Repurposing the Pathogen Box compounds for identification of potent anti-malarials against blood stages of Plasmodium falciparum with PfUCHL3 inhibitory activity

Malaria has endured as a global epidemic since ages and its eradication poses an immense challenge due to the complex life cycle of the causative pathogen and its tolerance to a myriad of therapeutics. PfUCHL3, a member of the ubiquitin C‐terminal hydrolase (UCH) family of deubiquitinases (DUBs) is cardinal for parasite survival and emerges as a promising therapeutic target. In this quest, we employed a combination of computational and experimental approaches to identify PfUCHL3 inhibitors as novel anti‐malarials. The Pathogen Box library was screened against the crystal structure of PfUCHL3 (PDB ID: 2WE6) and its human ortholog (PDB ID: 1XD3). Fifty molecules with better comparative score, bioavailability and druglikeliness were subjected to in‐vitro enzyme inhibition assay and among them only two compounds effectively inhibited PfUCHL3 activity at micro molar concentrations. Both MMV676603 and MMV688704 exhibited anti‐plasmodial activity by altering the parasite phenotype at late stages of the asexual life cycle and inducing the accumulation of polyubiquitinated substrates. In addition, both the compounds were non‐toxic and portrayed high selectivity window for the parasite over mammalian cells. This is the first comprehensive study to demonstrate the anti‐malarial efficacy of PfUCHL3 inhibitors and opens new avenues to exploit UCH family of DUBs as a promising target for the development of next generation anti‐malaria therapy.

www.nature.com/scientificreports/ In this study, we employed structure-based drug discovery approach to repurpose 400 compounds from MMV Pathogen Box as novel anti-malarial with PfUCHL3 inhibitory effect. Herein, we docked Pathogen Box compounds against a grid encompassing active site residues and neighboring residues of PfUCHL3 and HsUCHL3. Based on comparative dock score and ADME (absorption, distribution, metabolism and excretion) profiling, best fifty candidates were assessed against in-vitro activity of P. falciparum UCHL3. Further, in-silico approaches were employed to gain insight into the enzyme-inhibitor interactions and the efficacy of the most promising hits was evaluated against the blood stages of P. falciparum in culture. The selectivity of lead compounds for parasites over the mammalian cell lines was also assessed. We present the first report attempting to identify the novel inhibitors targeting PfUCHL3 with potent inhibition of parasite growth. In the process, this study provides new chemical scaffolds that may serve as a template to develop more specific and potent clinical therapeutics against malaria.

