Dual stage synthesis and crucial role of cytoadherence-linked asexual gene 9 in the surface expression of malaria parasite var proteins

Abstract

Plasmodium falciparum erythrocyte membrane protein 1 (PfEMP1) family members mediate the adherence of parasite-infected red blood cells (IRBCs) to various host receptors. A previous study has shown that the parasite protein, cytoadherence-linked asexual gene 9 (CLAG9), is also essential for IRBC adherence. However, how CLAG9 influences this process remains unknown. In this study, we show that CLAG9 interacts with VAR2CSA, a PfEMP1 that mediates IRBC adherence to chondroitin 4-sulfate in the placenta. Importantly, our results show that the adherent parasites synthesize CLAG9 at two stages—the early ring and late trophozoite stages. Localization studies revealed that a substantial level of CLAG9 is located mainly at or in close proximity of the IRBC membrane in association with VAR2CSA. Upon treatment of IRBCs with trypsin, a significant amount of CLAG9 (≈150 kDa) was converted into ≈142-kDa polypeptide. Together these data demonstrate that a considerable amount of CLAG9 is embedded in the IRBC membrane such that at least a portion of the polypeptide at either N or C terminus is exposed on the cell surface. In parasites lacking CLAG9, VAR2CSA failed to express on the IRBC surface and was located within the parasite. Based on these findings, we propose that CLAG9 plays a critical role in the trafficking of PfEMP1s onto the IRBC surface. These results have important implications for the development of therapeutics for cerebral, placental, and other cytoadherence-associated malaria illnesses.

Malaria caused by the Plasmodium species of protozoan parasites is a major public health problem around the globe with nearly half the population at risk for contracting the disease and ≈1 million people die annually (1, 2). Of several species of malaria parasites that infect humans, P. falciparum is the most deadly and is responsible for >80% of deaths due to malaria (3). A distinctive feature of P. falciparum compared with other malaria parasites is its ability to sequester in the microvascular capillaries of various organs by the adherence of infected red blood cells (IRBCs) to the endothelial cell surface molecules such as CD36, ICAM1, VCAM1, and PECAM1/CD31 and to chondroitin 4-sulfate in the placenta (4–6). This process leads to vascular obstruction, inflammation, endothelial damage, and organ dysfunction and failure. Hence, cytoadherence is central to the development of cerebral, placental, and other organ-related severe pathological conditions (7–9).

Studies have shown that a family of 150- to 400-kDa antigenic proteins, collectively called P. falciparum erythrocyte membrane protein 1 (PfEMP1), mediates the IRBC adherence (10). PfEMP1s are encoded by a repertoire of ≈60 var genes and are expressed on the IRBC surface in a mutually exclusive manner among parasite clonal populations (11). Different PfEMP1s exhibit distinctive adhesive property, which enable IRBCs to bind various host receptors (12), thereby sequestering in different organs and causing multiorgan pathology (7–9, 13, 14). Although IRBCs sequester mostly in the vascular capillaries, the process also occurs in the placental blood space during pregnancy (5–9, 11–14). The PfEMP1s that mediate cytoadherence in the vascular capillaries have not been characterized in detail. However, it is known that a specific PfEMP1 called VAR2CSA is the ligand for the chondroitin 4-sulfate (C4S)-mediated adherence of IRBCs in the placenta (15–19).

Studies have also shown that several proteins, including knob-associated histidine rich protein (KAHRP), PfEMP3, ring-infected erythrocyte surface antigen, mature parasite-infected erythrocyte surface antigen, and CLAG9 significantly influence cytoadherent property of the parasite (20, 21). Targeted deletion of cytoadherence-linked asexual gene 9 (clag9) demonstrated that CLAG9 is absolutely essential for IRBC binding to CD36 (21). CLAG9 is also essential for IRBC binding to C4S because the gene knockout parasites could not be selected for C4S binding (the present study), indicating that CLAG9 plays an important role in cytoadherence mediated by various PfEMP1s. However, how CLAG9 deletion results in the loss of IRBC binding to host receptors remains unknown.

