The Plasmodium life cycle is a sequence of alternating invasive and replicative stages within the vertebrate and invertebrate hosts. How malarial parasites exit their host cells after completion of reproduction remains largely unsolved. Inhibitor studies indicated a role of Plasmodium cysteine proteases in merozoite release from host erythrocytes. To validate a vital function of malarial cysteine proteases in active parasite egress, we searched for target genes that can be analyzed functionally by reverse genetics. Herein, we describe a complete arrest of Plasmodium sporozoite egress from Anopheles midgut oocysts by targeted disruption of a stage-specific cysteine protease. Our findings show that sporozoites exit oocysts by parasite-dependent proteolysis rather than by passive oocyst rupture resulting from parasite growth. We provide genetic proof that malarial cysteine proteases are necessary for egress of invasive stages from their intracellular compartment and propose that similar cysteine protease–dependent mechanisms occur during egress from liver-stage and blood-stage schizonts.
Malaria is caused by intracellular parasites of the phylum Apicomplexa that can enter and exit host cells. The characterization of parasite and host cell proteins involved in Plasmodium cell entry has provided a detailed understanding of the underlying mechanisms (1) and led to new intervention strategies (2). In contrast, the equally important process of Plasmodium release is less well understood. With the exception of ookinetes, invasive stages (i.e., sporozoites, liver-stage merozoites, and blood-stage merozoites) are formed by multiple fission in processes called sporogony and merogony, respectively. These stages then need to egress from their intracellular compartment and, shortly thereafter, from their host cell. Inhibitor studies suggested that multiple proteolytic events occur during rupture of schizont-infected erythrocytes and subsequent reinvasion of erythrocytes (3, 4). Treatment of intracellular schizonts with the cysteine protease inhibitor E64 resulted in accumulation of membrane-enclosed viable merozoites (5, 6). In support of active proteolytic events during parasite egress, stage-specific expression of cysteine and serine protease activities has been detected (7). In addition, several genes that encode potential cysteine proteases have been identified and characterized in Plasmodium (8). They include falcipain 1, a nonessential cathepsin L–like cysteine protease with yet undefined functions in oocyst development (9, 10), the food vacuole–resident hemoglobinases falcipain 2/2' and 3 (11–13), and a family of proteases that were termed serine repeat antigens (SERAs) (14–16). Members of this distinct Plasmodium protease family are clustered on chromosome II (17) and belong to papain-like cysteine proteases based on a central ∼30-kD protease domain. Reverse genetics showed that some members are vital for erythrocytic schizogony, whereas others are dispensable for asexual growth of Plasmodium (16). However, so far no function in parasite egress has been assigned to any of these proteins. We reasoned that inactivation of a member of the Plasmodium papain-like cysteine protease family for which expression is restricted to sporogenic stages might lead to an essential function that can be analyzed on the cellular level. Here, we show targeted disruption of an oocyst-specific papain-like cysteine protease in P. berghei. Mutant sporozoites fail to egress from midgut oocysts. Therefore, we termed the corresponding protein egress cysteine protease 1 (ECP1).
Results And Discussion
Identification of a stage-specific Plasmodium cysteine protease
Several members of papain-like cysteine proteases, also termed SERAs, were previously reported to be nonessential during asexual blood-stage development (16). We tested expression of the five cysteine proteases of the P. berghei SERA locus by RT-PCR (Fig. 1 A). Our analysis revealed that one member (ECP1) displayed an interesting restriction of gene transcription to sporozoite stages. Notably, ECP1 transcription is specific for mature oocysts, the stage that marks the final step of sporozoite generation, and is subsequently down-regulated in mature salivary gland sporozoites that are transmitted to the mammalian host (Fig. 1 B). The orthologous genes in P. falciparum (SERA8; PFB0325c) (17) and P. yoelii (PY02063) (18) show 54 and 81% overall amino acid sequence identity with P. berghei ECP1 (PbECP1; DQ000976), respectively (Fig. 1 C). In good agreement with our findings, the P. falciparum orthologue was reported recently to be expressed specifically in sporozoites (19) and absent from erythrocytic stages (20). All Plasmodium ECP1 proteins contain a central ∼250–amino acid papain-family cysteine protease domain (Fig. 1 C). Within the domain, conservation to PbECP1 is 70% and 93% for the P. falciparum and P. yoelii orthologues, respectively. A hallmark of papain-family cysteine proteases is the presence of the catalytic triad with invariant cysteine, histidine, and asparagine residues and the oxyanion-hole glutamine residue (8). Presence of these residues in the ECP1 proteins indicates that they might function as proteases (Fig. 1 D).
