Insight into antigenic diversity of VAR2CSA-DBL5ε domain from multiple Plasmodium falciparum placental isolates

Abstract

Background: Protection against pregnancy associated malaria (PAM) is associated with high levels of anti-VAR2CSA antibodies. This protection is obtained by the parity dependent acquisition of anti-VAR2CSA antibodies. Distinct parity-associated molecular signatures have been identified in VAR2CSA domains. These two observations combined point to the importance of identifying VAR2CSA sequence variation, which facilitate parasitic evasion or subversion of host immune response. Highly conserved domains of VAR2CSA such as DBL5ε are likely to contain conserved epitopes, and therefore do constitute attractive targets for vaccine development.

Methodology/principal findings: VAR2CSA DBL5ε-domain sequences obtained from cDNA of 40 placental isolates were analysed by a combination of experimental and in silico methods. Competition ELISA assays on two DBL5ε variants, using plasma samples from women from two different areas and specific mice hyperimmune plasma, indicated that DBL5ε possess conserved and cross-reactive B cell epitopes. Peptide ELISA identified conserved areas that are recognised by naturally acquired antibodies. Specific antibodies against these peptides labelled the native proteins on the surface of placental parasites. Despite high DBL5ε sequence homology among parasite isolates, sequence analyses identified motifs in DBL5ε that discriminate parasites according to donor's parity. Moreover, recombinant proteins of two VAR2CSA DBL5ε variants displayed diverse recognition patterns by plasma from malaria-exposed women, and diverse proteoglycan binding abilities.

Conclusions/significance: This study provides insights into conserved and exposed B cell epitopes in DBL5ε that might be a focus for cross reactivity. The importance of sequence variation in VAR2CSA as a critical challenge for vaccine development is highlighted. VAR2CSA conformation seems to be essential to its functionality. Therefore, identification of sequence variation sites in distinct locations within VAR2CSA, affecting antigenicity and/or binding properties, is critical to the effort of developing an efficient VAR2CSA-based vaccine. Motifs associated with parasite segregation according to parity constitute one such site.

Figure 1. High conservation of DBL5ε-VAR2CSA sequences.

(A) Plot of DBL5ε Shannon entropy (H): H = 0: Complete conservation, only one residue present at the given position. 0<H≤1: Considered highly conserved. 1<H≤2: Considered conserved. 2<H≤4.3 considered variable. (B) Three-dimensional model of DBL5ε showing the sequence variability. Heat-map colouring is dark blue (conserved) to red (variable).

Figure 2. VAR2CSA DBL5ε patterns distribution.

(A, B): Sequence logo showing the identified significantly distributed residues I, K and Q The sequence logo displays the residues present at each position, where at least one residue was identified as being significantly distributed with respect to associated numerical parameter. Each letter denotes a given residue and the height corresponds to increasing z-score. The residues are coloured according to: Acidic [ED]: red, Basic [RKH]: blue, Neutral [GNQSTY] = green, Hydrophobic [ACFILMPVW] = black. Numbers below each column denotes corresponding position in the multiple alignment. Letters positioned correctly are associated with high values and upside down letters with low. An asterisk denotes a deletion. It should be noted that in the sequence logos other residues appears (*, E, K), these are however not identified as significantly distributed (i.e. p>0.05). DBL5ε models showing the position of the identified significant residues (red), T₂₇₇, I₂₇₉ (C) and Q₃₀₃ (D).

Figure 3. Mapping of VAR2CSA-DBL5ε signatures.

Based on the identified region of interest and predominant motifs, two representative sequences were selected for homology modelling primigravidae CYK040 (deletion) and multigravidae CYK008 (TFKNI). Blue is CYK040 primigravidae sequence, red is CYK008 multigravidae sequence and dotted circle is deletion/TFKNI motif. The figure illustrates how the conformation of the region depends on the presence or absence of the TFKNI-motif. Using homology modelling, the motif is identified as being surface exposed and may thus alter the immunogenicity of the region.

Figure 4. Bacterial recombinant DBL5ε domain of VAR2CSA expression.

