Plasmodium falciparum: Cloned and Expressed CIDR Domains of PfEMP1 Bind to Chondroitin Sulfate A

Experimental Parasitology 95, 113–121 (2000) doi:10.1006/expr.2000.4512, available online at http://www.idealibrary.com on

Plasmodium falciparum: Cloned and Expressed CIDR Domains of PfEMP1 Bind to Chondroitin Sulfate A

Roland Degen, Niklaus Weiss, and Hans-Peter Beck Swiss Tropical Institute, CH 4002 Basel, Switzerland

Degen, R., Weiss, N., and Beck, H.-P. 2000. Plasmodium falciparum: Cloned and expressed CIDR domains of PfEMP1 bind to chondroitin sulfate A. Experimental Parasitology 95, 113–121. Adherence of erythrocytes infected with mature asexual Plasmodium falciparum parasites (iRBC) to microvascular endothelial cells contributes to the pathology of P. falciparum malaria. It has been shown that the variant P. falciparum erythrocyte membrane protein 1 (PfEMP1) confers adhesion to a wide range of cell surface receptors. Previously, the cysteine-rich interdomain region (CIDR) of PfEMP1 has been identified as binding site to CD36. We provide evidence that the same region can also mediate binding to chondroitin sulfate A (CSA). CIDR domains of two different parasite strains were expressed in Escherichia coli as a 6xHis-tagged protein. Purified recombinant protein bound to Chinese hamster ovary (CHO) cells which naturally express chondroitin sulfate A. Treatment of wild-type CHO cells with chondroitinase ABC reduced binding up to 94.4%. Competitive binding using soluble CSA inhibited binding to CHO cells by up to 100% at 2 mg/ml and by 62.4% at 0.5 mg/ml, whereas 1 mg/ml heparan sulfate had only a little effect (18.1%). In contrast, a recombinant 6xHis-tagged DBL1 domain showed no binding to wild-type CHO cells. Such an approach of analyzing various domains of PfEMP1 as recombinant proteins may elucidate their functions and may lead to novel anti-adherence therapeutics, especially for maternal malaria infections. q 2000 Academic Press Index Descriptors and Abbreviations: Plasmodium falciparum; cytoadherence; CHO, Chinese hamster ovary; CIDR, cysteine-rich interdomain region; CRM, cysteine-rich motif; CSA, chondroitin sulfate A; CSase ABC, chondroitinase ABC (EC 4.2.2.4.); DBL1 domain, Duffy binding like domain 1; His, histidine; FACS, fluorescence-activated cell scan; FITC, fluorescein isothiocyanate; ICAM1, intercellular adhesion molecule 1; iRBC, infected red blood cell; PECAM1, platelet/ endothelial cell adhesion molecule 1; PfEMP1, Plasmodium falciparum erythrocyte membrane protein 1; VCAM1, vascular cell adhesion molecule 1.

0014-4894/00 $35.00 Copyright q 2000 by Academic Press All rights of reproduction in any form reserved.

INTRODUCTION

Human erythrocytes infected with mature stages of Plasmodium falciparum (iRBC) adhere to postcapillary venular endothelium cells. This leads to sequestration of iRBCs and is believed to be an important contributor to the pathogenesis of malaria. While sequestration of iRBCs within the microvasculature might prevent their splenic clearance, the occlusion of these capillaries is thought to be correlated with the development of severe malaria syndromes, including cerebral malaria (Miller et al. 1994). P. falciparum-infected erythrocytes have been shown to bind to epithelium cells using a wide range of cell surface receptors such as thrombospondin (Roberts et al. 1985), CD36 (Barnwell et al. 1985), ICAM1 (CD54) (Berendt et al. 1989), VCAM1 (CD 106), and E-selectin (ELAM1 or CD62E) (Ockenhouse et al. 1992), chondroitin sulfate A (CSA) (Rogerson et al. 1995), PECAM1 (CD31) (Treutiger et al. 1997), and the integrin avb3 (Siano et al. 1998). This binding is believed to be conferred through the P. falciparum erythrocyte membrane antigen 1 (PfEMP1) and has been demonstrated for binding to CD36 and CSA. Being exposed to the immune system, PfEMP1 exhibits antigenic variation and is believed to be responsible for the chronicity of P. falciparum infections. PfEMP1 is encoded by a large multigene family, termed var-genes (Baruch et al. 1995; Smith et al. 1995; Su et al. 1995) with approximately 50 genes distributed throughout all 14 chromosomes in the haploid genome of P. falciparum (Su et al. 1995; Rubio et al. 1996). var-genes are located

