Invasion of erythrocytes by Plasmodial merozoites is a composite process involving the interplay of several proteins. Among them, the Plasmodium falciparum Cysteine-Rich Protective Antigen (PfCyRPA) is a crucial component of a ternary complex, including Reticulocyte binding-like Homologous protein 5 (PfRH5) and the RH5-interacting protein (PfRipr), essential for erythrocyte invasion. Here, we present the crystal structures of PfCyRPA and its complex with the antigen-binding fragment of a parasite growth inhibitory antibody. PfCyRPA adopts a 6-bladed β-propeller structure with similarity to the classic sialidase fold, but it has no sialidase activity and fulfills a purely non-enzymatic function. Characterization of the epitope recognized by protective antibodies may facilitate design of peptidomimetics to focus vaccine responses on protective epitopes. Both in vitro and in vivo anti-PfCyRPA and anti-PfRH5 antibodies showed more potent parasite growth inhibitory activity in combination than on their own, supporting a combined delivery of PfCyRPA and PfRH5 in vaccines.https://doi.org/10.7554/eLife.20383.001
Malaria is one of the deadliest infectious diseases worldwide, killing over 400,000 people a year. About 200 million people are infected every year, placing a huge social and medical burden especially on developing countries. Microscopic parasites known as Plasmodium are responsible for causing this disease. Plasmodium parasites have a complex life cycle involving both mosquito and mammal hosts. This includes a stage where the parasites infect the mammal’s red blood cells, which causes the symptoms of the disease. In 2012, a team of researchers discovered that a protein called CyRPA forms a group (or ‘complex’) with several other proteins to allow the parasites to enter red blood cells.
Developing a vaccine is one of the most promising approaches to prevent malaria. Vaccines help the body to recognise and fight an invading microbe by triggering an immune response that results in the production of proteins called antibodies, which can bind to specific molecules on the surface of the microbe. If the microbe later enters the body, these antibodies can be produced quickly to eliminate the microbe before it causes disease. However, efforts to develop a highly effective vaccine against malaria have so far been unsuccessful.
Favuzza et al. – including some of the researchers involved in the 2012 work – used a technique called X-ray crystallography to investigate the three-dimensional structure of the CyRPA protein. The experiments show that an antibody is able to bind to a region of CyRPA – a designated ‘protective epitope’ – that is similar in the CyRPA proteins of all Plasmodium falciparum strains. These antibodies can prevent the parasite from entering the red blood cells, and vaccines containing CyRPA may therefore be effective at protecting individuals from malaria. The findings of Favuzza et al. also suggest that using CyRPA in combination with another protein in the complex called RH5 could make the vaccine more powerful as it would make it harder for the parasite to become resistant.
The next step following on from this work is to design a vaccine containing protective CyRPA epitopes that triggers an immune response in mammals that is strong enough to reduce the numbers of parasites in the blood. A future challenge will be to develop a vaccine that combines several proteins involved in different stages of the parasite’s life cycle to provide full protection against malaria.https://doi.org/10.7554/eLife.20383.002
According to the World Health Organization 2015 Malaria Report (who.int/malaria/publications/world_malaria_report/en), malaria is estimated to have caused 214 million clinical cases and 438,000 deaths in 2015. The disease is transmitted by female Anopheles mosquitoes and caused by parasitic protozoans of the genus Plasmodium, of which P. falciparum and P. vivax are the most prevalent and P. falciparum is causing the most often fatal and medically most severe form of malaria. Debilitating clinical symptoms associated with the infection are caused by the multiplication of the asexual blood-stage parasites in erythrocytes. One of the most promising targets for malaria vaccine development is therefore at the stage where merozoites invade erythrocytes.
Invasion of host erythrocytes by merozoites is a complex process, conceptually divisible into four phases: (1) initial recognition of and reversible attachment to the erythrocyte membrane by the merozoite; (2) junction formation leading to irreversible attachment of the merozoite, parasitophorous vacuole formation, and release of the Plasmodium rhoptry-microneme secretory organelles; (3) invagination of the erythrocyte membrane around the merozoite, accompanied by the shedding of the merozoite’s surface coat; (4) closing of the parasitophorous vacuole and resealing of the erythrocyte membrane mark the completion of merozoite invasion (Pinder et al., 2000). The initial recognition and the active invasion of erythrocytes depend on specific molecular interactions between parasite ligands and receptors on the host erythrocyte membrane. Although several ligand-receptor interactions have already been identified, the entire network of molecular interactions involved in invasion is not yet fully disentangled. In addition, P. falciparum merozoite proteins are antigenically highly diverse and in part functionally redundant, to facilitate parasite escape from host immune surveillance and to ensure erythrocyte invasion via alternative pathways (Cowman et al., 2012).
Most efforts in malaria blood stage vaccine research and development have historically concentrated on immuno-dominant, polymorphic antigens that contribute to the invasion of red blood cells by merozoites. Despite major efforts, blood-stage vaccines based on merozoite surface antigens have so far shown limited efficacy in clinical trials (reviewed in Halbroth and Draper, 2015). Extensive antigenic polymorphism represents one major hurdle for the development of an effective blood-stage malaria vaccine (Takala et al., 2009; Dzikowski and Deitsch, 2009). Therefore, the identification of new candidate antigens that are able to induce broad strain-transcending immunity and that are not susceptible to ‘vaccine resistance’ has become a recent research focus.
Availability of pathogen genomes is facilitating the discovery of novel vaccine candidate antigens through ‘reverse vaccinology’ approaches (Rappuoli, 2001; Donati and Rappuoli, 2013). Sequencing and annotation of the P. falciparum genome (Gardner et al., 2002) has supported the identification of new blood-stage vaccine candidate antigens (Conway, 2015; Proietti and Doolan, 2014), among which the P. falciparum Cysteine-Rich Protective Antigen (PfCyRPA) has a number of noteworthy properties. While PfCyRPA is highly conserved among a plethora of P. falciparum isolates, it also is poorly immunogenic in the context of natural exposure (Dreyer et al., 2012). Moreover, PfCyRPA-specific monoclonal antibodies (mAb) inhibit parasite growth both in vitro and in vivo by blocking merozoite invasion (Dreyer et al., 2012; Favuzza et al., 2016).
PfCyRPA is a 42.8 kDa protein of 362 residues with a predicted N-terminal secretion signal. Orthologs of PfCyRPA have been found in the genomes of the human malaria parasite P. vivax and the primate pathogens P. knowlesi, P. cynomolgi, and P. reichenowi (Figure 3—figure supplement 1), but not in the sequenced genomes of other Plasmodium species. P. falciparum PfCyRPA shares on average 42% sequence identity with its orthologs, but within different P. falciparum isolates PfCyRPA is highly conserved: just 13 dimorphic amino acid positions (highlighted in Figure 3—figure supplement 1) were found in 227 P. falciparum field isolates (Manske et al., 2012), and only a single variant (Arg399 instead of Ser399) was found at a frequency of greater than 2%.
PfCyRPA is part of a multi-protein complex (Reddy et al., 2015; Volz et al., 2016) including also the PfRH5-interacting protein PfRipr and the reticulocyte binding-like homologous protein PfRH5, which binds to the erythrocyte receptor basigin (Baum et al., 2009; Crosnier et al., 2011; Chen et al., 2011b, 2014). PfRH5, PfCyRPA, and PfRipr colocalize during parasite invasion at the junction between merozoites and erythrocytes. The complex seems to be required both for triggering Ca2+ release and establishment of tight junctions (Volz et al., 2016). While merozoites deficient in PfCyRPA or PfRH5 can still bind to erythrocytes, they do not attach irreversibly and cannot invade the host cells (Volz et al., 2016). Like PfCyRPA (Dreyer et al., 2012), PfRH5 induces invasion-blocking antibodies that are effective across common genetic variants (Douglas et al., 2011; Bustamante et al., 2013; Douglas et al., 2014).
