antigenic characterization of an intrinsically ...€¦ · antigenic characterization of an...

10
Published Ahead of Print 10 September 2012. 2012, 80(12):4177. DOI: 10.1128/IAI.00665-12. Infect. Immun. Raymond S. Norton and Robin F. Anders Paul Masendycz, Michael Foley, James G. Beeson, Reiling, Kaye Wycherley, Michelle J. Boyle, Vivian Kienzle, Christopher G. Adda, Christopher A. MacRaild, Linda Protein 2 Plasmodium falciparum Merozoite Surface Intrinsically Unstructured Protein, Antigenic Characterization of an http://iai.asm.org/content/80/12/4177 Updated information and services can be found at: These include: SUPPLEMENTAL MATERIAL Supplemental material REFERENCES http://iai.asm.org/content/80/12/4177#ref-list-1 at: This article cites 57 articles, 22 of which can be accessed free CONTENT ALERTS more» articles cite this article), Receive: RSS Feeds, eTOCs, free email alerts (when new http://journals.asm.org/site/misc/reprints.xhtml Information about commercial reprint orders: http://journals.asm.org/site/subscriptions/ To subscribe to to another ASM Journal go to: on January 23, 2013 by Alfred Health http://iai.asm.org/ Downloaded from

Upload: others

Post on 14-Jun-2020

11 views

Category:

Documents


0 download

TRANSCRIPT

  Published Ahead of Print 10 September 2012. 2012, 80(12):4177. DOI: 10.1128/IAI.00665-12. Infect. Immun. 

Raymond S. Norton and Robin F. AndersPaul Masendycz, Michael Foley, James G. Beeson,Reiling, Kaye Wycherley, Michelle J. Boyle, Vivian Kienzle, Christopher G. Adda, Christopher A. MacRaild, Linda Protein 2Plasmodium falciparum Merozoite SurfaceIntrinsically Unstructured Protein, Antigenic Characterization of an

http://iai.asm.org/content/80/12/4177Updated information and services can be found at:

These include:

SUPPLEMENTAL MATERIAL Supplemental material

REFERENCEShttp://iai.asm.org/content/80/12/4177#ref-list-1at:

This article cites 57 articles, 22 of which can be accessed free

CONTENT ALERTS more»articles cite this article),

Receive: RSS Feeds, eTOCs, free email alerts (when new

http://journals.asm.org/site/misc/reprints.xhtmlInformation about commercial reprint orders: http://journals.asm.org/site/subscriptions/To subscribe to to another ASM Journal go to:

on January 23, 2013 by Alfred H

ealthhttp://iai.asm

.org/D

ownloaded from

Antigenic Characterization of an Intrinsically Unstructured Protein,Plasmodium falciparum Merozoite Surface Protein 2

Christopher G. Adda,a Christopher A. MacRaild,b Linda Reiling,c Kaye Wycherley,d Michelle J. Boyle,c Vivian Kienzle,a*Paul Masendycz,d Michael Foley,a James G. Beeson,c Raymond S. Norton,b and Robin F. Andersa

Department of Biochemistry, La Trobe University, Victoria, Australiaa; Monash Institute of Pharmaceutical Sciences, Monash University, Parkville, Australiab; The BurnetInstitute for Medical Research and Public Health, Melbourne, Australiac; and The Walter and Eliza Hall Institute of Medical Research, Parkville, Australiad

Merozoite surface protein 2 (MSP2) is an abundant glycosylphosphatidylinositol (GPI)-anchored protein of Plasmodium falcip-arum, which is a potential component of a malaria vaccine. As all forms of MSP2 can be categorized into two allelic families, avaccine containing two representative forms of MSP2 may overcome the problem of diversity in this highly polymorphic pro-tein. Monomeric recombinant MSP2 is an intrinsically unstructured protein, but its conformational properties on the merozoitesurface are unknown. This question is addressed here by analyzing the 3D7 and FC27 forms of recombinant and parasite MSP2using a panel of monoclonal antibodies raised against recombinant MSP2. The epitopes of all antibodies, mapped using both apeptide array and by nuclear magnetic resonance (NMR) spectroscopy on full-length recombinant MSP2, were shown to be lin-ear. The antibodies revealed antigenic differences, which indicate that the conserved N- and C-terminal regions, but not the cen-tral variable region, are less accessible in the parasite antigen. This appears to be an intrinsic property of parasite MSP2 and isnot dependent on interactions with other merozoite surface proteins as the loss of some conserved-region epitopes seen usingthe immunofluorescence assay (IFA) on parasite smears was also seen on Western blot analyses of parasite lysates. Further stud-ies of the structural basis of these antigenic differences are required in order to optimize recombinant MSP2 constructs beingevaluated as potential vaccine components.

Plasmodium parasites are antigenically complex, and conse-quently antibodies to many different parasite antigens are seen

in the adaptive immune response to malaria. Prominent amongthe antigens of Plasmodium falciparum, the cause of most malariadeaths, are numerous proteins associated with the merozoite sur-face or released from the merozoite apical organelles. Antibodiesto many of these antigens have been shown to inhibit invasion ofhost erythrocytes by P. falciparum merozoites using an in vitrogrowth inhibition assay (GIA) (3, 9, 27, 45). Inhibition of parasitedevelopment by antibodies to other merozoite antigens appears tobe mediated by monocytes activated by antibody-Fc receptorinteractions in a process known as antibody-dependent cellularinhibition (ADCI) (7, 44, 49, 50). To what extent either of these invitro assays of antibody activity represents correlates of functionalimmunity remains unclear, but both have been used extensively inthe evaluation of merozoite antigens as vaccine candidates.

Merozoite surface protein 2 (MSP2) is a merozoite antigen thatis a potential component of a malaria vaccine (20, 32). MSP2 is ahighly abundant glycosylphosphatidylinositol (GPI)-anchoredmembrane protein but is unusual in that orthologues are lackingin the other Plasmodium species that infect humans and in rodentparasites. For this reason there has been no assessment of MSP2 asa vaccine candidate using rodent models of the human disease. Amajor problem confronting the use of MSP2 and other merozoitesurface proteins as vaccine candidates is their extensive sequencediversity (17, 24, 46, 51). MSP2 is one of the most polymorphic ofall the merozoite surface proteins, with a central variable region,which comprises �60% of the �220-residue mature polypeptidechain, flanked by conserved N-terminal and C-terminal regions(17, 46). The variable region contains sequence repeats flanked bynonrepetitive dimorphic sequences, which have allowed all MSP2alleles to be categorized into the 3D7 and FC27 allelic families.There is much evidence that many of these polymorphisms have

arisen as a result of the selection pressure of protective host im-mune responses (4, 12, 54). Anti-MSP2 antibodies induced byinfection with P. falciparum are largely directed against epitopes inthe central variable region of the molecule (34, 48), and the strain-specific protection afforded by the Combination B vaccine in aphase 1/2b trial in Papua New Guinea (18, 20) indicates that anti-MSP2 antibodies to variable-region epitopes are protective. Littleis known about the mechanisms underlying that protection or thefine specificity of the effector immune responses. However, invitro studies have shown that human antibodies to MSP2 havefunctional activity in ADCI assays (19, 32).

