Epitope mapping of PfCP-2.9, an asexual blood-stage vaccine candidate of Plasmodium falciparum

Background Apical membrane antigen 1 (AMA-1) and merozoite surface protein 1 (MSP1) of Plasmodium falciparum are two leading blood-stage malaria vaccine candidates. A P. falciparum chimeric protein 2.9 (PfCP-2.9) has been constructed as a vaccine candidate, by fusing AMA-1 domain III (AMA-1 (III)) with a C-terminal 19 kDa fragment of MSP1 (MSP1-19) via a 28-mer peptide hinge. PfCP-2.9 was highly immunogenic in animal studies, and antibodies elicited by the PfCP-2.9 highly inhibited parasite growth in vitro. This study focused on locating the distribution of epitopes on PfCP-2.9. Methods A panel of anti-PfCP-2.9 monoclonal antibodies (mAbs) were produced and their properties were examined by Western blot as well as in vitro growth inhibition assay (GIA). In addition, a series of PfCP-2.9 mutants containing single amino acid substitution were produced in Pichia pastoris. Interaction of the mAbs with the PfCP-2.9 mutants was measured by both Western blot and enzyme-linked immunosorbent assay (ELISA). Results Twelve mAbs recognizing PfCP-2.9 chimeric protein were produced. Of them, eight mAbs recognized conformational epitopes and six mAbs showed various levels of inhibitory activities on parasite growth in vitro. In addition, seventeen PfCP-2.9 mutants with single amino acid substitution were produced in Pichia pastoris for interaction with mAbs. Reduced binding of an inhibitory mAb (mAb7G), was observed in three mutants including M62 (Phe491→Ala), M82 (Glu511→Gln) and M84 (Arg513→Lys), suggesting that these amino acid substitutions are critical to the epitope corresponding to mAb7G. The binding of two non-inhibitory mAbs (mAbG11.12 and mAbW9.10) was also reduced in the mutants of either M62 or M82. The substitution of Leu31 to Arg resulted in completely abolishing the binding of mAb1E1 (a blocking antibody) to M176 mutant, suggesting that the Leu residue at this position plays a crucial role in the formation of the epitope. In addition, the Asn15 residue may also play an important role in the global folding of PfCP-2.9, as its substitution by Arg lead to reduced binding of most mAbs and abolishing the binding of mAb6G and mAbP5-W12. Conclusions This study provided valuable information on epitopes of PfCP-2.9 vaccine candidate through generation of a panel of mAbs and a series of PfCP-2.9 mutants. The information may prove to be useful for designing more effective malaria vaccines against blood-stage parasites.


Background
Malaria remains a severe tropical disease. It causes about 247 million cases in the world and nearly a million deaths in 2006, most of which were in children under five years of age in sub-Saharan Africa [1]. The Plasmodium para-site of malaria has a complex life cycle, and the blood stage is believed to be responsible for the disease and directly associated with clinical manifestations and therefore considered as a primary target for developing malaria vaccines. Apical membrane antigen 1 (AMA-1) and merozoite surface protein 1 (MSP1) of Plasmodium falciparum play essential roles in the blood stage of malarial life cycle and are two promising vaccine candidates [2,3]. After a simi-lar proteolytic processing step by P. falciparum subtilisinlike protease 2 (PfSUB2) [4,5], their C-terminal regions remain anchored within the parasite membrane, and then are carried into the newly invaded red blood cells [6,7]. Plasmodium falciparum chimeric protein 2.9 (PfCP-2.9) constructed by fusing PfAMA-1 (III) with PfMSP1-19 in previous studies [8,9], was shown to enhance the product yield, immunogenicity, and antibody-mediated in vitro growth inhibition assays (GIA). When formulated with Montanide ISA720, the vaccine was also confirmed to be safe, tolerable and immunogenic in two Phase I clinical trials [10,11]. Although higher levels of antibody titers are usually associated with better protection, some researchers argued that fine specificity of antibodies was closely associated with biological functions [12][13][14]. As described by Uthaipibull et al [13], antibodies to PfMSP1-19 could be classified as inhibitory (inhibiting secondary processing and/or erythrocyte invasion), blocking (blocking the function of inhibitory antibodies) and neutral (neither inhibitory nor blocking). It was proposed that there may also exist malaria specific non-AMA-1 blocking antibodies, because it was found that a fraction of malaria specific immunoglobulins (IgGs) from Malians that did not bind to AMA-1 substantially reduced the growth-inhibitory activity of both total IgGs and anti-AMA-1 antibodies from the sera of U.S. vaccinees [15]. Therefore, fine epitope mapping of the PfCP-2.9 protein is needed for further development of the vaccine candidate. Some epitopes of PfMSP1-19 have been described previously, especially since the discovery of the three-dimensional structure of the molecule [16][17][18][19]. PfMSP1-19 consists of two epidermal-growth-factor (EGF)-like domains, forming a disc-like shaped molecule. Inhibitory monoclonal antibodies (mAbs), including 12.8 and 12.10, recognize conformational epitopes on one side of the molecule, while blocking or neutral mAbs seem to recognize epitopes localized on the other side. The difference in distribution of the inhibitory and neutral epitopes is consistent with the formation of a ring structure between PfMSP1-19 and both 2F10 (neutral) and 12.10 (inhibitory) antibodies [12]. The structure of PfAMA-1 (III) has also been resolved, although its functional epitopes remain unknown [20]. It was found in a previous study [9] that PfCP-2.9 maintained the important conformational epitopes on PfMSP1-19 through interaction with a panel of conformational mAbs and displayed elevated biological activity compared with the two components. However, functional epitopes on PfCP-2.9 remain unclear. In this study, a panel of anti-PfCP-2.9 mAbs and a series of mutated proteins with single amino acid substitution on PfCP-2.9 were prepared to clarify the distribution of epitopes on PfCP-2.9.

