Antiplasmodial potential of compounds isolated from Ziziphus mucronata and their binding to Plasmodium falciparum HGXPRT using biophysical and molecular docking studies

The stem bark of Z. mucronata was collected from Rhebokfontein (27°22′44.3″S 31°28′07.3″E) within the KwaZulu-Natal Province of South Africa between February and March 2022. The plant’s stem bark was air-dried and crushed into a fine powder in preparation for sequential extraction. A 150-g sample of dried Z. mucronata was macerated successively with four solvents of increasing polarity. The plant was initially extracted with hexane for a week at room temperature. The solvent extraction was filtered through Whatman No. 1 filter paper and concentrated using a rotary evaporator to yield crude extracts. The crude extracts were then further macerated with DCM, EtOAC, and methanol (MeOH), thus yielding viscous crude extracts.

A bioassay-guided isolation of antiplasmodial active compounds was performed. DCM and EtOAC extracts, which exhibited high activities (IC₅₀ = 5.49 and 7.22 µg/mL, respectively), were selected for further isolation of active compounds using chromatographic techniques. The crude extracts were separated by column chromatography (CC) over silica gel and eluted with a stepwise gradient. The solvent system initially contained 10% ethyl acetate in hexane, gradually increasing to 50% ethyl acetate, while collecting approximately 10–15 mL fractions. Progress in the separation of compounds was monitored by conducting thin-layer chromatography (TLC) of each fraction using hexane–ethyl acetate (7:3) solvent system and sulfuric acid in methanol for detection. Fractions with similar patterns were combined and collected as pooled fractions. The pooled fractions were then concentrated under reduced pressure to yield yellow needle-shaped powder, spindle-shaped white solid residues, and white powdery compounds, labeled as KZD60 (135 mg), KZ589 (8 mg), and KZD70 (7.56 mg), respectively. 1D ¹H, ¹³C, and 2D distortionless enhancement by polarization transfer (DEPT) NMR studies were used to determine the chemical structures of the isolated compounds.

The antiplasmodial activity of extracts and compounds was screened against the wild-type drug-sensitive P. falciparum (NF54) strain of the human malaria parasite Plasmodium falciparum at the H3D testing Center at the University of Cape Town, South Africa (H3D Testing Centre 2023). Continuous cultures of asexual erythrocyte stages of P. falciparum were maintained using the method described by Trager and Jensen (Trager and Jensen 1976), with minor modifications. The parasites were cultured in human red blood cells (O⁺, suspended at 3% hematocrit) using RPMI 1640 medium supplemented with 0.5% Albumax II, 200 M hypoxanthine, 24 µg/mL gentamicin, 25 mM HEPES, and buffered with 0.2% sodium bicarbonate (NaHCO₃).

The in vitro antiplasmodial activity of crude extracts and isolated compounds was assessed on cultured P. falciparum using the Plasmodium lactate dehydrogenase (pLDH) assay, adapted from Makler and co-workers (Makler et al. 1993). pLDH is an enzyme that converts lactate to pyruvate, a crucial step in the glycolytic pathway of Plasmodium parasites. Unlike human LDH, pLDH utilizes the co-enzyme 3-acetylpyridine adenine dinucleotide (APAD) instead of NAD⁺. In the presence of phenazine ethosulfate (PES), pLDH catalyzes lactate oxidation to pyruvate while reducing APAD⁺ to APADH. This, in turn, reduces the yellow tetrazolium dye, nitroblue tetrazolium (NBT), to a dark blue-purple formazan compound. The formation of formazan is proportional to pLDH activity and can be quantified spectrophotometrically at a specific wavelength (Markwalter et al. 2016).

Z. mucronata crude extracts and compounds were dissolved in dimethyl sulfoxide (DMSO) to produce stock solutions of 20 mg/mL, which were then diluted with complete growth media to generate the tested concentration range. Artesunate (ARTS) and chloroquine (CQ) were used as positive controls, starting from a 1-µg/mL concentration. The stock solutions were subsequently diluted in complete media to achieve the various test concentrations. Twofold serial dilutions were made in 96-well microtiter plates in duplicate, with 40 µL of each concentration transferred to another 96-well microtiter plate. The final volume of 160 µL of parasitized red blood cells (2% parasitemia) was added to each well. The plates were incubated at 37 °C for 72 h in a sealed gas chamber with 3% O₂ and 4% CO₂, and the balance N₂.