In contrast to other well-characterized vaccine candidates,MSP2 lacks a fold stabilized by intramolecular disulfide bonds(57). The amino acid composition of MSP2 has a marked deficit ofhydrophobic residues, and extensive physicochemical analyseshave established that MSP2 is an intrinsically unstructured pro-tein (1, 57). Nuclear magnetic resonance (NMR) studies on re-combinant FC27 MSP2 revealed a highly disordered polypeptidechain containing two short sequences with some helix propensity

Received 22 June 2012 Returned for modification 16 July 2012Accepted 4 September 2012

Published ahead of print 10 September 2012

Editor: J. H. Adams

Address correspondence to Robin F. Anders, [email protected].

* Present address: Vivian Kienzle, Queensland Institute of Medical Research,Herston, Queensland, Australia.

C.G.A. and C.A.M. contributed equally to this article.

Supplemental material for this article may be found at http://iai.asm.org/.

Copyright © 2012, American Society for Microbiology. All Rights Reserved.

doi:10.1128/IAI.00665-12

December 2012 Volume 80 Number 12 Infection and Immunity p. 4177–4185 iai.asm.org 4177

on January 23, 2013 by Alfred H

ealthhttp://iai.asm

.org/D

ownloaded from

and some motional restriction in the region of the single intramo-lecular disulfide bond toward the C terminus of the molecule (57).Like a number of other intrinsically unstructured proteins, MSP2has a propensity to form amyloid-like fibrils (1). It seems highlylikely that the conformation of the GPI-anchored merozoite sur-face antigen will be more constrained than monomeric recombi-nant MSP2, possibly as a result of oligomerization through inter-molecular �-strand interactions that are amyloid like or as a resultof interactions with the merozoite membrane (31).

If MSP2 is to be used as a component of a malaria vaccine, it isimportant to understand the structural relationship between re-combinant MSP2 and the target antigen on the merozoite surface.For this reason, we have produced a panel of monoclonal antibod-ies to MSP2, which we have used in a series of antigenic analyses ofrecombinant and parasite MSP2. The results show that there aresignificant conformational differences between the two forms ofMSP2 which are likely to have important implications for the ef-ficacy of a vaccine based on the recombinant antigen.

MATERIALS AND METHODSParasites and recombinant MSP2. P. falciparum strains 3D7 and FC27were cultured and used in Western blotting and immunofluorescenceassays (IFA) as described previously (1). With the exception of the antigenused to immunize mice (Ag1624) (see below), the recombinant forms of3D7 and FC27 MSP2 used in the antigenic analyses described here wereC-terminally His-tagged proteins expressed in Escherichia coli, as used in acombination MSP2 vaccine tested in a phase 1 clinical trial (32).

Western blotting, ELISA, and IFA. Western blotting, enzyme-linkedimmunosorbent assays (ELISA), and IFA were all carried out using stan-dard protocols as described previously (1). For semiquantitative IFA, par-asite smears were fixed with acetone-methanol (1:1, vol/vol), and mono-clonal antibodies (MAbs) were titrated in half-log serial dilutions from 3�g/ml to 0.03 �g/ml in 1% bovine serum albumin (BSA) in phosphate-buffered saline (PBS) (BSA-PBS). Each slide was divided into six squares,and each square was incubated with a different dilution of the primaryantibody followed by the secondary antibody (goat anti-mouse IgG con-jugated to Alexa Fluor 568 also diluted 1:200 in BSA-PBS). Fluorescencewas detected by excitation at 568 nm, and images were captured using aSpot-RT digital camera. Scoring was based on the penultimate dilution atwhich fluorescence was visible.

Epitope mapping with peptide array. An array of 84 biotinylated 13-mer peptides covering the entire mature 3D7 and FC27 MSP2 sequenceswith an overlap of eight residues (see Table S1 in the supplemental mate-rial) was purchased from Mimotopes Pty Ltd. (Melbourne, Australia).Peptides were solubilized in 80% dimethyl sulfoxide (DMSO) to a con-centration of 2 to 6 mg/ml. Nunc 96-well flat-bottom plates were coatedovernight with 1 �g/ml streptavidin in PBS. Blocking was performed for 2h with 1% casein at 37°C, followed by incubation with peptides diluted inPBS to 4 to 12 �g/ml. MSP2 MAbs at 1 �g/ml were incubated for 1 h atroom temperature, and binding was detected with anti-mouse horserad-ish peroxidase (HRP)-conjugated antibody (Invitrogen) at 1:2,500. Plateswere washed three times with PBS– 0.05% Tween between incubations.The ELISA plates were exposed to the substrate 2,2-azinobis(3-ethylben-zthiazolinesulfonic acid) (ABTS) (Sigma-Australia) for 30 min, and ab-sorbance was read at 405 nm after the reaction was stopped with 1% SDS.

Monoclonal antibodies. Hybridoma cell lines producing anti-MSP2MAbs were produced using standard procedures. Briefly, spleen cellsfrom female CBA mice immunized with polymeric recombinant 3D7MSP2 (Ag1624) formulated in Montanide ISA720 (1) were fused to Sp2/0myeloma cells, and hybridomas were selected by their growth in hypoxan-thine-aminopterin-thymidine(HAT) medium using 96-well microtiterplates. Polymeric MSP2 was used to increase the immunogenicity of theantigen and to increase the likelihood of polymer-specific MAbs beingproduced. Primary hybridoma supernatants were screened by ELISA on

plates coated with monomeric and polymeric Ag1624 (1), and positivehybridomas were subjected to at least two rounds of limiting-dilutioncloning before cell lines were stored in liquid nitrogen. Purified MAbswere isolated from the supernatants of roller bottle cultures on proteinG-Sepharose and were assayed by ELISA on plates coated with mono-meric 3D7 and FC27 MSP2.

NMR spectroscopy. Nontagged, 15N-labeled FC27 MSP2 was pre-pared for NMR studies as described previously (57). A C211S mutationwas introduced by Quikchange site-directed mutagenesis (Stratagene).Samples for NMR spectroscopy contained 44 �M 15N-labeled FC27MSP2, 20 mM sodium phosphate, and 10% 2H2O, pH 7.4, in a volume of280 �l in Shigemi microtubes. MAbs were concentrated to �10 mg/ml inPBS using Centricon centrifugal filtration units (Millipore) immediatelybefore titration into the NMR samples. Data were acquired at 5°C toreduce line broadening due to exchange with solvent and at 600 MHz 1Hfrequency using Bruker Avance II and Varian Inova spectrometers. 1H-15N band-selective optimized flip angle short transit heteronuclear mul-tiple-quantum coherence (SOFAST HMQC) spectra (41) were acquiredin the presence of increasing concentrations of MAb up to �20 �M.Backbone amide peak assignments under these experimental conditionswere derived from the published assignments (57) using pH titrations tofollow chemical shift changes. Peak heights were determined using theCcpNmr Analysis program (53), and data were normalized to the heightin the absence of antibody and corrected for the dilution of MSP2 due tothe addition of antibody.

