A conserved epitope in VAR2CSA is targeted by a cross-reactive antibody originating from Plasmodium vivax Duffy binding protein

Abstract

During Plasmodium falciparum infection in pregnancy, VAR2CSA is expressed on the surface of infected erythrocytes (IEs) and mediates their sequestration in the placenta. As a result, antibodies to VAR2CSA are largely restricted to women who were infected during pregnancy. However, we discovered that VAR2CSA antibodies can also be elicited by P. vivax Duffy binding protein (PvDBP). We proposed that infection with P. vivax in non-pregnant individuals can generate antibodies that cross-react with VAR2CSA. To better understand the specificity of these antibodies, we took advantage of a mouse monoclonal antibody (3D10) raised against PvDBP that cross-reacts with VAR2CSA and identified the epitopes targeted by this antibody. We screened two peptide arrays that span the ectodomain of VAR2CSA from the FCR3 and NF54 alleles. Based on the top epitope recognized by 3D10, we designed a 34-amino acid synthetic peptide, which we call CRP1, that maps to a highly conserved region in DBL3X. Specific lysine residues are critical for 3D10 recognition, and these same amino acids are within a previously defined chondroitin sulfate A (CSA) binding site in DBL3X. We showed by isothermal titration calorimetry that the CRP1 peptide can bind directly to CSA, and antibodies to CRP1 raised in rats significantly blocked the binding of IEs to CSA in vitro. In our Colombian cohorts of pregnant and non-pregnant individuals, at least 45% were seroreactive to CRP1. Antibody reactivities to CRP1 and the 3D10 natural epitope in PvDBP region II, subdomain 1 (SD1), were strongly correlated in both cohorts. These findings suggest that antibodies arising from PvDBP may cross-react with VAR2CSA through the epitope in CRP1 and that CRP1 could be a potential vaccine candidate to target a distinct CSA binding site in VAR2CSA.

Keywords: PvDBP; VAR2CSA; cross-reactivity; malaria; peptide; pregnancy; vaccine design.

Figure 1

3D10 recognizes a conserved epitope in VAR2CSA. (A) Sequence alignment of the most reactive epitope in the peptide arrays. (B) Logo plot of 722 VAR2CSA sequences across Africa and Asia depicting the degree of conservation at each amino acid position of the epitope. Color scheme is based on their chemical properties: polar = green, basic = blue, acidic = red, hydrophobic = black. (C) Cryo-EM of the core structure of NF54 VAR2CSA (DBL1-DBL4; PDB: 7JGH) in the presence of chondroitin sulfate A (magenta). (D) Crystal structure of the VAR2CSA DBL3X domain depicting the specific residues (blue) that interact with the sulfate ion from CSA (PDB: 3CPZ). The epitope from (A) is colored red.

Figure 2

3D10 targets an epitope shared between SD1_CLIPS and CRP1. (A) Titration of 3D10 against CRP1 and its variants coated at 5 µg/ml. (B) Titration of 3D10 against CRP1 and SD1_CLIPS coated at 1 µg/ml. In (A, B), OD values were subtracted from the OD of the isotype control immunoglobin G1 (IgG1) and presented as relative OD. Data are mean ± SD. (C, D) Competition ELISA of 3D10 with SD1_CLIPS (C) or VAR2CSA (D) as the capture antigen. Competing antigens are an unrelated peptide (neg. control; blue), SD1_CLIPS (black) and CRP1 (green). Data are the mean ODs with competitor relative to no competitor ± SD of 3 independent experiments. Significance was determined using a one-way ANOVA with Tukey’s multiple comparisons test (***P ≤ 0.0002; ****P < 0.0001).

Figure 3

CRP1 and its mutants bind to CSA. The left panels depict heat exchange during the binding events between the peptides and CSA. The right panels show the change in enthalpy plotted against the molar ratio of CSA to peptides. Binding isotherms of (A) CRP1, (B) CRP1-K1510A, (C) CRP1-AAA, (D) scrambled peptide, and (E) CRP1 in the presence of 100 mM NaCl. DP – Differential power (μcal/s), ΔH – enthalpy change (kcal/mol).

Figure 4

Anti-CRP1 IgG inhibits IE binding to CSA. (A) The percent inhibition was calculated from the number of IEs bound to CSA after pre-incubation with 200 µg/ml of anti-CRP1 IgG compared with 200 µg/ml IgG from rats immunized with the carrier protein conjugate. Controls for the binding assay included soluble CSA (sCSA) and IgG raised against ID1-ID2a (PAMVAC vaccine construct) as positive controls, and an unrelated peptide (P5) as a negative control. Two rats were immunized for each immunogen and denoted as ‘.1’ and ‘.2’. Data are mean ± SD from three independent experiments and analyzed with one-way ANOVA followed by Tukey’’s multiple comparisons test (****P < 0.0001 for each dataset relative to the anti-P5 controls). (B) Reactivity of rat sera (1/100 dilution) against CRP1, CRP1 mutants and a scrambled peptide generated based on the CRP1 sequence. Sera from rats immunized with only the carrier protein was used as a negative control. Anti-AE-BSA represents pooled sera from two rats.

Figure 5

Seroreactivity to CRP1 correlates strongly with SD1 reactivity. (A) Sera from different populations in Colombia were tested by ELISA for reactivity to CRP1 (1/400 dilution), and (B) SD1_CLIPS (1:200 dilution). Sera from malaria-exposed populations were selected based on VAR2CSA seropositivity. Antibody binding (OD) was converted to arbitrary units (AU) based on the positive control included on every plate. Values are expressed as mean AU ± SD. The dashed line represents the cut-off for CRP1 and SD1_CLIPS seropositivity based on the mean AU (+2 SDs) from sera of individuals living in a non-endemic region of Colombia. (C) Serum reactivity against CRP1 and SD1_CLIPS was correlated using Spearman rank correlation in the non-pregnant cohorts from Puerto Libertador and Tierralta (r_s= 0.8424, P<0.0001, n=137), and (D) pregnant populations (r_s= 0.9449, P<0.0001, n=56). r_s, Spearman rank coefficient.