Structural and Functional Studies of Interaction between Plasmodium falciparum Knob-associated Histidine-rich Protein (KAHRP) and Erythrocyte Spectrin*

,

Malaria caused by Plasmodium falciparum is the most serious parasitic disease of humans. Clinical symptoms occur during the asexual stage of the life cycle of the parasites, at which time it multiplies within the human erythrocyte. During intraerythrocytic growth, the parasite extensively modifies the host erythrocyte resulting in alterations of morphology, mechanical properties, and adhesive properties. It is generally believed that almost all the altered properties of parasitized erythrocytes are because of the action of a group of parasite proteins that become associated with erythrocyte membrane proteins (see Refs. 1-3 for recent reviews). Well studied mem-bers of this group include the knob-associated histidine-rich protein (KAHRP), 1 P. falciparum erythrocyte membrane protein 1 (PfEMP1), the ring parasite-infected erythrocyte surface antigen (RESA), and the mature parasite-infected erythrocyte surface antigen (MESA).
KAHRP, an 85-105-kDa protein that is expressed during the middle and later stages of the asexual cycle (4,5), has been shown to be critically important for knob formation in infected erythrocytes (6). It interacts with erythrocyte skeletal proteins such as spectrin, actin, and ankyrin (7,8) and also with erythrocyte membrane-associated parasite protein, PfEMP1 (9). PfEMP1 mediates the adhesion of parasitized erythrocytes to the vascular endothelium (10), a process strongly implicated in the pathology of cerebral malaria (11). The absence of the KAHRP protein at the erythrocyte membrane leads to a weakening of the interaction between PfEMP1 and the vascular endothelium at physiologic shear stresses (6). This decreased avidity of the interaction in the absence of KAHRP can mitigate the severity of vascular obstruction and hence complications of cerebral malaria. In contrast, little is known regarding the functional implication of KAHRP-skeletal protein interactions.
Spectrin is a major component of the erythrocyte membrane skeleton and plays a key role in maintaining erythrocyte membrane integrity and mechanical stability (12). It is composed of two subunits, an ␣ spectrin (280 kDa) chain and a ␤ spectrin (246 kDa) chain, both being made up of repeating units (the ␣ spectrin chain has 21, and the ␤ spectrin has 16) flanked by non-consensus structures. Each repeat is composed of ϳ106 amino acids and folds into a left-handed, antiparallel triplehelical coiled coil structure. The ␣ spectrin and the ␤ spectrin associate side to side in an antiparallel orientation to form heterodimers. Spectrin heterodimers then self-associate head to head to form spectrin tetramer, which exists as an elongated molecule with a contour length of 200 nm (see Refs. 13 and 14 for recent reviews). The function of terminal regions of spectrin chains has been well defined. For example, the N terminus of the ␣ spectrin chain and the C terminus of the ␤ spectrin chain are involved in tetramerization (15,16), and the N terminus of ␤ spectrin chain is responsible for binding to actin and protein 4.1R (17,18). 2 However, the function of spectrin re-peats, which make up the bulk of a spectrin molecule, remains largely unknown.
In the present study, we identified the binding site for spectrin in KAHRP and the reciprocal KAHRP-binding site in spectrin. We showed that spectrin bound to a specific region of KAHRP (residues 370 -441) and that this region of KAHRP specifically bound to a single repeat of ␣ spectrin (repeat 4, ␣R4). Furthermore, resealing of the ␣R4 polypeptide into erythrocytes prior to infection by malarial parasites prevented the membrane association of KAHRP. These findings enabled us to identify an important functional role for the interaction between KAHRP and spectrin in localizing KAHRP to the membrane of the infected erythrocytes.

EXPERIMENTAL PROCEDURES
Materials-Type O ϩ fresh blood was taken from healthy volunteers with informed consent. Serum from type A blood donor was obtained from Interstate Blood Bank (Memphis, TN). Parasite clone 3D7 was obtained from MR4 (Manassas, VA). pGEX vector and glutathione-Sepharose 4B were purchased from Amersham Biosciences. pET28b(ϩ) vector and nickel column were from Novagen (Madison, WI), and BL21(DE3) bacteria were from Stratagene (La Jolla, CA). Reduced form glutathione and isopropyl ␤-D-galactopyranoside were purchased from Sigma. Anti-spectrin antibody was prepared by injecting rabbit with spectrin purified from erythrocytes as antigen, and the resulting antibody was affinity-purified. Monoclonal anti-His antibody was from Roche, and HRP-conjugated anti-mouse IgG and HRP-conjugated antirabbit IgG were from Jackson ImmunoResearch Laboratory (West Grove, PA). Renaissance chemiluminescence detection kit was from Pierce Biotechnology, Inc. Percoll was from Sigma. TOPO-3 and fluorescein isothiocyanate-conjugated anti-rabbit IgG were from Molecular Probes (Eugene, OR). The CM-5 sensor chip, amino coupling kit, and other reagents for SPR assay were purchased from BIAcore (Piscataway, NJ). All other chemicals were reagent grade products from standard sources.