‘Structural vaccinology’, a combination of immunological, structural, and bioinformatics approaches, is increasingly used for the design of improved vaccine antigens (Dormitzer et al., 2008; Cozzi et al., 2013; Malito et al., 2015). To this end, the crystal structures of PfRH5 in complex with basigin and neutralizing inhibitory mAb have been determined (Chen et al., 2014; Wright et al., 2014). Here, we describe the crystal structure of the promising vaccine candidate PfCyRPA alone and in complex with the antigen-binding fragment (Fab) of the parasite growth inhibitory mAb c12 (Dreyer et al., 2012). The structure of PfCyRPA represents a step toward elucidating its biological function. Furthermore, definition of the specific epitope–paratope interactions from the crystal structure of the PfCyRPA/c12 complex will support rational design of an epitope-focused PfCyRPA-based candidate vaccine.
The finding that PfCyRPA and PfRH5 form a complex essential for parasite invasion prompted us to investigate the fine specificities of previously generated parasite inhibitory and non-inhibitory anti-PfCyRPA mAbs. The 16 anti-PfCyRPA mAbs available for analysis (Dreyer et al., 2012; Favuzza et al., 2016), showed six distinctive reactivity patterns with seven overlapping recombinant protein fragments of PfCyRPA, assigning them to the epitope groups A – F, with groups A, B, C, and F comprising the parasite inhibitory and groups D and E the non-inhibitory mAbs (Figure 1). All mAbs bound to the full-length PfCyRPA (without the signal sequence; fragment 1). mAbs belonging to epitope group A exclusively bound this fragment, indicating that they recognize conformational epitopes not present in any of the shorter PfCyRPA sequence stretches. Lack of binding to fragments 2 and 7 indicates that the epitope may comprise sequences from both ends of the polypeptide chains. Epitope group B antibodies, including mAb c12, bound only to fragments 1, 2, and 3. Epitope group F mAbs bound to fragments 1, 2, and 5, but not to fragment 3, indicating that in contrast to group B, residues located in the sequence stretch between aa 181–251 are required for their binding. The single mAb c04 constitutes the epitope group C showing binding to fragment 7 (only in IFA, not confirmed by Western blotting analysis). The non-inhibitory mAbs clustered into the distinct epitope groups D and E.
Since it may prove useful to incorporate a combination of PfCyRPA and PfRH5 in a multivalent malaria vaccine, we investigated whether mAbs against these two vaccine candidate antigens have additive or synergistic effects. In a first step, we tested a combination of inhibitory mAbs against the two antigens in an in vitro parasite growth inhibition assay. Parasites were cultured for one cycle of merozoite invasion in the presence of the anti-PfCyRPA c12 mAb with or without the anti-PfRH5 BS1.2 mAb at concentrations of 500, 250, and 125 μg/mL. Either mAbs showed potent inhibitory activity, consistently reducing parasite growth of all four tested P. falciparum strains in a concentration-dependent manner and to the same extent as the well characterized inhibitory anti-MSP-1 mAb 12.10 (Blackman et al., 1990) (Figure 2A and Figure 2—figure supplement 1). When combining the anti-PfCyRPA c12 mAb with the anti-RH5 BS1.2 mAb, we found a significantly enhanced inhibitory activity: while mAbs c12 and BS1.2 at a concentration of 250 µg/mL inhibited growth by 21% ± 2.2% and 31% ± 4.6%, respectively, the combination of both mAbs (250 µg/mL each) inhibited growth by 59% ± 1.4%; (Figure 2A). The functional activity of both mAbs was not enhanced by the addition of a malaria-unrelated control mAb.
In a second step, the in vivo parasite inhibitory activity of the mAbs was evaluated in the P. falciparum SCID murine model that employs non-myelodepleted NODscidIL2Rγnull mice engrafted with human erythrocytes (Dreyer et al., 2012; Jiménez-Díaz et al., 2009). Groups of three mice received 2.5 or 0.5 mg of mAbs c12 or BS1.2 or a combination of both mAbs by i.v. injection. The control groups received either 2.5 mg of an isotype-matched malaria-unrelated mAb or the same volume of PBS without Ab. Mice were infected with parasitized erythrocytes 1 day after the antibody transfer and parasitemia was subsequently monitored (Figure 2B). In the control groups, parasitemia reached 19.6% ± 0.8% on day 9 after mAb injection. Parasitemia in mice having received 2.5 mg c12 or BS1.2 mAb increased only marginally, reaching 2.2 ± 0.5 and 2% ± 0.3% on day nine after mAb injection, respectively. At the lower dose of 0.5 mg c12 and BS1.2 inhibited parasite growth to 10.1 ± 2.3 and 10.8% ± 6.9% parasitemia, respectively (Figure 2B). In accord with the in vitro data, parasitemia decreased significantly (p=0.0356; unpaired t test, 95% confidence interval, two-tailed) and reached only 4.8% ± 1.9% on day 9 if mice received 0.5 mg of the anti-PfRH5 BS1.2 mAb in addition to 0.5 mg of the anti-PfCyRPA c12 mAb.
These results demonstrated that anti-PfCyRPA and anti-PfRH5 antibodies have an additive parasite growth inhibitory effect, justifying the combination of both antigens in a subunit vaccine. While the structure of PfRH5 in complex with an inhibitory antibody has been determined, structural information for PfCyRPA is lacking. We therefore determined the crystal structures of PfCyRPA and its complex with the Fab of the growth inhibitory mAb c12.
The far-UV CD spectrum of PfCyRPA is consistent with an all-β structure connected by loop regions and the absence of α-helices (Figure 3—figure supplement 2A). Mass spectrometric analysis of proteolytic fragments of PfCyRPA revealed at least four disulfide bonds that are sequential along the sequence (Figure 3—figure supplement 2B). Intrinsic fluorescence showed that the two Trp residues present in PfCyRPA are buried in the native state. Also, the disulfide bonds seem to be buried, because addition of 50 mM reducing agent had no significant effect on the fluorescence of PfCyRPA (Figure 3—figure supplement 2C). Taken together, these data are consistent with PfCyRPA forming a compact, disulfide-stabilized molecule of predominantly β-sheet structure.
In order to crystallize PfCyRPA, we needed to pre-treat the protein with Actinase E (Figure 4—figure supplement 1A). The crystal structure of PfCyRPA, determined to a resolution of 2.5 Å (detailed in Supplementary file 1), confirmed our biophysical analyses (Figure 3). PfCyRPA adopts a six-bladed β-propeller structure that buries the disulfide bonds and the Trp residues. Each blade of the propeller is constructed by a four-stranded anti-parallel β-sheet (Figure 3B). The five disulfide bonds in PfCyRPA are located within blades 2–6 (Figure 3A), stabilizing each individual blade. The first blade has no disulfide bond; it is formed by β-strands from the N- and C-terminal regions of PfCyRPA, potentially enabling conformational changes in PfCyRPA by opening and closing. A domain alignment search (DALI; [Holm and Rosenström, 2010]) for related structures revealed that PfCyRPA adopts a heavily modified sialidase/neuraminidase fold. The closest structural relative is the catalytic domain of Vibrio cholerae sialidase (Moustafa et al., 2004) (Figure 3C). The two proteins have only 9% sequence identity and the structures have a large root mean square distance (rmsd) of 3.7 Å over 285 residues, clearly showing that while the overall fold is similar, the structures are very different with respect to the inclination of the blades (Figure 3C) and the length and conformations of the surface loops connecting the β-strands (Figure 3D). PfCyRPA also contains a signature sequence motif for sialidases known as an Asp-box (201-SHDKGETW-208; conserved residues are underlined) (Roggentin et al., 1989), which serve structural roles in the β-propellers of sialidases. However, while bacterial sialidases contain between three to five Asp-boxes, PfCyRPA contains only a single one.
