Plasmodium Ookinete-secreted Proteins Secreted through a Common Micronemal Pathway Are Targets of Blocking Malaria Transmission*

The mosquito midgut ookinete stage of the malaria parasite, Plasmodium, possesses microneme secretory organelles that mediate locomotion and midgut wall egress to establish sporogonic stages and subsequent transmission. The purpose of this study was 2-fold: 1) to determine whether there exists a single micronemal population with respect to soluble and membrane-associated secreted proteins; and 2) to evaluate the ookinete micronemal proteins chitinase (PgCHT1), circumsporozoite and TRAP-related protein (CTRP), and von Willebrand factor A domain-related protein (WARP) as immunological targets eliciting sera-blocking malaria parasite infectivity to mosquitoes. Indirect immunofluorescence localization studies in Plasmodium gallinaceum using specific antisera showed that all three proteins are distributed intracellularly with a similar granular cytoplasmic appearance and with focal concentration of PgCHT1 and PgCTRP, but not PgWARP, at the ookinete apical end. Immunogold double-labeling electron microscopy, using antisera against the membrane-associated protein CTRP and the soluble WARP, showed that these two proteins co-localized to the same micronemal population. Within the microneme CTRP was associated peripherally at the microneme membrane, whereas PgCHT1 and WARP were diffuse within the micronemal lumen. Sera produced against Plasmodium falciparum WARP significantly reduced the infectivity of P. gallinaceum to Aedes aegypti and P. falciparum to Anopheles mosquitoes. Antisera against PgCTRP and PgCHT1 also significantly reduced the infectivity of P. gallinaceum for A. aegypti. These results support the concept that ookinete micronemal proteins may constitute a general class of malaria transmission-blocking vaccine candidates.

Transmission of malaria occurs after a female mosquito ingests infected blood, thereby initiating subsequent parasite sexual and sporogonic development. Within minutes in response to midgut environmental cues, gametocytes fertilize to form zygotes that transform into motile ookinetes over the following 15-25 h. Mature midgut lumen ookinetes penetrate and traverse the peritrophic matrix and midgut epithelium and then develop into oocysts on the luminal side of the epithelial basement membrane. Strategies to block the transmission of parasites from vertebrate host to mosquitoes seek to interrupt parasite development at some point in the continuum from gametocytes to ookinete penetration of the mosquito midgut epithelium.
The Plasmodium ookinete contains a single type of specialized secretory organelle, the microneme, which is thought to be involved in host-cell recognition, binding, and motility via secretion of soluble and cell surface molecules involved in interaction with different compartments within the mosquito midgut. Ookinetes differ from other Plasmodium invasive stages, such as sporozoites and merozoites, in that they lack rhoptry and dense granule organelles. Plasmodium ookinetes appear to secrete micronemal contents constitutively, and it has not been determined whether ookinete apical secretion is regulated. The question remains, however, whether the numerous micronemes visible within the apical end of Plasmodium ookinetes are functionally equivalent or whether classes of proteins with different destinations are associated with distinct micronemal populations.
Previous studies suggest that soluble ookinete-secreted/ released proteins (1) as well as cell surface molecules (2,3) are transmission-blocking targets. This study focused on the following three micronemal proteins of Plasmodium gallinaceum ookinetes: the soluble, secreted chitinase PgCHT1 (4), the membrane-associated protein CTRP 1 (5)(6)(7)(8)(9), and the recently described von Willebrand adhesive (vWA) domainrelated protein (WARP) (10). In this study we test the hypothesis that micronemal secreted proteins are candidate targets of blocking parasite infectivity for mosquitoes. After verifying that the Escherichia coli-produced recombinant proteins had appropriate biological activity and elicited specific antisera, the antibodies were used to determine subcellular localization and transmission blocking activity in mosquito membrane feeding assays. Transmission blocking assays determined the ability of both single sera and combinations to inhibit P. gallinaceum ookinete infectivity for the Aedes aegypti mosquito midgut. The results have implications for furthering a mechanistic understanding of Plasmodium cell biology and parasite-mosquito interactions and suggest that ookinete micronemal proteins may be a general class of malaria transmission-blocking targets.

