**3.12 Molecular dynamic simulation of protein–ligand complex**

Molecular dynamics simulation allowed the early view of proteins as relatively rigid structures to be replaced by a dynamic model in which the internal motions and resulting conformational changes play an essential role in its function [59]. An RMSD plot generated after molecular dynamics simulation showed a deviation of about 0.25 Å (**Figure A7**). Further scrutinized with molecular dynamics simulations gave the protein a dynamic dimension to its 3D structural form producing a realistic environment for the ligand interactions that were carried out in the docking process.

Molecular dynamics simulations can also capture a wide variety of important biomolecular processes, including conformational change, ligand binding, and protein folding [60]. The stability of docked protein–ligand complexes was determined by their (RMSD) plots generated from the MD simulation output file. The backbones of the four complexes were observed to be stable over time (**Figure 3**). The fluctuations of the protein–ligand complexes were analyzed within the system to check for movement and structural stability during the course of the simulation. These movements and stability are significant for the complex functioning inside living systems. The backbone of the *Lm*TR-ZINC8782981 complex showed the greatest stability with an average RMSD of 0.25 nm amongst all the complexes. The *Lm*TR-CHEMBL1277380 complex was fairly stable with RMSD of 0.4 nm. Amongst the four leads, Pectachol was found with the lowest RMSD of 0.3 nm. In terms of stability, the compound Marmin and Colladonin proved to be very much stable around 0.73 nm over the production time of 100 ns. The RMSD of Karatavicinol was observed to show stability from 0 to 60 ns and increased its RMSD to 0.5 nm from 60 to 100 ns.

The flexibility of residues contribution by the *Lm*TR was assessed by the root mean square fluctuation (RMSF). RMSF indicates the flexibility of different regions of a protein, which can be related to crystallographic B factors [61]. The results of

#### **Figure 3.**

*RMSD values of the* Lm*TR-ligand complexes of the four leads (Karatavicinol, Marmin, Pectachol, and Colladonin) and the two known inhibitors after 100 ns. The complexes in the graph are color coded.*

the RMSF plots showed consistency for the docked complexes (**Figure A8**). The highest fluctuations exhibited was observed around residue numbers 70–90, with Karatavicinol and CHEMBL1277380 showing higher fluctuation levels followed by ZINC8782981 and Marmin complexes. Other regions where good fluctuations were observed include residues between 395 and 410 and 465–480. Overall, Marmin showed more fluctuations around most residues with a distinct difference at residue numbers 260 and 310.

The compactness of the complexes over simulation time is determined by the Rg. If a protein is folded well, it will likely maintain a relatively steady value of Rg, whereas its value will change over time if the protein unfolds [62]. Rg values of all complexes indicated stable complexes over 100 ns (**Figure A9**). The Rg graph showed most compounds experienced a fairly stable Rg. Marmin experienced the lowest Rg value around 2.33 nm compared to other complexes. This was followed by Colladonin, Pectachol and Karatavicinol with Rg values of 2.37, 2.42, and 2.45 nm, respectively. Between the known inhibitors, CHEMBL1277380 was observed to have an average Rg value of 2.46 nm whilst ZINC8782981 showed the average highest value of around 2.5 nm. Inferring from the Rg graph, the compactness of the *Lm*TR–Marmin, –Colladonin and –Pectachol complexes were maintained after complex formation.

#### **3.13 Evaluation of leads using MM-PBSA**

MM-PBSA was employed to calculate free binding energies by per-residue decomposition of the protein complexes. At a quantitative level, simulation-based methods provide substantially more accurate estimates of ligand binding affinities (free energies) than other computational approaches such as docking [63]. Residues contributing binding free energy greater than 5 kJ/mol or less than 5 kJ/mol are considered critical for binding of a ligand to a protein [64]. MM-PBSA results showed only Asp327 amongst the hydrogen bonding residues of Karatavicinol to contribute a per residue decomposition energy of 13.65 kJ/mol. Amino acid residue Asp35 (21.89 kJ/mol) was observed with such greater contribution (**Figure A10**). The complex of *Lm*TR–Marmin also showed surrounding hydrophobic residues Asp35 (8.62 kJ/mol), Ala46 (7.65 kJ/mol), Arg290 (8.83 kJ/mol), and Glu141

