In high‐transmission regions, protective clinical immunity to Plasmodium falciparum develops during the early years of life, limiting serious complications of malaria in young children. Pregnant women are an exception and are especially susceptible to severe P. falciparum infections resulting from the massive adhesion of parasitized erythrocytes to chondroitin sulphate A (CSA) present on placental syncytiotrophoblasts. Epidemiological studies strongly support the feasibility of an intervention strategy to protect pregnant women from disease. However, different parasite molecules have been associated with adhesion to CSA. In this work, we show that disruption of the var2csa gene of P. falciparum results in the inability of parasites to recover the CSA‐binding phenotype. This gene is a member of the var multigene family and was previously shown to be composed of domains that mediate binding to CSA. Our results show the central role of var2CSA in CSA adhesion and support var2CSA as a leading vaccine candidate aimed at protecting pregnant women and their fetuses.
Plasmodium falciparum causes the most severe form of human malaria, with more than two million deaths per year. Whereas adults in endemic areas usually develop immunity to clinical malaria, women, during their first pregnancy (primigravidae), become particularly susceptible to infection (Brabin, 1983). The pathologies are associated with massive sequestration of P. falciparum‐infected erythrocytes (IE) in the placenta. Pregnancy‐associated malaria (PAM) is mainly linked to a subpopulation of IE that adhere to chondroitin sulphate A (CSA) expressed by syncytiotrophoblasts in the placenta (Fried & Duffy, 1996). Placental isolates are functionally distinct because they do not bind to CD36, but instead bind to CSA (Fried & Duffy, 1996) and, to some extent, to hyaluronic acid (Beeson et al, 2000). With successive pregnancies, women develop broadly strain‐transcendent antibodies to the IE surface (Fried et al, 1998), suggesting that a vaccine against PAM is feasible. Although the search for the CSA ligands resulted in the identification of several different var gene products, conflicting data have emerged on their validity (reviewed by Rowe & Kyes, 2004).
The var multigene family consists of approximately 60 distinct members per haploid genome that encode P. falciparum erythrocyte membrane protein‐1 (PfEMP‐1; Gardner et al, 2002). Expression of the var genes is mutually exclusive, allowing the expression of only one PfEMP‐1 on the surface of each IE, mediating sequestration in different microvasculature sites (Chen et al, 1998; Scherf et al, 1998). Binding studies using recombinant PfEMP‐1 domains have shown interactions with various host receptors, such as CD36, ICAM‐1 and CSA (reviewed by Smith et al, 2001).
A large number of epidemiological studies on pregnant women in malaria endemic areas strongly support the concept that an anti‐disease vaccine that blocks the adhesion to CSA and thereby protects pregnant women is possible (reviewed by Duffy & Fried, 2003). To this end, it is crucial to define with certainty which parasite CSA‐binding ligands are expressed on the surface of IE during pregnancy. To assess the repertoire of CSA‐binding ligands, we generated disruption mutants of the var2csa gene that was previously reported to possess several CSA‐binding domains, and to be upregulated in placental parasites (Salanti et al, 2003; Gamain et al, 2005). The FCR3Δvar2csa mutant parasites did not recover the CSA‐binding phenotype, suggesting that a single member of the P. falciparum var gene family determines cytoadhesion to CSA.
Targeted disruption of the var2csa gene in FCR3 parasites
It has been reported that var2csa is transcriptionally upregulated and expressed at the surface of CSA‐binding parasites (Salanti et al, 2003, 2004; Gamain et al, 2005). To investigate the role of var2csa in P. falciparum IE adhesion to CSA, we established parasite lines with a disruption in the var2csa gene.
The pHTK‐var2csa vector contains the hDHFR gene flanked by the var2csa DBL3‐X and DBL5‐ε sequences (Fig 1A). Insertional disruptant mutants were generated by double‐crossover homologous recombination of the pHTK‐var2csa transfection construct, resulting in the replacement of the var2csa DBL4‐ε domain with the hDHFR expression cassette (Fig 1A). FCR3 parasites were transfected with pHTK‐var2csa and selected on WR99210 and ganciclovir to obtain FCR3Δvar2csa mutants. After selection of the FCR3Δvar2csa population for knob‐positive parasites using gelatin flotation, the mutants were cloned by limiting dilution and genetically characterized.