Two molecules of PfCyRPA are present in the asymmetric unit and have conformational differences in several surface loops, suggesting possible flexibility of these loops in solution (Figure 3B). From the point of peptidomimetics that could be derived from PfCyRPA for vaccine development, these surface loops connecting the blades on the back and front of the β-propeller are natural candidates. Although some of the loops, e.g. in blade 5 (arrows in Figure 3B), have significant structural plasticity, it should be possible to stabilize them in a conformation suitable to raise an immune response.
The sialidase fold of PfCyRPA and the presence of an Asp-box motif raised the question whether PfCyRPA exhibits sialidase activity. In view of the involvement of viral and microbial sialidases in the unmasking of cryptic host ligands, host cell adhesion, and invasion (Chen et al., 2011a; Lewis and Lewis, 2012; Matrosovich et al., 2015), sialidase activity could make sense for PfCyRPA-mediated invasion of erythrocytes. We detected no sialidase activity, however, when we tested recombinant PfCyRPA for neuraminidase activity using a colorimetric assay (Figure 3—figure supplement 3). The active site of sialidases usually contains a triad of Arg residues that bind to the substrate, a Glu/Tyr pair where the acid acts as a general base to activate the Tyr nucleophile, and a hydrophobic pocket with a conserved Trp that accommodates the acetyl group of sialic acid (Buschiazzo and Alzari, 2008). In addition, many sialidases bind to Ca2+ ions, of which one is close to the active site and is bound by three main-chain carbonyl groups and oxygen atoms in the side-chains of Asn, Thr, and Asp. PfCyRPA contains none of the residues necessary for catalysis, nor does it harbor a Ca2+ binding site, providing structural correlates of the absence of sialidase activity in PfCyRPA. PfCyRPA and sialidases may have evolved from a common ancestor, or PfCyRPA could have evolved from a genuine sialidase to adopt other functionalities.
The parasite inhibitory anti-PfCyRPA c12 mAb binds tightly to PfCyRPA with a Kd of ca. 1 nM as determined by surface plasmon resonance analysis (Dreyer et al., 2012). This mAb recognizes PfCyRPA independent of its glycosylation, as revealed by Western blot analyses (Figure 4A). Because of these favorable properties, we chose c12 for epitope mapping by determining the structure of a PfCyRPA/c12 complex. In a first step, we determined the crystal structure of c12 in isolation (Figure 4B and C). We obtained three different crystal forms containing a total of four crystallographically independent Fab molecules. Superposition of the structures showed very little structural plasticity of the CDR loops (Figure 4C), suggesting that they retain their structures upon epitope binding.
We then purified the PfCyRPA/c12 complex, analyzed it by limited proteolysis, and crystallized it. Actinase E treatment of the PfCyRPA/c12 complex resulted in the same proteolysis pattern as observed for PfCyRPA alone (Figure 4—figure supplement 1B), suggesting that the epitope for mAb c12 is distant from the Actinase E recognition site at Asp189. Actinase E treatment was not necessary to crystallize the PfCyRPA/c12 complex, however. We determined the crystal structure of the PfCyRPA/c12 complex at a resolution of 4.0 Å by molecular replacement using the individual high-resolution structures of c12 and PfCyRPA as search models (Figure 5A). Despite the limited resolution, novel molecular features were visible in the electron density maps of the complex, providing confidence in the relative orientation of PfCyRPA and c12 (Figure 5—figure supplement 1). First, the electron density visible for PfCyRPA after molecular replacement with c12 was used as a search model for structure determination of PfCyRPA alone (see Materials and methods). This strategy would have been impossible had the placement of c12 been wrong. Second, a loop region that is absent in the PfCyRPA search model due to Actinase E proteolysis (see above) exhibits omit electron density in the PfCyRPA/c12 complex (Figure 4—figure supplement 1). Third, consistent with the similar conformations of the CDR loops in the individual c12 Fab structures (Figure 4) there are negligible conformational changes in the c12 CDR loops when bound to PfCyRPA. Similarly, the epitope conformation in PfCyRPA is very similar in the unbound and complexed form. Minor adjustments of side-chains were required during rebuilding of the complex structure. While the resolution of the PfCyRPA/c12 complex is limited and many side-chains lack clear electron density, as is typical for this resolution, knowledge of the relative orientation of PfCyRPA with respect to c12 in the complex is sufficient for designing peptidomimetics to target the immune response to the protective epitope.
The epitope recognized by mAb c12 is a surface composed of blade two and part of blade 3 of PfCyRPA (Figure 5A). The most frequent amino acid dimorphism of PfCyRPA at position 339 is thus located outside the epitope recognized by mAb c12, consistent with the observation that mAb c12 binds to P. falciparum independent of the PfCyRPA variant they express (Dreyer et al., 2012). mAb c12 buries a total surface area of 950 Å2 on PfCyRPA with the major contributor being the light chain, which buries 520 Å2, while the heavy chain buries only 430 Å2. The surface complementarity coefficient Sc of the complex is 0.67, a typical value for antibody-antigen interactions (Lawrence and Colman, 1993); a value of one would denote perfect complementarity. The light chain also forms more potential hydrogen bonds to PfCyRPA than the heavy chain (Figure 5B). Seven hydrogen bonds are possible between PfCyRPA and the c12 light chain. Each side-chain of light chain CDR1 residues 50-RND-52 can form a hydrogen bond with the side-chains of PfCyRPA residues Asp66, Asn45, and Tyr78, respectively. A particularly large contribution to complex stability may stem from a salt bridge between PfCyRPA residue Asp66 and CDR1 residue Arg50, which by mutagenesis was found to be true (see below), a further confirmation that the PfCyRPA/c12 complex structure is correct. In addition, the heavy chain contributes significantly to antigen binding. Tyr120 of heavy chain CDR3 can engage in two hydrogen bonds with the side-chains of PfCyRPA residues Asp92 in blade two and Tyr144 in blade 3, thus establishing a mini-network that bridges blades 2 and 3. Together with heavy chain CDR3 residue Tyr122, Tyr120 also forms numerous van der Waals interactions with Asn91, Lys95, Glu96, and Phe136, Tyr144, Asn145, and Asn146 (Figure 5B and C). A complete listing of the derived potential interactions is available in Supplementary file 2, which due to the limited resolution of the complex structure must remain tentative. What is clear from the structure is that mAb c12 recognizes a discontinuous epitope that contains at least 17 residues distributed over four PfCyRPA sequence stretches (marked in Figure 3A).
The epitope extracted from the complex structure matches the binding pattern of mAb c12 to the PfCyRPA fragments in our epitope analysis (Figure 1): only fragments 1–3 containing all the seventeen interaction sites were recognized by mAb c12. Furthermore, fragment 4, lacking three out of these 17 residues (Tyr144, Asn145, and Asn146) and fragments 5 and 6, lacking four residues (Asn45, Val46, Pro47, and Asp66), were not recognized by mAb c12.
