Comparative Analysis of Different DNA-Binding Drugs for Leishmaniasis Cure: A Pharmacoinformatics Approach

Authors


Corresponding author: Raju Poddar, rpoddar@bitmesra.ac.in

Abstract

Several experiments have been performed to test DNA-binding drugs to cure Leishmania infection. However, there are no details of pharmacoinformatics study. Herein, we have selected a good number of compounds from experimentally verified studies and performed a comparative analysis based on pharmacoinformatics techniques. In silico docking study was performed to observe the molecular level interactions of these known ligands with the DNA receptor by automated computational docking using Glide. A comparison between the calculated interaction energies and in silico ADME/T study was made. In agreement with drug likeness rules, our study suggests that seco-hydroxy-aza-CBI-TMI (compound 4b; GScore, −12.058) is a potential molecule for targeting the DNA to cure leishmaniasis.

Parasitic infections caused by protozoan parasites remain a major public health concern affecting the lives of billions of people worldwidea (1). Among these diseases, leishmaniasis, the vector-borne parasitic diseases resulting from infection of macrophages by obligate intracellular parasites of the genus Leishmania, continues to threaten millions of people worldwide (2).

The World Health Organization has classified the leishmaniasis as a major tropical disease. Leishmaniasis comprises a group of diseases with extensive morbidity and mortality in most developing countries. An effective vaccine against leishmaniasis is not available and chemotherapy is the only effective way to treat all forms of disease. The classic and new experimental drugs for the treatment of leishmaniasis have several drawbacks, such as the need for long periods of medication, renal disruption or other side-effects, the emergence of resistance, toxicity, and high costs. Thus, the search for new and more effective chemotherapeutic agents against leishmaniasis with fewer or no side-effects continues. To this end, the rational design of new experimental antileishmanial drugs is an important goal.

Transcription and replication are vital to cell survival and proliferation as well as for smooth functioning of all body processes. DNA starts transcribing or replicating only when it receives a signal, which is often in the form of a regulatory protein binding to a particular region of the DNA. Thus, if the binding specificity and strength of this regulatory protein can be mimicked by a small molecule, then DNA function can be artificially modulated, inhibited, or activated by binding this molecule instead of the protein. Thus, this synthetic/natural small molecule can act as a drug when activation or inhibition of DNA function is required to cure or control a disease. Recent developments in the biology of protozoa have also demonstrated that acridines could exert a powerful toxicity toward Plasmodium, Trypanosoma, and Leishmania (3) parasites. However, their mechanisms of action toward protozoa are still poorly understood. Keeping this in mind, several experiments have been performed to test DNA-binding drugs to cure Leishmania infection (4–15). Experimentally, these compounds have shown positive result and proven to be good antileishmanials. Since there is no report on their docking study yet, so in this work, we have studied their interaction with DNA in silico. To find out the best compound for further studies in future, we have compared their docking results and predicted their ADME/T properties. For this study, selection of compounds was performed on the basis of their performance in experiments. Studies show that these molecules or their core structures specifically bind to DNA major or minor grooves (12,15,23,24,26,27,29).

Materials and Methods

Ligand and receptor preparation

The structure of all the compounds was drawn with ChemSketch 12.01 (http://www.acdlabs.com/resources/freeware/chemsketch/). Their common pharmacophores are shown in Table 3. For all the computational studies, Schrodinger LLC (New York, NY, USA) was used. The designed compounds then imported to maestro v8.5, a graphical user interface (GUI) for all of Schrödinger’s computational programs. Ligand preparation was carried out using LigPrep module. The Schrödinger ligand preparation product LigPrep was designed to prepare high-quality, all atom 3D structures for large numbers of drug-like molecules, starting with 2D or 3D structures in SD, Maestro, or SMILES format. Ligand preparation was carried out using OPLS 2005 force field and pH 7 ± 2. A total of 32 conformers per ligand were generated by retaining specific chiralities. Crystallographic structures of DNA-bound ligand were retrieved (Table 1) from PDB database (http://www.rcsb.org), and receptor preparation was carried out by Protein Preparation Wizard module in Maestro v8.5 by importing the structure in workspace. The different types of ligand attached to them were identified and removed and then were minimized by applying the force field OPLS2005 by keeping its default values. All the water molecules were removed (except present in the binding site) and then explicit H atoms were added to the structure. The van der Waals radii were scaled up by the default value of 1.00 Å for the atoms with the partial charges of <0.25. The receptor grid was generated around the centroid of the ligand contained by receptors file, and the ligands with maximum grid size were allowed to dock. The 1FD5 was used for generating docking grid to acridine and its derivatives. Same was done with other compounds to their corresponding receptors.