PbECP1 gene disruption
To study the role of PbECP1, we generated a loss-of-function parasite line. The endogenous ECP1 copy was targeted with an insertion plasmid (21). Homologous recombination was expected to lead to gene disruption by generation of two truncated nontranscribed ecp1 copies (Fig. 2 A). This strategy allows gene disruption without loss of genetic information and is likely to minimize cis effects on neighboring genes. The parental blood-stage population from the successful transfection was used for cloning three independent disruption parasite lines, termed ecp1(-). Insertion-specific PCR analysis confirmed the correct insertion at the predicted locus (Fig. 2 B). To verify PbECP1 deficiency of the mutant parasites, we performed RT-PCR and cDNA amplification using polyA+ RNA from oocyst sporozoites as templates (Fig. 2 C). We also confirmed that expression of the neighboring genes, SERA2 and ORF2, is not affected in the ecp1(-) disruptants. We next examined the phenotype of ecp1(-) parasites during the Plasmodium life cycle. As expected, ecp1(-) clones were indistinguishable from WT parasites in development and growth of asexual and sexual Plasmodium stages (unpublished data). Transmission to Anopheles mosquitoes and oocyst development were normal when compared with WT parasites (Table S1).
We next analyzed sporozoite development by examining oocyst morphology and comparing oocyst sporozoite numbers in WT and ecp1(-) parasites. No differences in generation of viable sporozoites were observed. Importantly, when the ecp1(-) sporozoites were liberated from dissected midgut oocysts, they showed the typical short residual gliding motility of WT oocyst sporozoites in vitro (Fig. 3 A). Together, our findings show that ECP1 is dispensable for Plasmodium cellular functions before sporozoite release. We conclude that ecp1(-) parasites form viable sporozoites in numbers comparable with WT parasites, in good agreement with our observation that ECP1 is developmentally up-regulated in mature oocysts (Fig. 1 A).
ecp1(-) sporozoites fail to egress from midgut oocysts
Upon closer examination by phase-contrast microscopy, we observed a peculiar arrangement of sporozoites within the oocysts (Fig. 3 B). Although WT sporozoites are arranged in a radial fashion, ecp1(-) sporozoites seemed to be organized in circles. Intriguingly, ecp1(-) sporozoites displayed a continuous circular movement around a central axis, in both clockwise and anticlockwise directions (Video 1). In WT oocysts, sporozoite bending and flexing is seen on rare occasions, presumably in preparation for egress from the oocyst (unpublished data). In general, no motility can be observed in WT oocysts (Video 2). In marked contrast, continuous circular motility was observed in all ecp1(-) oocysts examined. This previously unrecognized motility within midgut oocysts is likely a consequence of a defect after completion of sporogony. This observation prompted us to perform a detailed spatial and temporal analysis of sporozoite distribution within the Anopheles mosquito (Table I). Intriguingly, no sporozoites were detected in the hemocoel or in the salivary glands of infected mosquitoes despite efficient infection rates and high numbers of oocyst sporozoites. Although we continued to look for salivary gland sporozoites throughout the life span of the mosquitoes (∼55 d after feeding) we failed to detect ecp1(-) salivary gland sporozoites. In WT parasites, oocyst sporozoite numbers peak at ∼day 14 after infection. Thereafter, sporozoites are released continuously into the hemocoel, where they can be detected transiently (Table I). Sporozoites enter salivary glands rapidly and actively; their final destination in the invertebrate host (22). Accordingly, numbers of oocyst sporozoites decline over time, whereas salivary glands remain filled with sporozoites, rendering infected mosquitoes infectious for life. In striking contrast, ecp1(-) oocysts do not rupture, resulting in a remarkable accumulation of viable sporozoites (Table I). Hence, the observed intraoocyst motility is likely a consequence of the failure to egress the oocysts. We also noticed that none of the persisting ecp1(-) oocysts was melanized throughout the mosquito life span, nor were the survival rates of the infected mosquitoes affected (unpublished data). Together, our data indicate that oocysts are not breached passively by parasite growth. Instead, we propose that sporozoite egress is an active process that requires ECP1 functions.
