Cynaroside inhibits Leishmania donovani UDP-galactopyranose mutase and induces reactive oxygen species to exert antileishmanial response

Abstract Cynaroside, a flavonoid, has been shown to have antibacterial, antifungal and anticancer activities. Here, we evaluated its antileishmanial properties and its mechanism of action through different in silico and in vitro assays. Cynaroside exhibited antileishmanial activity in time- and dose-dependent manner with 50% of inhibitory concentration (IC50) value of 49.49 ± 3.515 µM in vitro. It inhibited the growth of parasite significantly at only 20 µM concentration when used in combination with miltefosine, a standard drug which has very high toxicity. It also inhibited the intra-macrophagic parasite significantly at low doses when used in combination with miltefosine. It showed less toxicity than the existing antileishmanial drug, miltefosine at similar doses. Propidium iodide staining showed that cynaroside inhibited the parasites in G0/G1 phase of cell cycle. 2,7-dichloro dihydro fluorescein diacetate (H2DCFDA) staining showed cynaroside induced antileishmanial activity through reactive oxygen species (ROS) generation in parasites. Molecular-docking studies with key drug targets of Leishmania donovani showed significant inhibition. Out of these targets, cynaroside showed strongest affinity with uridine diphosphate (UDP)-galactopyranose mutase with −10.4 kcal/mol which was further validated by molecular dynamics (MD) simulation. The bioactivity, ADMET (absorption, distribution, metabolism, excretion and toxicity) properties, Organisation for Economic Co-operation and Development (OECD) chemical classification and toxicity risk prediction showed cynaroside as an enzyme inhibitor having sufficient solubility and non-toxic properties. In conclusion, cynaroside may be used alone or in combination with existing drug, miltefosine to control leishmaniasis with less cytotoxicity.


Introduction
Leishmaniasis is a vector-borne disease caused by obligate intra-cellular protozoan parasite of genus Leishmania. The disease is endemic in more than 89 different countries worldwide. There are 20 different species of Leishmania which can infect animals, humans and sand-flies to cause species-specific complications in the host. The majority of Leishmania species responsible for causing leishmaniasis are Leishmania major, L. infantum, L. chagasi and L. donovani. The reservoirs of the disease are animals like canines and rodents (zoonotic cycle) and human (anthroponotic cycle) [1,2]. There are several molecules which are reported to have anti-leishmanial activity yet, but only a few are classified as effective antileishmanial drugs that are used for the treatment of leishmaniasis [3]. The first effective

Anti-promastigote evaluation and IC 50 determination
Log-phase promastigotes were enumerated and incubated at the density of 5 × 10 6 parasites in the absence and presence of cynaroside at two-fold serial dilutions starting at 120 μM for 48 h at 22 • C. The percentage parasite viability was calculated as (Mean parasite number of treated sample/Mean parasite of control) × 100. The 50% inhibitory concentration was determined by extrapolating the graph of % viability of parasites against concentration of the drug/compound. Morphological changes of the parasites were observed under 20× and 40× lenses using LMI U.K. vImage software.

Anti-amastigote and cytotoxicity evaluation of cynaroside
A total of 2 × 10 6 THP-1 differentiated macrophages were plated in 96-well tissue culture-grade plates in RPMI 1640 complete media with 5% CO 2 at 37 • C to evaluate the cytotoxic activity of cynaroside. Differentiated adherent cells were washed with plain RPMI 1640 media and exposed to two-fold serial dilution of cynaroside starting from 500 μM for 24 h. Cells were further incubated with 50 μM of 5 mg/ml of MTT for 3-4 h, thereafter the resulting formazan was dissolved in 150 μM of DMSO. The amount of formazan produced represented the relative number of viable cells which was recorded spectrophotometrically at 570 nm by ELISA plate reader. The cytotoxic concentration 50%, i.e. CC 50 value was determined by extrapolation of the dose-response curve of percentage cell viability vs concentration of the compound. For parasite load calculation, THP-1-differentiated macrophages were plated on coverslip in six-well plates and infected by L. donovani with 1:10. Infected macrophages were treated with different concentrations of the drugs/compounds for 48 h and then cells were fixed by chilled methanol and stained by Giemsa to calculate the % parasites load.

