Upon Plasmodium falciparum merozoites exposure to low [K+] environment in blood plasma, there is escalation of cytosolic [Ca2+] which activates Ca2+-Dependent Protein Kinase 1 (CDPK1), a signaling hub of intra-erythrocytic proliferative stages of parasite. Given its high abundance and multidimensional attributes in parasite life-cycle, this is a lucrative target for designing antimalarials. Towards this, we have virtually screened MyriaScreenII diversity collection of 10,000 drug-like molecules, which resulted in 18 compounds complementing ATP-binding pocket of CDPK1. In vitro screening for toxicity in mammalian cells revealed that these compounds are non-toxic in nature. Furthermore, SPR analysis demonstrated differential binding affinity of these compounds towards recombinantly purified CDPK1 protein. Selection of lead compound 1 was performed by evaluating their inhibitory effects on phosphorylation and ATP binding activities of CDPK1. Furthermore, in vitro biophysical evaluations by ITC and FS revealed that binding of compound 1 is driven by formation of energetically favorable non-covalent interactions, with different binding constants in presence and absence of Ca2+, and TSA authenticated stability of compound 1 bound CDPK1 complex. Finally, compound 1 strongly inhibited intra-erythrocytic growth of P. falciparum in vitro. Conceivably, we propose a novel CDPK1-selective inhibitor, step towards developing pan-CDPK kinase inhibitors, prerequisite for cross-stage anti-malarial protection.

Introduction

Plasmodium falciparum is the most prevalent malaria parasite which accounted for a majority of the total estimated malaria cases in WHO African Region (99.7%), the Western Pacific (71.9%), the Eastern Mediterranean (69%) and WHO region of Southeast Asia (62.8%) (statistics are as per 2017) [1]. Five countries in WHO sub-Saharan African Region and Republic of India have been reported to carry almost half of the global malaria burden. Children aged under 5 years constitute the most vulnerable group of the population which accounted for almost 61% of all malaria deaths worldwide in 2017 [1]. Clinical manifestations of malaria largely result from an exponential expansion of the parasite in blood wherein parasites invade into, multiply within and evade from erythrocytes for their propagation. Currently, there is no drug regime available that can target all the intra-erythrocytic proliferation stages of the parasite.

In this process of parasitic growth, exposure of P. falciparum merozoites from high (host cytosolic; 140 mM) to low (blood plasmic; 5.4 mM) potassium-ion (K+) environment triggers phospholipase-C mediated intracellular Ca2+ release from endoplasmic reticulum [2]. This escalation of cytosolic Ca2+ activates Calcium-Dependent Protein Kinases (CDPKs) of the parasite. In the genus Plasmodium, CDPK proteins are present as a multigene family comprised of seven members and each gene is predominantly expressed in a distinct phase of the parasite life cycle. For example, CDPK3 is expressed in ookinetes and P. berghei parasites without a functional cdpk3 gene show reduced ability to invade midgut of the mosquito host, while simultaneously showing reduction in the gliding motility phenotype [3,4]. Similarly, PbCDPK4 regulates sexual differentiation into male gametocytes and therefore, being crucial for transmission to the mosquito host and sexual reproduction [5]. Likewise, PfCDPK5 guides egress of merozoites from the host erythrocytes [6]. However, besides these members of cdpk multigene family [7,8], one of the most widely studied member is CDPK1, which is a signaling hub responsible for multistage cellular processes involved in schizogony, merozoite invasion, gametogenesis, and transmission to the mosquito host [9–13].

Mechanistic studies revealed that Protein Kinase G (PKG)-mediated phosphorylation of CDPK1 allows its localization to apical structures [14]. Another piece of study also depicted that CDPK1 regulates actin-myosin motor during gliding and invasion of the parasite by phosphorylating a 50-kDa Glideosome-Associated Protein (GAP-45) and Myosin Tail domain Interacting Protein (MTIP), two important components of Inner Membrane Complex (IMC) of P falciparum merozoites. Along with the proteins of IMC, it also phosphorylates regulatory subunit of cAMP-dependent Protein Kinase A (PKAr), thereby regulating cAMP-mediated signaling cascade module [13]. With this understanding of an indispensable role of CDPK1 in the parasite life cycle, we pursued this as a potential target for development of novel specific inhibitors that can block parasite growth targeting multiple stages in malarial infection. In this study, we have identified small molecule ligands that bind to and inhibit the functional (kinase) activity of CDPK1 in vitro and parasite growth ex vivo, by combining high-throughput virtual screening strategy, followed by experimental evaluation by adopting several biochemical and biophysical assays in vitro. This study can in principle be extended towards developing novel drug candidates against any eukaryotic parasite by systematic high-throughput screening of small molecule-based drug-like libraries.

Materials and methods

Culture of P. falciparum 3D7 and HepG2

P. falciparum strain 3D7 was cultured according to the previously described protocol [15]. Briefly, parasite culture was maintained in O+ erythrocytes at 4% hematocrit in RPMI 1640 medium (Gibco®, U.S.A.), supplemented with 0.5% AlbuMAXTM I (Gibco®, U.S.A.), 50 mg/L hypoxanthine, 10 mg/L gentamycin (Gibco®, U.S.A.) and 2 gm/L sodium bicarbonate. Culture was maintained in ambient hypoxic environment (5% O2, 5% CO2, balanced with nitrogen).

The human hepatocellular carcinoma cell line, HepG2 was cultured in Dulbecco's Modified Eagle's Medium (Gibco®, U.S.A.) supplemented with 10% fetal bovine serum (Gibco®, U.S.A.) and penicillin-streptomycin (100 units/ml and 100 mg/ml, respectively) solution. Culture was maintained at 37°C in a humidified atmosphere containing 5% CO2.

Purification of 6xHis-CDPK1, homology modeling and virtual screening

Cloning, overexpression and purification of 6xHis-CDPK1 was done as described previously [11]. To raise anti-sera, mice were immunized by injecting the emulsified protein. Comparative or homology modeling was employed to model 3D structure of CDPK1 from P. falciparum strain 3D7. To accomplish this, amino acid sequence of PfCDPK1 protein (PF3D7_0217500) was retrieved from PlasmoDB database [16]. To search for a suitable template for homology modeling, BLASTp search was performed using amino acid sequence of PfCDPK1 as query sequence, against Protein Data Bank (PDB) database [17,18]. Amino acid sequence identity between CDPK1 orthologs from P. falciparum 3D7 and P. berghei strain ANKA was found to be 93%, thus rendering X-Ray diffraction based structural model of PbCDPK1 (PDB ID: 3Q5I; resolution: 2.1 Å) as a suitable template to model 3D structure of PfCDPK1 [19]. Homology modeling was done by using Modeler v9.17 software [20]. Generated model was further subjected to structural refinement by using ModRefiner [21]. Reliability of the refined structural model of CDPK1 was assessed by examining backbone dihedral angles: phi (Ø) and psi (Ψ) of the amino acid residues lying in the energetically favorable regions of Ramachandran space [22]. This was done by using PROCHECK v.3.5 [23]. Percentage quality assessment of the protein structure was evaluated by utilizing four sorts of occupancies: core, additionally allowed, generously allowed and disallowed regions. The refined 3D structural model of PfCDPK1, thus generated, was subsequently used for docking studies.

Novel small molecule inhibitors of CDPK1 were identified by virtual screening of MyriaScreen II diversity collection of 10,000 drug-like screening compounds [24]. Structural Data Format (SDF) file of Myriascreen II library was requested from Sigma–Aldrich. Ligands were subjected to virtual screening against 3D structural model of CDPK1 using Autodock Vina Tools v1.5.6 [25]. As per the already established catalytically active ATP binding pocket in PfCDPK1, we ensured that the active site residues were covered while constructing the virtual 3D-grid for docking based virtual screening. Centre and volume of the grid was set to cover the entire ligand (ATP) binding pocket of the protein, with x, y, z co-ordinates of the center of energy, −1.534, −18.536, −20.028, respectively, through Autogrid module of AutoDock Tools, with default spacing. Top ranked compounds were selected based on the highest (most negative) binding affinity to the catalytically active site (ATP binding pocket) of CDPK1. Finally, compounds binding to the ATP binding pocket of CDPK1 were further filtered on the basis of their propensity to follow Lipinski's rule by using Drug Likeness Tool (DruLiTo) [26]. The PyMOL Molecular Graphics System, v2.1 by Schrödinger, LLC was used for molecular visualization of CDPK1 structural model and AutoDock Vina results [27].

Evaluation of cytotoxicity in HepG2 cell line

Cytotoxic effect on HepG2 was assessed by utilizing ability of live cells to cleave MTT [(3-[4,5- dimethylthiazol-2-yl]-2,5-diphenyl tetrazolium bromide)] into blue colored formazan crystals. Briefly, HepG2 cells were seeded in a 96-well microtiter plate at a density of 10,000 cells per 100 µl per well. Cells were allowed to proliferate at 37°C for 48 h in the presence of compounds at 100 µM concentration. Cells cultured in the absence of any compound were maintained as positive control. After treatment, MTT labeling reagent was added to each well to a final concentration of 0.5 mg/ml and incubated at 37°C for 4 h. Purple colored formazan crystals, thus formed, were dissolved in 100 µl DMSO solvent and optical density was measured spectrophotometrically by taking absorbance at 570 nm using Varioskan™ LUX multimode microplate reader (Thermo Fisher Scientific™). The absorbance of untreated cells was taken as 100%. The experiment was done twice.

