Identification of first-in-class plasmodium OTU inhibitors with potent anti-malarial activity

OTU proteases antagonize the cellular defense in the host cells and involve in pathogenesis. Intriguingly, P. falciparum , P.vivax , and P.yoelii have an uncharacterized and highly conserved viral OTU-like proteins. However, their structure, function or inhibitors have not been previously reported. To this end, we have performed structural modeling, small molecule screening, deconjugation assays to characterize and develop first-in-class inhibitors of P. falciparum , P.vivax , and P.yoelii OTU-like proteins. These Plasmodium OTU-like proteins have highly conserved residues in the catalytic and inhibition pockets similar to viral OTU proteins. Plasmodium OTU proteins demonstrated Ubiquitin and ISG15 deconjugation activities as evident by intracellular ubiquitinated protein content analyzed by western blot and flow cytometry. We screened a library of small molecules to determine plasmodium OTU inhibitors with potent anti-malarial activity. Enrichment and correlation studies identified structurally similar molecules. We have identified two small molecules that inhibit P. falciparum , P.vivax , and P.yoelii OTU proteins (IC50 values as low as 30nM) with potent anti-malarial activity (IC50 of 4.1- 6.5 μ M) . We also established enzyme kinetics, druglikeness, ADME, and QSAR model. MD simulations allowed us to resolve how inhibitors interacted with plasmodium OTU proteins. These findings suggest that targeting malarial OTU-like proteases is a plausible strategy to develop new anti-malarial therapies.


Introduction
Malaria is a life-threatening disease caused by parasites transmitted to humans through the bites of infected female Anopheles genus mosquitoes (1-4). There have been over 219 million malaria cases only in 2017 worldwide caused by different Plasmodium species. All known Plasmodium species not only exhibit similar biological and morphological characteristics but also may cause malaria disease. Intracellular parasites utilize various mechanisms for intracellular invasion and blocking cellular immune response (3,5). The Crimean-Congo hemorrhagic fever virus (CCHFV) Presence of viral OTU-like proteins in different Plasmodium species suggests that they may involve in malarial pathogenesis. Intriguingly, recent studies show that several OTUlike DUBs are expressed by malaria parasites (6). However, their function and potent inhibitors remain undetermined.
Here we studied three viral OTU-like proteins that we identified in P. falciparum (pf), P. vivax (pv) and P. yoelii (py) species. We hypothesized that these proteins pose DUB activity, which we have tested in vitro and in vivo. In addition, we computationally modeled their structures and performed small molecule library screening followed by in vitro validations. Identification of inhibition pocket in malarial OTU proteins also allowed us to develop novel, first-in-class small molecules that inhibit malarial OTU DUBs with shown antimalarial activity. determination of proteins transferred to nitrocellulose membranes. After washing steps, manufacturer's recommended amount of primary antibodies was diluted at the 1:1000 ratio in 5% milk-incubation solution. Anti-HisTag primary antibody was incubated with the membrane which contains proteins at 4°C for 16 hours. The next day, the membrane was washed 3 times with washing solution and incubated with the secondary antibody (HRP bound and antibody recognizing primary antibodies, at a 1: 2000 dilution) for 1 hour, on a rocking system, at room temperature. Subsequently, the membrane was washed 3 times with washing solution and the protein determination step was accomplished with chemiluminescence imaging system. The amount of expression of each protein sample was determined by Pierce ECL plus reagent and Bio-Rad ChemiDoc MP system.

Affinity purification of recombinant proteins
Bacterial samples with induced OTU expression were exploded with 5 minutes sonication in the solubilization solution (8 M Urea, 20 mM Tris-HCL pH: 8.0) as we did before (7). The samples were centrifuged at maximum speed and the formed precipitate was removed.
The supernatant was passed through 0.45 μm filters and HisTagged recombinant proteins were isolated by HisTrap HP columns (GE Healthcare, 17-5248-01) in accordance with the manufacturer's protocol using the ÄKTAprime plus protein purification system. Solution Pierce ™ BCA Protein Assay Kit was used to determine the amount and concentrations of the proteins obtained. Measurements were carried out by spectrophotometer at 562 nm wavelength.