Preparation of Spectrin, Recombinant Spectrin Fragments, and KAHRP Fragments-Spectrin from erythrocytes was prepared according to the method described by Tyler et al. (19). Spectrin fragments were cloned into pGEX-4T-2 (20) and KAHRP fragments were cloned into pGEX-KG vector (9). The cDNA encoding the desired polypeptide was transformed into BL21 bacteria strain. The expression of recombinant proteins was induced by 0.1 mM isopropyl ␤-D-thiogalactopyranoside at 16°C for 3-4 h. The GST-tagged spectrin and KAHRP polypeptides were purified using a glutathione-Sepharose 4B affinity column. Two His-tagged KAHRP fragments (K2 and K2aϩb) were subcloned into pET-28b(ϩ) vector using NcoI and XhoI cloning sites upstream and downstream, respectively. His-tagged KAHRP fragments were expressed as above and purified with a nickel column. Proteins were dialyzed against PBS (10 mM phosphate, pH 7.4, 150 mM NaCl). Protein concentrations were determined spectrophotometrically using extinction coefficients calculated from the tryptophan and tyrosine contents, taking the molar extinction coefficients of these amino acids at 280 nm as 5500 and 1340, respectively (21). All the proteins were clarified by ultracentrifugation at 230,000 ϫ g for 30 min at 4°C before use.
Pull-down Assays-For GST pull-down assay, GST-tagged recombinant polypeptides were coupled to glutathione-Sepharose 4B beads at room temperature for 30 min. Beads were pelleted and washed. Binding partner was added to the coupled beads in a final volume of 80 l. The final concentration of the coupled protein was 1 M. The mixture was incubated for 1 h at room temperature, pelleted, washed, and eluted with 10% SDS. The pellet was analyzed by SDS-PAGE. The binding of full-length spectrin to KAHRP fragments was detected by Western blot using anti-spectrin antibody, and similarly the binding of His-tagged KAHRP fragments to GST-tagged recombinant spectrin fragments was detected by Western blot using anti-His antibody. GST was used as negative control in all experiments. For His pull down, the procedure was basically the same except that the His-tagged protein was coupled to nickel beads.
Surface Plasmon Resonance Assay-Surface plasmon resonance assay was performed using a BIAcore 3000 instrument. His-tagged KAHRP polypeptide (K2 or K2aϩb) was covalently coupled to a CM-5

FIG. 1. Schematic representation (A) and SDS-PAGE of recombinant spectrin fragments (B) as well as spectrin single repeats (C).
The boundaries of all spectrin fragments and single repeats were defined by SMART annotations. 1 g of total purified GST fusion proteins was analyzed by 10% SDS-PAGE and stained with GelCode Blue. N and C indicate the N terminus or C terminus of spectrin chains, respectively.

FIG. 2. Schematic representation (A) and SDS-PAGE of recombinant KAHRP fragments (B).
The P. falciparum KAHRP gene contains two exons separated by an intron. The first exon encodes a putative signal sequence, and the second exon contains three repeat regions as indicated. Amino acid residue numbers and the lengths of fragments are indicated. GST-tagged KAHRP fragments were run in 10% SDS-PAGE and His-tagged KAHRP fragments were run in 15% SDS-PAGE. Binding of spectrin dimer to KAHRP fragments. Spectrin dimer was incubated with various GST-tagged KAHRP fragments at room temperature for 30 min. The interaction was assessed by GST pull-down assay, and the binding was detected by anti-spectrin antibody.
biosensor chip using an amino coupling kit. Binding reactions were done in HBS-EP buffer, containing 20 mM HEPES, pH 7.4, 150 mM NaCl, 3 mM EDTA, and 0.05% (v/v) surfactant P20. The surface was regenerated before each new injection using 50 mM NaOH. The BIAcore instrument was programmed to perform a series of binding assays with increasing concentrations of ␣N-5 or ␣R4 polypeptide over the same regenerated surface. Sensograms (plots of changes in response unit on the surface as a function of time) derived were analyzed using the software BIAeval 3.0. Affinity constants were estimated by curve fitting using a 1:1 binding model.
