6 K. L. JOSEPH LIBIN L ET AL.
coordinate length of approximately 5 nm. The potential of mean force (PMF) (Allen et al., 2006) profile
was generated using umbrella sampling simulation outputs, employing the Weighted Histogram Analysis
Method (WHAM) (Kumar et al., 1992) in conjunction with GROMACS tools (Yekeen et al., 2023). Standard
deviation criteria and block averaging across replicates were employed to assess histogram overlap and
WHAM convergence.
The PMF plot illustrates the binding free energy profile, with the y-axis indicating free energy in kcal/
mol and the x-axis representing the distance over which the system has been extended. Thus, the US
binding free energy (DGUS) was evaluated by calculating the difference between the energy minimum
and the plateau region observed in the potential of mean force (PMF) curve. 56 ns of umbrella sampling
simulations were conducted across 28 windows for each protein-ligand configuration, facilitating
improved sampling along the reaction coordinate to obtain precise free energy estimations. All simula-
tions were conducted using a high-performance GPU-enabled computing workstation. We ensured the
repeatability of trajectories and convergence of sampling by analysing block averages and standard devi-
ations across replicates.
3. Results and discussion
3.1. Molecular docking and active site analysis
The results of molecular docking calculations, initiated by the redocking of the co-crystallized ligand
SAM within the binding pocket of DENV MTase. The redocking (Santos-Martins et al., 2014) computation
yielded nine generated docking poses, with the optimal pose illustrated in Figure 2 exhibiting a docking
score of −8.2 kcal/mol. The active amino acid residues within the SAM binding domain of the dengue
methyltransferase target (PDB ID: 3P97) were identified in the residue range 55–148, as documented in
PDB site records. Fifteen amino acid residues within this binding domain exhibit non-covalent interac-
tions with the SAM ligand in its co-crystallized conformation. Out of the 15 examined residues, all
except, residue id 56 were present in the re-docking pose, suggesting their non-covalent interactions
with the SAM ligand. Hydrogen bonds and non-hydrogen bonds, such as van der Waals interactions,
carbon-hydrogen bonds, pi-sigma bonds, and pi-alkyl interactions, are the classes of non-covalent inter-
actions observed in these re-docking poses. The close proximity of this re-docking pose to its original
pose in the crystal structure confirms the calibration and validation of the docking protocol.
The docking poses generated correspond to each selected protein-ligand complex under investiga-
tion, with results summarised in Table 1. All selected flavonoids exhibit binding scores ranging from
−8.5 to −9.5 kcal/mol (Table 1) when docked with the DENV MTase target. Neohesperidin and hesperi-
din, sharing the same structural backbone, exhibit slightly larger negative docking scores of −9.5 and
−9.2 kcal/mol, respectively. The larger negative docking scores confirm that all selected flavonoids were
Figure 2. Best re-docking pose of the co-crystallized ligand SAM at the binding site of DENV MTase target protein.
Various protein-ligand non-covalent interactions were also highlighted.