Desalting and concentration of recombinant proteins
Elution solution was removed using Desalting columns (UFC501024 Amicon Ultra -0.5 mL Centrifugal Filters Ultracel -10K, 10,000 NMWL columns) and replaced with PBS solution.
Briefly, 400 μl of PBS was added to 100 μl of purified protein and centrifuged at 14.000 g for 30 minutes. An additional 400 μl of PBS was added when 100 μl remained, and the procedure was repeated two more times. In the next step, the centrifuge was repeated and eluted. The proteins remaining on the top of the columns were collected in PBS. BCA test was used for the determination of concentrations of recombinant protein samples after desalting. The experiment was carried out on the microplate using the Thermo Scientific ™ Pierce ™ BCA protein test kit following manufacturer's recommendations.

Solution and folding analysis
PBS and QuickFold ™ Protein Refolding Kit (Cat # 0600, Athenaes) solutions were used to determine the conditions under which proteins were folded. 25 μl of isolated proteins are added to 475 μl of each 15 different solutions that supplied in the kit. It was incubated at 4 o C for 1 hour. Then, it was centrifuged at maximum speed (14,000 rpm) for 5 minutes.
Unfolded proteins are expected to form precipitate. 475 μl of the supernatant was transferred to new 1.5 mL tube and water was added to the precipitate and SDS-PAGE analysis was performed. The gel was stained with coomassie staining solution. Then, gel photos were taken with the Bio-Rad ChemiDoc gel imaging device.

Determination of deubiqiutinase activities
Deubikiutinase activities of recombinant proteins were determined by using fluorescently linked Ubiquitin C-terminus peptide (UB-AMC substrate). The solution (pH: 7.5) containing 10 mM HEPES, 100 mM NaCl, and 2.5 mM DTT was used as the primary reaction solution (7). Reactions with different concentrations of recombinant enzymes and Ubiquitin-AMC (Boston Biochem, U550) were prepared as 50 μl in 96 well plates and protected from light. Measurements were made at 345 nm excitation and at 445 nm emission using Thermo Scientific ™ Varioskan™ LUX at 25 o C.

Cloning of OTU sequences into the pCDNA3.1 mammalian expression vector
Recombinant proteins in bacterial vectors were transferred to pCDNA3.1 mammalian expression vector by subcloning. For this purpose, recombinant DNA sequences were obtained after cleavage of pET26b vectors with XbaI (NheI compatible) and EcoRI enzymes and purified from the gel. The pcDNA3.1 vector was cut with NheI (XbaI compatible) and EcoRI restriction enzymes and dephosphorylated with alkaline phosphatase. Then the recombinant DNA sequences were cloned into the pcDNA3.1 vector by ligation. EcoRI-XbaI cleavage was performed for selection of positive clones.

In vitro small molecule analysis
Putative OTU inhibitors were dissolved in DMSO as 10 mM stocks and stored at -20 o C until use. In the first set of applications, 20 μM doses of molecules were tested ( Table 2).
Each compound was purchased from vendors provided by Molport.com. Ribavirin, an antiviral drug, was used as negative control. mOTU DUB activity measurements were made at 345 nm for excision and emission at 445 nm using Thermo Scientific ™ Varioskan ™ LUX at 25°C. DMSO given samples were used as controls. Experiments were conducted with wider range of (0.01-100 μM) inhibitor concentrations to determine the IC 50 values of the inhibitors that we found their inhibitory activity. A dose-response curve was generated for this purpose. IC 50 calculations were calculated with regard to linear regression and logarithmic transformation using the AAT Bioquest IC 50 tool .

Statistical analysis
Significance of the difference between control samples and treatments were calculated by Student's t test. Values less than p <0.05 were considered statistically different.