Generation of hybridomas and production of mAbs to PfCP-2.9
A panel of mAbs to the PfCP-2.9 were produced in this study by the hybridoma technology as previously described [21]. Briefly, BALB/c mice were immunized by administration of 50 μg PfCP-2.9 in the presence of FCA (Freund's complete adjuvant)/FIA (Freund's incomplete adjuvant), followed by three booster injections of the same dose at 15-day intervals. Spleen cells were separated from mice 3 days after the last immunization and fused with SP2/0 mouse myeloma cells. The hybrids were then selected in hypoxanthine-aminopterin-thymidine (HAT) medium and cloned by limiting dilution. For large-scale production, hybridoma cell lines were abdominally injected into mice to generate mAbs in ascites. The mAbs were further purified by protein G affinity chromatography, dialyzed against RPMI 1640 media, aliquoted and stored at -80°C.

Parasite cultures and GIA
Plasmodium falciparum FCC1/HN strain was maintained and used for the assay [9]. Synchronized parasites were cultured until the mature trophozoite stage of development. 100 μl culture containing 2% RBC with a 0.5% ± 0.1% initial parasitemia and 25 μl of test mAbs was added to 96-well flat-bottomed plates (Greiner Bio-one, Germany). The initial concentrations of the test mAbs were adjusted to 2 mg/ml before the assay. IgG of rabbit anti-AMA1-C1 (2 mg/ml) was used as an inhibitory control (gift of NIH) and mAb5.2 was used as a non-inhibitory control. All the mAbs were sterilized by filtration through a 0.22 μm Millex-GV filter (Millipore, USA). After 42-h incubation with the purified antibody, 50 μl well-mixed parasites from each assay well were dispensed into 0.25 ml cold phosphate buffered saline (PBS, pH 7.4) in a fresh 96-well U-bottomed ELISA microplate (Greiner Bio-one, Germany). The plates were loaded into a Centrifuge 5810R (Eppendorf, Germany) and centrifuged at 1,300 × g for 10 min at 4°C. A portion (0.24 ml) of the supernatant was removed from each well, and the plate was frozen at -80°C for 30 min to lyse the pelleted cells. After the plate was brought back to room temperature (RT), 0.1 ml complete pLDH substrate (100 mM Tris-HCl pH 7.5, 50 mM sodium L-lactate, and 0.25% Triton X-100 containing 0.2 mg nitroblue tetrazolium ml -1 , 0.05 mg 3-acetylpyridine adenine dinucleotide ml -1 , and 1 U Diaphorase ml -1 ) was added to each well. The plate was kept in the dark and read at 15 and 30 min. The absorbance of wells in each plate was measured at 650 nm on a Microplate reader (ELx800 ELISA reader; Bio-Tek Instruments, Winooski, VT), and the percentage of inhibition was calculated as 100 -(A 650 of the test sample -A 650 of the RBC only)/(A 650 of mAb5.2 -A 650 of RBC only) ×100.

