Novel putative glycosylphosphatidylinositol-anchored micronemal antigen of Plasmodium falciparum that binds to erythrocytes

Abstract

We have identified a new Plasmodium falciparum erythrocyte binding protein that appears to be located in the micronemes of the merozoite stage of the parasite and membrane linked through a glycosylphosphatidylinositol (GPI) anchor. The protein is designated GPI-anchored micronemal antigen (GAMA) and was identified by applying a set of selection criteria to identify previously uncharacterized merozoite proteins that may have a role in cell invasion. The protein is also present in the proteomes of the sporozoite and ookinete micronemes and is conserved throughout the genus. GAMA contains a novel domain that may be constrained by disulfide bonds and a predicted C-terminal hydrophobic sequence that is presumably replaced by the GPI. The protein is synthesized late during schizogony, processed into two fragments that are linked by a disulfide bond, and translocated to an apical location, which is probably the micronemes. In a proportion of free merozoites GAMA can also be detected on the parasite surface. Following erythrocyte invasion the bulk of the protein is shed in a soluble form, although a short C-terminal fragment may be carried into the newly invaded red blood cell. The protein was shown to bind reversibly to erythrocytes and therefore represents a new example of a host cell binding protein.

FIG. 1.

Schematic of the primary structure of GAMA. Indicated are the predicted signal peptide (residues 1 to 21) and C-terminal transmembrane domain (residues 715 to 738), the two asparagine-rich regions (residues 38 to 64 and 356 to 485), and the two regions at the N- and C-terminal ends of the protein (NTD, residues 68 to 166; CTD, residues 578 to 714) which were expressed in recombinant form and used to raise specific antisera. The protein consists of 738 amino acids (aa) with a calculated molecular mass of ∼83 kDa.

FIG. 2.

Transcriptional analysis of PF08_0008 expression by RT-PCR using gene-specific primers. The PF08_0008 transcript was detected in the late schizont stages, between 44 and 47 h postinvasion (hpi). No PCR products were obtained using samples prepared for each time point in the absence of reverse transcriptase (not shown).

FIG. 3.

Membrane anchorage of GAMA. (A) To provide evidence of a GPI anchor, antibodies were used to immunoprecipitate full-length GAMA (∼83 kDa, arrow) from [³H]glucosamine- and mannose-labeled parasites. As a positive control, the MAb 1E1 was used to immunoprecipitate full-length MSP1 (∼200 kDa) from labeled parasites. The proteins immunoprecipitated with these two antibodies were fractionated on different gels, and the film was subjected to different exposure times. No proteins were immunoprecipitated with normal mouse serum (NMS). (B) To examine whether or not GAMA is membrane bound, total schizont material (T) was treated with a high-pH carbonate buffer to yield carbonate-soluble (CS) and -insoluble (CP) fractions, which were then resolved by SDS-PAGE and analyzed by Western blotting. Anti-CTD antiserum detected two species, both of which were found in the carbonate-insoluble fraction, confirming the membrane association. The larger species is presumably full-length GAMA, and the 49-kDa polypeptide (p49) likely represents the product of a proteolytic cleavage event. Blots were also probed with an antibody raised against the type I transmembrane protein PTRAMP and the soluble MSP3 in order to verify the outcome of the extraction procedure.

FIG. 4.

Processing of GAMA. (A) Total schizont material was examined by Western blotting under reducing and nonreducing conditions using the anti-CTD antiserum to provide evidence of a disulfide-linked complex. Under reducing conditions, both full-length GAMA and p49 were evident in the sample, whereas under nonreducing conditions, p49 was absent. An additional band was observed under nonreducing conditions, which effectively comigrated with full-length GAMA and is consistent with p49 remaining associated with the N-terminal product of the initial processing event, by way of a disulfide linkage. (B) To determine the timing of the primary processing event, proteins were immunoprecipitated from BFA-treated (+BFA) or methanol-treated (−BFA) schizont material using anti-CTD antiserum and mouse preimmune serum (NMS). The level of GAMA processing was reduced in BFA-treated parasites, so that full-length GAMA was more abundant in the treated samples than in untreated samples. Although it was difficult to determine any difference in the level of p49 between the samples due to the interference of the comigrating immunoglobulin G heavy chain, the amounts of p42 and p37 fragments were considerably reduced in the BFA-treated samples. (C) A combination of immunoprecipitation and Western blotting was used to identify components of the complex. Proteins immunoprecipitated with anti-CTD antiserum from ³⁵S-labeled NP-40-lysed schizonts were separated by SDS-PAGE, transferred to nitrocellulose, and probed with anti-NTD antiserum on a Western blot (left panel). In the right panel is shown an autoradiograph of ³⁵S-labeled proteins immunoprecipitated from the same preparation of schizonts using either anti-CTD antiserum or mouse preimmune serum. Anti-NTD antiserum reacted in Western blotting with both the full-length GAMA and the 37-kDa polypeptide that were immunoprecipitated by the anti-CTD antiserum (both polypeptides are indicated by arrows and labeled), suggesting that the 37-kDa polypeptide is indeed the N-terminal product of GAMA processing. The labeled secondary antibody also reacted with the heavy and light chains of the immunoglobulin G present in the sample (marked with asterisks). The anti-NTD antiserum did not react with the 42-kDa polypeptide immunoprecipitated by the anti-CTD antiserum, and presumably not with the 49-kDa polypeptide, although the latter was hard to confirm due to reactivity of the secondary antibody with immunoglobulin G heavy chain. (D) The protein is shed from the parasite by proteolysis. Culture supernatant was analyzed by Western blotting under reducing (R) and nonreducing (NR) conditions with the anti-CTD and -NTD antisera. A 42-kDa doublet (p42) was detected by the anti-CTD antiserum, and a 37-kDa polypeptide (p37) was detected by the anti-NTD antiserum. The identical mobilities of the shed fragments recognized by the two different antisera under nonreducing conditions suggest that they are together in a complex.

