Mature parasite-infected erythrocyte surface antigen (MESA) of Plasmodium falciparum binds to the 30-kDa domain of protein 4.1 in malaria-infected red blood cells

During maturation of Plasmodium falciparum within human red blood cells (RBCs), several parasite proteins are synthesized and transported into the RBC cytoplasm where some associate with the RBC membrane skeleton (reviewed in Cooke et al¹ ). One of these proteins, the mature parasite-infected erythrocyte surface antigen (MESA), has a 19-residue sequence (DHLYSIRNYIECLRNAPYI² ) that binds the RBC membrane skeleton protein 4.1 (4.1R).³  Although the precise function of MESA within infected RBCs (IRBCs) is unknown, studies have shown that interruption of the MESA-4.1R interaction results in accumulation of unbound MESA within the RBC cytoplasm and parasite death through an unknown mechanism that may be related to the presence of unbound MESA.⁴  Interruption of this and other protein interactions at the IRBC membrane skeleton in vivo offers as yet unexplored avenues for the development of novel antimalarial therapies. Here, using recombinant 4.1R and MESA proteins, we mapped the region of 4.1R that binds the 19-residue region of MESA and also demonstrated competition between MESA and p55 for 4.1.

Fragments of 4.1R and MESA were amplified by polymerase chain reaction (PCR) as previously described⁵  (Figure 1), either from cDNAs of 4.1R⁸  or appropriate genomic DNA subfragments of the MESA gene.^3,7  The resulting fragments were cloned into the Escherichia coli protein expression plasmids pGEX-KG¹⁰  or pET-31b(+) (Novagen, Darmstadt, Germany), and glutathione S-transferase (GST)– and histidine-tagged fusion proteins were expressed and purified using standard affinity chromatography under native purification conditions.^3,8,11  Additionally, 4.1R was purified from healthy human RBCs.⁹  Purified proteins (Figure 1C) were dialyzed extensively against phosphate-buffered saline (PBS; 0.15 M NaCl containing 10 mM Na₂HPO₄/NaH₂PO₄, pH 7.4) and then used in protein interaction assays using an IAsys resonant mirror biosensor (Affinity Sensors, Cambridge, United Kingdom), as previously described^5,8  or in GST-pull down assays (Figure 2 and Table 1; methods are described in figure legend and table footnote). Approval for the study was obtained from the institutional review boards of Monash University, Tokyo Women's Medical University, and the New York Blood Center. Informed consent was provided according to the Declaration of Helsinki.

Quantitative binding data showed that GST-MF3(S)+19 bound purified 4.1R with moderate affinity (K_D(kin) = 2.7 × 10^–7 M; Table 1). Binding studies using recombinant GST-30 kDa, -16 kDa, -10 kDa, and -22/24 kDa proteins of 4.1R showed that, whereas GST-MF3(S)+19 bound the 30-kDa domain with moderate affinity (K_D(kin) = 1.9 × 10^–6 M), it did not bind the 10-kDa, 16-kDa, or 22/24-kDa domains. To map the MESA binding sequences within the 30-kDa region more precisely, 8 smaller subfragments of the 30-kDa domain were generated (Figure 1B). GST-MF3(S)+19 bound with moderate affinity to GST-30 kDa F5 (K_D(kin) = 1.17 × 10^–6 M) but not to GST-30 kDa F1, F2, F3, F4, F6, F7, or F8 recombinant proteins. These data enabled us to identify the 51 residues of the 30-kDa domain encoded by exon 10 of 4.1R to be responsible for the interaction of 4.1R with MESA. Comparison of the sequences of the nonbinding fragments F4, F6, and F8 with the binding fragment F5 suggests that the binding domain could be further mapped to the central 24 residues of the 4.1R 30-kDa region, although such an assignment would need to be further confirmed by a direct binding assay. All interactions presented here were confirmed, with similar affinities, using cuvettes coated with the synthetic 19-residue MESA peptide (data not shown). No binding of GST-4.1R fusion proteins or purified RBC 4.1R to GST-MF3(S)Δ19, GST-MF4(S) (data not shown), GST, or BSA was detected. Further, although the 16-kDa domain of 4.1R did not bind GST-MF3(S)+19, we found the 30 + 16-kDa fusion protein bound with higher affinity (K_D(kin) = 1.3 × 10^–7 M) to MESA than GST-30 kDa alone (K_D(kin) = 1.9 × 10^–6 M).

