Tracking Glideosome-Associated Protein 50 Reveals the Development and Organization of the Inner Membrane Complex of Plasmodium falciparum ^{▿ †}

Parasite culture, transfection, and PfGAP50-GFP constructs.

P. falciparum parasites (strain 3D7) were maintained in O-type human RBCs from the Melbourne Red Cross Blood Bank in RPMI-HEPES medium supplemented with 5% human serum and 0.5% Albumax (19). Tight synchronization of cultures was achieved using a modification of the method previously described (34). Briefly, schizont stage parasites were purified using 70% (wt/vol) Percoll. Purified parasites were immobilized as a monolayer in a dish coated with concanavalin A. Uninfected RBCs were overlaid for 30 min and then removed to begin a new culture of highly synchronous parasites. The timing of the synchronous culture was measured from the beginning of the 30-min incubation period. Aliquots were taken every hour for confocal microscopic analysis.

PfGAP50 (PlasmoDB gene identifier, PFI0880c) comprises four distinct domains, as shown in Fig. 1A. PfGAP50 is characterized by a hydrophobic N-terminal signal sequence with a predicted cleavage site after amino acid 25, an acid phosphatase homology domain, a transmembrane domain, and a 5-amino-acid predicted cytoplasmic domain (SKNMK). The full-length PfGAP50 gene was PCR amplified using primers 5′-TCAGAACTCGAGATGAATTATTGTAAAACCACGTTCC-3′ and 5′-TATGGTACCTTTCATATTTTTTGATAAAAAAGAGGAAGC-3′. A C-terminal PfGAP50_1-369 was constructed using the above forward primer and the alternative reverse primer 5′-ACCAATGGTACCTACAACTCTAACAAAGGTATC-3′. The DNA fragments were digested with XhoI/KpnI (underlined) and cloned into the P. falciparum transfection vector pGLUX1 (A. G. Maier, M. T. O'Neill, and A. F. Cowman, unpublished data), which incorporates the human dihydrofolate reductase (hDHFR) cassette for positive selection with the prodrug WR99210 (18). This vector drives transgene expression from the PfCRT 5′ promoter region, which has maximal expression in the late ring stages (11). A second mutant, in which the acid phosphatase domain of PfGAP50 is replaced with GFP (PfGAP50ΔAP-GFP), was constructed using overlapping PCR. The PfGAP50 N-terminal and C-terminal domains and the GFP coding sequence were amplified from 3D7 genomic DNA using the following primer pairs. N-terminal, 5′-GATATGTCAATATTTGTATAACACATGAAG-3′ and 5′-TTCTTCTCCTTTACTCCAATCACCCAAAGA-3′; C-terminal, 5′-GATGAACTATACAAATTTGTTAGAGTTGTA-3′ and 5′-TAACATTTATTCATTATAATATTTATCATATTATATATATA-3′; and GFP, 5′-AGTAAAGGAGAAGAACTTTTCACTGGAGTTGTC-3′ and 5′-TTTGTATAGTTCATCCATGCCATGTGTAATCCC-3′. The final construct was generated by PCR using the overlapping product as a template and 5′-TCAGAACTCGAGATGAATTATTGTAAAACCACGTTCC-3′ and 5′-GCGCTTAATTAATTATTTCATATTTTTTGATAAAA-3′. The PCR product was digested with XhoI/PacI (underlined), purified, and cloned into the pGLUX1 vector. 3D7 parasites were transfected with 100 μg of purified plasmid DNA as previously described (13). The episomal constructs were maintained by culturing them in the presence of 10 nM WR99210.

Western analysis of PfGAP50-GFP transgenic parasites.

Highly synchronized late-stage PfGAP50-GFP parasites were purified using 70% (wt/vol) Percoll and treated with streptolysin O (SLO) (1.5 U/μl final concentration) at 37°C for 15 min to remove hemoglobin (40). The cells were lysed with 3 rapid freeze-thaw cycles in phosphate-buffered saline (PBS), and the soluble and insoluble fractions were separated by centrifugation at 12,000 × g for 10 min at 4°C. Proteins were separated on 4 to 12% SDS NuPage gels (Invitrogen) under reducing conditions and transferred onto nitrocellulose membranes. The membranes were probed with either mouse anti-GFP (Roche; 1:500), anti-PfGAP45 (1/500 dilution), anti-PfGAP50 (1:500) (9) or anti-MTIP (22), or rabbit anti-PfGAPDH (36), followed by horseradish peroxidase-conjugated sheep anti-mouse or anti-rabbit secondary antibodies (Promega; 1:5,000). Western blots were developed using enhanced chemiluminescence.

