Protein and ligand preparation
For this investigation, the enzyme Decaprenylphosphoryl-β-D-ribose-2-oxidase (DprE1) from Mycobacterium tuberculosis (PDB ID: 4FDO) was chosen. DprE1 is an important enzyme that Mycobacterium tuberculosis (Mtb) uses to synthesize its cell walls. It is also a prospective target for the development of antituberculosis drugs. DprE1 (PDB ID: 4FDO) exhibited slightly higher qualities in terms of Rfree, class score, Ramachandran outliers, side chain outliers, and RSRZ outliers based on the preliminary protein quality assessment utilizing the wwPDB validation slider (Table 1).
Table 1: Information on reported DprE1 PDB crystal structures and their protein quality score
Protein
|
Reference
|
Resolution
|
wwPDB validation
|
Rfree
|
Clash sore
|
RMCD outliers
|
Side chain outliers
|
RSRZ outliers
|
4FDO
|
[23]
|
2.40
|
0.197
|
4
|
0.0%
|
2.8%
|
0.7%
|
4KW5
|
[24]
|
2.61
|
0.242
|
7
|
0.0%
|
5.7%
|
1.2%
|
4NCR
|
[25]
|
1.88
|
0.234
|
9
|
0.1%
|
9.1%
|
2.9%
|
4P8C
|
[26]
|
1.95
|
0.237
|
4
|
0.1%
|
4.5%
|
8.1%
|
4P8T
|
[27]
|
2.55
|
0.237
|
5
|
0.1%
|
4.9%
|
4.3%
|
4P8L
|
[28]
|
2.02
|
0.228
|
4
|
0.0%
|
4.0%
|
6.3%
|
4P8K
|
[29]
|
2.49
|
0.265
|
7
|
0.0%
|
7.1%
|
5.7%
|
6HEZ
|
[30]
|
2.30
|
0.243
|
6
|
0.1%
|
5.1%
|
9.3%
|
NB: RMCD: Ramchandran
The model was submitted to SAVE v6.0 (available at https://saves.mbi.ucla.edu/) in order to further evaluate its quality after the co-crystallized ligand, cofactors, and water molecules were removed. SAVES v6.0 is a complete toolbox that includes WHATCHECK, VERIFY3D, ERRAT, and PROVE, which is able to predict the structural stereochemical property of a protein model. The selected protein structure (4FDO) had an overall quality score of 95%, which is good, according to the ERRAT server. The result of ERRAT assessment of the protein is shown as a graph between residues and error values (Fig. 1a). On the error axis, two lines are drawn to indicate the confidence with which it is possible to reject regions that exceed that error value. Regions with grey color depicts good regions in the protein structure while regions with yellow and red represent poor or problematic part. VERIFY3D analysis of the protein structure revealed that 93% of the residues have average 3D-1D score > 0.1. Additionally the Ramachandran plot generated by the PROCHECK tool was employed to assess the structural stereochemical property of the protein structure. It plots the phi (ф) and psi (Ψ) dihedral angles to identify the sterically allowed and disallowed conformations of a protein's backbone, ensuring the quality of the protein structure and identifying any errors related to it. The Ramachandran plot of 4FDO, as shown in Fig. 1b indicates that 92.2% (329), 7.8% (28), 0%(0) and 0% (0) amino acid residues are present in the most favored regions, additionally allowed regions, generously allowed regions, and disallowed regions. The 30 prospective chalcones p-methyl acetophenone, p-methoxy acetophenone and p-chloro acetophenone series presented in (S1 Fig 1a, 1b and 1c) respectively were the ligands used in this investigation. After adding hydrogens and Gasteiger charges to the ligands, MM94FF was used to optimize their geometry.
Molecular docking
Molecular docking has grown in significance as a modeling technique for drug discovery [31]. Using molecular docking technique and other in silico methods, several investigations have revealed antitubercular activity of several molecules [32]. This computational approach has been used in the current investigation to elucidate the binding modes and interaction mechanisms of prospective chalcones on the Mtb DprE1 enzyme (4FDO). The redocked CT319 and the reference ligand CT319 obtained from PDB were compared in order to evaluate the accuracy of the docking process. Good alignment of the aromatic, trifluoromethyl, and nitro groups between the two structures suggests that our docking procedure was accurate. The RMSD value of redocked poses CT319 was 0.53 angstroms. RMSD values less than 2.0 Å are considered good, but values closest to zero are ideal. Hence, the docking process was deemed valid and could proceed. The 3D diagram of the superimposition of the reference ligand and the redocked pose is represented in Fig. 2a. The 2D protein-ligand interactions of reference ligand CT319 obtained from PDB and redocked CT319 is represented in Fig. 2b and Fig. 2c respectively (Fig. 2).