Surface plasmon resonance (SPR) analysis of binding affinity between compounds and CDPK1

To determine binding strength of compounds (N = 18) procured from virtual screening of MyriaScreen II library, real-time biomolecular interaction analysis with SPR was carried out at physiologically relevant concentrations by using AutoLab Esprit SPR. Kinetic rate constants: Ka and Kd (association and dissociation rate constants, respectively) as well as affinity constant, KD were measured at room temperature (RT, 293 K). SPR analysis was performed by following previously described protocols [28]. Briefly, interaction kinetics was studied by injecting compounds at different concentrations: 100, 75, 50, 25 and 12.5 µM over the CDPK1 (7.43 ng per mm2) immobilized SPR sensor chip surface, followed by comparing their respective kinetics & binding affinities at RT. A control flow cell was activated and blocked in a similar manner with injection of equivalent concentrations of DMSO to allow for reference subtraction. HEPES-NaOH buffer was used both as immobilization and binding solutions. Data were fit to the two-state conformational change model using AutoLab SPR Kinetic Evaluation software provided with the instrument. KD value was calculated using the Integrated Rate Law (IRL) equation. Two independent experiments were performed.

In vitro kinase assay with CDPK1

Functional activity of 6xHis-CDPK1 was monitored by utilizing ADP-Glo™ Kinase Assay (Promega Corporation), which is ATP regeneration based luciferase reaction system resulting from nascent ADP phosphorylation. The luminescence signal, thus generated, is proportional to the amount of ADP released in a given kinase reaction. Phosphorylation experiments were performed with 100 ng of 6xHis-CDPK1, per reaction, in assay buffer (100 mM Tris–Cl, pH 7.4; 2.5 mM DTT; 50 mM MgCl2 and 2.5 mM MnCl2), by following previously described protocol [29]. Enzymatic reaction was carried out in the absence and presence of calcium ions. An amount of 10 µg of dephosphorylated casein from bovine milk, per reaction, was used as exogenous substrate for the enzyme. 2.5 mM EGTA was added for conditions requiring absence of Ca2+ ions. Kinase reactions were initiated by adding 1 µM ATP and allowed to take place at 30°C for 1 h. To test for any inhibitory effect of compounds (N = 18) screened from MyriaScreen II Diversity Collection, CDPK1 was allowed to pre-incubate in the reaction buffer with 50 µM of each compound at room temperature for 1 h, followed by initiating the reactions by adding ATP. Upon completion of the reactions, ADP-Glo™ Kinase Assay was performed. Luminescence signal, thus generated, was measured with Lumat³ LB 9508 Ultra- Sensitive Tube Luminometer (Berthold Technologies, U.S.A.). Kinase activity of CDPK1 was further assessed in the presence of different concentrations of lead compounds: ST092793 (assigned as compound 1) and S344699 (assigned as compound 2) to evaluate their respective IC50 values. All experiments were done at least twice.

Phosphorylation status of CDPK1 and casein in presence of the lead compounds 1(ST092793) and 2(S344699) was checked by in-gel fluorescence staining with phosphoamino acid stain, Pro-Q Diamond (Thermo Fisher Scientific™). Protocol was followed as recommended by the supplier. Briefly, SDS–PAGE was performed for the phosphorylation reactions carried out in the presence of different concentrations: 100, 50, 25 and 12.5 µM of lead compounds 1(ST092793) and 2(S344699). Phosphostained gels were imaged using Typhoon™ FLA 9500 biomolecular imager (GE Healthcare Life Sciences). Excitation was set at 555 nm with emission filter of 580 BP 30. After visualization, gel(s) was stained with Coomassie Brilliant blue G-250 dye to confirm equal loading of CDPK1 and dephosphorylated casein in all the lanes. Dried coomassie gels were scanned on Epson CX5400 high resolution scanner. Estimates of band intensities of CDPK1 and casein from phosphostained gel images were made using ImageJ 1.52a, which is a public domain Java image processing program. Lanes and bands were auto-detected and manually size-adjusted to ensure full band inclusion and equivalent thickness across all lanes.

Competitive ATP affinity chromatography assay

To confirm if the lead compounds 1(ST092793) and 2(S344699) inhibit functional activity of CDPK1 by directly binding to its active site (ATP-binding pocket), affinity chromatography of CDPK1 pre-incubated with the compounds was performed by utilizing ATP Affinity Test Kit (Jena Bioscience). Protocol was followed as recommended by the manufacturer. Briefly, 10 µg of purified CDPK1 protein per reaction was incubated with 100 µM of compound 1(ST092793) and/or 2(S344699) in binding solution for 1 h at RT. This interaction was competed with 8AH-ATP-agarose {8-[(6-Amino) Hexyl]-amino-ATP-agarose} in which ATP is immobilized on the agarose matrix via the adenine base through C8 linker. Blank agarose served as negative control. Elution fractions were analyzed by SDS–PAGE and immunobloting by probing recombinant CDPK1 with monoclonal poly-Histidine-peroxidase antibody. CDPK1 present in flowthroughs of control and test samples served as loading controls for normalization. Signals were detected with SuperSignal™ West Femto Maximum Sensitivity Substrate (Thermo Fisher Scientific™). The experiment was done twice.

Isothermal titration calorimetry (ITC) and steady-state fluorescence spectroscopy

To calculate kinetic parameters such as binding affinity constant, Ka for the interaction of CDPK1 with relatively more potent inhibitor (i.e. compound 1(ST092793)), ITC experiments were performed by using MicroCal iTC200 (Malvern Instruments Ltd, U.K.). For this purpose, recombinant CDPK1 protein was dialyzed extensively against HEPES-NaCl buffer (10 mM HEPES-NaOH, pH 7.4 and 150 mM NaCl) using AmiconTM Ultra-15 Centrifugal Filter Unit (10 kDa cutoff) before its subsequent use in ITC. Dilutions of CDPK1 and compound 1(ST092793) were prepared in the same buffer to ignore contribution from buffer-buffer interaction, and degassed by vacuum for 10 min prior to use. ITC analysis was done in the absence and presence of 2.5 mM CaCl2 at RT, by following previously described protocols [30]. Briefly, syringe was loaded with 100 µM of the ligand (i.e. compound 1(ST092793)), and sample cell was filled with 10 µM of CDPK1. Background titration profiles were obtained by injecting ligand into the buffer under identical experimental conditions. The experiment was done twice. Amount of heat produced per injection (corrected for heat of dilution of the ligand) was analyzed by integration of area under individual peaks by MicroCal ORIGIN 7 software provided by the instrument manufacturer. Experimental data was presented as the corrected amount of heat produced per second (µcal/sec) following each injection of the ligand into the protein solution as a function of time (minutes). The non-linear experimental data was fit to the single-site binding model yielding molar binding stoichiometry (N), binding constant (Ka), enthalpy change (ΔH), entropy change (ΔS) and Gibbs free energy (ΔG).

To confirm the results obtained with ITC, fluorescence spectroscopic analysis was performed by using Cary Eclipse Fluorescence Spectrophotometer (Agilent Technologies, US). All experiments were done at RT using a quartz cell with 1 cm path length. Dilution of CDPK1 was prepared in HEPES-NaCl buffer, in the absence and presence of calcium ions. Intrinsic fluorescence emission spectra were recorded for free (unbound) CDPK1 protein (5 µM) as well as upon its complexation with increasing concentrations of compound 1(ST092793) (0.5 µM to 29 µM, in the absence of calcium; and, 1 µM to 38 µM in the presence of calcium). The emission spectra for CDPK1 were recorded in the range of 300–450 nm after excitation at 295 nm to selectively excite tryptophan residues of the protein. Excitation and emission slits were set at 10 nm. All spectra were corrected by subtracting the corresponding buffer baseline, obtained under similar conditions. The experiment was done twice. Data obtained from CDPK1 quenching experiments were fitted using Kaleidagraph 4.0.

Thermal stability assay (TSA)

Thermostability of CDPK1 was monitored for the detection of binding events between CDPK1 and Compound 1(ST092793) by following the protocol as described earlier [31]. Briefly, 10 µg of the purified recombinant protein per reaction (in HEPES-NaCl buffer, pH 7.5; 30 µl) was incubated with 100 µM of the compound, in the absence and presence of 2.5 mM CaCl2. Protein in control reactions was incubated with an equivalent volume of DMSO. Following incubation, reaction mixtures were heated on Mastercycler™ Nexus Thermal Cycler (Eppendorf™) at different temperatures ranging from 40°C to 90°C for 6 min and then cooled down to room temperature for 4 min. Following centrifugation at 16 800 g for 40 min at 4°C, SDS–PAGE of supernatants was performed. Estimates of band intensities of CDPK1 protein from Coomassie stained gel images were made using ImageJ 1.52a. For analysis of melting shift, CDPK1 signal intensity was normalized to the respective intensity at 40°C and quantified protein levels were analyzed. The experiment was done twice.

Evaluation of growth inhibition of malaria parasite

Lead compound 1(ST092793) and compound 2(S344699) with CDPK1-inhibitory activity were subjected to dose-dependent cytotoxic evaluation on parasite growth at concentrations ranging from 1.56 to 100 µM. Briefly, trophozoites at 1% initial parasitaemia were treated with the compounds for one complete intra-erythrocytic life cycle of the parasite. Untreated parasites served as control. Post incubation, parasites were stained with Ethidium Bromide (10 µM), followed by flow cytometry on BD LSRFortessa™ cell analyzer using FlowJo v10 software. Fluorescence signal (FL-2) was detected with the 590 nm band pass filter by using an excitation laser of 488 nm collecting 100,000 cells per sample. Following acquisition, parasitaemia levels were estimated by determining the proportion of FL-2 positive cells using Cell Quest. Percent parasite growth inhibition was calculated as follows: % Parasite growth inhibition = [1–(% Parasitaemia (treatment)/% Parasitaemia (Control)] × 100. IC50 value was calculated by using Origin® 2018b Graphing and Analysis Software (OriginLab Corporation). The experiment was done twice.

Statistical analysis

In the bar graphs, data are expressed as the mean ± standard deviation (s.d.) of three independent experiments done in duplicates. Unless stated, data analysis of the assay plots and calculation of biochemical parameters were done by using both Microsoft Excel and Origin® 2018b Graphing and Analysis Software (OriginLab Corporation) were used. Unless indicated, the differences were considered to be statistically significant at P < 0.05.