113

114 on all chromosomes in subtelomeric regions and on four chromosomes internally (Thompson et al., 1997), with only one var-gene expressed at a time in mature parasite forms (Rowe et al. 1997; Scherf et al. 1998; Chen et al. 1998). Antigenic switching has been estimated to occur at a rate of approximately 2% per generation in vivo (Roberts et al. 1992). var-genes are highly polymorphic and organized in two exons separated by a 1-kb intron down-stream of the transmembrane domain with exon 2 coding for a conserved acidic terminal sequence. The majority of the molecule, encoded by exon 1, is exposed on the iRBC’s surface in the electrodense protrusions of knobs. Exon 1 codes for 2-5 Duffybinding-like (DBL) domains and a cysteine-rich interdomain region (CIDR) interspersed between DBL1 and DBL2 (Baruch et al. 1995; Smith et al. 1995; Su et al. 1995). Apart from short conserved sequence motifs in the DBL1 domain and a conserved cysteine-rich motif in the CIDR domain, exon 1 displays extreme sequence diversity and this DBL1– CIDR organization has been proposed to form a conserved head structure (Su et al. 1995). To date, it has been shown in vitro that the CIDR domain mediates binding to CD36 (Baruch et al. 1997; Smith et al. 1998), whereas the DBL1 domains of two different PfEMP1 molecules have been shown to mediate rosetting via CR1 (Rowe et al. 1997) or via HPS (Chen et al. 1998). More recently, the adhesion of iRBCs to CSA has been identified to be mediated by the expression of a specific PfEMP1 variant and the binding was significantly inhibited by antisera against the DBL3 and the CIDR domain of that variant (Reeder et al. 1999). Using CHO cell expression Buffet et al. (1999) showed that a recombinant DBL3 domain could confer such binding. CSA as a possible receptor for iRBCs was first described by Rogerson et al. (1995) and Robert et al. (1995). Since CSA is expressed on the surface of syncytiotrophoblasts in placenta (Gysin et al. 1997), it has been speculated that despite frequent previous exposure primigravidae mothers may become highly susceptible to P. falciparum infections by providing a placental substrate, which selects for CSAbinding parasites, a subgroup which these mothers might not have encountered yet, leading to placental malaria (Fried and Duffy 1996; Maubert et al. 1997). In this study we cloned several PfEMP1–CIDR domains as 6xHis-tagged proteins in Escherichia coli from synchronized, schizont-enriched (Matile and Pink 1990) P. falciparum cultures of strain ItG2.F6 and K1. We show that these recombinant domains bind to nontransfected Chinese hamster ovary cells which naturally express CSA. The binding was dose-dependent, inhibited by soluble CSA, and was

DEGEN, WEISS, AND BECK

abolished after treatment with chondroitinase ABC (CSase ABC).

MATERIAL AND METHODS

Parasite and CHO cell culture. P. falciparum parasites (strain ItG2.F6 and K1) were cultured with human A1 erythrocytes in continuous motion with automated medium exchange twice daily (Matile and Pink 1990). Nonsynchronized parasites were harvested at a density of 5–10% parasitemia. Binding to CD36 and ICAM1, but not to CSA, of iRBCs of strain ItG2.F6 was confirmed in binding assays as previously described (Udeinya et al. 1985; Willimann et al. 1995). CD36- and ICAM1-transfected CHO cells (Hasler et al. 1993) were obtained from the American Type Culture Collection (Nos. CRL-2092 and CRL-2093, respectively). Wild-type CHO cells were a kind gift of Dr. H. Matile (Hoffmann–LaRoche, Basel, CH). Transfected and nontransfected CHO cells were cultured in RPMI 1640 (Gibco BRL Life Technologies, Basel, CH), containing 25 mM sodium carbonate and supplemented with 10% fetal calf serum. Cloning of var-gene-domains and expression as 6xHis-tagged proteins in E. coli. P. falciparum polyA-RNA of strain ItG2.F6 was isolated with the Fast Track 2.0 kit (Invitrogen, Leek, NL) and a cDNA library was constructed (Invitrogen) into the pcDNA3.1 vector. The cDNA library and genomic DNA from strain K1 was PCR-screened with primers CIDRFW 58-CGGGATCCAAATGGAAATGTTAT TATG-38 carrying a BamHI site and CIDRREV 58-GGGGTACCTTGTAGTAATTTATCAATT-38 carrying a KpnI site. The purified PCR products were cloned in-frame into the BamHI–KpnI sites of pQE-30 (Qiagen, Basel, CH) with the 6xHis-tag at the 58 end of the cloned fragment. Similarly, a DBL1 domain was amplified from the cDNA library using the following primers: DBL1FW 58-CGGGATCCGCACGAAGTTTTGCAGATATTGG-38 and DBL1REV 58-GGGGTACCTTCGGCCCATTCCTCGAACCA-38 also carrying BamHI and KpnI sites, respectively. The PCR product was cloned and sequenced as described below. Proteins were expressed in E. coli, strain M15[pREP4] (Qiagen), at room temperature for 4 h after ITPG induction. Cloned fragments were PCR cycle sequenced using the two QIAexpress primers Type III (forward) and type IV (reverse)(Qiagen) with the DNA sequencing dye terminator kit (Perkin–Elmer, Rotkreuz, CH). Sequence analysis was carried out with an ABI Prism 310 Genetic Analyser (Perkin– Elmer) and ABI Prism software was used to proofread and translate the data into amino acid sequences. Extraction and purification of recombinant proteins. The CIDR domains were extracted using the QiaExpress system under native conditions and purified according to the manufacturer’s recommendation (Qiagen). In brief, bacteria were pelleted, lysed with 1 mg/ml lysozyme for 15 min on ice, sonicated, and spun again at 14,000g for 30 min. The lysate was mixed with 1/4 vol of 50% Ni–NTA slurry (Qiagen) and loaded in a gravity-flow column. The column was washed with wash buffer (50 mM NaH2PO4, 300 mM NaCl) containing increasing concentrations of imidazole (23 0 mM, 23 20 mM, 13 35 mM) and protein was eluted with wash buffer containing 250 mM imidazole. The purified CIDR domains were dialyzed overnight against 2 L PBS at 48C. Purity was visually tested on SDS–PAGE gels and the final

