Taxifolin as dual inhibitor of Mtb DNA gyrase and isoleucyl-tRNA synthetase: in silico molecular docking, dynamics simulation and in vitro assays

Abstract

DNA gyrase and aminoacyl-tRNA synthetases are two essential bacterial enzymes involved in DNA replication, transcription and translation. Flavonoids are plant secondary metabolites with variable phenolic structures. In this study, eight flavonoids structurally similar to quercetin were selected and their ADMET properties were evaluated. Molecular docking and free energy calculations were carried out to examine the binding of these flavonoids to the ATP-binding site and editing domain of DNA gyrase and Isoleucyl-tRNA synthetase, respectively. Taxifolin was found out to be the top lead molecule in both the docking studies with a good number of interactions with the active site amino acids. Further, binding of taxifolin to the proteins was extensively studied using 50 ns molecular dynamics simulation. In vitro anti-tuberculosis activity of taxifolin was evaluated and compared with the standard drugs. Minimal inhibition concentration of taxifolin was found to be ≤ 12.5 μg/ml.

Electronic supplementary material

The online version of this article (10.1007/s40203-018-0045-5) contains supplementary material, which is available to authorized users.

Introduction

According to a report by World Health Organization (WHO), tuberculosis (TB) was one of the top ten causes of death worldwide in 2015. Over 95% of TB deaths occur in low- and middle-income countries, with India leading the count, followed by Indonesia, China, Nigeria, Pakistan and South Africa (Floyd et al. 2018).

Tuberculosis is caused by Mycobacterium tuberculosis, complex bacilli (Mtb) that most often affects the lungs. The first-line anti-TB drugs usually include rifampicin (RIF), isoniazid (INH), pyrazinamide (PZA) and ethambutol (EMB), with rifampicin and isoniazid are extended in combination until the patient is fully recovered. This drug combination therapy prevents the bacteria from developing drug resistance and has been in widespread use for over 20 years. RIF, discovered in 1959, was the last consequential compound developed specifically for the treatment of TB and marketed in 1967. Even though the current drug regimen is 95% effective in usual cases, has number of limitations (Koul et al. 2011). These include the side effects of individual drugs, long duration of therapy required for cure and mostly, the development of multidrug-resistant TB (MDR-TB) and extensively drug-resistant TB (XDR-TB) (Dawson and Diacon 2013). In this scenario, there is an urgent need to develop drugs with greater effectiveness, less side effects and resistance.

DNA gyrase, a type II topoisomerase present only in bacteria and plants has been an attractive target for developing antibacterial drugs. They introduce ATP-dependent topological changes to DNA and are crucial for bacterial survival. DNA gyrase consists of A and B subunits, which integrate to form an A₂B₂ complex in the active enzyme. The A subunit interacts with DNA and accommodate the active-site tyrosine responsible for DNA cleavage. The B subunit carries out hydrolysis of ATP that supply the energy required to perform enzyme function. DNA gyrase inhibitors like coumarins and quinolones are widely used for the treatment of bacterial infectious diseases (e.g., ciprofloxacin) (Karkare et al. 2013; Collin et al. 2011).

All types of cells require aminoacyl-tRNA synthetases (aaRSs) for protein synthesis. These enzymes connect the amino acid onto its cognate transfer RNAs (tRNAs) through the aminoacylation of tRNAs. Active site composition of a specific synthetase differs enough for a drug to differentiate a bacterial synthetase from its human counterpart. These features make AARSs attractive targets for inhibitors that act as anti-infective drugs (Sassanfar et al. 1996).

Flavonoids are a class of pigmentary compounds (secondary metabolite) widely distributed in plants, and are regularly consumed by humans. Flavonoids are hydroxylated polyphenolic compounds produced by plants to defend various microbial infections (Kumar and Pandey 2013). They are also a constituent of many of the traditional herbal medicines and have a variety of other applications (Plaper et al. 2003). Flavonoids have attracted extensive attention as a promising therapeutic compound due to its ever expanding bioactivities (Takekoshi et al. 2014). Out of the 14 structurally diverse groups of flavonoid, six are well known and characterized; the flavones, isoflavones, flavanones, flavonols, flavanols (catechins), and anthocyanidines (Hendrich 2006). Flavonoids known to have various beneficial properties like antibacterial, antiprotozoal, anti-inflammatory, dietary antioxidant, vascular and oestrogenic activities, primarily achieved via inhibition of oxidases and NADH usage (Brown et al. 2007).

