Calcium-Dependent Protein Kinase 5 Is Required for Release of Egress-Specific Organelles in Plasmodium falciparum

ABSTRACT The human malaria parasite Plasmodium falciparum requires efficient egress out of an infected red blood cell for pathogenesis. This egress event is highly coordinated and is mediated by several signaling proteins, including the plant-like P. falciparum calcium-dependent protein kinase 5 (PfCDPK5). Knockdown of PfCDPK5 results in an egress block where parasites are trapped inside their host cells. The mechanism of this PfCDPK5-dependent block, however, remains unknown. Here, we show that PfCDPK5 colocalizes with a specialized set of parasite organelles known as micronemes and is required for their discharge, implicating failure of this step as the cause of the egress defect in PfCDPK5-deficient parasites. Furthermore, we show that PfCDPK5 cooperates with the P. falciparum cGMP-dependent kinase (PfPKG) to fully activate the protease cascade critical for parasite egress. The PfCDPK5-dependent arrest can be overcome by hyperactivation of PfPKG or by physical disruption of the arrested parasite, and we show that both treatments facilitate the release of the micronemes required for egress. Our results define the molecular mechanism of PfCDPK5 function and elucidate the complex signaling pathway of parasite egress.

need to identify additional critical biological pathways that might be targeted by a new generation of antimalarial medications.
The major clinical manifestations of malaria result from exponential expansion of parasites during the blood stage of the life cycle (3)(4)(5). Asexual blood-stage development of Plasmodium parasites occurs via schizogony. In this process, multiple asynchronous nuclear divisions occur within a common cytoplasm. Mature and invasive merozoites are not formed until a final segmentation (or budding) wherein the newly formed daughter parasites are all formed simultaneously. Thus, the daughter merozoites are not fully mature until all daughter merozoites are mature (6). This requirement highlights the importance of a coordinated egress signaling pathway to prevent release of immature parasites.
In apicomplexan organisms, including both Plasmodium spp. and Toxoplasma gondii, release of specialized apical organelles known as micronemes is a critical step in parasite egress. The triggered discharge of these organelles depends on the activation of calcium signaling, cGMP signaling, and phosphatidic acid signaling (7)(8)(9)(10). Some of the protein mediators of microneme discharge have been identified, including calciumdependent protein kinases (7,11), C2 domain-containing proteins (12), and acylated pleckstrin homology domain-containing proteins (10). However, the specific roles played by these proteins in the activation of microneme discharge remain incompletely understood.
Calcium-dependent protein kinases (CDPKs) (13)(14)(15) and the cyclic GMP-dependent protein kinase (PKG) (8,16,17) are important for P. falciparum egress. Additionally, a protease cascade mainly driven by the serine protease P. falciparum SUB1 (PfSUB1) is essential for egress (8,18,19). P. falciparum PKG (PfPKG) activity is required for PfSUB1 release from specialized exoneme organelles (8), and once discharged into the parasitophorous vacuole, PfSUB1 proteolytically processes multiple parasite proteins, including the P. falciparum merozoite surface protein-1 (PfMSP1), which itself is essential for egress and allows PfMSP1 to interact with and destabilize the host erythrocyte cytoskeleton (20). An important role for the perforin-like protein PfPLP1 in blood-stage parasite egress has been hypothesized (21). However, PfPLP1 can be knocked out in the blood stage without a detectable replication defect, suggesting that its role is nonessential and/or redundant for asexual parasite egress (22). PfCDPK5 (PF3D7_1337800) is essential for P. falciparum egress (15). However, the mechanism behind the egress block in PfCDPK5-deficient parasites has remained unknown.