P. falciparum BINDING TO CHONDROITIN SULFATE A

concentration of protein was determined photometrically. The protein solutions were sterilized by filtration and aliquots were stored at 2208C. Antibodies and soluble receptors. Mouse mAb MEM112 (Anawa AG, Wangen, CH) recognizes human ICAM1, mouse mAb OKM*5 (Ortho Diagnostic System, Raritan, NJ, USA) recognizes human CD36, and mouse mAb anti-6xHistidine clone AD 1.1.10 (R&D Systems, Minneapolis, MN, U.S.A.) recognizes the 6xHis-tag. HRP-conjugated rabbit anti-mouse total IgG (DAKO corporation, Zug, CH) were used as secondary antibodies in Western blots and FITC-conjugated goat anti-mouse IgG F(ab8)2 (Cappel, Durham, NC, USA) in FACS analyses. CSA (sodium salt from bovine trachea) and HPS (sodium salt from bovine kidney) (both Sigma, Buchs, CH) were resuspended in PBS at a concentration of 2 mg/ml, aliquoted, and stored at 2208C. Chondroitinase ABC (from Proteus vulgaris) was obtained commercially (Sigma). Protein-binding assay to CHO cells and FACScan analysis. Transfected and nontransfected CHO cells were harvested with PBS/ 0.5 mM EDTA immediately prior to the assay to exclude possible effects on trypsin-sensitive receptors, and were resuspended in PBS/ 1% BSA to a cell density of 250,000 cells/FACS-vial (Falcon, Oxnard, CA, U.S.A.). Unspecific binding of cells was blocked by addition of 250 ml PBS/5% dry milk and incubated on ice for 1 h. Cells were spun at 300g for 3 min and the pellet was resuspended in 2 ml PBS/

115 1% BSA, pH adjusted to 7.0. The pH of PBS/1% BSA was adjusted to 7.0, since previous experiments showed no difference in binding at pH 6.8 and 7.3, but at pH 6.3 binding was decreased (data not shown) Subsequently, cells were pelleted again and resuspended in 100 ml PBS/1% BSA containing the protein (2–20 mg/ml) and incubated on ice for 1 h. Cells were washed twice in PBS/1% BSA as above, and 50 ml of the first antibody was added (a-human-ICAM1 and a-humanCD36 diluted 1:1000, a-His-tag diluted 1:100 in PBS/1% BSA) and incubated on ice for 1 h. Cells were washed again as above and the second antibody was added (50 ml a-mouse FITC-conjugated diluted 1:1000) and incubated for 1 h on ice. After two final washes, cells were resuspended in PBS and analyzed by FACS (Becton–Dickinson, San Jose, CA, U.S.A.). Ten thousand cells were analyzed in each experiment by laser agitation at 488 nm. The fluorescence emission intensity of FITC (520 6 30 nm) was measured for each cell. Two negative controls (without protein and without the first antibody) and two positive controls (mAb against human-CD36 and human-ICAM1) were analyzed in parallel. Chondroitinase ABC treatment. Naturally expressed CSA on the surface of CHO cells was removed with CSase ABC (Hasler et al. 1993). Monolayers of CHO cells were washed with PBS and incubated with 0.5 units CSase ABC/ml in PBS for 1 h at 378C. After treatment, cells were a round shape but not completely detached. PBS/CSase was aspirated and the cells were detached as described above.

FIG. 1. Sequence alignment of six cloned and sequenced CIDR domains, the rC1-2 CIDR domain described by Baruch et al. (1997), and the partial varCSA CIDR-domain (amino acids 661–729 in the original sequence) described by Reeder et al. (1999). Primer binding sites are boxed in rC1-2 and the tested sequences. The consensus sequence is given when $75% of sequences were identical for the respective amino acid. Asterisks below the consensus sequence indicate 100% homology. Letters in bold italic in the varCSA sequence indicate homologies to the CIDR domains described in this study.

116

DEGEN, WEISS, AND BECK

RESULTS Amplification of cDNA from ItG2.F6 for CIDR domains yielded two distinct bands of approximately 600 and 520 bp, and amplification from K1 genomic DNA yielded one band of approximately 510 bp. These bands were gel-eluted, purified, and subsequently cloned. Forty-two clones were picked, sequenced, and identified by sequence comparison as six different CIDR domains. All CIDR domains included the predicted CD36-binding domain (Baruch et al. 1997) and approximately 120 bp of up-stream sequence. The cysteine-rich motif (CRM) and four additional regions of homology to the CIDR domain described by Baruch et al. (1997) were present in all sequences. A fifth region of homology corresponded to the reverse primer binding site. Four different CIDR domains were identified from ItG2.F6, termed CIDR C1, CIDR C9, CIDR C10, and CIDR C32. The first 204 nucleotides of CIDR C32 were identical to CIDR C10 and were followed by 357 bp identical to CIDR C1 (Fig. 1). Another CIDR domain from ItG2.F6 was identified (CIDR C17) but contained 120 bp down-stream of the 58 primer a stop-codon (TAG) instead of a conserved tryptophan residue (TGG). It remains unclear whether this reflects a pseudogene sequence or a PCR artifact, although 14 independently picked clones showed this internal stop codon. Two additional CIDR domains were amplified and cloned from K1 genomic DNA (CIDR K39 and CIDR K42). All CIDR domains and a DBL1 control were expressed in E. coli and purified under native conditions, except CIDR C9, which was extracted under denaturing conditions (8 M urea) and was subsequently renatured. CIDR C17 was not expressed. Expressed fragments were tested on Western blots