Study of cell cycle of promastigotes
In brief, the promastigotes were cultured in the absence and presence of different concentrations of cynaroside and miltefosine as the positive control. The promastigotes were harvested after 48 h of incubation and washed thrice with PBS followed by fixation with 80% chilled ethanol and kept at 4 • C overnight. The fixed cells were washed twice with PBS and incubated with 200 μg/ml of RNase at 37 • C for 1 h followed by staining with 50 μM of 1 mg/ml of propidium iodide (PI) for 20 min in dark. Cells were analyzed through flow cytometer (BD FACS ARIA).

ROS estimation
To assess the cynaroside-induced ROS generation, 5 × 10 6 parasites were incubated with different concentrations of compounds/drugs at 22 • C for 48 h. The treated parasites were washed with PBS and incubated with 10 μM of fluorescent dye, 2,7-dichloro dihydro fluorescein diacetate (H 2 DCFDA) for 20 min in dark and analyzed through BD FACS ARIA. Data were represented in the form of histograms.

Homology modeling
The crystal structures of Pteridine reductase protein was retrieved from Protein Data Bank (PDB) [ID: 2XOX (PTR1)]. The PDB file used for docking-based virtual screening study was prepared by removing water molecules and adding hydrogen atoms. Due to the lack of solved 3D structure of LdUGM, LdPPDK and LdTS, homology modeling was opted to determine the structure of these enzymes. The protein sequence of these enzymes was retrieved from the NCBI protein database and PSI-BLAST was performed against PDB to find identical protein [21]. Homology modeling for the proteins was performed by SWISS-MODEL homology modeling server [22]. The template structures generated were assessed and the suitable template was selected for modeling the 3D structure. Out of all the modeled structures, one of the modeled structures was selected based on the sequence identity, Q-Mean, GMQE values and Ramachandran plot, and validated and analyzed through the RAMPAGE, PROCHECK and PDBsum servers [21,23]. The sequence identity between the target and template structure and the root-mean-square deviation (RMSD) value was calculated in Chimera [24].

Protein processing and ligand preparation
The downloaded and modeled protein structures were prepared by 'Discovery Studio visualizer 2020' . Protein molecules were loaded and the water molecules present in the proteins were removed and polar hydrogen atoms were added. The ligands attached to the proteins were selected and binding sites were defined followed by removing the ligands. The identification of the important residues of the catalytic pockets were taken from the native binding pockets of accessible crystal structure of proteins, various submitted literatures, from their homologous template proteins and investigation within the mechanism of inhibition. The active site residues of the protein are found and noted down. Preparation of ligand molecule was performed using PyRx.

Molecular docking and molecular dynamics simulation of protein-ligand complex
The protein molecule was loaded in the PyRx Virtual Screening Tool [25] and was converted into a macromolecule pdbqt format. The cynaroside to be screened was imported and converted into the ligand pdbqt format. In the Vina wizard the pdbqt macromolecules and pdbqt ligands are selected and the grid box is set up with all the active site residues containing within the grid box. The docking is performed by running Vina. The output file was analyzed to find binding energy. The best orientation of the cynaroside was selected and saved as a PDB file. The protein molecule and the best-oriented ligand molecule were loaded in Pymol [19] and the protein-ligand complex was visualized. The protein-ligand complex was loaded in Ligplot software [26] and the output 2D diagram was analyzed to find the number of hydrogen bonds and the binding site residues of the protein.
The most favourable binding poses of the cynaroside were analyzed by choosing the lowest free energy of binding ( G) and the lowest inhibition constant (K i ) which is calculated using the following formula: Additionally, we have performed molecular dynamics (MD) simulations using the GROMACS (version 5.1) software [18,24,25] for the best docking complex at 50 ns at 300 K. The resulting trajectories were analyzed, using RMSD, RMSF, RG and SASA by the utilities provided by GROMACS.