Resealing of Erythrocytes-To introduce recombinantly expressed proteins into erythrocytes, the washed erythrocytes were lysed and resealed using the dialysis method (22). Briefly, 2 ml of packed erythrocytes (in a dialysis tube, in the absence or presence of desired polypeptides) at a 50% hematocrit were dialyzed against 500 ml of cold hypotonic buffer (5 mM KPO 4 , pH 7.4, 20 mM KCl) for 80 min at cold. The isotonicity was restored, and erythrocytes were resealed by dialyzing the lysed cells against 500 ml of prewarmed isotonic buffer (5 mM KPO 4 , 160 mM KCl, 5 mM glucose) for 60 min at 37°C.
Invasion of Erythrocytes by Malaria Parasites-The 3D7 clone parasite was maintained using standard procedures in RPMI 1640 medium (23). Sorbitol synchronization of P. falciparum-infected erythrocytes was performed as described before (24). The synchronized infected cells were purified by Percoll gradient purification to nearly 100% purity (25). The fraction containing mature trophozoite-infected erythrocytes was added to resealed erythrocytes at a starting parasitemia of 2%. The cells were cultured using standard procedures in RPMI 1640 medium for 48 h. At the end of the culture, the infected erythrocytes were separated by Percoll gradient, and the purified late stage infected cells were examined by confocal microscopy.
Confocal Microscopy-Preparation of erythrocytes for immunofluorescence was performed as previously described (26). Briefly, Percoll gradient purified late stage infected erythrocytes were washed twice in PBS containing 5 mM glucose and then fixed for 5 min in 0.1% glutar-aldehyde in PBS. Cells were permeabilized in PBS containing 0.1 M glycine (rinsing buffer) plus 0.1% Triton-100 for 5 min followed by three rinses using rinsing buffer. Nonspecific binding was blocked by incubation for at least 60 min in blocking buffer (PBS containing 0.05 mM glycine, 0.2% fish skin gelatin, and 0.05% sodium azide). Staining of fixed, permeabilized cells was performed using anti-KAHRP antibody diluted in blocking buffer. After labeling, cells were resuspended in PBS and allowed to attach to cover slips coated with polylysine. The cover slips were mounted using Aqua-Mount medium. Samples were imaged with Zeiss 510 META confocal microscope using 63ϫ oil object.
SDS-PAGE and Western Blot-SDS-PAGE of samples (15 l of each) was performed on a 10% acrylamide gel. Proteins were transferred onto nitrocellulose membrane. After blocking for 1 h in blocking buffer (10 mM Tris, pH 7.4, 150 mM NaCl, 0.5% Tween 20, 5% nonfat dry milk), the blot was probed for 1 h with the desired primary antibody (polyclonal anti-spectrin antibody or monoclonal anti-His antibody). After several washes, the blot was incubated with anti-mouse (or anti-rabbit) IgG coupled to HRP. After final washes, the blot was developed using the Renaissance chemiluminescence detection kit. All steps were performed at room temperature.

Recombinant Protein Expression and
Purification-A schematic representation of the various spectrin and KAHRP fragments used in the present study are illustrated in Figs. 1A and 2A, respectively. All spectrin fragments and spectrin single repeats were cloned into pGEX-4T-2 protein expression vector (20). To detect by Western blotting the binding of KAHRP fragments to GST-tagged spectrin fragments using anti-His antibody, we constructed His-tagged K2 and K2aϩb fragments from previously described GST-tagged KAHRP fragments (9). At 16°C, all recombinant proteins are well expressed and soluble. The purity of various GST-spectrin fusion proteins and GST-tagged KAHRP, as well as His-tagged KAHRP fragments, used in the present study are shown in Figs. 1, B and C and 2B, respectively.
Mapping the Spectrin-binding Site in KAHRP-To define the spectrin-binding site in KAHRP, three recombinant GSTtagged KAHRP polypeptides (K1, K2, and K3), which encompass the full-length of KAHRP, were purified and examined for their ability to bind the spectrin dimer using the pull-down assay. As shown in Fig. 3A, under the binding conditions used in these experiments only the GST-tagged K2 fragment was able to pull down the spectrin dimer. Spectrin dimer did not bind to GST, GST-tagged K1, or K3 fragments. To further define the spectrin binding region in K2 fragment, the binding FIG. 4. Binding of K2 fragment to recombinant spectrin fragments and single repeats. His-tagged K2 fragment was incubated with various recombinant GST-tagged spectrin fragments and single repeats. The binding was assessed by GST pull-down assay and detected by Western blot using anti-His antibody. of spectrin dimer to subfragments of K2, K2A, and K2B, was examined. Interestingly, neither of them had the ability to bind spectrin dimer (Fig. 3B). Based on this finding, we reasoned that the spectrin-binding site in K2 fragment probably requires the C-terminal part of K2A and N-terminal part of K2B. To test this hypothesis, we constructed a fragment designated K2aϩb, which is composed of 36 amino acids from the C-terminal of K2A and 36 amino acids from the N-terminal of K2B. As shown in Fig. 3B, in contrast to K2A and K2B, K2aϩb did bind spectrin dimer. Thus, we localized a distinct spectrin-binding site in KAHRP to a 72-amino-acid region. This finding is in agreement with a previous report, which localized the spectrinbinding site in KAHRP to a much larger 271-amino-acid region that encompasses the domain we identified (27).