The PfCyRPA epitope recognized by mAb c12 was further verified by mutagenesis. The crystal structure suggests that Asp66 forms a salt bridge with Arg50 in the CDR1 loop of the c12 light chain (Supplementary file 2 and Figure 5—figure supplement 1). When Asp66 was changed to Lys, a capture ELISA detected significantly weaker binding of the mAb c12 to the Asp66Lys variant of PfCyRPA (Figure 6A). As expected, the Asp66Lys replacement did not affect binding of the control mAb SB1.6 (Figure 6B), which belongs to the unrelated epitope group F and is expected to engage in crucial interactions with PfCyRPA residues located between residues 181 and 251 (Figure 1).
PfCyRPA has been identified as a novel malaria blood-stage vaccine target in an endeavor to test predicted P. falciparum open reading frames for the capacity to elicit parasite-inhibitory mAbs (Dreyer et al., 2010). PfCyRPA is stage-specifically expressed in late schizonts and elicits mAbs that inhibit parasite growth in vitro and in a P. falciparum experimental infection model based on NODscidIL2Rγnull mice engrafted with human erythrocytes (Dreyer et al., 2012). Recent studies (Reddy et al., 2015) have placed PfCyRPA in a ternary complex with PfRH5 and PfRipr that as a whole is essential for erythrocyte invasion (Volz et al., 2016). Like PfRH5 and PfRipr (Baum et al., 2009; Chen et al., 2011b), PfCyRPA is refractory to genetic disruption (Reddy et al., 2015), and loss of PfCyRPA or PFRH5 function in conditionally expressing mutants blocks parasite growth due to the inability of merozoites to invade erythrocytes (Volz et al., 2016). Here, we show that antibodies against PfRH5 and PfCyRPA have additive parasite growth inhibitory activity. Inclusion of both antigens into a multivalent subunit vaccine therefore represents an attractive strategy to prevent emergence of escape mutants. While an ortholog of PfCyRPA is also present in the genome of the human parasite P. vivax, an ortholog of PfRH5 has only been found so far in the genome of P. reichenowi, a chimpanzee malaria parasite closely related to P. falciparum (Otto et al., 2014). This may indicate a function of PfCyRPA beyond its role as a protein-binding platform in the ternary complex with PfRH5 and PfRipr. While it has been claimed that PfCyRPA anchors the ternary complex through a GPI anchor to the parasite membrane (Reddy et al., 2015), recent results are not consistent with this suggestion (Volz et al., 2016). This inconsistency suggests that additional proteins might be involved in anchoring the complex to the parasite membrane and that PfCyRPA has an essential function other than that of an anchor. Further functional studies are required to elucidate the molecular mechanisms how the ternary complex is involved in the triggering of Ca2+ release and the establishment of tight junctions between the merozoite and the erythrocytes (Volz et al., 2016). In this context, it may be relevant that the five sequential disulfide bonds of PfCyRPA are in blades 1–5, leaving blade 6 of the β-propeller structure without such a local stabilization. Blade six is constructed by β-strands from the N- and the C-terminus, and in principle could act as a gate to allow PfCyRPA to undergo substantial conformational changes.
Structural vaccinology defines epitopes from antigen-antibody complex crystal structures and thus allows the design of surrogate antigens that elicit protective humoral immune responses. The approach may be of particular use for pathogens that cause chronic infections by presenting immuno-dominant antigens and epitopes to the immune system, which, however, only elicit non-protective antibody responses (Liljeroos et al., 2015; Loomis and Johnson, 2015; Robinson, 2013; Dormitzer et al., 2012). The structural studies of Neisseria meningitidis adhesin A and factor H binding protein (Malito et al., 2013), as well as of the Staphylococcus aureus manganese transport protein (Ahuja et al., 2015) represent remarkable examples in which crystal structure determination of antigens elucidated both the molecular mechanism of their biological functions and their immunological viability as vaccine antigens.
For the development of synthetic malaria subunit vaccines, conformationally stabilized and structurally optimized peptide antigens mimicking conserved protective epitopes (Mueller et al., 2003; Okitsu et al., 2007; James et al., 2006) may be more suitable components than entire recombinant proteins containing highly polymorphic immuno-dominant non-protective surface loops. Both PfCyRPA and PfRH5 elicit protective as well as non-protective antibodies (Dreyer et al., 2012; Douglas et al., 2011), making them candidates for structure-based approaches, that aim to focus the immune responses on protective epitopes.
Even at the comparatively moderate resolution of 4 Å, the crystal structure of the protective mAb c12 bound to PfCyRPA provides an ideal starting point for the design of a conformationally stabilized PfCyRPA-derived peptide antigen to elicit primarily protective antibodies. Particle-based antigen delivery systems such as virosomes (Cech et al., 2011; Pluschke and Tamborrini, 2012) are highly suitable to support, by repetitive epitope display, the development of strong immune responses against peptidomimetics. The epitope recognized by mAb c12 is monomorphic and, like PfCyRPA in general, little immunogenic in the natural context. Focusing the vaccine-induced immune response on this PfCyRPA epitope may confer a strong protective effect.
Malaria vaccine candidates based on single-antigens have been uniformly unsuccessful and an effective subunit vaccine will likely have to include antigens from more than one stage of the parasite life cycle (pre-erythrocytic, liver stage, and blood stage components), as well as sexual blood-stage gametocyte or mosquito-stage parasite antigens as transmission-blocking components. The Plasmodium genome contains more than 5000 open reading frames of which only <0.5% have been tested as vaccine candidates, so careful selection of new targets for an optimized multi-valent vaccine is called for (Proietti and Doolan, 2014; Dups et al., 2014; Longley et al., 2015; Wu et al., 2015). PfCyRPA represents one of the most promising new blood-stage vaccine candidate antigens identified with the support of –omics approaches, and there is renewed hope that an effective combination and formulation of such rationally selected antigens can lead to the development of a malaria subunit vaccine that offers high level strain-transcending protection (Halbroth and Draper, 2015). With the availability of structural analyses for both PfCyRPA and PfRH5 in complex with an inhibitory antibody, a structure-based approach is now possible. In view of the observed additive effect of antibodies, inclusion of immunogens representing both antigens in a subunit vaccine would make sense, thereby reducing the danger of selection of immune escape mutants.
Plasmids were amplified in E. coli strain Top10 (Life Technologies) grown in LB medium under 100 μg/mL ampicillin selection. The expression vector for secretion of recombinant PfCyRPA was generated by PCR using the plasmid BVM_PFD1130W_FLAG_GP_His as template (Dreyer et al., 2010). The resulting expression vector pcDNA3.1_BVM_CyRPA(26-362)_His6 encodes the bee-venom melittin (BVM) signal sequence to secrete PfCyRPA into the cultivation medium, and a C-terminal His6-tag. Three asparagine residues (N145, N322, and N338) were predicted to be potential sites for N-glycosylation of PfCyRPA in mammalian cells. Since Plasmodium proteins are not glycosylated (Dieckmann-Schuppert et al., 1992), an expression vector coding for the PfCyRPA triple variant N145Q/N322Q/N338Q comprising residues 29–362 and containing no N-glycosylation sites was derived from the first vector by site-directed mutagenesis (GenScript). The expression vector coding for the PfCyRPA Asp66Lys single amino acid variant (D66K) was also generated by site-directed mutagenesis (GenScript) resulting in the expression plasmid pcDNA3.1_BMV_CyRPA(29–362/D66K)_His6.
Plasmodium falciparum strains 3D7, K1, 7G8 and D6 were obtained from the Malaria Research and Reference Reagent Resource Center (MR4; Manassas, VA, USA) (MRA-102,–159, −154 and −285, respectively). Parasites were cultured essentially as described previously (Matile and Pink, 1990). The culture medium was supplemented with 0.5% AlbuMAX (Life Technologies) as a substitute for human serum (Dorn et al., 1995). Cultures were synchronized by sorbitol treatment (Lambros and Vanderberg, 1979). Erythrocytes for passages were obtained from the Swiss Red Cross (Switzerland).