After a 72-h incubation, the wells were gently resuspended. A 15-µL sample from each well was transferred to a duplicate plate containing 100 uL Malstat reagent and 25 µL of nitroblue tetrazolium solution per well, then incubated in the dark for 20 min. The NBT solution comprised 1 and 0.05 mg/mL of NBT and phenazine ethosulfate (PES) in water. APAD⁺ (0.664 mM) was prepared by adding it to 40 mL of MilliQ water and adjusting the pH to 9 with 1 M NaOH. To the APAD solution, sodium l-lactate (0.71 M), Tris base (0.22 M), and Triton X-100 (0.2%) were added, and the solution was then diluted to 50 mL with water. The optical density was measured using a microplate reader at a wavelength of 620 nm. The concentration required for 50% inhibition of parasite growth (IC₅₀) was calculated using a nonlinear regression curve in the Dotmatics software Platform.

A 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide (MTT) assay was used to assess the toxicity of the extracts and isolated compounds. Chinese hamster ovary (CHO) cells, a normal mammalian cell line, were seeded into 96-well plates (10⁵ cells/well) and incubated for 24 h at 37 °C with 5% CO₂. The cells were then treated with various concentrations of the extracts and test compounds, ranging from 50 µM to 16 nM, and incubated for 48 h at 37 °C. Subsequently, MTT solution was added to each well after 44 h, and the plate was incubated for an additional 4 h at 37 °C. Thereafter, the medium was removed and replaced with DMSO to solubilize the MTT formazan product, and the solutions were shaken for 20 min. The optical density was measured using a spectrophotometer at a wavelength of 540 nm. The 50% cytotoxicity concentration (CC₅₀) was calculated using the Dotmatics software Platform.

The gene construct, pET28-a/PfHGXPRT, was synthesized by Genscript Corporation (Nanjing, China). As described by Opoku and colleagues (Opoku et al. 2019), chemically competent Escherichia coli BL21 (DE3) cells were transformed with pET28-a/PfHGXPRT. The positively transformed cells harboring the PfHGXPRT gene were inoculated in 250 mL of Luria–Bertani (LB) broth enriched with 100 µg/mL kanamycin and incubated overnight at 37 °C. The overnight inoculum was sub-cultured in 1000 mL of supplemented LB broth and incubated at 37 °C with continuous shaking until the desired optical density (OD₆₀₀ = 0.4–0.6) was reached. PfHGXPRT expression was induced by the addition of 1 mM isopropyl-β-d-1-thiogalactopyranoside (IPTG). After overnight incubation, the cells were harvested by centrifugation, and the pellet fraction was suspended in lysis buffer (20 mM Tris–HCl, 500 mM NaCl, 25 mM imidazole, and 1 mM DTT at pH 7.4) on ice, followed by mild sonication. The resulting cell lysate was centrifuged at 4700 × g for 60 min. The PfHGXPRT supernatant was then collected and subjected to affinity purification.

The (His)₆-PfHGXPRT protein was purified using a HisPur nickel-charged nitriloacetic acid (Ni–NTA)-immobilized metal affinity chromatography (IMAC) column, following the manufacturer’s instructions. The bound protein was eluted with phosphate buffer containing 350 mM imidazole. The pure protein was then extensively dialyzed overnight at 4 °C against phosphate buffer without imidazole. The protein fractions were analyzed using 15% sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE). The Bradford assay was used to estimate the protein yield, with bovine serum albumin (Sigma-Aldrich, SA) serving as the standard. Results from this section are presented in the Supplementary data (Fig. S1).

Protein thermal shift assay was conducted using the Protein Thermal Shift™ Kit (ThermoFisher Scientific). The assay, as described by Niesen et al. (Niesen et al. 2007) with slight modifications, was followed. Briefly, PfHGXPRT was diluted to a final concentration of 100 µg/mL (3.57 µM) in 5 µL Protein Thermal Shift™ buffer and 2.5 µL 8 × Diluted Protein Thermal Shift ™ Dye. The melting temperature (T_m) was measured using a Rotor-Gene 6000 RT-PCR machine (Corbett Life Science). A temperature range of 25 to 95 °C was applied, with a ramp rate of 1 °C/min. The condition parameters were set for high-resolution melt (HRM) with excitation and emission wavelengths of 460 and 510 nm, respectively. All assays were performed both in the absence and presence of the isolated compounds at two different final concentrations: 0.5 and 1 mM for betulinic acid, methyl betulinate, and lupeol.

