In Silico Analysis of Some Phytochemicals as Potent Anti-tubercular Agents Targeting Mycobacterium tuberculosis RNA Polymerase and InhA Protein

Tuberculosis (TB) is a contagious disease, caused by Mycobacterium tuberculosis (MTB) that has infected and killed a lot of people in the past. At present treatments against TB are available at a very low cost. Since these chemical drugs have many adverse effects on health, more attention is now given on the plant-derived phytochemicals as potential agents to fight against TB. In this study, 5 phytochemicals, 4-hydroxybenzaldehyde, benzoic acid, bergapten, psoralen, and phydroxybenzoic acid, are selected to test their potentiality, safety, and efficacy against two Preprints (www.preprints.org) | NOT PEER-REVIEWED | Posted: 5 October 2020 doi:10.20944/preprints202010.0098.v1 © 2020 by the author(s). Distributed under a Creative Commons CC BY license. 2 potential targets, the MTB RNA polymerase and enoyl-acyl carrier protein (ACP) reductase, the InhA protein, using various tools of in silico biology. The molecular docking experiment, druglikeness property test, ADME/T-test, P450 SOM prediction, pharmacophore mapping, and modeling, solubility testing, DFT calculations, and PASS prediction study had confirmed that all the molecules had the good potentiality to inhibit the two targets. However, two agents, 4hydroxybenzaldehyde and bergapten were considered as the best agents among the five selected agents and they also showed far better results than the two currently used drugs, that function in these pathways, rifampicin (MTB RNA polymerase) and isoniazid (InhA protein). These two agents can be used effectively to treat tuberculosis.


Introduction
Tuberculosis (TB) is an ancient disease that plagued mankind many times in the past. It was responsible for many great epidemics. Mycobacterium tuberculosis (MTB) is the bacteria that is responsible for tuberculosis disease. This bacteria may have killed more people than any other microbial pathogens (Daniel, 2006). However, at present, tuberculosis is a preventable as well as a curable disease, which is possible at a very low cost. Tuberculosis is a highly contagious disease that can transmit via cough, spit, and sneezes of the infected person. MTB primarily infects the 3 lungs (. Grange and Zumla, 2002;Sepkowitz, 1996, Shah et al., 2015. If the disease is found in the lungs, then it is called pulmonary TB. However, TB can be found at other locations of the body. Such TB is called extra-pulmonary TB. Several antibiotics are used to fight against MTB. However, a new TB has emerged in recent years, which is resistant to multiple drugs that are commonly used in TB treatment. This new TB is called multidrug-resistant tuberculosis (MDR-TB). At present, rifampicin, isoniazid, pyrazinamide, and ethambutol are used all together to treat tuberculosis. However, as the MDR-TB is found to be resistant to multiple drugs that are used in the treatment of normal TB, other sets of drugs are used to treat the MDR-TB (Sreeramareddy et al., 2008;McIlleron et al., 2006;Ettehad et al., 2012). Rifampicin inhibits bacterial growth by inhibiting the RNA polymerase enzyme. RNA polymerase enzyme is responsible for synthesizing an RNA strand from a DNA strand by the process known as transcription (Figure 01). Ethambutol exerts its effects by inhibiting the transfer of mycolic acids into the cell wall of MTB as well as by changing the lipid metabolism of the bacteria. Pyrazinamide disrupts the membrane energetics and membrane transport. Thus, pyrazinamide shortens TB therapy (Rastogi and David, 1993;Zhang et al., 2003). Isoniazid, a drug used for treating TB, inhibits bacterial growth by inhibiting InhA protein, an enoyl-acyl carrier protein (ACP) reductase. The InhA protein is involved in the type II fatty acid biosynthesis pathway as well as mycolic acid synthesis which is an essential component of the bacterial cell membrane (Ducasse-Cabanot et al., 2004;He et al., 2007). In the mycolic acid synthesis pathway, two types of fatty acid synthase (FAS) enzymes are involved: FAS I and FAS II. The FAS I enzyme generates the starting material of the mycolic acid synthesis, acetyl-CoA. The acyl-CoA is converted to 3-ketoacyl-ACP by Kas III enzyme (betaketoacyl-ACP synthase III). The 3-ketoacyl-ACP then enters into a cyclic reaction catalyzed by the FAS II enzyme. 3-ketoacyl-ACP is converted into 3R-hydroxyacyl-ACP by beta-ketoacyl-5 the two commercially available and mostly used drugs, rifampicin against MTB RNA polymerase and isoniazid against the InhA protein, were used as controls. The mentioned five ligands: 4hydroxybenzaldehyde, benzoic acid, bergapten, psoralen and p-hydroxybenzoic acid, are used to dock against the MTB RNA polymerase and the InhA protein to test their efficacy and potentiality against the enzymes. Later, the two best ligands, each against one enzyme, were determined by analyzing the various tests that are conducted in the experiment and the two ligands were compared with the control to test their efficiency to inhibit TB.   Table 01. Table showing anti-tuberculosis agents with their respective plant sources.