In our attempt to identify novel anti-TB agent, we virtually screened a group of flavonoids against the crystal structure of DNA gyrase and Isoleucyl-tRNA synthetase (IleRS). Further, binding property of the topmost screened compound to both the enzymes was validated using molecular dynamic simulation. Anti-TB activity of the compound was confirmed by cell viability assay, raising the possibility of developing flavonoids as potential anti-TB therapeutic.

Methodology

Protein preparation

The X-ray crystal structure of the DNA gyrase in complex with small molecule inhibitor (PDB ID: 4DUH) and IleRS editing domain complexed with the pre-transfer editing substrate analogue (PDB ID: 1WK8) were obtained from the protein data bank and prepared for docking using protein preparation wizard of Schrödinger suite (Schrödinger, LLC, New York, NY). The protein crystalline structure obtained from Protein Data Bank (PDB) usually needs preprocessing due to the presence of an undesirable co-crystallized ligand, missing side chains, water molecules, metal ions, and cofactors. Some of the structure may have more than one subunit or repeated subunits. Missing information on connectivity, bond orders and formal charges are some of the other issues. The protein preparation wizard present in the Maestro suite can be used to solve most of the above mentioned problems automatically. The protein preparation wizard panel has three tabs, import and process, review and modify, then refine. Basic protein fixing tasks like bond order assignment, hydrogen bond addition and deletion of unwanted water molecule were performed using import and process tab. The review and modify tab was used to delete unwanted chains and fix or delete het groups. Finally, refine tab was employed to optimize orientations of hydrogen-bonded groups and minimize the structure with OPLS_2005 force field.

Grid generation at the active site

Enzymes are proteins that catalyze reactions converting substrates to products. The enzymatic reactions take place in the enzymes’ active sites (Pravda et al. 2014). Identification of the active site is a crucial step in structure-based drug design. Receptor grid generation option in the Glide was used for grid generation. The active site region of prepared proteins was identified and a grid box enclosing the entire active site was generated using the co-crystallized ligand. All the other options were kept default.

Ligand preparation

Quercetin, one of the most abundant natural flavonoids, can inhibit supercoiling activity of bacterial gyrase by binding to the active site of B subunit (Plaper et al. 2003). Eight flavonoids structurally similar to Quercetin were used as ligands in this study. The input ligands were prepared using Ligprep module of Schrödinger suite (Schrödinger, LLC, New York, NY). Each of the successfully processed ligand structure will be a three dimensional structure with proper chirality and low-energy. All the options were kept default. The processed ligands were subjected to extra precision (XP) docking using the Glide module of Schrödinger suite.

In silico ADMET prediction

Even though a chemical is highly active and has specific action, it cannot become a drug unless it is absorbed effectively into the body, distributed specifically to different regions of the body, retained in the body for a specific time without losing the activity and gets eliminated properly (Hodgson 2001). Hence, evaluating the absorption, distribution, metabolism, excretion, and toxicity (ADMET) properties of a compound is very essential in assessing its pharmacodynamic activities and drug likeness. A computational ADME study of selected flavonoids was carried out using QikProp (Schrödinger, LLC, New York, NY). In the present study, properties like molecular weight, predicted central nervous system activity, octanol/water partition coefficient, aqueous solubility, IC₅₀ value for blockage of HERG K+ channels, cell permeability, binding of a drug to human serum albumin, number of hydrogen bond donors and acceptors were calculated for each compound and drug-likeness was evaluated. Moreover, the properties of ligand with respect to toxicity and carcinogenicity were analyzed online using admetSAR (http://lmmd.ecust.edu.cn/admetsar1) tool.

Molecular docking

Glide (grid-based ligand docking with energetics) (Schrödinger, LLC, New York, NY) searches for possible conformation of the ligand in the active-site region of the receptor using a set of filters (Friesner et al. 2004). Extra precision (XP) glide methodology semi-quantitatively ranks candidate ligands according to their capability to bind to a particular conformation of the protein receptor (Friesner et al. 2006). The prepared ligand set was docked to the protein using the XP mode of Schrödinger’s glide software with default settings. The hydrophobic interactions and H-bonds of the best docked complex were identified using the Ligplot program (Laskowski and Swindells 2011).

