A Multidomain Adhesion Protein Family Expressed in Plasmodium falciparum Is Essential for Transmission to the Mosquito

The malaria parasite Plasmodium is a major cause of global human morbidity and mortality, whereas other apicomplexan parasites like Cryptosporidium parvum and Toxoplasma gondii are important opportunistic pathogens in immunocompromised individuals such as AIDS patients. An important aspect of apicomplexan biology is the molecular basis for their interaction with target cells, extracellular matrices, and hemo-lymphatic fluids of their animal hosts. Due to their extracellular localization, surface proteins proposed to be involved in these processes are likely accessible by host-immune sera and are therefore vaccine candidates stimulating antibody-mediated protection. Recent methodologies to identify extracellular proteins as research targets include “postgenomic” techniques such as genome sequence annotation, microarray screening, and proteomics. Of particular importance, completion of Plasmodium sp. genome sequences (1, 2) and the imminent completion of additional apicomplexan genomes have provided an unprecedented opportunity to identify novel antigens as prospective vaccine targets for study, as well as underpin a better understanding of Plasmodium and apicomplexan biology.

In the search for novel extracellular proteins as prospective vaccine candidates, we have exhaustively identified genes encoding extracellular proteins through genome sequence analysis of Plasmodium sp. and the recently completed C. parvum genome sequence, as well as whole genome comparison of the resulting catalogs of extracellular proteins (unpublished data). Here we describe a select subset of sexual stage–specific genes in the human malaria parasite, P. falciparum, that are striking for multidomain architectures composed of animal- and bacterial-like putative adhesive domains. Orthologues are present in the C. parvum genome sequence, indicating a cellular function likely conserved across the apicomplexan clade. These proteins are grouped as a family based upon a common Limulus coagulation factor C (LCCL) domain and are accordingly named CCp proteins, of which five genes are present in the P. falciparum genome sequence (PfCCps). One of these proteins, herein referred to as PfCCp3, was recently described as PSLAP in P. falciparum (3) and PbSR in Plasmodium berghei (4). These two studies presented, in part, contradictory conclusions regarding protein expression. Although PSLAP was shown to be expressed in the gametocyte stages, PbSR protein expression was described as sporozoite stage–specific and PbSR-targeted gene disruption led to inhibited sporozoite formation. In this study, we characterize three P. falciparum members of the LCCL family, PfCCp1, PfCCp2, and PfCCp3/PSLAP, based upon their striking multiadhesion domain structures. We show that PfCCp1, PfCCp2, and PfCCp3 are specifically expressed in gametocytes and released during gametogenesis. Disruption of the PfCCp2 or PfCCp3 gene loci via homologous recombination leads to a blockage in the P. falciparum sporozoite midgut to salivary gland transition. We suggest that the PfCCp protein family serves a role in multiple mosquito parasite stages and has an essential role in transmission to the mosquito, thus warranting further study as transmission-blocking vaccine candidates.

CCp genes were identified via screening of complete Plasmodium sp. and C. parvum genome sequence databases using the gapped version of the BLAST programs (BLASTPGP and TBLASTNGP, with both the single pass and the recently incorporated multipass profile option for translating searches of nucleotide databases). Sequence profile searches were performed using the PSI-BLAST program (5). Profiles were saved using the -C option and retrieved using the -R option. Complete domain architectures of identified proteins were determined by screening the nonredundant protein sequence database, the Expressed Sequence Tags database (National Center for Biotechnology Information), and the individual protein sequence databases of completely and partially sequenced genomes (accessible at http://www.ncbi.nlm.nih.gov/Microb_blast/unfinishedgenome.html). Multiple alignments of amino acid sequences were generated using a combination of PSI-BLAST and T-Coffee (6). Secondary structure predictions were produced using the PHD program (7), with multiple alignments of individual protein families used as queries. Preliminary phylogenetic trees were constructed using the neighbor-joining algorithm as implemented in the Mega2 package (8), or the least squares methods as implemented in the FITCH program of the PHYLIP package (9). These trees were used to obtain the maximum likelihood tree using the PROTML program of the MOLPHY package. The statistical significance of the trees was assessed using the relative estimates of the logarithmic likelihood (RELL) bootstrap values (10). Signal peptides in protein sequences were predicted using the SignalP program (11) and membrane topologies were predicted using the TMHMM2 package (12).