Materials and Methods
Ligand preparation, Grid generation and Glide docking, 2D representations of the best pose interactions between the ligands and their respective receptors were obtained using Maestro-Schrödinger Suite 2018-4 and the 3D representations of the best pose interactions between the ligands and their respective receptors were visualized using Discovery Studio Visualizer (Schrödinger Release 2015-1, 2015Visualizer, 2017). The 2D structures of ligands were downloaded from PubChem in SDF format (www.pubchem.ncbi.nlm.nih.gov) and the two receptors were downloaded from protein data bank (www.rcsb.org).

Protein Preparation
Three-dimensional structure of MTB RNA polymerase (PDB ID: 6M7J) and InhA protein (PDB ID: 2NSD) were downloaded in PDB format from protein data bank (www.rcsb.org). The proteins were then prepared and refined using the Protein Preparation Wizard in Maestro Schrödinger Suite 2018-4 (Sastry et al., 2013). Bond orders were assigned and hydrogens were added to heavy atoms.
Selenomethionines were converted to methionines as well as all the waters were deleted. Finally, the structure was optimized and then minimized using force field OPLS_2005.

Ligand Preparation and Receptor Grid Generation
The 2D  Grid usually confines the active site to shortened specific areas of the receptor protein for the ligand to dock specifically. In Glide, a grid was generated using default Van der Waals radius scaling factor 1.0 and charge cutoff 0.25 which was then subjected to the OPLS_2005 force field.
A cubic box was generated around the active site (reference ligand active site). Then the grid box volume was adjusted to 15×15×15 for the docking test.

Glide Standard Precision (SP) Ligand Docking and MM-GBSA Prediction
SP adaptable glide docking was carried out using Glide in Maestro Schrödinger Suite 2018-4. The Van der Waals radius scaling factor and charge cutoff was set to 0.80 and 0.15 respectively for all 9 the ligand molecules. The ligand with the lowest glide docking score was considered as the best ligand. The 2D and 3D pose interactions between the ligands and receptor were visualized by Maestro Schrödinger Suite 2018-4 and the interaction of the ligand molecule with various types of amino acids as well as their bonds was analyzed by Discovery Studio Visualizer. The molecular mechanics-generalized born and surface area (MM-GBSA) tool was used to determine the ΔGBind scores. The MM-GBSA study was carried out using Maestro-Schrödinger Suite 2018-4.

Ligand Based Drug Likeness Property and ADME/Toxicity Prediction
The molecular structures of every ligand were analyzed using the SWISSADME server  (Cheng et al., 2014;Dong et al., 2018).

Pharmacophore Modelling
The pharmacophore modelling of the 5 ligands was carried out using the Phase pharmacophore perception engine of Maestro-Schrödinger Suite 2018-4. The pharmacophore modelling was done manually. To carry out the process, the radii sizes were kept as the van der Waals radii of receptor atoms, the radii scaling factor was kept at 0.50, receptor atoms whose surfaces are within 2.00 Å of the ligand surface were ignored and the volume shell thickness was limited to 5.00 Å. The 2D and 3D pharmacophore modelling were carried out for all the ligand molecules.

Solubility Prediction
The solubility testing of the five ligands was performed using the QikProp wizard of Maestro-Schrödinger Suite 2018-4. In solubility prediction, the solubility of the selected ligands were determined in various interfaces like hexadecane/gas interface, octanol/gas interface, octanol/water interface etc.