Table 1.   PDB IDs of receptor–ligand complex retrieved for docking study
S. no.Receptor–ligand complexPDB ID
1DNA–Acridine1FD5
2DNA–Berberine3NP6
3DNA–Berenil2GVR
4DNA–Bisindolyl methane2B9S
5DNA–Duocarmycin1DSA
6DNA–Pentamidin3EY0

Docking

The automated molecular docking calculations for all the minimized structures were carried out using the Glide program (16). Glide docks flexible ligands into a rigid receptor structure by rapid sampling of the conformational, orientational, and positional degrees of freedom of the ligand. Glide uses MCSA (Monte Carlo-based simulated algorithm) based minimization. The ligands were docked flexibly to write up to 10 000 poses per docking run in the standard precision mode, and a total of five poses per ligand were used for postdocking minimization.

As a first step in its docking protocol, Glide carries out an exhaustive conformational search, augmented by a heuristic screen that rapidly eliminates conformations deemed not to be suitable for binding to a receptor, such as conformations that have long-range internal hydrogen bonds. This is performed by placing the ligand center at various grid positions of a 1 Å grid and rotates it around the three Euler angles. The next filter is a grid-based force field evaluation and refinement of docking solutions including torsional and rigid body movements of the ligand. This procedure eliminates high-energy conformers by evaluating the torsional energy of the various minima using a truncated version of the OPLS-AA molecular mechanics potential function and by imposing a cutoff of the allowed value of the total conformational energy compared to the lowest energy state. A small number of surviving docking solutions can then be subjected to a Monte Carlo procedure to minimize the energy score. The final energy evaluation is carried out with GlideScore, and five best poses were generated as the output for a particular ligand. (16).

GScore = 0.065*vdW + 0.130*Coul + Lipo + Hbond + Metal + BuryP + RotB + Site

(where vdW, van der Waal energy; Coul, Coulomb energy; Lipo, lipophilic contact term; HBond, hydrogen-bonding term; Metal, metal-binding term; BuryP, penalty for buried polar groups; RotB, penalty for freezing rotatable bonds; Site, polar interactions in the active site; and 0.065 and 0.130 are the coefficients of vdW and Coul).

Pharmacokinetics prediction of the compounds

Computational methods are increasingly used to streamline and enhance the lead discovery and optimization process. Though, accurate prediction of absorption, distribution, metabolism and excretion (ADME properties) is often difficult owing to the complexity of underlying physiological mechanisms. However, these properties have been studied by the ADME scoring using the QikProp module (W Jorgensen – Schrodinger LLC, 2003) of Maestro v8.5. QikProp™ v3.1 is a quick, accurate, easy-to-use absorption, distribution, metabolism, and excretion (ADME) prediction program. QikProp predicts physically significant descriptors and pharmaceutically relevant properties of organic molecules, either individually or in batches. QikProp settings determine which molecules are flagged as being dissimilar to other 95% of the known drugs. The compounds were neutralized before being used by QikProp, and the program was processed in normal mode. The neutralizing step is essential, as QikProp is unable to neutralize a structure, and no properties will be generated in the normal mode. The program was processed in normal mode and predicted 44 properties for all the molecules, consisting of principal descriptors and physiochemical properties with a detailed analysis of the log P (octanol/water), QP%, and log HERG, etc. It also evaluates the acceptability of the analogs based on Lipinski’s rule of 5 (17,18), which is essential for rational drug design. Lipinski’s rule of 5 is a rule of thumb to evaluate drug likeness or to determine whether a chemical compound with a certain pharmacological or biologic activity has properties that would make it a likely orally active drug in humans. The rule describes molecular properties important for a drug’s pharmacokinetics in the human body, including its ADME. However, the rule does not predict whether a compound is pharmacologically active (17). Toxicity (T) prediction of all molecules was done in Toxtree-v2.1.0 (Ideaconsult Ltd., Sofia, Bulgaria, 2010). Toxtree is a full-featured and flexible user-friendly open source application, which is able to estimate toxic hazard by applying a decision tree approach. If any compound fails to follow these rules or properties, is considered out of the category of drug.