RESULTSSpecificity of MSP2 MAbs by ELISA on recombinant MSP2. Ofthe nine MAbs generated in this study, six (1F7, 4D11, 6C9, 6D8,9G8, and 9H4) reacted by ELISA with both forms of recombinantMSP2, whereas three (2F2, 9D11, and 11E1) reacted with 3D7MSP2 but not FC27 MSP2. As expected, the control MAb 8G10,which reacts with an epitope in the 32-mer repeat of FC27 MSP2(15), reacted in the ELISA with FC27 but not with 3D7 MSP2(Table 1; see also Fig. S1 in the supplemental material). None ofthe MAbs was specific for polymeric MSP2 (data not shown), but,as described previously, MAb 6D8 reacted preferentially with mo-nomeric MSP2 (1).

All MSP2 MAbs recognize short linear sequences in MSP2.MAb epitopes were mapped using sets of biotinylated synthetic13-residue peptides that had an 8-residue overlap. One copy of thefirst three peptides common to both 3D7 and FC27 MSP2 wassynthesized, but because the central variable regions of 3D7 andFC27 MSP2 are different lengths, the two peptide sets (3D7 andFC27) extended through the conserved C-terminal region to givetwo sets of peptides covering the same sequence but out of framewith respect to each other by three residues (see Table S1 in thesupplemental material). All MAbs reacted strongly with one ortwo peptides, identifying a linear sequence that comprised all orthe major components of the corresponding epitopes (Fig. 1 andTable 1). The location of these linear sequences was consistentwith the specificities of the MAbs determined by ELISA usingrecombinant MSP2. Of the six MAbs that reacted with both formsof MSP2, one (6D8) reacted with peptide 3, which corresponds tothe C-terminal half of the conserved N-terminal region of MSP2.This was consistent with our earlier observation that this MAbreacts with the proteinase K-resistant core of MSP2 amyloid-likefibrils (1). The other five MAbs that reacted with both forms ofrecombinant MSP2 (9H4, 1F7, 6C9, 4D11, and 9G8) reacted withpeptides corresponding to conserved sequences close to the C ter-minus of MSP2.

From their reactivity with one 3D7 peptide (peptide 43) and

Adda et al.

4178 iai.asm.org Infection and Immunity

on January 23, 2013 by Alfred H

ealthhttp://iai.asm

.org/D

ownloaded from

two FC27 peptides (peptides 83 and 84) (Fig. 1 and Table 1), asequence of eight residues (NKENCGAA) was identified as theepitope for MAbs 4D11 and 9G8 (see Fig. S2 in the supplementalmaterial). Centrally located in this epitope, which is close to the Cterminus of the mature protein generated by cleavage at the GPIattachment site, is one of only two cysteine residues in MSP2,which form an intramolecular disulfide bond. Remarkably, thereactivity of 4D11 and 9G8 with MSP2 appeared to be unaffectedby the presence or absence of this bond (see below). Althoughthese two MAbs had very similar patterns of reactivity with thesepeptides in the conserved C-terminal region of MSP2, MAb 9G8but not MAb 4D11 had a weak but clearly significant reactivitywith peptide 75 in the FC27 set (Fig. 1A), which corresponds to asequence close to the C-terminal end of the central variable regionof FC27 MSP2. A modest sequence relationship in peptide 75 withthe 8-mer core of the major epitope (4 of 8 identical residues)(data not shown) is consistent with this being a cross-reaction. Incontrast, MAb 4D11 but not 9G8 reacted weakly with a sequencein the N-terminal conserved region (peptide 2). There was nosequence relationship with the 8-mer core of the 4D11/9G8epitope to explain this apparent cross-reaction.

The epitopes of the three other MAbs that reacted with theC-terminal conserved region of MSP2 were immediately N-termi-nal to the 4D11 and 9G8 epitopes (Fig. 1 and Table 1). These threeMAbs (1F7, 6C9, and 9H4) all reacted strongly with a single pep-tide in the 3D7 MSP2 set (peptide 42) but with no peptide in theFC27 MSP2 set. An alignment of the single reactive 3D7 peptide,which contained the two cysteine residues in MSP2, with the over-lapping 3D7 MSP2 peptides and the most closely correspondingFC27 peptide suggests that the epitope(s) for these three MAbs isdependent on the disulfide bond and several residues N-terminalto the disulfide bond (see Fig. S3 in the supplemental material).

The three MAbs that reacted by ELISA with 3D7 but not FC27MSP2 reacted with single peptides corresponding to central vari-able region sequences of MSP2. All three epitopes were locatedbetween the GGSA tandem repeats, which characterize the 3D7and some related forms of MSP2, and the conserved C-terminalregion (Fig. 1 and Table 1). As this region of the protein is dimor-phic, we would expect all three of these MAbs to react with most,if not all, forms of MSP2 belonging to the 3D7 allelic family. This

contrasts with the MAb 8G10, which was used here as a positivecontrol. Consistent with published data (15), MAb 8G10 reactedwith two peptides in the FC27 MSP2 set that comprised part of the32-mer repeat found in most alleles belonging to the FC27 allelicfamily (Fig. 1 and Table 1). There are polymorphic sites within thissequence, and, consequently, 8G10 would not be expected to reactwith all forms of MSP2 belonging to the FC27 allelic family.

Epitope mapping in recombinant FC27 MSP2 by NMR spec-troscopy. The epitopes of MAbs reactive against FC27 MSP2 werealso determined using NMR spectroscopy (38, 39). The strategyemployed relies on the unstructured nature of MSP2: backboneamide peaks of free MSP2 are sharp due to the extensive flexibilityof the molecule on the nanosecond time scale, but when MSP2 isbound to an antibody, these peaks are expected to be severelybroadened due to the slow rotational diffusion of the relativelyrigid antibody. The flexibility of MSP2 ensures that residues dis-tant from the epitope are decoupled from this slow diffusion andthus give rise to peaks of similar line shape to those of the freeprotein. Consistent with this expectation, when increasing sub-stoichiometric amounts of antibody were added to MSP2, thegradual disappearance of a few localized backbone amide peaks in1H-15N correlation spectra was observed, with the remainingMSP2 peaks unchanged in peak intensity or chemical shift (Fig. 2).

The epitopes mapped using this NMR technique are in agree-ment with those inferred from the peptide array data (Fig. 1 and2), further demonstrating that none of these MAbs recognizeepitopes that are conformational or otherwise cryptic in the pep-tide array. Furthermore, the minor cross-reactive epitopes identi-fied for MAbs 9G8 and 4D11 are also supported by the NMR data.To test the sensitivity of the antibodies binding to C-terminalepitopes to the presence of the disulfide bond, we repeated theseexperiments using MSP2 in which the C-terminal cysteine (C211in FC27 MSP2) was replaced with serine. MAbs 9G8 and 4D11bound both C211S and wild-type MSP2 indistinguishably, con-firming that the nearby intramolecular disulfide bond was irrele-vant to the epitope of these two antibodies (Fig. 2). In contrast,MAbs 1F7, 6C9, and 9H4 all failed to bind the mutant MSP2 underthese conditions, consistent with the inferred sensitivity of thisgroup of MAbs to the presence of the disulfide bond (Fig. 2).

