Structure of the C-terminal domains of merozoite surface protein-1 from Plasmodium knowlesi reveals a novel histidine binding site

The protozoan parasite Plasmodium causes malaria, with hundreds of millions of cases recorded annually. Protection against malaria infection can be conferred by antibodies against merozoite surface protein (MSP)-1, making it an attractive vaccine candidate. Here we present the structure of the C-terminal domains of MSP-1 (known as MSP-1(19)) from Plasmodium knowlesi. The structure reveals two tightly packed epidermal growth factor-like domains oriented head to tail. In domain 1, the molecule displays a histidine binding site formed primarily by a highly conserved tryptophan. The protein carries a pronounced overall negative charge primarily due to the large number of acidic groups in domain 2. To map protein binding surfaces on MSP-1(19), we have analyzed the crystal contacts in five different crystal environments, revealing that domain 1 is highly preferred in protein-protein interactions. A comparison of MSP-1(19) structures from P. knowlesi, P. cynomolgi, and P. falciparum shows that, although the overall protein folds are similar, the molecules show significant differences in charge distribution. We propose the histidine binding site in domain 1 as a target for inhibitors of protein binding to MSP-1, which might prevent invasion of the merozoite into red blood cells.     

Merozoite surface protein 1, immune evasion, and vaccines against asexual blood stage malaria

There is an urgent need for a vaccine against malaria and proteins on the surface of the merozoite are good targets for development as vaccine candidates because they are exposed to antibody. However, it is possible that the parasite has evolved mechanisms to evade a protective immune response to these proteins. Merozoite surface protein 1 (MSP-1) is a candidate for vaccine development and its C-terminal sequence is the target of protective antibody. MSP-1 is cleaved by proteases in two processing steps, the second step releases the bulk of the protein from the surface and goes to completion during successful red blood cell invasion. Antibodies binding to the C-terminus of Plasmodium falciparum MSP-1 can inhibit both the processing and erythrocyte invasion. Other antibodies that bind to either the C-terminal sequence or elsewhere in the molecule are 'blocking' antibodies, which on binding prevent the binding of the inhibitory antibodies. Blocking antibodies are a mechanism of immune evasion, which may be based on antigenic conservation rather than diversity. This mechanism has a number of implications for the study of protective immunity and the development of malaria vaccines, emphasising the need for appropriate functional assays and careful design of the antigen.     

Inhibitory and blocking monoclonal antibody epitopes on merozoite surface protein 1 of the malaria parasite Plasmodium falciparum

Merozoite surface protein 1 (MSP-1) is a precursor to major antigens on the surface of Plasmodium spp. merozoites, which are involved in erythrocyte binding and invasion. MSP-1 is initially processed into smaller fragments; and at the time of erythrocyte invasion one of these of 42 kDa (MSP-1(42)) is subjected to a second processing, producing 33 kDa and 19 kDa fragments (MSP-1(33) and MSP-1(19)). Certain MSP-1-specific monoclonal antibodies (mAbs) react with conformational epitopes contained within the two epidermal growth factor domains that comprise MSP-1(19), and are classified as either inhibitory (inhibit processing of MSP-1(42) and erythrocyte invasion), blocking (block the binding and function of the inhibitory mAb), or neutral (neither inhibitory nor blocking). We have mapped the epitopes for inhibitory mAbs 12.8 and 12.10, and blocking mAbs such as 1E1 and 7.5 by using site-directed mutagenesis to change specific amino acid residues in MSP-1(19) and abolish antibody binding, and by using PEPSCAN to measure the reaction of the antibodies with every octapeptide within MSP-1(42). Twenty-six individual amino acid residue changes were made and the effect of each on the binding of mAbs was assessed by Western blotting and BIAcore analysis. Individual changes had either no effect, or reduced, or completely abolished the binding of individual mAbs. No two antibodies had an identical pattern of reactivity with the modified proteins. Using PEPSCAN each mAb reacted with a number of octapeptides, most of which were derived from within the first epidermal growth factor domain, although 1E1 also reacted with peptides spanning the processing site. When the single amino acid changes and the reactive peptides were mapped onto the three-dimensional structure of MSP-1(19), it was apparent that the epitopes for the mAbs could be defined more fully by using a combination of both mutagenesis and PEPSCAN than by either method alone, and differences in the fine specificity of binding for all the different antibodies could be distinguished. The incorporation of several specific amino acid changes enabled the design of proteins that bound inhibitory but not blocking antibodies. These may be suitable for the development of MSP-1-based vaccines against malaria.     