Protein 4.1R functions to stabilize horizontal protein interactions between spectrin tetramers and actin filaments in the RBC membrane skeleton, as well as participating in several interactions that link the underlying spectrin-based membrane skeleton to the lipid membrane, including the 4.1R-glycophorin C (GPC)–p55 and 4.1R-band 3 interactions (reviewed in Gascard and Mohandas¹³  and Pinder et al¹⁴ ). In these and other membrane skeleton interactions, 4.1R has been identified as an important modulator of protein binding affinities. 4.1R has been shown to modulate the 4.1R-GPC-p55 and spectrin-actin–4.1R ternary complexes, by increasing the affinity of the p55-GPC⁸  and spectrin-actin interactions.¹⁵  Several studies have identified the specific residues in the 30-kDa domain of 4.1R that bind to GPC and p55.^8,16  The GPC binding site in 4.1R was mapped to 40 residues encoded by exon 8 of the 4.1R gene, whereas the region of 4.1R that binds to p55 was mapped to a 33-residue region encoded by exon 10.⁸  Our findings show that there is an overlap between the binding domains of MESA and p55 on the 4.1R protein. Recently, the 3D crystal structure of the 30-kDa domain (also known as the FERM [4.1-ezrin-radaxin-moesin] domain¹⁷ ) has been resolved (Figure 2A,¹² ), in which the exon 10 encoded sequences are located in the C-lobe of the 30-kDa domain. The location of the MESA binding domain in this structure is also shown. On the basis of our earlier findings, we hypothesized that binding of MESA to 4.1R may act as a competitive inhibitor in the p55-4.1R interaction. GST-pull down assays, using GST-4.1R 80 kDa coupled to glutathione Sepharose beads were performed to assess the binding of MESA to 4.1R in the absence or presence of p55. These data showed greatly reduced binding of MESA to 4.1R with increasing concentrations of p55 (Figure 2B). The competitive interaction of MESA and p55 for 4.1R may, therefore, result in modulation of the ternary complex between 4.1R-GPC-p55 and alter the stability of the membrane skeleton of P falciparum IRBCs.

Interestingly, fusion of the 30-kDa and 16-kDa regions of 4.1R enhanced the affinity of the interaction between 4.1R and MESA by an order of magnitude. Sequences that have little or no direct binding ability, but enhance the affinity of other binding sequences, perhaps through altered protein conformation, have previously been reported. For example, sequences from the repeat regions of the P falciparum circumsporozoite protein (CSP) have been shown to enhance the affinity of binding of CSP-derived peptides with human hepatoma cell lines.¹⁸  This, however, is the first report in which the affects of the 16-kDa domain of 4.1R on binding of the 30-kDa domain with a partner protein have been described.

The death of P falciparum when cultured in 4.1R-deficient RBCs is a clear indication of the extreme importance of the MESA-4.1R interaction for parasite survival.⁴  Here, we show data that MESA binds to a 51-residue region of the 4.1R 30-kDa domain and also that MESA and p55 compete for interaction with 4.1R. The death of parasites in 4.1R-deficient RBCs may, therefore, result from the accumulation of toxic levels of unbound MESA in the IRBC cytoplasm because of MESA's inability to associate with 4.1R in 4.1R-deficient RBCs⁴  and/or the subsequent inability of the parasite to correctly modulate the protein interactions that form the RBC membrane skeleton and facilitate parasite survival in vivo. Although the precise function of MESA at the membrane skeleton of IRBCs is unresolved (parasites that do not express MESA [resulting from spontaneous deletion of chromosome 5] cultured in normal human RBCs show that MESA is not critically required for growth⁴  or cytoadhesion¹⁹  in vitro), the experiments of Magowan et al⁴  have conclusively demonstrated that interruption of the MESA-4.1R interaction results in parasite death. Consequently, increased understanding of the MESA-4.1R interaction and its potential interruption in vivo may offer an as yet unexplored avenue for the future development of novel antimalarial therapies.

We thank Dr Philippe Gascard and Ms Marilyn Parra (Life Sciences Division, Lawrence Berkeley Laboratories, Berkeley, CA) and Ms Lisa M. Stubberfield (Department of Microbiology, Monash University, Melbourne, Victoria, Australia) for the gift of plasmids.