Mapping the KAHRP-binding Site in Spectrin-The same experimental approach was used to identify the KAHRP-binding site in the spectrin chains. Nine recombinant GST-tagged spectrin fragments encompassing the entire ␣ and ␤ spectrin chains were purified, and the binding of His-tagged K2 and His-tagged K2aϩb to these fragments was examined. As shown in Fig. 4A, the His-tagged K2 fragment bound specifically to only one ␣ spectrin fragment, ␣N-5 but not to any of the other eight spectrin fragments or to GST. Among the various structural elements that constitute the ␣N-5 fragment, K2 specifically bound to one single repeat, ␣R4 (Fig. 4B). The K2aϩb fragment exhibited the same behavior as the K2 fragment in that it specifically bound only the ␣N-5 fragment and the single repeat, ␣IR4 (data not shown). Thus among the 37 repeats of ␣ and ␤ spectrin, we have localized the KAHRP-binding site on spectrin to one single repeat, ␣R4.
Interactions between KAHRP Fragments and Spectrin Fragments as Assessed by Surface Plasmon Resonance Assay-To further confirm and characterize the interactions between KAHPR fragments and spectrin polypeptides, real-time plasmon resonance assays were performed. In these experiments, His-tagged KAHRP fragment (K2 or K2aϩb) was immobilized onto the surface of a sensor chip, and the binding of GST-tagged spectrin polypeptide (␣N-5 or ␣R4) was assessed. In agreement with the pull-down assay data, GST-␣N-5 and GST-␣R4 bound to the immobilized K2 fragment, whereas GST itself did not show any detectable binding (Fig. 5A). The dose-dependent binding of ␣N-5 to K2 is shown in Fig. 5B. Table I summarizes the characteristics of interactions between spectrin polypeptides (␣N-5, ␣R4) and KAHRP fragments (K2, K2aϩb). The binding affinities of all of these interactions are in the micromolar range, values comparable with those reported previously for various other protein-protein interactions in normal and infected erythrocytes (2).
Inhibition of Full-length Spectrin Binding to K2 Fragment by the Single Spectrin Repeat, ␣R4 -To confirm the specificity of the interaction between the single spectrin repeat ␣R4 and KAHRP, we performed a competitive inhibition assay. In this experiment, the K2 fragment was preincubated with various concentrations of ␣R4 prior to the addition of the full-length spectrin dimer. As shown in Fig. 6, the binding of spectrin dimer to the K2 fragment was decreased with progressively increasing concentrations of ␣R4. At a concentration of 1 M ␣R4, spectrin binding to KAHRP was inhibited by 50%.
Effect of ␣R4 on KAHRP Distribution and Localization in Infected Erythrocytes-To examine the functional implication of KAHRP-spectrin interaction, we examined the localization of KAHRP in infected erythrocytes that were resealed with excess amount of ␣R4 polypeptide prior to infection with P. falciparum parasites. It was anticipated that the exogenously added ␣R4 polypeptide would compete with native membrane-associated spectrin for interaction with KAHRP, which is expressed in the later stages of malaria infection. The location and distribution of KAHRP in the infected erythrocytes was monitored following 48 h of culture. As shown in Fig. 7A, KAHRP localized at the membrane in erythrocytes resealed with no added polypeptide as described previously (28,29). A very similar localization of KAHRP was seen in erythrocytes resealed with GST prior to infection (Fig. 7B). Strikingly, in erythrocytes resealed with GST-␣R4, KAHRP was diffusely distributed in the cytoplasm of infected erythrocytes (Fig. 7C). Uninfected erythrocytes did not show any staining with anti-KAHRP antibody (data not shown). DISCUSSION In the present study, we performed a detailed molecular characterization of the interaction between the P. falciparumencoded protein, KAHRP, and erythrocyte membrane skeletal protein, spectrin. A surprising feature of our finding is that only one unique repeat of ␣ spectrin bound KAHRP. Importantly, when this spectrin repeat was resealed into erythrocytes it blocked the membrane assembly of KAHRP in infected erythrocytes. This finding implies that spectrin plays a key role in localizing KAHRP at the membrane of infected erythrocytes.