Result and Discussion

Experimental studies shows that the ligands (Table 2) used in this study are good antileishmanials (4–15), but docking of these compounds revealed a great variation in their binding energy. Initially, totally five poses were saved from which best pose with lowest energy was chosen for each compound. Although the predicted free energy of binding is a useful descriptor of ligand–receptor complementarity, the choice of the ‘best’ docking model was ultimately dictated by various parameters of ADME/T study. Our results show that several compounds were not following the given range limit of few ADME/T properties. These include PISA (1/4 component, i.e. carbon and attached hydrogens of the SASA; range, 0–400 Å2), QlogPC16 (predicted log of hexadecane/gas partition coefficient; range, 4–18), QlogPo/w (predicted log of octanol/water partition coefficient; range, −2 to 6), QlogS (Predicted log of aqueous solubility; range, −6 to 0.5 m). QPPMDCK (predicted apparent MDCK cell permeability; range, poor < 25 nm/seconds and great > 500 nm/seconds), QPlogHERG (predicted IC50 value of blockage of HERG K+ channel; range, <−5) were also found to be violated by 2, 5-bis-(4-amidinophenyl)thiophene (5c) and berenil, respectively (not shown in Table 4).

Table 2.   Compounds investigated in this study Thumbnail image of
Table 3.   Common pharmacophore/pharmacophores among the compounds studied Thumbnail image of
Table 4.   Docking and ADME/T Score of the compounds under study
S. NoCompound idG scoreSASAFOSAFISAPISAWPSAVolQPpolrzQPlogPC16QPlogPoctQPlogPwQPlogPo/wQPlogSQPlogHERGPHOARule of fiveToxicity
  1. SASA, total solvent accessible surface area (range, 300–1000 Å2); FOSA, hydrophobic component of the SASA (range, 0–750 Å2); FISA, hydrophilic component of the SASA (range, 7–330 Å2); PISA, 1/4 component of the SASA (range, 0–400 Å2); WPSA, weakly component of SASA (range, 0–150 Å2); Volume, total solvent accessible area (range: 500–2000 Å3); QPpolrz, Predicted polarizability (range, 13–70 Å3); QPlogPC16, rredicted log of hexadecane/gas partition coefficient (range, 4–18); QPlogPoct, predicted log of octanol/gas partition coefficient (range, 8–43); QPlogPw, predicted log of water/gas partition coefficient (range, 5–48); QPlogPo/w predicted log of octanol/water partition coefficient (range, −2 to 6); QplogS, predicted log of aqueous solubility (range, −6 to 0.5 m); QPlogHERG, predicted IC50 value for blockage of HERG K+ channels (range, concern below −5); PHOS, per cent human oral absorption; GScore, Glide Score.

  2. The values in bold are out of range from the category of drug.

11a−5.987491560.887190.08127.05243.80945.4532.7279.97817.22211.9892.019−4.088−5.43688.7210NO
21b−4.727568N/AN/AN/AN/AN/AN/AN/AN/AN/AN/AN/AN/AN/AN/AN/ANO
31c−4.637121738.471098.531639.901291.949.08916.00222.97714.4764.7796.764−8.1951000NO
42a−5.621212N/AN/AN/AN/AN/AN/AN/AN/AN/AN/AN/AN/AN/AN/AN/ANO
52b−6.474291551.198374.143.851173.201002.434.3318.87913.4096.7343.049−2.499−5.0461000NO
62c−5.619455734.312512.31176.5345.470119639.57512.01223.79115.361.983−5.99−5.92367.1571NO
72d−5.509143N/AN/AN/AN/AN/AN/AN/AN/AN/AN/AN/AN/AN/AN/AN/AHigh
83−10.93399564.4250266.67297.80934.3128.30212.33423.99619.586−0.02−2.279−5.88340.1311NO
94a−11.09014767.378426.04121.51149.470.471374.145.76913.97923.24912.6984.0216.408−5.8211000NO
104b−12.05836573.2160220.97352.20972.2532.26712.57522.05415.9091.681−3.356−5.99557.8421NO
115a−10.94681851.3790125.74624.8100.91506.454.96319.51129.66416.426.0648.293−8.61486.7132NO
125b−11.71767889.017152.6888.545571.176.731632.158.19419.80929.37416.4276.5758.222−8.23596.0172NO
135c−11.4314429.265218.29210.9800692.7721.5716.54135.5386.1021.414−3.354−3.4457.9751NO
145d−11.41968911.9152.18102.79605.651.371647.759.14920.2729.73416.8056.5068.509−8.6393.1952NO
155e−10.88554603.1760225.4349.528.321020.434.16813.23922.19615.512.19−4.031−6.2360.071NO
166−9.29537.669484.3953.27200966.5531.2438.12714.1096.8192.467−2.056−4.9382.280NO