clag9 is a member of a family of five genes in P. falciparum; other members include clag2, clag3.1, clag3.2, and clag8 (22). CLAGs are thought to be expressed exclusively by the late trophozoite and schizont stages and targeted to the rhoptries of merozoites, where they are present as high molecular weight RhopH/CLAG complex (22). During merozoite invasion, rhoptry proteins are injected into the membrane junction between invading parasite and erythrocyte surface, thereby localizing to parasitophorous vacuolar membrane (PVM) (23). All of the clag genes have identical intron–exon structures, but CLAG9 is distinct from all other CLAGs in its amino acid sequence (22). Further, unlike other clag genes, which are quite divergent in sequences of different isolates, clag9 is highly conserved in parasites from different geographical locations (24), suggesting that it has an important function in parasite biology. Furthermore, parasites lacking CLAG9 normally invade erythrocytes, but IRBCs are impaired in adherence, indicating that CLAG9 is critical for cytoadherence, but not for erythrocyte invasion (21).

Here, we studied the role of CLAG9 in PfEMP1-mediated cytoadherence by using C4S-adherent parasites as a model system. We show that, in C4S-adherent parasites, CLAG9 is synthesized both at the early ring and late trophozoite stages. CLAG9 along with several other parasite proteins associates with VAR2CSA and, presumably, with other PfEMP1 members as well. Based on its interaction with VAR2CSA and localization of VAR2CSA in clag9^−/− parasites, we suggest that CLAG9 is involved in the transport of PfEMP1s onto the IRBC surface.

Results and Discussion

VAR2CSA Binds C4S.

Gene knockout studies have shown that VAR2CSA is essential for the binding of P. falciparum IRBCs to C4S (16, 17). However, binding of VAR2CSA to C4S has not been demonstrated unequivocally. To biochemically identify the parasite protein(s) involved in IRBC adherence, C4S-selected parasites (FCR3 strain) were metabolically labeled with [³⁵S]amino acids during the ring stage. The radiolabeled IRBCs at the early to midtrophozoite stages were first extracted with Triton X-100 buffer, followed by buffer containing 2% SDS to solubilize the tightly bound membrane proteins. SDS/PAGE of the SDS buffer extracts followed by fluorography revealed the presence of a prominent high molecular mass (≈300 kDa) protein that was expressed only by the C4S-adherent, but not by nonadherent, parasites; this protein was absent in Triton X-100 buffer extracts (Fig. 1 A and B). On the other hand, different high molecular mass proteins were expressed by the CD36-adherent FCR3 parasites (Fig. 1C). Western blotting using anti-VAR2CSA antibodies showed that the ≈300-kDa protein is VAR2CSA and is specifically expressed by C4S-adherent parasites (Fig. 1D). These results agree with the previous findings that VAR2CSA is highly up-regulated in C4S-adherent parasites and that PfEMP1 is extractable from IRBCs with SDS buffer but not with Triton X-100 buffer (15, 25).

Fig. 1.