In the current study, we perform a functional evaluation of the egress defect in PfCDPK5-deficient parasites. With superresolution microscopy, we demonstrate the dynamic localization of PfCDPK5 throughout schizogony and its colocalization with a subset of microneme organelles. We show that PfCDPK5-deficient parasites do not efficiently release a subset of micronemes and that this failure of microneme release is the cause of the egress block. In response to baseline activation of PfPKG, PfCDPK5deficient parasites initially show activation of the protease cascade that is similar to parasites without PfCDPK5 deficiency. However, this level of PfPKG activation is insufficient to induce the release of the required subset of micronemes for egress, and protease processing does not efficiently proceed to completion in PfCDPK5-deficient schizonts. Pharmacological enhancement of PfPKG activation facilitates microneme release and overcomes the PfCDPK5-mediated deficiency. Interestingly, mechanical release of PfCDPK5-deficient merozoites also overcomes this block. We demonstrate that a major function of PfCDPK5 is to act cooperatively with the PfPKG signaling pathway. Our study provides a new layer in our understanding of the P. falciparum egress pathway.

RESULTS
PfCDPK5 has dynamic localization during schizont development. We generated a new transgenic P. falciparum strain, 3D7-PfCDPK5 DD KnL, with three copies of the hemagglutinin (HA) epitope followed by the destabilizing domain (DD) at the endogenous PF3D7_1337800 locus ( Fig. 1A; also see Fig. S1 in the supplemental material). By placing the DD at the carboxy terminus of P. falciparum calcium-dependent protein kinase 5 (PfCDPK5), the resulting fusion protein is rapidly degraded when the stabilizing agent Shield-1 (Shld-1) is removed. As expected, when grown in the absence of Shld-1, these parasites have a Ͼ90% schizont-stage replication defect in the second cycle (Fig. 1A). These parasites also have a single copy of the mycobacteriophage Bxb1 attB sequence (23). To allow quantitative measurement of egress following parasitophorous vacuolar membrane (PVM) and red blood cell (RBC) plasma membrane rupture, we integrated an attP-containing nano-luciferase reporter cassette into this attB site (24,25). By fusing the first 60 amino acids from the knob-associated histidine-rich protein, the nano-luciferase reporter protein (referred to as KnL) is targeted via the secretory system to the parasitophorous vacuole (26). In parasite cultures grown with Shld-1 ([ϩ] Shld-1), bioluminescence accumulates in the cell pellet during asexual development and is released into the supernatant following egress (Fig. 1B). Thus, parasite biomass can be quantified by bioluminescence in cultured cell pellet, and egress is monitored by release of the KnL into the cell-free supernatant.
To evaluate the localization of PfCDPK5 during the asexual life cycle of P. falciparum, we utilized superresolution structured illumination microscopy (SR-SIM). Parasites were tightly synchronized, and at 44 h post-invasion (h.p.i.), schizonts were treated with E64, a cysteine protease inhibitor that prevents egress but still allows maturation of viable merozoites (26)(27)(28). Samples were collected for immunofluorescence analysis (IFA) by SR-SIM at~48 Ϯ 2 h.p.i. For an internal "timer" for parasite maturation, we monitored the expression and localization of the P. falciparum apical membrane antigen-1 (PfAMA1) (29)(30)(31). Consistent with previous studies of PfAMA1, we initially detected PfAMA1 expression in the micronemes and after a poorly characterized "egress trigger," observed PfAMA1 translocation to the plasma membrane of the fully mature daughter merozoites (8,32,33). In 3D7-PfCDPK5 DD KnL parasites with undetected PfAMA1 expression, we observed punctate cytoplasmic localization of PfCDPK5 (Fig. 1C, top row). In parasites with detected micronemal PfAMA1 localization, we observed two different staining patterns for PfCDPK5. The first was observed in parasites with weaker PfCDPK5 staining, where there was strong colocalization with micronemal PfAMA1 (Fig. 1C, second row). Given that PfCDPK5 has no known signal sequence, we hypothesize that it is located on the cytoplasmic side of these micronemes. In the second pattern, observed at higher levels of PfCDPK5 expression, PfCDPK5 retained colocalization with PfAMA1 but also exhibited additional diffuse apical staining (Fig. 1C, third row). In parasites that were beyond the "egress trigger," PfAMA1 was translocated to the merozoite surface, and PfCDPK5 relocalized to a peripheral cytoplasmic distribution with increased signal near the plasma membrane (Fig. 1C, E64-treated row). Localization was similar in free merozoites isolated from parallel cultures grown without E64 (Fig. 1C, Free merozoite row). Complete z-stacks for a representative parasite from each localization type are provided (see Movies S1 to S5 in the supplemental material). To verify that PfCDPK5 was localized inside the cytoplasm of "post-egress trigger" merozoites, we costained the parasites with antibodies directed against the PfGAP45 protein associated with the inner membrane complex (34) and confirmed that PfCDPK5 staining is located within PfGAP45 staining (Fig. S2A). This result demonstrates that PfCDPK5 is located within the parasite cytoplasm.
