|Home | About | Journals | Submit | Contact Us | Français|
The var multicopy gene family encodes Plasmodium falciparum erythrocyte membrane protein 1 (PfEMP1) variant antigens, which, through their ability to adhere to a variety of host receptors, are thought to be important virulence factors. The predominant expression of a single cytoadherent PfEMP1 type on an infected red blood cell, and the switching between different PfEMP1 types to evade host protective antibody responses, are processes thought to be controlled at the transcriptional level. Contradictory data have been published on the timing of var gene transcription. Reverse transcription-polymerase chain reaction (RT-PCR) data suggested that transcription of the predominant var gene occurs in the later (pigmented trophozoite) stages, whereas Northern blot data indicated such transcripts only in early (ring) stages. We investigated this discrepancy by Northern blot, with probes covering a diverse var gene repertoire. We confirm that almost all var transcript types were detected only in ring stages. However, one type, the well-conserved varCSA transcript, was present constitutively in different laboratory parasites and does not appear to undergo antigenic variation. Although varCSA has been shown to encode a chondroitin sulphate A (CSA)-binding PfEMP1, we find that the presence of full-length varCSA transcripts does not correlate with the CSA-binding phenotype.
Plasmodium falciparum is responsible for nearly all malaria-specific mortality and for a high proportion of overall malaria morbidity. The particular virulence of this species is partly attributed to modifications to the host erythrocyte membrane during asexual infection. Parasite proteins inserted into the infected red blood cell membrane mediate adhesion to a variety of host receptors on vascular endothelium, uninfected red blood cells, placental syncytiotrophoblast cells, platelets and dendritic cells (reviewed by Kyes et al., 2001). The downstream effects of these interactions are thought to underlie much of the pathogenesis of severe disease. One major group of parasite proteins involved is a variant antigen family, collectively termed P. falciparum erythrocyte membrane protein 1 (PfEMP1) and encoded by the var multicopy gene family (Baruch et al., 1995; Smith et al., 1995; Su et al., 1995). PfEMP1 proteins are also targets of host protective antibodies (Bull et al., 1998), but antigenic variation between PfEMP1-types leads to evasion of the host immune response during chronic infection.
Placental malaria is a special case of infected red blood cell adhesion, causing maternal anaemia and low birth-weight in the first pregnancies of women who already have partial antimalarial immunity (Brabin, 1983). In primigravidae, infected red blood cells can concentrate in the placenta, adhering to syncytiotrophoblast cells via chondroitin sulphate A (CSA; Fried and Duffy, 1996) or hyaluronic acid (Beeson et al., 2000). In subsequent pregnancies, a relative resistance develops, possibly caused by specific antibodies recognizing a restricted repertoire of antigens on the surface of infected red blood cells in the placenta (Fried et al., 1998; Buffet et al., 1999; Ricke et al., 2000; Lekana Douki et al., 2002). This situation contrasts with that in infections of non-pregnant hosts, where the repertoire of variant antigens expressed is extremely large (e.g. Marsh et al., 1986; Aguiar et al., 1992). It has thus become an important goal to define the subset of variant antigens expressed in the placenta, as they may form the basis of a specific vaccine against placental malaria.
Although there is evidence for non-PfEMP1-mediated ring-stage parasite adhesion that correlates with CSA binding at the pigmented trophozoite stage (Pouvelle et al., 2000), the var gene family is the best studied variant antigen responsible for adhesive events. Each haploid parasite genome contains ≈60 var genes (Gardner et al., 2002), which are highly diverse in sequence when compared within a single genome and between isolates (Freitas-Junior et al., 2000; Taylor et al., 2000; Fowler et al., 2002). The highly variable 5′ exon (3.5–10 kbp exon 1) codes for the extracellular portion of the protein plus a transmembrane segment, and the semi-conserved 3′ exon (1–1.5kb exon 2) codes for a cytoplasmic region. The genes for two different CSA-binding PfEMP1 variants have been identified in IT lineage laboratory parasites (equivalent to FCR3, Robson et al., 1992; var-CS2, Reeder et al., 1999; and FCR3.varCSA, Buffet et al., 1999). Although var-CS2 appears to be unique to the IT/FCR3 lineage, FCR3.varCSA is well conserved between isolates (Rowe et al., 2002; Salanti et al., 2002) and is referred to here as varCSA. Its unusually high degree of conservation between isolates makes varCSA a strong candidate for an antiplacental malaria vaccine. However, transcripts for varCSA are found in both placental and non-placental isolates (Fried and Duffy, 2002), raising some doubt as to how strictly varCSA is limited to placental isolates.
