Plasmepsin X activates the PCRCR complex of Plasmodium falciparum by processing PfRh5 for erythrocyte invasion

Abstract

Plasmodium falciparum causes the most severe form of malaria in humans. The protozoan parasite develops within erythrocytes to mature schizonts, that contain more than 16 merozoites, which egress and invade fresh erythrocytes. The aspartic protease plasmepsin X (PMX), processes proteins and proteases essential for merozoite egress from the schizont and invasion of the host erythrocyte, including the leading vaccine candidate PfRh5. PfRh5 is anchored to the merozoite surface through a 5-membered complex (PCRCR), consisting of Plasmodium thrombospondin-related apical merozoite protein, cysteine-rich small secreted protein, Rh5-interacting protein and cysteine-rich protective antigen. Here, we show that PCRCR is processed by PMX in micronemes to remove the N-terminal prodomain of PhRh5 and this activates the function of the complex unmasking a form that can bind basigin on the erythrocyte membrane and mediate merozoite invasion. The ability to activate PCRCR at a specific time in merozoite invasion most likely masks potential deleterious effects of its function until they are required. These results provide an important understanding of the essential role of PMX and the fine regulation of PCRCR function in P. falciparum biology.

Fig. 1. Mutation of PMX cleavage site in PfRh5 reveals alternative processing.

a Construction of P. falciparum parasites expressing mutant PfRh5 at the PMX cleavage recognition sequence. Wild-type cleavage is shown with the mutant forms listed in the grey box. Cas9 cut site in pfrh5 is shown. Pfrh5 genes containing wt or mutant PMX cleavage sites were HA-tagged. Right-pointing arrow is the endogenous pfrh5 promoter. Black stalk depicts endogenous pfrh5 terminator. White stalk depicts P. berghei dhfr-ts terminator. Selection for WR99210-resistant parasites encoded by human dhfr. The five transfected genes integrated into the genome (Y = Yes). b Schematic of PfRh5 cleavage site residues and updated sequence logo (weblogo.berkeley.edu) for PMX cleavage sites. c Mutation of PfRh5 cleavage site identifies alternative processing sites. Immunoblots were probed with anti-HA mAb and anti-hsp70 for protein quantitation. Right panels show schematics of predicted unprocessed protein (p64), predicted PMX-cleaved product (p50) and observed cleavage fragments in the mutant parasites (p54 and p53). Cleavage at NFLQ site generating p50 highlighted in blue. The mean values for unprocessed/processed band intensities are shown. Band intensities for PfRh5 unprocessed/processed were calculated from the five parasite lines in three independent experiments. d Effect of mutations in endogenous NFLQ motif on cleavage by PMX. Mutation of either the P1’ or P1 residue block NFLQ cleavage and generation of the p50 product. e Processing Inhibition Assays (PIA) with Rh5-NFAA parasites, using PMX-specific inhibitor WM4 and the PMIX/PMX dual inhibitor WM382 probed with anti-HA mAb. Lower panel shows schematic of observed products. Band intensities for PfRh5 unprocessed/processed were calculated from Control and WM4 merozoites for three independent experiments. Error bars represent standard deviation. f First 160 amino acids of PfRh5 showing putative alternative cleavage sites and mutated NFAA sequence. Lower-case letters are the signal sequence. g Construction of P. falciparum parasites with C-terminally HA-tagged prodomain (Rh5PD-CHA). HA-tag was placed N-terminal to the ‘NFLQ’ PMX cleavage site. h PIA with Rh5PD-CHA parasites using Compound 1 (C1) and probed with anti-HA mAb. Schematic shows expected and observed proteins. Putative non-preferred (alternative) cleavage sites (VFNQ & LLNE) and the known PMX site (NFLQ) are shown. The p17 product is the HA-tagged prodomain (PD), p7 product the HA-tagged ‘VFNQ-to-NFLQ’ protein and p6 the ‘LLNE-to-NFLQ’ protein. Source data are provided as a source data file.