Asexual and mature gametocyte NF54 isolate P. falciparum parasites were cultivated in vitro as described previously (13). Gametogenesis was induced by incubating mature gametocyte cultures in human serum for 5–30 min at room temperature as described previously (14). Zygote formation was achieved by incubation of emerged gametes in human serum overnight at room temperature. Zygotes and gametocytes were enriched via centrifugation over 6/11/16% Accudenz gradients and parasites were collected at gradient interfaces. Mosquito stages (retorts, ookinetes, oocysts, and salivary gland sporozoites) were obtained after membrane feeds of Anopheles freeborni mosquitoes on blood containing mature gametocytes as described previously (15). Retort and mature ookinetes were obtained by harvesting midgut bloodmeals 24 h after feeding. Oocysts were isolated from mosquito midguts 12 d after feeding. Sporozoite preparations were enriched by centrifugation from pooled, homogenized infected salivary glands isolated 18–20 d after feeding. To determine the efficiency of mosquito transmission, WT NF54- or PfCCp-KO–targeted gene-disruptant isolates were membrane fed to mosquitoes and oocyst numbers were assayed in mercurochrome-stained midguts 8–12 d after feeding. Similarly, the presence of salivary gland sporozoites was assayed 18–21 d after feeding.

Recombinant protein construct designs were guided by domain architecture determinations for PfCCp1 (sequence data are available from GenBank/EMBL/DDBJ under accession no. NP_702612.1), PfCCp2 (GenBank/EMBL/DDBJ under accession no. NP_702421.1), and PfCCp3 (GenBank/EMBL/DDBJ under accession no. NP_701955) amino acid sequences. For each PCCp gene, recombinant protein was produced corresponding to discrete domains, such as the LCCL domain (PfCCp1), the discoidin domain (PfCCp2), and the two tandem scavenger receptor (SR) domains (PfCCp3). Recombinant protein was expressed as fusions downstream of the maltose-binding protein (MBP) using the pMAL expression vector (New England Biolabs, Inc.). All constructs were generated by ligation of PCR products corresponding to discrete PfCCp domains amplified from 3D7 (clone of NF54 isolate) P. falciparum gametocyte stage cDNA. Recombinant protein was purified from bacterial extracts by sequential steps of amylose affinity (New England Biolabs, Inc.) and ion exchange chromatography (Source 15Q resin using an AKTA Prime chromatography system; Amersham Biosciences). 6-wk-old female CAF1 mice were immunized with 30–100 μg recombinant protein emulsified in Freund's complete adjuvant followed by two boosts at 3-wk intervals in Freund's incomplete adjuvant. Immune sera were collected 10 d after the third immunization. All antisera recognized cognate recombinant protein via Western blot analysis. Sera from prebleed and MBP protein alone–immunized mice served as controls for all antibody reagent studies. As a second source of antisera for double labeling experiments, antipeptide sera, directed against a PfCCp2 peptide sequence within the discoidin domain (CDKIKDTDDARDKYF), was produced in rabbits immunized with KLH-conjugated material produced using a maleimide-activated KLH kit (Pierce Chemical Co.). Immunizations of two rabbits were performed by Spring Valley Laboratories, Inc.

Mixed asexual stage parasites and Accudenz gradient–enriched gametocytes were saponin lysed and suspended in PBS and SDS-PAGE loading buffer. Salivary glands infected with sporozoites were dissected from An. freeborni mosquitoes and suspended in SDS-PAGE loading buffer. Parasite proteins were separated by electrophoresis (Novex PAGE system; Invitrogen) and transferred to Hybond ECL nitrocellulose membrane (Amersham Biosciences) according to the manufacturer's instructions. Membranes were blocked for nonspecific binding in Tris-buffered saline containing 3% skim milk and incubated for 2 h at room temperature with mouse immune sera specific to PfCCps, mouse mAb 2A10 recognizing the sporozoite surface circumSporozoite (CS) protein (16), MBP fusion partner protein control, or, as a control for loading, mouse sera recognizing the abundant P. falciparum endoplasmic reticulum (ER) protein, Pf39 (17, 18). After washing, membranes were incubated for 1 h at room temperature with an alkaline phosphatase–conjugated secondary antibody (Kirkegaard & Perry Laboratories, Inc.) and developed in NBT/BCIP solution (Sigma-Aldrich) for 5–15 min.