Identification and characterization of viral OTU-like proteins in plasmodium species
Increasing amount of nucleotide sequence data allowed us to identify CCHFV OTU like proteins in the three different plasmodium species. To this end, we have performed a constraint-based multiple alignment of plasmodium OTU-like Proteins with CCHFV OTU ( Figure 1A) and determined the highly conserved key residues. These conserved residues included highly conserved cysteine, histidine and aspartic acid, which are key component of OTU catalytic domain and Y-WG amino acids, which make inhibition pocket and UB interaction pocket of CCHFV OTU protein ( Figure S1). Note that identified Plasmodium proteins were not previously characterized as OTU proteins. Phylogenetic analysis of studied Plasmodium OTU proteins intriguingly showed that severity of disease might correlate with homology of studied proteins ( Figure 1B). We have also performed homology modeling of Plasmodium OTU-like proteins and found that they pose very similar Y-WG inhibition pockets compared to CCHFV OTU protein ( Figure 1C). Structural comparisons showed that distribution of amino acids in the Ramachandran's plot was appropriate. Alignment scores were 0.013, 0.044, and 0.049 and RSMD values were 0.579, 1.046 and 1.111 Å for pfOTU, pvOTU, and pyOTU, respectively ( Figure S2).
Protein docking studies done with human UB and plasmodium OTU-like proteins also showed that these plasmodium proteins could act as deubiquitinase (DUB) / OTU proteins similar to viral OTU DUB ( Figure 1D).
To further characterize identified proteins, we analyzed composition of proteins in terms of predicted molecular mass, theoretical pI, lipid anchors and predicted disulfide bonds analysis ( Figure 2C). Protein purification was carried out using histrap HP columns and ÄKTAprime plus protein purification system. Subsequently, the purity of the elution obtained from histag columns was investigated by SDS-PAGE. An increased purity relative to lysate samples was detected in SDS-PAGE analysis in the elution ( Figure 2D).
A sample UV graph corresponding to pyOTU is provided in Figure 2D. We

In vitro and in vivo deconjugating activity analysis of mOTU-like proteins
Deubiquitinase activities of malarial OTU proteins were measured by fluorometer using Ubiquitin-AMC (UB-AMC) substrate ( Figure 3A). OTU activity were determined with the increasing AMC emission. pfOTU ( Figure 3B), pvOTU ( Figure 3C) and pyOTU ( Figure   3D) proteins were shown to have significant de-ubiquitination activity in vitro. Initial velocity of pfOTU was higher compared to both pyOTU and pvOTU ( Figure 3E and Figure S4).
Next, we cloned the mOTU DNA into a mammalian expression vector to study their  Figure   4E). The amount of mono or poly-UB conjugated proteins in the central region were decreased after pvOTU and pyOTU treatments. In pfOTU applications, no effect was seen on mono or polyubiquitinated proteins content in the middle region.
We also utilized flow cytometry to further study poly or mono ubiquitinylated protein content in human cells post mOTU transfections. To this end, we transfected HEK cells with pcDNA (control) or pcDNA3.1 vectors expressing pfOTU, pvOTU, and pyOTU. Flow cytometry analysis of poly-UB by BML-PW8805 showed that pfOTU and pyOTU significantly decreased the polyUB content in the human cells (Figure S6A-B).
Furthermore, flow cytometry analysis of intracellular mono-poly-UB content by BML-PW8810 demonstrated that pfOTU, pvOTU and pfOTU lowered the mono-poly-UB content in human cells (Figure S6C-D). These analyses further confirmed the overall downregulation of ubiquitinated proteins in human cells by malarial OTU proteins.
We have also analyzed differentially expressed genes related to cellular immunity response in post mOTU treatments. To this end, we have studied expression of genes involved in pyroptosis pathway, dsDNA sensing, dsRNA sensing, cellular immunity (NF-κB and INFAs), and Interferon-simulated genes ( Figure 4F). interferon-simulated genes such as APOBEC3G and MX1.