Results
Structure based virtual screening (SBVS). SBVS approach was employed to identify novel anti-PfUCHL3 molecules. A flow chart depicting the screening methodology is illustrated in Fig. 1. Molecular docking algorithms were used to screen 400 compounds against the X-ray crystallographic structure of the enzyme (PDB ID: 2WE6) and with its human correspondent (PDB ID: 1XDT), which shares only 30% identity. Each molecule was docked 50 times in different conformations at the active site cleft of both the enzymes and the top hundred molecules were picked on the basis of higher selectivity towards P. falciparum UCHL3 than its mammalian counterpart (Table S1).
Next, the compounds exhibiting drug like properties, i.e. with acceptable range of pharmaceutically relevant properties and with zero violation of Lipinski rule of five were selected. Pharmaceutically relevant properties such as the partition coefficient (log Po/w) critical for estimation of absorption of drugs within the body, log Kp predicting skin permeability and log S exhibiting water solubility of the compounds were assessed, and eventually Pan-assay interference compounds which give false positive biological output were rejected 68 . As presented in Table S2, among the set of hundred compounds, based on SWISSDOCK-based ADME predictions, fifty distinct chemical entities that entirely fit inside the pink zone region of the bioavailability radar that displays the rapid appraisal of drug likeliness, were further investigated for their activity on recombinant PfUCHL3 in vitro.
Expression, purification and in-vitro assessment of compounds for enzyme inhibition. Plasmodium falciparum and human UCHL3 were expressed and purified to homogeneity as a single band around 30 kDa on SDS-PAGE (Fig. 2a-e). Both the enzymes were found to be catalytically active as evaluated by deubiquitinase assay (Fig. 2f). Subsequently, we evaluated the performance of the potential fifty compounds identified from in-silico studies against the catalytically active enzyme at a concentration of 100 μM with NEM (a known inhibitor of cysteine protease) as a positive internal control. At this concentration, the majority of compounds were ineffective on PfUCHL3 activity, whereas, four compounds showed more than thirty percent inhibition and only two of them (MMV676603 and MMV688704) exhibited an inhibition rate of more than fifty percent, as shown in Fig. 3.
Compounds with more than 50% enzyme inhibition were further evaluated for their inhibitory potential in a dose-dependent manner. As shown in Fig. 4a,b, compounds MMV676603and MMV688704 displayed an IC 50 value of 25.12 ± 1.80 μM and 87.14 ± 1.75 μM, respectively. Concomitantly, due to the limited availability of the compounds, we evaluated these compounds only at their respective IC 50 values against closely related human ortholog and observed no discernable effect on HsUCHL3 activity. The findings allude to the fact that both MMV676603 and MMV688704 hold the potential to inhibit in-vitro activity of PfUCHL3 over its mammalian ortholog with MMV676603 being a more potent inhibitor.   Table 1. Together, these results imply that compounds MMV676603 and MMV688704 harness more specific inhibitory action on PfUCHL3 than its mammalian ortholog. www.nature.com/scientificreports/ Molecular insights: unravelling interactions between identified leads and the enzyme. Protein-inhibitor interaction studies were performed to gain better insight into the binding mode of the compunds MMV676603 and MMV688704. As depicted in Fig. 5a, MMV676603 obstructs the catalytic center of PfUCHL3 by funneling into the binding groove of PfUCHL3. This complex is stabilized by the formation of a hydrogen bond with ASN61 and hydrophobic interactions with HIS164, PHE208, ASP179, LYS182, TYR56, VAL58 and ASP60 of PfUCHL3 (Fig. 5c). Similarly, MMV688704 aligns in the binding groove of PfUCHL3 and overpasses the catalytic center (Fig. 5b) by forming a hydrogen bond with ARG181 and hydrophobic interactions with ASN59, HIS164, ASP60, TYR56, VAL58 and ASN61 (Fig. 5d). On the contrary, both MMV676603   www.nature.com/scientificreports/ and MMV688704 sit in an entirely different groove (Fig. 6a,b) and show interactions with residues that do not play any interphase in enzymatic activity (Fig. 6c,d) and do not demonstrate any specific binding to human ortholog, supported by docking results. Hence, these results indicate the selective binding of the identified hits to PfUCHL3 over HsUCHL3 and provide scope for further refinement and lead optimization studies.

Anti-malarial effect of MMV676603 and MMV688704 on intra-erythrocytic stage of P. falciparum.
Since PfUCHL3 is cardinal for the parasite, therefore, we hypothesized that the inhibitors of PfUCHL3 would be effective in inhibiting parasite growth. To test this, we determined the inhibitory effect of the identified hits against the asexual intra-erythrocytic stage of P. falciparum 3D7 in culture using SYBR Green I assay. The standard anti-malarial, Chloroquine was used as an internal reference for validation of the assay. As expected, both the compounds were equally effective and caused complete inhibition of parasite growth at a concentration of 10 μM. Compounds MMV676603 and MMV688704 displayed a dose dependent inhibition of Chloroquine (CQ)-sensitive P. falciparum 3D7 strain with IC 50 values 450.5 ± 1.84 nM and 266.6 ± 1.77 nM, respectively ( Fig. 7a,b).

In-vitro phenotypic and speed of action analysis of the two hit molecules.
To investigate the effect of MMV676603 and MMV688704 on parasite growth and phenotype, ring stage parasites were treated with the compounds at ten-fold concentration of the IC 50 value and observed at various intervals for 56 h. Cultures treated with MMV676603 and MMV688704 displayed the appearance of young trophozoites at 16 h, exhibiting a similarity with the control set. However, both MMV676603 and MMV688704 treated parasites were arrested at the late trophozoite stage with shrunken, less granular and condensed appearance of the parasite. In comparison to control, both MMV676603 and MMV688704 treatment caused substantial delay in the transition from trophozoite to schizont stage. This results in complete arrest of the parasite growth at later stages of the   Fig. 8a. These results were further supported by quantification of parasitemia level at different time points, which revealed a continuous reduction in parasite count after 24 h accounting for almost ten-fold drop in parasitemia at 56 h (Fig. 8b). These results suggest that both MMV676603 and MMV688704 are slow acting inhibitors of parasite growth in culture (Fig. 8a).