and were all recognized by an HRP-conjugated a-His-tagantibody. Initially, FACS analysis showed that all six recombinant CIDR domains bound to CHO–CD36 cells in a dose-dependent manner (Table I, Fig. 2). Binding affinity varied, with CIDR C1 binding being the strongest (93 stronger than aHis-antibody without protein) and CIDR K39 the weakest (2.53 control). Subsequently, it was shown that all CIDR domains also bound to nontransfected CHO cells (Table I) and CHO–ICAM1 cells (data not shown), suggesting that binding was mediated through CSA. At a concentration of 20 mg/ml, all CIDR domains bound to CHO–CD36 cells but at a concentration of 2 mg/ml binding was only detectable by FACScan with CIDR C1 and CIDR C32 (Table I). The recombinant DBL1 domain with His-tag did not bind to any of the CHO cells (data not shown), excluding the possibility that binding was mediated through the His-tag. To confirm binding to CSA, CHO and CHO–CD36 cells were treated with CSase ABC, and binding of 0.2 mg CIDR C1 to treated cells was compared with binding to untreated cells. On both cell types binding was decreased after CSase ABC treatment by at least 90%. Increasing CSase ABC treatment (2 h) and increasing the concentration of CIDR domains (0.5 mg CIDR C1, 1 mg CIDR C10) resulted in a decrease of binding up to 94% when compared to untreated cells (Table II). There was no evidence of a difference in binding to CSase-ABC-treated CHO cells and CSase-treated CHO–CD36 cells indicating that CD36 was not involved in the binding of CIDR C1 (Table II, Fig. 3). In parallel, 0.7 mg of each of the six CIDR domains was tested against untreated CHO cells, CSase-treated CHO cells

TABLE I Binding of Different Recombinant His-Tagged CIDR Domains to CD36-Transfected CHO Cells and Nontransfected CHO Cells Measured by Cytofluorometry Amount of protein/tested cells 20 mg/ml; CHO–CD36 10 mg/ml; CHO–CD36 2 mg/ml; CHO–CD36 0 mg/ml; CHO–CD36 7 mg/ml; CHO 0 mg/ml; CHO

a-His-tag (negative control)

541

a-CD36 (positive control)

CIDR C1

CIDR C9

CIDR C10

CIDR C32

CIDR K39

CIDR K42

4571 4259 2128

2684 1482 724

1884 1813 829

2609 1838 1318

1300 826 616

1712 1613 858

1026

1563

806

995

842

911

4215

394

Note. The geometric mean fluorescence after staining with FITC-labeled a-His-tag antibody as measured directly by FACS analysis is depicted. As negative control a-His-tag antibody was incubated in the absence of protein; as positive control CD36-transfected CHO cells were incubated with OKM*5, a monoclonal antibody recognizing CD36. Both controls were visualized using a a-mouse IgG labeled with FITC.

117

P. falciparum BINDING TO CHONDROITIN SULFATE A

FIG. 3. Cytofluorogram of nontransfected CHO cells, untreated and treated with CSase ABC, and incubated with 5 mg/ml CIDR C1. Bound protein was detected using an a-His-tag monoclonal antibody followed by detection with FITC-labeled a-mouse IgG. As negative control, a-His-tag monoclonal antibody was incubated without prior incubation with protein. FIG. 2. Cytofluorogram of CHO cells transfected with CD36 and incubated with decreasing concentrations of CIDR C1 (20, 10, or 2 mg/ml). Bound protein was detected using an a-His-tag monoclonal antibody followed by detection with FITC-labeled a-mouse IgG. As negative control, a-His-tag monoclonal antibody was incubated without prior incubation with protein.

(1-h treatment), and untreated CHO cells with 200 mg soluble CSA as competitor. Binding was inhibited up to 100% (CIDR C10) by 200 mg soluble CSA, whereas CSase treatment resulted in 67 to 89% inhibition of binding. CIDR C32 was also tested against untreated CHO cells and with 50 mg CSA or 200 mg HPS as competitor. Competition with 50 mg CSA

TABLE II Comparison of Binding of Three Different CIDR Domains to CD36-Transfected CHO Cells and Nontransfected CHO Cells, either with or without Prior CSase ABC Treatment Tested cells CHO–CD36 CHO–CD36 CHO–CD36 CHO CHO CHO CHO CHO CHO CHO CHO CHO CHO

Treatment of cells None None CSase None None CSase None None CSase None CSase None CSase

(1 hr)

(1 hr)

(2 h) (2 h) (2 h)

Concentration/protein 0 mg/ml 2 mg/ml CIDR C1 2 mg/ml CIDR C1 0 mg/ml 2 mg/ml CIDR C1 2 mg/ml CIDR C1 0 mg/ml 5 mg/ml CIDR C1 5 mg/ml CIDR C1 10 mg/ml CIDR C10 10 mg/ml CIDR C10 10 mg/ml CIDR C32 10 mg/ml CIDR C32

Mean fluorescence

Percentage of binding

661 983 616 467 842 500 563 2768 687 1952 732 1691 765

0 100 213 0 100 8.8 0a 100a 5.6 100a 12.2 100a 17.9

Note. Binding is measured by cytofluorometry using a-His-tag antibody followed by incubation with FITC-labeled a-mouse IgG. The geometric mean fluorescence and remaining percentage of binding after treatment are depicted. The percentage of binding was calculated using the value from untreated cells as a maximum (100%) and the value of the negative control (a-His-tag only) as a minimum (0%). a Note that two independent experiments were carried out and the negative controls were used accordingly.