In silico screening and identification of potent mOTU inhibitors
Previous studies with CCHFV OTU protein led to identification of two different inhibitors as well as OTU DUB inhibition pocket (7). This allowed us modeling of malarial OTU proteins and their corresponding inhibition pockets as in CCHFV OTU protein ( Figure 5A). In addition, we prepared a library of relevant small molecules to perform in silico small molecule library screening. This library included over 77,000 potent viral OTU inhibitors and anti-malarial compounds curated from PubChem, MMV malaria box compounds, and drugs-now subset of compounds from ZINC database. Binding energies of these compounds were calculated automatically using AutodockVina program and paDEL-ADV platform. Enrichment analysis demonstrated that viral OTU inhibitors are more likely to inhibit mOTU DUBs compared to MMV, anti-malarial or random dataset from ZINC ( Figure 5B, Figure S7A). This enrichment was evident in the compounds as low as -6.6 kcal/mol docking scores ( Figure S7B). Following in silico screening with the small molecule library, we selected twenty different potent compounds with high affinity to pfOTU, pvOTU or pyOTU DUBs. Shape based clustering of hits with average distance allowed clustering hits into three clusters ( Figure 5C). We have identified structurally similar and dissimilar compounds ( Figure 5D). In addition, molecules having possible cardiotoxicity were determined by calculating the binding energies to the human Ether-àgo-go-Related Gene (hERG) potassium channel. Calculated affinities having <-7.8 kcal /

In vitro identification of mOTU inhibitors and their anti-malarial activity
Ubiquitin-AMC (DUB substrate) was used to test whether deubiquitinase activities of inhibited pvOTU DUB activity ( Figure 6B). Small molecules numbered #65, #66 and #76 inhibited pyOTU DUB activity ( Figure 6C). Small molecules numbered #60 and especially #65, #66 and #76 molecules were selected for further analysis as they can inhibit the activity of all three different malarial OTU-like proteins.
We have further investigated how these four different but structurally similar molecules interact with mOTU proteins. We found that these molecules are well embedded in the OTU DUB inhibition pocket (shown in blue, Figure 7A) and are positioned around reported anti-malarial activity and IC 50 values of as low as 3 nM. Overal, we have found that seven out ten pfOTU inhibitors that we identified (see Figure 6A) are known to inhibit P. falciparum growth. These findings suggest that malarial OTU proteins are potent targets for development of novel anti-malarial therapeutics.

Druglikeness analysis of confirmed mOTU inhibitors
We also addressed potent cytotoxicity of these compounds with normal human dermal fibroblasts (HDFs, Figure S8A) and endothelial cells (HUVECs, Figure S8B) at 10 μM and did not detect any significant effect in cell viability.
Selected small molecules were further confirmed with regard to fulfill the conditions of druglikeness, AMES, blood-brain barrier (BBB), ADME/Tox and pharmacokinetics by using and protein-ligand interactions and contacts were analyzed to find the interactions between all the protein-ligand complexes. As seen in the MD studies, local changes in protein-ligand RMSD and RMSF plots throughout simulation trajectory were acceptable.
Protein secondary structural element plots and analysis of these structures over residue index was observed as expected. RMSF, torsional profiles and properties of ligands during MD simulations were analyzed. Ligand RMSF values were generally stable.
The interactions between OTU protein and its inhibitors are realized by virtue of the highly conserved amino acids. Among these contacts, amino acid interactions that interact for more than 20% of the simulation time were selected. It was determined that each of the small molecules selected, numbered #60, #65, #66, #76 OTU inhibitors, were in the inhibition package characterized as Y-W in the pfOTU protein ( Figure S13) MD studies showed that all selected molecules were in the inhibition package of pvOTU, and had significant interaction with conserved Trp153 amino acids ( Figure S13) MD simulations showed that molecules #60, #65, #66, #76 occur in the inhibition package of pyOTU, that is around about Trp118 and Gly119 amino acids of pyOTU ( Figure S13).