Several parasite proteins, including CLAG9, associate with VAR2CSA that bound to immobilized C4S. Synchronous cultures of FCR3 parasites at 8–10 h after invasion were metabolically labeled with [³⁵S]amino acids for 12–14 h. The radiolabeled parasites were harvested, extracted with Triton X-100 buffer followed by SDS buffer, and extracts were analyzed by SDS/PAGE/fluorography. (A) SDS buffer extracts of nonadherent (lane 1) and C4S-adherent (lane 2) parasites. (B) SDS/buffer (lane 1) and Triton X-100/buffer (lane 2) extracts of C4S-adherent parasites. (C) SDS/buffer extracts of C4S-adherent (lane 1) and CD36-adherent (lane 2) parasites. Arrows indicate the proteins specifically present in C4S-adherent parasites (A-C). The protein bands indicated by arrowheads presumably represent PfEMP1s expressed by CD36-binding FCR3 parasites. (D) Western blotting of SDS extracts of nonadherent (lane 1) and C4S-adherent (lane 2) IRBCs by using rabbit anti-VAR2CSA, rabbit anti-CLAG9, mouse anti-KAHRP, and mouse anti-Pf39 (loading control) antibodies at dilutions given in SI Materials and Methods. (E) [³⁵S]-labeled nonadherent FCR3 parasites. Lane 1, total SDS-buffer extract. Proteins pulled down with C4S beads in the absence (lane 2) and presence (lane 3) of 1 mg/mL soluble C4S. (F) [³⁵S]-labeled C4S-adherent FCR3 parasites. Lane 1, proteins pulled down with C4S beads. Lanes 2 and 3, proteins pulled down by C4S-coupled beads in the presence of 1 and 5 mg/mL soluble C4S, respectively. Lanes 4–6, proteins pulled down with C4S beads in the presence of 1 mg/mL C6S, and 1 and 2 mg/mL HA, respectively. (G) Western blotting, using rabbit anti-VAR2CSA and rabbit anti-CLAG9 antisera, of parasite proteins from nonadherent (lane 1), C4S-adherent (lane 2), and CD36-adherent (lane 4) parasites that were pulled down with C4S-beads. Lane 3, total 2% SDS extract of CD36-adherent parasites.

To demonstrate the binding of VAR2CSA to C4S, we performed pull-down studies by using adherent and nonadherent FCR3 parasites and C4S-immobilized beads. Interestingly, in addition to VAR2CSA, several other proteins from SDS extracts of C4S-adherent parasites were specifically present in the C4S–pull-down fraction (compare Fig. 1F, lane 1, with Fig. 1E, lane 2). Further, the binding of VAR2CSA to immobilized C4S was inhibited by soluble C4S but not by chondroitin 6-sulfate (C6S) or hyaluronic acid, indicating specific binding of VAR2CSA to C4S (Fig. 1F, lanes 1–6). The parasite proteins with molecular mass of ≈150, ≈75, and ≈32–35 kDa found in the C4S-bound fraction were also specifically inhibited by soluble C4S. Mass spectrometry analysis of the ≈150-kDa band in the protein fraction (Fig. 1F, lane 1) that was pulled down by C4S beads revealed the presence of RhopH2, rhoptry neck protein (RON3), and a Plasmodium exported protein PF14_0758 (Table S1). The protein, PF14_0758, is maximally expressed at 8–12 h after invasion (PlasmoDB). Deletion of this gene resulted in the accumulation of PfEMP1 in the Maurer's cleft and, thus, it was proposed to be a member of PfEMP1 transport machinery (26). These results suggest that several parasite proteins are involved in forming a VAR2CSA multiprotein complex in which some proteins might interact directly with VAR2CSA and play roles in VAR2CSA transport.

CLAG9 Associates with VAR2CSA to Form a Complex.