EXPERIMENTAL PROCEDURES
Parasites-The P. gallinaceum 7A strain was maintained by mechanical blood passage and transmission through mosquitoes. Ookinetes were cultured in serum-free, chemically defined M199 medium and were obtained as described (11).
Membrane Feeding Assays-Female A. aegypti mosquitoes between 5 and 7 days post-emergence from pupae were starved for 24 h and then fed on membrane feeders containing 200 l of P. gallinaceum-infected blood (ϳ10% parasitemia) plus 70 l of antisera for 20 min. After feeding, nonengorged mosquitoes were removed, and engorged mosquitoes were maintained with 10% glucose at 26°C and 80% relative humidity. Mosquito midguts were dissected and oocysts enumerated 7 days after feeding. Statistical analysis was done using the Mann Whitney U test to compare differences in geometric mean oocyst density and the proportion of mosquitoes infected between groups, as described previously (1,(12)(13)(14)(15).
Construct Preparation, Production of Recombinant Protein, and Antibody Preparation-Enzymatically active recombinant PgCHT1 was produced in E. coli as described previously (4).
The full-length PgCTRP gene was isolated, and the DNA sequence was determined following amplification from P. gallinaceum genomic DNA using degenerate PCR primers designed based upon conserved regions identified following alignment of PfCTRP and PbCTRP. The first vWA domain of PgCTRP was expressed in E. coli and was amplified from P. gallinaceum genomic DNA template using gene-specific primers designed based on the full-length P. gallinaceum CTRP gene. The primers used are as follows: 5Ј-GCGCCATGGGTGAAAGTAA-CAAAGATGAATC-3Ј and 5Ј-GCGCTCGAGTGGTTTATGATCTGGTT-TTG-3Ј (NcoI and XhoI restriction sites are underlined, respectively). The PCR product was cloned into the E. coli expression vector, PET32b, and expressed in E. coli strain AD494 (DE3) (Novagen). Protein was induced at A 600 ϭ 0.6 with 0.1 mM isopropyl-1-thio-␤-D-galactopyranoside at 15°C in a shaking incubator for 16 h. Bacterial pellets were lysed by sonication in a buffer of 20 mM Tris, 50 mM NaCl, 10 mM imidazole, pH 8.0, in the presence of a protease inhibitor mixture (Roche Applied Science; Complete, EDTA-free Protease Inhibitor Mixture Tablets (catalog number 1 873 580) containing antipain, chymostatin, E-64, leupeptin, and 4-(2-aminoethyl)benzenesulfonyl fluoride). Soluble recombinant protein was purified by nickel-Sepharose affinity chromatography (Amersham Biosciences) by using an imidazole step gradient in 20 mM Tris, 50 mM NaCl, pH 8. The recombinant protein eluted at an imidazole concentration of 100 mM. Full-length P. falciparum WARP (minus the signal peptide sequence) was amplified from P. falciparum strain 3D7 genomic DNA using the following PCR primers designed to include the native stop codon: 5Ј-GCGGAATTCCGAAT-GAACGTAGTGTCTCATA-3Ј and 5Ј-GCGCTCGAGCATCTTATGATT-TATTCTTATCACA-3Ј (BamHI and XhoI restriction sites are underlined, respectively). The PCR product was cloned into the expression vector pGEX4T-2 (Amersham Biosciences) expressed as a fusion protein with Schistosoma mansoni glutathione S-transferase (GST). Protein was produced in E. coli strain AD494 (DE3) induced at A 600 ϭ 0.6 with 0.1 mM isopropyl-1-thio-␤-D-galactopyranoside at 15°C in a shaking incubator for 16 h. Bacterial pellets were lysed by sonication in PBS in the presence of a protease inhibitor mixture (Roche Applied Science). Soluble recombinant protein was eluted from a glutathione column (Amersham Biosciences) with 10 mM reduced glutathione in 50 mM Tris, pH 8.0, and dialyzed against PBS for further use.
Protein concentration was measured by the bicinchoninic acid assay (Bio-Rad). Mice were vaccinated with 50 g of protein emulsified in complete Freund's adjuvant (Sigma), followed by two booster injections at 3-week intervals with 25 g of protein emulsified in incomplete Freund's adjuvant.
