A multistage antimalarial targets the plasmepsins IX and X essential for invasion and egress

See allHide authors and affiliations

Science  27 Oct 2017:
Vol. 358, Issue 6362, pp. 522-528
DOI: 10.1126/science.aaf8675

Plasmodium parasite entrance and exit

Sweats and fever are the hallmarks of malaria. Red blood cells are the replication factories for malaria parasites. Fever occurs when the parasites' merozoite stages burst en masse from red blood cells into the circulation. Nasamu et al. and Pino et al. discovered that two parasite proteases, plasmepsin IX and X, are essential for mass cell exit (see the Perspective by Boddey). Plasmepsin X is also used by the merozoites to enter a fresh red blood cell to continue the replicative cycle. These two plasmepsins act by regulating the maturation of enzymes required to disrupt host cell membranes. Because these functions are essential for the parasite, the authors used protease inhibitors to show that plasmepsins provide potential drug targets.

Science, this issue p. 518, p. 522; see also p. 445


Regulated exocytosis by secretory organelles is important for malaria parasite invasion and egress. Many parasite effector proteins, including perforins, adhesins, and proteases, are extensively proteolytically processed both pre- and postexocytosis. Here we report the multistage antiplasmodial activity of the aspartic protease inhibitor hydroxyl-ethyl-amine–based scaffold compound 49c. This scaffold inhibits the preexocytosis processing of several secreted rhoptry and microneme proteins by targeting the corresponding maturases plasmepsins IX (PMIX) and X (PMX), respectively. Conditional excision of PMIX revealed its crucial role in invasion, and recombinantly active PMIX and PMX cleave egress and invasion factors in a 49c-sensitive manner.

Malaria remains a major cause of mortality worldwide, and resistance to existing antimalarials is a growing problem, which requires the development of new drugs urgently. Aspartic proteases are potential targets for chemotherapy (1) and key contributors to Plasmodium falciparum pathogenicity (2, 3). P. falciparum possesses a repertoire of 10 aspartic proteases, named plasmepsins (PfPMI to X). PfPMIX and PfPMX are expressed in mature blood-stage schizonts and invasive merozoites and fulfill indispensable but unknown functions. The activity of several serine and cysteine proteases promotes the destabilization of the parasitophorous vacuole membrane (PVM) and red blood cell (RBC) membranes that surround the parasite (4). Egress is followed by invasion of a fresh RBC, a process that takes 10 to 30 s. Invasion also crucially relies on serine proteases to activate or remove ligands involved in interactions with the host erythrocyte (5).

To study the role of aspartic proteases during egress and invasion, we used a hydroxyl-ethyl-amine scaffold that inhibits aspartic proteases by mimicking the tetrahedral intermediate of hydrolysis (6). Compound 49c (Fig. 1A) is such a peptidomimetic competitive inhibitor and has been found to be effective against P. falciparum in vitro and the rodent parasite Plasmodium berghei in vivo (7, 8). This compound has a modest effect after 24 hours of treatment [median inhibitory concentration (IC50) > 500 nM] and a considerably greater effect after 72 hours (IC50 of 0.6 nM), indicating inhibition occurs at a specific life-cycle stage. P. falciparum cultures treated at ring stage with 1 nM 49c showed no difference compared to controls during the first 24 hours, but, in contrast, did show a total disappearance of the parasites after three days (Fig. 1B). The killing profile of 49c is comparable to chloroquine, with a 99.9% parasite clearance (9) achieved at 48 hours of treatment (Fig. 1C). Importantly, 49c did not affect intraerythrocytic development and allowed the production of microscopically normal schizonts that were, however, not released from the host cell (Fig. 1D). Treatment 5 hours before egress was sufficient to inhibit egress, whereas treatment for 3 hours had no notable effect (Fig. 1E). Removal of 49c 1 hour before egress did not release the block, whereas washing it out 5 hours before egress totally rescued the phenotype (Fig. 1F), confirming that 49c acts during late schizogony to block egress but does not prevent intraerythrocytic development.

Fig. 1 Compound 49c prevents P. falciparum merozoite egress and invasion.