Results

Purification of 6xHis-CDPK1, three-dimensional structural model of PfCDPK1 and its quality assessment

Coomassie-stained gel shows the level of purity of recombinant PfCDPK1 used for all assays (Figure 1a). Anti-sera raised in mice was used to probe recombinantly purified CDPK1 protein and native protein from schizonts lysate (Figure 1a). PfCDPK1 contains an N-terminal KD tethered to a C-terminal CamLD via JD (Figure 1b). Three-dimensional structure of a protein can provide us with precise information about its single, most stable conformation, as dictated by its sequence. X-Ray diffraction based structural model of PbCDPK1 was used as a template to generate 3D co-ordinates of PfCDPK1 by homology modeling, one of the most common structure prediction methods in structural genomics and proteomics (Supplementary Figure S1a). After optimal rigid-body superimposition of PbCDPK1 with the generated structural model of PfCDPK1, overall Root-Mean-Square Deviation (RMSD) value of the C-alpha atomic co-ordinates was found to be 0.31 Å, suggesting a reliable 3D structure of PfCDPK1 (Supplementary Figure S1b). Assessment of stereochemical quality and accuracy of the generated homology model displayed 89.2% of amino acid residues lying in the most favored (‘core') regions, with 8.1%, 1.9%, and 0.8% residues in ‘additional allowed', ‘generously allowed' and ‘disallowed regions' of Ramachandran plot, respectively (Supplementary Figure S1c). Also, protein structure with ≥90% of its amino acid residues lying in the most favored regions of Ramachandran plot is considered to be as accurate as a crystal structure at 2 Å-resolution. This indicated that the backbone dihedral angles: phi and psi of the generated PfCDPK1 model were reasonably accurate. The comparable Ramachandran plot characteristics and RMSD value confirmed the reliability of the 3D-model of PfCDPK1 to be taken further for virtual screening.

Purified 6xHis-PfCDPK1 protein, three-dimensional structural model of PfCDPK1, virtual screening and cytotoxic potential of candidate compounds against HepG2 cell line.

Figure 1.
Purified 6xHis-PfCDPK1 protein, three-dimensional structural model of PfCDPK1, virtual screening and cytotoxic potential of candidate compounds against HepG2 cell line.

(a) Coomassie-stained gel showing the level of purity of recombinant PfCDPK1 used for all assays. Anti-sera raised in mice was used to probe recombinantly purified CDPK1 protein and native protein from schizonts lysate. (b) Domain architecture of PfCDPK1. It contains an N-terminal catalytic Kinase Domain (KD, shown in red) tethered to a C-terminal Calmodulin-Like regulatory Domain (CamLD, shown in blue) via auto-regulatory or Junction domain. Activation loop within the KD is shown in yellow. (c) Catalytically active ATP binding pocket. X-Ray diffraction based crystal structure of PbCDPK1 (PDB ID: 3Q5I) was used as template to generate 3D structure of PfCDPK1. (d) Virtual screening of MyriaScreen II library against ATP binding pocket of PfCDPK1. Compounds were ranked on the basis of their negative binding energies along the Y-abscissa. Additional filter was applied to screen for ligands with propensity to follow Lipinski's rule of five. 18 compounds were finally selected as probable hits and set to submit for experimental validation. (e) Cytotoxic potential of the candidate compounds (N = 18) obtained from virtual screening of MyriaScreen II library, on HepG2 cell line. At 100 μM concentration, all 18 compounds displayed negligible cytotoxicity upon 48 h of treatment. The experiment was done twice.

Figure 1.
Purified 6xHis-PfCDPK1 protein, three-dimensional structural model of PfCDPK1, virtual screening and cytotoxic potential of candidate compounds against HepG2 cell line.

(a) Coomassie-stained gel showing the level of purity of recombinant PfCDPK1 used for all assays. Anti-sera raised in mice was used to probe recombinantly purified CDPK1 protein and native protein from schizonts lysate. (b) Domain architecture of PfCDPK1. It contains an N-terminal catalytic Kinase Domain (KD, shown in red) tethered to a C-terminal Calmodulin-Like regulatory Domain (CamLD, shown in blue) via auto-regulatory or Junction domain. Activation loop within the KD is shown in yellow. (c) Catalytically active ATP binding pocket. X-Ray diffraction based crystal structure of PbCDPK1 (PDB ID: 3Q5I) was used as template to generate 3D structure of PfCDPK1. (d) Virtual screening of MyriaScreen II library against ATP binding pocket of PfCDPK1. Compounds were ranked on the basis of their negative binding energies along the Y-abscissa. Additional filter was applied to screen for ligands with propensity to follow Lipinski's rule of five. 18 compounds were finally selected as probable hits and set to submit for experimental validation. (e) Cytotoxic potential of the candidate compounds (N = 18) obtained from virtual screening of MyriaScreen II library, on HepG2 cell line. At 100 μM concentration, all 18 compounds displayed negligible cytotoxicity upon 48 h of treatment. The experiment was done twice.

Virtual screening of MyriaScreen II library against ATP binding pocket of CDPK1

MyriaScreen II diversity collection of drug-like screening compounds consists of 10,000 high-purity and diverse molecular candidates from Sigma–Aldrich and TimTec, Inc., created by combining medicinal chemistry and computational expertize, and selected on the basis of diversity and structural relevance for general screening. The collection is largely drug-like according to Lipinski's rule of five. Fast docking of MyriaScreen II library against the generated homology model of CDPK1 was done to computationally screen for ligands with propensity to bind catalytically active ATP binding pocket of the protein (Figure 1c). Compounds were arranged on the basis of their most negative free binding energies or binding affinity to the ATP binding pocket of the protein, and subsequently ranked by decreasing value of their negative binding energies along the abscissa (Figure 1d). Compounds were further filtered out on the basis of their propensity to follow Lipinski's rule of five pertaining to physicochemical properties including Molecular Weight (MW), Octanol/water partition coefficient (LogP), Hydrogen bond acceptors (H-acceptor) and Hydrogen bond donors (H-donors). 18 compounds were finally selected as probable hits and set to submit for experimental validation (Figure 1d). Besides, all filtered out compounds (N = 18) are stabilized by the formation of energetically favorable molecular interactions with the ATP binding pocket of CDPK1. Their binding energies, Lipinski's properties and chemical structures are shown in Supplementary Table S1. Supplementary Figure S2 shows predicted binding poses of the candidate compounds in the ATP binding pocket of CDPK1. Amino acid residues constituting active site of the protein, which interact with the compounds through polar contacts (H-bonds) are highlighted, alongwith the H-bond lengths.

Cytotoxic potential of the candidate compounds (N = 18) against HepG2 cell line

We further investigated their toxicity on human hepatocellular carcinoma cell line, HepG2 by using MTT cell-viability assay, as mentioned in materials and methods section. At 100 µM concentration, all 18 compounds displayed negligible toxicity on HepG2 cells upon 48 h of treatment, with treated cells as healthy as control ones (Figure 1e).

Compounds exhibit divergent binding affinities for CDPK1

Compounds (N = 18) filtered out from virtual screening of MyriaScreen II library were checked for their interaction with 6xHis-CDPK1 in vitro by utilizing SPR. Once immobilized on the SPR sensor chip surface, CDPK1 demonstrated good stability throughout the experiment. With increase in mass concentration of the compounds, gradual increase in SPR sensor signal was observed which linearly correlated with corresponding change in refractive index of the medium immediately adjacent to the SPR sensing surface. The concentration dependent real-time sensorgrams for the interactions, alongwith KD values of each compound are shown in Figure 2a. The hits displayed a wide range of binding affinities for CDPK1, with KD values varying from 5.4 × 1004 ± 5.8 × 1004 μM (ST018363) to 41.2 ± 0.13 μM (ST092794). Notably, interaction with ST092793 produced maximum Response Unit (RU) of ∼300.5 at the end of association phase (corresponding to injection of 100 μM concentration of the compound) with KD value of 51.9 ± 0.3 μM. Relative binding affinities of each compound for CDPK1 are depicted in Figure 2b. Sensor signals obtained from the biomolecular interaction analysis were further transformed into polynomial calibration curves to show relationship between Response at steady state (i.e. at equilibrium, denoted as Req) and concentrations of the compounds (data not shown).

Compounds exhibit divergent binding affinities for CDPK1.

Figure 2.
Compounds exhibit divergent binding affinities for CDPK1.

(a) Concentration dependent real-time sensograms for SPR based biomolecular interaction analysis of CDPK1 and compounds. Candidate compounds procured from virtual screening of MyriaScreen II library displayed a wide range of binding affinities for CDPK1, with KD values ranging from ∼5.4 mM (ST018363) to 41.2 µM (ST092794). Interaction with ST092793 (assigned as compound 1) produced maximum RU of ∼300.5, with KD value of 51.9 µM. (b) Relative binding affinities of each compound for CDPK1 are shown.

Figure 2.
Compounds exhibit divergent binding affinities for CDPK1.

(a) Concentration dependent real-time sensograms for SPR based biomolecular interaction analysis of CDPK1 and compounds. Candidate compounds procured from virtual screening of MyriaScreen II library displayed a wide range of binding affinities for CDPK1, with KD values ranging from ∼5.4 mM (ST018363) to 41.2 µM (ST092794). Interaction with ST092793 (assigned as compound 1) produced maximum RU of ∼300.5, with KD value of 51.9 µM. (b) Relative binding affinities of each compound for CDPK1 are shown.

SPR analysis suggests that most of the compounds filtered out from virtual screening of MyriaScreen II library are capable of interacting with CDPK1, indicating their probable tendency to inhibit kinase activity of the protein. To verify this inference, the compounds were further examined for their inhibitory effect (if any) on CDPK1 functional activity assays in vitro and parasite growth ex vivo.