DFT Calculation
Minimized ligand structures obtained from LigPrep were used for DFT calculation using the Jaguar panel of Maestro Schrödinger Suite v11.4 using Becke's three-parameter exchange potential and Lee-Yang-Parr correlation functional (B3LYP) theory with 6-31G* basis set (Lee et al., 1988;Becke, 1988). Quantum chemical properties such as surface properties (MO, density, potential) and Multipole moments were calculated along with HOMO (Highest Occupied Molecular Orbital) and LUMO (Lowest Unoccupied Molecular Orbital) energy. Then the global frontier orbital was analyzed and the hardness (η) and softness (S) of selected molecules were calculated using the following equation as per Parr and Pearson interpretation and Koopmans theorem (Pearson, 1986;Parr et al., 1989). η = (HOMOℇ-LUMOℇ)/2, S = 1/ η

PASS (Prediction of Activity Spectra for Substances) Prediction Study
The PASS (Prediction of Activity Spectra for Substances) prediction was carried out for only the two best-selected ligands that showed the best result in inhibiting their respective receptors, MTB RNA polymerase and InhA protein. PASS prediction was conducted by using the PASS-Way2Drug server (http://www.pharmaexpert.ru/passonline/) by using canonical SMILES from PubChem server (https://pubchem.ncbi.nlm.nih.gov/) (Filimonov et al., 2014). To carry out PASS prediction, Pa (probability "to be active") was kept greater than 70%, since the Pa > 70% threshold gives highly reliable prediction (Geronikaki et al., 1999). In the PASS prediction study, both the possible biological activities and the possible adverse effects of the selected ligands were predicted. The LD50 and Toxicity class were predicted using ProTox-II server (http://tox.charite.de/protox_II/) (Drwal et al., 2014).

Molecular Docking Study and Ramachandran Plot Analysis
All the selected ligand molecules and the controls were docked successfully with their target  Figure 03, Figure 04 and Figure 05 illustrate the 2D and 3D representations of the best interaction between the ligands and receptors as well as the various amino acids that take part in the interaction.

ADME/T Test
The results of the ADME/T test are listed in the molecules was hERG blocker and only bergapten was found to be human hepatotoxic as well as Ames positive. However, bergapten and psoralen showed drug-induced liver injury capability.

P450 Site of Metabolism (SOM) Prediction
The P450 SOM prediction was carried out for the five selected ligand molecules and the SOM

Pharmacophore Modelling
In the pharmacophore modelling experiment, all the ligands generated pharmacophore hypotheses

PASS (Prediction of Activity Spectra for Substances) Prediction Study
In the PASS prediction study, the predicted LD50 value and toxicity class of 4hydroxybenzaldehyde were not determined due to the unavailability of data in the server ProTox II. However, bergapten had the predicted LD50 value of 8100 mg/kg and toxicity class of 6.
However, the PASS prediction study was conducted for 10 intended biological activities and 5 toxic effects. To carry out the PASS prediction experiment, Pa > 0.7 was kept since this threshold give highly reliable prediction (Geronikaki et al., 1999). Both 4-hydroxybenzaldehyde and bergapten showed activities: aldehyde oxidase inhibitor, CYP2A6 substrate, CYP2A substrate, CYP2E1 substrate and CYP1A2 substrate. However, 4-hydroxybenzaldehyde also showed nitrilase inhibitory activity, thioredoxin inhibitory activity and chymosin activity and bergapten also showed activities: HIF1A expression inhibitor and CYP2A11 substrate. The toxic effects showed by 4-hydroxybenzaldehyde were: weakness, vascular toxicity and fatty liver and bergapten showed the toxic effects: hypothermic and carcinogenic group 3. The results of PASS prediction studies are listed in Table 10 and Table 11.

