Identification of Leishmania donovani Topoisomerase 1 inhibitors via intuitive scaffold hopping and bioisosteric modification of known Top 1 inhibitors

A library of arylidenefuropyridinediones was discovered as potent inhibitors of Leishmania donovani Topoisomerase 1 (LdTop1) where the active molecules displayed considerable inhibition with single digit micromolar EC50 values. This molecular library was designed via intuitive scaffold hopping and bioisosteric modification of known topoisomerase 1 inhibitors such as camptothecin, edotecarin and etc. The design was rationalized by molecular docking analysis of the compound prototype with human topoisomerase 1 (HTop1) and Leishmania donovani topoisomerase 1(LdTop1). The most active compound 4 displayed no cytotoxicity against normal mammalian COS7 cell line (~100 fold less inhibition at the EC50). Similar to camptothecin, 4 interacted with free LdTop1 as observed in the preincubation DNA relaxation inhibition experiment. It also displayed anti-protozoal activity against Leishmania donovani promastigote. Crystal structure investigation of 4 and its molecular modelling with LdTop1 revealed putative binding sites in the enzyme that could be harnessed to generate molecules with better potency.


Results and Discussion
Design and molecular docking. To begin with the design of novel inhibitors of topoisomerase 1 (human or Leishmania donovani), we scrutinized the structures of existing topoisomerase inhibitors such as camptothecin derivatives, E-ring modified camptothecin derivatives and non-camptothecin molecules (Fig. 1a) [21][22][23] .
Careful analysis of the structures in Fig. 1a revealed the presence of four common motifs, the phthalimide (A), indolizinone (B), pyridopyranones (C), and pyridooxepinones (D) (Fig. 1b) within the structural framework of the molecules. Next, these four motifs (A-D) were utilized to generate E by scaffold hopping (Fig. 1b). Subsequent bioisosteric modification of E afforded furopyridinedione derivative 13 as the library scaffold. The aryl moiety provides option for further diversification if required at a later stage (Fig. 1b).
To rationalize our design, 13 was docked with LdTop1. The experiment revealed several binding interactions such as intercalation at the DNA cleavage site, five hydrogen bonding with residues Arg190, Gln454 and Tgp11 and nonbonding interactions with Asn221, Dt10 and Dg12 (Fig. 2a) (refer SI for details of the docking experiment involving protein preparation). With HTop1 there were lesser H-bonding interactions (3 vs 5) (Fig. 2b). Consequently 13-LdTop1 complex (− 8.07 kcal/mole) was ~0.4 kcal/mole more stable than 13-HTop1 complex (− 7.70 kcal/mole). This was comparable to the binding interactions of camptothecin and edotecarin with both LdTop1 and HTop1 (refer SI).
Next, in a bid to understand the physicochemical indications rudimentary to the design we computed few relevant medicinal chemistry descriptors of 13 along with CPT, DFT, EDT and RST (Table 1). From the information it was revealed that 13, CPT, RST and DFT adhere to the Lipinski rules and that the physicochemical properties of 13 resides well within the range of these but EDT (Table 1). With very high molecular weight (MW), hydrogen bond donor and acceptor (HBD and HBA) ability, total polar surface area (TPSA) and rotatable bond, EDT stands apart from the rest. Interestingly compound 13 has the lowest Log P (Table 1) which indicated better membrane absorption and binding ability to the enzyme but poor aqueous solubility among all the molecules. Low log S w of compound 13 (almost at par with CPT [− 3.64 vs − 3.4]) indicated that substituting 13 with polar functionalities may improve the overall solubility. Finally 13 had the best oral bioavailability amongst all (Table 1).
Chemistry. Based on these docking results, we decided to generate a library based on 13 using combinatorial synthesis. Accordingly Knoevenagel condensation of furopyridinedione with appropriate aromatic and heteroaromatic aldehydes was explored with different solvents such as ethanol, methanol, isopropanol, n-butanol, acetonitrile and tetrahydrofuran. The reaction in ethanol happened to be most efficient where the product precipitates at the end of the reaction and could be isolated and purified by filtration and washing with ethanol without the need for chromatographic separation.