A common cross-species function for the double epidermal growth factor-like modules of the highly divergent plasmodium surface proteins MSP-1 and MSP-8

An understanding of structural and functional constraints on the C-terminal double epidermal growth factor (EGF)-like modules of merozoite surface protein (MSP)-1 and related proteins is of importance to the development of these molecules as malaria vaccines and drug targets. Using allelic replacement, we show that Plasmodium falciparum parasites can invade erythrocytes and grow efficiently in the absence of an MSP-1 protein with authentic MSP-1 EGF domains. In this mutant parasite line, the MSP-1 EGFs were replaced by the corresponding double EGF module from P. berghei MSP-8, the sequence of which shares only low identity with its MSP-1 counterpart. Hence, the C-terminal EGF domains of at least some Plasmodium surface proteins appear to perform the same function in asexual blood-stage development. Mapping the surface location of the few residues that are common to these functionally complementary EGF modules revealed the presence of a highly conserved pocket of potential functional significance. In contrast to MSP-8, an even more divergent double EGF module, that from the sexual stage protein PbS25, was not capable of complementing MSP-1 EGF function. More surprisingly, two chimeric double EGF modules comprising hybrids of the EGF domains from P. falciparum and P. chabaudi MSP-1 were also not capable of replacing the P. falciparum MSP-1 EGF module. Together, these data suggest that although the MSP-1 EGFs can accommodate extensive sequence diversity, there appear to be constraints that may restrict the simple accumulation of point mutations in the face of immune pressure in the field.     

Functional conservation of the malaria vaccine antigen MSP-119across distantly related Plasmodium species

The C-terminal region of Plasmodium falciparum merozoite surface protein 1 (MSP-119) is at present a leading malaria vaccine candidate. Antibodies against the epidermal growth factor-like domains of MSP-1 19are associated with immunity to P. falciparum and active immunization with recombinant forms of the molecule protect against malaria challenge in various experimental systems. These findings, with the knowledge that epidermal growth factor-like domains in other molecules have essential binding functions, indicate the importance of this protein in merozoite invasion of red blood cells. Despite extensive molecular epidemiological investigations, only limited sequence polymorphism has been identified in P. falciparum MSP-119 (refs. 9-11). This indicates its sequence is functionally constrained, and is used in support of the use of MSP-119 as a vaccine. Here, we have successfully complemented the function of most of P. falciparum MSP-119 with the corresponding but highly divergent sequence from the rodent parasite P. chabaudi. The results indicate that the role of MSP-119 in red blood cell invasion is conserved across distantly related Plasmodium species and show that the sequence of P. falciparum MSP-119 is not constrained by function.     

Construction and immunogenicity in mice of attenuated Salmonella typhi expressing Plasmodium falciparum merozoite surface protein 1 (MSP-1) fused to tetanus toxin fragment C

One strategy to develop a multi-antigen malaria vaccine is to employ live vectors to carry putative protective Plasmodium falciparum antigens to the immune system. The 19 kDa carboxyl terminus of P. falciparum merozoite surface protein 1 (MSP-1), which is essential for erythrocyte invasion and is a leading antigen for inclusion in a multivalent malaria vaccine, was genetically fused to fragment C of tetanus toxin and expressed within attenuated Salmonella typhi CVD 908. Under conditions in the bacterial cytoplasm, the fragment C-MSP-1 fusion did not form the epidermal growth factor (EGF)-like domains of MSP-1; monoclonal antibodies failed to recognize these conformational domains in immunoblots of non-denatured protein extracted from live vector sonicates. The MSP-1 was nevertheless immunogenic. One month following intranasal immunization of BALB/c mice with the live vector construct, four out of five mice exhibited > or =four-fold rises in anti-MSP-1 by ELISA (GMT=211); a single intranasal booster raised titers further (GMT=1280). Post-immunization sera recognized native MSP-1 on merozoites as determined by indirect immunofluorescence. These data encourage efforts to optimize MSP-1 expression in S. typhi (e.g. as a secreted protein), so that the EGF-like epitopes, presumably necessary for stimulating protective antibodies, can form.     