Effect of MMV676603 and MMV688704 on ubiquitination levels.
Considering the in-vitro inhibitory action of the two molecules against PfUCHL3 activity, we determined their effect on parasite ubiquitination level to ascertain that the parasiticidal activity of the identified compounds is mediated through PfUCHL3 inhibition. With this aim, asynchronous parasite culture was exposed to three times the IC 50 concentration of the two inhibitors and ubiquitination levels were detected post treatment. Treatment with both MMV676603 and MMV688704 induced a significant accumulation of high-molecular weight ubiquitin conjugates. In comparison to the vehicle control set, treatment with MMV676603 displayed the highest levels of ubiquitination laddering, followed by MMV688704 (Fig. 9). In contrast, Chloroquine, a known antimalarial exhibited negligible effect on the parasite ubiquitination levels ( Fig. 9), indicating that these compounds alter PfUCHL3 activity for effective parasite killing. and MMV688704 was not due to general toxicity, the compound mediated cytotoxic effect was determined in the human hepatic cell line (HepG2) and human embryonic kidney cells (HEK-293T) with the help of MTT assay. No cytotoxicity was observed up to a concentration of 20 μM (100 times the effective concentration) (Fig. S3). Furthermore, as summarized in Table 2, compound MMV676603 and MMV688704 displayed a Selectivity Index of 50 and 100, respectively, presumably indicating a high anti-parasitic selective profile of these hits.