118 resulted in 62.4% inhibition of binding, whereas binding of the same domain was only affected a little (18.1% inhibition) by 100 mg HPS (Table III, Fig. 4).

DISCUSSION

Cellular adhesion plays a major role in P. falciparum infections. There are adhesive motifs in proteins which are known to participate in host cell invasion such as in region 2 of the circumsporozoite protein, in the thrombospondinrelated adhesive protein (TRAP/SSP2), or the erythrocyte binding antigen (EBA175) (Sim et al. 1994; Sinnis and Sim 1997). The interaction between P. falciparum-infected erythrocytes and cells of the microvascular endothelium, known as sequestration, is most important in the pathogenesis of P. falciparum. Identification of molecules involved in cytoadherence and their domains which mediate adherence of iRBC to different host receptors is an essential step in understanding the molecular pathogenesis of P. falciparum. In the long run, this might lead to the development of new and potent anti-adherence therapeutics. Recently, such domains have been identified in PfEMP1 by Baruch and colleagues (1997), Reeder et al. (1999), and

FIG. 4. Cytofluorogram of nontransfected CHO cells, competitively incubated with 7 mg/ml CIDR C32 and 200 mg soluble CSA. Binding was compared against incubation of CHO cells without addition of soluble CSA. Bound protein was detected using an a-Histag monoclonal antibody followed by detection with FITC-labeled amouse IgG. As negative control, a-His-tag monoclonal antibody was incubated without prior incubation with protein.

DEGEN, WEISS, AND BECK

Buffet et al. (1999) as ligands for CD36 and CSA, respectively. The CD36-binding domain, denoted rC1-2, corresponded to amino acids 576 to 755 in the MC PfEMP1 molecule (Baruch et al. 1995), whereas CSA binding was inhibited with antibodies directed against the var-CS2 domain between amino acids 404 and 1123 (Reeder et al. 1999). Buffet et al. (1999) have recently had individual domains of the FCR3.varCSA transfected into CHO cells and showed that the DBL3 domain is involved in binding to CSA (amino acids 1270–1577). In order to identify additional binding domains, we PCRamplified and cloned various var-gene fragments and expressed the corresponding peptides as His-tagged proteins. We were able to amplify var-gene fragments from both adherent parasites (ItG2.F6) and nonadherent parasites (K1). Since the cultures were not kept under selection prior to mRNA preparation, the identification of more than one transcript in ItG2.F6 was expected. Similarly, the lack of cytoadherence of K1 is not due to a lack of var-genes but due to a deletion in chromosome 9 (Kemp et al. 1992) preventing the formation of structures mediating adherence (A. Cowman, pers. comm.). On the other hand, the identification of only five different transcripts from ItG2.F6 and two gene fragments from K1 with our primers gives evidence to a large sequence diversity within the var-gene family. The CIDR domains identified in this study showed significant homologies to the PfEMP1 gene described by Baruch et al. (1997) and included the region of the CD36-binding domain of rC1-2. However, the domains described here were expressed as His-tagged proteins and showed no binding to CD36 on transfected CHO cells. For this, it is important to note that our forward primer used for amplification recognized a sequence up-stream of the binding site described by Baruch et al. (1997) and coded for the amino acid sequence KWKCYY (aa 1–6 in Fig. 1). Baruch and colleagues used a forward primer coding for KEDKIMSY (aa 33–40 in Fig. 1), a sequence which they found to be important for the correct folding of the protein in E. coli. Due to the observed sequence diversity, amino acids 31–39 in our CIDR domains (XGX (2–3)V(K/T/M)SY) were not identical to those described in rC1-2. Possible incorrect folding, despite the complete presence of necessary cysteines, might be responsible for the lack of CD36 binding observed in our study. Baruch and colleagues (1997) studied 13 further GST-fusion proteins corresponding to rC1-2 from eight different parasite isolates and tested their binding to CD36. Three fusion proteins bound strongly to CD36, 7 proteins bound weakly, and 3 CIDR domains did not bind to CD36. None of the CIDR domains, which weakly bound or failed to bind to CD36, was tested for binding to CSA. The presence of additional

119

P. falciparum BINDING TO CHONDROITIN SULFATE A

TABLE III Competitive Binding of Recombinant CIDR Domains and Soluble CSA to CHO Cells Measured by Cytofluorometry Using a-His-tag Antibody Followed by Incubation with FITC-Labeled a-Mouse IgG Assay condition None CSase 1 200 mg CSA 1 50 mg CSA 1 100 mg HPS

a-His-tag antibody 3.94 (0) nd nd nd nd

CIDR C1

ClDR C9

CIDR C10

geometric mean flourescence (percentage of remaining 10.26 (100) 15.63 (100) 8.06 (100) 5.98 (32.3) 5.22 (10.9) 5.20 (30.6) 5.26 (20.9) 4.75 (6.8) 3.81 (23.2)