A previous study showed that CLAG9 is essential for IRBC adherence to CD36 (21). However, it remains unknown whether CLAG9 is also required for IRBC adherence to other host receptors such as C4S. Therefore, we panned 3D7 clag9^−/− parasites in parallel to wild-type 3D7 parasites on human placental chondroitin sulfate proteoglycan (CSPG)-coated plates. Although wild-type parasites yielded C4S-adherent IRBCs, we were unable to obtain adherent parasites from clag9^−/− parasites. Given that CLAG9 is essential for cytoadherence and it associates with RhopH2 (21, 22), which is identified here as a component of VAR2CSA complex, we investigated the possibility of CLAG9 being present as a component in the C4S-bound VAR2CSA fraction. We performed pull down of parasite proteins with C4S-coupled beads from FCR3 nonadherent, C4S-adherent, and CD36-adherent parasites. Upon Western blotting, VAR2CSA was detected in C4S-adherent parasites but not in nonadherent or CD36-adherent parasites (Fig. 1G Upper; compare lane 2 with lanes 1 and 4). Further, a ≈150-kDa protein band, the expected size for CLAG9, was strongly reactive to anti-CLAG9 antibodies specifically in C4S-adherent parasites, but not in those of nonadherent or CD36-adherent parasites (Fig. 1G Lower; compare lane 2 with lanes 1 and 4), even though CLAG9 is present in the SDS extracts of these parasites (Fig. 1 D and G Lower, lane 3). These results indicate that CLAG9 that is pulled down with C4S beads is associated with VAR2CSA that binds to C4S and, thus, is specific to C4S-adherent parasites. To confirm the association between CLAG9 and VAR2CSA, we performed immunoprecipitation by using anti-CLAG9 and anti-VAR2CSA antibodies. Anti-VAR2CSA antibodies pulled down both VAR2CSA and CLAG9 (Fig. 2, lane 3) but not Pf39 control protein (Fig. 2, lanes 3 and 5) from the SDS extracts of C4S-adherent IRBCs; Pf39 was found quantitatively in the supernatant after immunoprecipitation (Fig. 2, lane 6). Similarly, CLAG9 antibodies pulled down VAR2CSA and CLAG9 (Fig. 2, lane 4), demonstrating that CLAG9 interacts with VAR2CSA.

Fig. 2.

CLAG9 interacts with VAR2CSA. SDS extracts of the C4S-adherent FCR3 IRBCs were immunoprecipitated by using rabbit anti-VAR2CSA (lane 3) and rabbit anti-CLAG9 (lane 4) antibodies (at dilutions indicated in SI Materials and Methods) and the immunoprecipitates were probed, respectively, with rabbit anti-VAR2CSA (Top) antiserum and rabbit anti-CLAG9 (Middle) antiserum, and mouse anti-Pf39 antibodies (Bottom). Lane 1, immunoprecipitation with preimmune serum. Lane 2, protein A-Sepharose control used for immunoprecipitation. Lanes 5 and 6, the SDS buffer extract of C4S-adherent IRBCs were immunoprecipitated with anti-VAR2CSA antiserum. The immunoprecipitate (lane 5) and supernatants (lane 6) were analyzed by Western blotting using mouse anti-Pf39 antibodies.

C4S-Adherent Parasites Synthesize CLAG9 at Both Ring and Late Trophozoite Stages.

In parasites not selected for cytoadherence, it has been shown that CLAG9 is synthesized at the late trophozoite stage and is localized to the rhoptries of merozoites as a rhoptry protein complex and that the complex is transferred to the invaded erythrocytes (22, 23). To determine whether the newly synthesized CLAG9 or that transferred from merozoites to the invaded erythrocytes associates with VAR2CSA, we assessed clag9 expression at both transcriptional and protein levels in the C4S-adherent FCR3 parasites. Accordingly, first, we performed RT-PCR analysis at different stages of parasite development by using tightly synchronized culture. The results showed that CLAG9 is transcribed at both ring and late trophozoite stages in adherent parasites (Fig. 3A). Next, we metabolically labeled synchronous cultures of C4S-adherent parasites with [³⁵S]amino acids at 6- to 8-h intervals starting at 3–4 h after invasion. The radiolabeled cells were extracted with Triton X-100 buffer followed by SDS buffer and the extracts were analyzed by SDS/PAGE/fluorography (Fig. 3B Upper). A ≈150-kDa band was prominently radiolabeled both at the ring and late trophozoite stages. Upon Western blotting, this protein strongly reacted with anti-CLAG9 antibodies (Fig. 3B Lower). Thus, consistent with the results of mRNA levels (Fig. 3A), CLAG9 was also synthesized at significant levels during the ring stage by adherent parasites (9–10 h after invasion). The majority of CLAG9 synthesized at this stage is present in the SDS buffer extract, and only a low level was present in the Triton X-100 buffer extract (Fig. 3B Upper), indicating that soon after synthesis, CLAG9 is inserted into the detergent resistant membrane fraction. In contrast, in the late trophozoite and schizont stage parasites, a significant amount of CLAG9 was present in Triton X-100 buffer extracts as well. The synthesis of CLAG9 markedly decreased at the early trophozoite stage (22 h after invasion) and started increasing from the midtrophozoite stage (27 h after invasion) with maximum level of synthesis at the late trophozoite stage (36 h). Notably, the timing of early CLAG9 synthesis at peak levels (9–10 h after invasion) coincides with the beginning of VAR2CSA synthesis (27). Thus, it appears that the newly synthesized CLAG9 becomes associated with VAR2CSA soon after its synthesis. Note that the absence of ≈140-kDa band prominently present at the late stage (see below) indicates that the observed synthesis of CLAG9 at the ring stage is not due to the presence of late-stage parasites in the culture.