Previous reports have identified specialized apical organelles, including the PfSUB1containing exoneme (18) and the PfROM1-containing mononeme (35), that do not overlap with classic micronemes, rhoptries, and dense granules. In order to further identify the localization of PfCDPK5, we costained parasites with markers directed against proteins in the rhoptries and micronemes. Results from SR-SIM showed that PfCDPK5 did not consistently colocalize with the rhoptry neck marker PfRON4 (36) or the exoneme marker PfSUB1 (Fig. S2B). However, we note that there are limited areas of overlap between PfCDPK5 and all of the tested apical organelle markers, mostly observed when PfCDPK5 had strong expression levels. Given the strong colocalization with PfAMA1-containing micronemes, we expected to see strong colocalization with PfEBA175, an invasion ligand residing in the micronemes (37)(38)(39). Interestingly, we found that PfCDPK5 did not colocalize as strongly with PfEBA175 as it did with PfAMA1 ( Fig. 2A). The limited regions of colocalization seen with high levels of PfCDPK5 expression largely correspond to the diffuse apical staining of PfCDPK5 and not the more intense punctate regions of PfCDPK5 staining. To confirm this result, we costained parasites with antibodies against PfAMA1 and PfEBA175. As has been hypothesized before by epifluorescence and immunoelectron microscopy (40), we found that at superresolution, these two proteins show limited regions of colocalization and largely define different subsets of micronemes (Fig. 2B). Thus, when visualized at superresolution, we found that PfCDPK5 and PfAMA1 colocalize but that PfAMA1 and PfEBA175 appeared in different subsets of micronemes. These results provide further evidence for the existence of subsets of micronemes in P. falciparum, potentially with different functions.
PfCDPK5-deficient parasites do not release micronemes. Studies from both Plasmodium spp. and T. gondii have demonstrated a role for calcium-dependent protein kinases in apical organelle discharge. PfCDPK1 has been hypothesized to be important for microneme discharge (7). However, these experiments relied on chemical and peptide inhibition of the kinase in extracellular merozoites. In addition, inducible expression of the PfCDPK1 autoinhibitory junction domain caused a replication defect at the early schizont stage of P. falciparum (13). A recent characterization of a P. falciparum strain with an inducible knockdown of PfCDPK1 demonstrated no detected defects in parasite egress or PfAMA1 release but did demonstrate decreased invasion and release of PfEBA175 (41). The Plasmodium berghei ortholog of PfCDPK1 is not essential for blood-stage replication and therefore unlikely to be essential for microneme discharge (42). Recently, PfCDPK1 was successfully knocked out in P. falciparum, demonstrating that the protein is important but not essential for asexual growth (43). In T. gondii, CDPK1 (TgCDPK1) (11) and TgCDPK3 (44)(45)(46) are essential and important, respectively, for microneme secretion. Given these precedents, we evaluated the requirement of PfCDPK5 for microneme secretion.
To simultaneously evaluate the status of the PfAMA1-and PfEBA175-containing subsets of micronemes, we costained [ϩ] and [Ϫ] Shld-1 schizonts for the two markers (Fig. 3C). In both [ϩ] and [Ϫ] Shld-1 schizonts, when PfAMA1 was apical, we found that PfEBA175 was also apical. Interestingly, in the "post-egress trigger" [ϩ] Shld-1 schizonts, we found that while PfAMA1 had been translocated, on the merozoite surface, PfEBA175 apparently remained apical. When released, PfEBA175-containing micronemes expose the ectodomain of PfEBA175 onto the surface of the apical end of the merozoite. The exposed PfEBA175 is processed by the rhomboid-like protease PfROM4 on the surface of the merozoite, shedding the ectodomain into the culture supernatant (48,49). Shedding of processed forms of invasion ligands is detected in the supernatant from E64-treated schizonts (8). Therefore, to directly evaluate whether PfEBA175containing micronemes had been released, we measured the discharge of the processed protein into the supernatant. With E64-treated schizonts, we found that processed PfEBA175 was readily detected in the supernatant of [ϩ] Shld-1 schizonts but was largely absent from the supernatant of [Ϫ] Shld-1 schizonts (Fig. 3D). Thus, although the sensitivity of superresolution microscopy was insufficient to detect discharge of PfEBA175-containing micronemes, their discharge was detected by measurement of the release of the processed form of PfEBA175. Together, these results demonstrate that the signaling pathways controlling the release of PfAMA1-containing and PfEBA175-containing micronemes are regulated by PfCDPK5.