Prime MM/GBSA

The ligand strain energies and ligand binding energies were calculated according to molecular mechanics combined with the Generalized Born Model and Solvent Accessibility method (MM/GBSA), using Prime module of Schrödinger for all the docked receptor-ligand complexes.

The binding free energy (ΔG_bind) was calculated using the following equation (Lyne et al. 2006):

$$\mathrm{\Delta }{\text{G}}_{\text{bind}}=\mathrm{\Delta }{\text{E}}_{\text{MM}}+\mathrm{\Delta }{\text{G}}_{\text{SOL}}+\mathrm{\Delta }{\text{G}}_{\text{SA}}$$

where ΔE_MM is the difference in energies between the protein–ligand complex and the sum of the energies of the unliganded protein and ligand, using the OPLS force field. ΔG_SOL is the difference in the GBSA solvation energy of the protein–inhibitor complex and the sum of the solvation energies for the unliganded protein and ligand. ΔG_SA is the difference in surface area energies for the complex structure and the sum of the surface area energies for the unliganded protein and ligand.

Molecular dynamic simulation

Molecular dynamic (MD) simulation approach can precisely characterize interactions between protein and ligand at atomic resolution. MD simulations have been a consequential tool in examining structural information in a variety of areas, where experimental data by themselves are inadequate. However, the time taken to run such calculations makes its application limited for multiple molecules (Rohini and Shanthi 2017). Here in this study, MD simulation was used to determine the binding stability of taxifolin with DNA gyrase and IleRS. A 50 ns MD simulation was carried out for each system with the protein–ligand complex in a cubic box containing water molecules and counter ions using Gromacs v5.1.1 as described previously (Fayaz and Rajanikant 2014). The root mean square deviation (RMSD) value of the protein backbone and ligand atoms was calculated to evaluate the overall stability of the complex throughout the simulation. Also, radius of gyration (Rg), root mean square fluctuation (RMSF), number of hydrogen bonds and distance between protein and ligand were checked for further validation of protein–ligand stability.

Anti-tuberculosis activity test

Anti-TB activity of screened compound was assessed against M. tuberculosis H₃₇Rv strain using the microplate Alamar Blue assay (MABA) as described previously (Cho et al. 2015). The final drug concentrations of the ligand tested were 0.2–100 µg/ml. The minimal inhibition concentration (MIC) of the ligand was found out and compared with the MIC of standard anti-TB drugs.

Results

Combined computational procedures, including molecular docking and MD simulations, and in vitro biological assay were employed to identify potential lead from the selected flavonoids for DNA gyrase and IleRS inhibition.

The studied compounds

Eight flavonoids structurally similar to quercetin were selected in the study (Table 1). No new ligand poses were generated during the ligand preparation.

Table 1.

Details of selected eight flavonoids

ADMET prediction

At first, selected flavonoids were subjected to in silico ADMET analysis for finding their drug-likeness (Table 2, 3). Predicted central nervous system (CNS) activity of all the compounds was − 2 (inactive) except 4′-7-isoflavandiol (Equol), which had a neutral value. Molecular weight (MW) of all the molecules was in the acceptable range of 130.0–725.0. Number of hydrogen bond donors (donorHB) and the hydrogen bond acceptors (accptHB) were also in the recommended span of 0.0–6.0 and 2.0–20.0, respectively. Predicted octanol/water partition coefficient (QPlogPo/w) was in agreement with the recommended range of − 2.0 to 6.5. All the compounds had predicted aqueous solubility (QPlogS) value and predicted IC₅₀ value for blockage of HERG K+ channels (QPlogHERG) in the acceptable range of − 6.5 to 0.5 and above − 5, respectively, except Sigmoidin A. Predicted apparent Caco-2 cell permeability (QPPCaco) value is related to a drug’s metabolism and its permeability to the biological membrane. This was within the acceptable span of 25–500 for all compounds, except 4′-7-isoflavandiol (Equol) and taxifolin. Prediction of binding of a drug to human serum albumin (QPkhsa) and percentage of human oral absorption was in the acceptable range for all the compounds. All the eight molecules satisfied drug-like properties based on Lipinski’s rule of five and other parameters (Table 2). All flavonoids considered for this study were found out to be non-Ames toxic and Non-carcinogens. Other toxicity related parameters like acute oral toxicity, rat acute toxicity and hERG inhibition were in favor of the molecules to be screened as drugs (Table 3).

