing their docking performance. Ligand 24 exhibited the strongest binding affinity (−10.3 kcal/mol) among the series, with no Lipinski violations. Its favorable physicochemical characteristics, including relatively low molecular weight and optimal TPSA, highlight its promise as a drug-like candidate.

Beyond drug-likeness, medicinal chemistry profiling was conducted using the Molinspiration tool to explore the biological target classes for these ligands. This included predictions for GPCR binding, ion channel modulation (ICM), KI, NRL binding, PI, and EI. Quinine (D3) showed a GPCR activity value of 0.23, while ligands 24 and 37 had values of 0.19 and 0.20, respectively, suggesting comparable or superior drug-like potential. Table S49 details the Molinspiration-based bioactivity scores for ligands 1-43. Ligand 24 showed notable activity across multiple pharmacological targets, including protease and enzyme inhibition, with a high PI value (0.41) and EI score (0.18).

To further assess the therapeutic potential of the ligands, in silico pharmacokinetic and ADMET (absorption, distribution, metabolism, excretion, and toxicity) profiling was conducted using the admetSAR and SwissADME platforms, evaluating seven key descriptors: human intestinal absorption (HIA), blood-brain barrier (BBB) penetration, cytochrome P450 inhibition (CYP3A4 and CYP2C19), water solubility, hERG inhibition potential, and synthetic accessibility (SA). The analysis revealed that all ligands exhibited high predicted intestinal absorption and medium BBB penetration, with ligands 8, 9, and 13 showing the highest BBB values. Among them, ligand 24 emerged as the most favorable candidate, demonstrating strong intestinal absorption, moderate BBB penetration (0.8000), a safe hERG inhibition profile (pIC₅₀ = 0.7904), and excellent SA (1.00), indicating ease of synthesis.

Based on in silico predictions, ligands 24 and 37 exhibit superior bioavailability and safety profiles compared with quinine and hydroxychloroquine. While all compounds are predicted to have excellent oral absorption, ligand 24 shows a more favorable balance of physicochemical properties, including a higher oral bioavailability score and a lower clogP value (2.3 vs. 3.24 for quinine), indicating improved solubility and absorption. Both ligands also demonstrate a markedly safer hERG profile, suggesting a reduced risk of cardiotoxicity relative to traditional quinolines. Notably, ligand 24’s moderate BBB penetration offers potential efficacy against cerebral malaria while minimizing the neuropsychiatric side effects often associated with higher CNS penetration of quinine and hydroxychloroquine. Coupled with its simplified SA, these attributes position ligand 24, alongside ligand 37, though with particular promise, as a next-generation antimalarial candidate with optimized bioavailability and an improved safety window.

Table 3 summarizes the comprehensive ADMET predictions for all ligands. The results indicate that most quinoline-4-carboxamide derivatives satisfy the key pharmacokinetic criteria for bioavailability and therapeutic safety [35], thereby supporting their consideration as computational leads that warrant subsequent experimental validation. To visually interpret oral bioavailability potential, a bioavailability radar diagram was generated for the 43 ligands (Figure S50 in the SI). This tool graphically assesses six key biopharmaceutical properties, including lipophilicity, size, polarity, solubility, saturation, and flexibility. The radar plots revealed minimal deviations from optimal ranges, supporting the oral drug candidacy of these compounds.

Ligand 24 has been explicitly identified as the best-balanced lead, combining the highest docking affinity (−10.3 kcal/mol) with favorable ADMET predictions, including good solubility, high intestinal absorption, acceptable BBB penetration, and a safe hERG profile (clogP = 2.3, TPSA = 92 Ų, hERG pIC₅₀ = 0.79). It also demonstrates good SA (3.21). To situate its pharmacokinetic profile within the clinical context of quinoline antimalarials, ligand 24 was benchmarked against quinine and hydroxychloroquine using identical in silico protocols. As summarized in Table 3, ligand 24 exhibits superior oral bioavailability (SwissADME score 0.9879 vs. 0.821-1.000 for the comparators), maintains moderate BBB penetration (0.80) within the therapeutic window for cerebral malaria while remaining below the neurotoxicity threshold (≥0.9), and displays a markedly safer hERG profile (pIC₅₀ 0.79 vs. 0.61 for quinine). Furthermore, its SA score of 1.0, substantially lower than that of quinine (3.8), indicates a concise two-step synthetic route from commercially available reagents, offering scalability and cost advantages. Collectively, these comparative findings reinforce ligand 24 as the most promising lead compound identified in this computational study.