In vitro growth inhibition assays with P. falciparum strains 3D7, K1, 7G8 and D6 were conducted essentially as described (Persson et al., 2006). Each culture (trophozoite stage parasites, 0.5% hematocrit, 2% parasitemia) was set up in triplicate in 96-well flat-bottomed culture plates. After 48 hr of incubation (one cycle of merozoite invasion), viable parasites were stained with hydroethidine and analyzed in a FACSscan flow cytometer (Becton Dickinson) using CellQuest software. A total of 50,000 cells per sample were analyzed. Percent inhibition was calculated from the mean parasitemia of triplicate test and control wells as:
Parasitemia of control samples was also determined by counting GIEMSA stained parasites.
All procedures involving living animals were performed in strict accordance with the Rules and Regulations for the Protection of Animal Rights (Tierschutzverordnung) of the Swiss Federal Food Safety and Veterinary Office. The protocol was granted ethical approval by the Veterinary Office of the county of Basel-Stadt, Switzerland (Permit Numbers: 2375 and 2303). Monoclonal antibodies were tested in the murine P. falciparum model essentially as described (Dreyer et al., 2012; Jiménez-Díaz et al., 2009). Human erythrocytes (hE) were administered daily (0.75 mL) by the i.v. or i.p. route. Mice received a single dose of mAbs formulation by i.v. injection. The following day, mice were infected with 3 × 107 erythrocytes parasitized by P. falciparum PfNF540230/N3, a strain developed at GlaxoSmithKline (GSK) for growth in hE engrafted mice (Jiménez-Díaz et al., 2009). Parasitemia was monitored daily by flow cytometry over 6 days (days 4-9 after mAb injection).
FreeStyle 293 F cells (Invitrogen, R790-07), a variant of human embryonic kidney HEK cells, were cultured in suspension in serum-free medium (FreeStyle 293 Expression Medium, Thermo Fisher Scientific, Waltham, MA) at 37°C in a humidified incubator with 5% CO2 in volumes of 1 L shake flasks (Corning; 120 rpm, 5 cm diameter) or 10 L wave bioreactors (Sartorius; 30 rpm, pH 7.2, 30% dissolved oxygen). Cells were diluted 1:2 with fresh culture medium and transfected at 1.2·106 cells/mL with 0.4 mg/L expression plasmid using a riDOM-based transfection system (Québatte et al., 2014). 72 hr post-transfection, cells were removed by filtration and the supernatant was concentrated with a 10K Pellicon three cassette (Millipore). The His6-tagged recombinant proteins were purified by immobilized metal ion affinity chromatography on a HisTrap HP column (5 or 10 mL volume; GE-Healthcare) equilibrated with 50 mM HEPES/NaOH pH 7.2, 500 mM NaCl. After washing the column with the same buffer containing 20, 40, and 50 mM imidazole, the protein was eluted with a linear 50–500 mM imidazole gradient over 20 column volumes. The eluate was concentrated by ultra-filtration (Amicon Ultra-4 Ultracel 10K) and applied to a HiLoad 16/600 Superdex 200 gel permeation column (GE-Healthcare) equilibrated with 50 mM Tris/HCl pH 7.4, 150 mM NaCl. Homogeneity of PfCyRPA was assessed by reversed-phase chromatography (RP-HPLC) on a Poroshell 300 SB-C8 1 × 75 mm2 column using a H2O +0.01% TFA / Acetonitrile + 0.08% TFA gradient, and was confirmed by LC/MS intact mass analysis. Protein yields were 17 and 10 mg/Lr of culture for the glycosylated and non-glycosylated PfCyRPA, respectively.
Purified anti-PfCyRPA mAb c12 (Dreyer et al., 2012) was diluted to 0.5 mg/mL in 20 mM sodium phosphate pH 7.0 and cleaved overnight at 21°C with 0.01 mg/mL papain (Sigma-Aldrich) in a molar ratio of 1:20. The reaction was stopped with 0.001 mg/mL E64 inhibitor (Sigma-Aldrich) and the concentrated (Amicon Ultra-4 Ultracel 10K) hydrolysate was applied to a 1 × 5 cm2 Toyopearl protein A (Tosoh Biosciences) column equilibrated in 20 mM sodium phosphate pH 7.0. The flow-through containing the Fab was concentrated and chromatographed on a 21.5 × 60 cm2 TSKgel G3000SW column (Tosoh Bioscience) equilibrated with 20 mM bis-Tris propane/HCl pH 7.0, 200 mM NaCl. The purity of the Fab was determined by RP-HPLC as described above.
The complex of PfCyRPA with the Fab of mAb c12 (abbreviated as c12 in the following) was prepared with a 1.5-fold excess of PfCyRPA, which was incubated for 20 min at 21°C, concentrated as above and chromatographed via a Superdex 200 Increase 10/300 GL column (GE-Healthcare) equilibrated with 20 mM bis-Tris propane/HCl pH 7.0, 200 mM NaCl. Complex-containing fractions were pooled and analyzed for homogeneity by asymmetric flow field-flow fractionation with static multi-angle light scattering (SEC/AF4-MALS). For the expression of the Asp66Lys PfCyRPA variant and PfRH5, FreeStyle 293 F cells were cultured in suspension in serum-free medium (FreeStyle 293 Expression Medium, Thermo Fisher Scientific) at 37°C in a humidified incubator with 5% CO2 in 125 mL shake flasks. Cells were diluted 1:2 with fresh culture medium and transfected at 106 cells/mL with 0.5 mg/L expression plasmid using the 293fectin transfection reagent (Thermo Fisher Scientific). 72 hr post-transfection, cells were removed by centrifugation, and the His6-tagged recombinant Asp66Lys PfCyRPA variant was purified by immobilized metal ion affinity chromatography on a HisTrap HP column (1 mL volume; GE-Healthcare) equilibrated with 50 mM Na-phosphate buffer pH 7.2, 500 mM NaCl. After washing the column with the same buffer containing 20 mM imidazole, the protein was eluted (isocratic elution) with 500 mM imidazole over five column volumes.
FreeStyle 293 F cells were tested and shown to be free of mycoplasma using MycoAlert Mycoplasma detection kit (Lonza; LT07-318). Identity of cells was confirmed using STR-PCR (Qiagen; Investigator Idplex Pkus Kit, 381625) from genomic DNA purified using High Pure PCR Template Kit (Roche Applied Science 11796828001). The obtained profile was compared to the database available from DSMZ and found to match the HEK293 profile.
293 HEK cells (ATCC, CRL-1573) expressing PfCyRPA fragments on the cell surface were generated essentially as described previously by Dreyer et al. (2012). Briefly, DNA sequences coding for the fragments of PfCyRPA were amplified by PCR from the BVM_PFD1130W_FLAG_GP_His plasmid (Dreyer et al., 2010). These expression vectors allow the anchoring of the protein of interest on the cell surface via the transmembrane domain of mouse glycophorin-A (GP). In addition, they contain the secretion signal of bee-venom melittin (BMV), a FLAG tag located extracellularly, and a His6-tag located in the cytosol. The 293 HEK cells were transfected with the different expression vectors using JetPEI transfection reagent (PolyPlus) according to the manufacturer’s instructions. Transient transfectants were harvested 48 hr post-transfection. HEK cell lysates were prepared at 107 cells/mL in RIPA-Buffer (1 % NP40, 0.25% DOC, 10% glycerol, 2 mM EDTA, 137 mM NaCl, 20 mM Tris/HCl pH 8.0, plus protease inhibitors) and used for Western blot analysis as described in Favuzza et al. (2016).