Because of the importance of PfEMP1 in pathogenesis and immune evasion, much effort has been expended on determining the molecular mechanisms controlling var gene expression. However, investigations on the timing of transcription within the cell cycle and on the range of var genes transcribed within any one parasite population have led to contradictory results. Reverse transcription-polymerase chain reaction (RT-PCR) and nuclear run-on data (Chen et al., 1998a; Scherf et al., 1998) suggested that many different var gene variants are transcribed in early (10 h after invasion; ring) stages of the asexual life cycle, and exclusive transcription of the gene encoding the single cytoadherent variant occurs in later (24 h after invasion; pigmented trophozoite) stages. Our Northern blot data (Smith et al., 1998; Kyes et al., 2000) indicated that full-length transcripts encoding the cytoadherent variant are present only in ring (3–18 h after invasion) stages, and correlated well with the stage specificity of PfEMP1-mediated binding phenotypes, starting at about 16 h after invasion (Gardner et al., 1996).
We tried unsuccessfully to reproduce the RT-PCR results described above in 3D7, the laboratory parasite used for the Malaria Genome Sequencing Project, but recent developments have allowed us to reinterpret our data. We reported previously a comparison of 3D7 ring versus pigmented trophozoite var transcripts (Taylor et al., 2000), using RT-PCR to amplify variant-specific sequence tags representing a short fragment within the most conserved region of exon 1 [Duffy-binding-like domain alpha (DBLα); for a description, see Smith et al., 2000]. We found multiple var types at low frequency in both stages, but a much higher frequency of one type (3D7AFBR4; accession no. AF133860) was seen in pigmented trophozoites. This type did not correlate with the major full-length var transcript detected on Northern blots. We now know through data from the Malaria Genome Sequencing Project that 3D7AFBR4 happens to be the variant-specific tag for 3D7 varCSA, and also that the 3D7 varCSA is truncated, lacking exon 2 (3D7chr5var; Rowe et al., 2002). In this case, the predominance of a single transcript type detected by RT-PCR in pigmented trophozoites clearly cannot relate to PfEMP1 expression, as the 3D7 varCSA is a pseudogene.
We thus resolved to re-examine the timing of var transcription, paying particular attention to the varCSA gene.
To investigate the discrepancy between Northern blot and RT-PCR data regarding the timing of var gene transcription, we first re-examined the design of our published time course Northern blots (Kyes et al., 2000). These data compared equivalent amounts of total RNA, showing relative transcript levels at each time point. However, accumulated RNA levels increase at least fourfold from rings to pigmented trophozoite stages (Newbold et al., 1982; data not shown), so this approach may have given a misleading picture of absolute var transcript levels over time. Therefore, we repeated the time course analysis, but prepared RNA from known numbers of cells, at 4 h intervals, from a single culture of tightly synchronized, unselected A4 parasites (A4ICAM-U). Blots comparing total RNA from equivalent numbers of cells at each time point were hybridized at low stringency with the complex var exon 2 probe. Confirming our previous reports (Kyes et al., 2000), full-length var transcripts were only detected in ring stages (8–20 h after invasion), not in pigmented trophozoites (after 20 h; Fig. 1A). In stages 24 h after invasion and later, an ≈3 kb transcript is presumed to represent transcripts of Pf60 (Bonnefoy et al., 1997), another multigene family with high similarity to var exon2, or aberrant exon 2 transcripts (Su et al., 1995).