With the optimized protocol in place, next the Knovenagel reaction was conducted in parallel in custom made metallic reaction blocks containing pre-tared reaction vials with 1 mM solution of furopyridinedione at 80 °C (Fig. 3). As expected the product precipitated on completion of the reaction. The supernatant ethanol was removed and the precipitate was washed thrice with ethanol and the solvent was subsequently removed. The whole operation was executed in parallel using a Tecan Evo Freedom liquid handling system containing robotics arms. The vials were dried in a genevac and weighed to determine the amount. An assay plate was generated with the samples at 50 μ M concentration in dimethyl sulfoxide (DMSO), for in vitro screening. The average yield of the compounds ranged from 54-96%. The compounds were characterized by 1 H and 13 C nuclear magnetic resonance spectroscopy (NMR) and high resolution mass spectroscopy (HRMS). The proton NMR unveiled the  In vitro screening and structure activity relationship (SAR). To investigate the inhibitory effect of our library compounds on the DNA relaxation activity of the recombinant Human and Leishmania topoisomerase 1 ( Table 2), assays were performed involving supercoiled plasmid DNA, the enzymes (HTop1 and LdTop1) and our molecules. The enzymes were purified by the standard literature procedure 9 . The inhibitory activities of our molecules against HTop1 and LdTop1 (EC 50 ) were compared to that of the standard top1 inhibitor camptothecin and the results are shown in Table 2. Several molecules (1, 2, 3, 4, 9, 14 and 15) showed considerable activity against LdTop1. However none of the molecules from the library exhibited inhibition against the HTop1, rendering them selective towards LdTop1. This is interesting when compared to the original molecules such as Camptothecin, Diflomotecan, Edotecarin and Rosettacin, which were used to design this scaffold. Camptothecin a Top1 poison has comparable inhibitory effect on both LdTop1 and HTop1. Nevertheless the other three compounds which are clinical candidates against various cancers are selective HTop1 inhibitors. This is noteworthy because our molecules inspired from majorly selective HTop1 inhibitors displayed particular binding to LdTop1 over HTop1 during in vitro screening (Table 2), which was predicted during the molecular docking ( Fig. 2) exercise.
The structure activity relationship study demonstrated the importance of the substituted aromatic moiety towards the LdTop1 inhibitory activity (Fig. 4a). In general the heteroaromatic (5 and 7), naphthalene (11) and aliphatic (16) derivatives, exhibited no inhibitory activity at all. Among the monosubstituted aromatic analogs the electron withdrawing o-trifluoromethyl, 3 (EC 50 ~5 μ M) and electron donating m-bromo, 9 (EC 50 ~7.5 μ M) substituents on the ring favors the inhibitory activity. Other meta-substituted compounds such as chloro and iodo analogs (14 and 15) also exhibited inhibitory activities of 17.42 ± 1.86 and 35.57 ± 2.42 μ M, respectively. All the para substituted analogs rendered inactive. Among the di-substituted compounds, the 3,4-dihydroxyphenyl derivative 4 exhibited the best potency among all the analogs with an EC 50 of 4 μ M. Interestingly, protecting the dihydroxy functionality as in 6 or installing electron withdrawing groups on the same position as in 8 led to the complete loss of activity. The 2,5-dimethyl analog, 1 was moderately active with an EC 50 of 23 μ M, whereas the 2,3-dimethoxy analog 12 was inactive. The only tri-substituted derivative 2 was decently active with EC 50 of 10 μ M. Apart from the importance of the pyridone moiety, the present study unveils two distinct SAR (Fig. 4) trend: (a) may be polar protic functionalities capable for H-bond formation at C 3 and C 4 positions of the aromatic ring is beneficial for the activity of the molecules and (b) strong electron donating functionality at meta position of the aromatic ring augments the inhibitory activity.
Additionally we performed pharmacophore analysis of our library. Accordingly pharmacophore queries (3D arrangement of molecular features) were devised from both active and inactive compounds of our library against LdTOP1 such that all or most of active compounds satisfied the queries (Fig. 4b). The analysis was done using pharmacophore elucidator module of MOE TM .