Compounds 1(ST092793) and 2(S344699) inhibit functional activity of CDPK1 in vitro

Since MyriaScreen II library of drug-like small molecular candidates was virtually screened on the basis of probable propensity of the compounds to interact with ATP binding pocket of CDPK1, we utilized in vitro kinase assay to demonstrate enzymatic activity of the protein in presence of the filtered out candidates by measuring amount of ADP produced in the kinase reactions. All 18 compounds taken into consideration showed differential inhibition of CDPK1 activity which supports well with our in silico virtual screening approach and SPR interaction analysis (Figure 3a). Notably, two compounds: ST092793 (assigned as compound 1) and S344699 (assigned as compound 2) significantly inhibited phosphorylation activity of CDPK1, accounting for 87.34% ± 0.15% and 68.05% ± 0.48% inhibition, respectively. DMSO, used as solvent for compounds, had no inhibitory effect on the enzymatic reactions and was taken as negative control. Dose-response curves for inhibitory activity of compounds 1(ST092793) and 2(S344699) against CDPK1 protein were found to be sigmoidal, depicting their IC50 values to be 33.8 µM and 42.6 µM, respectively (Figure 3b). Taken together, these results suggest compounds 1(ST092793) & 2(S344699) as inhibitors of CDPK1 enzymatic activity.

ST092793 (compound 1) and S344699 (compound 2) inhibit functional activity of CDPK1 in vitro.

Figure 3.
ST092793 (compound 1) and S344699 (compound 2) inhibit functional activity of CDPK1 in vitro.

(a) All 18 compounds showed differential inhibition in CDPK1 activity. Compounds 1(ST092793) & 2(S344699) significantly inhibited kinase activity, accounting for 87.34% ± 0.15% and 68.05% ± 0.48% inhibition, respectively. DMSO, used as solvent for the compounds, had no inhibitory effect. Enzymatic activity of CDPK1 in control reaction was taken as 100%. (b) Dose response curves for inhibitory activity of the lead compounds. Compounds 1(ST092793) & 2(S344699) exhibited IC50 values of 33.8 µM and 42.6 µM, respectively. Kinase reactions were performed twice. (c) Lead compounds inhibit auto- and trans-phosphorylation status of CDPK1 and dephosphorylated casein, respectively. Phosphostaining for CDPK1 enzymatic activity in the presence of 100 & 50 μM concentrations of lead compounds 1(ST092793) and 2(S344699) (i). Phosphostaining was further confirmed using a range of concentrations (i.e, from 100 to 12.5 µM) of both compounds (ii). Estimates of band intensities of CDPK1 and casein from phosphostained gel images were made by using ImageJ 1.52a. Band intensities in reactions with Ca2+ were taken as 100%.

Figure 3.
ST092793 (compound 1) and S344699 (compound 2) inhibit functional activity of CDPK1 in vitro.

(a) All 18 compounds showed differential inhibition in CDPK1 activity. Compounds 1(ST092793) & 2(S344699) significantly inhibited kinase activity, accounting for 87.34% ± 0.15% and 68.05% ± 0.48% inhibition, respectively. DMSO, used as solvent for the compounds, had no inhibitory effect. Enzymatic activity of CDPK1 in control reaction was taken as 100%. (b) Dose response curves for inhibitory activity of the lead compounds. Compounds 1(ST092793) & 2(S344699) exhibited IC50 values of 33.8 µM and 42.6 µM, respectively. Kinase reactions were performed twice. (c) Lead compounds inhibit auto- and trans-phosphorylation status of CDPK1 and dephosphorylated casein, respectively. Phosphostaining for CDPK1 enzymatic activity in the presence of 100 & 50 μM concentrations of lead compounds 1(ST092793) and 2(S344699) (i). Phosphostaining was further confirmed using a range of concentrations (i.e, from 100 to 12.5 µM) of both compounds (ii). Estimates of band intensities of CDPK1 and casein from phosphostained gel images were made by using ImageJ 1.52a. Band intensities in reactions with Ca2+ were taken as 100%.

Auto-phosphorylation & trans-phosphorylation status of CDPK1 and dephosphorylated casein, respectively, were checked in the absence and presence of varying concentrations of compounds 1(ST092793) and 2(S344699). Figure 3c(i) shows staining for enzymatic activity, in the presence of 100 and 50 µM concentrations of lead compounds 1(ST092793) and 2(S344699), with Pro-Q Diamond phosphoamino acid and Coomassie Brilliant Blue G-250 stains. The phosphostaining for CDPK1 enzymatic activity was further confirmed using a range of concentrations (i.e, from 100 to 12.5 µM) of both compunds (Figure 3c(ii)). In both cases, decrease in phosphorylation status of both CDPK1 and casein was readily detectable with corresponding increase in concentrations of the compounds, further reaffirming the observation that compounds 1(ST092793) and 2(S344699) are inhibitors of CDPK1 activity. We attribute minor differences in apparent size to the gel shift resulting from multiple phosphorylation sites on CDPK1 protein.

Compounds 1(ST092793) and 2(S344699) inhibit CDPK1 activity by interacting with its ATP binding pocket

Since compounds 1(ST092793) and 2(S344699) inhibit kinase activity of CDPK1 in vitro, we wanted to confirm if they accomplish this by directly binding to catalytically active ATP binding pocket of the protein. For this purpose, recombinantly purified CDPK1 was allowed to interact with compound 1(ST092793) and/or 2(S344699), followed by competing this interaction with ATP immobilized on the surface of agarose matrix. Reduction in binding with the ATP-conjugated agarose matrix was observed in the presence of compound 1(ST092793), compound 2(S344699) and in combination of both compounds (Figure 4a). It was therefore estimated that CDPK1 in complexation with the compound(s) possessed diminished ability to bind ATP, indicating that compounds 1(ST092793) and 2(S344699) inhibit in vitro kinase activity of CDPK1 by directly binding to its ATP-binding pocket. Schematic representation of complex formation between catalytically active ATP binding pocket of CDPK1 and compounds 1(ST092793) & 2(S344699) is shown in Figure 4b. Y148 and K85 of CDPK1 were found to interact with compound 1(ST092793) and compound 2(S344699), respectively via polar interactions (H-bonds, indicated by dashed lines) with bond lengths, 2.9 Å and 3.4 Å, respectively.

Compounds 1(ST092793) & 2(S344699) inhibit CDPK1 activity by directly interacting with its ATP binding pocket.

Figure 4.
Compounds 1(ST092793) & 2(S344699) inhibit CDPK1 activity by directly interacting with its ATP binding pocket.

(a) Competitive ATP affinity chromatography assay. Reduction in binding with the ATP-conjugated agarose matrix was observed upon complexation of CDPK1 with compound 1(ST092793), compound 2(S344699) and in combination of both. The experiment was done twice. (b) Schematic representation of complex formation between catalytically active ATP binding pocket of CDPK1 and compounds 1(ST092793) & 2(S344699). Y148 and K85 of CDPK1 were found to interact with compound 1(ST092793) (green) and compound 2(S344699) (blue), respectively via polar interactions (H-bonds, indicated by dashed lines) with bond lengths, 2.9 Å and 3.4 Å, respectively.

Figure 4.
Compounds 1(ST092793) & 2(S344699) inhibit CDPK1 activity by directly interacting with its ATP binding pocket.

(a) Competitive ATP affinity chromatography assay. Reduction in binding with the ATP-conjugated agarose matrix was observed upon complexation of CDPK1 with compound 1(ST092793), compound 2(S344699) and in combination of both. The experiment was done twice. (b) Schematic representation of complex formation between catalytically active ATP binding pocket of CDPK1 and compounds 1(ST092793) & 2(S344699). Y148 and K85 of CDPK1 were found to interact with compound 1(ST092793) (green) and compound 2(S344699) (blue), respectively via polar interactions (H-bonds, indicated by dashed lines) with bond lengths, 2.9 Å and 3.4 Å, respectively.

Compound 1(ST092793) interacts with CDPK1 favorably in the absence of calcium

After affirming CDPK1-inhibitory activity of compounds 1(ST092793) and 2(S344699)in vitro, we further sought to establish the knowledge of stoichiometry details and binding modes of complexation between CDPK1 and the relatively more potent inhibitor, i.e. compound 1(ST092793), by employing ITC. The representative binding isotherms resulting from titration of compound 1(ST092793) with CDPK1 are represented in Figure 5a. The compound showed differential binding with CDPK1 in the absence and presence of Ca2+. Binding isotherm in the absence of Ca2+ was monophasic in nature, reaching a plateau phase at 1 : 1 stoichiometry (∼1.0 molar ratio) which indicates that compound 1(ST092793) interacts with only a single independent binding site (i.e. catalytically active ATP binding pocket) of CDPK1 protein. As further injections continued, gradual decline in exothermic heat resulted in sigmoidal curve ending near zero baseline, indicating saturation of binding sites. On the basis of the nature of the curve, the data were fitted using single-site binding model. In contrast, linear thermogram was observed in the presence of Ca2+, indicative of non-specific binding of compound 1(ST092793) with CDPK1. In Figure 5a, the solid line shows the best fit, and the model reproduces experimental data fairly well. Ka values demonstrated that compound 1(ST092793) showed ∼100 times stronger binding affinity towards CDPK1 in the absence of Ca2+ (Ka = 3.6 × 105 ± 2.4 × 105 M−1) than in its presence (Ka = 2.6 × 103 ± 25% M−1), with the interactions mainly driven by enthalpy: ΔH = −8.1 ± 5.2 kcal mol−1, TΔS = −0.53 kcal mol−1 (±15%–20%) and ΔG = −7.57 kcal mol−1 (±15%–20%). Linear thermogram was obtained in the presence of Ca2+, indicative of non-specific and very weak binding of compound 1 with CDPK1.

Compound 1(ST092793) interacts with CDPK1 favorably in the absence of calcium.