UV–vis absorption spectroscopy is widely used to study proteins and their interactions with different molecules. UV–vis spectroscopy was performed using previously reported methods (Opoku et al. 2019; Salomane 2023) with slight modifications. A 35-mM solution of PfHGXPRT was exposed to 0.5- and 0.1-mM concentrations of betulinic acid, methyl betulinate, and lupeol for 30 min at room temperature. The spectrum readings were obtained using a UV-1800 spectrophotometer (Shimadzu, Japan) after base corrections, within the wavelength range of 200 to 600 nm. The absorption peak analysis was performed using UV Probe Software, version 2.

The binding mode of compounds from Z. mucronata to the PfHGXPRT protein structure was determined using AutoDock Tools4 (Morris et al. 2009). The crystal structure of PfHGXPRT was retrieved from the Protein Data Bank (PDB ID: 7TUX) (Kim et al. 2022), and ligand molecules were retrieved from the PubChem database (Hanwell et al. 2012). Energy minimization of the ligands was performed using the steepest descent method and MMFF99s, as implemented in the Avogadro program (Minnow et al. 2022). The PfHGXPRT receptor was structurally prepared using UCSF Chimera (Pettersen et al. 2004) software. AutoDock Tools were then used to determine the size of the grid box, which encapsulated the entire protein to facilitate efficient binding site identification. The grid box dimensions were set as X = 40, Y = 40, Z = 40, with a center at X = − 3.16, Y = − 39.361, Z = − 26.97, a grid spacing of 0.375 Å, and an exhaustiveness value of 10. The Lamarckian genetic algorithm, implemented in AutoDock, was used to analyze the docking of compounds to the receptor. The docked ligand conformations with the highest docking scores and an RMSD value of zero were chosen to create protein–ligand complexes. Generated complexes were visually inspected using UCSF Chimera, and the ligand–receptor interactions were presented using the LigPlot⁺ (Clarkson et al. 2004) program.

Four crude extracts and three major compounds were tested against the wild-type drug-sensitive P. falciparum strain (NF54), compared to chloroquine and artemisinin. Table 2 shows that the bark ethyl acetate and DCM extracts exhibited the highest activity and were considered potent against the P. falciparum strain with IC₅₀ values of 5.49 and 7.22 µg/mL, respectively. Therefore, these two extracts were selected for the subsequent isolation of active compounds.

Antiplasmodial activity is defined as follows: high activity at IC₅₀ \(\le\) 5 µg/mL, good activity at 5–10 µg/mL, moderate activity at 15 µg/mL < IC₅₀ ≤ 50 µg/mL, and inactive at 50 µg/mL. A pure compound is highly active when IC₅₀ ≤ 1 µg/mL (Memvanga et al. 2015; Basco and Heseltine 2007; Alsayari and Wahab 2021). Compounds explored in this study are isolated from the stem bark of Z. mucronata. Traditionally, plants from this genus have played a crucial role in treating and managing various diseases, including their use as antipyretics—one of the primary symptoms associated with malaria (Singh et al. 2019). This study demonstrated that lupeol (KZD70) showed good antiplasmodial activity (IC₅₀: 7.56 µg/mL), which is consistent with the findings of peer researchers (Ziegler et al. 2004), who reported that lupeol from Ficus benjamina also exhibited potent in vitro antiplasmodial activity. The proposed mechanism of action of lupeol is such that the compound is incorporated into the cell membrane of the host’s erythrocytes and altered, which prohibits parasite invasion as opposed to a direct toxic effect on the malaria parasite (Isah et al. 2016). Methyl betulinate (IC₅₀: 10.11 µg/mL) is a naturally occurring structural analog of betulinic acid. In a study by Ziegler and colleagues (Isah et al. 2016), the antiplasmodial activity of methyl betulinate showed high activity with an IC₅₀ value of 3.3 µg/mL. Other structural derivatives of betulinic acid were extensively researched (Steele et al. 1999), the majority of which possessed IC₅₀ values < 10 µg/mL, further suggesting that derivatization of the compound plays a role in enhancing its antiplasmodial activity (Steele et al. 1999). Compound KZD60 (betulinic acid) showed moderate activity with an IC₅₀ value of over 20 µg/mL. Similar findings were reported in a study by Kopra and co-workers (Kopra et al. 2022), where betulinic acid isolated from a Tanzanian tree, Uapaca nitida Mull-Arg (Euphorbiaceous), showed antiplasmodial activity translated by an IC₅₀ value of 25.9 µg/mL when investigated against the K1 multidrug-resistant P. falciparum strain. The antiplasmodial activity of betulinic acid is similar to betulinic acid isolated from a Tanzanian tree, Uapaca nitida Mull-Arg (Euphorbiaceous), where the compound was evaluated against the K1 chloroquine-resistant P. falciparum strain, with an IC₅₀ value of 25.9 µg/mL (Kopra et al. 2022). The results of the cytotoxicity study showed that the compounds were non-toxic compared to the standard cytotoxic drug, Emetine (an alkaloid), which was found to be highly cytotoxic.