Knob-like, electron-dense protrusions located at the membrane surface of infected erythrocytes are localized at the points of adhesive interaction between infected erythrocytes and vascular endothelial cells (30). KAHRP is associated with these knob-like structures. Although PfEMP1, the adhesive receptor expressed on the surface of infected erythrocytes is the mediator of adhesive interactions, KAHRP plays an important was added, and its binding to the K2 fragment was measured by GST pull-down assay. The binding was detected by anti-spectrin antibody and quantitated by densitometry. role in modulating the avidity of these interactions, presumably by clustering the receptors. Strong support for this thesis comes from the finding that in the absence of KAHRP, infected erythrocytes adhere very weakly to endothelial cells under physiological flow conditions (6). As adherence of infected erythrocytes to vascular endothelium is a major determinant of the pathogenicity of P. falciparum (11), reducing or eliminating adhesive interactions by interfering with the interaction of KAHRP with ␣ spectrin is a potential therapeutic strategy to prevent serious complications of P. falciparum malaria.
Previous studies have identified interactions between a number of parasite proteins including RESA (31-33), MSP-1 (34), MESA (35), and PfEMP1 (7) and erythrocyte membrane proteins. Two recent studies have suggested that over 400 parasite-encoded proteins are exported into erythrocyte cytoplasm (36,37), and a number of these are likely to interact with membrane proteins. Most studies, to date, on the interactions of parasite proteins with erythrocyte membrane proteins have focused on defining the domains in malaria proteins that participate in these interactions. For example, the 4.1R-binding site in MESA has been localized to 19 residues located in the N-terminal region of MESA (35). The spectrin-binding domain in RESA has been mapped to a 48-residue region (33), whereas the spectrin-binding domain in MSP-1 has been localized to a 30-residue region (34).
A unique feature of many malaria proteins is the presence of extensive regions of tandemly repeated sequences (38). The functional role of these repeats is not clear. It is interesting to note that the binding regions of MESA, RESA and MSP-1 mentioned above are all found in non-repetitive domains (33)(34)(35). Previous work that mapped the spectrin-binding domain of KAHRP to a 271-residue region (27), and this region contained the 5Ј-repeat. However, in the present study we found that K2A, which is composed of the complete 5Ј-repeat and 21 amino acids downstream of this 5Ј-repeat sequence did not bind spectrin. This suggests that the repeat sequence in KAHRP is not involved in KAHRP binding to spectrin. The binding site of some sporozoite proteins for hepatocytes has also been mapped to non-repetitive regions (39,40). Thus it appears that the repeat sequences of parasite proteins are probably not involved in association of malaria proteins with host proteins. Another common feature of binding motifs in malaria proteins for erythrocyte skeleton proteins is the relatively short length of these sequences. This reliance on short linear sequences may arise from the necessity of the exported malaria protein to interact with a preexisting network of proteins. However, the above considerations may not be relevant for malarial protein-protein interactions, as multiple binding regions in KAHRP have been identified to interact with PfEMP1, and the binding regions do contain repetitive sequence (41).
In contrast to our detailed understanding of the binding motifs on malaria proteins much less is known regarding the binding motifs on host cell proteins. To date, only the MESA-binding site on 4.1R has been mapped to a 51-residue region encoded by exon 10 of the 4.1R gene (39). Spectrin is a long, rod-like molecule with a contour length of 200 nm. It has been traditionally thought that the major function of spectrin is to provide the structural basis for the flexibility of erythrocyte membranes because of its unique repetitive triple ␣-helical structure. The successful localization of the KAHRP binding region to a single repeat (of 37 repeats) of spectrin highlights the specificity of KAHRP-spectrin interaction. In fact, this is the first demonstration that a specific spectrin repeat could serve as a docking site for malaria proteins in infected erythrocytes. This finding further implies that spectrin may serve as a scaffold for assembly of malarial proteins in infected erythrocytes.
The most striking outcome of this study is the observation that resealing of ␣R4 polypeptide into erythrocytes alters the localization of KAHRP in infected cells. This is probably because of the failure of interaction of KAHRP with membraneassociated spectrin in the presence of excess amount of the competitive polypeptide ␣R4. Thus, binding of KAHRP to spectrin is required for its assembly onto the erythrocyte membrane skeleton. These findings suggest that molecules that are able to block the binding of KAHRP (or other malaria proteins) to spectrin (or other red cell membrane skeleton proteins) could provide a novel therapeutic approach to mitigate the severity of malaria.