Lysates of untransformed (lane 1) bacteria, DBL5ε_CYK49 [uninduced (lane 2), induced 1mM IPTG (lane 3), induced 0.1 mM IPTG (lane 4)], DBL5ε_CYK49 (lane 5) and DBL5ε_CYK39 (lane 6) after two purification steps were subjected to SDS/PAGE and either stained with Coomassie blue (A) or immunoblotted with either purified IgG multigravidae plasma (B), antisera from mice vaccinated with DBL5ε_CYK39 (C), antisera from DBL5ε_CYK49 vaccinated mice (D) or monoclonal anti-histidine antibodies (E). 30 µg of bacteria-expressed-extract proteins and 2 µg of purified domains were used for analysis. Immune complexes were detected with appropriate horseradish peroxidase coupled antibodies.

Figure 5. CSPG binding of the DBL5ε domain of the VAR2CSA from parasite isolates.

(A): Increasing concentrations of protein were added to wells coated with 5 µg/ml of CSPG. CSPG-binding of the DBL5ε_CYK39 (circle), DBL5ε_CYK49 (triangle) and the non CSA-binding VARO NTS-DBL1α domain used as control (square). Results are the means of three binding assays and the error bars indicate the standard deviation. (B) Inhibition assay. Recombinant DBL5ε variants (5 µg/ml) were pre-mixed with increasing amounts of soluble CSA 0.25–400 µg/ml, and binding to CSPG-coated plates was determined. Results are the means of three inhibition binding assays and error bars indicate the standard deviation. (C): Sequence comparison of VAR2CSA DBL5ε domains from CYK39 and CYK49. Asterisks and circles indicate respectively Cystein residues and Lysine. Conserved amino acids are shown in red and polymorphic residues in black. The 7 loops (L1–L7) identified according to Andersen P et al. are underlined.

Figure 6. Plasma reactivity against DBL5ε domains of VAR2CSA.

(A): Plasma levels of IgG with specificity for DBL5ε domain of VAR2CSA in 8 French unexposed men (M), 16 French unexposed pregnant women (Fra), 75 Senegalese pregnant women (Sen) and 160 Beninese pregnant women (Ben). DBL5ε variants CYK39 and CYK49 were tested. (B): Plasma levels of VAR2CSA DBL5ε domain according to parity. DBL5ε antibodies levels were quantified in the same groups of malaria-exposed pregnant women (Benin and Senegal) as in A. 24 primigravidae (P), 51 multigravidae (M) from Senegal; 80 primigravidae and 80 multigravidae from Benin. (C): Plasma levels of VAR2CSA DBL5ε domain according to parity range. Malaria exposed women used in (B) were separated in three groups; primigravidae (P), women whose parity level is lower or equal to 3 (M≤3) [Beninese women: n = 48, n = 26 for Senegalese women] and those whose parity status is higher than 3 (M>3) [Beninese women: n = 32, n = 25 for Senegalese women]. (D) Correlation between the reactivity to each DBL5ε variant in a given plasma.

Figure 7. Cross-reactive antibody target between VAR2CSA DBL5ε variants.

Cross-reactivity was determined by competition ELISA using either a multigravid plasma pool with high titer of VAR2CSA-specific antibodies (Beninese or Senegalese women), plasma from DBL5ε_CYK39 or CYK49 DNA genetic vaccinated mouse (A). NTS-DBL1α domain of VARO was used as negative control (B). Each colour shows the reactivity with the indicated antibodies.

Figure 8. Reactivity of human specific conserved DBL5ε affinity purified antibodies with P. falciparum infected erythrocytes.

(A): IgG recognition of 3D7-DBL5ε peptides library. (B): Mapping of P4 and P13 peptides on DBL5ε model . (C): Senegalese women antibodies were affinity purified on peptides P4 and P13 and tested for reactivity against PAM Beninese parasite isolates. Flow cytometry analysis of human affinity-purified IgG against peptides P4 and P13 against PAM parasite isolates. Each colour shows the reactivity to native parasites with the indicated antibodies. Four isolates were tested with each IgG. Sample without primary antibody (blank), non-exposed women plasma pool, and exposed women plasma pool are used as control respectively.