Discussion
Regardless of the extensive efforts to combat this infectious disease, malaria continues to torment the world with its severity and complexity 69 . Limited protection by licensed vaccine and development of accelerated resistance to current clinical anti-malarials has overcome the efforts to contain this disease 70,71 . Such a scenario, stimulates an urgent need for new drug discovery programs to mitigate associated problems and develop new and cost effective anti-malarial therapies with least side effects. DUBs are a group of highly sophisticated enzymes known to regulate a number of cellular functions and their deregulation contribute to the development of severe clinical manifestations 28,72 . Likewise, compelling evidences from literature highlight the critical importance of DUBs in parasite biology. However, till recent times, studies  www.nature.com/scientificreports/ exploiting DUBs as a therapeutic target in Plasmodium are limited. PfUCHL3 is one such promising member of Plasmodium deubiquitinase family, whose absence exhibits a lethal effect on parasite survival, thereby, validating its utmost importance as a therapeutic target 46 . A report by Franco et al., identified inhibitors against PfUCHL3 from the ZINC database by virtual screening 73 . Unfortunately, the study lacked enzymatic validation and biological assay of the identified hits. Even though, PfUCHL3 shares some identity with HsUCHL3, the subtle differences in the active site pocket and substrate binding groove could serve as a fundamental scaffold for the development of selective therapeutics 55 . Therefore, in the present study, we focused to query Pathogen Box compounds for identification of novel PfUCHL3 inhibitors as potent anti-malarials. Herein, we performed targeted docking in search of selective inhibitors of PfUCHL3 over HsUCHL3 using AutoDock suite. The molecules, exhibiting selectivity towards P. falciparum UCHL3 and displaying drug like properties were evaluated for their in vitro enzyme inhibition assay at 100 μM. Interestingly, MMV676603 and MMV688704 were identified as potent inhibitors that potentially impeded PfUCHL3 activity at sub-micromolar concentration, which was in close association with the predicted in-silico IC 50 values and was further supported by predicted dock score values. We also performed a detailed analysis of the docked complexes to gain better insight into the binding mechanism of both the hits. Notably, MMV676603 blocks the binding pocket of PfUCHL3 by residing in the binding groove with its azapiro ring facing the catalytic unit (CYS92, HIS164 and ASP179) whereas the benzo ring faces the crossover loop. Such affirmative orientation positions the hydrogen acceptor and donors in the close vicinity thus contributing to the formation of strong non-covalent interactions, including hydrogen bonds with the catalytic triad along with ASP60, ASN61, van der waal interactions with PHE208 and TYR56 and ionic interactions with LYS182. The presence of three fluorines in MMV676603 further strengthens these interactions, thereby, enhancing the selectivity and permeability. Incidentally, fluorination of ligand molecules is a common practice to elevate the potency of the inhibitors. Whilst, these are short interactions, yet they impart a significant effect on protein ligand binding 74 . On the contrary, MMV676603 lacks proper orientation in the binding groove and interacts with the amino acid residues (LEU220-ALA224) lining the C-terminal region of HsUCHL3. Together, these collated evidences advocate for more proficient suppressive effects of MMV676603 on PfUCHL3 activity as compared to its mammalian ortholog. Additionally, MMV688704 also aligns itself in the binding groove of PfUCHL3 with the tail of the benzene ring facing the crossover loop and the pyrimidine ring aligns towards the catalytic center of PfUCHL3. This allows MMV688704 to form a strong network of non-covalent interactions with the amino acid residues surrounding the binding pocket. In contrast, the horizontal orientation of the MMV688704 (facing the C-terminal region) in HsUCHL3 hampers its interaction with the catalytic groove, thereby, impeding its potency towards human UCHL3. Both the compounds obey all druglikeliness parameters as depicted by bioavailability radar and even exhibit zero violation of Lipinski rule of five, which makes them suitable for oral absorption.
The compounds MMV676603 and MMV688704 were evaluated for their anti-malarial potential and compounds MMV676603 and MMV688704 were highly efficient against P. falciparum 3D7 (CQ-sensitive strain) in culture with an IC 50 value of 450.5 ± 1.84 nM and 266.6 ± 1.77 nM, respectively. Both MMV676603 and MMV688704 displayed asexual stage dependent growth inhibition after prolonged exposure of drug (after 24 h) and caused phenotypic and developmental aberrations in parasites. The treated parasite was completely arrested at late stages (trophozoites/schizont) with a compact and shrunken appearance as evidenced by lack of ring stage even after 56 h. The trophozoite stage marks the initiation of a highly replicative phase in the parasite life cycle, which indicates high dependency on the ubiquitin regulatory machinery for invasion and differentiation 16 . Therefore, it can be envisaged that DUBs like PfUCHL3 might regulate the critical processes in late asexual stages of the parasite and its inhibition by MMV676603 and MMV688704 hampers the parasite growth by altering the protein homeostasis and other essential pathways. Additionally, to affirm whether the biological action of these compounds is exerted through PfUCHL3 inhibition, the parasite ubiquitination levels were determined upon exposure to the selected hits. As compared to control and CQ, a known antimalarial with different mode of action, both MMV676603 and MMV688704 treated parasites exhibited a marked increase in the ubiquitination levels, thereby, suggesting their inhibitory role towards the deubiquination process. In comparison to MMV676603, a less intense ubiquitination pattern in MMV688703 treated set could be an attribute of differential impact on PfUCHL3 activity, rendering its dexterity to curb the deubiquitination event in the parasite milieu. Despite such variable biological effects, the enhanced ubiquitination laddering in compounds treated sets provide compelling evidences for PfUCHL3-mediated parasite killing by the identified hits. This, coupled with other results, make it very likely that the compounds MMV676603 and MMV688704 might be equally effective against the resistant strains of the parasite and possibly be used as an adjunct therapeutics in ACTs., Pathogen Box classifies MMV676603 and MMV688704 against tuberculosis and toxoplasmosis, respectively. However, this is the first report of their identification as an anti-malarial. MMV676603 belongs to the nitrobenzothiazinones family and is effective against the clinical isolates of tuberculosis 75 . Currently, it is under pre-clinical trials for the treatment of drug-resistant tuberculosis 76 . Some studies have also identified its potential against N. brasiliensi 77 and Corynebacterineae 78 . Alongside, this drug acts by forming a covalent bond with CYS387 of DprE1of an oxidoreductase enzyme that is vital for cell wall arabinan synthesis and ultimately death 78,79 . Consequently, we can infer that MMV676603 inhibits PfUCHL3 activity by altering the key interaction of cysteine residue present in the active site pocket of the enzyme. Such inhibition renders the enzyme non-functional by obstructing the overall parasite growth. In contrast, MMV688704 belongs to the family of pyrimidine, mainly known as DHFR inhibitors 80 with minimal evidence of it as anti-toxoplasmosis. In our study, MMV688704 displayed a better efficacy against the parasites in culture with enzyme inhibition at the micro molar range, which could be a result of an additional mode of action. Hence, the results necessitate further lead optimization and chemical engineering to improve its efficacy. These studies were further complemented by cytotoxicity assay on mammalian cell lines. Expectedly, compound MMV676603 and MMV688704 selectively inhibited parasites over the mammalian cell lines (HEK-293T and HepG2 cells) with S.I value of more than 50, which was in corroboration with our in-silico www.nature.com/scientificreports/ studies. However, in-vivo and dynamic mechanistic studies are warranted to provide a better insight into the potential of PfUCHL3 as a therapeutic target. Concurrently, our study identified MMV676603 and MMV688704 as promising antimalarial with appreciable inhibition of PfUCHL3 activity.