Discussions
Molecular docking generates a score based on the binding of ligand and receptor. The higher binding energy represents the lower bonding affinity and vice versa Sarkar et al., 2020a). Studies have proved that the lowest glide energy corresponds to the best result (Raj and Varadwaj, 2016 should be the best molecules to inhibit their targets. This is further confirmed by the MM-GBSA study. In the MM-GBSA study, the ΔGBind score is taken and the lowest (most negative) ΔGBind Score is always appreciable (Zhang et al., 2017, Ullah et al., 2020aSarkar et al., 2020b).  (Table 02 and Table 03).
Estimation of druglikeness properties aims to improve the drug discovery and development process. Molecular weight and topological polar surface area (TPSA) affect the permeability of the drug molecule through the biological barrier. Higher molecular weight and TPSA reduce the permeability and vice versa. LogP is expressed in the context of lipophilicity. It is described as the logarithm of partition coefficient of the candidate molecule in organic and aqueous phase.
Lipophilicity influences the absorption of the drug molecule inside the body. Higher LogP represents lower absorption and vice versa . LogS value influences the solubility of the candidate molecule and the lowest value is always preferred. Moreover, the more the number of hydrogen bonds, the greater the strength of interaction is and vice versa (Lipinski et al., 1997, 42 Pollastri, 2010Hubbard and Kamran Haider, 2001). Moreover, according to the Ghose filter, a candidate drug should have logP value between -0.4 and 5.6, molecular weight between 160 and 480, molar refractivity between 40 and 130 and the total number of atoms between 20 and 70, to qualify as a successful drug (Ghose et al., 1999). Veber rule describes that the oral bioavailability of a candidate drug depends on two factors: 10 or fewer numbers of rotatable bonds and the polar surface are which should be equal to or less than 140 Å2 (Veber et al., 2002). Furthermore, according to the Egan rule, the absorption of a candidate drug molecule depends on two factors: the polar surface area (PSA) and AlogP98 (the logarithm of partition co-efficient between noctanol and water) (Egan et al., 2000). And according to the Muegge rule, for a drug like chemical compound to become a successful drug, it has to pass a pharmacophore point filter, which was developed by the scientists (Muegge et al., 2001). According to the druglikensess property experiment, p-hydroxybenzoic acid should be considered as the best molecule since it had quite low molecular weight (138.12 g/mol), the lowest LogP value of 1.05, the highest druglikeness 43 ADME/T tests evaluate the pharmacological and pharmacodynamic properties of a candidate drug molecule within biological system. Blood brain barrier is very important for those drugs that target primarily the brain cells. Since, the oral delivery system is the most commonly used route of drug administration, therefore, it is expected that the drug is highly absorbed in intestinal tissue. Pglycoprotein protein in the cell membrane facilitates the transport of many drugs. Therefore, its inhibition affects the drug transport. In vitro study of drug permeability test utilizes Caco-2 cell line and its permeability reflects that the drug is easily absorbed in the intestine. Orally absorbed drugs travel through the blood circulation and deposit back to liver. In the liver, they are metabolized by group of enzymes of cytochrome P450 family and excreted as bile or urine.
Therefore, inhibition of any of the enzymes of this family might affect biodegradation of the drug molecule (Li, 2001;Guengerich, 1999;Sarkar et al., 2020d).  Swierczewska et al., 2015;Smalling, 1996;Sarkar et al., 2020c). HERG is a protein in the heart muscle which mediates the rhythm of the heart. HERG can be blocked by many blocking agents. This may lead to the cardiac arrhythmia and sometimes death. Human liver is the primary site of metabolism and it is extremely vulnerable to the harmful effects of various xenobiotic agents. Human hepatotoxicity (H-HT) involves any type of injury to the liver that may lead to organ failure and even death. Ames test is a mutagenicity assay that is 44 used to detect the mutagenic chemicals. The mutagenic chemicals can cause mutations and also capable of cancer development. Drug induced liver injury (DILI) is the injury to the liver that are caused by administration of drugs. DILI is one of the reasons that may lead to various liver problems (Sanguinetti et al., 1995;Aronov, 2005;Cheng and Dixon, 2003;Mortelmans and Zeiger, 2000;Holt and Ju, 2006).
In the absorption section, all the ligands performed quite similarly, however, based on the probability values, it can be concluded that 4-hydroxybenzaldehyde and benzoic acid are the best performers of the absorption section. In the distribution section, benzoic acid, bergapten and psoralen showed high plasma protein binding capability and all of the ligands were blood brain barrier permeable. Psoralen should be considered as the best performer in the distribution section, based on the probability values. No ligand showed satisfactory results in the metabolism section.
However, 4-hydroxybenzaldehyde, benzoic acid and p-hydroxybenzoic acid showed relatively good results since they were not inhibitory to any of the CYP450 isoenzymes. Benzoic acid could be considered as the best ligand in the metabolism section with good probability values. In the excretion section, 4-hydroxybenzaldehyde is the best ligand with the highest half-life of 1.7 hours.