Somewhat surprisingly, MAb 8G10 caused significant line-

TABLE 1 Reactivity of anti-MSP2 MAbs with recombinant proteins and peptides by ELISA and of smears of the P. falciparum cloned lines 3D7 andD10 by immunofluorescence microscopy

MAb

ELISA reactivityIFA reactivity by P.falciparum straina

Recombinant MSP2form(s)

Peptide

Name Isotype Specificity No. Sequence(s) 3D7 D10

6D8 IgG1 Conserved N-terminal region 3D7, FC27 3 FINNAYNMSIRRS � �11E1 IgG1 3D7 variable region 3D7 23 NPKGKGEVQEPNQ ���� �9D11 IgG2a 3D7 variable region 3D7 29 KSNVPPTQDADTK ����� �2F2 IgG1 3D7 variable region 3D7 35 QTESPELQSAPEN ����� �8G10 IgG2b FC27 variable region FC27 52, 58 SGSQRSTNSASTS,

IASGSQRSTNSAS� �����

1F7 IgG1 Conserved C-terminal region 3D7, FC27 42 SQKECTDGNKENC �/� �/�6C9 IgG1 Conserved C-terminal region 3D7, FC27 42 SQKECTDGNKENC � �9H4 IgG1 Conserved C-terminal region 3D7, FC27 42 SQKECTDGNKENC � �4D11 IgG1 Conserved C-terminal region 3D7, FC27 43, 83, 84 NKENCGAA ��� ���9G8 IgG2b Conserved C-terminal region 3D7, FC27 43, 83, 84 NKENCGAA ��� ���a Scoring was based on the penultimate dilution at which fluorescence was visible above background fluorescence: �, 3 �g/ml; ��, 1 �g/ml; ���, 0.3 �g/ml; ���� 0.1 �g/ml;�����, 0.03 �g/ml; �/�, reactivity with some parasites in the smear at 3 �g/ml; �, no detectable reactivity.

Antigenicity of Merozoite Surface Protein 2

December 2012 Volume 80 Number 12 iai.asm.org 4179

on January 23, 2013 by Alfred H

ealthhttp://iai.asm

.org/D

ownloaded from

FIG 1 Epitopes mapped on MSP2 by ELISA. (A) Epitope mapping with the anti-MSP2 MAbs on 13-mer peptides overlapping by eight residues covering thefull-length 3D7 MSP2 sequence (peptides 1 to 45) and the FC27 MSP2 sequence (peptides 46 to 84) lacking the first 25 residues which are identical to those in3D7 MSP2. OD, optical density. (B) Schematic diagram showing the location of the MAb epitopes identified by the peptide mapping in the two forms of MSP2.

Adda et al.

4180 iai.asm.org Infection and Immunity

on January 23, 2013 by Alfred H

ealthhttp://iai.asm

.org/D

ownloaded from

broadening in the conserved N-terminal region of MSP2, in addi-tion to the expected epitope in the 32-mer repeat in the centralvariable region of FC27 MSP2. This additionally affected regiondoes not correspond to the previously identified secondaryepitope of 8G10 (SNTNS, corresponding to residues 28 to 32)(40). Because 8G10 does not recognize 3D7 MSP2, it seems un-likely that this reflects a genuine interaction with the conservedregion. Rather, it may reflect antibody-mediated self-associationof the N-terminal region (1, 56) as a result of the duplication of theprimary epitope.

Only some MAb epitopes are accessible on the merozoitesurface. All MAb epitopes in the central variable region (11E1,9D11, and 2F2 in 3D7 MSP2 and 8G10 in FC27 MSP2) were ac-cessible in MSP2 on the surface of parasites when fixed smears ofinfected erythrocytes were examined by immunofluorescence mi-croscopy (Table 1; see also Fig. S5 in the supplemental material).The two MAbs (4D11 and 9G8) that recognized an epitope closeto the C terminus also gave a strong IFA signal on the parasiteantigen. In contrast, the other three MAbs reacting with the con-served C-terminal region and MAb 6D8, which reacts with theconserved N-terminal region, gave no signal or a very weak IFAsignal on the parasite antigens.

MAbs reveal antigenic differences between recombinant andparasite MSP2. When the anti-MSP2 MAbs were used to probeWestern blots of samples that had been subjected to SDS-PAGEunder nonreducing conditions, antigenic differences between therecombinant proteins and the corresponding parasite proteinswere also apparent. The three 3D7 MSP2-specific MAbs (11E1,9D11, and 2F2) reacted strongly with both recombinant and par-asite MSP2 although in each case the signal on parasite MSP2 wasclearly stronger (Fig. 3A). Similarly, the FC27 MSP2-specific MAb8G10 reacted strongly with both forms of FC27 MSP2, but againthere was a more intense signal on the parasite antigen. In con-trast, the MAbs that recognized conserved regions of MSP2 gen-erally reacted much more strongly with recombinant MSP2 thanthe parasite antigens (Fig. 3B). MAb 6D8, which reacts with anepitope in the short N-terminal conserved region of MSP2, re-acted equally strongly with both forms of recombinant MSP2, asseen by ELISA, but very weakly with parasite FC27 MSP2 and notat all with the 3D7 parasite antigen (Fig. 3B). Similarly, the threeMAbs (6C9, 1F7, and 9H4) that were shown by peptide mappingto react with an epitope centered on the disulfide bond in theC-terminal conserved region of MSP2 gave very poor, if any, sig-nals when used to probe the two forms of parasite MSP2 on West-ern blots. However, although these three MAbs reacted stronglywith recombinant 3D7 MSP2, they reacted very weakly with re-combinant FC27 MSP2 (Fig. 3B) even though the epitope identi-fied by peptide mapping was common to the two forms of MSP2.The other two MAbs that recognized epitopes in the C-terminalconserved region of MSP2 (4D11 and 9G8) reacted more stronglywith parasite MSP2 but still less strongly than with recombinantMSP2 (Fig. 3B). Although these two MAbs reacted with the sameMSP2 peptides (Fig. 1 and Table 1), their specificities were not

FIG 2 Epitopes mapped on MSP2 by NMR. Backbone amide peak heights for15N FC27 MSP2 (44 �M) in the presence of �15 �M MAb, normalized to thepeak height in the absence of MAb. Data for wild-type MSP2 are shown in blue,and data for C211S MSP2 are shown in red. Vertical bars indicate uncertaintiesin peak heights as estimated from noise levels in 15N-1H SOFAST-HMQCspectra. The locations of epitopes identified using the peptide array data in Fig.1A are indicated as black bars (gray bars are for weaker, cross-reactiveepitopes).