Clones were screened by PCR analysis for the disruption of the var2csa gene as well as for the absence of contaminating wild‐type var2csa and the presence of the HRP1 (KAHRP) gene (data not shown). The presence of the HRP1 gene, taken together with the enrichment by gelatin flotation, argues for the presence of knobs on the surface of the FCR3Δvar2csa IE. To confirm that pHTK‐var2csa had integrated into var2csa, Southern blots were performed using genomic DNA previously digested with BamHI, EcoRV or PvuII derived from parental FCR3 or recombinant parasites, and were hybridized with var2csa DBL3 or DBL5 radiolabelled probes (Fig 1B). These hybridizations showed bands of the expected size, indicating that the integration occurred at the predicted site within the var2csa gene (Fig 1A,B). Pulsed‐field gel electrophoresis (PFGE) was performed to further support the integration of the selectable marker cassette within the var2csa locus on chromosome 12 (Fig 1B). A clathrin heavy chain probe was used as a chromosome‐12‐specific marker. After the complete characterization of several mutant clones by PCR, and Southern blotting of both restriction enzyme digests and size‐fractionated chromosomal DNA (Fig 1B,C), two clones (1F1 and 2A5) were selected for further analysis.
FCR3Δvar2csa clones cytoadhere to CD36
To test the ability of the FCR3Δvar2csa mutants to cytoadhere, adhesion of the FCR3Δvar2csa mutants to CSA and CD36 was examined (Fig 2A). Equal numbers of erythrocytes infected with trophozoites of the FCR3Δvar2csa 1F1 and 2A5 mutant clones or control parasites were seeded on Petri dishes coated with different molecules. FCR3‐CSA and FCR3‐CD36 were used as controls. Whereas FCR3‐CSA IE bound in high numbers to CSA but not to CD36, no adhesion to CSA was observed for 1F1, 2A5 and FCR3‐CD36 IE. In contrast, 1F1, 2A5 and FCR3‐CD36 IE adhered strongly to CD36. These results show that the FCR3Δvar2csa mutants are still able to mediate binding to another host receptor. No cytoadhesion to BSA and chondroitin sulphate C was observed (data not shown).
Total RNA was isolated from ring‐ and trophozoite‐stage parasites to investigate var gene expression in the FCR3Δvar2csa and the parental FCR3 parasites selected for a CSA‐ or CD36‐binding phenotype. Whereas a full‐length var2csa transcript (∼13 kb) was observed in the FCR3‐CSA parasites, a nonfunctional truncated transcript (∼7 kb) was detected in the mutant clones 1F1 and 2A5 (Fig 2B).
Using a semiconserved varT11.1 exon II probe, larger transcripts of around 9 kb were identified in ring‐stage RNA of FCR3‐CD36 and in the two mutant clones, showing that full‐length var genes are transcribed in the CD36‐binding FCR3Δvar2csa parasites. This result, taken together with the presence of a nonfunctional var2csa truncated transcript, shows that mutually exclusive transcription of var genes can be overcome under certain conditions. As blots were washed using high‐stringency conditions, and because of the divergence in the var2csa exon II sequence, the var2csa transcript was not detected with the exon II probe in FCR3‐CSA parasites. Sterile exon II transcripts were detected at the trophozoite stages for all the parasite clones (Su et al, 1995). These results show that full‐length var genes mediating CD36 binding are transcribed in the FCR3Δvar2csa clones 1F1 and 2A5 and that disruption of the var2csa locus does not interfere with IE cytoadhesion to receptors such as CD36.
No adhesion of FCR3Δvar2csa clones to CSA
To determine the ability of the FCR3Δvar2csa mutants to recover cytoadherence to CSA, the parasites were re‐selected on CSA using different systems. Switching of var genes occurs in in vitro‐cultured parasites, and variants that are able to adhere to a large number of different host receptors have been isolated using receptor‐specific panning assays (Roberts et al, 1992; Scherf et al, 1998). FCR3Δvar2csa IE (clones 1F1 and 2A5) were first selected on recombinant human thrombomodulin‐coated plastic dishes (Parzy et al, 2000). After four pannings, no specific enrichment was observed (Fig 3A). However, FCR3‐CD36 wild‐type parasites could be selected for binding to CSA.
In addition, FCR3Δvar2csa IE were selected on Saimiri brain microvasculature endothelial cell clone Sc1707, which was previously described to express exclusively the adhesion receptor CSA and to be a useful cell system for selecting CSA‐binding parasites (Pouvelle et al, 1997). No adhesion of the FCR3Δvar2csa mutants to the Sc1707 cells was observed after five rounds of selection, whereas wild‐type FCR3‐CD36 population began to adhere to CSA after only one round of selection (Fig 3B). Parasites selected on Sc1707 were renamed FCR3‐CSA1707, FCR3‐CD361707, 1F11707 and 2A51707. Similar results were obtained by selecting the four parasite lines on CHO‐K1 cells (data not shown).