CIDR C32 binding) 9.95 (100) 4.90 (16.0) 5.16 (20.3) 6.20 (37.6) 8.86 (81.9)

CIDR K39

CIDR K42

8.42 (100) 5.10 (25.9) 5.23 (28.8)

9.11 (100) 4.64 (13.5) 5.29 (26.1)

Note. The geometric mean fluorescence and remaining percentage of binding are depicted. The percentage of binding was calculated using the value from assays without CSase treatment or the addition of soluble CSA as a maximum (100%) and the value of the negative control (a-Histag only) as a minimum (0%). In all assays CIDR domains were used at a concentration of 7 mg/ml.

var-genes not conferring binding to the receptor selected for suggests that a larger population of var-genes might exist in a parasite culture. This could explain our finding of CSAbinding domains from cultured parasites which were tested for binding to CD36. Such findings demonstrate that caution must be used when interpreting data from in vivo binding assays. Apart from the cysteine-rich motif, Baruch and colleagues identified five additional regions of homology, which they suspected to be involved in CD36 binding (Fig. 1). All regions of homology were also present in our CIDR domains, except region 4, which was only found in CIDR K39 (Fig. 1). The lack of binding to CD36-transfected CHO cells of our CIDR domains questions the importance of these motifs in CD36 binding. However, sequence diversity in our CIDR domains may well have modified the binding characteristics by generating CSA-binding motifs, lacking in recombinant rC1-2 protein, which promoted binding to CHO–CD36 cells but not to CHO–ICAM1 or wild-type CHO cells (Baruch et al. 1997). Recently, two var-genes have been identified conferring binding to CSA (Scherf et al. 1998; Reeder et al. 1999). In both cases, expressed var-genes were isolated after selection on CHO cells, and a single var-gene transcript was identified. The FCR3.varCSA gene was located on chromosome 10 and had an unusual small transcript of 4.4 kb (Scherf et al. 1998), whereas the var-CS2 transcript (strain ItG2) was 8.1 kb long (Reeder et al. 1999). In both cases the loss of the original binding phenotype (CD36 or CD36 and ICAM1, respectively) was accompanied by a change to a new var-gene variant. Yet, no var-gene has been found, which mediates both binding to CD36 and CSA, thus implying that the motifs might be quite different. If the CIDR domain is involved in both CD36 and CSA binding, then coexistence of binding

motifs is expected. The location of binding motifs might overlap and might be completely exclusive, thereby leading to subgroups of var-genes binding either to CSA or to CD36. Using recombinant gene fragments, as in this study, might provide evidence for the involvement of various regions in binding. However, such an approach will not allow the analysis of more complex molecular interactions, i.e., the concerted action of two or more domains. Reeder and colleagues (1999) showed that the major inhibition of binding was observed with antibodies against DBL3; however, they also showed inhibition of CSA binding with antibodies against the CIDR domain. In contrast Buffet et al. (1999) found no evidence of involvement of the CIDR domain in his transfection experiments. The published sequence from varCS2 (GenBank Accession No. AF 134154) also contains homologies identified in this study (Fig. 1) and might be responsible for the binding of the CIDR domains to CSA. The observed regions of homology in the CIDR domain predict a conserved three-dimensional structure. These conserved regions might be involved in binding of iRBCs to CSA, but it cannot be excluded that other var-gene domains are involved in this binding. It could be speculated that the binding motif is rather short and might be present in most, if not all, PfEMP1 molecules. Robert et al. (1995) showed that mAb 1G11 and 4B2 against CSA could block adhesion of two different P. falciparum strains to Saimiri brain epithelium. A low prevalence of CSA-binding phenotype was observed in field isolates from Papua New Guinea (Rogerson et al. 1995) and Thailand (Chaiyaroj et al. 1996) and might reflect the fact that a CSA-binding phenotype is less frequent. It is possible that the primer design and identification of CIDR domains in this study selected a restricted var-gene repertoire with particular homologies. Furthermore, binding

120 to other receptors might be a much more complex process, but could result in binding with higher avidity. Therefore, CSA binding might be observed in vitro only and might be less frequently observed in vivo due to stronger binding forces to other receptors. And finally, correct folding and molecular interactions of PfEMP1 certainly will influence the adherence phenotype. A study on cytoadherence with human placentas (Fried and Duffy 1996) revealed that iRBCs from all placentas bound predominantly to CSA, whereas iRBCs from the peripheral blood of the same pregnant women bound to some degree to CSA, to CD36, or to both. This could be explained that CSA might be a minor receptor, which only becomes relevant in placental infections due to the lack of alterative receptors. With this study we provide further evidence for the involvement of var-gene CIDR domains in the binding of PfEMP1 to CSA. Further motif mapping, expression analysis, and ex vivo cytoadherence studies might lead to the rational design of new tools for the treatment of severe or placental malaria.

ACKNOWLEDGMENT

The authors are grateful to Dr. H. Matile for providing wild-type CHO cells (K1) and P. falciparum cultures and Dr. I. Felger for critically reading the manuscript. Mrs. M. Vogel gave valuable technical support to the project. This project was supported by the Swiss National Foundation, Grant 3100-049763.96.