All the library compounds were sketched and energy minimized using MOE TM . The overlap score of 6.6636 indicated that major active molecules (7) were aligned in the design. The best query on the basis of overlap score  (alignment score) divulged 2 aromatic/pi ring centers (Aro|PiR) at the arylidene and furanone moiety, one hydrophobic centroid (Hyd) at the pyridone domain and two hydrogen-bond acceptor projection (Acc2):2) that are strategic and could be harnessed for further diversification (Fig. 4b).
Studies with compound 4. In a bid to justify highest in vitro activity of 4, it was further docked with LdTop1 enzyme (Fig. 5b). The experiment revealed that at the best pose, 4 and LdTOP1-DNA complex intercalates at the DNA cleavage site (Fig. 2, SI). The key binding sites in 4 as depicted in Fig. 5a are the pyridine nitrogen, the amide carbonyl and the two aromatic hydroxy functionalities. They form hydrogen bonding interactions with Arg190, Asn221 and DA113 and nonbonding interactions with Thr217, DT10, TGP11, DC112 residues of the LdTop1 enzyme (Fig. 5a). These were comparable to the interactions of camptothecin and LdTop1 (Fig. 5b). Further to comprehend the interactive nature of 4 a detailed three-dimensional structural investigation was carried out. The X-ray crystal structure determination revealed that compound 4 has crystallized along with dimethyl sulfoxide (DMSO); the solvent used for the crystallization. The molecule is in planar geometry and its m-hydroxy group forming strong O-H… O (1.824 Å) and O-H… S (2.900 Å) hydrogen bonds with the Oxygen and Sulfur atoms of DMSO, respectively, as depicted in Fig. 6.
The molecular packing diagram in the crystal lattice of 4 as shown in Fig. 7 reveals the presence of dimer motifs forming via N-H… O hydrogen bonds (2.054 Å) between the furopyridine moieties. The dimers are noticed at the corner and middle of the unit cell. The O-H… O hydrogen bonds forms between the hydroxyl groups of 4 and the Oxygen atom of DMSO molecules are also evident from the packing diagram.
The possible interactions in 4 were further visualized by performing Hirshfeld surface analysis [24][25][26] . which highlights the contact points of a molecule with its neighboring molecules in a crystal. The dimer formed via strong N-H… O hydrogen bonds and the strong O-H… O bonds formed with DMSO molecule are highlighted as large dark red circular surfaces (Fig. 8). Other weaker interactions (C-H… O) are also highlighted as small red  To interrogate whether compound 4 interacts with the enzyme, it was preincubated with LdTop1 at different concentration for 5 minutes at 37 °C before the addition of substrate DNA (Fig. 9a). The inhibitory effect of 4 in preincubation condition was further compared with its inhibition when incubated simultaneously with LdTop1 and super coiled DNA in the relaxation experiment (Fig. 9b). Nearly 100% inhibition was observed at 20 μ M of 4 under preincubation compared to ~92% inhibition under simultaneous condition. And the EC 50 values for simultaneous and preincubation assays were 3.77 ± 0.27 and 1.72 ± 0.23 μ M respectively. From this results we conclude that the inhibition profile was marginally improved by preincubation of enzyme with 4.
Next, compound 4 was screened against extracellular promastigotes 27 . The results (depicted in Table 3), revealed that it inhibited the proliferation of Leishmania donovani with an IC 50 of 4.21 ± 0.21 μ M. Standard leishmanial drug, miltefosine was used as a positive control and exhibited IC 50 of ~1 μ M. It was subsequently tested for its cytotoxicity against healthy COS7 mammalian cell lines and exhibited toxicity at a concentration nearly 100 times its IC 50 values (Table 3).

Conclusion
In this manuscript we reported the design of a series of furopyridinedione derivatives as LdTop1 inhibitors via intuitive scaffold hopping and bioisosteric modification of known top1 inhibitors. The design was rationalized by   (Fig. 8b). The active compounds showed minimal toxicity when screened against mammalian COS7 cells. Pharmacophore modelling, X-ray crystallographic investigation of 4 and molecular docking studies of 4/LdTop1-DNA ternary complex recommended the following structural attributes necessary to attend optimum LdTop1 inhibitory activity for our molecules: (i) functionalized pyridinedione as the central scaffold; (ii) polar protic functionality at the C 3 and C 4 and/or (iii) an electron donating substituent at C 3 position of the benzene ring attached to the pyridinedione moiety.