FIG 3 Western blotting on recombinant (lanes 1 and 2) and parasite (lanes 3and 4) 3D7 (lanes 1 and 3) and FC27 (lanes 2 and 4) MSP2. (A) Analyses of thethree 3D7 MSP2-specific MAbs (9D11, 11E1, and 2F2) and the FC27 MSP2-specific MAb 8G10. (B) Analyses of the six MAbs that recognize epitopes in theconserved N-terminal (6D8) or C-terminal (1F7, 6C9, 9H4, 4D11, and 9G8)regions of MSP2 as well as control blots probed with polyclonal sheep anti-3D7and anti-FC27 MSP2 antisera. �, anti.

Antigenicity of Merozoite Surface Protein 2

December 2012 Volume 80 Number 12 iai.asm.org 4181

on January 23, 2013 by Alfred H

ealthhttp://iai.asm

.org/D

ownloaded from

identical; they both reacted more strongly with recombinant 3D7MSP2 than with recombinant FC27 MSP2; but MAb 9G8 reactedequally well with both forms of parasite MSP2, whereas MAb4D11 reacted more strongly with the 3D7 form of parasite MSP2.Further evidence that these two MAbs had different fine specific-ities was the observation that 9G8, but not 4D11, cross-reactedwith a protein of � 200 kDa (Fig. 3B), which from its size andpresence in uninfected red cells (data not shown) was assumed tobe spectrin. Because the amount of MSP2 in parasite lysatesloaded on the gels in these Western blot analyses is not known, wecannot conclude that epitopes in the central variable region ofparasite MSP2 are more available than in the recombinant pro-teins, but we can conclude that they are preferentially displayed,compared with some conserved-region epitopes, in parasiteMSP2.

The failure of some MAbs to give a clear and reproducibleWestern blot signal on the parasite antigens was surprising, giventheir strong reactivity on at least one form of recombinant MSP2.Although some of these MAbs have very low binding affinities,this does not appear to be the sole explanation for this poor reac-tivity because MAb 11E1, which has one of the lowest bindingaffinities (36), gave a strong Western blot signal on the 3D7 para-site MSP2.

The MAbs revealed antigenic differences between the con-served regions of the two forms of MSP2. As discussed above,MAb 6D8 usually reacted weakly with FC27 but not at all with 3D7parasite MSP2 when SDS-PAGE was carried out under nonreduc-ing conditions (Fig. 3B and 4). When this experiment was re-peated using reducing sample buffer, reactivity of MAb 6D8 withFC27 parasite MSP2 was enhanced, but there was still no reactionwith 3D7 parasite MSP2 (Fig. 4, left panel). This antigenic differ-ence between the two forms of MSP2 must reflect differences inthe conformations adopted by the two different parasite forms ofthis largely unstructured antigen. Furthermore, as the 6D8epitope is in the N-terminal conserved region and as the onlydisulfide bond in MSP2 is in the conserved C-terminal region, thisresult suggests an interaction between these regions of MSP2,which may differ between the two forms of parasite MSP2. How-ever, there was an unexpected variability in these results in that the6D8 epitope was not detected in parasite FC27 MSP2 in someexperiments under nonreducing or reducing conditions for SDS-PAGE. This may be a reflection of the unstructured nature ofMSP2, which could result in different conformational ensemblesof MSP2 being analyzed in replicate experiments.

DISCUSSION

The results of these antigenic analyses of recombinant and parasiteMSP2 with a panel of MAbs have provided evidence consistentwith the conserved regions of both the 3D7 and FC27 forms ofparasite MSP2 being more ordered than in the corresponding re-combinant antigens. Previous physicochemical analyses carriedout on the monomeric recombinant proteins showed them to behighly disordered in solution and largely lacking in secondarystructure, characteristics of intrinsically unstructured proteins (1,57). Consistent with this structural characteristic of MSP2, theepitopes of all MAbs studied here mapped to short linear se-quences contained within the variable central region of MSP2 orwithin either the N- or C-terminal conserved regions. In the en-semble of equilibrium conformations that comprise intrinsicallyunstructured proteins such as MSP2, there will be molecules that,at least transiently, are more compact, with disparate regions ofthe polypeptide chain in close proximity. This could possibly re-sult in the generation of conformational B-cell epitopes, but ourresults suggest that this is not a frequent occurrence, and in con-trast to the dominance of conformational B-cell epitopes in glob-ular proteins, linear epitopes predominate in MSP2 and presum-ably other intrinsically unstructured proteins. This is of particularrelevance to malaria vaccine development, given the abundance ofunstructured proteins in Plasmodium proteomes (16, 33).

The linear nature of the epitopes was confirmed by the NMRtechnique also used here to map the epitopes. Upon the additionof MAbs to FC27 MSP2, the loss of backbone amide peaks in1H-15N correlation spectra was largely restricted to the regions ofthe MSP2 sequence identified by the peptide mapping as the cor-responding epitope. The NMR chemical shift is an exquisitely sen-sitive probe of the conformational properties of a molecule. Thus,the fact that no chemical shift changes were observed for anyMSP2 peaks on the addition of antibody suggests that the confor-mation of MSP2, beyond the immediate epitope, was not signifi-cantly perturbed by antibody binding.

Three of the MAbs generated in this study reacted withepitopes in the 3D7 family-specific sequence between the GGSArepeats and the conserved C-terminal region of MSP2. In contrast,the epitope of the control MAb 8G10 has been mapped previouslyto a sequence within the 32-mer repeat that is found in forms ofMSP2 belonging to the FC27 allelic family (15) and that was con-firmed here. All four of these MAbs reacted strongly by IFA withMSP2 in situ on the merozoite surface and gave strong and appro-priately strain-specific signals when used to probe Western blotsof recombinant and parasite MSP2. Thus, the constraints imposedon MSP2 by its anchoring into the merozoite plasma membraneor by possible protein-protein interactions involved in the forma-tion of the merozoite surface coat allow four widely spacedepitopes in the central variable region of the antigen to be readilyaccessible to antibodies. This is consistent with the finding that themajority of naturally occurring anti-MSP2 antibodies react withepitopes in the central variable region of the antigen (14, 42, 55).

Two (4D11 and 9G8) of the six MAbs that reacted withepitopes in conserved regions of MSP2 also gave strong IFA andWestern blot signals on parasite MSP2. The specificities of thesetwo MAbs could not be distinguished by either the peptide orNMR epitope mapping procedures used, but they clearly had dif-ferent fine specificities as one (9G8) but not the other (4D11)cross-reacted with red blood cell spectrin (Fig. 3B and data not

FIG 4 Western blots of parasite 3D7 MSP2 (lanes 1 and 3) and FC27 MSP2(lanes 2 and 4) subjected to SDS-PAGE under reducing (lanes 1 and 2) andnonreducing (lanes 3 and 4) conditions. The 6D8 epitope was enhanced byreduction of FC27 MSP2, the 6C9 epitope was destroyed by reduction of FC27and 3D7 MSP2, and the 9G8 epitope was unaffected by reduction of bothforms of MSP2.

Adda et al.