Mutagenesis
The pfcp-2.9 gene [9] was mutated by site-directed mutagenesis using Pichia optimized codons by overlap extension polymerase chain reaction (PCR) [22], based on the rule of "safe substitution" [23], which would minimally affect the global folding of the target protein. The constructed plasmid vector of pfcp-2.9/pBluescriptKS was amplified by KOD-plus DNA polymerase (Toyobo, Osaka, Japan) with the mutant primers listed in Table 1. Then the mutated genes were transformed into competent Escherichia coli DH5α cells (Life technologies). After screening by analysis of restriction enzyme digests, the sequence of selected mutant clones was confirmed using an Applied Biosystems ABI 377 DNA sequencer according to the manufacturer's instructions.

Protein expression and purification
Each mutated 6×His tagged gene was cloned into the expression vector of pPIC9K and linearized via SacI digestion. After transformation into Pichia pastoris by electroporation, selection of His + transformants and G418-resistant clones was performed according to the manufacturer's instructions (Invitrogen, San Diego, CA). The selected clone was grown in minimal glycerol medium at 30°C overnight and then resuspended in buffered methanol-complex medium containing 0.5% methanol to induce expression of the target gene. After 72-h induction, the supernatant was collected by centrifugation at 10,000 × g for 10 min, and then dialyzed against lysis buffer (50 mM NaH 2 PO 4 , 300 mM NaCl, pH 8.0) at 4°C overnight. Filtration was performed with a microporous membrane (0.22 μm in diameter) before the protein was purified by Ni-NTA affinity chromotography. Elution was performed with elution buffer (lysis buffer containing 50~250 mM imidazole, pH 8.0) after the column was equilibrated by lysis buffer and washed extensively with wash buffer (lysis buffer containing 10 mM imidazole, pH 8.0).

Analysis of antibody binding to PfCP-2.9 and mutated proteins by Western blot and ELISA
Purified protein samples were subjected to sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) after boiling in the sample buffer without a reducing agent. The protein bands were electrophoretically transferred to the nitrocellulose membrane (Huashun, China) by the Trans-blot SD semi-dry transfer cell (Biorad, USA). Blots were blocked with Tris-buffered saline (TBS, pH 7.4) containing 3% skim milk (TBS-M) at RT for 1 h; washed with TBS containing 0.05% Tween 20 (TBS-T) for three times and then probed by the primary antibody in TBS-M at RT for 2 h. The blots were washed and incubated with alkaline phosphatase (AP)-conjugated goat anti-mouse IgG (H+L) (Jackson ImmunoRe-search, USA) at a 1:1000 dilution for 1 h, followed by three washes in TBS-T and in AP buffer (100 mM NaCl, 5 mM MgCl 2 6H 2 O, 100 mM TrisCl, pH 9.5) for 10 min. Finally, antibody binding was detected by nitrotetrazolium blue chloride (NBT)/5-bromo-4-chloro-3-indolylphosphate (BCIP) (Genview, USA) substrate in AP buffer (66 μl NBT/33 μl BCIP/10 ml AP buffer) and terminated by washing with double distilled water. For qualitative comparison, proteins were analyzed on a SDS-PAGE gel to obtain an equal amount. Antibody binding to PfCP-2.9 and the mutated proteins was compared and scored as follows: "++" indicates approximately the same amount of antibody binding, "+" indicates reduced binding to the mutated protein compared with PfCP-2.9, and "-" indicates no detectable antibody binding to the mutated protein.
ELISA was also performed to examine the difference in binding intensity between PfCP-2.9 and its mutated proteins as described previously [9]. To avoid errors between the plates, PfCP-2.9 and its mutated proteins were coated in the same plate, while comparing with the same mAbs. MAb5.2 was adopted as an internal reference to assure that the proteins were equally coated, as the mapping of mAb5.2 had no overlapping epitope with the substitutions that we synthesized on PfCP-2.9 [12]., The absorbance of PfCP-2.9 for each mAb was used to obtain the percentage of the reduction rate (Rr) for the mutated proteins. The Rr value for the binding of a certain mAb was calculated as 100% -(A 450 of a mutated protein/A 450 of PfCP-2.9) ×100%. The recording method was similar to that in Western blot: "++" indicates about the same binding between PfCP-2.9 and mutated proteins with a Rr value within 25%; if the Rr value was between 25% and 75%, a signal of "+" would be marked; finally, "-" was assigned when the Rr value was above 75%, indicating no detectable antibody binding of the mutated protein as compared with PfCP-2.9.