293 HEK cells were regularly tested by PCR with mycoplasma-specific primer GPO-1 (5'- ACTCCTACGGGAGGCAGCAGTA-3') and MGSO (5'-TGCACCATCTGTCACTCTGTTAACCTC-3') and shown to be mycoplasma-free.
To locate the disulfide bonds in PfCyRPA, peptides derived from proteolysis under native conditions by the endo-proteases LysC, AspN, and trypsin were analyzed by UPLC-tandem mass spectrometry using a Dionex UltiMate3000 RRLC system (Thermo Scientific) coupled to a Synapt G2 HDMS mass spectrometer (Waters) with an electrospray ion source and working in resolution mode under default parameters. 10 µg PfCyRPA was hydrolyzed for several hours at 37°C with 0.3 µg of LysC, trypsin, or AspN. Peptides were separated on a Waters BEH130 C18 1.7 µm UPLC column (0.3 × 150 mm2) by a 5–45% gradient of acetonitrile in water containing 0.1% HCOOH at a flow rate of 10 µL/min over 90 min and eluted into the mass spectrometer. Data were analyzed using BiopharmaLynx software Ver. 1.3 (Waters). The data derived from the proteolysis are summarized in Figure 3—figure supplement 2.
The secondary structure composition of PfCyRPA was estimated by far-UV circular dichroism (CD) spectroscopy on a Jasco J-815 spectropolarimeter using 10 µM PfCyRPA in phosphate buffered saline (PBS) at 20°C. A cell of 1 mm path length was used to record four spectra between 200 and 320 nm with a step size of 0.1 nm and an integration time of 1 s. The spectral average was corrected for buffer contributions.
Intrinsic tryptophan fluorescence of a 5 μM PfCyRPA solution under native (PBS) and denaturing (70°C or PBS with 9 M urea) conditions was measured on an ISS PC-1 photon counting spectrofluorimeter (ISS, Inc.) at 20°C. Fluorescence emission was excited at 295 nm with 4.8 nm band width and path length 1 mm. Emission was monitored in 2 nm steps between 300 and 400 nm with a band width of 16 nm, a path length 5 mm, and an integration time of and 1 s.
All crystallization were done at 21°C in the sitting drop vapor diffusion setup using a Mosquito LCP crystallization robot (TTP Labtech). If not stated otherwise crystals were cryo-protected with paraffin oil and vitrified in liquid N2. A non-glycosylated variant of PfCyRPA (residues Asp29–Glu362, N145Q/N322Q/N338Q) did not yield well-diffracting crystals unless pre-treated with proteases. 1 mg/mL PfCyRPA in 50 mM Tris/HCl pH 7.4, 150 mM NaCl was diluted 1:1 with 10 mM HEPES/NaOH pH 7.5, 500 mM NaCl, subjected for 16 hr to a panel of twelve proteases (Hampton Research Proti-Ace Kit) at a final concentration of 0.01 mg/mL, and analyzed by reducing SDS-PAGE (Figure 4—figure supplement 1). Of the proteases that changed the apparent molecular mass of PfCyRPA, subtilisin and proteinase K produced multiple bands, while Glu-C, elastase, and thermolysin produced a single band of slightly smaller apparent molecular mass. The thermolysin-treated PfCyRPA was subjected to crystallization, yielding triclinic crystals upon mixing of 120 nL 7.3 mg/mL PfCyRPA in 50 mM Tris/HCl pH 7.4, 150 mM NaCl with 80 nL of precipitant consisting of 0.2 M LiOAc and 20% (w/v) PEG 3350. These crystals are consistent with four molecules per unit cell but did not diffract X-rays beyond a resolution of 3.3 Å. Actinase E-treatment of PfCyRPA results in a different proteolysis pattern with two bands of apparent molecular mass 17 kDa and 20 kDa (Figure 4—figure supplement 1). Thin, plate-shaped monoclinic crystals diffracting to a maximum resolution of about 2.5 Å were obtained from mixing of 100 or 140 nL 15 mg/mL Actinase E-treated PfCyRPA in 50 mM HEPES/NaOH pH 7.5, 150 mM NaCl with 100 nL and 60 nL precipitants consisting of 0–0.5 M MgCl2 and 20–25% PEG 3350. Crystals of the same habit were also obtained from 7.6 mg/mL PfCyRPA mixed with 19% PEG 3350, 0.3 M MgCl2 and additional 5 mM CaCl2, SrCl2, and BaCl2. Data collected from crystals grown in the presence of the heavier cations were tested for anomalous and isomorphous information content for phasing, which turned out to be negative.
Fab of mAb c12 crystallized from several conditions and its structure was determined in three crystal settings. Crystals were obtained by mixing of 80 nL 16.6 mg/mL c12 in 20 mM bis-Tris propane/HCl pH 7.0, 200 mM NaCl with 120 nL of precipitants consisting of 1.6 M sodium citrate pH 6.5 (hexagonal crystal form) or 0.1 M HEPES/NaOH pH 7.0, 20% w/v PEG 8000 (monoclinic and orthorhombic crystal forms). Monoclinic crystals were cryo-protected with reservoir solution supplemented with 20% glycerol.
Needle-shaped crystals of the PfCyRPA/c12 complex were obtained by mixing 120 nL of a 15.1 mg/mL solution in 50 mM Tris/HCl pH 7.4, 150 mM NaCl with 80 nL reservoir consisting of 0.1 M Sodium citrate pH 5.5, 17 % w/v PEG 5000 MME, 0.2 M NDSB-201 (non-detergent sulfobetaine). Cryo-protection was achieved with reservoir solution supplemented with 20% ethylene glycol.
Diffraction data were collected at Swiss Light Source beamline PX-II on a Pilatus 6M single photon counting detector using 1 Å radiation over a total range of at least 180° in fine slicing mode (Δϕ = 0.25°). Data were indexed, integrated, and scaled with XDS (Kabsch, 2010), except for data from the hexagonal form of c12, which was integrated with MOSFLM and scaled with AIMLESS (The collaborative computational project number 4, 1994). The high-resolution limit was chosen as the shell where the correlation coefficient for half datasets CC1/2 dropped below 70%. Data sets were tested for internal symmetry by self-Patterson and self-rotation function analyses (not shown). Twinning was excluded based on L-values and second moments. Data collection and also refinement statistics are collected in S1 Table.