Pharmacokinetics studies
The selected ligand was evaluated for pharmacological profiles by analyzing for Lipinski's rule of violation-5, which was analyzed by Molsoft L.L.C.: Drug-Likeness and molecular property prediction for drug-likeness (http://www. molsoft.com/mprop/) [27]. The bioactivity of cynaroside was checked by Molinspiration (https://molinspiration. com/cgi-bin/properties). The cynaroside was further evaluated for ADMET (absorption, distribution, metabolism, excretion and toxicity) properties by GUSAR [28] and SwissADME database [29]. SwissADME software helps us to identify the selected drug molecule by applying different virtual screening methods. Different components of lipophilicity (iLOGP, WLOGP, XLOGP3, MLOGP, Log Po/w), pharmacokinetics (GI absorption, BBB permeant, P-gp substrate, Log (K p )), water solubility also helped in the preliminary testing of the suitable drug molecule. OSIRIS Property Explorer programme was used to evaluate the mutagenic, tumorigenic, irritant and reproductive risks, and which also provides information on the compound's toxicity, solubility (LogS), hydrophilicity (LogP), molecular weight, drug-likeness and drug score [30].

Statistical analysis
All the experiments were performed in technical triplicate and done at least thrice, and the results represented are the mean of the triplicate with SEM. Statistical analysis was performed using GraphPad Prism 7.0 software and the statistical significance was calculated using one-way ANOVA followed by Dunnett's multiple comparison test. P<0.5 was considered statistically significant.

Cynaroside inhibits growth and survival of promastigotes in vitro
The cynaroside treatment showed the dose-dependent inhibition of parasite growth with 50% of inhibitory concentration (IC 50 ) of 49.49 + − 3.515 μM ( Figure 1A). Though, antileishmanial drug miltefosine showed similar trend with IC 50 value of 6.439 + − 0.5032 μM ( Figure 1A). Cynaroside inhibited only 20% of parasites at 20 μM concentration though it inhibited more than 50% of parasites when used in combination with non-toxic dose of miltefosine, 4 μM ( Figure 1B). The morphological analysis of treated L. donovani shows a characteristic difference as compared with untreated parasites. Parasites were visualized as cylindrical, elongated and flagellated in untreated control though oval and reduced size with degenerated flagella in treated samples ( Figure 1C).

Cynaroside inhibited intra-macrophagic parasites with less cytotoxicity
Cell cytotoxicity of cynaroside along with the miltefosine as the positive control was evaluated against THP-1-differentiated macrophages. The macrophages were incubated at different concentrations of cynaroside, miltefosine and both in combination. The cell viability was assessed through MTT assay and 50% cell cytotoxic (CC 50 ) concentration of cynaroside was determined as 65.33 + − 5.272 μM while that of miltefosine as 20.39 + − 1.69 μM ( Figure  2A). Infected macrophages were treated with cynaroside, miltefosine and both in combination at the least cytotoxic dose to assess the effect of cynaroside on the intra-macrophagic amastigotes ( Figure 2B). Cynaroside reduced the number of intra-macrophagic amastigotes by 50% at 20 μM concentration though it reduced to 80% when used in combination with 4 μM miltefosine ( Figure 2B). Similar effects were visualized in the images of stained slides ( Figure  2C). We have used the higher concentration of cynaroside because it is less toxic up to higher doses in comparison with miltefosine.

Cynaroside checks the parasites in G 0 /G 1 phase of cell cycle
Cell cycle analysis portrays the effect of treatments on its progression in different stages. Cynaroside arrested the promastigotes progressing from S to G 2 /M at 20 μM concentration (Figure 3). There was approximately 7% surge in S phase population when both cynaroside (20 μM) and miltefosine (4 μM) were used in combination (Figure 3). It suggested that cynaroside in combination with low dose of miltefosine inhibited the cell cycle progression as does higher dose of miltefosine. Higher doses of miltefosine are highly cytotoxic and can not be used safely.