Utilizing our bioluminescent transgenic reporter parasites (Fig. 1B), we found that [Ϫ] Shld-1 parasites, even with a Ͼ90% block as schizonts, still release some KnL into the supernatant. We note that the amount of bioluminescence in the supernatant of [Ϫ] Shld-1 parasites was not further decreased when these parasites were treated with a mixture of protease inhibitors (including serine, cysteine, and metalloprotease inhibitors) at 44 h.p.i. (Fig. S3A). Thus, it is likely that the PfCDPK5-deficient arrested schizonts have partial permeabilization of their surrounding PVM and RBC plasma membrane that is not dependent on ongoing protease activity. This finding is consistent with previous reports that the PVM and RBC plasma membrane of segmented schizonts are partially permeable to small proteins even in the presence of E64 (8,28). Nonetheless, the significant difference in the quantity of bioluminescence in the supernatant between [ϩ] and [Ϫ] Shld-1 schizonts allowed a quantitative measure of membrane permeability as a surrogate for relative egress efficiency. To test whether BIPPO activation of PKG induced PfCDPK5-deficient parasites to egress, we measured the bioluminescence in the supernatant of treated parasites. As expected, the amount of bioluminescent reporter released into the supernatant from [ϩ] and [Ϫ] Shld-1 parasites was equivalent following BIPPO treatment (Fig. 3F). Pretreatment of [Ϫ] Shld-1 parasites with C1 prior to BIPPO prevented this release (Fig. S3A). For additional confirmation of this result, we treated PfCDPK5-deficient parasites with zaprinast, a structurally distinct phosphodiesterase inhibitor that increases cGMP levels and activates PfPKG, and obtained similar release of bioluminescent activity into the supernatant (Fig. S3B). This result demonstrates that enhanced PKG activity induces microneme secretion and secondarily mediates full permeabilization of the PVM and RBC plasma membrane.
PKG activity has been linked to downstream calcium release in both Plasmodium species and T. gondii (52)(53)(54). It is formally possible that PKG-induced increases in cytoplasmic calcium may be sufficient to activate a residual amount of PfCDPK5 in [Ϫ] Shld-1 parasites. To test whether the BIPPO-induced microneme discharge in PfCDPK5deficient parasites was entirely dependent on calcium release, we evaluated PfAMA1 localization (Fig. 4A) and release of the KnL reporter in the supernatant (Fig. 4B) in three

biological replicates of [ϩ] and [Ϫ]
Shld-1 3D7-PfCDPK5 DD KnL schizonts treated with either BIPPO, a calcium ionophore (A23187), or both. In PfCDPK5-deficient schizonts, A23187 treatment did not induce PfAMA1 translocation, suggesting that calcium release by itself was not sufficient for microneme discharge. In [ϩ] Shld-1 control parasites, A23187 treatment did not inhibit PfAMA1 translocation, showing that the elevated level of calcium was not toxic or inhibitory for microneme discharge. This conclusion is further supported by the observation that the combination of A23187 and BIPPO in the parasites had the same result as that for BIPPO alone in both [ϩ] and [Ϫ] Shld-1 schizonts. Therefore, the BIPPO-induced microneme release and resulting egress in PfCDPK5-deficient schizonts are not solely due to a secondary PfPKG-induced calcium release.