Production and characterization of mAbs against PfCP-2.9
To identify epitopes of PfCP-2.9, a total of 12 mAbs recognizing the chimeric protein were prepared and characterized. Determination of binding regions on PfCP-2.9 for each mAb was investigated using recombinant PfAMA-1 protein (NIH) and PfMSP1-19 expressed from P. pastoris. The results showed that five mAbs (mAb2G, mAb7G, mAb10G, mAbG11.12 and mAbW9.10) recognized the AMA-1 portion, and four (mAb13, mAb16, mAbW17 and mAbP5-W12) recognized PfMSP1-19 epitopes, while the remaining three (mAb6G, mAb17, mAb18.20) failed to recognize either component of the chimeric protein in Western blot. PfCP-2.9 was further treated to determine whether its recognition by the mAbs was conformation-dependent. The results showed that except for mAb2G, mAb16, mAb17 and mAb18.20, interactions of the rest mAbs were reduction-sensitive. In addition, all the mAbs were detected for their ability to inhibit parasite growth in vitro by GIA and the results showed that six mAbs (mAb16, mAb17, mAbW17, mAb7G, mAb10G, and mAb13) had various levels of inhibitory activities from 9~32% The properties of the mAbs are summarized in Table 2.

Design and production of PfCP-2.9 mutants with single amino acid substitution
Seventeen mutants of the PfCP-2.9 with single amino acid substitution on either component of AMA-1 (III) or MSP1-19 were produced in Pichia pastoris. Figure 1A shows changes in location of the residues. Substitutions on the AMA-1 (III) component were selected as the highly conserved amino acids in different Plasmodium species [24] or amino acids within the epitope of mAb F8.12.10, which cross-reacts with several species of the Plasmodium [25]. These mutations were designed to discover possible functional epitopes on the AMA-1 (III) component. Some amino acid residues that participated in the formation of definitive blocking epitopes or those surface-exposed on the crystal structure of PfMSP1-19 were also selected to investigate the corresponding epitopes on PfCP-2.9. A general description of the substitutions is shown in Table 3.
To generate the PfCP-2.9 mutant genes, two overlapping fragments that carried the same point mutation were first produced by PCR amplification using the pfcp-2.9 gene as the template and then combined by the second round of PCR. The fragments and the final product of the mutant gene of m176 are representatively shown in Figure 1   sequenced before being transformed into P. pastoris for expression. The 17 PfCP-2.9 mutants were produced in Pichia system in a secreted form at various expression levels. The proteins were then purified by Ni-NTA affinity chromotography. SDS-PAGE and Western blot analyses of the PfCP-2.9 mutant proteins are shown in Figure  1(D) and 1(E), respectively.

Effects of single amino acid substitution of PfCP-2.9 on mAb binding
The PfCP-2.9 and its mutants were assayed for mAb binding by both ELISA and Western blot. Epitope mapping of PfCP-2.9 with the mAbs is summarized in Table 4, and the main results of Western blot and ELISA are shown in Figure 2. Mutations in the AMA-1 (III) component of PfCP-2.9 did not change the antibody binding significantly, except that M62 (Phe 491 TAla) reduced the binding of mAb7G, mAbG11. 12      "++": strong binding; "+": reduced binding; "-": no binding that had no effect on the binding in the immunoblot had an Rr value below 25%.