Parasite preparations for IFA microscopy included NF54 isolate–mixed asexual stages, mature gametocytes, emerging gametes, mosquito midgut blood smears 24 h after feeding, midgut oocysts 12 d after feeding, or salivary gland sporozoite smears 18 d after feeding. Preparations were either fixed for 4 d in 4% paraformaldehyde in PBS at 4°C or were air dried on slides and fixed for 10 min in −80°C cold methanol. For membrane permeabilization and blocking of nonspecific binding, fixed cells were incubated in 0.02% saponin, 0.5% BSA, and 1% neutral goat serum (Sigma-Aldrich) in PBS twice for 30 min each. Preparations were then incubated for 2 h at 37°C with either mouse immune sera recognizing specific PfCCps, mouse mAb 2A10 recognizing the CS protein, rabbit anti–α-tubulin II serum (American Type Culture Collection [ATCC]; reference 19), or rabbit immune serum recognizing Pfs25 (ATCC; references 20 and 21). Binding of primary antibody was visualized using fluorescence-conjugated goat anti–mouse or goat anti–rabbit secondary antibodies (Alexa Fluor 488 or Alexa Fluor 594; Molecular Probes). To highlight unlabeled parasites, selected specimens were further incubated with the nuclear dye TOTO-3 (Molecular Probes) for 30 min at 37°C. Paraformaldehyde-fixed samples were embedded in 1% agarose in PBS, coated onto slides, and air dried. In selected preparations, those not requiring Alexa 594 conjugation, counterstaining was performed using 0.05% Evans Blue (Sigma-Aldrich) in PBS. Labeled specimens were examined by confocal laser scan microscopy using a Zeiss LSM 510, and digital images were processed using Adobe Photoshop 6.0 software.

For preembedding immunolabeling, mixed asexual and gametocyte stages were fixed in 4% paraformaldehyde in PBS for 4 d at 4°C. Cells were blocked in 0.02% saponin, 0.5% BSA, and 5% neutral goat serum (Sigma-Aldrich) in PBS for twice for 30 min each, and then incubated for 2 h at 37°C with mouse immune sera against PfCCps, followed by incubation with a goat anti–mouse secondary antibody conjugated with either 10 nm gold (Amersham Biosciences) or alkaline phosphatase. Alkaline phosphatase labeling was detected by incubating cells in BCIP/NBT solution for 20 min. Cells were then embedded in agar, fixed in 1% glutaraldehyde and 4% paraformaldehyde in PBS for 1 d at 4°C, postfixed in 1% osmium tetroxide in PBS for 1 h, and dehydrated in successive washes at increasing concentrations of ethanol. Specimens were embedded in Epon (Electron Microscopy Sciences) at 60°C for 2 d. Photographs were taken with a JEOL 100CX-II transmission electron microscope and scanned images were processed using Adobe Photoshop 6.0 software. For the morphological examination of PfCCp2-targeted gene-disruptant (PfCCp2-KO) oocysts, infected mosquito midguts were fixed in 1% glutaraldehyde and 4% paraformaldehyde in PBS for 2 d, and postfixed in 1% osmium tetroxide and 1.5% K₃Fe(CN)₆ in PBS for 2 h, followed by incubation in 0.5% uranyl acetate for 1 h. Midguts were dehydrated in ethanol and embedded in Epon as described above.

A PfCCp2 KO construct, pPfCCp2-KO, was designed such that after homologous integration a “pseudo-diploid” locus is generated in which the first copy of the disrupted gene is truncated and possesses an in-frame stop codon at the 3′ end, whereas the second copy is truncated at the 5′ end and lacks a start methionine, a signal peptide sequence, and the start of the coding region. The disruption construct contained a 523-bp amino-terminal region segment of PfCCp2 that was PCR amplified from NF54 P. falciparum genomic DNA template using gene-specific 5′ primer 5′-TAGCGGCCGCACCCTAATGGCACCGAAT-3′ and 3′ primer 5′-ATGCGGCCGCTTAAGGAGAACTAGGTAACGC-3′. Each primer introduced a NotI restriction site (underlined) to facilitate ligation of restriction enzyme–digested PCR product into a NotI-cut pDT-Tg23 vector (22, 23). The PfCCp3 KO construct, pPfCCp3-KO, was similarly designed and contained a 565-bp amino-terminal region segment that was amplified from NF54 WT DNA using gene-specific 5′ primer 5′-TAGCGGCCGCAGCTGGATAGGTGCTCCT-3′ and 3′ primer 5′-ATGCGGCCGCTTATACACAAACGGTACCCCA-3′. The resulting constructs were loaded into uninfected erythrocytes via electroporation, and the erythrocytes were then inoculated with NF54 isolate parasites (24). At the time of the first parasite passage, this process was repeated such that the parasite cultures were diluted into freshly electroporated erythrocytes. After 3 d, 10 mM pyrimethamine drug pressure was initiated to select for TgDHFR expression in parasites that have taken up and retained plasmid. After ∼90 d of pyrimethamine drug pressure a plasmid-integrant population predominated and the presence of homologous gene disruptants was demonstrated by a diagnostic PCR assay of a disrupted locus (described in Fig. 7). At this time, transformed parasite clones were isolated by dilution cloning in microtiter plates using the Malstat reagent assay (25) as an indicator of parasite-positive wells. Homologous recombination in clonal isolates was confirmed by the above PCR assay and in select isolates was further confirmed by Southern blot analysis (see Supplemental Materials and Methods).