Figure 5.
Compound 1(ST092793) interacts with CDPK1 favorably in the absence of calcium.

(a) Representative ITC binding isotherms resulting from titrations of CDPK1 with compound 1(ST092793). The compound showed differential binding with CDPK1 in the absence (Ka = 3.6 × 105 ± 2.4 × 105 M−1) and presence (Ka = 2.6 × 103 ± 25% M−1) of Ca2+. Binding isotherm in the absence of Ca2+ reached a plateau phase at 1 : 1 stoichiometry indicating that compound 1(ST092793) interacts with only a single independent binding site of CDPK1. Also, interaction of compound 1(ST092793) with CDPK1 was enthalpically driven with ΔH = −8.1 ± 5.2 kcal mol−1, TΔS = −0.53 kcal mol−1 (±15%–20%) and ΔG = −7.57 (±15%–20%) kcal mol−1. In contrast, linear thermogram was obtained in the presence of Ca2+, indicative of non-specific and very weak binding of compound 1 with CDPK1. Solid line represents the best fit of the non-linear experimental data fitted to the single-site binding model. The experiment was done twice. (b) Fluorescence spectroscopy supports ITC interaction analysis. The tryptophan fluorescence maximum for CDPK1 was observed at 333 nm. In the absence of Ca2+, titration of CDPK1 with compound 1(ST092793) resulted in sharp fluorescence displacement curves. However, titration in the presence of Ca2+ resulted in small variations in the fluorescence emission intensity indicating that compound 1(ST092793) preferentially binds with open or ‘inactive’ conformation of CDPK1. The experiment was done twice.

Figure 5.
Compound 1(ST092793) interacts with CDPK1 favorably in the absence of calcium.

(a) Representative ITC binding isotherms resulting from titrations of CDPK1 with compound 1(ST092793). The compound showed differential binding with CDPK1 in the absence (Ka = 3.6 × 105 ± 2.4 × 105 M−1) and presence (Ka = 2.6 × 103 ± 25% M−1) of Ca2+. Binding isotherm in the absence of Ca2+ reached a plateau phase at 1 : 1 stoichiometry indicating that compound 1(ST092793) interacts with only a single independent binding site of CDPK1. Also, interaction of compound 1(ST092793) with CDPK1 was enthalpically driven with ΔH = −8.1 ± 5.2 kcal mol−1, TΔS = −0.53 kcal mol−1 (±15%–20%) and ΔG = −7.57 (±15%–20%) kcal mol−1. In contrast, linear thermogram was obtained in the presence of Ca2+, indicative of non-specific and very weak binding of compound 1 with CDPK1. Solid line represents the best fit of the non-linear experimental data fitted to the single-site binding model. The experiment was done twice. (b) Fluorescence spectroscopy supports ITC interaction analysis. The tryptophan fluorescence maximum for CDPK1 was observed at 333 nm. In the absence of Ca2+, titration of CDPK1 with compound 1(ST092793) resulted in sharp fluorescence displacement curves. However, titration in the presence of Ca2+ resulted in small variations in the fluorescence emission intensity indicating that compound 1(ST092793) preferentially binds with open or ‘inactive’ conformation of CDPK1. The experiment was done twice.

Overall, the resultant 1 : 1 stoichiometric complex of compound 1(ST092793) and CDPK1 in the absence of Ca2+ was found to be highly stable, supported by stronger Ka and favorable binding enthalpy and entropy.

Fluorescence spectroscopy supports ITC interaction analysis of compound 1(ST092793) and CDPK1

To confirm ITC interaction analysis of compound 1(ST092793) and CDPK1, fluorescence spectroscopic study was performed. Figure 5b shows typical fluorescence ‘displacement’ titrations, in the absence and presence of Ca2+. The tryptophan fluorescence maximum for CDPK1 was observed at 333 nm. In the absence of Ca2+, titration of CDPK1 with serial concentrations of compound 1(ST092793) resulted in progressive decline in the tryptophan fluorescence emission intensity throughout the wavelength range of 300–450 nm. In this case, the fluorescence displacement curve was sharp and the compound bound to CDPK1 protein even at lower concentrations. However, in the presence of Ca2+, lack of any trend and small variations in the fluorescence emission intensity indicates that compound 1(ST092793) preferentially binds with open or inactive conformation of CDPK1, as confirmed by ITC analysis.

Compound 1(ST092793) confers thermal stability to CDPK1

Thermal Stability Assay (TSA) is a method for detecting target engagement by monitoring thermostability of a given protein in the presence of its ligand [32]. This is based on the principle that target-ligand interactions result in alteration in the thermodynamic parameters of the protein, affecting its stability with corresponding increase in temperature. TSA was optimized for binding analysis between purified recombinant CDPK1 protein and Compound 1(ST092793). Melting curves at temperatures ranging from 40°C to 90°C demonstrated significant thermo-stabilization of CDPK1 in the presence of Ca2+ (Figure 6a). Importantly, in the opted temperature range, CDPK1 treated with Compound 1(ST092793) (100 µM) in the absence of Ca2+ showed abundances of the protein higher than the abundances of those treated with DMSO as control, suggesting ligand-dependent thermo-stabilization of CDPK1. No alteration in thermal stability was observed upon treatment of the protein with Compound 1(ST092793) in the presence of Ca2+ (Figure 6b). Furthermore, TSA analysis at 70°C with all 18 compounds (at 100 µM concentration) procured from virtual screening of MyriaScreen II library of small molecules demonstrated that thermal stability of CDPK1 gets relatively more enhanced in the presence of Compound 1(ST092793) (Figure 6c), further supporting our in vitro findings. For quantification of thermostable CDPK1, the signal intensity was normalized to the respective intensity at 40°C.

Compound 1(ST092793) confers thermostability to CDPK1.

Figure 6.
Compound 1(ST092793) confers thermostability to CDPK1.

(a) Representative melting curves showing thermo-stabilization of CDPK1 in the presence of Ca2+. (b) Compound 1(ST092793) mediated thermo-stabilization of CDPK1 is dependent on Ca2+ ions. (c) TSA analysis of 18 compounds procured from virtual screening of MyriaScreen II library. At 70°C, thermal stability of CDPK1 gets relatively most enhanced in the presence of Compound 1(ST09279), further supporting our in vitro findings. Estimates of band intensities of CDPK1 protein from Coomassie stained gel images were made by using ImageJ 1.52a. CDPK1 signal intensity was normalized to the respective intensity at 40°C. (d) Evaluation of growth inhibitory effect on the parasite. Lead compound 1(ST092793) with CDPK1-inhibitory activity in vitro showed IC50 value of 9.5 µM over 72 h of intra-erythrocytic life cycle of the parasite. Parasitemia level of untreated parasites was taken as 100%. The experiment was done twice. X-axis: Log[inhibitor], M; Y-axis: Percent parasite growth.

Figure 6.
Compound 1(ST092793) confers thermostability to CDPK1.

(a) Representative melting curves showing thermo-stabilization of CDPK1 in the presence of Ca2+. (b) Compound 1(ST092793) mediated thermo-stabilization of CDPK1 is dependent on Ca2+ ions. (c) TSA analysis of 18 compounds procured from virtual screening of MyriaScreen II library. At 70°C, thermal stability of CDPK1 gets relatively most enhanced in the presence of Compound 1(ST09279), further supporting our in vitro findings. Estimates of band intensities of CDPK1 protein from Coomassie stained gel images were made by using ImageJ 1.52a. CDPK1 signal intensity was normalized to the respective intensity at 40°C. (d) Evaluation of growth inhibitory effect on the parasite. Lead compound 1(ST092793) with CDPK1-inhibitory activity in vitro showed IC50 value of 9.5 µM over 72 h of intra-erythrocytic life cycle of the parasite. Parasitemia level of untreated parasites was taken as 100%. The experiment was done twice. X-axis: Log[inhibitor], M; Y-axis: Percent parasite growth.

Cytotoxic effect of compounds with high selectivity for parasite over host cells

Having observed no cytotoxic effect on HepG2 cell line, lead compound 1(ST092793) and compound 2(S344699) with propensity to bind and inhibit the enzymatic activity of CDPK1 were subjected to dose response over 72 h of intra-erythrocytic life cycle of the parasite. Compound 1(ST092793) and compound 2(S344699) showed IC50 values of 9.5 µM and 17.8 µM, respectively (Figure 6d and Supplementary Figure S3). Control parasites were healthy and completed its life cycle producing trophozoites at 72 h post-incubation.

Discussion

In Plasmodium falciparum, Ca2+-mediated signaling pathways are translated into cellular responses by CDPKs, among which CDPK1 is the most widely studied member of the cdpk multigene family [7,8,11,33,34]. Insights into the physiological role of CDPK1 in cellular biology of P. falciparum depicts importance of this protein in late schizogony, subsequent invasion of merozoites into host RBCs, formation of male and female gametes, and transmission to the mosquito host [9–12]. During gliding and invasion of the parasite, CDPK1 regulates actin-myosin motor by phosphorylating components of IMC, i.e. GAP45 and MTIP. It also regulates cAMP signaling module by phosphorylating and activating PKAr [13]. With its high abundance, CDPK1 stands alone as a multistage signaling regulator of P falciparum and has been depicted as the most sought after molecular target for drug discovery against malaria.