The discrepancy in observed timing of var gene transcription was not resolved by comparing absolute var transcript levels (per equivalent number of cells) with relative levels (per equivalent amounts of RNA) on Northern blots, so we tested whether the complex var exon 2 probe detects all var genes. In analysing var gene transcripts in the IT lineage, we had already noticed that R29var1 (Rowe et al., 1997) did not hybridize with a ‘generic’ probe, varC, which was based on the exon 2 sequence and presumed to be semi-conserved between all var gene types (Rubio et al., 1996). Indeed, sequence analysis showed that the exon 2 of R29var1 cannot be amplified with the varC primers. Therefore, for a more comprehensive detection of var transcripts, we had devised the ‘complex var exon 2’ probe, by mixing varC with a cloned R29var1 exon 2 fragment. However, of the published IT/FCR3 lineage var genes, the exon 2 of FCR3.varCSA is also unique. To be certain that we could detect this particular var transcript, and any other possible var genes with related exon 2 types, we designed primers to amplify FCR3.varCSA exon 2 and hybridized the labelled fragment to the A4ICAM-U time course blots (Fig. 1B). Two distinct transcript patterns were detected. In ring-stage parasites (8–16 h after invasion), barely detectable transcripts of 9–13 kb were present, similar to those detected by the complex var exon 2 probe. In late rings and pigmented trophozoites, however, a single large transcript of ≈15 kb was present at high levels (from 20 h after invasion). The same transcript was present only at very low levels during ring stages. Again, the 3 kb transcript in pigmented trophozoites probably represents Pf60/aberrant exon 2. Thus, the conflict in published results on var transcriptional timing could lie in the specific type of var gene being investigated, as distinguished by different exon 2 probes.
We also characterized the 15 kb transcript detected by the varCSA exon 2 probe. It corresponds to the relatively large size of FCR3.varCSA (Buffet et al., 1999), and its identity was confirmed by high-stringency hybridization of the same blot with various probes specific for FCR3.varCSA exon 1 (Fig. 1C; DBL7 is representative; DBL4 + DBL5 probe was also tested; data not shown). These probes only detected the 15 kb band, and it was present in all asexual stages. The pigmented trophozoite stages appear to have much higher levels of transcripts per infected red cell than ring stages. The 3 kb (Pf60/aberrant exon 2) band is not detected with this exon 1 probe, as expected.
The varCSA gene transcript profile that we observed in this time course could be a unique feature of the A4ICAM-U parasite, so we compared this parasite with the genetically related FCR3CSA-U and with unrelated TM180. On blots loaded with equivalent levels of RNA per lane and probed at high stringency with varCSA DBL4 + DBL5, a similar pattern is found in A4ICAM-U and FCR3CSA-U parasites. The 15 kb varCSA transcript is present at low levels in early to mid-rings and at much higher levels in late rings/pigmented trophozoites (Fig. 1E and F). With slightly less stringent washing conditions (60°C instead of 65°C to accommodate differences in levels of sequence similarity between TM180 and FCR3 varCSA genes), TM180 parasites have two transcripts, a 12 kb band detected only in rings and the 15 kb band detected in all stages (Fig. 1G). The 12 kb transcripts are ring stage specific, detected with the complex var exon 2 probe from 6 to 18 h after invasion, but not from 24 to 48 h after invasion (Fig. 1H).
We compared two further distinct parasite genotypes, 3D7 and HB3, for their var gene transcription pattern. Mid- to late-ring and pigmented trophozoite stage RNA samples were loaded with equivalent amounts of total RNA per lane, as well as equivalent cell numbers per lane (data not shown). As before, the complex var exon 2 probe detects transcripts of varying sizes, restricted to ring stages in all parasites (Fig. 2A). This probe detects no full-length var transcripts in pigmented trophozoite stages, regardless of how much RNA was loaded (data not shown). The varCSA exon 2 probe (Fig. 2B) detects multiple var transcripts in ring stages for the three unrelated parasites, and a 15 kb transcript in A4ICAM-U and HB3, in both rings and pigmented trophozoites. Hybridization with the varCSA DBL7 probe detects only a 15 kb band in A4ICAM-U and HB3, matching the timing of the 15 kb band detected by the varCSA exon 2 probe (Fig. 2C). No full-length transcript for the varCSA gene was detected with either probe in 3D7, at either stage, as expected because of the truncation of this gene in 3D7 (Rowe et al., 2002).