Conclusion
The present study represents first comprehensive attempt to repurpose the collection of MMV Pathogen Box compounds targeting UCHL3 of P. falciparum for the development of anti-malarial chemotherapy. In this study, we identified two novel promising non-cytotoxic small molecules, MMV676603 and MMV688704, as PfUCHL3 antagonists with biological activity against P. falciparum as an outcome of a stringent screening. In comparison to MMV676603, MMV688704 was found to be more potent against the parasite, but less effective against recombinant PfUCHL3. Therefore, it is tempting to dissect its probable mode of action. Besides this, both MMV676603 and MMV688704 hold the potential as potent anti-malarial. Therefore, improvement towards selectivity for PfUCHL3 over human ortholog and in-vivo studies will complement these findings. The outcome of this study provides a scaffold for future optimization of more specific, effective and selective compounds with its implication as a therapeutic as well as a tool to identify the biological functions of the enzyme. Overall, inhibitors targeting the UCHL3 family of DUBs can enlighten the drug development path and discovery against a range of parasitic and infectious diseases.

Material and methods
Pathogen Box compounds. The Pathogen Box was procured from Medicine for Malaria Venture (MMV) foundation (Geneva, Switzerland). The library of compounds was supplied in 96-well microtitre plate at a concentration of 10 mM in dimethyl sulfoxide (DMSO) (10 μL) and further dilutions were prepared according to MMV instructions. The details about plate layout, chemical formula and other biological activity are available online as Pathogen Box supporting information in a form of an Excel spreadsheet (https:// www. patho genbox. org/ about-patho gen-box/ suppo rting-infor mation).

Heterologous expression and purification of recombinant proteins. Expression vector encoding
PfUCHL3 and HsUCHL3 was transformed in E. coli BL21(DE3) cells. Protein expression and purification of the recombinant proteins were performed as described previously with slight modification 46 . In brief, protein expression was induced with 1 mM IPTG for 16 h at 30 °C. The cells were harvested and the pellet was lysed with lysis buffer (50 mM Tris, 150 mM NaCl, 2 mM DTT) containing protease inhibitor cocktail, lysozyme and DNase. The lysate was sonicated and the clear lysate was loaded onto a pre-equilibrated Ni-NTA column and protein was eluted with increasing concentration of imidazole hydrochloride. Subsequently, the eluants were subjected to size exclusion chromatography (AKTA Prime, GE Healthcare, Björkgatan, Uppsala, Sweden) and the fractions were collected and purity of the protein was determined by SDS-PAGE analysis.