Most of the small molecule kinase inhibitors have generally been identified either by high-throughput virtual or fragment-based screening of chemical libraries. Among the drug designing platforms, virtual or molecular docking screening utilizes in silico robotic platforms for identification of new lead compounds in the process of drug discovery. In this approach, large libraries of unique organic molecules are computationally screened for compounds, collectively referred to as biologically relevant chemical space, that complements ‘structural code’ (i.e. ligand binding pocket) of target protein(s), followed by predicting the binding affinity by experimental validation. Filters may be applied to ensure that the hit compounds meet standard of biological relevance or drug-likeness, exploting the druggability of the target proteins. Once promising candidates are identified, one can utilize medicinal chemistry reactions and synthetic strategies to dynamically generate more diverse compounds around the initial hits, with enhanced efficacy. This whole process of structure-based screening and identifying lead candidates frees medicinal chemistry from the tyranny of empirical screening which includes substrate-based design and incremental modification [35–38]. Previous studies by, Hou et al. [39] identified a phenyl thiadiazole derivative as lead compound by virtual screening against protein kinase CK2α, a serine/threonine kinase that has been implicated in pathologies of a variety of human diseases. Similarly, Ma et al. [40] identified Amentoflavone analogs as Type II inhibitors against JAK2, by structure-based virtual screening of a natural product library, followed by in silico optimization and kinetic experiments. Furthermore, Su et al. [41] identified two potent Rho kinase inhibitors, namely phloretin and baicalein, by integrating high-throughput virtual screening of natural products library with Quantitative Structure-Activity Relationship (QSAR)-based rescoring and kinase assay in order to attenuate Pulmonary Hypertension (PH). Later, Zhao et al. [42] identified potential Protein Kinase C (PKC) inhibitors, Fisetin and Tetrahydropapaverine, with activities at nanomolar level, by high-throughput virtual screening against a natural products library to modulate general anesthetic effects.

Earlier attempts have also been made to identify specific inhibitors against PfCDPK1 protein. First evidence came from Kato et al. [9] who identified a series of structurally related 2,6,9-trisubstituted purines by in vitro biochemical screening, by using recombinant PfCDPK1 protein, against a library of 20,000 compounds. Purfalcamine, a representative member of the series resulted in sudden developmental arrest at the late schizont stage in P. falciparum. Later, Lemercier et al. [43] identified two distinct chemical series of small molecule inhibitors, compounds 1 and 2 by biochemical screening of 54,000 compounds against recombinant PfCDPK1 protein. Also, an earlier study by our group (Bansal et al. [11]) utilized a peptidic region (P3 peptide; Leu356 to Thr370) of PfCDPK1 Junction Domain (JD) which could specifically inhibit functional activity of the protein, block microneme discharge and erythrocyte invasion by P. falciparum merozoites. Later, a series of imidazopyridazines was identified by Chapman et al. [44] as potent inhibitors of PfCDPK1, which showed promising in vitro anti-parasite activity, ADME and kinase selectivity profiles; and, in vivo pharmacokinetics in mouse, with modest in vivo efficacy in P. berghei mouse model of malaria. In the subsequent year, group of Chapman et al. (2014) used a structure-guided design approach using PfCDPK1 homology model to optimize the imidazopyridazine series of CDPK1 inhibitors. Structural modifications resulted in compounds with potent kinase-inhibitory & anti-parasitic activities, suitable ADME properties and modest in vivo efficacy in the mouse model [45]. Finally, Ansell et al. [46] screened a library of 35,000 small molecules by high-throughput in vitro biochemical assay followed by chemical optimization to identify chemical series of PfCDPK1-inhibitors.

Due to the emergence of parasite resistance to existing frontline anti-malarial drugs, there is a pressing need to identify and characterize new anti-malarial molecules that would target the parasite at multiple proliferative stages during its life cycle. In this context, since PfCDPK1 has been previously identified as one such target, we have utilized a structurally diverse collection of 10,000 drug-like molecules from MyriaScreen II library to screen for ligands complementing ATP binding pocket of CDPK1, as represented schematically in Figure 7. In the absence of an experimentally determined atomic structure of PfCDPK1, earlier attempts have been made to resolve homology modeling based three-dimensional structure of PfCDPK1 [9,11,43]. In all of these studies, distantly related orthologous sequences of PfCDPK1 either from humans (death-associated protein kinase, PDB ID: 1JKK) [9] or from another apicomplexan parasite, Cryptosporidium parvum (calcium/calmodulin-dependent protein kinase, PDB ID: 2qg5; and, CpCDPK3, PDB ID: 3LIJ) [11,43] were used as templates for structural modeling of PfCDPK1. Owing to the high sequence identity between CDPK1 orthologs from P. falciparum strain 3D7 and evolutionarily closely related species, P. berghei strain ANKA, we used X-Ray diffraction based structural model of PbCDPK1 (the most recently resolved crystal structure of CDPK1 protein, PDB ID: 3Q5I) as template to generate 3D structural co-ordinates of PfCDPK1. Stereochemical quality and accuracy of the structurally refined model of PfCDPK1 indicated that the backbone dihedral angles: ɸ and Ψ were reasonably accurate, thus confirming reliability of the generated model for performing docking-based virtual screening. Candidate compounds were filtered out from MyriaScreen II library on the basis of their most negative free binding energies or binding affinity towards ATP binding pocket of the protein. After applying additional filter of Lipinski's rule of five, 18 compounds were finally selected as probable hits and set to submit for experimental validation.

Overall schematic representation.

Figure 7.
Overall schematic representation.

Structurally diverse collection of 10,000 drug-like molecules from MyriaScreen II library was utilized to screen for ligands complementing ATP binding pocket of CDPK1, thus inhibiting parasite growth in vitro.

Figure 7.
Overall schematic representation.

Structurally diverse collection of 10,000 drug-like molecules from MyriaScreen II library was utilized to screen for ligands complementing ATP binding pocket of CDPK1, thus inhibiting parasite growth in vitro.

Real-time biomolecular interaction analysis using SPR indicated wide range of binding affinities of the hits (N = 18) for purified CDPK1, at physiologically relevant concentrations. Enzymatic activity assays indicated that the compounds showed differential inhibition of CDPK1 activity in vitro. Dose-response curves for CDPK1-inhibitory activity of lead compounds 1(ST092793) and 2(S344699) depicted their IC50 values to be 33.8 µM and 42.6 µM, respectively. Also, decrease in phosphorylation status of CDPK1 and casein was observed with increase in concentrations of both of these compounds, which suggests that compounds 1(ST092793) and 2(S344699) are inhibitors of CDPK1 protein. Furthermore, competitive ATP-affinity chromatography of CDPK1 pre-incubated with compound 1(ST092793) and/or 2(S344699) indicated that both of these compounds inhibit the enzymatic activity by directly binding to the catalytically active ATP-binding pocket of CDPK1.

Binding affinity and stoichiometry details of the complex formation between relatively more potent compound 1(ST092793) and CDPK1, as examined by using ITC, demonstrated that compound 1(ST092793) binds ∼100 times stronger with CDPK1 in the absence of Ca2+ than in the presence, with the interactions mainly driven by enthalpy. ITC interaction analysis was further confirmed by measuring intrinsic tryptophan fluorescence of CDPK1 in the presence of increasing concentrations of compound 1(ST092793). The interaction of compound 1(ST092793) with CDPK1 in the absence of Ca2+ seems counterintuitive, since Ca2+ activates the enzyme. Notably, however, we want to emphasize that previous finding by our group indicated that the interactions between the CamLD), JD and KD of PfCDPK1 are completely different in the presence and absence of Ca2+ [11]. Mechanistic insights into the regulation of PfCDPK1 by Ca2+ suggests that upon binding with Ca2+, PfCDPK1 undergoes a conformational change from ‘open’ or inactive state to ‘closed’ or active state, in which the CamLD interacts with the C-terminal region of the JD and exposes the substrate binding catalytic site in the KD, leading to activation of its functional activity [11]. In this context, our rationale behind identification of CDPK1-selective inhibitors by adopting a structure-based virtual chemical library screening approach combined with biochemical & biophysical characterization based tools, was likely to identify those lead candidate compounds capable of binding with the ‘open’ or inactive conformation of CDPK1, (i.e. in the absence of Ca2+). This would not allow CDPK1 to undergo the catalytically active conformational change and, thus, render CDPK1 in its inactive state throughout the intra-erythrocytic proliferative stages during its life cycle. And, as confirmed with ITC and fluorescence spectroscopic experiments in the presence and absence of Ca2+, we successfully achieved this feat by identifying compound 1(ST092793) which shows stronger binding affinity towards CDPK1 in the absence of Ca2+, and binding becomes non-specific and weaker in the presence of Ca2+. Moreover, our rationale is in consistent with the findings by Scheele et al. [47] who reported that Bumped Kinase Inhibitor (BKI), RM-1-132, that complements the enlarged ATP-binding pocket of CDPK1 from Toxoplasma gondii, stabilizes the enzyme (enhances on-target engagement or interaction) in the absence of Ca2+. This was confirmed by using Thermal Shift Assay (TSA) in which interaction of purified TgCDPK1 protein with the BKI inhibitor resulted in increase in melting Temperature (Tm) of the enzyme in the absence of Ca2+, as confirmed by titration of both Ca2+ and EGTA.

Furthermore, TSA analysis demonstrated enhanced thermo-stabilization of CDPK1 upon treatment with compound 1 (ST092793), in the absence of Ca2+. Our TSA analysis is in consistent with the findings by Scheele et al. [47] as mentioned above. Furthermore, TSA analysis with all 18 compounds procured from virtual screening of MyriaScreen II library demonstrated that thermal stability of CDPK1 gets relatively highly enhanced in the presence of Compound 1 (ST092793), further supporting our in vitro findings. Finally, lead compound 1(ST092793) and compound 2(S344699), with CDPK1 inhibitory activity in vitro, showed intra-erythrocytic growth inhibition of the parasite, depicting IC50 value of 9.5 µM and 17.8 µM, respectively. Table 1 compares the numerical values of affinity constant KD, PfCDPK1 IC50, and whole cell EC50 for compound 1(ST092793) and compound 2(S344699). Experimental binding affinity for both compounds correlates fairly well with their biochemical inhibition and cellular activities. CDPK1 inhibitory activity of the lead compound 1(ST092793) correlates relatively weakly with whole cell (parasite) potency. But, we can never rule out the possibility that additional intracellular targets may contribute to the cellular activity. Therefore, further optimization of biochemical properties of lead compound 1(ST092793) by using structure-based drug design approach would be a powerful tool to expedite the discovery of derivatives that exhibit greater specificity and selectivity towards PfCDPK1. Furthermore, to confirm that compound 1(ST092793) is killing the parasite through CDPK1, transgenic parasite line expressing CDPK1 mutated at its gatekeeper residue (T145M) can be exploited.