Methods regarding RNA isolation, transcript level analysis, and Southern Blot analysis are described in Supplemental Material and Methods. Fig. S1 demonstrates gametocyte stage–specific transcript expression of the six genes encoding the multidomain adhesion proteins PfCCp1 through PfCCp5 and PfFNPA. Fig. S2 shows analysis of PfCCp2- and PfCCp3-targeted gene disruptions by Southern blotting (Fig. S2, A and C), and the resulting loss of PfCCp2 and PfCCp3 protein expression analyzed by Western blotting (Fig. S2, B and D). Supplemental Materials and Methods and Figs. S1 and S2 are available.

Annotation of apicomplexan genomes reveals numerous genes encoding extracellular proteins having multidomain architectures reminiscent of, but in no case identical to, extracellular proteins found in multicellular organisms. We have recently conducted a whole genome comparison of P. falciparum and C. parvum genome sequences (unpublished data) and have exhaustively identified genes remarkable for multidomain “animal-like” architectures, in some instances conserved in orthologues present across the apicomplexan clade. Five of these proteins exhibit striking multiple adhesion modules and are herein grouped based upon the presence of a shared LCCL-like domain. The P. falciparum versions were accordingly termed PfCCp1 through PfCCp5 (see schematic in Fig. 1). PCCp3 has recently been variously described as PSLAP in P. falciparum and as PbSR in P. berghei to highlight the presence of tandem SR-like domains in addition to four copies of the LCCL domain (3, 4). CCp orthologues are conserved in C. parvum, as well as across the Plasmodium clade, with versions identified in the Plasmodium yoelii genome sequence and isolation of P. berghei versions, including PbCCp2 (sequence data are available from GenBank/EMBL/DDBJ under accession no. AF491294) and PbCCp3 (PbSR; GenBank/EMBL/DDBJ under accession no. AY034780). All PCCp genes possess signal peptide sequences as predicted by the SignalP algorithm (11). Transmembrane domains were not identified in PfCCp1 through PfCCp5, as predicted by the TMHMM2 program (12), nor are GPI anchor signal sequences present. Thus, given the predicted signal peptides, lack of ER retention or other organelle trafficking signals and the presence of disulfide bond containing extracellular adhesive domains combine to suggest that all PCCp proteins are secreted and serve extracellular functions, perhaps in parasite–parasite or parasite–host interactions.