Discussion
PfCP-2.9 is a promising vaccine candidate among many worthy vaccines under development for prevention of malaria. It is made by fusing two blood-stage antigens (PfAMA-1 (III) and PfMSP1-19) through a hinge.
Although animal experiments have demonstrated that this vaccine is much more immunogenic than either component or the mixture of the two components, its functional epitopes remain unclear.
To identify epitopes of PfCP-2.9, a panel of 12 mAbs were prepared in the present study, of which six showed various levels of inhibitory activity as identified by GIA. The Western blot analysis showed that most of the inhibitory mAbs were sensitive to reduction, indicating that the disulfide-bond stabilized conformation is still critical for the chimeric protein, which is consistent with the previous work [9]. Interestingly, three of the mAbs only recognized PfCP-2.9, but not either component. Possibility of novel epitopes created by interaction of the two components was excluded through analysis of the nuclear magnetic resonance (NMR) solution structure of PfCP-2.9 [26]. The specificity of the mAbs needs to be further investigated.
Epitope mapping was carried out by comparing mAb binding to PfCP-2.9 and its mutant proteins through Western blot and ELISA. Apparent binding changes are shown in Figure 2 and the epitopes associated with the changes were plotted onto the published three-dimensional structure of PfAMA-  [27] as shown in Figure 3. The published structures of the individual components were used for the plotting analysis based on the fact that NMR solution structures of the individual components of PfCP-2.9 were not changed. These findings may have the following implications: (1) Of the six mAbs whose binding was reduced or abolished by the mutations synthesized, only mAb7G exibited a mild inhibitory effect. This inhibitory epitope of mAb7G was found on the AMA-1 (III) component. The binding of mAb7G was reduced by any one of the three substitutions (M62: Phe 491 TAla, M82: Glu 511 TGln andM84: Arg 513 TLys). These three amino acids were all surface-exposed and clustered in the C-terminal unstructured region of AMA-1 (III), in close proximity to each other in conformation. The amino acid of Phe 491 was involved in an epitope recognized by mAbF8.12.19, which was not inhibitory but involved in interspecies cross-reactivity [25]. Besides the direct steric blockage of AMA-1 function such as displayed by mAb4G2 [28], cross-linking and the consequent inhibition of AMA-1 dispersion across the parasite surface have also been linked to invasion inhibition [28,29]. The mutation of M62 also disrupted the formation of a six-residue ring structure formed by Cys 490 -Cys 507 , Cys 492 -Cys 509 , Phe 491 and Arg 508 , which may be of functional importance on AMA-1 (III) [20]. These findings suggest that the amino acid of Phe 491 is tightly associated with the formation of the inhibitory epitope of mAb7G. Glu 511 and Arg 513 are both highly conserved amino acids across different Plasmodium species [24], and may also contribute to the formation of the epitope. The binding of mAbG11.12 was reduced by M62 or M82, while the binding of mAbW9.10 was only reduced by M62. Therefore, there is possibility that mAbG11.12 and mAbW9.10 could block the inhibitory effect of mAb7G by competing at the position of both Phe 491 and Glu 511 and at the position of Phe 491 , respectively.
(2) Some information about blocking epitopes was obtained. It was found that four definitive amino acids (Asn 15 , Glu 27 , Leu 31 and Glu 43 ) were involved in the formation of blocking epitopes on MSP1-19. It was reported [13] that mutation of these amino acids abolished the binding of blocking mAbs of mAb7.5, mAb2.2, mAb1E1 and mAb111.4, respectively. ELISA was performed to detect mAb1E1 binding to PfCP-2.9 mutant, and the results showed that its binding was abolished by mutant M176 (Leu 31 TArg) and reduced by mutant M188 (Glu 43 TLeu), which is consistent with the results from the mutants of PfMSP1-19 reported in other study [13].
There is a slight difference in the binding of mAb1E1 to the mutants by Western blot, which showed that the bindings to M176 and M188 were completely abolished. These results suggested that the residues of Leu 31 and Glu 43 were involved in the epitope of mAb1E1 on PfCP-2.9. In addition, the bindings of both mAb6G and mAbP5-W12 were abolished by M159 (Gln 14 TGly) or M188 (Glu 43 TLeu), and reduced by M185 (Lys 40 TIle). However, these antibodies showed no biological activity in GIA.
(3) A vital amino acid of Asn 15 , found in the present study, may be of critical importance for the correct folding of PfCP-2.9, since the substitution of Asn 15 TArg on PfCP-2.9 reduced the binding of most mAbs and abolished the binding of both mAb6G and mAbP5-W12.

Conclusions
This study focused on epitope mapping of PfCP-2.9 through generation of a panel of mAbs against PfCP-2.9 and a series of mutated proteins of PfCP-2.9 with single amino acid substitution. An inhibitory epitope on the AMA-1 (III) component, recognized by mAb7G, was associated with three amino acids, Phe 491 , Glu 511 and Arg 513 . The amino acid residues of Gln 14 and Glu 43 were involved in the formation of a non-inhibitory epitope recognized by two mAbs, mAb6G and mAbP5-W12. The blocking epitope of PfCP-2.9 recognized by mAb1E1 was the same as that described previously on PfMSP1-19.
These findings provide precise functional epitopes on PfCP-2.9 and may prove useful in directing the design of more effective malaria vaccines.