Hydrophobic drugs with high partition coefficients are preferentially distributed to hydrophobic compartments such as lipid bilayers of cells while hydrophilic drugs (low partition coefficients) preferentially are found in hydrophilic compartments such as blood serum. As adequate solubility and permeability is a prerequisite for drug absorption from the gastrointestinal tract, it plays a significant role for the resulting bioavailability of orally administered drugs. By violating these properties, a compound may be associated with side-effects, high or poor solubility (QlogS) that can result in poor absorption and distribution in the body.

The interactions and other scores resulted from docking studies for each compound are summarized in Table 5. The results are briefly described in the following section.

Table 5.   Details of interactions between seco-hydroxy-aza-CBI-TMI and DNA
LigandReceptorInteraction
LIG.NH (37)DT4.A.O2 (148)Hydrogen bond
LIGDA6.AHydrophobic
LIGDT7.AHydrophobic
LIGDT18.BHydrophobic
LIGDA19.BHydrophobic

Acridine and its derivatives

Acridine family includes a wide range of tricyclic molecules with various biologic properties. Considered as potential antiparasitic agents since the 1990s, numerous acridine derivatives have been synthesized and successfully assessed for their antimalarial, trypanocidal or antileishmanial properties (19,20,21). According to a previous study, N-[6-(acetylamino)-3-acridinyl] acetamide (compound 1c) and N-[6-(benzoylamino)-3-acridinyl] (compound 1b) benzamide demonstrated highly specific antileishmanial properties against the intracellular amastigote form of the parasite (12). But our docking score suggests that only acridine (compound 1a; GScore, −5.987491) was the best with no violation of any ADME/T descriptor (Table 4). Compound 1c was found to violate few ADME/T properties when analyzed by QikProp.

Berberin and its derivatives:

Berberine, a quaternary alkaloid and a minor groove binder (23), has demonstrated experimental and clinical efficacies against both visceral (4,6,7) and cutaneous (4) leishmaniases. Despite the apparent potential of this compound as an antileishmanial drug, only Putzer (22) has described the antileishmanial activity of derivatives of berberine and failed to present any biologic or chemical data to substantiate his claims. Further, in support of this study, Vennerstrom et al. (9) suggested that tetrahydroberberine (2c) was proven less toxic and more potent against Leishmania donovani than berberine (2a), N-methyl tetrahydro berberinium iodide (2b), berberine chloride (2d). In contrast to these results, we have found major variation in the data when applied in silico automated docking calculations. Our results show that N-methyl tetrahydro berberinium iodide (compound 2b) was having the best GScore (−6.474291) among all with no violence of ADME/T parameters. When studied using ToxTree software, berberine chloride (compound 2d) was analyzed as highly toxic among the all compounds (Table 4).

Berenil

Berenil (diminazene aceturate), a minor groove binder in AT-rich domain (24), has been proven a good antileishmanial previously (5,8). Our in silico docking calculation with a GScore of −10.93398 and ADME/T study with no violence of any parameters is in agreement with previous studies (Table 4).