Immunoprecipitation of IMC components.

Highly synchronized late-stage parasites were purified over 70% (wt/vol) Percoll, and proteins were extracted by incubating them on ice for 15 min in RIPA buffer (1% octylphenoxypolyethoxyethanol [IGEPAL CA-630], 150 mM NaCl, 0.5% sodium deoxycholate [DOC], 0.1% sodium dodecyl sulfate [SDS], 50 mM Tris, pH 7.4). Soluble and insoluble fractions were separated by centrifugation at 12,000 × g for 10 min at 4°C. Soluble lysates were incubated for 1 h at 4°C using the rabbit antibodies anti-PfGAP45, anti-GFP, and anti-MTIP and preimmune serum. A 1:1 ratio of protein-A/G-Sepharose was used for immunoprecipitation (IP). IP lysates were separated by SDS-PAGE and analyzed by Western blotting using mouse antibodies against GFP, PfGAP45, or PfMTIP. Western blots were developed as described above.

Light microscopy and immunofluorescence.

For immunofluorescence, air-dried thin smears were fixed for 10 min at room temperature in 100% acetone, and antibody labeling was performed as explained below. Alternatively, aliquots of culture were washed in PBS and fixed for 20 min at room temperature in 2% paraformaldehyde, 0.008% glutaraldehyde. The cells were permeabilized with 0.1% Triton X-100 for 5 min at room temperature prior to antibody incubation. The prepared cells were incubated for 1 h with the following antibodies diluted in 3% bovine serum albumin (BSA) in PBS: mouse anti-GFP (1:500; Roche), rabbit anti-PfGAP45 (1:500), rabbit anti-PfGAP50 (1:500), rabbit anti-RhopH3 (41) (1:500), and rabbit anti-Bet3 (2) (1:500), followed by Alexa Fluor 488 or 568 anti-IgG (Molecular Probes). Cells were washed in PBS, resuspended in Dako mounting medium, and viewed using a Leica TCS-SP2 confocal fluorescence microscope. For labeling of the endoplasmic reticulum (ER) cultures, ER-Tracker Blue-White DPX (0.25 μM; Invitrogen Corporation) was added to an aliquot of a culture for 30 min at 37°C. The cells were resuspended in fresh complete medium, and the ER fluorescence pattern was examined using an Olympus IX81 fluorescence microscope with a DAPI (4′,6-diamidino-2-phenylindole) filter set.

For live-cell confocal imaging of transfectants, 3 μl of culture (3% hematocrit; 5 to 10% parasitemia) was sealed under a coverslip and viewed on a Leica TCS-SP2 or a Zeiss LSM 510/FCS confocal microscope within 20 min of mounting. Fluorescence recovery after photobleaching (FRAP) was performed on a Zeiss LSM 510/FCS, as previously described (28). Selected regions were photobleached using a 488-nm laser at 80% intensity for 0.1 s. Fluorescence images were acquired before, immediately after, and 3 min after bleaching.

3D-SIM (37) was implemented on a DeltaVision OMX 3D-SIM (optical microscope, experimental) at the Microbial Imaging Facility of the University of Technology Sydney. The samples were excited using 405-nm and 488-nm lasers and imaged using emission filters of 419 to 465 nm and 500 to 550 nm, with a 100× 1.4-numerical aperture (NA) oil lens and multichannel simultaneous-imaging capability (electron-multiplying charge-coupled device cameras; >90% quantum efficiency; 512 by 512 pixels). Aliquots of culture were adjusted to a final concentration of 2% paraformaldehyde, 0.008% glutaraldehyde and resuspended in 50% Vectashield antifade mounting medium (Vectorlabs), containing DAPI, in PBS. All images were processed using NIH Image J version 1.42 (http://rsbweb.nih.gov/ij/).

Electron microscopy.

Transmission electron microscopy (EM) and tomography were performed as previously described (14, 23). Briefly, tightly synchronized early- or late-schizont-stage 3D7 parasites were fixed, embedded, and stained. Seventy-nanometer-thick sections were observed at 120 keV using a JEOL-2010HC (Japan) transmission electron microscope (La Trobe University, EM Unit). For electron tomography, 200- or 300-nm sections were cut, stained, and observed at 200 keV on a Tecnai G² F30 (FEI Company) transmission electron microscope (Bio21 Institute, Melbourne, Australia). Tilt series were collected and reconstructed to generate cell models as described previously (1).

Full-length PfGAP50-GFP is expressed in transfected P. falciparum.