The predicted moderate BBB penetration represents a critical pharmacological feature that requires careful risk-benefit evaluation. On the one hand, CNS accessibility is advantageous in treating cerebral malaria, a severe neurological complication of Plasmodium falciparum infection, as it enables drug exposure at sufficient concentrations to target sequestered parasites in the brain. On the other hand, CNS penetration also carries the potential for adverse neurological effects, a well-documented liability of several quinoline derivatives, which are associated with neurotoxicity and neuropsychiatric events at higher CNS levels. Therefore, the “moderate” BBB penetration predicted for ligand 24 is of particular significance, as it suggests the possibility of achieving therapeutic CNS concentrations while remaining below the threshold of severe neurotoxicity. This balanced profile highlights BBB permeability as a risk-benefit parameter that warrants further experimental validation in efficacy and neurotoxicity models to confirm the therapeutic window.

3.4. Molecular dynamics simulation

To further validate the stability and interaction profiles of promising ligands identified via molecular docking, MD simulations were conducted for 20 ns. The protein target aminopeptidase N (PDB ID: 3EBH) was analyzed in complex with ligands 24, 37, 38, and the reference drug quinine. Previous docking analyses demonstrated strong binding affinities of −10.3, −10.0, −9.8, and −8.3 kcal/mol, respectively. The primary goal of the MD simulations was to evaluate the temporal evolution, flexibility, and structural stability of these complexes under physiological conditions.

Key simulation parameters included root-mean-square deviation (RMSD), root-mean-square fluctuation (RMSF), radius of gyration (Rg), hydrogen bonding profiles, potential energy, and temperature variations. Simulations were performed at four temperatures (300, 305, 310, and 320 K) to explore thermal stability and conformational adaptability. Figures 8-10 and Supplementary Figures S51-S77 present the dynamic behavior and thermodynamic profiles of the protein-ligand complexes.

Close

Figure 8.

The progression of the hydrogen bond count (HBs) for (a) Ligands 24, 37, 38, and reference drug quinine at 300 K; (b) for ligand 24 during the 20 ns MD simulation.

Export to PPT

Close

Figure 9.

Temperature (K) versus time (ps) plots for (a) the combined system at four different temperature conditions (300, 305, 310, and 320 K), and (b) RMSD for the combined docked complex involving the protein (PDB: 3EBH) and ligand 24 (300 K: black line, 305 K: red line, 310 K: green line, 320 K: blue line) throughout the 20 ns MD simulation.

Export to PPT

Close

Figure 10.

PCA of MD trajectories for the target protein (PDB: 3EBH) and ligand 24 complex at (a) 300 K, (b) 305 K, (c) 310 K, and (d) 320 K (Intermediate states are marked by white dots, energetically unstable conformations are represented by black dots with scattering, and stable conformation states are denoted by red dots).

Export to PPT

3.5. Limitations

While this study provides a comprehensive in silico framework for antimalarial drug discovery, several limitations must be acknowledged. First, all binding affinity estimates are derived from static crystal structures and force-field approximations, which may not fully capture protein flexibility or the influence of physiological post-translational modifications. Similarly, though widely validated, docking scores and ADMET predictions carry inherent false-positive and false-negative risks; subtle chemical features outside the scope of the training sets could affect absorption, metabolism, or off-target activity.

A significant finding is the high predicted binding affinity of ligand 24 for both the malarial target (3EBH, –10.3 kcal/mol) and the human epidermal growth factor receptor (EGFR, PDB: 4UV7, –10.2 kcal/mol). This dual activity raises concerns regarding selectivity, as unintended EGFR inhibition could result in host toxicities. While such activity may open avenues for oncology-related applications, it poses a significant liability for antimalarial development. Therefore, future work must prioritize experimental assessment of selectivity by directly comparing inhibitory activity against both parasite and human targets to define the therapeutic window better.

Additionally, the MD simulations (20 ns) used here may not sufficiently capture slower conformational changes relevant to binding kinetics. The absence of experimental validation, such as biochemical assays, parasite growth inhibition studies, microsomal stability tests, and in vivo pharmacokinetic or toxicity evaluations, further limits the ability to draw definitive conclusions regarding efficacy and safety. Consequently, the designation of ligand 24 as a “lead” compound should be considered provisional until robust experimental data are available to support its translational potential.