Cynaroside treatment induced ROS generation
To study the impact of cynaroside treatment in inducing ROS generation in L. donovani promastigotes, H 2 DCFDA was used which is fluorescent green in presence of OH radicals and H 2 O 2 . Fluorescence intensity is directly proportional to ROS generation and peak shifting to right. In untreated control no green fluorescence was observed though at 20 μM of cynaroside the peak was shifted to right and approximately 12.8% of cells showed ROS production ( Figure  4). On the other hand cynaroside in combination with low dose of miltefosine induced the ROS generation in approximately 22.2% of cells (Figure 4). The result suggested, cynaroside treatment induced ROS production that lead to stress in the parasites. Cynaroside in combination with low dose of miltefosine was found to be more effective than alone (Figure 4).

Cynaroside blocks important drug targets of L. donovani
Molecular docking is one among the foremost popular methods within the field of computer-aided drug designing (CADD) for the identification of latest drug leads [31,32]. In the present time, CADD is getting used to annotate and analyze big drug libraries rapidly and hence saving an enormous amount of energy, time and costs [33]. Among all kinds of various interactions like amide-π interactions, π-π, H-bond etc., the binding efficiency is being evaluated on the idea of hydrogen bonding [31,34]. First, the 3D structures of the enzymes were constructed and validated by homology modeling from template structures (Supplementary Figures S1 and S2 and Tables S1-S3). Molecular   Figure S4). On the basis of binding affinity, cynaroside has been found to possess binding energy of −10.4 , −8.1, −7.8 and−7.7 kcal/mol, with of LdUGM, LdPTR1, LdPPDK and LdTS enzymes, respectively ( Table 1). The binding energy (kcal/mol) is employed to compare and study the binding affinity of various compounds/ligands with their respective target molecule, i.e. lower the binding energy, higher the affinity of the ligand for the receptor. So, the ligand with the highest affinity can be chosen as the potential drug for further studies. The enzyme UGM shows the lowest binding energy. The binding pattern of cynaroside with LdUGM may hamper the substrate accessibility and its subsequent inhibition as shown in ( Figure 5A) where the binding energy is −10.4 kcal/mol (Table 1). Cynaroside interacts with Val 61 , Tyr 98 , Tyr 157 , Arg 177 , Tyr 328 , Arg 338 , Tyr 407 , Tyr 442 (2) binding site residues of LdUGM by forming nine intermolecular hydrogen bonds with bond length ranging from 3.26, 3.10, 3.08, 3.10, 3.22, 2.80, 2.78, 2.96 and 3.26Å, respectively, as shown in Ligplot ( Figure 5B). It has been observed that residues of LdUGM such as Leu 43 , Ser 44 , Gly 59 , His 60 , Arg 436 , Ala 445 , Asn 446 , Gln 447 (n=8) are showing significant interactions with cynaroside ( Figure 5B). The enzyme LdPTR1 shows binding to cynaroside by forming four hydrogen bonds with residues Ser 40 (3) Figure  S5C). The minimum binding energy of −10.4 kcal/mol and the value of predicted inhibitory constant of the Cynaroside with respect to LdUGM was found to be 7.65 μM, which could be considered as its leishmanicidal activity (Table  1). Further, MD simulation of the highest affinity target, UDP galactopyranose mutase (UDPGM) upon cynaroside has been performed for 50 ns to understand the mechanism of interaction. After analyzing MD trajectories (using RMSD, RMSF, RG and SASA), we found that cynaroside perfectly bound into the active site UDPGM (Figure 6A-D). However, the trend of slight fluctuation continues in the cynaroside complexes till the end of simulation.