Fig. 2. PfRh5 N-terminal domain and processing by PMX are essential.

a Rh5-NFLQ parasite were grown+/− Rapa and probed with anti-HA and anti-FLAG mAbs. Right schematics show expected proteins in Control and Rapa conditions. Rh5-NFLQ proteins (uncleaved p63 and cleaved p49) detected with FLAG mAbs and unprocessed p64 and processed p50 PfRh5 proteins detected with HA mAbs. Protein knockdown signified by dimmed schematic. b Parasitaemia following+/− Rapa-treatment at 4 hr post-invasion (hpi) for Rh5iKO/Rh5-NFLQ parasites. Mean of three technical replicates+/− SD and representative of two experiments. c Rh5iKO/Rh5-NAAQ parasite grown+/− Rapa probed with anti-HA and anti-FLAG mAbs. Schematics show expected proteins in Control and Rapa conditions. Rh5-NAAQ proteins (uncleaved p63) and cleaved at alternative sites (p53/p52) are detected with FLAG mAbs and the unprocessed p64 and processed p50 PfRh5 proteins detected with HA mAbs. Knockdown of these proteins signified by a dimmed schematic. d Parasitaemia following+/− Rapa-treatment at 4 hr post-invasion (hpi) for the Rh5iKO/Rh5-NAAQ parasite. Mean of three technical replicates+/− SD and representative of two experiments. e Rh5iKO/Rh5-tmut parasite grown/− Rapa were probed with anti-HA, anti-FLAG and PfRh5 mAbs. Shown are schematics for expected proteins under Control and Rapa conditions. Rh5-tmut proteins (uncleaved p63 and cleaved p53/p54) detected with FLAG mAbs and the unprocessed p64 and processed p50 Rh5 proteins detected with HA mAbs. PfRh5 mAb detects both Rh5iKO and Rh5-tmut proteins to confirm the p53/54 bands detected with FLAG mAbs. Knockdown of proteins under Rapa conditions signified by dimmed schematic. f Parasitaemia following+/− Rapa-treatment at 4 hr post-invasion (hpi) for the Rh5iKO/Rh5-tmut parasite. Mean of three technical replicates+/− SD and representative of two experiments. g Rh5iKO/Rh5ΔPDcomp parasites grown+/− Rapa probed with anti-HA and anti-FLAG mAbs. On right are schematics for expected proteins in Control and Rapa conditions. Rh5ΔPD protein (p49) is detected with FLAG mAbs and unprocessed p64 & processed p50 Rh5 proteins are detected with HA mAbs. Protein knockdown signified by dimmed schematic. h Parasitaemia following+/− Rapa-treatment at 4 hr post-invasion (hpi) for Rh5iKO/Rh5ΔPDcomp parasite. Mean of three technical replicates+/− Standard Deviation (SD) and representative of two experiments. Source data are provided as a source data file.

Fig. 3. Post-Golgi trafficking of PfRh5 does not require interaction with a PfRh5-specific protein.

a P. falciparum parasites with a N-terminally HA-tagged prodomain (Rh5PD-NHA) were generated (Fig. S6). Synchronised 3D7 and Rh5PD-NHA were grown with WM382 (to prevent Rh5PD cleavage) to the schizont stage, uninfected erythrocytes lysed with saponin, then parasites treated (+/−) the DSP crosslinker. Twelve samples representing four in triplicate for Mass Spectrometry (MS) were prepared. The samples were: 3D7, 3D7 + DSP, Rh5PD-NHA and Rh5PD-NHA + DSP. Proteins were immunoprecipitated with agarose-bound anti-HA antibodies, trypsin digested and analysed by MS. The results of the comparison between Rh5PD-NHA + DSP and 3D7 + DSP are shown. All components of the PCRCR complex were detected (ie. PTRAMP, CSS, Ripr, CyRPA, Rh5). In addition, four other proteins with high significance were found: 10TM (PF3D7_1208100), GRP170 (PF3D7_1344200), Sel1 repeat protein (PF3D7_0204100) and P113 (PF3D7_1420700). b, f, i, l Representations for the four proteins targeted for inducible knockout using the DiCre system are shown. For P113 and GRP170, the HA-tag was placed N-terminal to the GPI anchor or the ‘KDEL’ ER retention signal, respectively. For Sel1 protein and 10TM, the HA-tag was placed at the C-terminus of the protein. The position of the loxP sites shows how much of each protein is removed in the presence of Rapa. c, g, j, m Constructs for each gene were transfected into the Pfs47-DiCre line to allow inducible knockout in the presence of Rapa. Parasitaemia following+/− Rapa-treatment at 4 hr post-invasion (hpi) for the four parasites is shown. Shown are data points for one experiment. d, e, h, k, n Proteins from synchronised schizonts from each parasite grown with or without Rapa, were blotted, then probed with anti-HA, anti-Rh5 and anti-hsp70 Abs. Below the immunoblots are schematics for the expected proteins to be produced upon probing with the anti-HA and anti-Rh5 mAb. Note that a processing event was detected for the P113 and Sel1 proteins. o Giemsa-stained images of the GRP170-iKO parasite at t = 47 hr, in the control and Rapa conditions. p Giemsa-stained images of the 10TM-iKO parasite through 2 asexual life cycles, under both control and Rapa conditions. Source data are provided as a source data file.