Interactions between merozoite surface proteins 1, 6, and 7 of the malaria parasite Plasmodium falciparum

Merozoites of the malaria parasite Plasmodium falciparum expose at their surface a large multiprotein complex, composed of proteolytically processed, noncovalently associated products of at least three genes, msp-1, msp-6, and msp-7. During invasion of erythrocytes, this complex is shed from the surface except for a small glycosylphosphatidylinositol-anchored portion originating from MSP-1. The proteolytic cleavage separating the C-terminal portion of MSP-1 is required for successful invasion. Little is known about the structure and function of the abundant and essential multipartite complex. Using heterologously produced MSP-1, MSP-6, and MSP-7 in precursor and with the exception of MSP-7 in processed form, we have studied in vitro the complex formation between the different proteins to identify the interaction partners within the complex. Both MSP-6(36) and MSP-7 bind only to MSP-1 subunits that are shed, but although MSP-6(36) contacts just subunit p38, MSP-7 interacts with p83, p30, and p38. The intact C-terminal region of MSP-6 is required for the association with p38 as well as for its multimerization into tetramers. Furthermore, our data suggest that only the processed form and not the precursor form of MSP-1 interacts with MSP-6(36). MSP-6- as well as MSP-7-specific rabbit antibodies inhibit parasite multiplication in vitro as shown previously for antibodies directed against MSP-1. Our findings raise interesting questions with regard to proteolysis-mediated mechanisms of maturation of the MSP-1-MSP-6-MSP-7 complex and to the mode by which antibodies directed against this complex interfere with parasite multiplication.     

Crystal structure of a Fab complex formed with PfMSP1-19, the C-terminal fragment of merozoite surface protein 1 from Plasmodium falciparum: a malaria vaccine candidate

Merozoite surface protein 1 (MSP1) is the major protein component on the surface of the merozoite, the erythrocyte-invasive form of the malaria parasite Plasmodium. Present in all species of Plasmodium, it undergoes two distinct proteolytic maturation steps during the course of merozoite development that are essential for invasion of the erythrocyte. Antibodies specific for the C-terminal maturation product, MSP1-19, can inhibit erythrocyte invasion and parasite growth. This polypeptide is therefore considered to be one of the more promising malaria vaccine candidates. We describe here the crystal structure of recombinant MSP1-19 from P.falciparum (PfMSP1-19), the most virulent species of the parasite in humans, as a complex with the Fab fragment of the monoclonal antibody G17.12. This antibody recognises a discontinuous epitope comprising 13 residues on the first epidermal growth factor (EGF)-like domain of PfMSP1-19. Although G17.12 was raised against the recombinant antigen expressed in an insect cell/baculovirus system, it binds uniformly to the surface of merozoites from the late schizont stage, showing that the cognate epitope is exposed on the naturally occurring MSP1 polypeptide complex. Although the epitope includes residues that have been mapped to regions recognised by invasion-inhibiting antibodies studied by other workers, G17.12 does not inhibit erythrocyte invasion or MSP1 processing.     

Epidermal growth factor-like motifs 1 and 2 of Plasmodium vivax merozoite surface protein 1 are critical domains in erythrocyte invasion

Plasmodium vivax merozoite surface protein 1 (PvMSP1) is believed to be important in erythrocyte invasion. However, the detailed mechanism of PvMSP1-mediated invasion has been unclear. We demonstrate that the C-terminal 19 kDa domain (PvMSP119) of PvMSP1, the 42-kDa fragment of PvMSP1 is further cleaved to a 33 kDa N-terminal polypeptide and a 19 kDa C-terminal fragment in a secondary processing step, is a critical domain in the binding between parasite ligand and erythrocyte receptor. Also, its cytoadherence was successfully blocked by naturally acquired immunity, was partially sensitive to neuraminidase and trypsin. When expressed separately epidermal growth factor (EGF)-like motifs 1 and 2, subunits of the PvMSP119, mediated 64% and 66% of the erythrocyte-binding activity, respectively, relative to their expression together as a single intact ligand domain. These results suggest that the EGF-like motifs 1 and 2 of PvMSP119 function as a core-binding portion in the attachment of PvMSP1 to erythrocytes.     