Pharmacodynamic studies showed non-toxic properties of cynaroside
The pharmacological studies were done on cynaroside for a good oral administration established through the Lipinski's rule of five, which was evaluated by Molsoft L.L.C.: Drug-likeness and molecular property prediction. Cynaroside followed all the parameters of Lipinski's rule of five, except HBD and HBA criteria which are exceeding by 2 and 1 atoms, respectively. The lipophilicity of cynaroside showed value of 0.47 that indicates moderate sublingual absorption as observed from Table 2. Lipinski's 'rule of five' [35] is an analytical approach for predicting drug-likeness stating that molecules had molecular weight (M.W. ≤ 500 Da), high lipophilicity expressed as LogP (LogP ≤ 5), hydrogen bond donors (HBD ≤ 5) and hydrogen bond acceptors (HBA ≤ 10) have good absorption or permeation across the cell membrane. For choosing the selected drug molecules through virtual screening, sometimes we found violation of some selection rules like Ro5, Veber etc. At present, in drug industry there are several important drugs available in the market which violate some likeness rules. Among the very popular drugs, some like fosinapril, bromocriptine mesylate, dabigitranetexilate, olmesartanmedoxomil and reserpine etc revealed two Ro5 rule violations [36]. The SwissADME was used for pharmacodynamic study of cynaroside to understand the action of drug inside a host's body. Cynaroside possess low gastrointestinal absorption and good solubility with value −3.65, which is higher than −4 (≥ −4) [37]. Cynaroside is not permeable to the blood-brain barrier. The leadlikeness criteria is violated only for molecular mass (≥350) and possess a moderate bioavailability score. The results are summarized in Table 2. The ADMET study focused on the parameters that can define absorption, distribution, metabolism, excretion, toxicity, gastrointestinal absorption (GIA), solubility (LogS), P-glycoprotein substrate inhibition, cytochrome substrate/inhibitor. Cynaroside was evaluated as an active enzyme inhibitor with value 0.42. The predicted bioactivity    by molinspiration is shown in Table 3. Molinspiration was used to evaluate the bioactivity of cynaroside by calculating the activity against GPCR ligand, kinase inhibitor, ion channel modulator, protease inhibitor, nuclear receptor ligand and enzyme inhibitor [37]. The bioactivity values were interpreted as follows: inactive (bioactivity score ≤ −5.0), moderately active (bioactivity score: −5.0 to 0.0) and active (bioactivity score ≥ 0) [38]. As per the Organisation for Economic Co-operation and Development (OECD) chemical classification, cynaroside was found to be a non-toxic as mentioned in Table 4. The principal aim of predicting the acute toxicity is to evaluate undesirable side effects of a compound after single or multiple exposures to an organism via a known administration route (oral, inhalation, subcutaneous (sc), intravenous (iv) or intraperitoneal (ip)). GUSAR was used to determine the acute toxicity of the successfully docked cynaroside based on the Prediction of Activity Spectra for Substances algorithm and Quantitative Neighborhoods of Atoms descriptors. The obtained result was compared with Toxicity Database to categorise on the basis of OECD chemical classification manual [28]. The criteria used for cynaroside to elicit toxicity based upon the administration route when the compound dose is more than 7000 mg/kg for intravenous route, more than 500000 mg/kg in case of the oral route, and more than 20000 mg/kg for intraperitoneal route and subcutaneous database as shown in Table 4. All the predicted toxicity risk factors for cynaroside were low and molecular weights less than 500, implied that it is likely to be absorbed and are capable to reach the place of action when administered as drugs [39]. It was predicted that cynaroside possessed no mutagenic, tumorigenic, irritant and reproductive effective toxicity risks as shown in Table 5.

Conclusion
Here, it was observed that cynaroside has the potential antileishmanial activity. It showed better response when used in combination with low concentrations of miltefosine. Through in silico study, we have found that the binding energy and the binding site residues of the LdUGM exhibited best interaction with the inhibitory flavonoid, cynaroside. It was confirmed by MD simulation study also. Our results suggested that cynaroside may be used as food constituent after the detailed in vivo studies on experimental visceral leishmanaisis, to fight against leishmaniasis.

Data Availability
Data are with the authors and will surely be provided on request through corresponding author.