Egress protease cascade is slowed in PfCDPK5-deficient parasites. Previously, we observed that the egress protease cascade was similarly initiated in both [ϩ] and [Ϫ] Shld-1 schizonts (15). However, that observation was done using a single time point. Thus, we evaluated the extent to which the protease cascade in PfCDPK5deficient parasites progressed with time-as has been demonstrated for wild-type parasites (8). For this set of experiments, we chose to use the previously published 3D7-PfCDPK5-DD parasites without an epitope tag (15). This strain was selected, because it has the strongest replication block of our existing strains. To evaluate the protease cascade, we focused on PfSUB1 processing of the P. falciparum merozoite surface protein-1 (PfMSP1), as this protein is known to be important for parasite egress (20). Following exoneme release, PfSUB1 proteolytically processes the~200-kDa PfMSP1 precursor and releases a 42-kDa product, PfMSP1 42 . We isolated 44 h.p.i. schizonts from parallel cultures that had been maintained with and without Shld-1. These isolated schizonts were sampled immediately and then allowed to incubate for an additional 6 h in the presence of E64. As seen previously, PfMSP1 42 was already detected at the starting time point in both [ϩ] and [Ϫ] Shld-1 samples (Fig. 4C). In the [ϩ] Shld-1 schizonts, the amount of fully processed PfMSP1 42 relative to the unprocessed precursor increased from Ͻ20% at 44 h.p.i. to Ͼ80% 6 h later. In contrast, in the [Ϫ] Shld-1 parasites, the relative amount of fully processed PfMSP1 42 remained Ͻ20% at the later time point.
To evaluate whether exoneme release was completed at the initial time point, [ϩ] Shld-1 parasites were treated with C1 to prevent any further PKG activity, and samples were collected at the later time point. Compared to the untreated [ϩ] Shld-1 schizonts and the starting samples, the relative amount of fully processed PfMSP1 42 was lower, indicating that protease processing was incomplete at the starting time point and that processing required PfPKG activity. As expected, treatment of [Ϫ] Shld-1 schizonts with C1 did not change the relative amount of the fully processed PfMSP1 42 . However, when the [Ϫ] Shld-1 schizonts were treated with BIPPO, the relative amount of fully processed PfMSP1 42 increased to Ͼ40% (Fig. 4C). These results demonstrate a previously unrecognized finding-that protease processing is incomplete in [Ϫ] Shld-1 schizonts and that full processing is increased when PfPKG activity is chemically enhanced by BIPPO. We evaluated PfSERA5, another substrate of PfSUB1 processing (18), in parallel (Fig. S3C). While the processing differences are smaller, we observed an increase in the relative amount of the mature (p50/p56) forms of PfSERA5 in the [ϩ] Shld-1 schizonts after 6 h and did not observe the same increase in the [Ϫ] Shld-1 schizonts (Fig. S3D). As was seen with PfMSP1, the amount of mature PfSERA5 increased in both [ϩ] and [Ϫ] Shld-1 schizonts following BIPPO treatment.
Physical release of merozoites overcomes microneme release defect. A recently developed procedure to obtain significant quantities of viable merozoites has allowed for direct studies of mechanically released parasites (27,55,56). We have previously shown that PfCDPK5-deficient merozoites were capable of reinvasion following physical disruption of the blocked schizonts (15). However, the invasion obtained by this method was inefficient compared to [ϩ] Shld-1 schizonts. We therefore modified the published "viable merozoite isolation" (27) procedure to directly evaluate the status of PfCDPK5-deficient merozoites (Fig. 5A). Applying this technique, we found that invasion efficiency of physically released merozoites from [ϩ] and [Ϫ] Shld-1 schizonts was similar and more robust than previously reported (Fig. 5B). We evaluated PfAMA1 localization in the schizont prior to shearing and in the merozoites following physical release. Once again, the [ϩ] Shld-1 schizonts had populations of both micronemal and translocated PfAMA1, while [Ϫ] Shld-1 schizonts had almost exclusively micronemal localization. Following shearing, translocated PfAMA1 was readily detected on physically released merozoites from both [ϩ] and [Ϫ] Shld-1 schizonts (Fig. 5C). We repeated the shearing in the presence of C1 to evaluate whether PfAMA1 translocation induced by physical release required PfPKG activity. By IFA, PfAMA1 remained micronemal in released merozoites from C1-treated [ϩ] or [Ϫ] Shld-1 schizonts (Fig. 5C; shown with nuclear counterstain in Fig. S4). Thus, the process of physical disruption and/or exposure of PfCDPK5-deficient merozoites to the extracellular milieu was sufficient to facilitate microneme release, as measured by PfAMA1 translocation, and this release required PfPKG activity. By preventing PfAMA1 translocation, inhibition of PfPKG prevented reinvasion and new ring formation. The mechanism responsible for the enhanced PfPKG activity in the sheared merozoites remains unknown. However, these findings convincingly demonstrate that a primary role of PfCDPK5 is to trigger microneme release, allowing parasite egress.