Fig. 3.

CLAG9 synthesis occurs both at the early ring and late trophozoite stages of adherent P. falciparum. (A) RT-PCR analysis of CLAG9 mRNA isolated from FCR3 parasites at 10, 22, and 36 h after invasion. Seryl tRNA synthetase mRNA was analyzed as an internal control. (B Upper) Synchronous cultures of FCR3 parasites were metabolically labeled with [³⁵S]amino acids for 6–8 h at different developmental stages starting at 3–4 h after invasion. The radiolabeled parasite cultures were harvested at the indicated time points, IRBCs extracted with Triton X-100 followed by SDS buffer, and extracts electrophoresed on 4–15% SDS-polyacrylamide gradient gels. The proteins in gels were transferred onto nitrocellulose membranes and radioactivity was visualized by fluorography. (B Lower) Western blotting of proteins in Upper by using 1:1,000 diluted rabbit anti-CLAG9 antibodies (SI Materials and Methods). PI, after invasion. Note: The seemingly lower amounts of CLAG9 and RhopH2 in lanes corresponding to 43–44 h after invasion are because these samples were analyzed on a different gel and, hence, exposure times were different. (C) The SDS buffer extracts of FCR3 nonadherent (NS) and C4S-adherent (CS) [³⁵S]-radiolabeled parasites (1 × 10⁶ cpm in each case) were immunoprecipitated by using 1:100 diluted rabbit anti-VAR2CSA antibodies, and the immunoprecipitates were analyzed by SDS/PAGE/fluorography.

To address the hypothesis that CLAG9 synthesized at the ring stage interacts with VAR2CSA, we performed immunoprecipitation of the SDS extract of IRBCs metabolically labeled at 10–12 h after invasion with [³⁵S]amino acids by using anti-VAR2CSA antibodies. The antibodies immunoprecipitated radiolabeled CLAG9 from C4S-adherent parasites but not from nonadherent parasites (Fig. 3C). These results are consistent with those from the coimmunoprecipitation of CLAG9 by anti-VAR2CSA antibodies (Fig. 2, lane 3). Thus, these data demonstrate that the newly synthesized CLAG9 interacts with VAR2CSA.

In contrast to the dual-stage expression of CLAG9, a ≈140-kDa protein was highly expressed only at the late trophozoite stage (Fig. 3B Upper). Considering that rhoptry-targeted proteins are expressed at high levels by the late-stage trophozoites (28), we were interested in identifying this protein. Trypsin digestion of this band and mass spectrometry analysis showed that the protein is RhopH2 (Table S2), a rhoptry-targeted protein. RhopH2 has been shown to complex with CLAG9 (22, 23). Thus, these results suggested that CLAG9 synthesized by the late-stage trophozoites complexes with RhopH2. It has been demonstrated that the proteins are targeted to rhoptries and, during the next cell cycle, the RhopH2-CLAG9 complex is carried by merozoites into the newly invaded RBCs and localizes mainly at PVM (23). A recent study reported that the RhopH2 acquired from the previous cell cycle is translocated to the erythrocyte cytoplasm and transiently associates with Maurer's cleft (29), suggesting that RhopH2 plays a role in protein transport. Because VAR2CSA also associates with Maurer's cleft during its transport to IRBC surface, it is possible that RhopH2 from the previous cell cycle interacts with VAR2CSA–CLAG9 complex at the PVM or in the erythrocyte cytoplasm, thereby involved in VAR2CSA transport.