REFERENCES

Barnwell, J. W., Ockenhouse, C. F., and Knowles, D. M. 1985. Monoclonal antibody OKM5 inhibits the in vitro binding of Plasmodium falciparum-infected erythrocytes to monocytes, endothelial and C32 melanoma cells. Journal of Immunology 135, 3494–3497. Baruch, D. I., Pasloske, B. L., Singh, H. B., Bi, X., Ma, X. C., Feldman, M., Taraschi, T. F., and Howard, R. J. 1995. Cloning the P. falciparum gene encoding PfEMP-1, a malarial variant antigen and adherence receptor on the surface of parasitized human erythrocytes. Cell 82, 77–87. Baruch, D. I., Ma, X. C., Singh, H. B., Pasloske, B. L., and Howard, R. J. 1997. Identification of a region of PfEMP1 that mediates adherence of Plasmodium falciparum infected erythrocytes to CD36: Conserved function with variant sequence. Blood 90, 3766–3775. Berendt, A. R., Simmons, D. L., Tansey, J., Newbold, C. I., and Marsh, K. 1989. Intercellular adhesion molecule 1 is an endothelial receptor for Plasmodium falciparum. Nature 341, 57–59.

DEGEN, WEISS, AND BECK

Buffet, P. A., Gamain, B., Scheidig, C., Baruch, D., Smith, J. D., Hernandez-Rivas, R., Pouvelle, B., Oishi, S., Fujii, N., Fusai, T., Parzy, D., Miller, L. H., Gysin, J., and Scherf, A. 1999. Plasmodium falciparum domain mediating adhesion to chondroitin sulfate A: A receptor for human placental infection. Proceedings of the National Academy of Sciences of the USA 96, 12743–12748. Chaiyaroj, S. C., Angkasekwinai, P., Buranakiti, A., Looareesuwan, S., Rogerson, S., and Brown, G. 1996. Cytoadherence characteristics of Plasmodium falciparum isolates from Thailand: Evidence for chondroitin sulfate A as a cytoadherence receptor. American Journal of Tropical Medicine and Hygiene 55, 76–80. Chen, Q., Barragan, A., Fernandez, V., Sundstro¨m, A., Schlichtherle, M., Sahlen, A., Carlson, J., Datta, S., and Wahlgren, M. 1998. Identification of Plasmodium falciparum erythrocyte membrane protein 1 (PfEMP1) as the rosetting ligand of the malaria parasite P. falciparum. Journal of Experimental Medicine 187, 15–23. Fried, M., and Duffy, P. E. 1996. Adherence of Plasmodium falciparum to chondroitin sulfate A in the human placenta. Science 272, 1502–1504. Gysin, J., Pouvelle, B., Le Tonqueze, M., Edelman, L., and Boffa, M. C. 1997. Chondroitin sulfate of thrombomodulin is an adhesion receptor for Plasmodium falciparum-infected erythrocytes. Molecular and Biochemichal Parasitology 88, 267–271. Hasler, T., Albrecht, G. R., Van Schravendijk, M. R., Aguiar, J. C., Morehead, K. E., Pasloske, B. L., Ma, C., Barnwell, J. W., Greenwood, B., and Howard, R. J. 1993. An improved microassay for Plasmodium falciparum cytoadherence using stable transformants of Chinese hamster ovary cells expressing CD36 or intercellular adhesion molecule-1. American Journal of Tropical Medicine and Hygiene 48, 332–347. Kemp, D. J., Thompson, J., Barnes, D. A., Triglia, T., Karamalis, F., Petersen, C., Brown, G. V., and Day, K. P. 1992. A chromosome 9 deletion in Plasmodium falciparum results in a loss of cytoadherence. Memorias do Instituto Oswaldo Cruz 87, 85–89. Matile, H., and Pink, J. R. L. 1990. Plasmodium falciparum malaria parasite cultures and their use in immunology. Immunological Methods IV, 221–234. Maubert, B., Guilbert, L. J., and Deloron, P. 1997. Cytoadherence of Plasmodium falciparum to intercellular adhesion molecule 1 and chondroitin-4-sulfate expressed by the syncytiotrophoblast in the human placenta. Infection and Immunity 65, 1251–1257. Miller, L. H., Good, M. F., and Milon, G. 1994. Malaria pathogenesis. Science 264, 1878–1883. Ockenhouse, C. F., Tegoshi, T., Maeno, Y., Benjamin, C., Ho, M., Kan, K. E., Thway, Y., Win, K., Aikawa, M., and Lobb, R. R. 1992. Human vascular endothelial cell adhesion receptors for Plasmodium falciparum-infected erythrocytes: Roles for endothelial leukocyte adhesion molecule 1 and vascular cell adhesion molecule 1. Journal of Experimental Medicine 176, 1183–1189. Pouvelle, B., Fusai¨, T., Lepolard, C., and Gysin, J. 1998. Biological and biochemical characteristics of cytoadhesion of Plasmodium falciparum-infected erythrocytes to chondroitin-4-sulfate. Infection and Immunity 66, 4950–4956. Reeder, J. C., Cowman, A. F., Davern, K. M., Beeson, J. G., Thompson, J. K., Rogerson, S. J., and Brown, G. V. 1999. The adhesion of

121

P. falciparum BINDING TO CHONDROITIN SULFATE A

Plasmodium falciparum-infected erythrocytes to chondroitin sulfate A is mediated by P. falciparum erythrocyte membrane protein 1. Proceedings of the National Academy of Sciences of the United States of America 96, 5198–5202.