ELISA to Test Ligand Binding by Recombinant PfWARP and the First vWA Domain of PgCTRP-Surveys of binding of recombinant PfWARP and the first PgCTRP vWA domain to potential ligands were performed similar to a protocol published previously (16) regarding E. coli-produced P. falciparum TRAP vWA domain, with the exception that biotinylated heparin bovine serum albumin (BSA) was used in the PgCTRP binding assays. Briefly, Dynex 4 plates (Fisher) were coated with 100 l of 100 g/ml collagen type I, collagen type IV, fibrinogen, unconjugated BSA, or heparin-conjugated BSA in Tris-buffered saline (20 mM Tris, 150 mM NaCl, 1 mM CaCl 2 , 1 mM MgCl 2 , pH 7.5) for 1 h at 37°C. As a positive control, a construct expressing the P. falciparum TRAP vWA domain fused to GST (kindly provided by Dr. C. J. McCormick (16)) was expressed in E. coli strain AD494 (DE3) cells and purified by glutathione affinity chromatography as above. For the rPf-WARP binding assays, recombinant (r) GST was expressed and purified for use as a negative control. For the PgCTRP binding assays, recombinant His 6 -tagged P. falciparum chitinase PfCHT1 (17) was expressed in E. coli for use as a negative control. The protein concentrations of the recombinant proteins were measured with the BCA assay (Bio-Rad) and adjusted to 10 g/ml for use in the ELISA. Binding of recombinant protein was assessed by using alkaline phosphatase-conjugated rabbit anti-GST (Amersham Biosciences), with para-nitrophenyl phosphate (Sigma) as substrate, and A 405 nm measured in a HTS7000ϩ microplate reader (Hewlett-Packard).
Immunofluorescence Microscopy-Ookinetes were dried onto 10-well Teflon-coated slides. The samples were blocked and permeabilized by incubation overnight in PBS containing 20% human serum and 3% Triton X-100. The samples were incubated with primary antisera recognizing the first vWA domain of PgCTRP and PfWARP or PgCHT1 and PfWARP for 1 h at room temperature. After washing five times with PBS, the samples were incubated with Alexa Fluor 546 goat anti-rabbit IgG conjugation and Alexa Fluor 488 goat anti-mouse IgG conjugation for 30 min at room temperature and washed five times with PBS and one time with water. Cellular immunolocalization was examined by using an Olympus BX51 fluorescence microscope. As negative controls, preimmune mouse and rabbit sera were used as primary antibody.
Immunoelectron Microscopy-Cultured ookinetes were fixed in 1% paraformaldehyde, 0.1% glutaraldehyde in PBS (pH 7.4) and embedded in LR White resin (Polysciences, Warrington, PA). Sections were blocked in PBS containing 5% nonfat milk, 0.01% Tween 20 and incubated with primary antibodies diluted in PBS/milk/Tween 20. Grids were incubated with 5 or 15 nm gold particle-labeled antimouse or anti-rabbit IgG diluted 1/20 with PBS/milk/Tween 20, stained in 2% uranyl acetate in 50% methanol, rinsed with 50% methanol, and stained with Reynold's lead citrate. Sections were carbon-coated in a vacuum evaporator and observed in a Hitachi H-800 electron microscope.