(A) Structure of compound 49c. (B) Wild-type (WT) P. falciparum parasites were treated with 1 nM 49c or dimethyl sulfoxide (DMSO), and parasitemia was quantified daily over a 6-day period by counting from Giemsa-stained blood smears. Error bars show the SD of three replicates from three independent experiments. (C) P. falciparum viability time-course profiles in the presence of 49c (10 nM), chloroquine (CQ, 100 nM), and atovaquone (ATV, 10 nM). (D) WT P. falciparum parasites were treated with 1 nM 49c for 6 or 40 hours. DMSO-treated control parasites reinvaded, whereas 49c-treated parasites were blocked at a fully mature schizont stage. Scale bars, 2 μm. (E and F) P. falciparum cultures were treated with 1 nM 49c at different time points before egress; ring and schizont stages were quantified. Compound 49c acts on its aspartic protease targets between 5 and 3 hours before egress. In (F), the compound was washed out 5 or 1 hours before egress. (G) Scheme of PfSUB1 and PfAMA1 maturation steps; the cleavage sites and the proteases responsible are indicated when known. The resulting products are shown, and their molecular masses are indicated. The pro-sequence of PfAMA1 is shown in pink. Ab1 and 2 show the major forms recognized by the two antibodies used. (H) Immunoblot evaluating the processing of PfSUB1 upon treatment with either DMSO or 49c. P. falciparum blood-stage parasites were treated with 49c for 6 hours. Parasites were allowed to egress for 15 or 30 min. The p54 precursor and the p47 active form of PfSUB1 are indicated with arrows. A very small proportion of the p54 PfSUB1 precursor was converted to the p47 active form when parasites were treated with 1 nM 49c (indicated with an asterisk). The lower 40-kDa band likely corresponds to a degradation product of PfSUB1. (I) Electron micrograph of DMSO- or 49c-treated P. falciparum schizont stage parasites. WT parasites were used and fixed at the time of egress initiation or 30 min after. Bottom images show zoom-ins of red dashed areas in the images above. Black arrows highlight the RBC plasma membrane. Red arrowheads label the PVM. Scale bars, 2 μm. (J) WT parasites were treated with 1 nM 49c for 40, 5, or 1 hours before being mechanically released and allowed to invade. (K) Immunoblots evaluating the processing of PfAMA1 after treatment with either DMSO or 49c. P. falciparum blood-stage parasites were treated with 49c (1 nM) for 6 hours. Parasites were allowed to egress for 15 or 30 min. The p83, p66, and p44 forms are indicated with arrows. A polyclonal anti-PfAMA1 antibody serum recognizing predominantly the precursor of PfAMA1 was used (39). In the immunoblot on the right, a serum recognizing the p44 processed form was used (40).

Plasmodium egress from infected red blood cells (iRBCs) is a two-step process, initiated by the disruption of the PVM followed by the erythrocyte membrane. These two steps require the serine protease PfSUB1 (10, 11), which undergoes at least two proteolytic processing events during its maturation to produce the mature p47 (47-kDa) form (Fig. 1G) (12). Treatment of parasites with 10 nM 49c prevented the p54-to-p47 transition, whereas 1 nM 49c resulted in traces of mature p47 PfSUB1 (Fig. 1, G and H). 49c had no effect on the trafficking and secretion of PfSUB1 from exonemes, as no difference was observed when comparing 49c-treated and control egressing schizonts (fig. S1A), and did not inhibit the enzymatic activity of recombinant PfSUB1 in vitro (fig. S1B). PfSUB1 governs egress by processing the merozoite surface protein PfMSP1 (11) and serine-rich antigen (SERA) family proteins in the parasitophorous vacuole (PV) (fig. S1I) (13). Both PfMSP1 and PfSERA5 remained unprocessed in parasites treated with 49c, indicating that PfSUB1 was inactive (fig. S1, C to E). We visualized the effect of 49c on PVM breakdown using parasites expressing a green fluorescent protein (GFP) fusion of the soluble PV protein PfPVI (PfPVI-GFP) (14). When the PVM ruptures, pores form in the iRBC membrane, leading to the disappearance of the GFP signal (15). In control parasites, the extremely short period between PVM rupture and egress could not be observed. By contrast, parasites treated with 1 nM 49c were able to break the PVM but remained trapped within the RBCs, whereas 10 nM 49c completely blocked PVM rupture (fig. S1F). These results were confirmed by transmission electron microscopy (Fig. 1I).