To examine the genesis and origin of the IMC, P. falciparum transfectants expressing full-length PfGAP50 tagged at the C terminus with GFP were generated (Fig. 1A). Mature-stage parasite cultures of the parent (3D7) and transfected (PfGAP50-GFP) lines were harvested and selectively lysed with streptolysin O to remove the host cell hemoglobin. The isolated parasites were disrupted by freeze/thawing, and the supernatant and pellet fractions were subjected to Western blotting. Anti-PfGAP50 recognizes a band of 42 kDa in the pellet fraction of 3D7-infected RBCs (Fig. 1B, left), as reported previously (9), and an additional band of 64 kDa in the PfGAP50-GFP transfectant (Fig. 1B, middle). The relative intensity of the upper band indicates that PfGAP50-GFP is expressed at a lower level than endogenous PfGAP50. When the blotted lysates of the PfGAP50-GFP transfectants were probed with anti-GFP, only the 64-kDa protein species was detected (Fig. 1B, right), with no reactivity in 3D7 parasites or uninfected RBCs (data not shown). The apparent molecular masses of PfGAP50 and PfGAP50-GFP are in rough agreement with the predicted masses after cleavage of the signal sequence (41.5 kDa and 68.3 kDa, respectively). The presence of PfGAP50 and PfGAP50-GFP in the pellet fraction indicates that they are membrane associated. The blots were stripped and reprobed with anti-PfGAPDH to compare loading levels and to validate the lysis procedure (Fig. 1B, bottom).

Full-length PfGAP50-GFP is incorporated into the glideosome complex.

Endogenous PfGAP50 and PfGAP45 form part of the glideosome complex that is associated with the IMC (9). To determine whether full-length PfGAP50-GFP participates in this complex, we immunoprecipitated IMC components from the transfectants. Mature-stage parasite-infected RBCs were enriched and solubilized using a detergent-containing buffer. Cleared soluble lysates were immunoprecipitated with a range of antibodies (Fig. 2). When a Western blot of the precipitated samples was probed with mouse anti-GFP antibodies, a band corresponding to full-length PfGAP50-GFP (64 kDa) was detected in samples precipitated with anti-PfGAP45, anti-MTIP, and anti-GFP sera, but not with the preimmune serum (Fig. 2). The lysate sample precipitated using anti-GFP contained both PfGAP45 and PfMTIP (Fig. 2, right lanes). Mouse anti-PfGAP45 antiserum detected a doublet at 37 or 39 kDa, while anti-PfMTIP recognized a 27-kDa band. This demonstrates that full-length PfGAP50-GFP is in complex with endogenous PfGAP45 and PfMTIP.

Live-cell imaging reveals marked rearrangements of the IMC during parasite development.

To follow the movement and redistribution of PfGAP50 during P. falciparum development, we synchronized a parasite culture to within a 2-h window (34). Confocal microscopy analysis of cells reveals a marked reorganization of the PfGAP50-GFP as the parasite matures (Fig. 3; see Fig. S1 in the supplemental material). In ring and early- to mid-trophozoite stage parasites, PfGAP50-GFP is present within the parasite in a pattern that is reminiscent of that in the ER (Fig. 3; see Fig. S1, top, in the supplemental material). In the late-trophozoite/early-schizont stage, the fluorescence pattern becomes more looped (Fig. 3, 38 to 40 h). At 40 to 42 h postinfection, the PfGAP50-GFP starts to accumulate into foci, which by 44 h have developed into distinct puncta at the periphery of the parasite (Fig. 3). The punctate fluorescent structures are close to bright dots in the differential interference contrast (DIC) image that probably represent the rhoptry organelles (see Fig. S1, schizonts, in the supplemental material). In the final stage of schizogony (∼46 h postinvasion), the fluorescence appears to redistribute around the periphery of the daughter merozoites (Fig. 3; see Fig. S1 in the supplemental material).

Full-length PfGAP50-GFP is trafficked via the ER to apical regions of developing merozoites.

In an effort to determine the nature of the subcellular compartments with which PfGAP50-GFP is associated, we undertook colabeling with different markers. In the early stages of development, ER-Tracker labels a reticular structure that largely overlaps with the PfGAP50-GFP location (Fig. 4 A, top). The profile is dispersed and coincident with the pattern for the ER-resident protein PfERC (ER calcium binding protein) in trophozoite stage parasites (Fig. 4B, top).