Purification and Identification of P. gallinaceum WARP-Pooled culture supernatants and soluble P. gallinaceum ookinete extracts were subjected to high pressure liquid chromatography under the conditions as described (4). Fractions were size-fractionated by SDS-PAGE on a 4 -20% gel (NOVEX gels, Invitrogen), which was stained with Coomassie Blue R-250 (Bio-Rad) and destained for 4 h with Gel-Stain Destain Solution (NOVEX) with three changes of destaining solution. Stained bands were excised from the gel, rinsed twice with 50% acetonitrile in HPLC-grade water, and frozen on dry ice. Subsequent protein sequencing steps were performed at the Harvard University Microchemistry Laboratory (Cambridge, MA). The band was subjected to in-gel reduction, S-carboxy-amidomethylation and tryptic digestion (Promega) (18), and a 10% aliquot of the resultant mixture was analyzed. Sequence information was determined by capillary (180 m ϫ 15 cm column, LC Packings, Amsterdam, The Netherlands) reverse-phase chromatography coupled to the electrospray ionization source of a quadrupole ion trap mass spectrometer (Finnigan LCQ, San Jose, CA). The instrument was programmed to acquire successive sets of three scan modes consisting of full scan MS over the m/z 395-1200 atomic mass units, followed by two data-dependent scans on the most abundant ion in that full scan. These data-dependent scans allowed the automatic acquisition of a high resolution (zoom) scan to determine charge state and exact mass, and MS/MS spectra for peptide sequence information. The remainder (90%) of the peptide mixture was separated by microbore high performance liquid chromatography using a Zorbax C18 1.0 ϫ 150 mm reverse-phase column on a Hewlett-Packard 1090 HPLC/1040 diode array detector. Optimum fractions were chosen based on differential UV absorbance at 205, 277, and 292 nm, peak symmetry, and resolution, and then further screened for length and homogeneity by matrix-assisted laser desorption time-of-flight mass spectrometry on a Thermo BioAnalysis Lasermat 2000 (Hemel, UK). Strategies for peak selection, reverse-phase separation, and Edman microsequencing have been described previously (18). MS/MS spectra were used as the basis of the final sequence interpretation.
The full-length PgWARP sequence was assembled from raw shotgun sequencing data from the Sanger Centre, Cambridge, UK. Other Plasmodium genomic data were obtained from www.plasmodb.org. The ClustalW algorithm in the T-Coffee software suite (www.ch.embnet.org/ software/TCoffee.html) (19) was used for alignment, and www.bork. embl-heidelberg.de/Alignment/consensus.html was used to identify the 85% consensus sequence of the WARP proteins (Fig. 2). The Signal P server was used to predict the secretory signal peptide for the WARP proteins (20).

RESULTS
Identification of WARP in the Ookinete Stage-As part of a previous study of P. gallinaceum ookinete-secreted chitinases (4), an unidentified ϳ220-kDa Coomassie Blue-stained protein band was nearly co-purified through multiple chromatographic steps from P. gallinaceum ookinete lysates combined with spent culture supernatant by SDS-PAGE (Fig. 1). This band was excised and subjected to in-gel tryptic digestion followed by HPLC separation. Amino acid sequences of two peptides were determined by de novo MS/MS sequencing as follows: one with sequence (L/I)(L/I)(L/I)VG(L/I)G(L/I)GNK, and the other with sequence V(L/I)(F/M)YYGYDYGANK. BLAST search of the Plasmodium genome data bases demonstrated that these peptide sequences belonged to a new gene encoding a protein containing a predicted secretory signal sequence (Fig. 2, underlined amino sequences) followed by a single vWA domain. This family of Plasmodium proteins was recently identified by computational genome analysis and a subtractive hybridization approach (10,21), demonstrated to be present within P. berghei micronemes, and named WARP (10). WARP is highly conserved between Plasmodium species (Fig. 2). WARP is the third vWA domain-containing protein, after TRAP and CTRP (Fig.  3), to be described as a Plasmodium mosquito stage protein (Fig. 2). In contrast to TRAP and CTRP, WARP notably lacks a transmembrane domain and is either secreted or remains as-sociated with the surface membrane via interaction with membrane-associated proteins.
A P. falciparum WARP cDNA was amplified by reverse transcriptase-PCR from total RNA of in vitro cultivated P. falciparum gametocytes. Comparison with P. falciparum genomic sequence information showed no nucleotide sequence differences and indicated that the gene is encoded on a single exon. Alignment of the predicted amino acid sequences of the WARP proteins from P. falciparum, P. gallinaceum, Plasmodium berghei, Plasmodium yoelii, Plasmodium knowlesi, and Plasmodium vivax shows significant cross-species conservation of sequence (PfWARP versus PgWARP 62% identical; PfWARP versus PyWARP 58% identical; and PfWARP versus PvWARP 67% identical), including rigorous conservation of seven cysteines (indicated by an asterisk above the cysteines in Fig. 2). The four most carboxyl-terminal cysteines delineate the vWA domain and are conserved with the vWA domain of the Plasmodium TRAP gene. The amino-terminal regions (in PfWARP, amino acids 25-71) of all WARP proteins share significant sequence similarity, but similar proteins were not identified in GenBank TM (nonredundant) via reiterative PSI-BLAST queries. This region contains an odd number of conserved cysteines, one of which lies within a nonconserved proline-rich low complexity region (in PfWARP, amino acids 69 -90) that separates the conserved amino-terminal region and the vWA domain.