Table 2. Interaction of co-crystalized ligand, standard drugs (isoniazid and pyrazinamide) and docked chalcones with favorable binding affinity, against Mtb DprE1 (PDB ID 4FDO)
The binding energy of the reference ligand (CT319) with 4FDO was -10.0 kcal/mol. Analysis of DprE1-CT319 docked complex revealed non-covalent bonding interaction such as conventional hydrogen bonds, carbon hydrogen bonds, halogen, pi-alkyl and alkyl bonds as the main types of interactions mediating the binding of the reference ligand (CT319) to Mtb DprE1. All of the prospective chalcones hit the active site of Mtb DprE1 with estimated binding energies ranging from -6.5 to -10.4 kcal/mol. Table 2 depicts the interactions of chalcones with favorable binding affinity. In addition, comparing prospective chalcones to the standard antitubercular drugs such as isoniazid and pyrazinamide, all of them demonstrated a higher binding affinity towards the Mtb DprE1target. We discovered that the chalcones scaffolds bearing indole moiety in the methylacetopheneone (9A) methoxyacetophenone (9B) and chloroacetophenone (9C) series exhibited higher docking score than the reference ligand. This finding highlights how important it is to combine chalcones with aromatic heterocyclic rings, such indoles, to further enhance their pharmacological and biological properties.
The binding pocket containing key amino acid residues allows the compound to form a non-covalent bonding contact with Mtb DprE1 (4FDO).
As depicted in Fig. 3, we observed Leu317, Phe320, Trp230, Leu363, and Val365 residues in the tail region of DprE1 active site were involved in the formation of a hydrophobic cavity that is occupied by the non-polar p-choro and p-methyl substituent of chalcone-indole complex derivatives. By forming van der Waal contacts with the terminal phenyl ethyl group, these residues appeared to play significant role in stabilizing the reference ligand (CT319) in this region of DprE1 active site. This could possible suggest that the introduction of halogens and non-polar alkyl moieties on aromatic substituents of chalcones could makes them more attractive to this portion of the active site of DprE1 protein. Additionally, the aromatic heterocyclic ring of chalcone-indole complex derivatives nicely fit into a hydrophobic cavity created by His132, Gly133, Val365, Lys367, Phe369 and Asn385 within the head region of the DprE1 active site. DprE1's inhibitory activity is highly driven by the occupancy of this region [33]. The head region is known to accommodate very hydrophobic functional groups such as those containing trifluoromethyl groups. For instance, the reference ligand (CT319) possess a –CF3 moiety, which ideally pack well in this region by forming dense van der Waal interaction with these residues. Aside these functional groups, the docking studies revealed that an indole moiety could also occupy this region and interact fully with the active site residues of DprE1 protein. Hence, our theoretical hypothesis was that chalcones' indole ring could strongly drive and influence the ligand receptor affinity and inhibitory activity of DprE1. It is also worth noting that, the chalcone-indole complex derivatives again bonded with the backbone of Cys387, Val365, Gly117 and Lys418 in the trunk region of the DprE1 active site. Per previous reports, [33] Cys387 plays a crucial role in forming covalent bond with the nitro groups of compounds such as pBTZ169, BTZ043, among others and Lys418 is involved in the stabilization of the adduct by forming an additional hydrogen bond. This hydrogen bond interaction in this case involved Lys418 and the enone ketone moiety of the derivatives of the chalcone-indole complex. Once more, the binding mechanism of the hit compounds also revealed the formation of an extra conventional hydrogen bond between the -NH moiety of the indole ring and the amide oxygen of the ASN385 side chain. Although this residue is found in the DprE1 protein's binding pocket, its interactions with known DprE1 inhibitors, including pBTZ169, BTZ043, and CT319 has not been fully established. Nevertheless, we think that this extra hydrogen bond interaction also played a role in the chalcone-indole complex derivatives' binding and stability to the DprE1 protein. From the docking studies, compounds 9A, 9B and 9C appear to share an interesting correlation with known inhibitors of DprE1 protein. This suggests that the compounds may be just as stable in vitro, given that the interaction level seems to be the same as that of DprE1 inhibitors. Hence, this prediction can serve as the basis for additional in vitro experiments, including enzymatic ones.