We assessed the effect of 49c on erythrocyte invasion by mechanically releasing merozoites (16). Treatment with 49c for >5 hours impaired invasion, whereas a 1-hour treatment had no notable impact (Fig. 1J). Invasion critically relies on the formation of a moving junction composed of the apical membrane antigen 1 (PfAMA1) and rhoptry proteins (PfRONs) (17). PfAMA1 (18) is a microneme integral membrane protein, which is processed by the action of an unknown protease at its N terminus to generate the secreted p66 form from a p83 precursor (fig. S1G). This event occurs before exocytosis and appears to be a prerequisite for PfAMA1 secretion (19). Consistently, 49c abrogated the processing of PfAMA1, resulting in accumulation of the p83 precursor (Fig. 1K) without affecting PfAMA1 trafficking to the micronemes (fig. S1H).

Of these erythrocytic stages, PMIX and PMX are predominantly expressed at the schizont stage, suggesting a role in egress and/or invasion and implicating them as plausible targets for 49c (fig. S2A) (20). The PMIX and PMX genes appeared to be refractory to genetic ablation in both P. falciparum and P. berghei (21). We therefore opted for conditional expression systems, DiCre (22) for PfPMIX and the auxin-inducible degron approach (23) for PbPMIX and PbPMX. The P. berghei inducible knockdowns (KDs) showed only low levels of protein destabilization (fig. S3, A and B). We modified the PfPMIX locus to insert loxP sites and a C-terminal Ty epitope tag in DiCre-expressing parasites (22) (fig. S4, A and B). PfPMIX-Ty only partially colocalized with the secreted proteins PfCyRPA (cysteine-rich protective antigen) and PfRhopH3 (high–molecular weight rhoptry protein 3) but not with PfAMA1 (Fig. 2A). The localization at the proximity of the rhoptries was confirmed by immune electron microscopy (fig. S4, C and D) and is concordant with transcriptome information indicating that they are expressed before secretory organelle proteins (24). Induction of DiCre activity by rapamycin led to the complete disappearance of PfPMIX (Fig. 2B). PfPMIX-deficient parasites became undetectable three days after rapamycin treatment (Fig. 2C). In the absence of PfPMIX, intracellular development occurred normally until the schizont stage (Fig. 2D). Consistent with its expression in schizonts (fig. S2A), ring-stage parasites were considerably reduced by the second cycle after rapamycin treatment, showing a severe default in invasion. This decline in erythrocyte invasion was more pronounced when the samples were treated with trypsin to remove noninvasive adherent merozoites from the host-cell surface (Fig. 2E). Any delay in the egress process results in noninvasive merozoites, probably because of the exhaustion of their secreted protein set (16). The replication defect in parasites lacking PfPMIX did not result from impaired egress (25), as time-lapse video microscopy (11) did not reveal delayed egress (fig. S4C and movies S1 and S2), confirming that PfPMIX is essential for invasion only.

Fig. 2 PfPMIXs play a critical role in RBC invasion.

(A) Immunofluorescence assay (IFA) results showing the localization of PfPMIX relative to the microneme proteins PfAMA1, PfCyRPA, and PfRhopH3. Scale bar, 2 μm. (B) Immunoblot evaluating schizont-stage lysates from the 3D7-DiCre parental line, PfPMIX-Lox, and PfPMIX-KD after rapamycin treatment. Antibodies against PfPRF were used as a loading control. (C) Representative replication curves. 3D7-DiCre parental line and PfPMIX-KD lines were treated with rapamycin at the ring stage, whereas the PfPMIX-Lox control was not. (D) 3D7-DiCre and PfPMIX-Lox parasites were treated with DMSO/rapamycin at the ring stage, and intraerythrocytic development was monitored on Giemsa-stained blood smears 5, 20, and 44 hours postinvasion. (E) The invasion ability of 3D7-DiCre, PfPMIX-Lox, and PfPMIX-KD lines was quantified with (+) and without (–) trypsin treatment to remove noninvasive merozoites adherent on RBCs. Cultures were treated with rapamycin at the ring stage and allowed to mature to the schizont stage. Schizonts were allowed to egress and reinvade for 4 hours before quantification. (F and G) Immunoblots evaluating the processing of PfRAP1 (F) and PfASP (G) after PfPMIX deletion and treatment with either DMSO or 49c. PfPMIX-Lox parasites were treated at the ring stage with rapamycin or 1 nM 49c for 10 hours. When fully mature, schizonts were collected. The p86, p82, and p64 forms of PfRAP1 are indicated with arrows. (H) Typical progress curves showing cleavage in vitro of fluorogenic PfRAP1 substrate by recombinant rPfPMIX. 49b (1 μM), an inactive compound structurally related to 49c, and the aspartyl protease inhibitor pepstatin (10 μM) had no inhibitory effect on rPfPMIX catalytic activity at these concentrations, whereas a robust inhibition was observed in the presence of 1 μM of 49c. rPfPMIX D/A and rPfPMX are used as controls for the assay. RFU, relative fluorescence unit. rPfPMX cleavage of (I) PfAMA1 and (J) PfSUB1 peptides and their inhibition in the presence of 49c (10 nM or 1 μM). The mutant PfSUB1 peptide was not cleaved by rPfPMX. Neither rPfPMX D/A nor rPfPMIX cleaved the PfSUB1 peptide. PfSUB1 amino acid residue sequence, DABCYL-G-SMLEVENDAE-G-EDANS; mutant PfSUB1 sequence, DABCYL-G-SMAAVENDAE-G-EDANS. Amino acid residues: A, Ala; C, Cys; D, Asp; E, Glu; G, Gly; L, Leu; M, Met; N, Asn; S, Ser; V, Val; and Y, Tyr.