The presence of a short sequence motif in PfCyRPA that is characteristic for sialidases suggested that a six-bladed β-propeller might serve as a molecular replacement model. However, all attempts using several hundred β-propeller structures as search models both with and without loop regions and/or as poly-Ala models were unsuccessful. Sialidases often contain structural Ca2+ ions, which prompted co-crystallization of PfCyRPA with the earth alkali cations Ca2+, Sr2+, and Ba2+ for SIRAS phasing, but none of these provided useful derivatives. Attempts to phase the PfCyRPA diffraction data by sulfur-SAD were also unsuccessful. Packing density estimations indicated two molecules of PfCyRPA in the monoclinic asymmetric unit. The PfCyRPA data could be phased by molecular replacement using the unexplained electron density of the PfCyRPA/c12 complex (see below) as a search model. The volume of density was separated using a mask, placed in a large cubic cell, and back transformed to yield structure factor amplitudes, which were then used in a PHASER (The collaborative computational project number 4, 1994) molecular replacement search for two molecules. The solution had a non-random log-likelihood gain of LLG = 175 and the orientations of the two entities replicated the 180° two-fold NCS (non-crystallographic symmetry) estimated from the self-rotation function. Density averaging using PARROT (The collaborative computational project number 4, 1994) and the NCS operator defined by the two PHASER solutions resulted in manually interpretable electron density maps. The final model has a discontinuity at Asp189, which is consistent with cleavage of a surface loop by Actinase E. A symmetry-related PfCyRPA occupies the space liberated by the cleaved loop, explaining the necessity of the proteolytic pre-treatment for crystal formation in this setting. The model was refined using automatically generated NCS restraints excluding diverging surface loop regions. 99.5% of all residues are in the favored regions of the Ramachandran plot. The hexagonal dataset of c12 (Supplementary file 1) was phased by molecular replacement using a homology model generated from the c12 sequence in MOE (Chemical Computing Group) as the search model. Separate searches for the variable and constant parts of the Fab were performed in PHASER, which, as anticipated, placed the VHVL and CHCL domain boundaries close to each other to generate a complete Fab. The final model has 99.2% of all residues in the favored regions of the Ramachandran plot. The monoclinic and orthorhombic c12 datasets were phased using the refined hexagonal c12 structure as molecular replacement search model. Both structures have 98% of their main-chain torsion angles in the favored region of the Ramachandran plot. Elbow angles were calculated with PHENIX (Adams et al., 2010).
According to the Matthews parameter, the 4 Å dataset collected from the PfCyRPA/c12 crystal has enough space per asymmetric unit to harbor a single complex. Molecular replacement readily placed c12, and initial maps calculated from this partial solution revealed additional electron density at the tips of the Fab (Figure 5—figure supplement 1), which occupied a volume large enough to host PfCyRPA. The density resembled a β-propeller but was not interpretable due to the limited resolution of the data. However, this density provided enough phasing power to solve the 2.5 Å PfCyRPA structure (see above), showing that the placement of the Fab was correct. The same arrangement of VHVL and CHCL was obtained when using the VHVL and CHCL parts of the Fab as separate search models, indicating that the elbow angle of c12 does not change upon binding to PfCyRPA. The final model of PfCyRPA was used to complete the molecular replacement phasing of the PfCyRPA/c12 complex (log-likelihood gain of >1260), density for which is shown in Figure 5—figure supplement 2. A loop region not present in the PfCyRPA model could be traced in the electron density of the complex (Panel B in Figure 5—figure supplement 1), establishing confidence in the correctness of the molecular replacement solution. The model of the complex was refined according to established protocols for lower resolution structures (DeLaBarre and Brunger, 2006), that is using simulated annealing, group ADP values, automatically assigned TLS definitions, secondary structure restraints, and external restraints provided by the higher resolution individual structures of PfCyRPA and c12. No real-space refinement was applied during building and refinement of the PfCyRPA/c12 complex. Strong geometric restraints were applied throughout model building of the PfCyRPA/c12 complex. All models were built with COOT (Emsley et al., 2010) and refined with PHENIX (Adams et al., 2010).
PHENIX: a comprehensive Python-based system for macromolecular structure solutionActa Crystallographica Section D Biological Crystallography 66:213–221.https://doi.org/10.1107/S0907444909052925
Reticulocyte-binding protein homologue 5 - an essential adhesin involved in invasion of human erythrocytes by plasmodium falciparumInternational Journal for Parasitology 39:371–380.https://doi.org/10.1016/j.ijpara.2008.10.006
Antibodies inhibit the protease-mediated processing of a malaria merozoite surface proteinJournal of Experimental Medicine 180:389–393.https://doi.org/10.1084/jem.180.1.389
The many blades of the β-propeller proteins: conserved but versatileTrends in Biochemical Sciences 36:553–561.https://doi.org/10.1016/j.tibs.2011.07.004
Toward understanding tryptophan fluorescence in proteinsBiochemistry 37:9976–9982.https://doi.org/10.1021/bi980274n
Paths to a malaria vaccine illuminated by parasite genomicsTrends in Genetics 31:97–107.https://doi.org/10.1016/j.tig.2014.12.005
The cellular and molecular basis for malaria parasite invasion of the human red blood cellThe Journal of Cell Biology 198:961–971.https://doi.org/10.1083/jcb.201206112
Structural vaccinology: a three-dimensional view for vaccine developmentCurrent Topics in Medicinal Chemistry 13:2629–2637.https://doi.org/10.2174/15680266113136660187
Considerations for the refinement of low-resolution crystal structuresActa Crystallographica Section D Biological Crystallography 62:923–932.https://doi.org/10.1107/S0907444906012650
Apparent lack of N-glycosylation in the asexual intraerythrocytic stage of plasmodium falciparumEuropean Journal of Biochemistry 205:815–825.https://doi.org/10.1111/j.1432-1033.1992.tb16846.x
Reverse vaccinology in the 21st century: improvements over the original designAnnals of the New York Academy of Sciences 1285:115–132.https://doi.org/10.1111/nyas.12046
Structure-based antigen design: a strategy for next generation vaccinesTrends in Biotechnology 26:659–667.https://doi.org/10.1016/j.tibtech.2008.08.002
Neutralization of plasmodium falciparum merozoites by antibodies against PfRH5The Journal of Immunology 192:245–258.https://doi.org/10.4049/jimmunol.1302045
Antibody and B cell responses to plasmodium sporozoitesFrontiers in Microbiology 5:625.https://doi.org/10.3389/fmicb.2014.00625
Genetics of antigenic variation in plasmodium falciparumCurrent Genetics 55:103–110.https://doi.org/10.1007/s00294-009-0233-2
Improved murine model of malaria using plasmodium falciparum competent strains and non-myelodepleted NOD-scid IL2Rγ0 mice engrafted with human erythrocytesAntimicrobial Agents and Chemotherapy 53:4533–4536.https://doi.org/10.1128/AAC.00519-09
Synchronization of plasmodium falciparum erythrocytic stages in cultureThe Journal of Parasitology 65:418–420.https://doi.org/10.2307/3280287
Shape complementarity at protein/protein interfacesJournal of Molecular Biology 234:946–950.https://doi.org/10.1006/jmbi.1993.1648
Host sialoglycans and bacterial sialidases: a mucosal perspectiveCellular Microbiology 14:1174–1182.https://doi.org/10.1111/j.1462-5822.2012.01807.x
Structural and computational biology in the design of immunogenic vaccine antigensJournal of Immunology Research 2015:1–17.https://doi.org/10.1155/2015/156241
Malaria vaccines: identifying plasmodium falciparum liver-stage targetsFrontiers in Microbiology 6:965.https://doi.org/10.3389/fmicb.2015.00965
Protein crystallography in vaccine research and developmentInternational Journal of Molecular Sciences 16:13106–13140.https://doi.org/10.3390/ijms160613106
Plasmodium Falciparum Malaria Parasite Cultures and Their Use in ImmunologyAcademic Press.