Biological Activity of Recombinant PgCHT1, PfWARP, and the First PgCTRP A Domain-Recombinant PgCHT1, Pf-WARP, and the PgCTRP vWA domains were assayed for binding activity, in part to assess if protein was properly folded after expression in E. coli. Recombinant PgCHT1 was able to hydrolyze 4-methylumbelliferone chitotrioside, as reported previously (4). In a solid phase assay, E. coli-expressed recombinant PfWARP (Fig. 4) and the first vWA domain of PgCTRP (Fig. 5) (expressed according to the schematic in Fig. 3), were tested for their ability to bind to a variety of components found in intercellular matrices. Recombinant P. falciparum TRAP was used as a positive control in these assays; either recombinant GST or an irrelevant His 6 -tagged protein (PfCHT1) were used as negative controls. Recombinant PfWARP (Fig. 4) and the PgCTRP first vWA domain (Fig. 5) did not demonstrably bind to collagen, fibrinogen, or, as a negative control, bovine serum albumin. Both proteins had significant binding to heparin conjugated to bovine serum albumin, suggesting specific binding to a highly negatively charged carbohydrate-type ligand. Although these data do not demonstrate the in vivo ligand recognized by these proteins, they suggest that the recombinant proteins are properly folded and bind to an appropriate matrix substrate that might be predicted to be present on target cells/basement membrane encountered during ookinete transit through the midgut wall.
In Vivo Expression of PgCHT1, PgCTRP, and PgWARP in P. gallinaceum Ookinetes-Western immunoblotting assays were performed to determine whether hyperimmune mouse polyclonal antisera recognized P. gallinaceum ookinete-produced, native PgCHT1, PgCTRP, and PgWARP. An appropriately sized band was detected for all three proteins (Fig. 6). As predicted, the proteins were not detected in 3-4-h-old zygotes (Fig. 6); the parasite stage formed immediately after gamete fertilization but appeared in ookinetes, which begin to differentiate 10 -15 post-fertilization. PgWARP was detected as a high molecular mass protein (ϳ250 kDa) under nonreducing conditions; under reducing conditions, the protein appeared as ϳ45 kDa. This observation strongly suggested that the anti-WARP antiserum did not cross-react with the high molecular weight, monomeric protein PgCTRP, despite the presence of  (lanes 61-64) Two-color immunofluorescence microscopy was performed to determine PgCHT1, PgCTRP, and PgWARP localization within mature P. gallinaceum ookinetes and to provide insight into the overall intracellular distribution and spatial relationships of these three proteins (Fig. 7). PgCHT1 and PgCTRP were present in a granular appearance throughout the ookinete cytoplasm, with notable concentration at the apical end of the parasite. When visualized simultaneously with PgCHT1, Pg-WARP was also present in a granular pattern throughout the cytoplasm, but with less focal concentration at the apical end of the parasite in comparison with the appearance of PgCHT1 and PgCTRP. These findings were reproducible in more than 40 morphologically mature ookinetes examined. As a negative control, polyclonal antisera raised against E. coli-produced recombinant His 6 -tagged P. vivax chitinase PvCHT1 (from inclusion bodies; an irrelevant antigen because there was no reactivity to the protein on Western immunoblot (data not shown)) produced no signal by immunofluorescence microscopy.