Significant Level of CLAG9 Is Located at the IRBC Membrane and CLAG9 Is Involved in VAR2CSA Transport onto the IRBC Surface.

Further, we studied the localization of CLAG9 in the late ring and early trophozoite stages of C4S-adherent FCR3 parasites. Immunofluorescence analysis using anti-CLAG9 antibodies showed a strong staining of the IRBC membrane (Fig. 4A). A substantial level of CLAG9 is located at or in close proximity to the erythrocyte membrane. Nonpermeabilized IRBCs also showed considerable levels of staining of the erythrocyte surface with anti-CLAG9 antibodies (Fig. 4B), but the intensity of staining was substantially lower than that observed with permeabilized cells (Fig. 4A). Anti-VAR2CSA antibodies also localized VAR2CSA on the erythrocyte membrane of both permeabilized and nonpermeabilized IRBCs of the C4S-adherent parasites (Fig. 4 C and D); the nonadherent parasites were not stained (Fig. S1). Interestingly, in nonadherent IRBCs lacking VAR2CSA expression, most of the KAHRP is found in the IRBC cytoplasm, suggesting that functional transport machinery is absent because of the lack of proteins involved in PfEMP1 trafficking. We also examined VAR2CSA expression in 3D7 clag9^−/− and wild-type parasites. When 3D7 wild-type parasites were selected for C4S binding, VAR2CSA expression was observed on the IRBC surface (Fig. 4 E and F). Although not selected for C4S-adherence, ≈2.7% of the clag9^−/− IRBCs stained with anti-VAR2CSA antibodies (Fig. S2). In these cells, VAR2CSA was not detected on the IRBC surface, but present mainly within parasites and, to a certain extent, in the erythrocyte cytoplasm (Fig. 4G). These data demonstrate that CLAG9 is essential for the efficient transport of VAR2CSA from parasites to the erythrocyte cytoplasm and then to the IRBC surface. Further, in clag9^−/− parasites, KAHRP was found throughout the erythrocyte cytoplasm, resembling the distribution pattern that was observed in nonadherent parasites (compare Fig. 4G with Fig. S1A), suggesting that the transport machinery is not functional. Thus, these data indicate that the newly synthesized CLAG9 associates with VAR2CSA during its synthesis and is essential for the VAR2CSA surface expression.

Fig. 4.

VAR2CSA and CLAG9 localize predominantly to IRBC membranes of C4S-adherent parasites. Immunofluorescence and light micrographs (LM) of membrane-permeabilized (A and C) and intact (B and D) IRBCs of C4S-adherent FCR3 parasites, harvested at the late ring/early trophozoite stage, were stained by using rabbit anti-CLAG9 (A and B), rabbit anti-VAR2CSA (C and D), and mouse anti-KAHRP (control) antibodies at dilutions given in SI Materials and Methods. (E and F) Immunofluorescence and LM of permeabilized (E) and intact (F) C4S-adherent 3D7 IRBCs stained as above with anti-VAR2CSA antibodies. (G and H), Immunofluorescence and LM of permeabilized clag9^−/− parasites stained with anti-VAR2CSA antibodies (G) and anti-CLAG9 antibodies (H).

To gain further insight into the relative localization of CLAG9 and VAR2CSA, we performed immunoelectron microscopy of C4S-adherent IRBCs by using 12- and 18-nm gold particle-conjugated antibodies, respectively, to localize VAR2CSA and CLAG9. The results show that, even in the early stage parasites, CLAG9 is located mainly at or in close proximity to the erythrocyte membrane (Fig. 5 and Fig. S3). A significant amount of VAR2CSA colocalized with CLAG9 in the erythrocyte membrane. However, the density of gold particles corresponding to VAR2CSA was substantially low, likely because of a significant loss of anti-VAR2CSA antibody binding epitopes during glutaraldehyde/paraformaldehyde fixation of cells.