MSP-1 malaria pseudopeptide analogs: biological and immunological significance and three-dimensional structure

Merozoite Surface Protein-1 (MSP-1) has been considered as a malaria vaccine candidate. It is processed during the Plasmodium falciparum invasion process of red blood cells (RBCs). A conserved MSP-1 C-terminal peptide was identified as a high-activity erythrocyte-binding peptide (HAEBP) termed 1585. Since conserved HAEBPs are neither antigenic nor immunogenic we decided to assess the significance of a single peptide bond replacement in 1585. Thus, two pseudopeptides were obtained by introducing a Y[CH2-NH] reduced amide isoster into the 1585 critical binding motif. The pseudopeptides bound to different HLA-DR alleles, suggesting that backbone modifications affect MHC-II binding patterns. Pseudopeptide-antibodies inhibit in vitro parasite RBC invasion by recognizing MSP-1. Each pseudopeptide-induced antibody shows distinct recognition patterns. 1H-NMR studies demonstrated that isoster bonds modulate the pseudopeptides' structure and thus their immunological properties, therefore representing a possible subunit malaria vaccine component.     

Structural requirements for ligand binding by a probable plant vacuolar sorting receptor

How sorting receptors recognize amino acid determinants on polypeptide ligands and respond to pH changes for ligand binding or release is unknown. The plant vacuolar sorting receptor BP-80 binds polypeptide ligands with a central Asn-Pro-Ile-Arg (NPIR) motif. tBP-80, a soluble form of the receptor lacking transmembrane and cytoplasmic sequences, binds the peptide SSSFADSNPIRPVTDRAASTYC as a monomer with a specificity indistinguishable from that of BP-80. tBP-80 contains an N-terminal region homologous to ReMembR-H2 (RMR) protein lumenal domains, a unique central region, and three C-terminal epidermal growth factor (EGF) repeats. By protease digestion of purified secreted tBP-80, and from ligand binding studies with a secreted protein lacking the EGF repeats, we defined three protease-resistant structural domains: an N-terminal/RMR homology domain connected to a central domain, which together determine the NPIR-specific ligand binding site, and a C-terminal EGF repeat domain that alters the conformation of the other two domains to enhance ligand binding. A fragment representing the central domain plus the C-terminal domain could bind ligand but was not specific for NPIR. These results indicate that two tBP-80 binding sites recognize two separate ligand determinants: a non-NPIR site defined by the central domain-EGF repeat domain structure and an NPIR-specific site contributed by the interaction of the N-terminal/RMR homology domain and the central domain.     

Proteome analysis reveals a large merozoite surface protein-1 associated complex on the Plasmodium falciparum merozoite surface

Plasmodium merozoite surface protein-1 (MSP-1) is an essential antigen for the merozoite invasion of erythrocytes. A key challenge to the development of an effective malaria vaccine that can block the erythrocyte invasion is to establish the molecular interaction(s) among the parasite surface proteins as well as with the host cell encoded receptors. In the present study, we applied molecular interactions and proteome approaches to identify PfMSP-1 associated complex on the merozoite surface. Proteomic analysis identified a major malaria surface protein, PfRhopH3 interacting with PfMSP-1(42). Pull-down experiments with merozoite lysate using anti-PfMSP-1 or anti-PfRhopH3 antibodies showed 16 bands that when identified by tandem mass spectrometry corresponded to11 parasite proteins: PfMSP-3, PfMSP-6, PfMSP-7, PfMSP-9, PfRhopH3, PfRhopH1, PfRAP-1, PfRAP-2, and two RAP domain containing proteins. This MSP-1 associated complex was specifically seen at schizont/merozoite stages but not the next ring stage. We could also identify many of these proteins in culture supernatant, suggesting the shedding of the complex. Interestingly, the PfRhopH3 protein also showed binding to the human erythrocyte and anti-PfRhopH3 antibodies blocked the erythrocyte invasion of the merozoites. These results have potential implications in the development of PfMSP-1 based blood stage malaria vaccine.     

The crystal structure of C-terminal merozoite surface protein 1 at 1.8 A resolution, a highly protective malaria vaccine candidate

The C-terminal proteolytic processing product of merozoite surface protein 1 (MSP1) appears essential for successful erythrocyte invasion by the malarial parasite, Plasmodium. We have determined the crystal structure at 1.8 A resolution of a soluble baculovirus-recombinant form of the protein from P. cynomolgi, which confers excellent protective efficacy in primate vaccination trials. The structure comprises two EGF-like domains, and sequence comparisons strongly suggest that the same conformation is present in all species of Plasmodium, including P. falciparum and P. vivax, which are pathogenic in man. In particular, conserved interdomain contacts between the two EGF modules should preserve the compact form of the molecule in all species. Implications of the crystal structure for anti-malarial vaccine development are discussed.     