PfCCp1 and PfCCp2 have similar architectures and are paralogs arising via relatively recent gene duplication after domain accretion. C. parvum also has paralogs, CpCCp1 and CpCCp2, and thus the gene duplication event likely occurred before the divergence of the apicomplexan genera. Within PfCCp1 and PfCCp2, we identified three predicted polysaccharide-binding domains, namely ricin, discoidin, and levanase-type lectin domains (26, 27). COOH terminal to these modules are two copies of an apicomplexan-specific cysteine-rich module (herein termed ApicA) that has not been identified in any other gene. Between the discoidin and levanase lectin domains there are two distinct cysteine-rich modules, the described animal- and fungal-specific LCCL domain (28), and a novel module, termed NEC, which appears in a wide range of animal proteins such as neurexins (29), fibrillar collagen α globular domain (30) from vertebrates and sponges, and the fibrinogen family of proteins (31). The NEC domain has not been identified previously as a distinct module because in all the animal instances it occurs in combination with other cysteine-rich segments. This has resulted in it being considered a portion of several other distinct modules such as the fibrinogen globular domain (31, 32) or the fibrillar collagen α COOH-terminal globular domain (30). We term this module the NEC domain after neurexins (29) and collagens, the two ancient animal surface proteins in which they are represented. This is the first instance in which the NEC domain occurs in a stand alone form and is consistent with it forming a distinct NH₂-terminal folding α-β3 unit stabilized by a cysteine bridge in the fibrinogen crystal structure (Fig. 2). The structure of this domain reveals a distinct cleft that is likely to function as a binding pocket (32). Several animal proteins functioning as lectins contain the NEC domain, such as ficolin and intelectin (33, 34), suggesting that it probably represents a fourth sugar-binding domain in the PfCCp proteins. Phylogenetic analysis suggests that the levanase lectin and discoidin domains group closer to the bacterial forms as opposed to those from other organisms (Fig. 3 A), whereas the NEC domain groups with animal affinity (Fig. 3 B). Thus, the extracellular adhesion modules can be attributed to multiple distinct heritages, those originally derived in bacteria or animals and laterally transferred to Apicomplexa, and those “invented” within the lineage leading to the Apicomplexa.

The domain architecture of PfCCp3 (PSLAP/PbSR) has been recently described in detail (3, 4) and contains, in addition to four LCCL domains, a tandem repeat of an SR-like domain, an LH2 domain with a predicted role in lipid binding, and a region similar to pentraxin. Pentraxin and the SR domain have thus far been found only in the extracellular proteins of animals other than apicomplexans. PfCCp4 possesses a single LCCL domain followed by tandem levanase domains, plus additional anonymous cysteine-rich globular domains within the amino-terminal region. PfCCp5 encodes a region at the amino terminus that is similar to a domain found within fibronectin (FN2), followed by a domain similar to the amino-terminal region of anthrax-protective antigen that may bind host cell polysaccharides, and a single LCCL domain near the carboxy terminus. An additional P. falciparum gene, herein named PfFNPA, is similar to PfCCp5 in overall architecture but lacks an LCCL domain. Given that PfFNPA has an apparent C. parvum orthologue (Fig. 1), the two genes are likely to have diverged in P. falciparum from a common ancestor that was similar in architecture to PfFNPA, followed by accretion of an LCCL domain in PfCCp5.

Transcript expression was studied by real time RT-PCR analysis using RNA isolated from asexual and gametocyte intraerythrocytic in vitro–cultivated P. falciparum parasites. Expression was gametocyte stage specific for PfCCp1 through PfCCp4 and PfFNPA, in comparison with transcript expression controls for asexual (AMA-1) and gametocyte (Pfs48/45) stage–specific genes (Fig. S1). PfCCp5 showed transcript expression in asexual parasites in addition to its expression in gametocyte stages (Fig. S1). PfCCp1, PfCCp2, and PfCCp3 were chosen for further study of life cycle expression and cellular localization due to their striking multidomain architectures and the presence of orthologues in other apicomplexans. Protein expression for PfCCp1, PfCCp2, and PfCCp3 was shown to be gametocyte stage specific as assayed by Western blot analysis using mouse sera generated against PfCCp1, PfCCp2, and PfCCp3 recombinant protein (Fig. 4 A). Protein bands detected in gametocyte preparations had the expected molecular weights of roughly 185 kD for PfCCp1 and PfCCp2, and 150 kD for PfCCp3. Control mouse sera directed against the MBP recombinant protein fusion partner showed no labeling in gametocyte and asexual preparations, whereas the loading control mouse antibody Pf39 recognized a 39-kD ER protein (17, 18) abundantly expressed in both asexual and gametocyte parasites.