Malaria is a lethal infection disease caused by
RTS, S / AS01 (RTS, S) is the only vaccine available to date to provide partial malaria protection in young children (37). Considering all these situations, there is a need for new treatment methods and the discovery of new drugs.
Though P. falciparum is responsible from most of the deaths, P. vivax can lead to serious, even fatal infections and result in mortality. The symptoms of malaria caused by P. vivax are generally the same as for others. There are symptoms such as fever, vomiting, diarrhea, headache and falling from strength. One of the most important problems is the spleen growth. P. vivax enters the bloodstream, move to the liver and form merozoites.
After separation from the liver, merozoites invade red blood cells and begin to multiply.
After 48 hours, red blood cells explode due to excessive proliferation of the parasites, resulting in increased fever in the patients. P. vivax differs from P. falciparum in several respects. The P. vivax parasite invades younger and smaller red blood cells. They can wait for months in a dormant condition in the liver. P. vivax causes less complications but is more common than P. falciparum. More importantly, P. vivax cannot bind to endothelial cells in blood vessels and is rarely fatal. P. yoelii is another intracellular parasite and can cause malaria. Fatal and non-fatal subtypes have been identified.
We have previously identified inhibition pocket of CCHFV OTU deubiquitinase that could be pharmacologically targeted. Intriguingly, three different species of Plasmodium, which are known to cause malaria, pose previously uncharacterized proteins similar to CCHFV OTU deubiquitinase with corresponding conserved residues and inhibition pocket. Viruses utilize OTU deubiquitinase to shut down any cellular defense mechanisms and induce pathogenesis. Extensive in silico studies were performed to identify important amino acid sites in the inhibition of viral OTU protein. During these studies, three new protein sequences in distinct plasmodium species were identified that matched with viral OTU protein and we hypothesized that they may be important actors in intracellular invasion of plasmodium. We have showed DUB activity of viral OTU-like proteins and developed firstin-class drug-like small molecules, which inhibits malarial OTU deubiquitinase activity.
Here we showed that four small molecules could effectively block DUB activity of pfOTU, pvOTU and pyOTU proteins. In addition, several of them are already pose anti-malarial activity. These studies suggest that targeting malarial OTU proteins is a plausible strategy to develop new antimalarial drugs.
A large number of putative deubiquitinases in Plasmodia have been identified based on sequence homology, albeit without proof of expression or their function. In a study, the deubiquitinating and deNeddylating enzyme activity of PfUCH54 was shown experimentally (38). However, the importance and necessity of PfUCH54 (NP_701037) and other parasitic Ub-reactive proteins in Plasmodium species have not yet been thoroughly studied. According to a recent study based on the morphological and growth analysis, another of plasmodium falciparum OTU cysteine protease (PF3D7_1031400, which encode OTU domain, Pfam Accession No. PF02338) was found to be associated with parasitic apicoplasty (39). After its downregulation, it was determined that parasite growth was decreased and the intererythrocytic parasite cycle was disrupted along with the deformation of parasite forms. These results showed that plasmodium falciparum OTU proteins have vital importance for parasitic growth and maintenance (39). Here, we have characterized XP_001352102, XP_001614763, and XP_726550 Plasmodium OTU-like proteins due to their similarity to viral OTU protein. In addition, novel anti-malarial OTU inhibitors discovered in this project may be used to inhibit different types of potential OTU deubiquitinases in parasites.
Inhibition of proteasome function leads to both toxicity in P. falciparum and inhibition of parasite growth (40,41). This suggests that the ubiquitin-proteasome pathway is a good target. Each element in this pathway can be taken as an individual target and inhibited.
Thus, the therapies become more specific and targeted. In some cases, pathogens such as viruses and bacteria have the ability to manipulate the ubiquitin system to their  (42)(43)(44)(45). All these findings suggest that OTU DUBs in the ubiquitin-proteasome pathway are good targets for the treatment of viral and parasitic diseases.