Table 2.

ADME properties of selected eight flavonoids

Ligand name	CNS activitya	MW	donorHB	accptHB	QPlogPo/w	QPlogS	QPlogHERG	QPPCaco	QPlogKhsab	Human oral absorptionc (%)
Afzelehin	− 2	274.273	4	4.7	1.149	− 2.857	− 4.837	147.741	− 0.262	72.505
Aromadendrin	− 2	288.256	3	5.7	0.775	− 2.894	− 4.89	64.077	− 0.292	63.821
4′-Hydroxywogonin	− 2	300.267	2	4.5	1.667	− 3.224	− 4.643	128.902	− 0.039	74.477
4′-7-Isoflavandiol (Equol)	0	242.274	2	2.25	2.761	− 3.578	− 4.863	910.081	0.171	96.073
Licoflavone C	− 2	338.359	2	3.75	3.067	− 4.637	− 4.927	145.05	0.485	83.592
3-O-Methylquercetin	− 2	316.267	3	5.25	1.135	− 3.23	− 4.902	52.046	− 0.165	64.309
Sigmoidin A	− 2	424.493	3	4.75	4.303	− 7.028	− 5.928	92.205	0.941	87.307
Taxifolin	− 2	304.256	4	6.45	0.126	− 2.694	− 4.802	23.126	− 0.422	52.097

^aPredicted central nervous system activity on a − 2 (inactive) to + 2 (active) scale

^bPrediction of binding to human serum albumin (acceptable range − 1.5 to 1.5)

^cPredicted human oral absorption on 0–100% scale (> 80% is high and < 25% is poor)

Table 3.

Toxicity profile of selected eight flavonoids

Ligand	Ames toxicity	Carcinogens	Acute oral toxicity	Rat acute toxicity	hERG inhibitor 1, hERG inhibitor 2
Afzelehin	Non AMES toxic	Non-carcinogens	IV	2.0532 LD50, mol/kg	Weak inhibitor, non-inhibitor
Aromadendrin	Non AMES toxic	Non-carcinogens	II	3.0825 LD50, mol/kg	Weak inhibitor, non-inhibitor
4′-Hydroxywogonin	Non AMES toxic	Non-carcinogens	III	2.7192 LD50, mol/kg	Weak inhibitor, non-inhibitor
4′-7-Isoflavandiol (Equol)	Non AMES toxic	Non-carcinogens	III	2.7268 LD50, mol/kg	Weak inhibitor, non-inhibitor
Licoflavone C	Non AMES toxic	Non-carcinogens	III	2.9772 LD50, mol/kg	Weak inhibitor, non-inhibitor
3-O-Methylquercetin	Non AMES toxic	Non-carcinogens	III	2.6388 LD50, mol/kg	Weak inhibitor, non-inhibitor
Sigmoidin A	Non AMES toxic	Non-carcinogens	III	3.2421 LD50, mol/kg	Weak inhibitor, non-inhibitor
Taxifolin	Non AMES toxic	Non-carcinogens	II	3.0200 LD50, mol/kg	Weak inhibitor, non-inhibitor

Molecular docking study

To understand the structural basis of the selected ligand’s binding to DNA gyrase and IleRS, each prepared ligand was individually docked with the receptors at their active site using Glide module. The DNA gyrase structure obtained from PDB (ID: 4DUH) had a co-crystallized 4-{[4′-methyl-2′-(propanoylamino)-4,5′-bi-1,3-thiazol-2-yl]amino}benzoic acid molecule that was used to position the grid at the ATP-binding site in B subunit. Some of the key information obtained from XP pose viewer file includes Glide score, Glide energy and Prime MM/GBSA ΔG bind. Docking results were further analyzed and ranked according to the ascending order of their Glide score (Table 4). In the case of DNA gyrase, all the docked flavonoids have their Glide score in the range of − 8.22 to − 4.00 and Glide energy − 46.09 to − 31.26 kcal/mol. taxifolin was at the top of the list with a best Glide score of − 8.22.