The A4ICAM-U, HB3 and TM180 parasites had never been selected for CSA binding. The A4 clone had originally been selected for intercellular adhesion molecule-1 (ICAM-1) binding, but this particular culture had not been selected for any binding phenotype for at least 20 cycles before RNA extraction. We therefore compared RNA from related parasites before and after selection for specific phenotypes. We compared A4 with low (A4ICAM-U) and high (A4ICAM-1C1) ICAM-1 binding phenotypes, and FCR3CSA before (FCR3CSA-U) and after (FCR3CSA+) selection on CSA. The complex var exon 2 probe (Fig. 3A) detected two bands in A4ICAM-U and A4ICAM-IC1. The large (13 kb) band corresponds to previously reported ICAM-1-binding PfEMP1 transcript A4var (Smith et al., 1998). A diffuse band at around 9 kb probably represents multiple var transcripts of a similar size range. The 13 kb band was enriched by ≈13% (relative to the diffuse band) in the higher ICAM-1-binding parasite, A4ICAM-IC1 (as judged by densitometry), corresponding well with the modest increase in ICAM-1 binding. The same probe detected two bands in FCR3CSA-U (9.5 kb and barely visible 14 kb), and two bands (equal intensities of 9.5 kb and 14 kb) in FCR3CSA+. The varCSA DBL7 probe (Fig. 3B) detects a 15 kb transcript in both rings and pigmented trophozoites of all parasites, regardless of the binding phenotype.
All parasites tested, except for 3D7, which has a truncated version of the gene, seemed to contain the varCSA-type transcripts either in all asexual stages or mainly in pigmented trophozoite stages. It remained possible that all parasites bind CSA to some extent, thus explaining the apparent universal presence of the conserved varCSA transcripts. We performed standard infected red blood cell binding assays to ICAM-1, CSA and 1% BSA spotted on plastic dishes (Fig. 3C). Several parasites bound to ICAM-1 significantly [>1000 infected RBC mm−2; TM180 (not shown), A4ICAM-U, A4ICAM-1C1]. Three had low binding to ICAM-1 [50–600 infected RBC mm−2; 3D7 and HB3 (not shown); FCR3CSA-U]. As described above, FCR3CSA-U bound to both ICAM-1 (low) and CSA (>1000 mm−2), but FCR3CSA+ bound only to CSA (>8000 iRBC mm−2). No parasite other than FCR3CSA (unselected and CSA selected) bound significantly to CSA. These results confirm that FCR3.varCSA transcript levels are not upregulated in parasites selected for enhanced CSA binding, and show that parasites containing full-length varCSA transcripts do not necessarily bind to CSA.
Our study of var gene transcripts in laboratory parasites was designed to resolve an apparent discrepancy between two methods of analysing RNA. Our Northern blot data had suggested that full-length var transcripts are present only in ring stages (Kyes et al., 2000), and that these transcripts correlate with the PfEMP1 type expressed (Smith et al., 1998). RT-PCR data had suggested that many var types are transcribed in ring stages, and that only the major PfEMP1-encoding type is transcribed in pigmented trophozoites (Chen et al., 1998b; Scherf et al., 1998). Recent investigations, which focus on single-cell RT-PCR with increased sensitivity of var gene detection, have suggested that many var transcripts can still be detected in pigmented trophozoite stages, and that truly exclusive var gene transcription never occurs (Duffy et al., 2002). Although it is powerful for determining which var transcripts are present, we felt that RT-PCR is too sensitive for making generalizations about transcription timing. For example, sporozoite stage-specific transcripts can be detected by RT-PCR in asexual stages (Fidock et al., 1994; Chen et al., 1998b; Scherf et al., 1998), suggesting that either stage specificity is not controlled at the transcription level or RT-PCR is too sensitive to distinguish a low background level from full specific activation. Critically, our own RT-PCR experiments on the naturally occurring varCSA ‘knock-out’, 3D7, showed that the major var transcript detected in pigmented trophozoites is the varCSA pseudogene (Taylor et al., 2000). Therefore, until we understand more about how to interpret RT-PCR data, Northern blot analysis paired with phenotype data is essential for unravelling transcriptional control processes. We have concentrated here on verifying our own Northern blot results.