The adhesion phenotype of the Sc1707‐selected parasites was examined on adhesion receptors coated to plastic Petri dishes. Whereas FCR3‐CD361707 bound strongly to CSA, FCR3Δvar2csa clones 1F11707 and 2A51707 maintained the CD36‐binding phenotype that had already been observed before the selection procedure (Fig 3C,D). No CSA‐specific adhesion was detected after five pannings of the FCR3Δvar2csa clones 1F1 and 2A5 on Sc1707. Taken together, our experiments suggest that the var2CSA protein is essential for cytoadhesion of late‐stage FCR3‐IE to CSA, as no other protein emerged to compensate var2CSA loss.
In FCR3Δvar2csa mutant parasites, transcription of the full‐length var2csa and the expression of the protein on the surface of the IE are prevented. Disruption of var2csa did not, however, affect the ability of the mutants to switch expression to another var gene, and show adhesive phenotypes. This is illustrated by the CD36‐binding phenotype and the transcription of var transcripts in both FCR3Δvar2csa mutant clones. Similar conclusions were reached by previous studies, showing that targeted disruption of var genes only affected expression of this particular gene, but not that of other variants (Horrocks et al, 2002; Andrews et al, 2003).
The finding that the var2csa promoter remains active in both mutant clones, leading to the truncated var2csa transcripts, might be explained by the presence of the active hDHFR expression cassette that is introduced in the var2csa coding region. As each var gene represents a transcriptional unit, the active promoter of the selectable marker could disturb this silencing mechanism, with the silencer being either located in the var intron (Deitsch et al, 2001) or, alternatively, in a specific chromatin environment that leads to var gene repression (Freitas‐Junior et al, 2005). Importantly, it shows that the mutually exclusive transcription of var genes can be overcome under certain conditions, which are potentially related to the specific transcription mechanism of telomere‐associated genes (Duraisingh et al, 2005; Ralph et al, 2005).
Our FCR3Δvar2csa mutant clones 1F1 and 2A5 did not recover the CSA‐binding phenotype, after selection on different CSA‐expressing cells or recombinant human thrombomodulin. In previous reports, several var gene products other than var2CSA have been linked to cytoadhesion of IE to CSA, var‐CS2 and var1CSA (Scherf et al, 1998; Buffet et al, 1999; Reeder et al, 1999; Degen et al, 2000; Andrews et al, 2003). The presence and integrity of the var1csa and var‐CS2 genes in the genome of the FCR3Δvar2csa mutants was shown by PCR analysis (supplementary Fig S1 online). As var1csa is also transcribed in the mutant parasites, our results indicate that var1csa and var‐CS2 are not involved in CSA‐specific adhesion of late‐stage IE. Doubts on a functional role of var1CSA in cytoadhesion to CSA were raised earlier, because of the unusual transcription pattern of var1csa in CSA‐ and CD36‐binding parasites (Kyes et al, 2003). Also, in contrast to our FCR3Δvar2csa mutants, var1csa disruption mutants recovered the CSA‐binding phenotype after selection on CSA (Andrews et al, 2003), leading to the transcription of var2csa (Gamain et al, 2005). However, experimental evidence strongly supported a role for var1CSA in cytoadhesion to CSA (reviewed by Rowe & Kyes, 2004). First, var1CSA possesses a combination of a CSA‐binding DBLγ domain and a non‐CD36‐binding CIDR1 (Gamain et al, 2002). Second, antibodies raised against the var1CSA DBL3γ domain label placental parasite isolates and inhibit IE cytoadhesion to CSA (Costa et al, 2003). These antibodies also recognize the var2CSA protein on the surface of the IE (Andrews et al, 2003). At this stage of our knowledge, we assume that crossreactivity of the antibodies between var1CSA and var2CSA may be explained by similar three‐dimensional structures in the CSA‐binding domains of both genes. This hypothesis is supported further by the presence of significant homologies between the var1CSA DBLγ minimal binding domain and the var2CSA DBL3‐X binding domain (Gamain et al, 2004). Even if the var1csa gene product possesses all the characteristics that would make it a CSA‐binding ligand, its biological role remains uncertain. It is possible that the protein is not exported to the surface of the mature asexual‐stage IE or is expressed at another developmental stage of the parasite, such as the sporozoites.