We identified the rhoptry associated protein 1 (PfRAP1) and the apical sushi protein (PfASP) as substrates of PfPMIX. Both proteins are targeted to the rhoptry and are extensively processed during their maturation (26, 27). PfRAP1 is converted from a short-lived, 86-kDa precursor into an 82-kDa (p82) form which is then converted to a 67-kDa (p67) form during schizont maturation (28). Importantly, the p86 precursor accumulated in the absence of PfPMIX, as well as upon treatment with 100 nM and 1 μM 49c, suggesting that 49c targets PfPMIX (Fig. 2F) but apparently less potently than the protease responsible for PfSUB1 activation (Fig. 1F). PfASP processing was also impaired in PfPMIX-KD parasites and a 1 nM 49c treatment resulted in similar inhibition (Fig. 2G). A difference in 49c activity against various substrates of PfPMIX is not unexpected, as 49c acts as a peptidomimetic competitive inhibitor (6).

To ensure that both rhoptry proteins represent direct substrates of PfPMIX, we expressed recombinant active PfPMIX and PfPMX (rPfPMIX, rPfPMX) as well as the catalytically dead mutants (rPfPMIX D/A, rPfPMX D/A), where the first catalytic aspartic acid (D) is converted to alanine (A), in baculovirus-infected insect cells. Both rPfPMIX and rPfPMX were active against a Toxoplasma gondii rhoptry protein 1 (ROP1) peptide (29) and sensitive to 49c (fig. S5, A to D). Notably, rPfPMIX, but not rPfPMIX D/A or rPfPMX, showed activity on PfRAP1-derived peptide, but not on a mutant peptide, and was inhibited by 1 μM 49c (Fig. 2H). Concordantly, rPfPMIX was active against PfRAP1 immunoprecipitated from PfPMIX-KD parasites (fig. S6A) but not rPfPMIX (fig S6, G and H). Similarly, PfASP purified from PfPMIX-KD parasite supernatant was efficiently processed by rPfPMIX but not when 1 μM 49c was added to the assay or with rPfPMX D/A or rPfPMX (fig. S6B).

The reticulocyte-binding homolog 5 (PfRh5) binds to erythrocyte basigin, is essential for merozoite invasion (30), and acts in concert with PfRipr (Rh5-interacting protein) and PfCyRPA (31) as well as Pfp113 (32). Whereas PfRh5, PfRipr, and Pfp113 (surface protein P113) are processed and released normally in the absence of PfPMIX, very little PfCyRPA is detectable in the supernatant during egress (fig. S6, C to E). CyRPA is not known to be processed, and the defect observed in PfCyRPA release remains unexplained but might contribute to the loss of invasiveness in the absence of PfPMIX.