FIG. 5.

Analyses of the distribution of GAMA in asexual blood stage parasites by immunofluorescence. (A) Formaldehyde-fixed P. falciparum 3D7 schizonts were probed with mouse anti-CTD (top panel) or anti-NTD (bottom panel) antiserum and an anti-mouse-fluorescein isothiocyanate conjugate (green). Parasite nuclei were stained with DAPI (blue). Staining with both antisera resulted in a strong punctuate pattern, with a single point of fluorescence located away from the nucleus and consistent with a location in the apical organelles. Scale bars represent 2 μm. No signal was detected when parasites were probed with normal mouse serum (not shown). (B) Formaldehyde-fixed free P. falciparum merozoites were probed with rabbit anti-CTD antiserum and either mouse polyclonal anti-AMA1 antibodies (panels i, iii, and iv) or mouse MAb 61.3 (panel ii). Mouse antibodies were detected with an anti-mouse Alexa Fluor 594 conjugate (red) and rabbit antibodies with an anti-rabbit Alexa Fluor 488 conjugate (green). Parasite nuclei were stained with DAPI (blue). Scale bars are 5 μm in panel i and 2 μm in the other panels. In panels i and ii there is a clear apical localization of AMA1, RhopH2, and GAMA, with AMA1 and GAMA appearing to be colocalized; panel iii shows a circumferential localization of both GAMA and AMA1 on the surface of free merozoites; and panel iv shows a “cap”-like distribution of GAMA on the surface of free merozoites (arrows), in contrast to the circumferential distribution of AMA1. (C) Newly invaded ring stage parasites were probed with the mouse anti-MSP1₁₉ MAb 1E1 and either rabbit anti-CTD (top panel) or anti-NTD (bottom panel) antiserum. Mouse antibodies were detected with an anti-mouse Alexa Fluor 594 conjugate (red) and rabbit antibodies with an anti-rabbit Alexa Fluor 488 conjugate (green). Parasite nuclei were stained with DAPI (blue). There was no reactivity exhibited by mouse preimmune serum with ring stage parasites (not shown). Scale bars are 2 μm. There is a clear reactivity of anti-CTD antibodies, but not anti-NTD antibodies, with ring stage parasites. Anti-CTD and anti-MSP1₁₉ signals appear to colocalize (top panel), suggesting that the “stub” part of GAMA that is produced upon shedding of the p42-p37 complex is carried into the host cell and remains associated with the ring stage parasite.

FIG. 6.

GAMA binds to erythrocytes. (A) Samples of culture supernatant that were incubated in the absence (−RBC) or presence (+RBC) of erythrocytes were analyzed by Western blotting with antibodies specific to EBA175 (top panel), MSP3 (middle panel), and GAMA (bottom panel). The defined erythrocyte adhesin EBA175 is depleted from culture supernatant upon the addition of erythrocytes, as is GAMA. MSP3 is not depleted. (B) Proteins bound to the erythrocyte surface were eluted with 0.5 M NaCl and analyzed by Western blotting with anti-GAMA antibodies. The binding and elution of GAMA indicate a reversible interaction with erythrocytes.

FIG. 7.

Schematic representation of the processing and shedding of GAMA. Primary processing of full-length protein gives rise to the p49 and p37 fragments. These fragments are held together in a disulfide-linked complex and are released into culture supernatant following alternative cleavage events near the membrane anchor, leaving behind a residual “stub.” The regions of GAMA expressed in a recombinant form in order to raise antisera are indicated in darker shading.