4182 iai.asm.org Infection and Immunity

on January 23, 2013 by Alfred H

ealthhttp://iai.asm

.org/D

ownloaded from

shown). The epitope(s) recognized by 4D11 and 9G8 was unusualin two respects: first, it was located within a few residues of the at-tachment site for the C-terminal GPI anchor, and, second, theeight-residue sequence identified as the core of the epitope by thepeptide mapping included Cys219 in 3D7 MSP2 (Cys211 in FC27MSP2). This residue forms an intramolecular disulfide bond withCys211 (Cys203 in FC27 MSP2), but the pattern of peptide reac-tivities and NMR epitope mapping with the FC27 C211S mutant(Fig. 2) showed that the reactivity of these two MAbs with MSP2was independent of this disulfide bond being in place. We haveshown previously that this disulfide bond is responsible for somelocalized motional restriction in FC27 MSP2 (57). Nonetheless, itappears that even in the absence of this bond, the local conforma-tion necessary for binding to the 4D11 and 9G8 MAbs is still pop-ulated to a significant extent.

The epitopes of the other four MAbs located in conserved re-gions of MSP2 were either not detected or only weakly stainedwhen parasite smears were examined by IFA. One of these MAbs,6D8, is the only MAb identified that reacts with the short con-served N-terminal region of MSP2, which was previously estab-lished to be responsible for the formation of amyloid-like fibrils byMSP2 (1, 30). The failure to detect this epitope on the merozoitesurface was not due to removal of this region of MSP2 by proteo-lytic processing because polyclonal rabbit antibodies to the N-ter-minal conserved region of the antigen did react with the merozoitesurface by IFA (data not shown). Partial loss of the 6D8 epitopeupon fibril formation by recombinant MSP2 (1) led us to suggestthat amyloid-like intermolecular interactions involving this re-gion of MSP2 may be responsible for the generation of homo-oligomers of MSP2 on the merozoite surface. Cross-linking stud-ies have provided evidence of oligomerization of parasite MSP2(1), and this could explain the loss of the 6D8 epitope althoughother evidence supporting this hypothesis is lacking. Alterna-tively, it is possible that interactions between the N-terminal re-gion of MSP2 and the merozoite membrane may render the 6D8epitope cryptic (31, 57), but this would not explain why the 6D8epitope is not accessible in parasite MSP2 on Western blots.

MAbs 1F7, 6C9, and 9H4 recognized an epitope in the con-served C-terminal region of MSP2 just N-terminal to the 9G8and 4D11 epitope but reacted much more weakly than MAbs9G8 and 4D11 by IFA with the merozoite surface. In contrast tothe 9G8/4D11 epitope, the 1F7/6C9/9H4 epitope was depen-dent on the disulfide bond being in place. However, the weakIFA reactivity of MAbs 1F7, 6C9, and 9H4 was not due to thelack of this disulfide bond, which forms spontaneously in re-combinant and parasite MSP2 (Fig. 4 and data not shown).

Studies with merozoite surface protein 1 (MSP1) and apicalmembrane antigen 1 (AMA1), which have been extensively as-sessed as potential vaccine components, have shown that antibodyfine specificity is an important determinant of GIA activity (11,25). MAbs that block the secondary processing of MSP1 inhibitmerozoite invasion, whereas other MAbs binding to C-terminalepidermal growth factor (EGF)-like domains of MSP1 do not in-hibit (6). A subset of noninhibitory MAbs block the binding ofinhibitory MAbs, with the consequence that their GIA activity isabolished (6, 21). Antibodies that block proteolytic processing ofAMA1 are also inhibitory (13), but other anti-AMA1 antibodiesinhibit merozoite invasion, presumably because they block thebinding of RON2 to AMA1, an interaction which is necessary forthe formation of the moving junction that forms between the

merozoite and erythrocyte membranes to facilitate invasion (8,28, 37, 52). The C-terminal domain of MSP1 and the extracellulardomains of AMA1 have globular folds stabilized by multiple in-tramolecular disulfide bonds (5, 23). Consistent with this, inhib-itory MAbs to these antigens bind to conformational epitopes onthe surface of the proteins, and protective antibody responses areinduced in vivo only by protein constructs with the native disulfidebonds in place (2, 29).

In contrast to MSP1 and AMA1, little information has beenavailable about the fine specificities of anti-MSP2 antibodies orabout the mechanisms by which they are protective. Althoughthere have been reports of antibodies to MSP2 having GIA activity(10, 15, 35), polyclonal rabbit and human antisera we have raisedto MSP2 have no significant inhibitory activity, and in this regardthey contrast markedly with polyclonal rabbit antisera raised toAMA1 using similar procedures (22). Likewise, none of the MAbsstudied here directly blocked merozoite invasion or inhibited in-tracellular growth of the parasite (data not shown). This includedMAb 8G10, which had been reported previously to inhibit mero-zoite invasion (15). Recent studies have reported that human an-tibodies induced by P. falciparum infections or vaccination withMSP2 can mediate ADCI (19, 32, 47), raising the prospect that thismay be a better correlate of protective efficacy in antibodies di-rected against MSP2. It will be important to determine if antibod-ies to the epitopes defined in this study mediate other immuneeffector mechanisms such as ADCI.

MSP1, the most abundant and extensively characterized pro-tein on the merozoite surface, forms a protein complex with twoperipheral membrane proteins, MSP6 and MSP7 (26). Interac-tions with other proteins could result in a more ordered structurefor MSP2 on the merozoite surface with consequences for its an-tigenicity, and/or they could result in the loss of some antigenicepitopes because of steric effects. However, there is no evidence ofMSP2 forming hetero-oligomeric complexes with other merozo-ite surface proteins, and the fact that some conserved-regionepitopes were less accessible in Western blotting as well as the IFAof the parasite antigens suggests that the different conformation isan intrinsic property of the parasite MSP2. This may be a reflec-tion of the GPI moiety attached to the C terminus of parasiteMSP2. GPI anchors have been reported to affect antigenicity orenzymatic activity of proteins to which they are attached (43), butthe results of our preliminary experiments examining the role ofthe GPI anchor on MSP2 antigenicity have been inconclusive.Experiments to explore this further as well as the impact of otherpossible posttranslational modifications of parasite MSP2 areplanned.

The results of this study provide support for the strategy ofdeveloping an MSP2 vaccine comprising the 3D7 and FC27 formsof recombinant MSP2, which would induce antibodies to epitopesin the dimorphic family-specific region of MSP2 that are accessi-ble in MSP2 in situ on the merozoite surface. However, it is ofinterest that some conserved region epitopes (4D11 and 9G8) arealso accessible on the parasite surface, and a vaccine that also in-duced high-titer antibody responses to these epitopes may bemore effective against a wide range of parasite genotypes. Ulti-mately, it will be important to carry out structural analyses of theisolated parasite protein to determine its conformational charac-teristics. This will allow the design of a recombinant MSP2 vaccinecomponent that more closely matches the conformational andantigenic properties of the parasite antigen. The results obtained

Antigenicity of Merozoite Surface Protein 2

December 2012 Volume 80 Number 12 iai.asm.org 4183

on January 23, 2013 by Alfred H

ealthhttp://iai.asm

.org/D

ownloaded from

here, which suggest that the antigenic differences between parasiteand recombinant MSP2 are an intrinsic property of the parasiteantigen and not a consequence of interactions with other mero-zoite surface proteins, encourage us to believe that this may befeasible.