Table 4.

Docking score, glide energy and binding free energy of eight flavonoids

Ligand name	DNA gyrase			IARS
	Docking score	Glide energy (kcal/mol)	ΔG bind (kcal/mol)	Docking score	Glide energy (kcal/mol)	ΔG bind (kcal/mol)
Taxifolin	− 8.22	− 39.21	− 27.41	− 10.79	− 41.79	− 29.69
3-O-Methylquercetin	− 7.54	− 46.09	− 27.48	− 7.84	− 41.16	− 27.54
4′-Hydroxywogonin	− 7.81	− 42.10	− 29.89	− 7.02	− 31.12	− 14.73
Aromadendrin	− 7.75	− 33.91	− 23.44	− 6.26	− 40.03	− 33.58
Licoflavone C	− 7.46	− 44.00	− 36.65	− 4.75	− 40.09	− 27.65
Afzelehin	− 5.72	− 33.96	− 19.18	− 5.51	− 35.59	− 30.58
Sigmoidin A	− 4.00	− 31.26	− 23.62	− 5.23	− 43.63	− 25.29
4′-7-Isoflavandiol (Equol)	− 5.60	− 34.74	− 23.86	− 1.00	− 29.50	− 25.03

Spatial conformation and key interactions of taxifolin in the active site of DNA gyrase are schematically presented in Fig. 1a, c respectively.

Fig. 1.

Binding site residues and the docking poses of taxifolin at the ATP-binding site of a DNA gyrase and editing site of b IleRS. Inter-molecular interactions of c DNA gyrase and d IleRS with taxifolin

Taxifolin made hydrogen bond with Gly77 and hydrophobic contacts were made with amino acids Asn46, Arg76, Ile78, Pro79, Gly101, Lys103 and Thr165 (Fig. 1c).

Co-crystallized 5′o-[n-(l-valyl)sulphamoyl] adenosine molecule present at the editing site of IleRS was used to position the grid at the active site. Information received from the XP pose viewer file were analyzed and ligands were listed according to the ascending order of their Glide score (Table 4). Highest and lowest docking scores of ligands were − 10.79 and − 1.00, respectively, when docked with IleRS (PDB ID: 1WK8). Molecular docking with IleRS also demonstrated taxifolin as the top ligand hit with a Glide score of − 10.79. Thus, taxifolin was the top common hit molecule in both the docking studies and has the potential to bind to two key proteins essential for the survival of M. tuberculosis. Thus, taxifolin was taken for further in silico and in vitro studies.

IleRS-taxifolin interactions revealed that amino acid Trp227, Thr228, Asp313 and Val318 are involved in the hydrophobic interactions with taxifolin. Simultaneously, taxifolin made hydrogen bonds with Thr229, Thr230, Ser 310, Thr315, Gly316, Ile317 and His 319 (Fig. 1d). Spatial conformation of taxifolin in the active site of IleRS is depicted in Fig. 1b. Binding mode of rest of the selected flavonoids with DNA gyrase and IleRS is shown in the online resource 1 and 2, respectively.

The strength by which a ligand binds to a biomolecule can be determined from binding affinity values. Relative binding affinity for the selected ligands to the receptor was estimated using the MM/GBSA ΔG bind values. Binding affinity of selected ligands for DNA gyrase and IleRS is given in the Table 4.

MD simulation

Molecular dynamic simulation technique has favorably been adopted to analyze complex molecular behaviors like molecular motions and function, modeling enzyme mechanisms and complex biomolecular assemblies at the atomic level (Dodson et al. 2008).

The root mean square deviation value is an indication of the conformational integrity of a structure in the course of the simulation. The system containing DNA gyrase and taxifolin was well converged and the mean RMSD values for protein backbone and ligand atoms were ~ 0.35 and ~ 0.06 nm, respectively (Fig. 2a, b). Relatively steady RMSD values of ligand deduced that, taxifolin was stable at the active site. The minimum distance between the atoms of protein and ligand remained constant at a value of ~ 0.23 nm throughout the 50 ns simulation (Fig. 2f). Even though there were slight changes, RMSF and Rg values were relatively stable at ~ 0.2 and ~ 1.64 nm (Fig. 2d, e). Further, the number of H-bonds formed between the atoms of DNA gyrase and taxifolin also showed the stability of the protein–ligand interaction (Fig. 2c).