Fig. 4. The PCRCR complex is processed by PMX and formed in micronemes or a microneme subset.

Immunoblots on proteins found in different compartments of the merozoite: microneme, rhoptry neck, rhoptry body and exoneme. Processing Inhibition Assays were set up with either 3D7, PMX-HA and PMIX-HA parasites. The derivation of PMIX-HA, and PMX-HA parasites has been described. All assays included C1. Supernatant (S or supe) and merozoite (M or mero) fractions are shown. Below each blot is a diagram of the expected products under Control and C1 conditions. a Immunoblot of the rhoptry neck protein, PfRh5. b Immunoblot of the microneme protein, PTRAMP. c Immunoblot of the microneme protein, PfRipr. d, e Immunoblots of the microneme proteins, PfCSS and PfCyRPA. f Immunoblot of the exoneme and microneme protease, PMX. g Immunoblot of the rhoptry body protein, PMIX. h, i Immunoblots of the microneme proteins, AMA1 and EBA140. j Immunoblot of the rhoptry neck protein, PfRh2. All proteins were separated under reducing conditions, except the PfRipr proteins. The approximate location of the mAb or pAb used for the immunoblots is shown in the schematic for each protein, together with the calculated size of the unprocessed or processed product in kDa. The asterisk (*) indicates a cross-reaction of some Abs with albumin in supernatant fractions. k WM4 blocks exoneme and microneme discharge. The merozoite fractions from a Processing inhibition assay on PMX-HA parasites treated with either C1 or the PMX inhibitor, WM4, were saponin-lysed. Proteins from equal volumes of the saponin supernatant (S) and saponin pellet (P) were separated by SDS-PAGE and probed with either HA or Rh5 mAbs. The illustrations of C1 and WM4-treated merozoites are identical, since the data shows that WM4 just like C1, blocks exoneme and microneme discharge. The blue boxes delineate the proteins detected in the saponin pellets (i.e., remaining within the merozoites) under both treatments. Source data are provided as a source data file.

Fig. 5. The role of the N-terminal prodomain of PfRh5 in the PCRCR complex.

a Binding of CyRPA to either Rh5_unproc or Rh5_proc was determined by Biolayer Interferometry (BLI). Biotinylated CyRPA was pre-added to a Streptavidin biosensor, before dipping into twofold dilutions of Rh5_unproc and Rh5_proc at t = 0 s. Binding occurred for 120 s followed by dissociation for a further 120 s. The K_D values were calculated using Octet Data Analysis v11.0 software. b Binding of Rh5_unproc and Rh5_proc to the RCR complex by BLI. CyRPA was added at t = 0 s to biotinylated Ripr immobilised on a Streptavidin biosensor for 120 s before allowing dissociation for a further 120 s. Either Rh5_unproc or Rh5_proc were added to Ripr-CyRPA and allowed to bind, then dissociate, before PTRAMP-CSS (PC) was added at t = 480 s. Binding and dissociation were allowed to occur until t = 720 s. c Binding of Rh5_unproc or Rh5_proc and PCRC-Rh5_unproc or PCRC-Rh5_proc to erythrocytes measured by flow-cytometric analysis. The assay was done three times and means (+/−) standard error of the mean (SEM) is shown. One-way ANOVA with Sidak’s multiple comparisons test was used to calculate p values. d Representation of the binding results in c using known structures of CSS (PDB ID: 7UNZ), PfRipr, CyRPA, PfRh5 (PDB ID: 6MPV), basigin (BSG) (PDB ID: 3B5H) and the AlphaFold Monomer v2.0 structures of the PTRAMP ectodomain (Uniprot: Q8I5M8) and Rh5PD (UniProt: Q8IFM5). Structures were assembled in ChimeraX (https://www.rbvi.ucsf.edu/chimerax). Source data are provided as a source data file.