Next, we tested whether viable motile ecp1(-) oocyst sporozoites are infectious to the mammalian host. We injected highly susceptible Sprague/Dawley rats with either WT or ecp1(-) oocyst sporozoites (Table S2). As expected, we achieved consistent blood-stage infections with 100,000 WT oocyst sporozoites. In striking contrast, animals injected with even 10-fold higher doses of ecp1(-) sporozoites remained malaria free, suggesting additional functions of ECP1 after oocyst rupture.
Lack of ECP1 results in protected oocysts
Upon midgut dissection we also observed that ecp1(-) oocysts were resistant to light mechanical stress such as gentle grinding to free sporozoites. Occasionally we detected free-floating oocysts that were detached from the midgut (Fig. 3 C). The oocyst capsule has a bipartite structure with the inner layer being of parasite origin and the outer thick layer deriving from the basal lamina of the mosquito midgut (23). The inner oocyst membrane is covered by the circumsporozoite protein (CSP) (24) and, hence, can serve as a marker for oocyst permeabilization. To test if ecp1(-) oocysts are more rigid than WT oocysts, we dissected midguts and permeabilized oocysts with the detergent saponin (Fig. 4 A). Although all oocysts can be permeabilized with methanol and displayed strong circumferential CSP staining, only WT oocysts could be permeabilized by the natural surfactant saponin (Fig. 4 A). This finding suggests that developing Plasmodium oocysts are protected by an impermeable oocyst wall that is processed actively after sporozoite maturation. To control for CSP levels in ecp1(-) oocyst sporozoites, we performed a Western blot analysis of sporozoites dissected in the absence or presence of the cysteine protease inhibitor E64 (Fig. 4 B; reference 25). We detected a previously unrecognized additional CSP band that was specific for oocyst sporozoites. Notably, this signal was abundant only in the ecp1(-) mutant or when WT midguts were dissected in the presence of E64. These findings may indicate a role of ECP1 in CSP processing. Although the substrate of ECP1 is not known yet, CSP is a likely candidate for a direct or indirect downstream proteolytic processing event, particularly because it seems to be one of the dominant parasite-derived components lining the inner side of the oocysts (24).
Collectively, our data suggest that ECP1 plays a central role in the egress of the sporozoites from midgut oocysts. Lack of ECP1 proteolytic activity blocked the life cycle of malaria parasites inside the mosquito vector at the oocyst stage. Therefore, inhibition of oocyst rupture provides an additional target for transmission-blocking strategies. Oocysts stand out in the Plasmodium life cycle, because they represent the longest developmental phase and the only replicative phase of the malaria parasites that does not need host cells for its expansion. Despite their importance, oocysts remain the least-characterized mosquito stage of Plasmodium. Purification of protected ecp1(-) oocysts may provide a rare resource for a detailed analysis of the molecular repertoire of mature oocysts. Our study may pave the way for the identification of similar egress cysteine proteases that drive merozoite release from liver-stage and blood-stage schizonts by targeted gene disruption. This possibility is supported by the inhibitory effect of cysteine protease inhibitors on merozoite egress from host erythrocytes (5, 6). Identifying an essential function of cysteine proteases, such as the one of ECP1 for sporozoite egress, is fundamental for drug-target validation and rational design of inhibitors.
Materials And Methods
Animals were from Charles River Laboratories. All animal work was conducted in accordance with European regulations and approved by the state authorities (Regierungspräsidium Karlsruhe).