Fig. 2.

RMSD of backbone atoms of a DNA gyrase and b taxifolin compared to the initial position. c Number of H-bonds formed between DNA gyrase and taxifolin. d RMSF plot of DNA gyrase—taxifolin. e Radius of gyration as a function of time for DNA gyrase. f Minimum distance calculated between DNA gyrase and taxifolin

Taxifolin remained stable at the active site of IleRS showing a similar affinity as in the case of DNA gyrase (Fig. 3b). In the initial stage of simulation, RMSD value for protein backbone atoms was fluctuating; from 0 to 10 ns and then it remained constant at a value of ~ 0.5 nm (Fig. 3a). RMSD value of ligand atoms was stable at a value of ~ 0.05 in the beginning of simulation but showed slight variation afterwards (Fig. 3b). Minimum distance between protein and ligand atoms, RMSF and Rg values were stable at ~ 0.2, ~ 1.63 and ~ 0.2 nm (Fig. 3f, d, e), respectively. Number of H-bond formed between the ligand and protein was also satisfactory to prove a strong binding between them (Fig. 3c).

Fig. 3.

RMSD of backbone atoms of a IleRS and b taxifolin compared to the initial position. c Number of H-bonds formed between IleRS and taxifolin. d RMSF plot of IleRS—taxifolin. e Radius of gyration as a function of time for IleRS. f Minimum distance calculated between IleRS and taxifolin

Anti-tubercular assay

Anti-TB potential of taxifolin was evaluated in vitro against the M. tuberculosis H₃₇ Rv strain using a microplate Alamar blue assay (MABA) with Pyrazinamide, Streptomycin and Ciprofloxacin as positive controls (Fig. 4). Lowest drug concentration, which interrupted bacterial growth and prevented the color change from blue/black to pink, was taken as MIC. MIC of standard anti-TB drugs, Pyrazinamide, Ciprofloxacin and Streptomycin was ≤ 3.125, 3.125 and 6.25 μg/ml. MIC of taxifolin was found to be ≤ 12.5 μg/ml and which is promising (Table 5).

Fig. 4.

Microplate Alamar blue assay showing MIC of pyrazinamide, ciprofloxacin, streptomycin and taxifolin. Blue or black color—sensitive; pink—resistant

Table 5.

MIC value of taxifolin compared with standard anti-tuberculosis drugs

Compound name	MIC (μg/ml)
Pyrazinamide	≤ 3.125
Ciprofloxacin	≤ 3.125
Streptomycin	≤ 6.25
Taxifolin	≤ 12.5

Discussion

In the present study, dual-target pharmacological strategy was employed to explore the potential of flavonoids as anti-TB agents. Binding of a set of flavonoids to two of the most essential bacterial proteins, DNA gyrase and IleRS was studied using in silico methods and anti-tubercular activity of the compound (taxifolin) was confirmed in vitro.

In the past 20 years, there have been significant advances in the effort to contain TB, but still global elimination of the disease has not been achieved yet. Current treatment regimens proposed by WHO for drug-sensitive and drug-resistant TB have certain basic drawbacks like long period of treatment—lasting at least 6 months (Zumla et al. 2014). The major reason for the continued prevalence of TB is the spread of multidrug-resistance TB (MDR-TB), mostly derived from the use of improper dosages and regimens, insufficiency of quality-assured pharmaceutical products, and little efforts to support patient’s adherence (Raviglione and Sulis 2016). Apart from multidrug-resistance TB, extensively drug-resistant TB (XDR-TB), which is resistant to at least four of the core anti-TB drugs, has emerged as another threat to TB management. Inherent problems like these have made new anti-TB drugs and treatment regimens a clinical and public health priority.

There are two ways to tackle this scenario; firstly, novel drug targets can be determined for the formulation of new drugs with no preexisting resistance mechanisms. Secondly, new drugs or modification of existing drug classes can be made for known and clinically validated targets to exclude the chance of cross resistance with the available drugs for which resistances have elicited. The first method has disadvantages like low hit rates from high throughput screens and difficulty in examining novel targets without inhibitors that can be tested in infection models. In this study, we adopted the second approach as it has better success rate as it aims validated molecular targets whose inhibition is known to cause prevention of bacterial growth (Mdluli and Ma 2007).