Presence of Micronemal Proteins in the Same Secretory Pathway and Micronemes of Ookinetes-A previous report (10) suggested that the P. berghei CTRP and WARP homologues were present in different micronemal populations, based upon single FIG. 2. Alignment of WARP proteins from P. falciparum, P. gallinaceum, P. yoelii, P. berghei, P. vivax, and P. knowlesi. Identically conserved amino acid residues are indicated by an asterisk; highly conserved amino acid residue by a colon, and weakly conserved by a period. An 85% consensus sequence is shown; identical residues are shown in capital letters. Below the protein sequences are consensus positions identified as follows: s, amino acids with small side chains (Ala, Gly, Ser, Pro, Asp, Asn, and Val); a, amino acids with aromatic side chains (Phe, Tyr, and Trp); p, amino acids with polar side chains (Lys, Arg, His, Glu, Asp, Gln, and Asn); h, amino acids with hydrophobic side chains (Leu, Ile, Val, Met, Tyr, Phe, Trp, and Ala); l, amino acids with aliphatic side chains (Ala, Leu, Ile, and Val); c, amino acids with charged side chains (Glu, Asp, Lys, His, and Arg); ϩ, amino acids with basic side chains (Lys, His, and Arg); and u, amino acids with tiny side chains (Gly, Ala, and Ser). Predicted signal peptide sequences are indicated in boldface at the amino terminus. Experimentally determined peptides initially identifying the protein as PgWARP are underlined.
staining immunoelectron microscopy and morphological appearance of the micronemes. To investigate whether another soluble protein, PgCHT1, was present in a similar or distinct population of micronemes compared with CTRP, dual immunogold electron microscopy was performed on P. gallinaceum ookinetes (Fig. 8). CTRP, a protein with a predicted single transmembrane ␣-helix, was found in a circumferential "rim" distribution within micronemes concentrated at the ookinete apical end (large (15 nm) gold dots). In contrast, PgCHT1, a protein that is not cell surface-associated and lacks predicted transmembrane domains, was found soluble within the central lumen of the same micronemes as those containing CTRP (small (5 nm) gold dots). Similarly, PgCHT1 and PgWARP are both secreted in close proximity through the electron-dense area of the apical complex and appear within the same micronemes (PgCHT1 labeled with large (15 nm) gold dots; Pg-WARP labeled with small (5 nm) gold dots) (Fig. 9, inset). PgWARP, despite the absence of identifiable transmembrane domains or glycosylphosphatidylinositol-anchoring motifs, appears to be present on the ookinete cell surface, including at the posterior end of the parasite. These observations suggest that there is likely a single pathway for extracellular secretion of proteins through the micronemal pathway and that the same set of micronemes transports soluble, membrane-associated, and complexed proteins for secretion through the electrondense region of the ookinete apical complex.
Effect of Antibodies against Micronemal Proteins on the Infectivity of P. gallinaceum for A. aegypti Mosquitoes-Membrane feeding experiments were performed to test the hypothesis that the micronemal proteins chitinase, WARP, and CTRP would be immunological targets of blocking malaria transmission. Antisera to all three P. gallinaceum proteins significantly inhibited the formation of oocysts in A. aegypti. Polyclonal anti-rPgCHT1 antisera reduced the number of oocysts by 67-96% and the proportion of infected mosquitoes by 37-76% (Table I), statistically strongly significant compared with both infectious blood meal sham controls and with negative control antisera obtained from mice vaccinated with two different irrelevant proteins (administered with Freund's adjuvant to control for the effects of the vaccination procedure). Polyclonal anti-PgCTRP antisera reduced the number of oocysts by 78 -88% and the proportion of mosquitoes infected by 38 -68% (Table II). Polyclonal PfWARP antisera reduced oocyst counts 70 -82% and the proportion of mosquitoes infected by 34 -51% (Table II). To determine the effect antiserum combinations on P. gallinaceum infectivity to Ae. aegypti, mosquitoes were fed with either a combination of anti-PgCTRP ϩ anti-PfWARP or all three antisera mixed together in the membrane feeder (Table III). The combination of polyclonal antisera against PgC-TRP and PfWARP further reduced the number of oocysts in a reproducible and statistically significant fashion. The presence of all three antisera in the blood meal did not additionally reduce transmission, compared with combinations of two anti- Effect of Antibodies against Micronemal Proteins on Infectivity of P. falciparum for Anopheles Mosquitoes-Membrane feeding assays using cultured P. falciparum strain 3D7 gametocytes were performed to determine whether antibodies to the P. falciparum chitinase PfCHT1 and the P. falciparum WARP orthologue inhibited parasite infectivity to Anopheles mosquitoes similar to what was observed with homologous proteins in the P. gallinaceum/A. aegypti system.