Generation of the ecp1(-) parasite line
For targeted disruption of ECP1, an integration vector was generated by amplification of a PCR fragment using P. berghei genomic DNA as template and primers ECP1for (5′-GGACTAGTGAGCATATAGAAAGCCATATTCAAC-3′; SpeI site is underlined) and ECP1rev (5′-TCCCCGCGGGCACCTTGCTCAATTATGTAATCTTTTAAG-3′; SacII site is underlined). Cloning into the P. berghei transfection vector (21) resulted in plasmid pAA05. The targeting plasmid was linearized with EcoRV, and parasite transfection, positive selection, and parasite cloning were performed as described previously (21). Integration-specific PCR amplification of the ecp1(-) locus was generated using specific primer combinations. We obtained three independent ecp1(-) clonal parasite populations that were phenotypically identical. Detailed analysis was performed with one representative clone.
For RT-PCR analysis, we isolated poly (A+) RNA from 5 × 105 WT salivary gland sporozoites and 106 WT and ecp1(-) oocyst sporozoites, respectively, using oligo dT-columns (Invitrogen). For cDNA synthesis and amplification, we performed a two-step PCR using random decamer primers (Ambion) and subsequent standard PCR reactions, using gene-specific primers.
Phenotypical analysis during the Plasmodium life cycle
Blood-stage development was analyzed in vivo in asynchronous infections using Naval Medical Research Institute mice. Gametocyte differentiation and exflagellation of microgametes were detected in mice before mosquito feedings. Sporozoite populations were separated and analyzed as described previously (26, 27). Adherent sporozoites were incubated with a mAb against P. berghei circumsporozoite protein (PbCSP) (28) and a polyclonal anti–P. berghei thrombospondin-related anonymous protein (TRAP) antiserum (29). Bound antibodies were detected using Alexa Fluor 546–conjugated anti–mouse antibodies and Alexa Fluor 488–conjugated anti–rabbit antibodies, respectively (Molecular Probes).
In vivo infectivity of sporozoites
For determination of the infectivity of oocyst sporozoites, infected midguts were dissected at days 15–17 after feeding. Sporozoites were liberated and injected i.v. at the numbers indicated into young Sprague/Dawley rats. Patency was checked daily by Giemsa-stained blood smears.
For the analysis of CSP localization in the oocysts, infected midguts were fixed in 2% formaldehyde/0.2% glutaraldehyde, permeabilized with 1% saponin in PBS/1% FCS or with ice-cold methanol and incubated with primary anti-PbCSP (1:1,000; reference 28). At the high-antibody dilution, internal sporozoites are not visualized. Bound antibodies were detected using Alexa Fluor 488–conjugated anti–mouse antibodies.
Western blot analysis
For detection of CSP levels in WT and ecp1(-) oocysts, we dissected midguts of infected mosquitoes at day 15 after infection. Infected midguts were isolated, ground, and pelleted in the presence or absence of freshly prepared 100-μM E64 (Sigma-Aldrich; reference 5). Total oocyst lysates equivalent to 100,000 oocyst sporozoites and, as a control, 100,000 WT salivary gland sporozoites were separated on a 10% SDS PAGE and transferred to a nitrocellulose membrane. CSP was detected with primary anti-PbCSP (1:8,000; reference 28). Bound antibodies were detected using horseradish peroxidase–conjugated anti-mouse antibodies (Sigma-Aldrich).
Online supplemental material
Table S1 shows that oocyst development of ecp1(-) parasites is not affected compared with WT parasites. Table S2 shows that ecp1(-) oocyst sporozoites are noninfectious to the mammalian host. Video 1 shows real-time live-imaging of ecp1(-) oocysts. ecp1(-) sporozoites lack the capacity to egress oocysts and instead display continuous circular motility. Video 2 shows the corresponding WT oocysts with no detectable internal motility.
We gratefully acknowledge A. Kunze for expert technical assistance and A. Baumm and R. Mosbach for professional help with the video documentation. We are grateful to Dr. V. Nussenzweig for sharing unpublished results and critically reviewing the text.
Our work was supported by grants from the research focus “Tropical Medicine Heidelberg” of the Medical Faculty of Heidelberg University and the European Commission (BioMalPar 23).
The authors have no conflicting financial interests.