Differential scanning fluorimetry (DSF) is a versatile technique that reports protein thermal unfolding via fluorogenic dyes (typically SYPRO Orange or ANS), often used for protein stability and protein–ligand interactions (PLI) studies (Celej et al. 2003). Ligand binding drastically affects protein stability by inducing changes in the conformation of the target protein, which, in turn, produces a given response (Acosta et al. 2020). This study explored the PfHGXPRT thermal stability in the presence and absence of the study compounds at two differing concentrations (Table 3).

The stability of a protein is temperature-dependent due to Gibbs’ free energy during unfolding. Adding betulinic acid at both concentrations resulted in minimal changes to the observed T_m values of PfHGXPRT compared to PfHGXPRT alone. The presence of 0.5 mM methyl betulinate decreased PfHGXPRT thermal stability by reducing the T_m by approximately 1 °C, while 1 mM increased the T_m by 2 °C, suggesting that methyl betulinate causes concentration-dependent thermal stability. PfHGXPRT remained stable in the presence of methyl betulinate and betulinic acid individually. This may be due to the nature of the protein; purine phosphoribosyl transferases have oligomerization states that contribute to the protein’s overall stability (Gonçalves et al. 2022). The stability of PfHGXPRT during thermal denaturation in the presence of lupeol exhibited a peculiar trend; this behavior contrasts with the typical concentration-dependent thermal stability, where higher concentrations of ligands enhance the stability of the protein. In this case, at lower concentrations of lupeol, the T_m increased slightly to 62.0 °C, while the opposite was observed at higher concentrations (1 mM), with a decrease in thermal stability resulting in a drastic change of T_m from 61.90 to 30.70 °C. This is considered a negative T_m shift, meaning lupeol caused structural changes in the protein toward a more disordered conformation (Yu et al. 2015).

The absorption spectra of PfHGXPRT, like many proteins, are primarily due to the aromatic amino acids. The optimum PfHGXPRT concentration was 30 mM (Fig. 2A). The presence of the compounds had varying influences on the absorption peak of PfHGXPRT at different concentrations, as shown in Fig. 2b–d. The maximum absorption peak of unbound PfHGXPRT was at 280 nm, mainly due to the tyrosine residues in the protein (10 tyrosine residues and 1 tryptophan residue). The phenolic hydroxyl group of tyrosine gives the residue a unique ability to interact with non-protein atoms, especially in enzymatic reactions, where it donates both an electron and a proton (Fernandes et al. 2018; Subramanyam et al. 2009). In an unbound state, tyrosine residues are unoccupied, resulting in a high absorption peak. When bioactive compounds directly interact with the protein, they influence the peak, exposing fewer or more tyrosine residues (Opoku et al. 2019). Figure 2b shows a sharp peak of PfHGXPRT at 280 nm, which increased in intensity as the concentration of betulinic acid increased, resulting in a hyperchromic shift. A similar effect was observed by Subramanyam and colleagues (Eftink and Pedigo 2003), where betulinic acid independently increased the intensity of the peak of human serum albumin. The absorbance of betulinic acid suggests that more tyrosine residues were exposed (Opoku et al. 2019).

Bild vergrößern

Methyl betulinate (MA) had a similar effect on PfHGXPRT at both concentrations (Fig. 2c). MA lowered the absorbance and shortened the wavelength, resulting in a blue shift. A study by Opoku et al. (Opoku et al. 2019) showed a similar trend, whereby the interaction of iso-mukaadial acetate and ursolic acid acetate with the same protein (PfHGXPRT) decreased absorbance. The blue shift was due to the phenol rings of the tyrosine residues, which are usually buried within the hydrophobic core, being exposed during protein during unfolding (Antosiewicz and Shugar 2016). Although similar, MA had an intensified effect at low concentrations (0.1 mM). Additionally, the changes in the absorption spectra were similar to those observed in a study by Antosiewicz and Shugar (Schmid 2001), where high pH values caused protein denaturation, indicating ionization of the buried residues. This suggests that MA may cause protein denaturation in a similar manner; however, further studies must confirm this.