This is perhaps surprising, given that PfGAP50 has previously been reported to be a component of the IMC, which lies under the merozoite plasma membrane. To ensure that this was not an artifact due to expression of PfGAP50-GFP under a heterologous promoter, we examined samples of the 3D7 parental strain. An antiserum raised against endogenous PfGAP50 (9) gives a labeling pattern that shows almost complete overlap with that observed for PfGAP50-GFP in both trophozoite and schizont stage parasites (see Fig. S2A in the supplemental material). As described above, in schizont stage parasites, the GFP chimera redistributes to punctate structures located around the periphery of the dividing parasite (Fig. 4A, middle). At this stage, ER-Tracker labels reticular structures, presumably the ER, within the parasite with a diffuse staining pattern quite unlike that of PfGAP50-GFP. Similarly, in immunofluorescence images of schizont stage 3D7 parasites, PfGAP50 and PfERC are not colocated (Fig. 4B, bottom). The PfGAP50-containing structures are not Golgi apparatus, as the labeling patterns for PfGAP50-GFP and the Golgi marker, PfBet3 (2), are quite distinct (see Fig. S2B in the supplemental material).

In rupturing schizonts, PfGAP50-GFP is located at the periphery of individual daughter cells and appears to extend from one pole, while ER-Tracker labels the inside of individual merozoites (Fig. 4A, bottom). These data indicate that PfGAP50-GFP is redistributed from the ER to separate structures within the dividing cell. Colabeling with the rhoptry marker RhopH3 (41) confirms the location of PfGAP50-GFP, extending from the apical region of the daughter merozoites (see Fig. S2C in the supplemental material).

We compared the location of PfGAP50-GFP in the late schizont stage with that of the known IMC-resident protein PfGAP45 (Fig. 4C, top). In schizont stage parasites, these proteins appear to be largely colocated in a region of each daughter merozoite away from the nucleus (i.e., at the apical end). This is consistent with coparticipation in the IMC. Interestingly, in released merozoites, the locations of PfGAP45 and PfGAP50-GFP are not completely overlapping (Fig. 4C, bottom), suggesting that PfGAP50-GFP is reorganized after merozoite release.

In an effort to determine the physical organization of PfGAP50 within the IMC, we performed FRAP studies of PfGAP50-GFP-labeled structures at different stages of parasite development. In the trophozoite stage, application of an intense laser pulse caused selective bleaching of PfGAP50-GFP in a region of the ER (Fig. 5 A). There was partial loss of fluorescence from adjacent regions during the bleach pulse and no recovery of the signal from other regions of the ER over a period of 3 min. This indicates that at least part of the PfGAP50-GFP is present in a state that shows restricted diffusion. When the PfGAP50-GFP had redistributed to the apical caps, there was no recovery of the signal after photobleaching, indicating that PfGAP50-GFP from other caps cannot diffuse into this compartment (Fig. 5B). When the PfGAP50-GFP has redistributed around the periphery of the daughter merozoites (Fig. 5C), loss of signal is restricted to the bleach point and does not recover. This indicates that PfGAP50-GFP is present in a protein complex that restricts lateral diffusion within the membrane.

PfGAP50 is located in an apical IMC cap in developing merozoites.

The punctate PfGAP50-GFP-labeled structures observed in developing merozoites are intriguing; however, their small size limits the information that can be obtained using conventional optical microscopy. 3D-SIM has recently been developed as a method for imaging fluorescently labeled samples with improved resolution (37). We used 3D-SIM to image PfGAP50-GFP in parasites that are colabeled with DAPI (Fig. 6).

The characteristic PfGAP50-GFP-labeled punctate structures are apparent in a parasite with 8 nuclei (Fig. 6A and C). At higher resolution, it is evident that the PfGAP50-GFP-containing caps are located at the apical ends of nascent merozoites and form flat, ellipsoid structures, each with two unlabeled “pores.” The average dimensions of these pores were estimated to be 110 ± 40 nm by 140 ± 40 nm (n = 18). The nature of these structures and their organization relative to the parasite nuclei can be best appreciated in a rotation of the 3D-SIM reconstruction presented in movie S1 in the supplemental material.

As the eight nuclei start to divide, the PfGAP50-GFP-labeled structures appear to be pulled apart, with the pores opening to give a claw-like appearance (Fig. 6B and D). These structures are organized in pairs, with the cavities facing each other (Fig. 6D, arrow; for a rotation of the 3D-SIM reconstruction, see movie S2 in the supplemental material). The average length of the cavity is 230 ± 40 nm, while the width in its central region is 130 ± 30 nm (n = 18). These may represent the apical caps of separating daughter cells in which the nascent IMC structures are pulled apart.