ADME and toxicity assessment
The physicochemical, pharmacokinetic, and safety characteristics of chalcone-indole complex derivatives were estimated in silico (Table 3). The compounds displayed appreciable gastrointestinal absorption profile. Notably, all the three compounds conform to Lipinski, Ghose, Verber, Egan and Muegge rule for orally active drugs hence has the potential to be developed into orally active lead compounds. Additionally, they were not substrates for p-glycoprotein therefore there is less chance that they will be pushed back into the gut lumen, which would reduce their efficacy. The score on the synthetic accessibility of chalcone-indole complex derivatives was below 3, suggesting that they are easily synthesized, on a scale ranging from 1 (very easy) to 10 (very difficult).
Table 2a. In-silico ADMET screening for selected compounds (9A, 9B and 9C)
Table 2b: In-silico ADMET screening for selected compounds (9A, 9B and 9C)
Key: BBB: blood-brain barrier; SA: Synthetic Accessibility; MW: molecular weight; HBD: hydrogen bond donor; HBA: hydrogen bond acceptor; RB: Rotatable bonds; GIA: gastrointestinal absorption; TPSA: topological polar surface area; L-Ro5: Lipinski’s rule of five; L-Ro5: HBD ≤ 5, HBA ≤ 10, MW ≤ 500, log P ≤ 5; % Abs = 109-0.345 TPSA.
Many xenobiotics, including pharmaceuticals are detoxify by microsomal heme containing superfamily of enzymes called cytochrome P450 found mainly in the liver [34]. These enzymes usually will hydroxylate, oxidize or dealkylate chemical moities present in molecules during phase I metabolic reactions [35]. We investigated the hit chalcones on the CYP3A4 reactivity site, denoted by green circles, to see if the compounds would be able to access the site of metabolism after entering the bloodstream. We calculated only the intrinsic reactivity and accessibility, labelled as overall site of metabolism (SOM) score. This was displayed as green circles (Fig. 1), in which the radius was proportional to the score, where larger scores meant higher reactivity.
The compounds have also been determined by in silico toxicity assessment to be non-hepatotoxic and not to inhibit the human Ether-à-go-go-related gene, making them less likely to contribute to cardiotoxicity. The chalcone-indole complex derivatives were highly hydrophobic and hence have the propensity to penetrate the lipid-rich cell wall barrier that protects M. tuberculosis against most antibiotics and antitubercular drugs. However, this feature could also allow them to cross the blood-brain barrier, thereby increasing the risk of adverse effects in the central nervous system. This allow them for future optimization to attain the desired drug-like properties.
AntiTB sensitivity prediction
Further, the mycoCSM website (available at https://biosig.lab.uq.edu.au/myco_csm/) was used to predict the compounds' minimum inhibitory concentrations (MIC) against Mycobacterium tuberculosis. The predicted MIC values are presented in the Table 3. The compounds recorded lower predicted minimum inhibitory concentrations (MIC) compared to pyrazinamide. Antimicrobial agents with lower minimum inhibitory concentrations (MIC) are more effective because less of the compound is required to inhibit the growth of the organism. We proposed that these compounds might exhibit antitubercular activity by inhibiting the DprE1 enzyme.