Cellular localization was studied via indirect IFA confocal laser scanning microscopy and immunoelectron microscopy. For PfCCp1, PfCCp2, and PfCCp3, a punctate rim fluorescence pattern was observed in P. falciparum gametocytes, whereas asexual stage parasites did not show any labeling (Fig. 4 B). Labeling was first observed in gametocytes stage III and remained throughout maturation (unpublished data). Colocalization of the three PfCCp proteins was investigated using rabbit sera generated against PfCCp2 peptide sequences in combination with mouse antisera against PfCCp1 and PfCCp3, respectively. Experiments revealed colocalization of all three proteins at the gametocyte surface (unpublished data). Double labeling experiments, using rabbit sera recognizing α-tubulin II as a marker for microgametocytes (19), further indicated that PfCCp proteins are expressed in both microgametocytes and macrogametocytes (unpublished data). The ultrastructural localization of PfCCp1 was additionally investigated by immunoelectron microscopy using two different preembedding labeling techniques. Immunogold labeling revealed localization of PfCCp1 associated with the parasite surface membrane in the parasitophorous vacuole of mature gametocytes (Fig. 5, A and C). Similar results were obtained by alkaline phosphatase labeling, which results in a loose precipitation at the site of protein expression (Fig. 5, B and D). Alkaline phosphatase labeling further showed protein localization in the cytoplasm of gametocytes and attached to the outside of erythrocytes, indicating possible release of the protein. No labeling was detected in asexual parasites or erythrocytes (Fig. 5 E). We further investigated the role of PfCCps during gamete emergence, triggered by incubation at room temperature in the presence of human serum. Although mature gametocytes exhibited PfCCp labeling associated with the parasite surface (Fig. 6 A), during emergence labeling was detected outside the infected erythrocytes (Fig. 6 B). Consistent with a secreted release of PfCCps, we have observed that PfCCp expression markedly decreases in emerged gametes and protein localizes extracellularly, adhering to complexes of macrogametes (Fig. 6 C). These complexes are readily observable in vitro due to the high gametocytemia of the parasite cultures and are formed by the inherently “sticky” macrogametes associating with exflagellating microgametes, mature gametocytes, and erythrocytes.

PfCCp expression ceases after emergence of gametes and formation of zygotes. Double labeling using rabbit sera recognizing Pfs25, a marker to identify unfertilized macrogametes, zygotes, and ookinetes (20, 21), revealed a low remaining expression for PfCCps in zygotes 20 h after gametocyte emergence in vitro (Fig. 6 D). Bloodmeal preparations isolated from mosquito midguts 24 h after feeding on gametocyte cultures did not show PfCCp protein expression in retort stages (Fig. 6 E) nor in ookinetes (Fig. 6 F). Macrogamete complexes present in the bloodmeal retained PfCCp labeling, suggesting that released PfCCp proteins persist within an extracellular matrix surrounding the macrogamete aggregates. PfCCp expression was not found in mature oocysts 12 d after feeding, nor in midgut or salivary gland sporozoites, using mAb 2A10 recognizing P. falciparum CS protein (16) as a positive immunolabeling control (unpublished data). Reexpression of PfCCp proteins in salivary gland sporozoites was additionally investigated by Western blot analysis. No labeling for PfCCp1, PfCCp2, or PfCCp3 was detected in a sporozoite lysate, whereas screening with mAb 2A10 detected two expected protein bands of ∼45–50-kD molecular weights (unpublished data).

The PfCCp2 and PfCCp3 gene loci were independently disrupted to begin characterizations of PfCCp family function, as well as to provide confirmations of life cycle stage of protein expression. Fig. 7 shows a schematic of the targeted gene KO construct, pPfCCp2-KO, designed to generate a disrupted pseudo-diploid locus after homologous integration. A pPfCCp3-KO construct was designed similarly for PfCCp3-targeted disruption. After transformation of P. falciparum parasites, pyrimethamine-resistant polyclonal lines were established and clonal lines were isolated. PfCCp2 gene disruption was verified in 12 isolated clones and PfCCp3 gene disruption was verified in 6 isolated clones by a PCR assay diagnostic of an integrated locus (using primers P1 and P2 in Fig. 7; unpublished data). Further characterizations focused on two clones for each disruptant gene, termed PfCCp2-KO isolates D-11H and F-1D and PfCCp3-KO isolates H-3D and I-9C. Southern blotting analysis confirmed disrupted gene loci in all PfCCp2-KO and PfCCp3-KO clones (Fig. S2, A and C). Protein expression was not detected by Western blot analysis in gametocyte protein extracts isolated from disruptant isolates when compared with a WT NF54 isolate control (Fig. S2, B and D). In vitro emergence assays using PfCCp2-KO and PfCCp3-KO mature gametocytes, followed by IFA using α-tubulin II or Pfs25 labeling, indicated a normal exflagellation behavior and typical zygote and retort formation, respectively, compared with WT cultures (unpublished data). Thus, the ability to disrupt the PfCCp2 or PfCCp3 gene loci in asexual parasites and the abrogation of protein expression in gametocytes indicate that neither PfCCp2 nor PfCCp3 are essential for proliferation of asexual stages, gametocytogenesis, maturation of gametocytes in vitro, and efficient gametogenesis.