Fig. 5.

VAR2CSA and CLAG9 are present in the IRBC membranes of C4S-adherent parasites. The C4S-adherent FCR3 parasites, at the late ring to early trophozoite stages, were analyzed by immunoelectron microscopy by using rat anti-VAR2CSA and rabbit anti-CLAG9 primary antibodies followed by biotin-conjugated goat anti-rat IgG and 18-nm gold particle-conjugated goat anti-rabbit IgG secondary antibodies (SI Materials and Methods). The bound biotin-labeled antibodies were probed with 12-nm gold particles-conjugated streptavidin. Arrows and arrowheads indicate, respectively, 12- and 18-nm gold particles. (Scale bars: 200 nm.)

Significant Level of CLAG9 Is Embedded in the IRBC Membrane and a Portion of the Peptide at Either N or C Terminus Is Surface Exposed.

To confirm the surface exposure of CLAG9, we performed trypsinization of IRBCs. Treatment of the C4S-adherent FCR3 IRBCs with trypsin and Western blotting analysis of the SDS buffer extract showed that a portion of CLAG9 at either its N or C terminus was cleaved by the enzyme, whereas the remainder was unaffected (Fig. 6). On the other hand, CLAG9 that was extracted with Triton X-100 buffer was completely resistant to trypsin, suggesting that a substantial amount of CLAG9 is present in the PVM and/or beneath the parasite membranes. The enzyme-cleaved fragment of CLAG9 was ≈8 kDa lower in size than intact CLAG9 (≈150 kDa) and no other fragments of CLAG9 were observed. These data demonstrate that a considerable amount of CLAG9 is embedded in the IRBC membrane with a portion at either the N or C terminus is exposed on the surface. Because the trypsin-cleaved ≈8-kDa polypeptide corresponds to an ≈70-amino acid stretch and anti-CLAG9 antibodies used here recognize peptide epitopes at either end of CLAG9 (390-406 and 1,282–1,294), information about whether N- or C-terminal end was cleaved could not be obtained.

Fig. 6.

A significant amount of CLAG9 is in the IRBC surface such that a small portion at its either N- or C-terminal end is extracellular. C4S-adherent FCR3 IRBCs at the late ring to early trophozoite stage were treated with trypsin at the indicated conditions, extracted with Triton X-100 buffer (A) followed by SDS buffer (B), and extracts were analyzed by Western blotting using rabbit anti-CLAG9 antibodies (SI Materials and Methods). Arrow indicates the trypsin-cleaved fragment seen only in the SDS buffer extracts. Pf39 protein loading control analyzed by Western blotting using anti-Pf39 antibodies is also shown.

The results of this study show that VAR2CSA associates with CLAG9. Soon after synthesis, most of the newly synthesized CLAG9 is present in the Triton X-100-resistant membrane fraction, presumably in lipid rafts, where clustering proteins and transmembrane proteins localize (30). Such lipid rafts have been shown to be involved in trafficking of proteins from the Golgi apparatus to plasma membrane (30). Based on this information and our findings described here, we propose the following model for the function of CLAG9 in the PfEMP1-mediated adherence of IRBCs. The parasites synthesize CLAG9 at the early ring stage such that a significant pool of protein is available for interaction with VAR2CSA. The newly synthesized CLAG9, PF14_0758, and possibly other parasite proteins associate with VAR2CSA in the lipid rafts in the ER/Golgi apparatus and the complex traffics to the parasite plasma membrane and then to PVM. Eventually, the protein complex moves through PVM into RBC cytoplasm, and then finally VAR2CSA and CLAG9 are inserted into the RBC membrane such that the Duffy binding-like domains of VAR2CSA are surface exposed and CLAG9 polypeptide is partially exposed either at the N or C terminus on the IRBC surface. The CLAG9 that is synthesized by the late stage parasite is targeted to rhoptries and, subsequently, transferred to the invaded erythrocytes as CLAG9–RhopH2 complex (23). This complex has been shown to move to erythrocyte membrane (29), and it might be involved in protein transport. Because CLAG9 is also important for the binding of PfEMP1s to CD36 and likely to other host receptors (21), presumably CLAG9 also interacts with other PfEMP1s that are expressed on the IRBC surface and involved in IRBC adherence to vascular endothelia. Because the ectodomains of PfEMP1s are quite variable, it is logical that CLAG9 interacts with the conserved cytoplasmic portion of PfEMP1s. Furthermore, it has been shown that the cytoplasmic portion of PfEMP1s also interacts with KAHRP, which in turn interacts with RBC cytoskeletal proteins such as spectrin and F-actin to form rigid knob structures (31, 32). Thus, our results provide important insights into the processes involved in PfEMP1 transport. These data have important implications for the development of therapeutics, based on the abolition of PfEMP1 surface expression by using small molecule inhibitors, against cerebral, placental, and other organ-related severe malaria pathology.