DISCUSSION
PfCDPK5 is essential for P. falciparum egress out of an infected RBC. It localizes to the apical ends of newly forming daughter merozoites, and in mature, segmented schizonts, PfCDPK5 colocalizes with PfAMA1; later, PfCDPK5 has a diffuse apical staining  that overlaps with several apical markers. PfCDPK5 is present in the membraneassociated fraction following sodium carbonate extraction, and this localization may be due to potential palmitoylation sites within the protein (15). However, the trafficking determinants for PfCDPK5 remain unknown. By superresolution microscopy, we demonstrate multiple nonoverlapping subsets of micronemes. This result has been previously noted for PfSUB1 (the "exoneme") and PfROM4 (the "mononeme") (18,35). However, we suggest that the current list of apical organelles should include additional subsets of egress-specific micronemes. We hypothesize that a subset of micronemes is triggered for release by cooperative activation of PfPKG by PfCDPK5. This defect is most convincingly demonstrated for the PfAMA1-containing micronemes. The complete identities of which parasite proteins reside in these micronemes remain unknown but likely include proteases, perforin-like proteins, and potentially other unknown proteins. We hypothesize that PfCDPK5 interacts with and likely phosphorylates proteins that are associated with apical organelles to transmit a "release" signal.
We note that parasites physically released following PfPKG-blockade by compound 1 (C1) do not translocate PfAMA1 to the merozoite surface ( Fig. 5C; also see Fig. S4 in the supplemental material) and are, therefore, not invasive. In contrast, PfCDPK5deficient parasites can reinvade new RBCs following physical release. Thus, blocking parasites prior to exoneme release prevents both egress and invasion. Our current findings demonstrate an essential role of PfCDPK5 for a downstream release of egressmediating micronemes, and we show that chemically enhanced PfPKG activation can overcome the egress block in PfCDPK5-deficient parasites.
We, therefore, propose a model of egress wherein PfCDPK5 cooperates with PfPKG activation (Fig. 6). In this model, a basal level of PfPKG activation allows some release of exonemes and partial activation of the protease cascade. This initial step is independent of PfCDPK5. In a later step, PfCDPK5 cooperates with PfPKG to facilitate PfAMA1 translocation and discharge of micronemes required for egress. This second step, requiring PfPKG and PfCDPK5, induces full activation of the protease cascade and allows parasite egress. Finally, supraphysiological activation of PfPKG by BIPPO treatment can bypass the requirement for PfCDPK5. One alternative to the model is that PfCDPK5 activity does not affect exoneme secretion directly but rather functions to increase the efficiency and/or completeness of protease processing after exoneme release.
Our methods to obtain free PfCDPK5-deficient merozoites require either BIPPO (or zaprinast) treatment or physical disruption. These methods facilitate PfAMA1 translocation and likely the release of other egress-specific factors. We cannot formally rule out a role for PfCDPK5 in the invasion process that was not also "overcome" by these treatments. Nonetheless, PfCDPK5 is the only calcium-dependent protein kinase in Plasmodium spp. that has a direct role in the release of micronemes required for parasite egress.
