A recombinant expression system for the Plasmodium falciparum proteasome enables structural analysis of its assembly and the design of selective inhibitors

Abstract

The Plasmodium falciparum 20S proteasome (Pf20S) has emerged as a promising antimalarial target. Development of therapeutics to this target has previously relied on native purifications of Pf20S, which is challenging and has limited the scope of previous efforts. Here, we report an effective recombinant Pf20S platform to facilitate drug discovery. Proteasome assembly was carried out in insect cells by co-expressing all fourteen subunits along with the essential chaperone homolog, Ump1. Unexpectedly, the isolated proteins consisted of both a mature and an immature complex. Cryo-EM analysis of the immature complexes revealed structural insights detailing how Ump1 and the propeptides of the β2 and β5 subunits coordinate β-ring assembly, which differ from human and yeast homologs. Biochemical validation confirmed that β1, β2, and β5 subunits of the mature proteasome were catalytically active. Clinical proteasome inhibitors, bortezomib, carfilzomib and marizomib were potent but lacked Pf20S selectivity. However, the tripeptide-epoxyketone J-80 inhibited Pf20S β5 with an IC₅₀ of 22.4 nM and 90-fold selectivity over human β5. Structural studies using cryo-EM elucidated the basis for the selective binding of J-80. Further evaluation of novel Pf20S-selective inhibitors such as the reversible TDI-8304 and irreversible analogs, 8304-vinyl sulfone and 8304-epoxyketone, confirmed their potency and selectivity over the human constitutive proteasome. This recombinant Pf20S platform facilitates detailed biochemical and structural studies, accelerating the development of selective antimalarial therapeutics.

Figure 1.. Characterization of recombinant Pf20S proteasome.

a. Native-PAGE analysis of insect cell lysates following baculovirus co-infection, labeled with Me4BodipyFL-Ahx3Leu3VS fluorescent activity-based probe. b. Native-PAGE gel of purified recombinant Pf20S proteasome. Gels were visualized by silver staining (total protein) and fluorescent scanning (active proteasome complexes). STD is a protein standard. c. Proteasome subunit abundance in the upper (U) and lower (L) excised gel bands. Detailed proteomic data are provided in the Source Data file.

Figure 2.. Incorrectly assembled half-Pf20S displays distinct β2 and β5 propeptides compared to the human 20S.

a. Cryo-EM map of incorrectly assembled half-Pf20S, showing top-down view and bottom-up view. Scale bar, 13 Å. b. Atomic model of the half Pf20S, illustrating Ump1 and β2 and β5 propeptides positioned within the antechamber formed by the α-ring and β-ring. Scale bar, 13 Å. c. Interaction between Pf20S Ump1 and N-terminal helix of β2 propeptide. Scale bar, 5 Å. d. Overlay of β2 propeptides from human 20S (grey) (PDB ID: 7NAN) and Pf20S (blue) (PDB ID: 9Y0L). Red box highlights the N-termini of the two β2 propeptides, which adopt different conformations. Scale bar, 10 Å. e. Stabilization of the β2 propeptide by Ump1. Scale bar, 20 Å. f. Atomic model of Pf20S β5 (grey) wth the propeptide region highlighted (pink). Red box highlights the N-terminal region of β5 propeptide. Scale bar, 10 Å. g. An α helix (pink) within the N-terminal region of the β5 propeptide interacts with the β6 (green) and β5 (grey) subunits. Scale bar, 20 Å.

Figure 3.. Biochemical characterization of proteasome activity.

a. SDS-PAGE analysis of proteasome samples. Upper panel: fluorescence scan of protein gel labelled with Me4BodipyFL-Ahx3Leu3VS activity-based probe revealing the β2 and β5 catalytic subunits. Lower panel: silver-stained gel showing molecular weight marker and equal protein loading in each lane. b. Time-course analysis of proteasome caspase-like (β1) activity using Ac-nPnD-amc fluorogenic substrate. Activity measured in the presence of proteasome inhibitors. Data shown as the change in relative fluorescence units (RFUs) over time. c. Concentration-response curves showing the effect of PA28α regulatory subunit on proteasome activity. Different proteasome subunits (β5, β2, β1) show distinct responses to increasing PA28α concentrations (0-800 nM). Assays were performed in technical replicates (n = 3).

Figure 4.. Testing clinical proteasome inhibitors.

Chemical structures and concentration response assays for a. bortezomib, b. carfilzomib and c. marizomib against Pf20S (solid lines) and human c20S (dashed lines). Assays were performed in technical triplicate reactions using β5, β2 and β1 specific substrates. Assays were performed in technical replicates (n = 3).

Figure 5.. J-80 covalently interacts with the β5 substrate binding pocket of Pf20S.

a. Structure of J-80 and concentration response assay against Pf20S (solid lines) and human c20S (dashed lines). Assays were performed in technical triplicate reactions using β5, β2 and β1 specific substrates. b. The cryo-EM density map with the β5 subunit highlighted in light blue. Scale bar, 10 Å. c. Ribbon structure of Pf20S. In the inset, J-80 (orange) (PDB ID: 9Y0K) is bound to the β5 catalytic Thr1, forming a ring. Interactions occur between J-80 and the substrate binding pocket of β5 (light blue ribbon) and β6 (grey ribbon). Scale bar, left: 10 Å, right: 2.5 Å. Enzymatic assays were performed in technical replicates (n = 3)

Figure 6:. Development of Pf20S-selective inhibitors.

Chemical structure and concentration response assays with a. EY-4-78 and b. TDI-8304. c. Structure of TDI-8304 bound to the β5 subunit of Pf20S (PDB ID: 8G6E). Scale bar, 2.5 Å. d. Structure of J-80 bound to the β5 subunit of Pf20S (PDB ID: 9Y0K). Scale bar, 2.5 Å. e. Irreversible binding of J-80 to Thr1. f. TDI-8304 binds reversibly to β5 but does not directly interact with Thr1. Chemical structure and concentration response assays with g. 8304-epoxyketone and h. 8304-vinyl sulfone. All assays were performed in technical triplicate reactions using β5, β2 and β1 specific substrates.