Table 1
The table compares the numerical values of affinity constant KD, PfCDPK1 IC50, and whole cell EC50 for compound 1(ST092793) and compound 2(S344699)
S. No.ParameterCompound 1(ST092793)Compound 2(S344699)
Affinity constant, KD 51.9 ± 0.3 µM 200 ± 5.6 µM 
PfCDPK1 IC50 33.8 µM 42.6 µM 
Whole cell EC50 9.5 µM 17.8 µM 
S. No.ParameterCompound 1(ST092793)Compound 2(S344699)
Affinity constant, KD 51.9 ± 0.3 µM 200 ± 5.6 µM 
PfCDPK1 IC50 33.8 µM 42.6 µM 
Whole cell EC50 9.5 µM 17.8 µM 

Experimental binding affinity for both compounds correlates fairly well with their biochemical inhibition and cellular activities.

In conclusion, we adopted a structure-based virtual chemical library screening approach in combination with extensive biochemical and biophysical characterization based tools to identify novel lead candidates, Compound 1(ST092793) and Compound 2(S344699), capable of inhibiting CDPK1 activity in vitro. The inhibitory potency of these compounds was authenticated by favorable interactions with the catalytically active ATP binding pocket of the protein. These compounds exhibited cytotoxicity against malaria parasite ex vivo, indicating their promising candidacy for malaria treatment. Further studies can be planned to improve CDPK1 inhibitory activity and kinase selectivity of these compounds by designing new analogs of the same. Altogether, our study proposes a novel CDPK1-selective inhibitor with strong anti-malarial activity and represents a combinatorial high-throughput platform involving in silico and in vitro applications as a blueprint for the rational identification of novel selective inhibitors.

Competing Interests

The authors declare that there are no competing interests associated with the manuscript.

Funding

Funding from Science and Engineering Research Board (SERB) (EMR/2016/005644) and Drug and Pharmaceuticals Research Programe (DPRP) (Project No. P/569/2016-1/TDT for SS is acknowledged. SS is a recipient of the Innovative Young Biotechnologist Award (IYBA) and National Bioscientist Award from Department of Biotechnology (DBT). RJ acknowledges University Grant Commission (UGC) for doctoral research fellowship.

Author Contributions

R.J. and S.S. designed the experiments. R.J. and S.G. performed the experiments. S.P. and S.S. contributed reagents. R.J., M.M., S.S. analyzed data. R.J., S.P. and S.S. wrote the manuscript.

Acknowledgements

Authors acknowledge SPR facility of Advanced Instrumentation Research Facility (AIRF), Jawaharlal Nehru University (JNU), and central facility of Special Center for Molecular Medicine (SCMM), JNU for flow cytometry. The lab facilities of Shiv Nadar University and SCMM, JNU are also acknowledged.