We focused on two possible alternative interpretations of our previously published data: that the RNA sample amounts for pigmented trophozoites were insufficient to detect a low relative level of var transcripts, or that the probe was not detecting all var transcripts. We began with testing the timing of steady-state full-length var transcripts within asexual stages, in phenotypically unselected A4 parasites. Time course Northern blots comparing either relative transcript levels (in equivalent amounts of total RNA) or absolute levels (in RNA from equivalent numbers of cells) were similar for the majority of var transcript types. However, we did find a bias in our ‘generic’ var probe. The use of a new type of var probe, varCSA exon 2, led to the observation of more var transcript types and two different steady-state transcript patterns. One pattern, identical to our previous findings, is exhibited by multiple var transcripts of varied sizes, detected only in rings. The other pattern is represented by a unique 15 kb transcript, detected either in all stages or predominantly in pigmented trophozoites. We confirmed, by hybridization with gene-specific probes at high stringency, that the unique 15 kb var transcript is the highly conserved varCSA gene. Thus, in unselected A4 parasites, there are two steady-state var transcript patterns, and the varCSA gene pattern is different from that for many other var genes.
As this result could be explained by some unique feature of the A4 parasite, and as the varCSA gene is highly conserved between isolates (Rowe et al., 2002; Salanti et al., 2002), we investigated whether the varCSA transcripts are present in both ring and pigmented trophozoite stages in other genetically related and unrelated laboratory parasites. We compared equal amounts of total RNA in full 48 h time courses and showed that the timing of varCSA transcript detection in A4 and related FCR3CSA is similar (low levels in early rings, high levels at late rings and in pigmented trophozoites). In the unrelated TM180 parasite, varCSA transcripts are present at similar levels in both rings and pigmented trophozoites. Two-point time courses showed that 3D7 does not transcribe detectable levels of a full-length version of the pseudogene, 3D7chr5var, and that, in HB3, similar levels of varCSA are detected in rings and pigmented trophozoites. This demonstrates that the varCSA gene could have two slightly different steady-state transcript patterns, depending on the parasite lineage, with transcripts detectable from early ring stage onwards or mainly in pigmented trophozoites.
Most strikingly, the CSA-binding phenotype did not correlate with the presence of full-length varCSA transcripts. Only the FCR3CSA-U and FCR3CSA+ infected red blood cells bound to CSA. In these two parasite lines, a sixfold increase in CSA binding did not correlate with any change in the level of the varCSA transcript. Additionally, the varCSA transcripts were detected in A4, TM180 and HB3, which did not bind to CSA. Therefore, all parasites tested with an intact varCSA gene contain full-length varCSA transcripts at some point in the 48 h life cycle, regardless of binding phenotype.
One var transcript appears to correlate well with CSA binding in FCR3CSA-U and FCR3CSA+. Our starting culture of FCR3CSA-U bound to both ICAM-1 and CSA and contained at least three var transcripts in ring stages, of 9.5 kb, 14 kb and 15 kb. After CSA selection of FCR3CSA-U, the enrichment of the 14 kb var transcript correlates with an increase in CSA binding and a decrease in ICAM-1 binding. Combined with our data that all parasites can contain the 15 kb varCSA transcripts regardless of CSA-binding phenotype, this suggests that the var transcript unique to FCR3CSA+ is a 14 kb band. We are currently characterizing this transcript. In retrospect, our findings reconcile data from Duffy et al. (2002), which suggested the presence of varCSA transcripts in both a CSA-selected IT/FCR3 lineage parasite and its non-CSA-binding parent, and indicated a novel var transcript type in the CSA-selected parasites (Duffy et al., 2002). The novel var transcript agrees in relative size (and stage specificity) with the ‘14 kb’ band that we identified in FCR3.CSA CSA-selected parasites, possibly representing the same var gene.