Fig. 6. Model of protein processing and organellar discharge during merozoite egress and invasion.

a Merozoite of a schizont bounded by parasite plasma membrane (PPM) within the PVM and erythrocyte membrane. PfSUB1 and PMX are stored in exonemes. C1 blocks exoneme and microneme discharge, while PMX inhibitor WM4 and dual PMX/PMIX inhibitor WM382 block parasitophorous vacuole membrane (PVM) and erythrocyte membrane degradation as SUB1 is inactive. b Following exoneme fusion and discharge, SUB1 and PMX are discharged into the parasitophorous vacuole (PV). Here, SUB1 processes several PV-resident proteins (such as SERA5) culminating in a ruptured PVM and a permeabilised erythrocyte membrane. A subclass of microneme is postulated to contain EBAs, PfRhs, AMA1, PMX and PCRCR. In this study, ~30% of PfRh5 was PMX-processed to the Rh5_proc form, while 70% remains unprocessed, as depicted in the microneme contents. Rh5_proc: PMX-processed PfRh5; Rh5_unproc: Unprocessed PfRh5; Rh5PD; PfRh5 prodomain. c Following microneme fusion and discharge, the EBAs, PfRhs and PCRCR complex are translocated to the parasite membrane at the apical tip, while AMA1 initially inserts into the merozoite membrane, then spreads over the merozoite surface. Further membrane lytic events result in a degraded PVM and perforated erythrocyte membrane. d, e Merozoites egress from the erythrocyte ready to engage fresh erythrocyte. PfRh and EBAs are involved in early invasion steps, followed by binding of PCRCR to basigin (BSG), which culminates in release of RON complex proteins into the RBC membrane. Processed AMA1 on the merozoite surface engages with inserted RON complex, to form the tight junction for merozoite entry. We propose that SUB2 ‘sheddase’ is contained in a second microneme subset also containing PMX, that fuses at the rhoptry neck and merozoite membrane, releasing activated SUB2 to shed proteins from the merozoite surface. f Proteins from the merozoite fraction of a 3D7 processing inhibition assay treated (+/−) C1, were probed with an anti-Rh5 mAb and the PfRh5 unprocessed and processed band intensities calculated. Mean values are shown for the level of PfRh5 processing, shown in the histogram below. Band intensities for PfRh5 unprocessed/processed were calculated from Control and C1-treated merozoites in three independent experiments. Error bars represent standard deviation. g, h Data comparing PfRh5 processing levels in C1-treated merozoites (Fig. 6f) suggests two possible models for microneme heterogeneity. The first model (g) posits there are two early microneme subsets (type 1 and 2), only one of which contains PMX. Under control conditions, this subset fuses at the rhoptry neck before PfRh5 has been completely PMX-processed, giving a mix of 28% processed and 34% unprocessed. The other subset has no PMX, so upon rhoptry neck fusion, contributes a further 38% unprocessed PfRh5. Under C1 conditions where there is no rhoptry fusion, PfRh5 processing in the microneme subset goes to completion at 62% and leaving 38% unprocessed. The second model (h) is like the microneme subset model, except that the microneme is compartmentalised into two sections, only one of which contains PMX, but both compartments contain PfRh5. Source data are provided as a source data file.