Monoclonal antibody 1C3, raised against enzymatically active recombinant PfCHT1 (13), when added to infectious blood meals fed to colonized Anopheles gambiae strain G3 and Anopheles stephensi mosquitoes, significantly inhibited oocyst formation and reduced the proportion of mosquitoes infected in a dose-dependent manner (Table IV). In A. gambiae, oocyst counts were reduced by 85-94% with a concomitant reduction in the proportion of mosquitoes infected by 73-86%. In A. stephensi mosquitoes, the effect of mAb 1C3 was even more profound, with 92-100% reduction in oocyst counts and 84 -100% reduction in proportion of mosquitoes infected.
Mouse and rabbit polyclonal antisera raised to biologically active recombinant PfWARP had nearly complete transmission blocking activity in membrane feeding assays of P. falciparum to An. gambiae mosquitoes (Table V). DISCUSSION Most strategies for identifying malaria transmission-blocking vaccine targets have focused on early sexual stage parasite surface proteins, i.e. present on gametocytes, gametes, or zy-gotes (23,24). In this study, we demonstrate that antibodies directed against three late mosquito midgut stage proteins have significant transmission blocking activity, namely the ookinete-secreted proteins PgCHT1 chitinase, WARP, and CTRP. The transmission blocking activity of antisera persisted despite the presence of high levels of proteolytic activity in blood meal digestion, secreted by epithelial cells into the midgut. The antisera likely act on ookinete stages of the parasite because the proteins first appear in ookinetes (Fig. 6). Immunolocalization studies demonstrate that all three ookinete-secreted proteins are present in micronemes (7,10,25). Furthermore, evidence presented here suggests that PgCHT1, PgWARP, and PgCTRP are present within the same microneme population and are secreted through a common pathway through the electron-dense apical end of the ookinete. Finally, transmission-blocking results in the avian model of malaria transmission were confirmed by parallel studies of the lethal human malaria parasite, P. falciparum. Taken together, these observations suggest that as a class, ookinete-secreted micronemal proteins are candidate target antigens blocking malaria transmission. One could speculate that interruption of microneme formation or trafficking of secreted/chaperone proteins into micronemes may be useful as a novel approach to interventions against malaria, including transmission-blocking approaches.
The experiments presented here used the following E. coliproduced recombinant proteins: PfCHT1, PgCHT1, PgCTRP (the first vWA domain), and PfWARP. Assays of adhesive activity performed in this study indicate that the proteins were properly folded and had appropriate biological activities. Chitinase activity for rPfCHT1 and rPgCHT1 has been reported previously (4,17); additionally in the presence of 100 mM dithiothreitol, chitinase activity was maintained indicating that the enzymatic activity of these proteins for small oligosaccharide substrates is not disulfide-dependent (data not shown). We demonstrate here that the first vWA domain of PgCTRP and recombinant PfWARP bound to highly negatively charged glycosaminoglycans (heparin) but not to other proteins found in basement membrane such as fibrinogen and collagen type I or IV or to highly negatively charged albumin. These findings do not confirm that heparin is the natural ligand of these ookinete proteins. Rather, in the limited present study we interpret these results simply to indicate that these disulfide-dependent proteins, as produced in E. coli, likely induce conformationally dependent, neutralizing protective antibodies. Experiments are in progress to use recombinant PfWARP to identify whether it has a natural mosquito midgut ligand, which would  be of substantial interest as a potential ookinete invasion receptor. Fundamental studies of Plasmodium ookinete cell biology are likely to yield novel, practical, and pre-clinical development of malaria transmission-blocking strategies and identification of potential vaccine targets. We demonstrate that three micronemal proteins of the ookinete are secreted via a common organelle, through a common pathway in the electron-dense region of the apical end. This finding stands in contrast to a previous report that seemed to indicate that membrane-bound versus soluble proteins were present in different micronemal populations (10). Indeed, we found that PgWARP is present as a reduction-sensitive, high molecular weight complex (Fig. 2) and located within the same micronemal population as PgCHT1 (soluble) and PgCTRP (membrane-bound by a single transmembrane pass). Furthermore, PgWARP localizes to the ookinete cell surface as a high molecular weight complex, despite the absence of recognizable membrane-localizing motifs. This character is reminiscent of the chitinase, PgCHT2, which also associates in multimer complexes on the ookinete surface (4,13). PgCHT2 is the second chitinase of P. gallinaceum and is orthologous to PfCHT1. 2 Ongoing experiments will address the mechanism(s) by which WARP (and PgCHT2) associates with the cell surface and seek to determine the multimeric structure of the membrane surface high molecular weight complexes. It is also important to determine whether the transmission blocking activity of anti-WARP antibodies, as shown here and elsewhere (21), is because of reaction sith serface-associ-ated or secreted, soluble WARP protein or whether the antiserum is involved in disrupting a putative interaction of WARP with a mosquito or parasite ligand.