In the toxicity section, both 4-hydroxybenzaldehyde, benzoic acid, p-hydroxybenzoic acid showed the best performances since they were not hERG blockers, human hepatotoxic, Ames mutagenic as well as they were also DILI negative. Bergapten was, however, human hepatotoxic, Ames mutagenic and DILI positive and psoralen was only DILI positive.
The Cytochrome P450 (Cyp450) is a family of enzymes and comprises 57 isoforms of P450 enzymes. These enzymes catalyze the phase-I metabolism of almost 90% of the marketed drugs and are heme-containing (Glue and Clement, 1999;Tyzack et al., 2014). The functions of these enzymes are to catalyze the conversion of lipophilic drugs to more polar compounds (Danielson, 45 2002 for QPlogPw is 4.0 -45.0, the acceptable range for QPlogPo/w is -2.0 -6.5, the acceptable range for QPlogS is -6.5 -0.5, the acceptable range for CIQPlogS is -6.5 -0.5 (Hussain and Verma, 2018 be concluded that bergapten showed the best results in the solubility test, with its best scores among all the selected ligand molecules (Table 08).
Frontier orbitals study is an essential method of understanding the pharmacological properties of various small molecules (Matysiak, 2007). HOMO and LUMO are the globally studied orbitals that help to understand the chemical reactivity and kinetic stability of small molecules. The term 'HOMO' indicates the regions on a small molecule that may donate electrons during a complex formation and the term 'LUMO' indicates the regions on a small molecule that may receive electrons from the electron donating species. The difference between HOMO and LUMO energy is known as gap energy. Gap energy corresponds to the electronic excitation energy. The compound that has the greater orbital gap energy, tends to be energetically unfavourable to undergo a chemical reaction and vice versa (Zhan et al., 2003;Hoque et al., 2015). Moreover, the HOMO-LUMO gap also correlates with the hardness and softness of a molecule (Ayers et al., 2006 hydroxybenzaldehyde had the score that were much lower than the scores of rifampicin. On the other hand, Bergapten was selected as the best ligand to inhibit the InhA protein and it was also far superior than the InhA protein inhibitor or control, isoniazid. Isoniazid had docking score of -6.018 Kca/mol and ΔGBind score of -25.120 Kcal/mol, whereas, bergapten had much lower scores of -8.068 Kcal/mol and -57.590 Kcal/mol, respectively. For this reason, it can be concluded that, both 4-hydroxybenzaldehyde and bergapten had very performance and good efficiency to inhibit TB, when compared to the widely used drugs that are used to treat TB. The PASS prediction was study was conducted on only these two best ligands to determine their various biological and toxicological effects. ProTox-II server evaluates the toxicity of a chemical compound and classifies the compound into a toxicity class ranging from 1 to 6. The server  Secretariat, 2005). And ProTox-II server adds one more class to the 5 classes, making them 6 classes in total, class VI: non-toxic (LD50 > 5000) (http://tox.charite.de/protox_II/index.php?site=home. Accessed on: 09, August, 2019). The predicted LD50 value of bergapten was 8100 mg/kg and the toxicity class was 6. For this reason bergapten is non-toxic. The PASS prediction study was carried out for 10 intended biological activities and 5 toxic effects. Both 4-hydroxybenzaldehyde and bergapten showed good biological activities like aldehyde oxidase inhibitor, CYP2A6 substrate, CYP2A substrate, CYP2E1 substrate and CYP1A2 substrate. However, 4-hydroxybenzaldehyde also showed nitrilase inhibitory activity, thioredoxin inhibitory activity and chymosin activity and bergapten also showed activities: HIF1A expression inhibitor and CYP2A11 substrate. The toxic effects showed by 4hydroxybenzaldehyde were: weakness, vascular toxicity and fatty liver and bergapten showed the toxic effects: hypothermic and carcinogenic group 3. These toxic effects may interfere with the successful approval and marketing of the drugs. The PASS prediction study had confirmed the superiority of the two best ligands in the toxicity and adverse effects section. For this reason, it can be declared that, both the two selected agents showed satisfactory performances in the tests when compared to the controls.

Conclusion
5 agents known to have potential anti-tubercular properties were used to analyse in the experiment.
Considering all the parameters, it is clear that, all the plant derived anti-tubercular agents had very good inhibitory activities on the MTB. The various tests of in silico biology, that were used in the experiment, like the molecular docking study, druglikeness property experiment, ADME/T test, pharmacological property analysis, solubility and DFT calculations as well as the PASS prediction study had confirmed that 4-hydroxybenzaldehyde and bergapten were best agents among the selected ligands as well as their superiority over the two commercial, widely used drugs, rifampicin and isoniazid. For this reason, these two agents can be used effectively to fight against tuberculosis.
4-hydroxybenzaldehyde can be acquired from a variety of sources form the nature, like the plant Cinnamomum kotoense and bergapten can be acquired from the plant Fatoua pilosa. For this reason, these plants can be used effectively to treat tuberculosis. Moreover, in nature, a lot of other 50 plants can also be found containing these agents. However, more in vivo and in vitro researches should be carried out to finally confirm their activities. Moreover, more researches should be conducted on the other agents to identify their efficacy against TB since they also gave quite good results in the tests carried out in the experiment. Hopefully, this study will help the researchers in identifying the potential anti-tubercular phytochemicals.