Abbreviations

     
  • {8-[(6-Amino) Hexyl]-amino-ATP-agarose

    8AH-ATP-agarose

  •  
  • ACT

    Artemisinin-based combinatorial therapy

  •  
  • AMA-1

    apical membrane antigen-1

  •  
  • AML

    acute myeloid leukemia

  •  
  • AZ-VL

    AZ-virtual library

  •  
  • BI

    Boehringer Ingelheim

  •  
  • CaMKs

    calmodulin-dependent protein kinases

  •  
  • CamLD

    calmodulin-like regulatory domain

  •  
  • CDD

    conserved domain database

  •  
  • CDPK1

    Ca2+-dependent protein kinase 1

  •  
  • CDPKs

    calcium-dependent protein kinases

  •  
  • CETSA

    cellular thermal shift assay

  •  
  • DruLiTo

    drug likeness tool

  •  
  • EBA-175

    175 kDa Erythrocyte Binding Antigen

  •  
  • FL-2

    fluorescence signal

  •  
  • FLT3

    FMS-like tyrosine kinase 3

  •  
  • FS

    fluorescence spectroscopy

  •  
  • GAP-45

    50 kDa glideosome-associated protein

  •  
  • GMS

    greater Mekong sub-region

  •  
  • IMC

    inner membrane complex

  •  
  • IRL

    integrated rate law

  •  
  • ITC

    isothermal titration calorimetry

  •  
  • J

    junction domain

  •  
  • Ka

    association rate constant

  •  
  • Ka

    binding affinity constant

  •  
  • Kd

    dissociation rate constant

  •  
  • KD

    kinase domain

  •  
  • MDR

    multi-drug resistance

  •  
  • MTIP

    myosin tail domain interacting protein

  •  
  • N

    stoichiometry

  •  
  • PGVL

    Pfizer global virtual library

  •  
  • PIC

    protease inhibitor cocktail

  •  
  • PKAr

    regulatory subunit of cAMP-dependent Protein Kinase A

  •  
  • PKC

    protein kinase C

  •  
  • PKG

    protein kinase G

  •  
  • QSAR

    quantitative structure-activity relationship

  •  
  • Req

    response at equilibrium

  •  
  • RMSD

    root-mean-square deviation

  •  
  • RT

    room temperature

  •  
  • RU

    response unit

  •  
  • SD

    standard deviation

  •  
  • SDF

    structural data format

  •  
  • SPR

    surface plasmon resonance

  •  
  • TSA

    thermal stability assay

  •  
  • TSA

    thermal stability assay

  •  
  • ΔG

    Gibbs free energy change

  •  
  • ΔH

    enthalpy change

  •  
  • ΔS

    entropy change

References

References
1
2018, W. malaria report. G. W. H. O
. https://www.who.int/malaria/publications/world-malaria-report-2018/report/en/,
accessed in April 2019
2
Singh
,
S.
,
Alam
,
M.M.
,
Pal-Bhowmick
,
I.
,
Brzostowski
,
J.A.
and
Chitnis
,
C.E.
(
2010
)
Distinct external signals trigger sequential release of apical organelles during erythrocyte invasion by malaria parasites
.
PLoS Pathog.
6
,
e1000746
3
Ishino
,
T.
,
Orito
,
Y.
,
Chinzei
,
Y.
and
Yuda
,
M.
(
2006
)
A calcium-dependent protein kinase regulates Plasmodium ookinete access to the midgut epithelial cell
.
Mol. Microbiol.
59
,
1175
1184
4
Siden-Kiamos
,
I.
,
Ecker
,
A.
,
Nybäck
,
S.
,
Louis
,
C.
,
Sinden
,
R.E.
and
Billker
,
O.
(
2006
)
Plasmodium berghei calcium-dependent protein kinase 3 is required for ookinete gliding motility and mosquito midgut invasion
.
Mol. Microbiol.
60
,
1355
1363
5
Billker
,
O.
,
Dechamps
,
S.
,
Tewari
,
R.
,
Wenig
,
G.
,
Franke-Fayard
,
B.
and
Brinkmann
,
V.
(
2004
)
Calcium and a calcium-dependent protein kinase regulate gamete formation and mosquito transmission in a malaria parasite
.
Cell
117
,
503
514
6
Dvorin
,
J.D.
,
Martyn
,
D.C.
,
Patel
,
S.D.
,
Grimley
,
J.S.
,
Collins
,
C.R.
,
Hopp
,
C.S.
, et al (
2010
)
A plant-like kinase in Plasmodium falciparum regulates parasite egress from erythrocytes
.
Science
328
,
910
912
7
Harper
,
J.F.
and
Harmon
,
A.
(
2005
)
Plants, symbiosis and parasites: a calcium signalling connection
.
Nat. Rev. Mol. Cell Biol.
6
,
555
566
8
Billker
,
O.
,
Lourido
,
S.
and
Sibley
,
L.D.
(
2009
)
Calcium-dependent signaling and kinases in apicomplexan parasites
.
Cell Host Microbe
5
,
612
622
9
Kato
,
N.
,
Sakata
,
T.
,
Breton
,
G.
,
Le Roch
,
K.G.
,
Nagle
,
A.
,
Andersen
,
C.
, et al (
2008
)
Gene expression signatures and small-molecule compounds link a protein kinase to Plasmodium falciparum motility
.
Nat. Chem. Biol.
4
,
347
356
10
Azevedo
,
M.F.
,
Sanders
,
P.R.
,
Krejany
,
E.
,
Nie
,
C.Q.
,
Fu
,
P.
,
Bach
,
L.A.
et al (
2013
)
Inhibition of Plasmodium falciparum CDPK1 by conditional expression of its J-domain demonstrates a key role in schizont development
.
Biochem. J.
452
,
433
441
11
Bansal
,
A.
,
Singh
,
S.
,
More
,
K.R.
,
Hans
,
D.
,
Nangalia
,
K.
,
Yogavel
,
M.
et al (
2013
)
Characterization of Plasmodium falciparum calcium-dependent protein kinase 1 (PFCDPK1) and its role in microneme secretion during erythrocyte invasion
.
J. Biol. Chem.
288
,
1590
1602
12
Bansal
,
A.
,
Molina-Cruz
,
A.
,
Brzostowski
,
J.
,
Liu
,
P.
,
Luo
,
Y.
,
Gunalan
,
K.
et al (
2018
)
Pf CDPK1 is critical for malaria parasite gametogenesis and mosquito infection
.
Proc. Natl Acad. Sci. U.S.A.
115
,
774
779
13
Kumar
,
S.
,
Kumar
,
M.
,
Ekka
,
R.
,
Dvorin
,
J.D.
,
Paul
,
A.S.
,
Madugundu
,
A.K.
, et al (
2017
)
PfCDPK1 mediated signaling in erythrocytic stages of Plasmodium falciparum
.
Nat. Commun.
8
,
63
14
Alam
,
M.M.
,
Solyakov
,
L.
,
Bottrill
,
A.R.
,
Flueck
,
C.
,
Siddiqui
,
F.A.
,
Singh
,
S.
, et al (
2015
)
Phosphoproteomics reveals malaria parasite protein kinase G as a signalling hub regulating egress and invasion
.
Nat. Commun.
6
,
7285
15
Trager
,
W.
and
Jensen
,
J.B.
(
1976
)
Human malaria parasites in continuous culture
.
Science
193
,
673
675
16
Aurrecoechea
,
C.
,
Brestelli
,
J.
,
Brunk
,
B.P.
,
Dommer
,
J.
,
Fischer
,
S.
,
Gajria
,
B.
, et al (
2009
)
PlasmoDB: a functional genomic database for malaria parasites
.
Nucleic Acids Res.
37
,
D539
D543
17
Altschul
,
S.F.
,
Gish
,
W.
,
Miller
,
W.
,
Myers
,
E.W.
and
Lipman
,
D.J.
(
1990
)
Basic local alignment search tool
.
J. Mol. Biol.
215
,
403
410
18
Berman
,
H.M.
,
Westbrook
,
J.
,
Feng
,
Z.
,
Gilliland
,
G.
,
Bhat
,
T.N.
,
Weissig
,
H.
et al (
2000
)
The protein data bank
.
Nucleic Acids Res.
28
,
235
242
19
Wernimont
,
A.K.
,
Hutchinson
,
A.
,
Sullivan
,
H.
,
Panico
,
E.
,
Crombet
,
L.
,
Cossar
,
D.
et al
Crystal Structure of CDPK1 from Plasmodium Bergheii, PBANKA_031420; To be published
20
Webb
,
B.
and
Sali
,
A.
(
2016
)
Comparative protein structure modeling using MODELLER
.
Curr. Protoc. Bioinformatics
54
,
5.6.1
5.6.37
21
Xu
,
D.
and
Zhang
,
Y.
(
2011
)
Improving the physical realism and structural accuracy of protein models by a two-step atomic-level energy minimization
.
Biophys. J.
101
,
2525
2534
22
Ramachandran
,
G.N.
and
Sasisekharan
,
V.
(
1968
)
Conformation of polypeptides and proteins
.
Adv. Protein Chem.
23
,
283
437
23
Laskowski
,
R.A.
,
MacArthur
,
M.W.
and
Thornton
,
J.M.
(
1993
)
PROCHECK: a program to check the stereochemical quality of protein structures
.
J. Appl. Crystallogr.
26
,
283
291
24
Lipinski
,
C.A.
,
Lombardo
,
F.
,
Dominy
,
B.W.
and
Feeney
,
P.J.
(
2002
)
Experimental and computational approaches to estimate solubility and permeability in drug discovery and development settings
.
Adv. Drug Deliv. Rev.
46
,
3
26
25
Oleg
,
T.
and
Olson
,
A.J.
(
2010
)
Autodock Vina: improving the speed and accuracy of docking with a new scoring function, efficient optimization, and multithreading
.
J. Comput. Chem.
31
,
455
461
26
Steinbeck
,
C.
,
Han
,
Y.
,
Kuhn
,
S.
,
Horlacher
,
O.
,
Luttmann
,
E.
and
Willighagen
,
E.
(
2003
)
The chemistry development kit (CDK): an open-source Java library for chemo- and bioinformatics
.
J. Chem. Inf. Comput. Sci.
43
,
493
500
27
Delano
,
W.L.
(
2002
)
The PyMOL Molecular Graphics System
,
Schrodinger
28
Rehman
,
S.A.A.
,
Verma
,
V.
,
Mazumder
,
M.
,
Dhar
,
S.K.
and
Gourinath
,
S.
(
2013
)
Crystal structure and mode of helicase binding of the C-terminal domain of primase from Helicobacter pylori
.
J. Bacteriol.
195
,
2826
2838
29
Jain
,
R.
,
Garg
,
S.
and
Singh
,
S.
(
2019
)
Screening of chemical library against whole cell kinome activity via non-radioactive, high throughput kinase assay
.
MethodsX
6
,
162
168
30
Gupta
,
S.
,
Tiwari
,
N.
and
Munde
,
M.
(
2019
)
A comprehensive biophysical analysis of the effect of DNA binding drugs on protamine-induced DNA condensation
.
Sci. Rep.
9
,
5891
31
Hashimoto
,
M.
,
Girardi
,
E.
,
Eichner
,
R.
and
Superti-Furga
,
G.
(
2018
)
Detection of chemical engagement of solute carrier proteins by a cellular thermal shift assay
.
ACS Chem. Biol.
13
,
1480
1486
32
Pantoliano
,
M.W.
,
Petrella
,
E.C.
,
Kwasnoski
,
J.D.
,
Lobanov
,
V.S.
,
Myslik
,
J.
,
Graf
,
E.
, et al (
2001
)
High-density miniaturized thermal shift assays as a general strategy for drug discovery
.
J. Biomol. Screen.
6
,
429
440
33
Sharma
,
P.
and
Chitnis
,
C.E.
(
2013
)
Key molecular events during host cell invasion by Apicomplexan pathogens
.
Curr. Opin. Microbiol.
16
,
432
437
34
Zhao
,
Y.
,
Kappes
,
B.
and
Franklin
,
R.M.
(
1993
)
Gene structure and expression of an unusual protein kinase from Plasmodium falciparum homologous at its carboxyl terminus with the EF hand calcium- binding proteins
.
J. Biol. Chem.
268
,
4347
4354
PMID:
[PubMed]
35
Shoichet
,
B.K.
(
2004
)
Virtual screening of chemical libraries
.
Nature
432
,
862
865
36
Lionta
,
E.
,
Spyrou
,
G.
,
Vassilatis
,
D.
and
Cournia
,
Z.
(
2014
)
Structure-based virtual screening for drug discovery: principles, applications and recent advances
.
Curr. Top. Med. Chem.
14
,
1923
1938
37
Cheng
,
T.
,
Li
,
Q.
,
Zhou
,
Z.
,
Wang
,
Y.
and
Bryant
,
S.H.
(
2012
)
Structure-based virtual screening for drug discovery: a problem-centric review
.
AAPS J.
14
,
133
141
38
Oprea
,
T.I.
(
2002
)
Virtual screening in lead discovery: a viewpoint
.
Molecules
7
,
51
62
39
Hou
,
Z.
,
Nakanishi
,
I.
,
Kinoshita
,
T.
,
Takei
,
Y.
,
Yasue
,
M.
,
Misu
,
R.
, et al (
2012
)
Structure-based design of novel potent protein kinase CK2 (CK2) inhibitors with phenyl-azole scaffolds
.
J. Med. Chem.
55
,
2899
2903
40
Ma
,
D.L.
,
Chan
,
D.S.H.
,
Wei
,
G.
,
Zhong
,
H.J.
,
Yang
,
H.
,
Leung
,
L.T.
et al (
2014
)
Virtual screening and optimization of type II inhibitors of JAK2 from a natural product library
.
Chem. Commun.
50
,
13885
13888
41
Su
,
H.
,
Yan
,
J.
,
Xu
,
J.
,
Fan
,
X.Z.
,
Sun
,
X.L.
and
Chen
,
K.Y.
(
2015
)
Stepwise high-throughput virtual screening of Rho kinase inhibitors from natural product library and potential therapeutics for pulmonary hypertension
.
Pharm. Biol.
53
,
1201
1206
42
Zhao
,
J.
and
Zhou
,
C.
(
2015
)
Virtual screening of protein kinase C inhibitors from natural product library to modulate general anaesthetic effects
.
Nat. Prod. Res.
29
,
589
591
43
Lemercier
,
G.
,
Fernandez-Montalvan
,
A.
,
Shaw
,
J.P.
,
Kugelstadt
,
D.
,
Bomke
,
J.
,
Domostoj
,
M.
et al (
2009
)
Identification and characterization of novel small molecules as potent inhibitors of the plasmodial calcium-dependent protein kinase 1
.
Biochemistry
48
,
6379
6389
44
Chapman
,
T.M.
,
Osborne
,
S.A.
,
Bouloc
,
N.
,
Large
,
J.M.
,
Wallace
,
C.
,
Birchall
,
K.
, et al (
2013
)
Substituted imidazopyridazines are potent and selective inhibitors of Plasmodium falciparum calcium-dependent protein kinase 1 (PfCDPK1)
.
Bioorganic Med. Chem. Lett.
23
,
3064
3069
45
Chapman
,
T.M.
,
Osborne
,
S.A.
,
Wallace
,
C.
,
Birchall
,
K.
,
Bouloc
,
N.
,
Jones
,
H.M.
, et al (
2014
)
Optimization of an imidazopyridazine series of inhibitors of Plasmodium falciparum calcium-dependent protein kinase 1 (Pf CDPK1)
.
J. Med. Chem.
57
,
3570
3587
46
Ansell
,
K.H.
,
Jones
,
H.M.
,
Whalley
,
D.
,
Hearn
,
A.
,
Taylor
,
D.L.
,
Patin
,
E.C.
, et al (
2014
)
Biochemical and antiparasitic properties of inhibitors of the Plasmodium falciparum calcium-dependent protein kinase PfCDPK1
.
Antimicrob. Agents Chemother.
58
,
6032
6043
47
Scheele
,
S.
,
Geiger
,
J.A.
,
DeRocher
,
A.E.
,
Choi
,
R.
,
Smith
,
T.R.
,
Hulverson
,
M.A.
, et al (
2018
)
Toxoplasma calcium-dependent protein kinase 1 inhibitors: probing activity and resistance using cellular thermal shift assays
.
Antimicrob. Agents Chemother.
62
,
e00051
e00118