ACKNOWLEDGMENTS

This work was supported by the National Health and Medical ResearchCouncil of Australia (Project Grant 637368 to R.S.N. and R.F.A., ProgramGrant to J.G.B., Infrastructure for Research Institutes Support SchemeGrant to the Burnet Institute, and a Research Fellowship to R.S.N.), theAustralian Research Council (Future Fellowship to J.G.B.), and a Victo-rian State Government Operational Infrastructure Support grant (BurnetInstitute). M.J.B. was supported by an Australian Government APA and aTop-up Award from University of Melbourne, Department of Medicine,Dentistry and Health Sciences.

We have no conflicts of interest to report.

REFERENCES1. Adda CG, et al. 2009. Plasmodium falciparum merozoite surface protein

2 is unstructured and forms amyloid-like fibrils. Mol. Biochem. Parasitol.166:159 –171.

2. Anders RF, et al. 1998. Immunisation with recombinant AMA-1 protectsmice against infection with Plasmodium chabaudi. Vaccine 16:240 –247.

3. Angov E, et al. 2003. Development and pre-clinical analysis of a Plasmo-dium falciparum merozoite surface protein-1(42) malaria vaccine. Mol.Biochem. Parasitol. 128:195–204.

4. Barry AE, Schultz L, Buckee CO, Reeder JC. 2009. Contrasting popu-lation structures of the genes encoding ten leading vaccine-candidate an-tigens of the human malaria parasite, Plasmodium falciparum. PLoS One4:e8497. doi:10.1371/journal.pone.0008497.

5. Blackman MJ, Ling IT, Nicholls SC, Holder AA. 1991. Proteolyticprocessing of the Plasmodium falciparum merozoite surface protein-1produces a membrane-bound fragment containing two epidermal growthfactor-like domains. Mol. Biochem. Parasitol. 49:29 –33.

6. Blackman MJ, Scott-Finnigan TJ, Shai S, Holder AA. 1994. Antibodiesinhibit the protease-mediated processing of a malaria merozoite surfaceprotein. J. Exp. Med. 180:389 –393.

7. Bouharoun-Tayoun H, Attanath P, Sabchareon A, Chongsuphajaisid-dhi T, Druilhe P. 1990. Antibodies that protect humans against Plasmo-dium falciparum blood stages do not on their own inhibit parasite growthand invasion in vitro, but act in cooperation with monocytes. J. Exp. Med.172:1633–1641.

8. Cao J, et al. 2009. Rhoptry neck protein RON2 forms a complex withmicroneme protein AMA1 in Plasmodium falciparum merozoites. Parasi-tol. Int. 58:29 –35.

9. Chang SP, Gibson HL, Lee-Ng CT, Barr PJ, Hui GS. 1992. A carboxyl-terminal fragment of Plasmodium falciparum gp195 expressed by a recom-binant baculovirus induces antibodies that completely inhibit parasitegrowth. J. Immunol. 149:548 –555.

10. Clark JT, et al. 1989. 46 –53 kilodalton glycoprotein from the surface ofPlasmodium falciparum merozoites. Mol. Biochem. Parasitol. 32:15–24.

11. Coley AM, et al. 2001. Rapid and precise epitope mapping of monoclonalantibodies against Plasmodium falciparum AMA1 by combined phage dis-play of fragments and random peptides. Protein Eng. 14:691– 698.

12. Conway DJ, Polley SD. 2002. Measuring immune selection. Parasitology125(Suppl):S3–S16.

13. Dutta S, Haynes JD, Moch JK, Barbosa A, Lanar DE. 2003. Invasion-inhibitory antibodies inhibit proteolytic processing of apical membraneantigen 1 of Plasmodium falciparum merozoites. Proc. Natl. Acad. Sci.U. S. A. 100:12295–12300.

14. Ekala MT, Jouin H, Lekoulou F, Mercereau-Puijalon O, Ntoumi F.2002. Allelic family-specific humoral responses to merozoite surface pro-tein 2 (MSP2) in Gabonese residents with Plasmodium falciparum infec-tions. Clin. Exp. Immunol. 129:326 –331.

15. Epping RJ, et al. 1988. An epitope recognised by inhibitory monoclonalantibodies that react with a 51 kilodalton merozoite surface antigen inPlasmodium falciparum. Mol. Biochem. Parasitol. 28:1–10.

16. Feng ZP, et al. 2006. Abundance of intrinsically unstructured proteins in

P. falciparum and other apicomplexan parasite proteomes. Mol. Biochem.Parasitol. 150:256 –267.

17. Fenton B, et al. 1991. Structural and antigenic polymorphism of the 35-to 48-kilodalton merozoite surface antigen (MSA-2) of the malaria para-site Plasmodium falciparum. Mol. Cell. Biol. 11:963–971.

18. Fluck C, et al. 2004. Strain-specific humoral response to a polymorphicmalaria vaccine. Infect. Immun. 72:6300 – 6305.

19. Flueck C, et al. 2009. Evaluation of two long synthetic merozoite surfaceprotein 2 peptides as malaria vaccine candidates. Vaccine 27:2653–2661.

20. Genton B, et al. 2002. A recombinant blood-stage malaria vaccine reducesPlasmodium falciparum density and exerts selective pressure on parasitepopulations in a phase 1-2b trial in Papua New Guinea. J. Infect. Dis.185:820 – 827.

21. Guevara Patino JA, Holder AA, McBride JS, Blackman MJ. 1997.Antibodies that inhibit malaria merozoite surface protein-1 processingand erythrocyte invasion are blocked by naturally acquired human anti-bodies. J. Exp. Med. 186:1689 –1699.

22. Hodder AN, Crewther PE, Anders RF. 2001. Specificity of the protectiveantibody response to apical membrane antigen 1. Infect. Immun. 69:3286 –3294.

23. Hodder AN, et al. 1996. The disulfide bond structure of Plasmodiumapical membrane antigen-1. J. Biol. Chem. 271:29446 –29452.

24. Hoffmann EH, et al. 2001. Geographical patterns of allelic diversity in thePlasmodium falciparum malaria-vaccine candidate, merozoite surfaceprotein-2. Ann. Trop. Med. Parasitol. 95:117–132.

25. Holder AA. 2009. The carboxy-terminus of merozoite surface protein 1:structure, specific antibodies and immunity to malaria. Parasitology 136:1445–1456.

26. Kadekoppala M, Holder AA. 2010. Merozoite surface proteins of themalaria parasite: the MSP1 complex and the MSP7 family. Int. J. Parasitol.40:1155–1161.

27. Kennedy MC, et al. 2002. In vitro studies with recombinant Plasmodiumfalciparum apical membrane antigen 1 (AMA1): production and activityof an AMA1 vaccine and generation of a multiallelic response. Infect.Immun. 70:6948 – 6960.