Synergy between calcium-dependent kinases and PKG has been suggested by previous studies with both T. gondii and Plasmodium parasites (57)(58)(59). Our current results provide strong genetic evidence for this hypothesis. Brochet and colleagues evaluated the relationship between P. berghei PKG (PbPKG) and calcium signaling, providing a molecular pathway for connection of the two pathways (52). Alam and colleagues demonstrated that PfCDPK1 was a likely substrate of PfPKG and that phosphorylated PfCDPK1 localized to the apical area of daughter merozoites (60). Bansal and colleagues demonstrated that increased PfPKG activity was able to compensate for a less active PfCDPK1 (57). Similarly, Moon and colleagues demonstrated that increased PbPKG signaling was able to overcome a defect in gliding motility observed in PbCDPK3-deficient ookinetes (59). Our current study adds to the complexity of the egress signaling pathway by demonstrating a cooperative relationship between PfCDPK5 and PfPKG activity.
In summary, we have demonstrated a functional explanation for the egress block observed in PfCDPK5-deficient parasites. In the absence of normal PfCDPK5 activity, physiologic activation of the PfPKG pathway is insufficient to release a subset of micronemes that are essential for parasite egress. Thus, the molecular mechanism for the egress block observed in PfCDPK5-deficient schizonts is secondary to the failure to discharge micronemes that are required for egress. Parasite culture and transfection. The 3D7 strain (Walter & Eliza Hall Institute) and transgenic derivatives were cultured in human red blood cells (RBCs) in RPMI 1640 supplemented with 0.5% Albumax II, 50 mg/liter hypoxanthine, 0.21% sodium bicarbonate, and 25 mM HEPES as previously described (62). To generate the 3D7-PfCDPK5 DD strain, sorbitol-synchronized ring-stage parasites were electroporated with 100 g of plasmid DNA of the single-crossover plasmid (see Text S1 for details). Following transfection, parasites were maintained with 250 nM Shield-1 (Shld-1), and stable singlecrossover parasites were selected by cycling on and off WR99210 (Jacobus Pharmaceutical Company). Individual transgenic clones were obtained by limiting dilutions. To generate the 3D7-PfCDPK5 DD KnL parasites with the KnL reporter, a clone was transfected with pJDD250 that expresses the Bxb1 integrase (23) and the KnL reporter and selected on blasticidin (while maintaining WR99210 and Shld-1).

MATERIALS AND METHODS
Parasite phenotypic assays. For replication curves, ring-stage parasites were washed to remove Shld-1, replated in the presence or absence of 0.25 mM Shld-1 at 0.4% parasitemia and 1% hematocrit, and parasitemia was monitored by flow cytometry using SYBR green I staining. For measurements of bioluminescence throughout the asexual development cycle, 100 l of resuspended culture was removed from ring, trophozoite, schizont, and reinvaded rings and centrifuged at 14,000 rpm in a microcentrifuge. The supernatant was transferred to a fresh tube, and 20 l was mixed 1:1 with nano-Glo buffer/substrate (Promega). The pelleted cells were resuspended in 100 l of phosphate-buffered saline Absalon et al. For all indirect immunofluorescence assays (IFAs), parasites were synchronized by Percoll purification of late-stage schizonts, followed by sorbitol synchronization of newly invaded rings 2 h later. In the following cycle, sorbitol synchronization was repeated, and the cultures were washed and replated with or without Shld-1. IFAs were performed as previously described (47) with minor modifications. Thin smears were made on glass slides, air dried, and fixed with 1% paraformaldehyde, permeabilized with 0.1% Triton 100 for 10 min, and blocked with 3% bovine serum albumin (BSA) overnight at 4°C. Primary antibodies were incubated overnight in a cold room at the dilutions indicated: anti-PfAMA1 (1:200), anti-PfEBA175 (1:500), anti-PfGAP45 (1:5000), anti-HA (1:50), anti-PfSUB1 (1:500), and anti- PfRON4 (1:200). Subsequently, cells were washed three times with PBS and incubated for 45 min with the Alexa Fluor 488 or 555 secondary antibodies (1:1,000) (Molecular Probes). After removal of unbound antibodies with three PBS washes, slides were mounted with Vectashield containing 4=,6=-diamidino-2-phenylindole (DAPI) (Vector Laboratories Inc.) with coverslips and kept at 4°C until evaluation. Wide-field images were obtained with a Nikon E800 epifluorescence microscope using a 100ϫ (oil) objective, and images were captured using SPOT Imaging software and then processed using Adobe Photoshop. Superresolution structured illumination microscopy (SR-SIM) Z-stacks were captured using an ELYRA PS.1 microscope (Carl Zeiss Microscopy). The ELYRA microscope was used with a 100ϫ (oil) objective and excitation wavelengths of 405, 488, and 561 nm. SIM images were collected at 100-to 200-nm z-axis steps, with five rotations of the structured illumination grid per channel. The resulting stacks were processed using default reconstruction parameters in ZEN 2012 Black software.