Methods
Chemistry. All reactions were carried out in flame-dried sealed tubes with magnetic stirring. Unless otherwise noted, all experiments were performed under argon atmosphere. All reagents were purchased from Sigma Aldrich, Acros or Alfa Aesar. Solvents were treated with 4 Å molecular sieves or sodium and distilled prior to use. Purifications of reaction products were carried out by column chromatography using Chem Lab silica gel (230-400 mesh). Infrared spectra (IR) were recorded on a Thermoscientific Neoled IS5 FTIR spectrophotometer and are reported as wavelength numbers (cm −1 ). Infrared spectra were recorded by preparing a KBr pellet containing the title compound. 1 H NMR and 13 C NMR spectra were recorded with tetramethylsilane (TMS) as internal standard at ambient temperature unless otherwise indicated on a Varian 300/400 and JEOL JNM-ECX500 MHz at 500 MHz for 1 H NMR and 100 MHz for 13 C NMR. Chemical shifts are reported in parts per million (ppm) and coupling constants are reported as Hertz (Hz). Splitting patterns are designated as singlet (s), broad singlet (bs), doublet (d), triplet (t). Splitting patterns that could not be interpreted or easily visualized are designated as multiple (m). The Mass Spectrometry analysis was done on the 6540 UHD Accurate-Mass Q-TOF LC/MS system (Agilent Technologies) equipped with Agilent 1290 LC system obtained by the Dept. of Chemistry, School of Natural Sciences, Shiv Nadar University, Uttar Pradesh 203207, India.
Plasmid relaxation assay. The type I DNA topoisomerase are assayed by decreased mobility of the relaxed isomers of supercoiled pBS (SK + ) [pBlue-script (SK + )] DNA in agarose gel. The relaxation assay was carried out as described previously with LdTop1B and HTop I, diluted in the relaxation buffer (25 mM Tris-HCl, pH 7.5, 5% glycerol, 0.5 mM DTT, 10 mM MgCl 2 , 50 mM KCl, 25 mM EDTA and 150 mg/ml BSA) and supercoiled plasmid pBS (SK + ) DNA (85-95% were negatively supercoiled, with remainder being nicked circles) 9,28,29 . The reconstituted enzyme LdTop1B was assayed at 50 mM KCl concentration as described before 30 . For all kinetic studies, the reaction mixtures containing the buffer and DNA were heated to 37 °C before addition of the enzymes. The reactions were rapidly quenched using stop solution and kept on ice. The gels were stained with ethidium bromide (EtBr) (0.5 mg/ml) and the amount of supercoiled monomer DNA band fluorescence were quantified by integration using Gel Doc 2000 under UV illumination (Bio-Rad Quantity One Software), as described previously 30 . Purification of recombinant human topoisomerase I (HTop1). The wild-type HTop1 (91 kDa) was purified from Sf-9 insect cells infected with the recombinant baculovirus (a kind gift from Prof. J. J. Champoux). Approximately, 1 * 10 9 Sf-9 cells were infected with the recombinant virus, and cells were harvested after 48-h of infection. The cells were lysed and enzyme was purified as described 31 . Briefly, at 48 h post-infection, approximately 3 * 10 9 Sf-9 cells were harvested by centrifugation for 5 min at 400 * g. The cells were resuspended in 1 liter of ice-cold phosphate-buffered saline and centrifuged for 5 min at 400 * g. The nuclei were washed twice in 80 ml of lysis buffer minus Triton X-100 and resuspended in 40 ml of resuspension buffer (50 mM KCl, 10 mM Tris-Cl, pH 7.5, 2 mM MgCl 2 , 25 mM DTT, 0.4 mg/ml phenylmethylsulfonyl fluoride, 0.12 mg/ml aprotinin). The nuclear extract was stirred for 5 min at ~200 rpm. With continued stirring, 50 ml of polyethylene glycol (PEG) buffer (18% PEG 8000, 1 M NaCl, 10% glycerol) was added dropwise in order to precipitate the DNA. The PEG precipitate was pelleted by centrifugation and the resulting PEG supernatant was dialyzed against 4 liters of potassium phosphate buffer. The dialyzed PEG supernatant passed over a 10-ml bed volume of phosphocellulose (Whatman P11) equilibrated with PPB. The HTop1 flowed through the PS column and then passed over a Mono Q column followed