Pentamidine and its derivatives

Pentamidine (compound 5a) is one of the few antileishmanial drugs currently available. It belongs to the diamidine class of drugs, which has been suggested to exert antiparasitic activity by binding to DNA, interfering with polyamine metabolism and disrupting of mitochondrial membrane potential (25). Typically, these dicationic molecules bind to DNA by selectively interacting with AT-rich regions of the minor groove. The complementarity of the curvature of the dicationic molecules with that of the minor groove of DNA has been considered an important determinant in the interaction of such groove binders with DNA (26,27). Pentamidine is already in clinical use but limited by toxicity, administration by injection, and development of resistance (10). In addition to pentamidine, many of diamidine molecules –N-[2-(methylsulfanyl)-4-{5-[3-(methylsulfanyl)-4-(pyridine-2-imidamido) phenyl] fura N-2-Yl} phenyl] pyridine-2-carboximidamide (5b), 2 5-bis-(4-amidinophenyl)thiophene (5c), N-(3-chloro-4-{5-[2-chloro-4-(pyridine-2-imidamido) phenyl] fura N-2-yl} phenyl) pyridine-2-carboximidamide (5d), 2 4-bis-(4-amidinophenyl)furan (5e) – exhibited promising activity against Leishmania. The antileishmanial activity of several such dicationic molecules was reported earlier (10,11,13,28).

Docking studies of these compounds including pentamidine reveal that all these compounds were scored a good GScore, but only compounds 5c and 5e satisfied the ADME/T parameters’ range and other three compounds viz. 5a, 5b, 5d seemed to violate few filters.

Diindolyl methane

(3,3′)-Diindolylmethane (DIM, compound 6) is a natural compound, product of acid condensation of indole-3-carbinol (I3C), found in most cruciferous vegetables of the genus Brassica. DIM inhibits an unusual bisubunit topoisomerase I from L. donovani and has been found to interact both with free enzyme and with DNA. A study has been carried out to verify the interaction between DIM-DNA (15). Their data and our result of its in silico interaction with DNA (GScore, −9.29) and ADME/T study showed that the compound can be a future drug molecule.

Duocarmycine and its derivative

The duocarmycins bind to the minor groove of DNA and alkylate the nucleobase adenine at the N3 position (29). Seco-hydroxy-aza-CBI-TMI, analog of duocarmycin (4a), was shown to inhibit the growth of the protozoan parasites L. donovani, Leishmania mexicana, in culture (14). Seco-hydroxy-aza-CBI-TMI (4b) was also shown (Figures 1 and 2) to score a good binding energy (GScore, −12.058362), and the study of ADME/T parameters with no violence of any parameters is in agreement with the previous studies (Table 4).

Figure 1.

 Illustration of binding pocket of seco-hydroxy-aza-CBI-TMI in DNA: (A) and (B) Ligand is shown as stick and DNA residues as wire. Possible hydrogen bond is shown as yellow thick line with the atoms and distance (2.844 Å).

Figure 2.

(A) Ligand is shown in active site pocket. Active site residues are shown as surface. (B) 2D view of the complex (generated by Poseview v1.0.0; BioSolveIT GmbH, Sankt Augustin, Germany). Hydrogen bond is shown as dashed line, and residues taking part in hydrophobic interaction are in green color. Hydrophobic interactions are shown as green line.

Conclusion

In this study, we have conducted comparative docking analysis by automated docking study using Glide that has allowed us to determine the receptor–ligand interactions of the most representative DNA-binding ligands available from experimental analyses. A detailed comparative analysis of the interactions between proposed ligands and DNA has pointed out which compound with best ADME/T score can be taken into account for the design of new compounds. Taken together, these docking results suggest that an ideal ligand must be the one that has a less docking score and that comes under the parameters designed for satisfying its drug likeness. In agreement with these statements, with the satisfaction of the ADME/T parameters like Lipinski’s rule of five (17,18), toxicity, etc., and with a lowest GScore (−12.058, Figures 1 and 2), we suggest that seco-hydroxy-aza-CBI-TMI (4b) is proven as a potential molecule for leishmania treatment by targeting the DNA.

Footnotes

Acknowledgments

The authors are thankful to the Sub-Distributed Information Center (BTISnet SubDIC), Department of Biotechnology and Department of Pharmaceutical Science, BIT, Mesra, Ranchi, for their kind support.

Ancillary