The different pattern of steady-state transcripts for the 15 kb varCSA gene suggests that it is unique in its regulation compared with the other var genes detected by the complex var exon 2 probe, and is in a class of its own for studying rules of var gene transcriptional regulation and switching. This is supported by the difference in its upstream regulatory sequence (Vazquez-Macias et al., 2002) compared with other var genes (Voss et al., 2000).
Preliminary Northern blot data (not shown) suggest that the full-length conserved varCSA gene is transcribed in ring-stage field isolates from children, and RT-PCR detects these transcripts in non-pregnant donors (Fried and Duffy, 2002). VarCSA transcripts are therefore not strictly a characteristic of isolates from pregnant women. This does not rule out the possibility that the varCSA-encoded PfEMP1 is directly involved in CSA binding, but it does strongly suggest that the varCSA-type gene will be detected by RT-PCR in many isolates, whatever the stage and phenotype. In theory, the ring stages will have a higher relative level of the relevant PfEMP1-encoding var transcripts. In practice, pigmented trophozoites yield more RNA, so both stages should be valid for investigating repertoires of var transcription by RT-PCR. However, our data have important implications for the interpretation of field study data. The most conservative view cautions against correlating the presence of varCSA-type transcripts with CSA binding.
Most laboratory parasites were from local cryopreserved stocks. All parasite genotypes were confirmed by PCR. The A4 parasites are genetically identical to FCR3 (Robson et al., 1992). A4 was originally derived from ITO4 selected for binding to ICAM-1 (Roberts et al., 1992). The A4 parasites used in these experiments termed A4ICAM (unselected) or A4ICAM-U had not been selected for any binding phenotype for over 20 cycles in culture. A4ICAM-1C1 was chosen as a control for high ICAM-1 binding; it had been derived from a high ICAM-1 binding culture of A4 by limiting-dilution cloning. The parasite FCR3CSA was obtained from the Malaria Research and Reference Reagent Resource Center (MR4; depositor A. Scherf) as a positive control for a CSA-binding parasite, with FCR3.varCSA transcription. This frozen parasite stock was thawed, cultured for several weeks, then cryopreserved for storage. After thawing again, it was cultured for three cycles and analysed for RNA and binding phenotype. These parasites, after a total of two thaws and 22 cycles in culture, bound not only to CSA but also to ICAM-1. At 28 cycles (from the MR4 source), the parasites were panned once on CSA, cultured for seven cycles and analysed again for RNA and binding phenotype (a total of 35 cycles in culture with one CSA selection; referred to as FCR3CSA+). A parallel unselected culture was compared (at 35 cycles; FCR3CSA-U). Panning on CSA bound to plastic (Sigma C9819, 1 mg ml−1 in phosphate-buffered saline; 1× PBS: 0.01 M Na-phosphate, 0.0027 M KCl, 0.138 M NaCl, pH 7.4), was performed as described by Roberts et al. (1992). TM180, HB3 and 3D7 are unrelated, are not related to IT/FCR3 and had all been grown for at least 10 cycles with no particular phenotype selection.
Parasites were cultured and sorbitol synchronized with standard techniques (Trager and Jensen, 1976; Lambros and Vanderberg, 1979). All cultures were mycoplasma negative (ATCC mycoplasma PCR test kit, according to instructions).
Parasites were sorbitol synchronized then cultured for at least 2 h before harvesting the first ring-stage RNA samples. Cultures were at 3–10% parasitaemia, and 200–500 μl of packed infected red blood cells were processed for each RNA sample. Cells were spun directly from warm culture medium, the supernatant removed, and the appropriate volume of TRIzol (Invitrogen) was added. Samples were processed as described by Kyes et al. (2000).
For the A4ICAM-U time course, samples were taken every 4 h, from the same culture at 9% parasitaemia. The samples were resuspended in equivalent volumes of formamide, and roughly 2 × 107 cell equivalents of total RNA were loaded per lane on the agarose gels in Fig. 1A–D.