Sialic acid recognition by Vibrio cholerae neuraminidaseJournal of Biological Chemistry 279:40819–40826.https://doi.org/10.1074/jbc.M404965200
Development and optimization of high-throughput methods to measure plasmodium falciparum-specific growth inhibitory antibodiesJournal of Clinical Microbiology 44:1665–1673.https://doi.org/10.1128/JCM.44.5.1665-1673.2006
The case for a rational genome-based vaccine against malariaFrontiers in Microbiology 5:741.https://doi.org/10.3389/fmicb.2014.00741
riDOM, a cell penetrating peptide. interaction with phospholipid bilayersBiochimica Et Biophysica Acta (BBA) - Biomembranes 1838:968–977.https://doi.org/10.1016/j.bbamem.2013.10.017
Max Bergmann lecture protein epitope mimetics in the age of structural vaccinologyJournal of Peptide Science 19:127–140.https://doi.org/10.1002/psc.2482
Conserved sequences in bacterial and viral sialidasesGlycoconjugate Journal 6:349–353.https://doi.org/10.1007/BF01047853
The CCP4 suite: programs for protein crystallographyActa Crystallographica. Section D, Biological Crystallography 50:760–763.https://doi.org/10.1107/S0907444994003112
Development of malaria transmission-blocking vaccines: from concept to productAdvances in Parasitology 89:109–152.https://doi.org/10.1016/bs.apar.2015.04.001
Stephen C. HarrisonReviewing Editor; Harvard Medical School, United States
In the interests of transparency, eLife includes the editorial decision letter and accompanying author responses. A lightly edited version of the letter sent to the authors after peer review is shown, indicating the most substantive concerns; minor comments are not usually included.
Thank you for submitting your article "Structure of the malaria vaccine candidate antigen CyRPA and its complex with a parasite invasion inhibitory antibody" for consideration by eLife. Your article has been favorably evaluated by Wenhui Li (Senior Editor) and four reviewers, one of whom, Stephen C. Harrison (Reviewer #4), is a member of our Board of Reviewing Editors. The following individual involved in review of your submission has agreed to reveal their identity: Matthew K Higgins (Reviewer #2).
The reviewers have discussed the reviews with one another and the Reviewing Editor has drafted this decision to help you consider preparing a revised submission.
CyRPA is an important red cell invasion protein that forms a complex with RH5 and RIPR. P. falciparum parasites use this complex to engage the host cell receptor, basigin, during parasite invasion of red cells. CyRPA and RH5 are both candidate vaccine immunogens, as antibodies directed against these proteins potently neutralize parasites. This manuscript extends our knowledge of antibody neutralization by showing that single monoclonals against CyRPA and RH5 act additively to provide enhanced neutralization. This result suggests that combined use of both proteins may be beneficial for vaccine development.
The authors report the structure of CyRPA at 2.5 Å resolution, which shows that it has a neuraminidase fold but lacks sialidase activity, thus indicating a non-enzymatic function. The authors also describe a structure of CyRPA bound with the Fab of neutralizing antibody c12, at to a reported 3.6 Å resolution, and validate the interaction site using a single point mutant and truncation analysis. For the most part the antibody data and the structure of CyRPA alone appear to be well done, but the reviewers have substantive concerns with the structure of CyRPA in complex with c12. Both structures are timely and relevant for the future of malaria vaccines, and the manuscript would have been suitable for publication in eLife were the structural results for the c12 complex more solid. If the structure with c12 has problems, the vaccine development focus of the manuscript is not justified and the manuscript less compelling overall. There are also concerns about the experiments on additivity of antibodies directed against RH5 and CyRPA, the other novel claim. We invite the authors to consider responding with a revised manuscript (and appropriate experiments and computations) to the concerns summarized below. We caution that relatively extensive new information will be required for a favorable outcome.
1) The expression of fragments of a single domain, even if it turns out to be one that is sequential in its fold (a β-propeller), generally gives unfolded polypeptide chain. Unless the fragments in Figure 1 have the correct S-S bonds within the blades that are present (are the chosen limits at positions between blades?), then they can only define linear epitopes – likely to be the least interesting ones. Moreover, in the cases in which there is no binding, have the authors demonstrated that the fragment is even expressed on the cell surface?
2) In Figure 1B the GIA data are shown simply as 'yes' or 'no'. The GIA assays used throughout seem to diverge from the standardised GIA assay in the NIH GIA reference laboratory used by many studies. This divergence makes it hard to compare how inhibitory the mAbs are when compared with other mAbs from published studies. The selling point for this paper is the ability to use this information for 'structural vaccinology' but without a real quantitative comparison of the mAbs used with the best from other studies, it is hard to know whether this c12 mAb is worth recapitulating in a vaccination attempt. More standardised and quantitative GIA would be important to make this conclusion. Indeed, the data in Figure 2 do suggest that rather high concentrations of these mAbs are required to see an inhibitory effect.
3) Because the CyRPA is a conserved protein across several Plasmodium species as well as P. falciparum isolates, at least one reviewer would have liked to see data showing whether the inhibitory mAb inhibits strains other than 3D7. There is no mention of the specificity or the istotype of the control mAbs. Finally, it is important to show that mAbs specific for proteins other than RH5 do not have the same additive effect with anti-PfCyRPA. That control is missing here.
4) A key point for vaccine development is to show sterilizing immunity by targeting RH5 and CyRPA. In Figure 3B, however, the authors do not show the sterilizing immunity that would be expected if the 2.5 mg CS1.2 and 2.5 mg c12 combination was tested. Was this performed? If not, why not? Since the authors have shown previously that c12 inhibits invasion, the only novelty here is the combination.
5) The authors seem to have little experience with relatively low resolution structures, and their description of the refinement of the complex between Fab and PfCyRPA as "unsatisfactory" (subsection “Characterization of the epitope recognized by the parasite growth inhibitory anti-PfCyRPA mAb c12”, second paragraph) undermines confidence in the rest of the analysis, as one wonders whether the crystallographers involved (who are they?) really know how to look at a lower resolution map and how to avoid the pitfalls. Since both structures have been refined at quite high resolution, and since even the CDR loops of the Fabs do not vary much from crystal form to crystal form, it seems quite odd that the complex is so poorly determined. Specific points to consider are some red flags from data reduction statistics. Using the CC1/2 as the resolution cutoff criterion gives Rmeas of 59.5 in the outer bin (high for a low symmetry space group like P21) and I/sig(I) of 0.4 (and only 2.6 for the entire dataset). Since the elbow varies a bit among Fab c12 crystal forms, did the authors use the variable and constant modules as separate MR search objects when determining the structure of the complex?
Summary of required revisions:
1) Response to point (1), above, either with new data or – if a better structure of the c12:PfCyRPA complex can be achieved – by eliminating those experiments altogether, as a fully convincing structure will make them redundant.
2) Explicit defense of the variant GIA assay, or use of a more standard, quantitative assay.
3) Invasion inhibition assay with at least one more strain and inclusion, for both strains, of the stipulated control.
4) What combination of concentrations will give sterilizing immunity?
5) A better structure of the complex. Careful re-refinement of the structure seems in order, including careful re-examination of the initial MR solution. It might be better to collect more data, as redundancy, even at the current resolution, will result in more accurately determined intensities. Please provide some supplementary figures with maps and models, as was once standard, to restore confidence that the structure is correct. Show us the original MR-phased map in the region of the interface, together with the final model inside it, and then also the final 2Fo-Fc map, likewise with the model. One could still have problems even if these maps looked clear and well fit, but they will help.https://doi.org/10.7554/eLife.20383.036
- Gerd Pluschke
The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.
We thank Marcello Foggetta and Martin Siegrist for cell transfection. We also thank the staff at the Swiss Light Source for beamline support and Expose GmbH for data collection.
Animal experimentation: All procedures involving living animals were performed in strict accordance with the Rules and Regulations for the Protection of Animal Rights (Tierschutzverordnung) of the Swiss Federal Food Safety and Veterinary Office. The protocol was granted ethical approval by the Veterinary Office of the county of Basel-Stadt, Switzerland (Permit Numbers: 2375 and 2303).
- Stephen C. Harrison, Reviewing Editor, Harvard Medical School, United States
© 2017, Favuzza et al
This article is distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use and redistribution provided that the original author and source are credited.