Crystal structures of the photosystem II D1 C-terminal processing protease

We report here the first three-dimensional structure of the D1 C-terminal processing protease (D1P), which is encoded by the ctpA gene. This enzyme removes the C-terminal extension of the D1 polypeptide of photosystem II of oxygenic photosynthesis. Proteolytic processing is necessary to allow the light driven assembly of the tetranuclear manganese cluster, which is responsible for photosynthetic water oxidation. The X-ray structure of the Scenedesmus obliquus enzyme has been determined at 1.8 A resolution using the multiwavelength anomalous dispersion method. The enzyme is monomeric and is composed of three folding domains. The middle domain is topologically homologous to known PDZ motifs and is proposed to be the site at which the substrate C-terminus binds. The remainder of the substrate likely extends across the face of the enzyme, interacting at its scissile bond with the enzyme active site Ser 372 / Lys 397 catalytic dyad, which lies at the center of the protein at the interface of the three domains.     

Src-mediated tyrosine phosphorylation of NR2 subunits of N-methyl-D-aspartate receptors protects from calpain-mediated truncation of their C-terminal domains

Src-mediated tyrosine phosphorylation of N-methyl-d-aspartate receptor subunits has been shown to modify the functional properties of N-methyl-d-aspartate receptors. Moreover, calpain-mediated truncation of N-methyl-d-aspartate receptor subunits has been found to alter the structure of the receptors. In the present study, we first used immunoprecipitation with a variety of antibodies against N-methyl-d-aspartate receptor subunits and anti-phosphotyrosine antibodies to show that tyrosine-phosphorylated subunits of N-methyl-d-aspartate receptor are protected against calpain-mediated truncation of their C-terminal domains. A GST fusion protein containing the C-terminal domain of NR2A was used to identify the calpain cutting sites in the C-terminal domain. One site was identified at residues 1278-1279, corresponding to one of the preferred calpain truncation sites. This site is adjacent to a consensus sequence for Src-mediated tyrosine phosphorylation, and Src-mediated tyrosine phosphorylation of the GST-NR2A C-terminal fusion protein also inhibited calpain-mediated truncation of the fusion protein. We propose that phosphorylation of NR2 subunits and the resulting inhibition of calpain-mediated truncation of their C-terminal domains provide for the stabilization of the N-methyl-d-aspartate receptors in postsynaptic structures.     

Structure of C-terminal fragment of merozoite surface protein-1 from Plasmodium vivax determined by homology modeling and molecular dynamics refinement

One current vaccine candidate against Plasmodium vivax targeting asexual blood stage is the major merozoite surface protein-1 of P. vivax (PvMSP-1). Vaccine trials with PvMSP-1₁₉ and PvMSP-1₃₃ have succeeded in protecting monkeys and a large proportion of individuals, naturally exposed to P. vivax transmission, develop specific antibodies to PvMSP-1₁₉. This study presents a model for the three-dimensional structure of the C-terminal 19 kDa fragment of P. vivax MSP-1 determined by means of homology modeling and molecular dynamics refinement. The structure proved to be consistent with MSP-1₁₉ of known crystal or solution structures. The presence of a main binding pocket, well suited for protein–protein interactions, was determined by CASTp. Corrections reported to the sequence of PvMSP-1₁₉ Belem strain were also inspected. Our model is currently used as a basis to understand antibody interactions with PvMSP-1₁₉.     

Allelic recombination and linkage disequilibrium within Msp-1 of Plasmodium falciparum, the malignant human malaria parasite