For the comparison of ring (16 h after invasion) and pigmented trophozoite (30–36 h after invasion) RNA from different parasite lineages, samples were first electrophoresed on a small 1% agarose gel [containing 1× TBE (0.089 M Tris, 0.089 M boric acid, 2 mM EDTA) and 5 mM guanidine thiocyanate (Goda and Minton, 1995)] to compare relative RNA content, then on 0.8% agarose–1× TBE–5 mM GSCN gels for blotting. Aliquots of 3–6 μg of total RNA were loaded per lane and transferred to Hybond-N+ (Amersham).
Standard PCR conditions for amplification from P. falciparum genomic DNA were used to produce probe fragments. Primers were varC as reported by Rubio et al. (1996), varCSA exon 2 (spanning amino acids 3313–3536, referring to map in Buffet et al., 1999) and varCSA individual DBL fragments [DBL4, DBL5 and DBL7, as reported by Buffet et al. (1999)]. Perkin-Elmer Amplitaq DNA polymerase was used with manufacturer’s buffer, 2 mM MgCl2. Reactions were cycled: 95°C for 3 min, followed by 30 cycles of 94°C for 30 s, 50°C for 30 s, 65°C for 1 min. Template for all amplification was A4 genomic DNA. For the varC primers, this results in a mix of var exon 2 fragments. All double-stranded DNA fragments were purified on Qiagen PCR columns, according to the manufacturer’s instructions. The R29var1 exon 2 plasmid was derived by first cloning a larger PCR fragment amplified from A4 genomic DNA, then subcloning a DraI fragment. Its identity to R29var1 exon 2 was confirmed by hybridization to Northern blots containing RNA from R29var1-transcribing parasites and by restriction mapping, and the sequence (GenBank accession no. AJ535777) matches the partial exon 2 sequence (reported previously, accession no. Y13403). The complex var exon 2 probe was a mix of varC and plasmid R29var1 exon 2. Aliquots of 20–40 ng of DNA were α-32P-dATP labelled (Megaprime; Amersham, according to the manufacturer’s instructions).
Prehybridization (>1 h) and hybridizations (overnight) were performed in 7% SDS–0.5 M NaPO4–2mM EDTA, pH 7.2 (Church and Gilbert, 1984) Low-stringency hybridizations for varC and exon 2 probes were at 50°C; high stringency for varCSA DBL probes was at 60°C. Washes were at 5°C more than the hybridization temperature, in 0.5× SSC 0.1% SDS, except where noted otherwise. Exposures to autoradiography film ranged from 1 to 7 days. Band sizes given are approximate and are only indicated to identify bands according to relative size. Absolute size is difficult to determine, as most var transcripts are well above the range of the molecular size markers. geldoc software variously puts the varCSA transcript size at 14.3–15.2 kb, depending on the gel.
Densitometry was performed using Bio-Rad geldoc software.
All laboratory parasites were cryopreserved, thawed, then analysed for binding phenotype according to methods described by Roberts et al. (1992). CSA (in 1× PBS, 10 μg ml−1 and 100 μg ml−1; Sigma C8529), ICAM-1-FC (40 μg ml−1; Berendt et al., 1992) and 1% bovine serum albumin (BSA) were placed, 3 μl per spot on a Petri dish, incubated at 4°C overnight, then blocked with 1% BSA in 1× PBS for 1 h at 37°C. Parasites were adjusted to 5–6% parasitaemia, resuspended in binding medium (pH 7), then incubated for 1 h on the Petri dishes (with gentle mixing every 10 min). Non-adherent cells were removed by gentle washing with 1× PBS; plates were fixed in 2% glutaraldehyde, stained with Giemsa, then air dried. Parasite binding levels were calculated as numbers of infected red blood cells mm−2, with duplicate plates in each experiment and triplicate spots per plate. BSA binding levels were subtracted as background for each spot. RNA samples were collected at both ring (16–20 h before assay) and pigmented trophozoite stages (at time of binding assay). Binding assays were repeated several times in subsequent cycles, and the data shown are representative.
The following reagent was obtained through the Malaria Research and Reference Reagent Resource Center, Division of Microbiology and Infectious Disease, NIAID, NIH: FCR3.CSA, from A. Scherf. This work was funded by The Wellcome Trust [J.A.R. career development fellowship (055167) and senior research fellowship, and C.I.N. Programme grant].