To characterize PfCCp2-KO and PfCCp3-KO phenotypes during mosquito transmission, mature gametocyte cultures of WT parasites and gene-disruptant clones were fed to An. freeborni mosquitoes via membrane feeding. Six independent feeding experiments for PfCCp2-KO cultures and two independent feeds for PfCCp3-KO cultures revealed a normal infection rate of mosquito midgut oocysts compared with mosquitoes fed with WT parasites (Table I). The number of oocysts per midgut was comparable in WT and gene-disruptant lines. Ultrastructural examinations of PfCCp2-KO (Fig. 8 A) and PfCCp3-KO (Fig. 8 B) oocysts indicated apparent normal differentiation, in which developed sporozoites (Fig. 8, C and D) showed no morphological differences compared with WT midgut sporozoites. However, no sporozoites were detected 19 d after feeding within or associated with salivary glands, nor within the hemocoel of mosquitoes fed with the KO lines, whereas mosquitoes fed with WT cultures exhibited salivary gland infection rates ranging from 13 to 86% (Table I). To address the fate of sporozoites, in one experiment, >50 mosquito midguts were pooled 20 d after feeding with PfCCp2-KO line F-1D, homogenized, and enriched for sporozoites via differential centrifugation. Specimens were screened by IFA using the mAb 2A10 recognizing the CS protein and revealed CP protein expression in sparse, intact midgut PfCCp2-KO sporozoites having CS protein precipitate trails indicative of motility on the glass slides (unpublished data). Thus, PfCCp2- and PfCCp3-targeted gene disruption results in a developmental blockage in the sporozoite transition from the mosquito midgut to salivary gland sporozoites.

This study focused on the identification of novel antigens through sequence profile analysis and comparative genomics, followed by characterization of select extracellular proteins throughout the malaria parasite life cycle. Here we describe six Plasmodium genes unique to the Apicomplexa but having multiple domain architectures reminiscent of animal extracellular proteins. Five of these genes contain one or more LCCL domains, with four genes having orthologues in the distantly related coccidian, C. parvum. The conservation of multidomain architectures in orthologues across the apicomplexan clade implies that the proteins serve functions that have not diverged within specific apicomplexan lineages. Because Plasmodium and Cryptosporidium have highly diverged parasitic life strategies, it is anticipated that the multidomain adhesion proteins are not involved directly in host recognition, such as binding to host cell surface proteins or interacting with components of the host immune system. Otherwise, it would be necessary to propose that recognition of a specific host molecule is conserved across the apicomplexan clade. It is more likely that a proposed host ligand is “generic,” such as the sugar moieties of heparin sulfate or other proteoglycans within host cell surfaces or extracellular matrices. This proposal is supported by the presence of multiple putative polysaccharide-binding domains in the architectures of the family members. This is also consistent with the observed mutant phenotype in which PfCCp2 and PfCCp3-KO parasites fail to progress to the salivary glands, via possible disruption of interactions with host glycoproteins mediating this process. On the other hand, an alternative explanation might be that at least some of the proteins were recruited to serve an “ancient” parasite-specific cellular function, such as participation in a stage-specific aspect of the parasitophorous vacuole, formation of structures associated with the parasite pellicule, or serving in the formation of an extracellular matrix surrounding one or more emerged parasite stages. Either of the two proposed models must take into account a role for the multiple putative polysaccharide-binding domains within the LCCL domain proteins. It is likely that the expression stage specificity is conserved across genera and support for CCp function might ultimately be provided by the study of life cycle stage of expression and localization in additional apicomplexans, such as Cryptosporidium and Toxoplasma.

Transcript expression studies show abundant gametocyte-specific expression for five of the six multidomain adhesion proteins. Protein expression and immunolocalization studies for three selected proteins, PfCCp1, PfCCp2, and PfCCp3, indicate that the proteins colocalize and are released during gametogenesis and remain associated with gamete aggregates. Gametocyte-specific expression of PfCCp proteins is supported by high throughput mass spectrometry proteomic analysis of P. yoelii rodent malaria sexual stage parasites (35), as well as a recent study showing gametocyte-specific transcript and protein expression of PSLAP/PfCCp3 in P. falciparum (3). In the latter study, cellular localization revealed submembranous expression or secretion to the parasitophorous vacuole and erythrocyte cytoplasm of mature gametocytes. This localization pattern is similar to that observed in our indirect IFA and immunoelectron microscopy studies.