We performed EM on schizont stage parasites in an effort to find an ultrastructural correlate of these doubly punctured ellipsoids. In early-schizont stage parasites, apical prominences are observed on the developing merozoites at either side of a single nucleus (Fig. 6G and H). The apical prominence is formed close to a mitotic spindle (Fig. 6H, m). The multiple membranes and cytoskeletal components (microtubules) of the nascent pellicle give this region an electron-dense appearance (Fig. 6G and H, arrowheads). Later in development, separate structures can be observed at the apical ends of the dividing merozoites (Fig. 6I and J, arrowheads), which likely correspond to the separating claw-like structures seen by 3D-SIM. The “pores” in the PfGAP50-GFP-labeled cap presumably correspond to the central region of the cytoskeletal annulus at the apex of the cell. An electron tomogram was generated from a section through a mature merozoite, and the tubulin rings are observed in cross-section in a virtual section from the tomogram (Fig. 6K, arrows). Four serial electron tomograms collected through mature merozoites were rendered to generate a model of a mature merozoite (Fig. 6L), with the apical rings depicted in blue. These polar rings provide a structure through which the rhoptry necks (indicated in burnt orange) are connected to the parasite surface. The inner, middle, and outer polar rings have diameters of approximately 120 nm, 180 nm, and 240 nm. We interpret the 3D-SIM data to suggest that PfGAP50 is recruited to the developing IMC and that the IMC forms a membrane collar surrounding the apical pore.

As the schizont develops further, the PfGAP50-GFP appears to extend around the surface of the parasite (Fig. 6E and F). The apical pores remain visible, but due to the extension of the IMC around the periphery of the merozoite, which dilutes the PfGAP50-GFP, they are less prominent. The change in IMC organization is best appreciated by comparing the dividing schizonts in movies S1 and S2 in the supplemental material with the more mature schizonts in movies S3 and S4.

The C-terminal transmembrane domain of PfGAP50 is needed for correct trafficking to the IMC.

In T. gondii, the transmembrane domain and the short C-terminal segment of PfGAP50 are needed for correct assembly into the glideosome complex (20). In an effort to determine whether the same region of PfGAP50 is needed for correct trafficking and assembly in P. falciparum, we generated transfectants expressing GFP chimeras of two different PfGAP50 mutants (Fig. 7 A) and examined their intracellular locations.

Live-cell fluorescence microscopy of the PfGAP50_1-369-GFP transfectants (lacking the C-terminal hydrophobic and cytoplasmic domains) (Fig. 7B) indicates that the chimera is in the PV, with some protein being transferred to the digestive vacuole (Fig. 7B, middle). Colabeling of these cells with ER-Tracker (Fig. 7B, bottom) confirms that the PfGAP50_1-369-GFP fluorescence is outside the region of the ER. The data suggest that the N-terminal signal peptide directs entry into the secretory pathway and that, in the absence of amino acids 370 to 396, it follows the default trafficking pathway to the PV (3). This confirms that the C-terminal transmembrane and cytoplasmic domains are required for correct trafficking of PfGAP50 to the IMC in P. falciparum.

Mature stage PfGAP50_1-369-GFP transfectants were released with streptolysin O and then ruptured, and the pellet and supernatant fractions were subjected to Western analysis. The transfectants express a protein of 61 kDa (Fig. 7C), which is similar to the anticipated size of PfGAP50_1-369-GFP after cleavage of the signal sequence, as well as a protein of 26 kDa, which likely represents the protease-resistant core of GFP. The full-length chimera is largely released into the supernatant fraction, suggesting that it is a soluble protein. The cytoplasmic protein PfGAPDH shows a similar solubility profile (Fig. 7C, bottom).

The acid phosphatase homology domain of PfGAP50 is not needed for correct trafficking to the IMC.

PfGAP50 has a central domain that exhibits ∼25% identity to different purple acid phosphatases of plants and animals. To determine whether this domain is needed for trafficking to the IMC, we generated P. falciparum transfectants episomally expressing a construct in which the acid phosphatase homology domain is replaced by GFP (Fig. 7A). Live-cell fluorescence microscopy reveals that the PfGAP50ΔAP-GFP chimera has a fluorescence pattern similar to that of full-length PfGAP50-GFP (Fig. 7D). This suggests that the phosphatase domain is not required for correct trafficking of PfGAP50 to the IMC. Western analysis reveals a protein with an apparent molecular mass of 34 kDa (as expected) that is present in the pellet fraction (Fig. 7E), as for full-length PfGAP50-GFP. Unfortunately, these domain-swapped transfectants grew poorly and were lost within 2 to 3 weeks of the first appearance of fluorescent parasites, and we were unable to obtain sufficient material to confirm an interaction with other IMC proteins by immunoprecipitation analyses. The data suggest that, while not essential for trafficking to the IMC, the acid phosphatase domain may play an important role in PfGAP50 function.