Because PfAMA1, PfSUB1, PfSERA5, and PfMSP1 are processed normally in the absence of PfPMIX but not upon 49c treatment (fig. S6, C to E), we hypothesized that PfPMX is responsible for the maturation of PfAMA1 and PfSUB1. We were, however, unable to conditionally knock down PfPMX expression in either P. falciparum or P. berghei (fig. S3, A and B). Instead, we show that rPfPMX cleaves in vitro fluorogenic peptides corresponding to the PfAMA1 p83-to-p66 (33) (Fig. 2I) and the PfSUB1 p54-to-p47 cleavage sites (Fig. 2J), as well as the recombinant PvSUB1 (fig. S6F). By contrast, rPfPMIX is inactive on PfAMA1 and PfSUB1 peptides (Fig. 2, I and J). Notably, taken together (fig. S6, G and H), PfPMIX cleaves rhoptry proteins and PfPMX cleaves microneme and exoneme proteins, respectively. Importantly, 49c inhibited rPfPMX activity in vitro, validating PfAMA1 and PfSUB1 as substrates for PfPMX. Decisively, 49c dually targets PfPMIX and PfPMX and hence provides a rationale for the failure to isolate resistant parasites (fig. S6I).

The antiplasmodial activity of 49c in vivo was characterized by using the rodent model P. berghei. On the basis of the pharmacokinetics of 49c in mice (fig. S7A), we opted for intraperitoneal injection of 100 mg/kg 49c, which sustained blood concentrations higher than 0.2 μM over a 24-hour window. Daily treatment for 4 days cleared parasites from peripheral blood (Fig. 3, A and B). After initial treatment, circulating schizonts accumulated in the blood, confirming that 49c also blocked P. berghei egress from RBCs (Fig. 3C and fig. S7B). No parasites were detectable after 2 weeks of treatment.

Fig. 3 Compound 49c prevents malaria parasite transmission.

(A and B) 49c cures malaria in vivo. Mice were infected with GFP-expressing P. berghei parasites at day 0. From days 2 to 5, mice were injected intraperitoneally (ip) daily with 100 mg/kg 49c. Parasitemia and survival curves are shown. n = 5. pi, postinfection. (C) Mice were infected with WT P. berghei parasites and treated with 100 mg/kg 49c for two consecutive days. Parasitemia and gametocytemia were quantified. (D) Gametocytes isolated from control and 49c-treated mice were activated ex vivo. The exflagellation rate at 20 min postactivation was calculated on the basis of microscopic observations. Note that active flagellated gametes remained trapped in the host erythrocyte. n = 5. (E) Gametocytes isolated from control and 49c-treated mice were activated ex vivo. Egress was quantified by flow cytometry on the basis of the presence of the erythrocyte membrane marker Ter-119 at the time of activation (T0) and 5 and 10 min postactivation. n = 4. (F) Immunoblot showing the expression of PbPMIX-HA and PbPMX-HA in blood-stage schizonts, gametocytes, and ookinetes. Actin was used as a loading control. (G) Immunoblots evaluating the processing of PbSUB1 upon treatment with either DMSO or 49c at the gametocyte stage. Mice infected with P. berghei expressing PbSUB1-HA were treated with 49c (100 mg/kg for 48 hours) and sulfadiazine. Gametocytes were purified and processed for immunoblot analysis. (H) Gametocytes isolated from untreated mice were activated ex vivo in the presence of 1 nM of either DMSO or 49c. The exflagellation rate at 20 min postactivation is shown. n = 3. (I and J) Gametocytes isolated from 49c-treated (I) or untreated (J) mice were activated ex vivo in the presence or absence of 1 nM 49c. Conversion to ookinetes was quantified 24 hours postactivation. n = 5. (K) Immunoblots evaluating the processing of CelTOS and its secretion upon treatment with either DMSO or 49c. Gametocytes expressing a HA-tagged CelTOS were activated ex vivo with or without 1 nM 49c. Tubulin was used as loading control. Arrow, processed PbCelTOS; asterisk, full-length PbCelTOS. (L) Immunoblot showing the cleavage of PbCelTOS-HA immunoprecipitated from 49c-treated ookinetes in the presence of rPfPMX. Processing of PbCelTOS-HA is abrogated in the presence of 10 nM 49c but not in the presence of 1 μM 49b or 10 μM pepstatin. rPfPMX D/A and PbCelTOS ex vivo samples and no enzyme are used as controls. The lower panel shows the presence of rPfPMX and rPfPMX D/A identified by Flag-tag antibodies. Arrow, processed PbCelTOS; asterisk, full-length PbCelTOS. (M) Female An. stephensi mosquitos were fed on mice infected with mCherry-expressing P. berghei either treated or not treated with 49c (ip injection, 100 mg/kg, at 30 hours prior to blood feeding). At day 7 postfeeding, from each mosquito cage (n = 3), 15 mosquito midguts were dissected, fluorescent micrographs of the individual midguts were recorded, and the number of oocysts was determined. Representative pictures are shown. Scale bar, 500 μm. (N) The fluorescent oocysts from the individual microscopic pictures were counted. The graph shows the distribution of oocyst numbers per mosquito that fed on the three control and three drug-treated mice.