The C-terminal, cysteine-rich 19kDa domain of merozoite surface protein-1 (MSP-1) of Plasmodium falciparum is a target of the host's humoral immunity and thus a malaria vaccine candidate. Although variation in the 19kDa domain is limited among parasite isolates, tertiary structure-dependent intramolecular associations between the 19kDa domain and other parts of MSP-1 are suggested to be involved in immune evasion by allowing competitive binding of protective and non-protective antibodies directed to their epitopes, which are conformationally in close proximity but separated at the primary structure. Since allelic recombination can account for the major variability of the Msp-1 gene, we examined whether linkage disequilibrium occurs between polymorphic loci in the 5'- and the 3'-region, the latter encoding the 19kDa domain. From 184 Thai field isolates, we selected 69 isolates with a single allelic type in six variable blocks of Msp-1 as determined by PCR-based allelic typing. All the isolates showed no evidence of recombination in blocks 6 to 16, whereas recombination was apparent in blocks 2 to 6. Sequencing of the 3'-region revealed two potential recombination sites in block 17. Strong linkage disequilibrium was seen between polymorphic loci in the 5'- and 3'-regions. The strength of this disequilibrium did not correlate with distance between loci. We discuss the possible role of epistatic selection on particular association types (haplotypes) of Msp-1.     

Synthesis, activity, and preliminary structure of the fourth EGF-like domain of thrombomodulin.

The fourth EGF-like domain of thrombomodulin (TM4), residues E346-F389 in the TM sequence, has been synthesized. Refolding of the synthetic product under redox conditions gave a single major product. The disulfide bonding pattern of the folded, oxidized domain was (1-3, 2-4, 5-6), which is the same as that found in EGF protein. TM4 was tested for TM anticoagulant activity because deletion and substitution mutagenesis experiments have shown that the fourth EGF-like domain of TM is essential for TM cofactor activity. TM4 showed no TM-like activity in two assay systems, both for inhibition of fibrin clot formation, and for cofactor activity in thrombin activation of protein C. A preliminary structure of TM4 was determined by 2D 1H NMR from 519 NOE-derived distance constraints. Distance geometry calculations yielded a single convergent structure. The structure resembles the structure of EGF and other known EGF-like domains but has some key differences. The central two-stranded beta-sheet is conserved despite the differences in the number of amino acids in the loops. The C-terminal loop formed by the disulfide bond between C372 and C386 in TM4 is five amino acids longer than the analogous loop between C33 and C42 of EGF protein. This loop appears to have a different fold in TM4 than in EGF protein. The loop forms the two outside strands of a broken, irregular tri-stranded beta-sheet, and amino acids H384-F389 lie between the two strands forming the middle strand of the sheet. Thus, although the C-terminus of EGF protein forms one of the outside strands of a tri-stranded antiparallel sheet, the C-terminus of TM4 forms the inside strand of an irregular tri-stranded parallel-anti-parallel sheet. The residues D349, E357, and E374, which were shown to be critical for cofactor activity by alanine scanning mutagenesis, all lie in a patch near the C-terminal loop, and are solvent accessible. The other critical residues, Y358 and F376, are largely buried and appear to play essential structural rather than functional roles.     

Structure of domain III of the blood-stage malaria vaccine candidate, Plasmodium falciparum apical membrane antigen 1 (AMA1)

Apical membrane antigen 1 of the malarial parasite Plasmodium falciparum (Pf AMA1) is a merozoite antigen that is considered a strong candidate for inclusion in a malaria vaccine. Antibodies reacting with disulphide bond-dependent epitopes in AMA1 block invasion of host erythrocytes by P.falciparum merozoites, and we show here that epitopes involving sites of mutations in domain III are targets of inhibitory human antibodies. The solution structure of AMA1 domain III, a 14kDa protein, has been determined using NMR spectroscopy on uniformly 15N and 13C/15N-labelled samples. The structure has a well-defined disulphide-stabilised core region separated by a disordered loop, and both the N and C-terminal regions of the molecule are unstructured. Within the disulphide-stabilised core, residues 443-447 form a turn of helix and residues 495-498 and 503-506 an anti-parallel beta-sheet with a distorted type I beta-turn centred on residues 500-501, producing a beta-hairpin-type structure. The structured region of the molecule includes all three disulphide bonds. The previously unassigned connectivities for two of these bonds could not be established with certainty from the NMR data and structure calculations, but were determined to be C490-C507 and C492-C509 from an antigenic analysis of mutated forms of this domain expressed using phage display. Naturally occurring mutations in domain III that are located far apart in the primary sequence tend to cluster in the region of the disulphide core in the three-dimensional structure of the molecule. The structure shows that nearly all the polymorphic sites have a high level of solvent accessibility, consistent with their location in epitopes recognised by protective antibodies. Even though domain III in solution contains significant regions of disorder in the structure, the disulphide-stabilised core that is structured is clearly an important element of the antigenic surface of AMA1 recognised by protective antibodies.