There is still debate in the literature concerning the role of the peritrophic matrix as a potential barrier to ookinete invasion of the mosquito midgut. Whereas disruption of the P. falciparum chitinase gene, PfCHT1, completely prevented P. falciparum oocyst development in Anopheles freeborni mosquitoes, knockout of the P. berghei chitinase gene PbCHT1 had only a partial effect on reducing oocyst counts. This discrepancy may be attributable to the fact that P. berghei ookinetes may develop sooner than those of P. gallinaceum and P. falciparum and escape the midgut before the peritrophic matrix fully forms (26 -28). Furthermore, previous reports have suggested that the peritrophic matrix in An. stephensi mosquitoes and perhaps other Anopheles spp. mosquitoes lack chitin (29), thus reducing the potential importance of ookinete-secreted chitinase as a transmission-blocking target. In the experiments presented here, mAb 1C3 had strong transmission blocking activity in both An. gambiae and An. stephensi. It is possible that the mechanism by which mAb 1C3 reduced P. falciparum infectivity for An. gambiae and An. stephensi was at a level other than at the peritrophic matrix, i.e. by interacting with the surface of the ookinete directly rather than neutralizing the chitinase activity (13). Regardless of the mechanism by which mAb 1C3 reduced P. falciparum oocyst counts in two species of Anopheles mosquito, the data presented here are consistent with PfCHT1 being a potential target of blocking parasite transmission to mosquitoes.
Possible cross-reactivity could have limited interpretation of 2 F. Li and J. M. Vinetz, manuscript in preparation. the membrane-feeding experiments using anti-PfWARP antisera in both the P. gallinaceum and P. falciparum systems. We showed that anti-PfWARP antisera reacted specifically with a single, reduction-sensitive, high molecular weight protein present in P. gallinaceum ookinetes but not in 3-4-h-old zygotes (Fig. 6), consistent with previous reports (10). Furthermore, numerous negative control antisera had no transmission blocking activity in the experiments where anti-PfWARP antisera were added to membrane feeds. These results confirm that the anti-PfWARP antisera were specific for PgWARP and reduce the potential concern whether such antibodies might crossreact with CTRP (30). Similar to the experience of others in using membrane feeding assays to determine the effect of various antisera on Plasmodium infectivity to mosquitoes, we observed experiment-to-experiment variability in the numbers of oocysts obtained, with relatively low geometric means oocyst counts observed in both P. gallinaceum and P. falciparum experiments. Such variability has been observed by others, with P. gallinaceum, P. berghei, P. vivax, and P. falciparum (15,22,31,32). Each of the membrane feeding experiments presented here, both of P. gallinaceum and P. falciparum, was analyzed independently with internal negative controls included in each assay. Each membrane feeding assay was performed at least three times for each treatment group, and similar magnitudes of statistically significant oocyst reduction were observed, despite varying numbers of oocysts seen in the negative control groups. Therefore, we conclude that despite the variability of oocyst counts in the experiments presented, statistically robust conclusions about the effect of the anti-micronemal protein antisera can be made. In summary, we present evidence that validates, as proof-ofprinciple, three Plasmodium ookinete-secreted proteins (chitinase, WARP, and CTRP) as malaria transmission-blocking targets in both an avian model system and with the human pathogen P. falciparum. Despite the contrasting final destinations of these proteins, they are secreted by a common micronemal pathway through the electron-dense region of the apical complex of the ookinete. With ongoing genome-scale efforts to delineate the proteome and gene expression repertoire of Plasmodium sexual stage genes, novel insights into the fundamental biology of malaria parasites will be gained, with the anticipation of applying such results to the amelioration and control of human malaria.