28. Lamarque M, et al. 2011. The RON2-AMA1 interaction is a critical step inmoving junction-dependent invasion by apicomplexan parasites. PLoSPathog. 7:e1001276. doi:10.1371/journal.ppat.1001276.

29. Ling IT, Ogun SA, Holder AA. 1994. Immunization against malaria witha recombinant protein. Parasite Immunol. 16:63– 67.

30. Low A, et al. 2007. Merozoite surface protein 2 of Plasmodium falciparum:expression, structure, dynamics, and fibril formation of the conservedN-terminal domain. Biopolymers 87:12–22.

31. MacRaild CA, Pedersen MØ, Anders RF, Norton RS. 2012. Lipid inter-actions of the malaria antigen merozoite surface protein 2. Biochim. Bio-phys. Acta 1818:2572–2578.

32. McCarthy JS, et al. 2011. A phase 1 trial of MSP2-C1, a blood-stagemalaria vaccine containing 2 isoforms of MSP2 formulated with Mon-tanide ISA 720. PLoS One 6:e24413. doi:10.1371/journal.pone.0024413.

33. Olugbile S, et al. 2009. Vaccine potentials of an intrinsically unstructuredfragment derived from the blood stage-associated Plasmodium falciparumprotein PFF0165c. Infect. Immun. 77:5701–5709.

34. Osier FH, et al. 2010. Allele-specific antibodies to Plasmodium falciparummerozoite surface protein-2 and protection against clinical malaria. Par-asite Immunol. 32:193–201.

35. Ramasamy R. 1987. Studies on glycoproteins in the human malaria par-asite Plasmodium falciparum. Identification of a myristilated 45kDa mero-zoite membrane glycoprotein. Immunol. Cell Biol. 65:419 – 424.

36. Reddy SB, et al. 2012. High affinity antibodies to Plasmodium falciparummerozoite antigens are associated with protection from malaria. PLoS One7:e32242. doi:10.1371/journal.pone.0032242.

37. Richard D, et al. 2010. Interaction between Plasmodium falciparum apicalmembrane antigen 1 and the rhoptry neck protein complex defines a keystep in the erythrocyte invasion process of malaria parasites. J. Biol. Chem.285:14815–14822.

38. Rosen O, Anglister J. 2009. Epitope mapping of antibody-antigen com-plexes by nuclear magnetic resonance spectroscopy. Methods Mol. Biol.524:37–57.

39. Sasakawa H, et al. 2007. Ultra-high field NMR studies of antibody bind-ing and site-specific phosphorylation of alpha-synuclein. Biochem. Bio-phys. Res. Commun. 363:795–799.

40. Saul A, et al. 1989. Cross-reactivity of antibody against an epitope of the

Adda et al.

4184 iai.asm.org Infection and Immunity

on January 23, 2013 by Alfred H

ealthhttp://iai.asm

.org/D

ownloaded from

Plasmodium falciparum second merozoite surface antigen. Parasite Im-munol. 11:593– 601.

41. Schanda P, Brutscher B. 2005. Very fast two-dimensional NMR spectros-copy for real-time investigation of dynamic events in proteins on the timescale of seconds. J. Am. Chem. Soc. 127:8014 – 8015.

42. Scopel KK, da Silva-Nunes M, Malafronte RS, Braga EM, Ferreira MU.2007. Variant-specific antibodies to merozoite surface protein 2 and clin-ical expression of Plasmodium falciparum malaria in rural Amazonians.Am. J. Trop. Med. Hyg. 76:1084 –1091.

43. Sharom FJ, Lehto MT. 2002. Glycosylphosphatidylinositol-anchoredproteins: structure, function, and cleavage by phosphatidylinositol-specific phospholipase C. Biochem. Cell Biol. 80:535–549.

44. Shi YP, Udhayakumar V, Oloo AJ, Nahlen BL, Lal AA. 1999. Differentialeffect and interaction of monocytes, hyperimmune sera, and immuno-globulin G on the growth of asexual stage Plasmodium falciparum para-sites. Am. J. Trop. Med. Hyg. 60:135–141.

45. Singh S, et al. 2006. Immunity to recombinant Plasmodium falciparummerozoite surface protein 1 (MSP1): protection in Aotus nancymai mon-keys strongly correlates with anti-MSP1 antibody titer and in vitro para-site-inhibitory activity. Infect. Immun. 74:4573– 4580.

46. Smythe JA, et al. 1990. Structural diversity in the 45-kilodalton merozoitesurface antigen of Plasmodium falciparum. Mol. Biochem. Parasitol. 39:227–234.

47. Stubbs J, et al. 2011. Strain-transcending Fc-dependent killing of Plas-modium falciparum by merozoite surface protein 2 allele-specific humanantibodies. Infect. Immun. 79:1143–1152.

48. Taylor RR, Smith DB, Robinson VJ, McBride JS, Riley EM. 1995.Human antibody response to Plasmodium falciparum merozoite surface

protein 2 is serogroup specific and predominantly of the immunoglobulinG3 subclass. Infect. Immun. 63:4382– 4388.

49. Tebo AE, Kremsner PG, Luty AJ. 2001. Plasmodium falciparum: a majorrole for IgG3 in antibody-dependent monocyte-mediated cellular inhibi-tion of parasite growth in vitro. Exp. Parasitol. 98:20 –28.

50. Theisen M, et al. 1998. The glutamate-rich protein (GLURP) of Plasmo-dium falciparum is a target for antibody-dependent monocyte-mediatedinhibition of parasite growth in vitro. Infect. Immun. 66:11–17.

51. Thomas AW, Carr DA, Carter JM, Lyon JA. 1990. Sequence comparisonof allelic forms of the Plasmodium falciparum merozoite surface antigenMSA2. Mol. Biochem. Parasitol. 43:211–220.

52. Tonkin ML, et al. 2011. Host cell invasion by apicomplexan parasites:insights from the co-structure of AMA1 with a RON2 peptide. Science333:463– 467.

53. Vranken WF, et al. 2005. The CCPN data model for NMR spectroscopy:development of a software pipeline. Proteins 59:687– 696.

54. Weedall GD, Conway DJ. 2010. Detecting signatures of balancing selec-tion to identify targets of anti-parasite immunity. Trends Parasitol. 26:363–369.

55. Weisman S, et al. 2001. Antibody responses to infections with strains ofPlasmodium falciparum expressing diverse forms of merozoite surfaceprotein 2. Infect. Immun. 69:959 –967.

56. Yang X, et al. 2007. A partially structured region of a largely unstructuredprotein, Plasmodium falciparum merozoite surface protein 2 (MSP2),forms amyloid-like fibrils. J. Pept. Sci. 13:839 – 848.

57. Zhang X, et al. 2008. Solution conformation, backbone dynamics andlipid interactions of the intrinsically unstructured malaria surface proteinMSP2. J. Mol. Biol. 379:105–121.

Antigenicity of Merozoite Surface Protein 2

December 2012 Volume 80 Number 12 iai.asm.org 4185

on January 23, 2013 by Alfred H

ealthhttp://iai.asm

.org/D

ownloaded from