Transmission is mediated by an obligatory sexual life-cycle phase (fig. S7D), and drugs blocking transmission to the mosquito vector are potentially valuable for malaria eradication. Although 49c did not affect differentiation from asexual stages into microscopically mature gametocytes (Fig. 3C), it prevented further development into fertile gametes. Upon mosquito ingestion, each male gametocyte differentiates into eight spermlike microgametes that are released in a process termed exflagellation. A 48-hour treatment with 49c led to a 10-fold decrease in exflagellation (Fig. 3D) and prevented the lysis of the RBC membranes surrounding both male and female parasites (Fig. 3E). Both PbPMX and its substrate PbSUB1 are expressed in mature gametocytes, whereas PbPMIX could not be detected in these stages (Fig. 3, F and G). Consistently, a 48-hour treatment with 49c strongly reduced PbSUB1 processing, pointing to a conserved proteolytic cascade required for the egress of both asexual and sexual erythrocytic stages. Treatment at the time of gametogenesis activation had no effect on exflagellation, indicating that PMX activity is required for parasite egress during gametocytogenesis before mosquito ingestion (Fig. 3H).

Egress of gametes from the host erythrocyte is followed by fertilization. Within 24 hours, zygotes transform into ookinetes, which colonize the epithelial monolayer of the mosquito midgut. Treatment with 49c during in vivo gametocytogenesis completely blocked ookinete formation (Fig. 3I). Conversely, treatment at the onset of gametogenesis did not prevent the development of ookinetes (Fig. 3J). We were not able to detect PbSUB1 in ookinetes, but 49c inhibited processing of the micronemal protein PbCelTOS that occurs during the late stages of ookinete development (Fig. 3K). PbCelTOS is crucial for ookinetes to traverse host cells into the site of oocyst development (34). Similar to gametocytes, PbPMX, but not PbPMIX, was detected in ookinetes (Fig. 3F), and in vitro assays revealed that rPfPMX cleaved immunoprecipitated hemagglutinin (HA) epitope–tagged PbCelTOS (PbCelTOS-HA) (Fig. 3L). Conversely, rPfPMIX is inactive on this CelTOS substrate (fig. S6, G and H). Given the potent inhibitory effect of 49c on gametogenesis and ookinete biology, we assessed the transmission-blocking potential of 49c in vivo. A single treatment administered to infected mice 30 hours before mosquito blood meal completely blocked oocyst formation in the midgut of Anopheles mosquitos (Fig. 3, M and N).

Several studies have highlighted the commonalities between the blood and hepatic stages with regard to egress and invasion strategies, as illustrated for SUB1 (13, 35, 36) and AMA1 (37), respectively. The effects of 49c on hepatic stage development with the focus on egress of P. berghei were examined by using HeLa as well as HepG2 cells infected with mCherry-expressing P. berghei sporozoites. 49c added 2 hours postinfection affected neither the number of infected cells (fig. S8, A, E, and F) nor the size of intrahepatic parasites in vitro after 48 hours (fig. S8, B, G, and H). Infected cells detach upon rupture of the PVM, which typically occurs between 55 and 60 hours postinfection (38). 49c led to a dramatic reduction of detached cells at doses as low as 6 nM and no detached cells at doses ≥ 25 nM (Fig. 4A and fig. S8, C and D). Merozoite development was normal, but progression to detached cells was hampered, resulting in accumulation of merozoite stage parasites at the time of cell detachment (Fig. 4B). Staining with MSP1, a marker for successful liver stage development, confirmed that 49c does not affect merozoite development (Fig. 4, B and C). An in vivo time-course experiment was conducted with mice infected with luciferase-expressing P. berghei sporozoites and either treated twice with 100 mg/kg 49c or left untreated (Fig. 4D). The livers of drug-treated and control mice were comparably infected at 44 hours postinfection, as revealed by bioluminescence imaging. Control mice exhibited the typical disappearance of signal from the liver after 55 hours and the concomitant appearance of signal in blood after 65 hours (Fig. 4E). The liver load was prolonged in the presence of 49c, likely as a result of impaired egress, and blood-stage development was strongly delayed, as analyzed by flow cytometry in the blood of infected animals (Fig. 4F). Treatment with 49c had a strong effect on the establishment of blood-stage parasites, although, at the administered doses, a complete block was not achieved.