Materials and Methods

Clag9^−/− parasites were of 3D7 background, and the results obtained from these knockout parasites were compared with wild-type 3D7 strain having the same genotype. All other experiments were performed by using C4S-adherent, CD36-adherent, and nonadherent FCR3 parasites of same genotype. Rabbit and rat antisera against VAR2CSA were produced by using, respectively, the recombinant DBL5ε and DBL4ε expressed in insect cell (33). The rabbit anti-CLAG9 antibodies were produced by using synthetic CLAG9 peptides as reported (23). Anti-KAHRP mouse monoclonal antibody (mAb89) and primers for parasite genotyping were generous gifts by Diane W. Taylor. Anti-Pf39 mouse antibody (MRA-87) was provided by the Malaria Research and Reference Reagent Resource Center (Manassas, VA).

Coupling of C4S and CSPG to Sepharose.

The CSPG was purified from human placenta as described (34) and conjugated to CNBr-activated Sepharose 4B beads. The density of coupled CSPG was ≈1.6 mg/mL Sepharose beads as estimated by determining the amount of uncoupled CSPG using the uronic acid assay (35). A CSPG form of recombinant thrombomodulin (TM; ref. 36) was coupled to Sepharose 4B to a density of 0.6 mg of TM/mL of beads. The CSPG–TM was also conjugated to magnetic Dynabeads M-450 (1 × 10⁸). Bovine tracheal chondroitin 4-sulfate (bCSA) was derivatized with adipic dihydrazide (37) and then coupled to CNBr-activated Sepharose 4B to a density of 2.4 mg of CSA/mL of gel.

C4S Pull-Down Analysis of Parasite Proteins.

The 2% SDS buffer extract (20 μL; SI Materials and Methods) of IRBCs was diluted to 8 mL with RPMI medium1640 containing 1% BSA, 0.5% Triton X-100, and protease inhibitors to adjust SDS concentration to 0.005%. To the diluted extracts (0.5 mL per sample) were added bCSA- or CSPG-conjugated Sepharose 4B gel (5- to 10-μL pellet) or TM-coupled Dynabeads (1 × 10⁷ beads), which were mixed and incubated in a rotator at 4 °C overnight. For inhibition studies, C4S, C6S, and hyaluronic acid were added to extracts before incubation with C4S-conjugated beads. The gel suspensions were centrifuged, and gel pellets or beads were washed with PBS until radioactivity in washings was at background levels. In the case of Dynabeads, the beads were separated by a magnetic field and washed. The bound materials in gel pellets or beads were eluted with 10 μL of 2× SDS/PAGE reducing sample buffer, aliquots were measured for radioactivity in Beckman liquid scintillation counter LS6000IC, and samples analyzed by SDS/PAGE/fluorography (SI Materials and Methods).

Descriptions of the remaining methods are given in SI Materials and Methods.