For the P. falciparum AMA1 (PfAMA1) translocation IFAs shown in Fig. 3, 10 M E64 was added to the schizonts 46 h postinvasion (h.p.i.) and incubated for 4 h, then 2 M BIPPO was added for an additional 2 h, and schizonts were harvested for analysis. For the bioluminescence and PfAMA1 translocation IFAs shown in Fig. 4, 10 M E64 was added to 46 h.p.i. schizonts and incubated for 150 min, 1 M A23187 (or nothing) was added for 90 min, then 2 M BIPPO was added for 2 h, and schizonts/supernatants were harvested for analysis. A total of 100 schizonts were scored for micronemal or plasma membrane translocation, and bioluminescence was measured as described above. For PfEBA175 processing experiments, synchronized ring-stage parasites were washed three times to remove Shld-1, and grown at 4% hematocrit in the presence and absence of 250 nM Shld-1. At 44 h.p.i., schizonts were purified by passage through a magnetically activated cell sorting (MACS) column, spun down to a pellet, and resuspended in 100 l of medium containing 10 M E64 in the presence or absence of Shld-1, and plated in a 96-well plate. Heparin (50 g/ml) was added to all wells to prevent reinvasion. BIPPO (2 M) was added 4 h after plating, and parasites were collected 2 h after the addition of BIPPO. Samples were spun down to separate supernatant and pellet. Parasite pellets were washed once with RMPI 1640, and supernatants were spun again to clear any remaining parasites. Washed pellets and cleared supernatants were boiled in sample buffer, analyzed by immunoblotting with anti-PfEBA175 antibodies or anti-PfLDH and anti-H3 antibodies to control for loading, probed with Li-Cor secondary antibodies, and visualized on a Li-Cor Odyssey CLx system. Concurrently, samples of parasites were prepared for immunofluorescence assays by preserving 300 l each of Shld-1 cultures grown with Shld-1 ([ϩ] Shld-1 cultures) and Shld-1 cultures grown without Shld-1 ([Ϫ] Shld-1 cultures) prior to MACS purification. Cultures were plated along with MACS-purified schizonts, and samples were treated as described above. At the time of sample collection, cultures were spun down, and smears were made from the RBC pellet. IFA preparation was performed as described above, and PfAMA1 translocation was determined.
For PfMSP1 and PfSERA5 processing assays, 44 h.p.i. schizonts from [ϩ] and [Ϫ] Shld-1 cultures were purified by magnetic separation, placed in medium without Albumax II with or without Shld-1. A sample of the purified schizonts were lysed immediately in sample buffer, and the remaining schizonts were incubated for an additional 6 h before harvesting. E64 (10 M) was added to all samples at 44 h.p.i. (for C1-treated samples, 2.5 M C1 was also added). At 48 h.p.i., 2 M BIPPO was added to indicated samples. All treated samples were lysed at 50 h.p.i. Lysates were separated on a TGX 4 to 20% gradient cell (Bio-Rad), transferred to polyvinylidene difluoride (PVDF) by wet transfer, immunoblotted with rabbit anti-PfMSP1 42 or mouse anti-PfSERA5 (and anti-PfLDH and anti-H3 as loading controls), probed with Li-Cor secondary antibodies, and visualized on a Li-Cor Odyssey CLx system.
For physical disruption assays, viable merozoites were performed as previously described (27,55) with minor modifications. Late-stage schizonts were magnet purified, incubated in medium with 10 M E64 and without Albumax II for 6 h, and sheared with 20 strokes through a 28-gauge needle. The sheared merozoites were incubated with or without fresh RBCs at 37C for 15 min. Reinvasion was determined by flow cytometry, and PfAMA1 translocation was determined by IFA.