Structure based virtual screening (SBVS).
The dataset comprising of 400 molecules from MMV, Geneva were docked against the active site of PfUCHL3 and its human ortholog whose crystal structure was retrieved from Protein Data Bank (PDB ID: 2WE6 and 1XD3 respectively). Autodock Racoon software 81 was used for docking. Before commencing for docking, the protein and ligands were prepared separately. Water molecules and cofactors were removed from the protein. However, the SMILES of 400 compounds were used to retrieve the 2D structures from ChEMBL database, followed by their conversion to PDB via Pymol and lastly to the PDBQT format. Subsequently, the grid was generated by keeping the active site residues (CYS92, HIS164 and ASP179) of PfUCHL3 as highlights, which are dissimilar in positioning when compared with human UCHL3 whose binding pocket comprises of CYS95, HIS169 and ASP184. Top 100 compounds having a comparative binding energy value of ≥ 3 kcal/mol were selected for ADME analysis. ADME screening. The assortment of flawed molecular entities in the early phases of drug discovery diminishes the error rate and increases the efficacy in the initial stage. Hence, the pharmacokinetic activities of the leads were estimated using SWISS ADME 82 . Only drug-like candidates with zero violation of Lipinski rule of five and with easy absorption, distribution, metabolism and excretion were taken forward. www.nature.com/scientificreports/ In-vitro DUB assay. The selected compounds were initially screened against recombinant PfUCHL3 at a concentration of 100 μM using fluorimetric assay as described by Artavanis et al. 46 with relevant modification. In brief, the assay was performed in a 60 μl reaction volume containing 10 pM of enzyme in reaction buffer (50 mM Tris-Cl, pH-8.0, 150 mM NaCl, 2 mM DTT, 2 mM EDTA and 0.1 mg/mL BSA) and 100 μM of compounds. The fluorogenic peptide substrate Ub-AMC (Ubiquitin C-terminal tagged 7-amido-4-methylcoumarin)(Boston Biochem,Cambridge, MA, USA) was added at a final concentration of 125 nM and the release of AMC was continuously monitored (Excitation: 485 nm and Emission: 535 nm) for a period of 30 min at 25 °C. The enzyme inhibition assay for each compound was performed in triplicates. The compounds inhibiting ≥ 50% of PfUCHL3 activity were identified as potent hits. The IC 50 of best compounds was determined through the dose-response workspace of GraphPad Prism Software (GraphPad Co. Ltd., San Diego, CA, USA).
In-vitro culturing and synchronization of parasite. Plasmodium falciparum 3D7 was cultured with human O + human erythrocyte (5% hematocrit) in RPMI-1640 (Gibco, ThermoFisher Scientific, Waltham, MA, USA) medium supplemented with 0.5% (w/v) Albumax (Gibco, ThermoFisher Scientific, Waltham, MA, USA), Hypoxanthine (50 mg/L), Gentamicin (10 mg/L) and Ampicillin (10 mg/L) in mixed gas environment (5% O 2 , 5% CO 2 and 90% N 2 ) at 37 °C 83 . Human whole blood was procured from the Rotary Blood Bank, New Delhi. Under sterile conditions, erythrocytes were obtained by removal of plasma and peripheral blood mononuclear cells (PBMCs) using histopaque gradient. Parasitemia levels were routinely assessed using Giemsa staining of blood smears. Synchronization was performed using the sorbitol lysis method 84 . The parasite culture was treated with 5% D-sorbitol for 10 min at 37 °C for the enrichment of ring stage parasite. The culture was pelleted by centrifugation and washed thrice with media. In addition, the high synchrony level of the parasite was maintained by performing two consecutive sorbitol treatments at an interval of 4 h. The parasite was visualized by microscopic analysis of Giemsa stained blood smears.
Drug susceptibility assay for asexual stages of P. falciparum. In-vitro anti-malarial efficacy of compounds was determined by SYBR-Green I based proliferation assay as reported previously 85  Phenotypic evaluation and speed of action studies. Synchronized culture containing ring stage parasites at 1% parasitemia and 2% hematocrit was treated with test compounds at tenfold the IC 50 concentration. After addition of compounds, the parasitemia level was quantified and morphological changes were studied by microscopic analysis of Giemsa stained thin blood smear at 0, 8, 16, 24, 36, 48 and 56 h. The speed of action of the compounds was determined on the basis of their effect on the early or late stage of the erythrocytic stages of the parasite as described by Terquille et al. 86 .
Effect of selected hits on parasite ubiquitination levels. Asynchronus parasite cultures at 5% hematocrit and 4-5% parasitemia were incubated in the absence (vehicle control) or presence of three times the IC 50 concentration of the identified compounds or Chloroquine (CQ) for 24 h at 37 °C. Post treatment, cultures were harvested and pellets were extracted with 0.1% (w/v) saponin for 15 min on ice. Cells were pelleted and parasite pellets were washed with PBS until the supernatant became clear. Parasite pellets were resuspended in RIPA buffer (150 mM NaCl, 1% Nonidet P-40, 0.5% sodium deoxycholate, 0.1% SDS, 50 mM Tris pH 8.0) and lysed on ice for 45 min, followed by mild 6 cycles of sonication (2 s On and 2 s Off) and the clear lysate was obtained post centrifugation at 13,800×g for 15 min at 4 °C. Protein concentration was estimated by Bradford and an equal amount of parasite lysates were resolved on SDS-PAGE and subjected to immunodetection using anti-ubiquitin antibody and anti β-actin antibody (loading control).
Cytotoxicity assessment against mammalian cell lines. HepG2 (human hepatocellular carcinoma) and HEK-293T (human embryonic kidney) cells were cultured DMEM (Invitrogen, Carlsbad, CA) medium supplemented with 10% fetal bovine serum and antibiotics (100 U/mL of penicillin and 100 mg/mL streptomycin) in a humidified 5% CO 2 chamber at 37 °C.The cytotoxicity of top hits was assessed against HepG2 and HEK-293T cells using MTT assay as described previously 87 . In brief, cells were seeded in a 96 well microplate and incubated for 24 h at 37 °C in a CO 2 incubator. After 24 h, cells were exposed to different dilutions of drugs (prepared in respective medium) for  www.nature.com/scientificreports/