Many research groups have shown the antibacterial activity of flavonoids either as crude extracts from plants or purified compound. Researchers have also reported the effectiveness of using naturally occurring flavonoids in combination with other antibacterial agents against resistant strains of bacteria (Cushnie and Lamb 2005). Zheng et al. (2014) demonstrated that flavonoid structure is a suitable template for the design of effective M. tuberculosis (Mtb) proteasome inhibitors. In our study, eight flavonoids structurally similar to quercetin were selected and used for the in silico studies. All the eight compounds were satisfying the ADMET parameters and they had comparable drug-likeness.

Mycobacterium tuberculosis DNA gyrase is a proven target for anti-tubercular drug discovery as its inhibition results in considerable mycobactericidal activity. Inhibitors of DNA gyrase can be used for reducing the duration of TB therapy as they are effective against non-replicating, persistent mycobacteria. A novel inhibitor of DNA gyrase can be effectively used against multidrug and fluoroquinolone resistant TB (Mdluli and Ma 2007). Binding of quercetin to the ATP binding site of DNA gyrase B subunit had been characterized by Plaper et al. (2003) and quercetin made bonds with amino acids Asn46, Gly 77, Thr165, Arg76, Ile78 and Pro79. We have also observed similar binding of taxifolin with the B subunit of DNA gyrase. The binding of taxifolin to the B subunit of DNA gyrase was also analogous to the binding of compound 18 in the study by Brvar et al. Both taxifolin and compound 18 made similar interactions with Gly77 and Thr165 and bound to the active site in an identical manner (Brvar et al. 2012).

As in the case of DNA gyrase, aminoacyl-tRNA synthetase is also extensively studied for its potential as target for anti-TB drugs. In prokaryotes, this enzyme is majorly conserved, making it an appropriate target for the production of broad-spectrum antibiotics. Apart from FDA approved drug Mupirocin, there are other IleRS inhibitors like SB-234764, CB-432 and Icofungipen, which are in the experimental stage (Vondenhoff and Van Aerschot 2011). Taxifolin made interactions with residues Trp227, Thr228, Thr229, Thr230, Ser310, Asp313 and His319 of the editing domain of IleRS. Similar interactions were observed between editing domain of IleRS and a substrate analogue in the work by Fukunaga and Yokoyama (Fukunaga and Yokoyama 2006).

Further, the binding of taxifolin to DNA gyrase and IleRS was analyzed at atomic scale using MD simulation studies. To ascertain the stability of the docking conformation, it has to be supported by a relatively lengthy MD simulation study (Maharaj and Soliman 2013). The RMSD for protein backbone atoms showed that in both the simulations, systems reached equilibrium before 50 ns. Even though the RMSF values of both the systems were relatively stable, there was a slight deviation in the RMSF values of some of the atoms (atom number ~ 650–800) in DNA gyrase (Fig. 2d). In agreement with the docking results, all the other parameters studied during the course of MD simulation were relatively stable and substantiated our hypothesis.

In this present study, in vitro study was carried out with H₃₇Rv strain, which is the most common M. tuberculosis virulent strain used in the clinical and research laboratory setting. Even though the MIC of taxifolin was slightly higher than the MIC of standard anti-TB drugs, it had comparable MIC with the other experimental drugs (Tripathi et al. 2006; Veau et al. 2016). Further studies are warranted for the in vivo evaluation of the efficacy of taxifolin as anti-TB drug. On the light of current demand for a better medication for TB, multi-target pharmacological strategy has immense potential to be successful. These multi target drugs will have better efficacy with less possibility of invoking drug resistance.

Conclusion

Molecular docking and binding energy studies of eight flavonoids with DNA gyrase and IleRS show that taxifolin can bind to these essential proteins alike. This was further validated with the help of MD simulation studies. In vitro anti-TB activity of taxifolin was tested and found out to have comparable MIC with the standard anti-TB drugs. To our knowledge, this is the first study aims to target two essential bacterial proteins with a flavonoid with the intention of developing an anti-TB drug.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Funding

None.

Compliance with ethical standards

Conflict of interest

The authors declare that there is no conflict of interest.