by loaded onto a Mono S column. The Mono S column was eluted with a 25-ml 50-200 mM KPO 4 , pH 7.4, gradient. HTop1 eluted as a single major peak at 150 mM KPO 4 . The peak Mono S fractions were pooled, concentrated with an Amicon ultrafiltration cell, dialyzed into storage buffer (50% glycerol, 10 mM Tris-hydrochloride, pH 7.5, 1 mM DTT, 1 mM EDTA), and stored at − 20 °C. 4.21 ± 0.21 390.58 ± 3.15~100 OD 600 = 0.6 with 0.5 mM IPTG (isopropyl β -D-thiogalactoside) at 22 °C for 12 has described previously 9 . Cells harvested from 1 litre of culture were separately lysed by lysozyme/sonication, and the proteins were purified through Ni-NTA (Ni 2+ -nitriloacetate-agarose column (Qiagen, Hilden, Germany) followed by a phospho-cellulose column (P11 cellulose; Whatman, Maidstone Kent, UK) as described previously 9 . Finally, the purified proteins LdTOP1L and LdTOP1S were stored at − 70 °C. The concentrations of each protein were quantified by using Bradford Reagent (PIERCE, Thermo Fisher Scientific Inc., Rockford, USA). Purified LdTOP1L was mixed with purified LdTOP1S at a molar ratio of 1:1 to a total protein concentration of 0 The crystal structure had various missing residues and side chains which were added using structure preparation module of MOE software. 3D protonation of the proteins were executed in order to prepare them for docking. Water molecules were removed and hydrogen atoms were added and energy was minimized by MOE energy minimization module using MMFF94x force field. Further we used the procedure described by Roy et al. in which the 22 DNA double strand is substituted with 22 DNA double strand present in the human Top1-DNA complex (PDB ID: 1K4T) by fitting the two molecules on the backbone atoms 11 . The substitution has done since the presence of a cleaved site in the 22 double strand present in the 1K4T X-ray structure that contains the topotecan molecule permits to investigate the ability of the compound 15 to interact with the LdTOP1LSDNA cleavable complex.

Purification and reconstitution of two subunits of
A grid of 60 × 60 × 60 with 0.375 Å were constructed around the protein DNA complex area using Autogrid v 4.2 software. The 2D structure of compounds was built using MOE-builder tool and minimized using MMFF94x force field. Docking analysis were done using Autodock v 4.2 software in which top 10 docked conformation were taken using Genetic Algorithm. Each docking calculation consist of 25 × 10 6 energy evaluations with 250 population size. A mutation rate of 0.02 and a crossover rate of 0.8 were used to generate new docking trials. X-Ray crystal analysis. The compound 4 was crystallized by slow evaporation of its solution in dimethyl sulfoxide (DMSO) at room temperature, after seeding with its poor quality crystals grown in (N-Methyl-2-pyrrolidone (NMP). Crystal structure of 4 was determined by measuring X-ray diffraction data on a D8 Venture Bruker AXS single crystal X-ray diffractometer equipped with CMOS PHOTON 100 detector having monochromatised microfocus sources (Mo-Kα = 0.71073 Å) at room temperature. The structure was solved using SHELX program implemented in APEX3 32,33 . The non-H atoms were located in successive difference Fourier syntheses and refined with anisotropic thermal parameters. All the hydrogen atoms were placed at the calculated positions and refined using a riding model with appropriate HFIX commands. The program Mercury was used for molecular packing analysis 34 . Hirshfeld surface analysis was performed using CrystalExplorer package 35 . The structural details can be found from the deposited CIF with CCDC numbers 1451904.
Pharmacophore Modelling. All the library compounds were sketched and energy minimized using MOE.
A library database was constructed consisting of structures of compound 1-21 and EC 50 values in MOE. This database was given as an input to pharmacophore elucidator and pharmacophore queries were searched. The best query on the basis of overlap score (Alignment score) was chosen.