At both tested concentrations, lupeol caused a blue shift. However, the 0.1-mM concentration decreased the peak, while the 0.5-mM concentration had the opposite effect, increasing the intensity (Fig. 2d); these results correlate with the protein thermal shift assay. According to Eftnik and Pedigo (Antosiewicz and Shugar 2016), a blue shift indicates protein unfolding, which consequently increases the exposure of aromatic side chains to the environment. These shifts are primarily due to solvent polarity. At low concentrations, the hypochromic effect of lupeol on PfHGXPRT could be due to multiple reasons, including aggregation induced by the compound, denaturation of the protein, which leads to the loss of its native structure/or changes in the local environments causing a reduction in the tyrosine absorbance at the wavelength (Cassera et al. 2011).

PfHGXPRT is part of a crucial metabolic pathway in P. falciparum and is essential for the parasite’s survival. This enzyme displays a high affinity for all three substrates, including xanthine, which is not typically preferred by HGXPRT enzymes in most organisms. This distinction sets PfHGXPRT apart from its counterpart in human red blood cells (Downie et al. 2008). Combined with its well-characterized structure and the promising results shown in previous studies (Minnow et al. 2022; Opoku et al. 2019; Roy et al. 2015; Grube et al. 2022; Karmakar et al. 2016), PfHGXPRT is a suitable target for this study, among other proteins. The in vitro studies using BA, LU, and MBA with PfHGXPRT showed promising antiplasmodial activity. Consequently, the binding interactions of these compounds with PfHGXPRT were further investigated through in silico molecular docking simulations. The binding energy and the interacting amino acid residues of PfHGXPRT with each compound are shown in Table 4. The predicted binding modes of each of the compounds with PfHGXPRT, along with the 2D and 3D interactions, are shown in Fig. 1. Lupeol, which exhibited the most potent antiplasmodial activity against P. falciparum (IC₅₀ = 7.56 µg/mL), showed the best binding affinity to PfHGXPRT, with a low binding energy of − 7.6 kcal/mol. This compound had hydrophobic interactions with the Phe18, Ser192, Ile175, Trp181, Lys185, Pro16, Val189, and Asp15 residues. Betulinic acid (20 µg/mL) and methyl betulinate (10.11 µg/mL) showed the same binding affinity of − 7.2 kcal/mol and, similar to lupeol, had hydrophobic interactions. Betulinic acid formed hydrophobic interactions with Lys114, Leu231, Ser230, Thr229, Asn206, Lys227, Ala228, Tyr226, and Glu207, while methyl betulinate formed interactions with residues Lys114, Leu231, Ser230, Thr229, Asn206, Lys227, Ala228, Tyr226, and Glu207.

Betulinic acid and methyl betulinate interacted with two amino acid residues, Glu207 and Asn206, which are part of the active loop IV of the enzyme, with the latter forming a Cα bond with Asn118, part of loop II. Loop II is involved in gatekeeping for substrate binding and product release processes (Patil et al. 2010). Methyl betulinate had an additional interaction with Tyr116, which also forms part of loop II and is predicted to interact with the ribose ring of the purine, thereby stabilizing the carbocation intermediate during enzyme catalysis (Patil et al. 2010). Lupeol docked on the protein’s outer surface, forming hydrophobic interactions with Phe18, Ser192, Ile175, Trp181, Lys185, Pro16, Val189, and Asp15 (Fig. 1). The compound primarily interacted with loop III, which contains amino acid residues 180 to 184 (Patil et al. 2010). Loop III acts as the connection between the hood and domain of the protein and is responsible for regulating catalytic events, specifically by inhibiting water molecules from entering the reaction center (Patil et al. 2010). The results suggested that the compounds could bind to the allosteric sites of the protein, as they displayed similar binding modes and shared fundamental hydrophobic interactions. The average number of hydrophobic atoms in marketed drugs is 16, highlighting the importance of hydrophobic interactions in drug design (Martínez-Peinado et al. 2021). According to Martinez-Peinado and colleagues, hydrophobic interactions occur primarily in high-efficiency ligands, making them the main driving force in drug–receptor interactions. These interactions increase binding affinity and stabilization between target–drug interfaces (Martínez-Peinado et al. 2021).