Sinnis, P., and Sim, B. K. L. 1997. Cell invasion by the vertebrate stages of Plasmodium. Trends in Microbiology 5, 52–57.

Robert, C., Pouvelle, B., Meyer, P., Muanza, K., Fujioka, H., Aikawa, M., Scherf, A., and Gysin, J. 1995. Chondroitin-4-sulphate (proteoglycan), a receptor for Plasmodium falciparum-infected erythrocyte adherence on brain microvascular endothelial cells. Research in Immunology 146, 383–393.

Smith, J. D., Chitnis, C. E., Craig, A. G., Roberts, D. J., HudsonTaylor, D. E., Peterson, D. S., Pinches, R., Newbold, C. I. and Miller, L. H. 1995. Switches in expression of Plasmodium falciparum var genes correlate with changes in antigenic and cytoadherent phenotypes of infected erythrocytes. Cell 82, 101–110.

Roberts, D. D., Sherwood, J. A., Spitalnik, S. L., Panton, L. J., Howard, R. J., Dixit, V. M., Frazier, W. A., Miller, L. H., and Ginsburg, V. 1985. Thrombospondin binds falciparum malaria parasitized erythrocytes and may mediate cytoadherence. Nature 318, 689–692.

Smith, J. D., Kyes, S., Craig, A. G., Fagan, T., Hudson-Taylor, D. E., Miller, L. H., Baruch, D. I., and Newbold, C. I. 1998. Analysis of adhesive domains from the A4VAR Plasmodium falciparum erythrocyte membrane protein-1 identifies a CD36 binding domain. Molecular and Biochemical Parasitology 97, 133–148.

Roberts, D. J., Craig, A. G., Berendt, A. R., Pinches, R., Nash, G., Marsh, K., and Newbold, C. I. 1992. Rapid switching to multiple antigenic and adhesive phenotypes in malaria. Nature 357, 689–692. Rogerson, S., Chaiyaroj, S. C., Ng, K., Reeder, J., and Brown, G. 1995. Chondroitin sulfate A is a cell surface receptor for Plasmodium falciparum-infected erythrocytes. Journal of Experimental Medicine 182, 15–20. Rowe, A., Moulds, J. M., Newbold, C. I., and Miller, L. H. 1997. P. falciparum rosetting mediated by a parasite-variant erythrocyte membrane protein and complement-receptor 1. Nature 388, 292– 295. Rubio, J. P., Thompson, J. K., and Cowman, A. F. 1996. The var genes of Plasmodium falciparum are located in the subtelomeric region of most chromosomes. EMBO Journal 15, 4069–4077. Scherf, A., Hernandez-Rivas, R., Buffet, P., Bottius, E., Benatar, C., Pouvelle, B., Gysin, J., and Lanzer, M. 1998. Antigenic variation in malaria: in situ switching, relaxed and mutually exclusive transcription of var genes during intra-erythrocytic development in Plasmodium falciparum. EMBO Journal 17, 5418–5426.

L. H. 1994. Receptor and ligand domains for invasion of erythrocytes by Plasmodium falciparum. Science 264, 1941–1944.

Su, X. Z., Heatwole, V. M., Wertheimer, S. P., Guinet, F., Herrfeldt, J. A., Peterson, D. S., Ravetch, J. V., and Wellems, T. E. 1995. The large diverse gene family var encodes proteins involved in cytoadherence and antigenic variation of Plasmodium falciparuminfected erythrocytes. Cell 82, 89–100. Thompson, J. K., Rubio, J. P., Caruana, S., Brockman, A., Wickham, M. E., and Cowman, A. F. 1997. The chromosomal organization of the Plasmodium falciparum var gene family is conserved. Molecular and Biochemical Parasitology 87, 49–60. Treutiger, C. J., Heddini, A., Fernandez, V., Muller, W. A., and Wahlgren, M. 1997. PECAM-1/CD31, an endothelial receptor for binding Plasmodium falciparum-infected erythrocytes. Nature Medicine 3, 1405–1408. Udeiyna, I. J., Leech, J., Aikawa, M., and Miller, L. H. 1985. An in vitro assay for sequestration binding of Plasmodium falciparuminfected erythrocytes to formalin-fixed endothelial cells and amelanotic cells. Journal of Protozoology 32, 88–90.

Siano, J. P., Grady, K. K., Millet, P., and Wick, T. M. 1998. Short Report: Plasmodium falciparum: Cytoadherence to avb3 on human microvascular endothelial cells. American Journal of Tropical Medicine and Hygiene 59, 77–79.

Willimann, K., Matile, H., Weiss, N. A., and Imhof, B. A. 1995. In vivo sequestration of Plasmodium falciparum-infected human erythrocytes: A severe combined immunodeficiency mouse model for cerebral malaria. Journal of Experimental Medicine 182, 643–653.

Sim, B. K. L., Chitnis, C. E., Wasniowska, K., Hadley, T. J., and Miller,

Received 10 August 1999; accepted with revision 4 April 2000