Fig. 4 Compound 49c prevents hepatic merosomes formation.

(A) The detached cell–formation rate was reduced in 49c-treated P. berghei–infected HeLa cells in a dose-dependent manner. The results were statistically evaluated by a one-way analysis of variance (ANOVA) test with Dunnet’s multiple comparisons (*P ≤ 0.05; ***P ≤ 0.001; n.s., not significant >0.05). (B) Distribution of the different hepatic stages by IFA of mCherry-expressing P. berghei in HeLa cells treated with either DMSO or 49c 54 and 65 hours postinfection (hpi). An anti-Msp1 antiserum was used to detect maturation of merozoites. Classification was done as reported before (41). (C) Positive MSP1 staining of infected HeLa cells 65 hours postinfection indicates normal merozoite development. At the cytomere stage, clear invaginations are visible, and at the merozoite stage, individual merozoites are labeled with an MSP1 antibody. EXP1, a PVM protein, is shown to visualize the PVM. Scale bar, 10 μm; inserts are magnified 3×. (D) Mice were infected by intravenous injection with mCherry-Fluc–expressing P. berghei sporozoites and treated with 100 mg/kg 49c at 20 and 40 hpi (n = 3). At indicated time points RediJect D-Luciferin was administered, and whole-body luminescence was detected by using the in vivo imaging system. In control mice, the luminescence signal from the liver disappears at 55 hours, the time at which hepatic parasites egress. From 65 hpi, blood-stage parasites are detectable. In drug-treated mice, the signal increased at 52 hpi and then remained longer in the liver (still present at 55 hpi), most likely because egress is blocked. (E) The measured total flux (photons/second, p/s) in the head and chest region (blood-stage parasites) of the mice in (D) is shown for the different time points. (F) At each time point, blood samples were analyzed by flow cytometry detecting the mCherry signal of the parasite in 106 whole blood cells and confirming the results obtained by bioluminescence imaging (E).

Curative and preventive strategies for malaria treatment should ideally target three malarial life-cycle stages: exoerythrocytic forms, the asexual blood stages, and the transmission stages. Here we show that the pleiotropic plasmepsin inhibitor 49c inhibits malarial PMIX and PMX, resulting in a block in blood-stage parasite egress and invasion as well as hepatic-stage egress and transmission. Taken together, PMIX and PMX qualify as very promising dual targets toward malaria eradication.

Supplementary Materials

Materials and Methods

Figs. S1 to S8

References (4163)

Movies S1 and S2

References and Notes

  1. Acknowledgments: We are grateful to C. Boss (Actelion Pharmaceuticals Ltd.) and S. Wittlin for providing us with the initial 49c stock and for their help with the chemical synthesis. We thank V. Polonais for her exploratory work on the project and G. Wright, M. Lebrun, D. Gaur, and A. Cowman for the gift of numerous antibodies. We are grateful to O. Billker for the gift of compound 2 and F. Hackett, J. B. Marq, and J. Xu for their technical assistance. We thank the Biocenter Oulu Mass Spectrometry Core Facility for their services. We would like to thank the PlasmoGEM (Plasmodium genetic modification project) team (Wellcome Trust Sanger Institute) for providing the PlasmoGEM vectors. We would like to thank the Netherlands Cancer Institute Protein Facility for provision of the ligation-independent cloning vector, which was acquired by material transfer agreement and the Nederlandse Organisatie voor Wetenschappelijk Onderzoek (NWO) for financial support to the facility (grant 175.010.2007.012). This work was funded by Carigest SA (D.S.-F.), the Swiss National Foundation (grants 156825 to P.P., 310030B_166678 to D.S.-F., 310030_159519 to V.H., and BSSGI0_155852 to M.B.), (grant 51TRPO_151032 to V.H. and D.S.-F.), and the Academy of Finland (grants 257537 and 292718 to I.K.). Bo.M. is funded by the European Research Council under the European Union’s Horizon 2020 Research and Innovation program under grant agreement no. 695596. Funding to M.J.B. was from the Francis Crick Institute ( and Wellcome Institutional Strategic Support Fund (ISSF2) funding to the London School of Hygiene & Tropical Medicine. All the data required to understand and interpret the paper are available in the main text and the supplementary materials.

Stay Connected to Science


Navigate This Article