(5.12 kJ/mol) with their energy decomposition to be greater than 5 kJ/mol and less than 5 kJ/mol. The only hydrogen bonding residue that showed a relevant contribution of energy decomposition was Thr51 (16.36 kJ/mol) (**Figure 4**). Hydrophobic amino acid residues Lys61 (5.16 kJ/mol), Tyr198 (11.29 kJ/mol), Asp327 (5.56 kJ/mol), and Arg331 (6.38 kJ/mol) showed relevant contribution to the total binding energy of the *Lm*TR–Colladonin complex (**Figure A11**). Moreover, only hydrogen bonding residue Lys60 (13.32 kJ/mol) in *Lm*TR–Pectachol complex showed to be a critical residue in binding. Other surrounding residues contributed substantially to the per residue energy decomposition in the *Lm*TR–Pectachol complex. This included Lys61 (10.97 kJ/mol), Arg287 (6.91 kJ/mol), Asp327 (9.30 kJ/mol), Met333 (6.72 kJ/mol), Leu334 (8.91 kJ/mol), Lys361 (6.46 kJ/ mol), and Cys364 (6.07 kJ/mol) (**Figure A12**). Deducing from the substantial contribution of energy per decomposition of residues, we propose Asp35, Thr51, Lys61, Tyr198, and Asp327 to be critical in intermolecular bonding and stabilization of ligands at the FAD active site.

## **3.14 Other energy terms**

Van der Waals forces, electrostatic and polar solvation energies, and SASA are relevant energy terms contributing to the overall free binding energy of the complex. The van der Waals energy refers to the weak attraction existing between the intermolecular forces. The van der Waals energy observed in our study showed Karatavicinol and CHEMBL1277380 to have the lowest and highest energy of 228.565 and 171.823 kJ/mol, respectively. Colladonin, Marmin, and Pectachol also showed relatively low van der Waals energy of 189.289, 189.229, and 209.538 kJ/mol, respectively as compared with ZINC8782981 with 222.123 kJ/mol. Electrostatic energy refers to the potential energy of a system consisting of different electric charges [65–67]. The lowest electrostatic energy was exhibited by Marmin (386.401 kJ/mol) followed by Pectachol (286.260 kJ/mol), and Colladonin (249.067 kJ/mol). Karatavicinol and the other two inhibitors were observed with high electrostatic energy (**Table 2**). Some studies have observed that

#### **Figure 4.**

*MM-PBSA plot of the binding free energy decomposition contribution per residue of* Lm*TR–Marmin complex. Coded red lines represent surrounding active site amino acid residues.*

van der Waals and electrostatic forces contribute favorably to the energetics of binding along with simulations that favor the binding of complexes [66, 68].

Polar solvation energy on the other hand refers to the electrostatic interaction that exists between the solute and the continuum solvent [69]. The highest polar solvation energy amongst the leads was exhibited by Marmin (484.074 kJ/mol) and the lowest by Karatavicinol (227.483 kJ/mol). Solvent accessible surface area (SASA) energy was calculated after MD. This represents the non-polar solvation energy [69]. This energy measures the interactions that exist between the complex and the solvents. Amongst the leads, Karatavicinol obtained the lowest SASA energy followed by Pectachol, Colladonin, and Marmin (**Table 2**). Relative to these were the low SASA energies of the inhibitors ZINC8782981 and CHEMBL1277380.

The total contribution of these energies enabled the final estimation of the free binding energies in the complexes (**Table 2**). The lowest free binding energy contributing to more stability of the protein–ligand complex was observed by Marmin (109.114 kJ/mol). Next amongst the four complexes was Pectachol (63.487), followed by Karatavicinol (57.644 kJ/mol), and Colladonin (48.936 kJ/mol). The low binding energy of Marmin was much closer to that of CHEMBL1277380 (111.732 kJ/mol) with that of Pectachol higher than ZINC8782981 (54.399 kJ/ mol). These energies address the potential of Marmin and Pectachol to bind most effectively at the active site of *Lm*TR. *Lm*TR–Marmin's free binding energy correlated with the low binding energy (9.3 kcal/mol) from docking. That of Pectachol showed a good free binding energy than that obtained from docking. This was better than that of Karatavicinol and Colladonin (**Table 1**).