Table 3: The predicted MIC values of the prospective chalcones
Compound
|
MIC (log µM)
|
MIC (um)
|
9A
|
-4.232
|
0.0145433
|
9B
|
-4.115
|
0.0163380
|
9C
|
-4.213
|
0.0148131
|
Pyrazinamide
|
-3.705
|
0.0246166
|
Isoniazid
|
-4.899
|
0.0074606
|
Molecular dynamics
Based on the molecular docking results, the chalcone-indole complexes derivatives pack nicely within the hydrophobic cavities of the head, trunk and tail regions of the DprE1 active site, with compound 9C recording the most favorable free energy of binding. Using Gromacs 2023.2, a 200 ns molecular dynamics simulation was carried out to evaluate the time-dependent stability of the DprE1-9C complex. The protein topology file was created using the gmx pdb2gmx module, and the force field used in the MD simulation was CHARMM36-jul2022. Prior to creating the ligands' topology files, hydrogen atoms were added to the docked ligand and the bond orders were further sorted. The output file was fed into Cgeff server (available at https://cgenff.silcsbio.com/) to generate the ligand stream file. This file was opened in a notepad to assess ligand quality using the highest penalty score of the associated parameters. Penalties lower than 10 indicate the analogy is fair, penalties between 10 and 50 mean some basic validation is recommended and penalties higher than 50 indicate poor analogy and mandate extensive optimization and validation (https://cgenff.silcsbio.com/). Our ligands were deemed suitable for topology file generation based on the evaluation. Additionally, the protein and ligand topologies were merged to produce the protein-ligand complex file. A dodecahedron box filled with simple point charge (SPC) water model and 0.15M NaCl were to neutralize the charge on the protein surface. Steepest Descent integrator was used to minimize the energy of the system. For Equilibration of the systems at 300 K and 1.0 bar temperature and pressure, respectively, constant-temperature, constant-pressure (NPT) and constant-temperature, constant-volume (NVT) ensemble were utilized simultaneously. Finally, in order to determine if the complex's stability would be maintained, a 200 ns MD simulation was performed. Utilizing gromacs inbuilt tools to analyze the MD trajectories, it was found that the root mean square deviation of 9C increased slightly from 0.10 -0.19 nm for the first 5 ns and then remind steady (0.1-0.17 nm) for 135 ns which may be due to the interaction of the molecule with active site amino acids. Therefore, the molecule can bind with the mtb DprE1 protein and form a stable complex with it. The maximum and average RMSD of 9C did not exceed 0.25 nm (~2.5 Å) throughout the entire 200 ns MD simulation. This suggests a high likelihood of the complexes being stable in vitro. The RMSD of the docked 9C and CT9319 during a 200 ns MD simulation is displayed in Fig. 5
The RMSD of the docked complex was further analyzed at 50 ns intervals as seen in table 4. In these time intervals, the compound 9C displayed less variability than the reference ligand (CT319). This implies that compared to reference ligand the hit compound is less likely to result in structural instability when bound to Mtb DprE1 protein.
Table 4: The RMSD of the docked complex compared to the reference ligand
|
Average RMSD (nm)
|
Time (ns)
|
DprE1-9C
|
DprE1-CT319
|
0-50
|
0.1133
|
0.1565
|
50-100
|
0.1105
|
0.1825
|
100-150
|
0.1274
|
0.2604
|
150-200
|
0.1674
|
0.3007
|
0-200
|
0.1297
|
0.2265
|
To explore residue fluctuation upon binding the root mean square fluctuation (RMSF) of the amino acids of the protein-bound and unbound state with 9C and CT319 was calculated from their 200 ns. The change of RMSF values of amino acids is almost similar in both systems. All the active site residues found in molecular docking were found to be stabilized throughout the simulation period and this could be due to the direct or indirect interactions with the compound. Hence, these amino acids can form a stable complex with compound 9C in the active site of the target protein (Fig. 6).
We noticed significant fluctuations in two particular regions of DprE1 (residues 269–303 and 316–330) in both free and ligand-bound states. These are very flexible regions of DprE1 and represent the distorted loops of the tail portion of the protein. These regions of the target protein usually cover the binding pocket and function as a gatekeeper, allowing the binding site to open and close. Aside this, no significant fluctuation seemed to be present in the residue binding especially those implicated in the binding of the compound.
We also computed the radius of gyration, Rg, and utilized the parameter to study the conformational changes of DprE1 protein between its free and ligand-bound states. Analysis of radius of gyration plot (Fig. 7) indicated that the compactness of the protein in both free and ligand-bound states is almost similar. The average Rg value of the apoprotein protein and its complex with compound 9C was found to be approximately 2.16 nm and 2.18 nm respectively
The maximum and average number of hydrogen bonds formed between the compounds and Mtb DprE1 protein during the MD simulation were ascertained by hydrogen bond analysis. Formation of hydrogen bonding and optimized hydrophobic interaction between a ligand and a target is essential for ligand stability within the active site and help alter binding affinity and drug efficacy [36]. A maximum of three hydrogen bonds were found in DprE1-9C. However, throughout the simulation time, two hydrogen bonds predominated (Fig. 8)
As a result, and in line with the interactions observed in the docking investigation, van der Waals interactions played a major role in the compounds' binding and stability with the Mtb target protein.