Characterization and Inhibition of the Chaperone Function of Plasmodium falciparum Glucose-Regulated Protein 94 kDa (Pf Grp94)

Abstract

Plasmodium falciparum expresses four heat shock protein 90 (Hsp90) members. Among these, one, glucose-regulated protein 94 (PfGrp94), is localized in the endoplasmic reticulum (ER). Both the cytosolic and ER-based Hsp90s are essential for parasite survival under all growth conditions. The cytosolic version has been extensively studied and has been targeted in several efforts through the repurposing of anticancer therapeutics as antimalarial drugs. However, PfGrp94 has not been fully characterized and some of its functions related to the ER stress response are not fully understood. Structural analysis of the recombinant full-length PfGrp94 protein showed a predominantly α-helical secondary structure and its thermal resilience was modulated by 5'-N-ethyl-carboxamide-adenosine (NECA) and nucleotides ATP/ADP. PfGrp94 exhibits ATPase activity and suppressed heat-induced aggregation of a model substrate, malate dehydrogenase, in a nucleotide-dependent manner. However, these PfGrp94 chaperone functions were abrogated by NECA. Molecular docking and molecular dynamics (MD) simulations showed that NECA interacted with unique residues on PfGrp94, which could be potentially exploited for selective drug design. Finally, using parasites maintained at the red blood stage, NECA exhibited moderate antiplasmodial activity (IC₅₀ of 4.3, 7.4, and 10.0 μM) against three different P. falciparum strains. Findings from this study provide the first direct evidence for the correlation between in silico, biochemical, and in vitro data toward utilizing the ER-based chaperone, PfGrp94, as a drug target against the malaria parasites.

Keywords: Plasmodium falciparum; NECA; characterization; endoplasmin; glucose regulated protein 94 (Grp94); inhibition; malaria.

FIGURE 1

PfGrp94 protein expression and purification. SDS‐PAGE analysis of PfGrp94 expression in E. coli XL1‐Blue cells (A) and the purification of recombinant PfGrp94 protein using immobilized metal‐affinity chromatography (B). Western blot analysis of the expression and purification of the protein using polyclonal rabbit‐raised α‐PfGrp94 antibodies (lower panels). Lane M: protein molecular weight marker (kDa); Lane C: control (untransformed E. coli XL1‐Blue cells); Lane 0: total extract for cells transformed with pQE30/PfGrp94 before IPTG induction; Lanes 6 and O/N: total lysate obtained 6 and 16 h respectively post‐induction with 1 mM IPTG at 16°C; Lane L: cell lysate; Lane F: flowthrough; Lane W: wash containing unbound proteins, Lanes E1 and E2: fraction eluted at 100 and 500 mM imidazole, respectively.

FIGURE 2

Secondary structure analysis of recombinant PfGrp94. The CD spectrum of PfGrp94 at 25°C (A). The relative folded protein fraction in phosphate buffer at different pH (B) and in the presence of varying amounts of chaotropic agents urea and guanidine HCl (C) monitored at 194, 208, and 222 nm. The folded fraction of PfGrp94 subjected to variable incubation temperature which was raised monotonically from 25°C to 90°C in 5°C increments in the presence and absence of nucleotides or inhibitor NECA (D). The relative folded fraction was determined from the fraction of the ellipticity at 222 nm measured at increasing temperatures compared to that recorded at ambient temperature.

FIGURE 3

Tertiary structure analysis of PfGrp94. The differential scanning fluorimetry melting curves of PfGrp94 in buffer supplemented with 10 μM ATP, ADP, or NECA, as indicated (A) were used to determine the inflection points, which relate to the melting temperatures, are indicated with arrows the corresponding first derivative plots of the sigmoidal melting curves of PfGrp94 (B). The peaks of the curves reflect the melting temperatures. The SDS‐PAGE analysis of PfGrp94 proteolytic degradation (C) and the densitometric analysis of the band intensities (D). Lane MW: molecular weight ladder (kDa), Lane 0: apo‐PfGrp94 without protease, lanes marked 5 and 15 indicate PfGrp94 digested with proteinase K for 5 or 15 min, respectively in the absence (apo‐PfGrp94) and presence of ADP, ATP or NECA, as indicated. The results are representative of three independent repeats and error bars indicate standard deviation.

FIGURE 4

Analysis of the chaperone activity of PfGrp94. (A) The dose–response curves for the inhibition of the ATPase activity of PfGrp94 with NECA. (Inset) Michaelis–Menten's analysis of the PfGrp94 basal ATPase hydrolysis rate was determined by a colorimetric malachite green assay. The ATPase activity was calculated in nmol Pi/min/μM of PfGrp94 protein. (B) The suppression of MDH‐aggregation assay in the presence of varying amounts of PfGrp94. The aggregation suppression assay of equimolar MDH and PfGrp94 in the presence of varying amounts of nucleotides ATP (C), ADP (D), and the inhibitor NECA (E), normalized to the aggregation percentage of MDH alone. Error bars represent the standard deviation.

FIGURE 5

The predicted docking poses for NECA binding on PfGrp94 and Hsp90 NTDs. The 3D binding pocket and 2D representation of the interacting residues of PfGrp94 with nucleotides (A) ATP, (B) ADP, and (C) NECA within a distance of less than 4 Å stabilizing the complex, as predicted using Schrödinger Maestro 2022.1.

FIGURE 6

NECA forms stable interactions with PfGrp94 during MD simulation. (A) The R _g of PfGrp94 Cα atoms. (B) The RMSD plot of apo‐PfGrp94 Cα atoms and with nucleotides or NECA over a simulation period of 250 ns. (C) Protein RMSF of apo‐PfGrp94 and in complex with nucleotides or NECA. (D) The ligand RMSD graph showing the stability of the NECA or nucleotides with respect to the PfGrp94 binding site. PfGrp94 amino acid residues that interact with the ligand (E) ATP, (F) ADP, and (G) NECA which are marked with green‐colored vertical bars; H‐bonds are categorized into backbone acceptor; backbone donor; side‐chain acceptor; side‐chain donor. Hydrophobic interactions are categorized into π–cation; π–π*; and other, nonspecific interactions. The stacked bar charts are normalized throughout the 250 ns trajectory. The images were generated from Maestro v13.1.