To test the ability of the FCR3Δvar2csa parasites to be recognized by plasma antibodies from malaria‐exposed pregnant women, flow cytometry using intact IE was performed (supplementary Fig S2 online). Whereas FCR3‐CSA IE are recognized by sera of malaria‐exposed multigravidae, FCR3‐CD36 and FCR3Δvar2csa IE are not (see the supplementary information online). In contrast, sera of malaria‐exposed men do not recognize any of these IE. These results confirm the previously published data that only var2csa‐expressing parasites are recognized in a parity‐dependent manner (Salanti et al, 2004).
In conclusion, we show that a single member of the var repertoire is required for binding to CSA in FCR3 parasites. Given that FCR3Δvar2csa disruptant mutants do not recover this binding phenotype, even after several rounds of panning selection, we conclude that no other parasite gene can compensate for the loss of function under the experimental CSA selection conditions of our work. Thus, our demonstration of the central role of var2CSA in CSA adhesion is important for the future design of a vaccine against the complications of malaria during pregnancy.
Parasites and cells. P. falciparum FCR3 clones were cultivated according to standard conditions (Trager & Jensen, 1976). Knob‐positive parasites were routinely selected by gelatin flotation using Plasmion (Fresenius Kabi, France). Saimiri brain microvasculature endothelial cell clone Sc1707 was cultured, as described earlier (Pouvelle et al, 1997).
Plasmids and transfection. Fragments corresponding to the DBL3‐X and DBL5‐ε domains of var2csa were PCR amplified from FCR3 genomic DNA using the primer combinations DBL3‐F/DBL3‐R and DBL5‐F/DBL5‐R (primer sequences are shown in supplementary Table S1 online). These PCR fragments were sequentially cloned into pHTK (Duraisingh et al, 2002) using the SacII/SpeI sites, as well as the EcoRI/AvrII sites, to derive pHTK‐var2csa.
Ring‐stage FCR3 parasites were transfected with 100 μg plasmid DNA and selected with 2.5 nM WR99210 (Jacobus Pharmaceutical Co. Inc., Princeton, NJ, USA) and 4 μM ganciclovir (Sigma, Saint Quentin, Fallavier, France), as described previously (Duraisingh et al, 2002).
Pulsed‐field gel electrophoresis and Southern blot. Genomic DNA was digested and size fractionated on 0.8% agarose gels. PFGE and Southern blot were performed, as described previously (Hernandez‐Rivas & Scherf, 1997). The chromosome‐12‐specific probe for clathrin heavy chain was obtained by PCR amplification using the primers CHC‐F and CHC‐R. The hDHFR probe was obtained by restriction of pHTK with BamHI and HindIII. Var2csa probes were used, as described in the Plasmids and transfection. Membranes were hybridized at high‐stringency conditions at 60°C overnight and washed twice with 0.2 × SSC and 0.1% SDS at 60°C for 30 min.
Northern blot analysis. Total RNA was prepared from synchronized parasite cultures approximately 10 and 30 h after invasion. RNA preparation, electrophoresis, membrane transfer and hybridization were carried out, as described previously (Kyes et al, 2000). Membranes were hybridized at high‐stringency conditions at 60°C overnight and washed twice with 0.5 × SSC and 0.1% SDS at 60°C for 30 min. Radiolabelled probes for FCR3 var2csa DBL3‐X or FCR3 var1csa DBL3‐γ, or a probe based on the var semiconserved exon II (varT11.1 gene, 7,930–9,147 base pairs; GenBank accession number U67959) were generated, as described previously (Gamain et al, 2005).
P. falciparum adhesion assays and pannings. Trophozoite‐stage IE were purified using gelatin flotation and selected for CSA binding on Sc1707 cells and on recombinant human thrombomodulin carrying CSA chains, as described previously (Pouvelle et al, 1997). For Sc1707, the number of IE bound per square millimetre was determined for four different fields and the mean±s.d. was calculated. Cytoadhesion assays on receptors immobilized on plastic Petri dishes were carried out, as described previously (Baruch et al, 1999; Buffet et al, 1999). The average number of adherent IE (±standard error of the mean (s.e.m.)) for four different fields was determined in three independent experiments.
Supplementary information is available at EMBO reports online (http://www.nature.com/embor/journal/vaop/ncurrent/extref/7400466‐s1.pdf).
This work was supported by grants from the Gates foundation and the network of excellence ‘BioMalPar’ funded by the European Union.
- Copyright © 2005 European Molecular Biology Organization