To begin to address function of the LCCL domain family members, we have disrupted the PfCCp2 and PfCCp3 gene loci and have shown that neither gene is essential for asexual proliferation or required for efficient gametocytogenesis. Despite abundant gametocyte stage expression in WT parasites, gene-disruptant lines formed gametes efficiently and, after membrane feeds to anophelene mosquitoes, formed midgut wall oocysts with efficiency similar to WT. Oocysts mature normally and intact sporozoites are released, but a block in transmission occurred during the mature midgut oocyst to salivary gland sporozoite transition. Hemocoel sporozoites were not detected, nor were sporozoites adhering to salivary glands. The fate of the gene-disruptant sporozoites is not yet known, with one possibility being that the sporozoites remain and subsequently degrade within the mosquito midgut. It is thus far not possible to interpret the discrepancy between the prominent PfCCp transcript and protein expression observed in gametocyte stages and the manifestation of a phenotype in PfCCp2 and PfCCp3 gene disruptants during late stages of midgut oocyst development. It is unlikely that the phenotype is due to an “epigenetic artifact” because it would require proposing that a similar epigenetic phenotype has occurred after disruption of either of the genetically unlinked genes, PfCCp2 and PfCCp3. The oocyst stage phenotype may either be linked in an as yet unknown fashion to protein function in gametocytes or might be due to abrogation of undetected low level sporozoite stage protein expression. The latter hypothesis is supported by recent description of a CpCCP1/2 homologue expressed in Cryptosporidium sporozoites (36, 37). CpCCP1/2 protein also localized to the parasitophorous vacuole in 72-h infected HCT-8 cells, thus drawing a parallel with Plasmodium gametocyte cellular localization in this study and providing support that the LCCL domain family members serve roles in multiple parasite stages.

Targeted gene disruption of PbSR (PbCCp3/PbSLAP) in the rodent malaria model, P. berghei, also results in a block in late stage oocyst maturation (4) and the phenotype is superficially similar to PfCCp2-KO and PfCCp3-KO. In contrast to the PbSR study (4), we were unable to detect PfCCp3 protein expression in sporozoites by IFA and Western blot assays using mouse sera directed against recombinant protein corresponding to the tandem SR domains. Whereas Claudianos et al. (4) describe 40- and 60-kD proteins by Western blot and immunolocalization in sporozoites using SR antipeptide sera, we detect a protein with the expected full-length molecular weight of 150 kD in P. falciparum gametocyte protein extracts, although we have not addressed possible proteolytic maturation of the protein during gamete emergence. The discrepancy in the two studies might be due to either sporozoite-specific processing resulting in a truncated protein contaminating protease activity in the sporozoite preparations, or nonspecific reactivity of one or more antibody reagents. Similar to our observations, Delrieu et al. (3) show that PSLAP/PfCCp3 is expressed in gametocytes and if gametocyte expression holds for P. berghei, it would suggest that expression of multiple PfCCp proteins are essential for sporozoite development, whereas expression in gametocytes is marked but not essential for in vitro development.

Transmission-blocking vaccines target sexual phase parasite antigens, thereby relying on human antibodies to inhibit parasite development within the mosquito host (38–40). The sexual stage expression and extracellular localization of PfCCp1, PfCCp2, and PfCCp3 make them potential vaccine candidates, and targeted gene disruption of PfCCp2 and PbSR/PfCCp3 further demonstrate that these highly conserved apicomplexan proteins are essential for P. falciparum development in the mosquito vector. PfCCp transmission-blocking assays are currently being pursued in our laboratory to investigate PfCCps as candidates for study as subunits of a transmission-blocking vaccine.

We thank Heather Eisen, Darrow Deluca, and Amina Kurtovic for technical assistance. We also thank Dr. Elizabeth Nardin for kindly providing the mAb 2A10, and David Keister, Olga Muratova, and Andre Laughinghouse for mosquito and parasite material, mosquito rearing, and mosquito feeds. Drs. Kirk Deitsch and Carl Nathan are thanked for critically reviewing the manuscript.

This work was supported by the Niarchos Foundation. The Department of Microbiology and Immunology at Weill Medical College of Cornell University acknowledges the support of the William Randolph Hearst Foundation.