A Library of Plasmodium vivax Recombinant Merozoite Proteins Reveals New Vaccine Candidates and Protein-Protein Interactions

Abstract

Background: A vaccine targeting Plasmodium vivax will be an essential component of any comprehensive malaria elimination program, but major gaps in our understanding of P. vivax biology, including the protein-protein interactions that mediate merozoite invasion of reticulocytes, hinder the search for candidate antigens. Only one ligand-receptor interaction has been identified, that between P. vivax Duffy Binding Protein (PvDBP) and the erythrocyte Duffy Antigen Receptor for Chemokines (DARC), and strain-specific immune responses to PvDBP make it a complex vaccine target. To broaden the repertoire of potential P. vivax merozoite-stage vaccine targets, we exploited a recent breakthrough in expressing full-length ectodomains of Plasmodium proteins in a functionally-active form in mammalian cells and initiated a large-scale study of P. vivax merozoite proteins that are potentially involved in reticulocyte binding and invasion.

Methodology/principal findings: We selected 39 P. vivax proteins that are predicted to localize to the merozoite surface or invasive secretory organelles, some of which show homology to P. falciparum vaccine candidates. Of these, we were able to express 37 full-length protein ectodomains in a mammalian expression system, which has been previously used to express P. falciparum invasion ligands such as PfRH5. To establish whether the expressed proteins were correctly folded, we assessed whether they were recognized by antibodies from Cambodian patients with acute vivax malaria. IgG from these samples showed at least a two-fold change in reactivity over naïve controls in 27 of 34 antigens tested, and the majority showed heat-labile IgG immunoreactivity, suggesting the presence of conformation-sensitive epitopes and native tertiary protein structures. Using a method specifically designed to detect low-affinity, extracellular protein-protein interactions, we confirmed a predicted interaction between P. vivax 6-cysteine proteins P12 and P41, further suggesting that the proteins are natively folded and functional. This screen also identified two novel protein-protein interactions, between P12 and PVX_110945, and between MSP3.10 and MSP7.1, the latter of which was confirmed by surface plasmon resonance.

Conclusions/significance: We produced a new library of recombinant full-length P. vivax ectodomains, established that the majority of them contain tertiary structure, and used them to identify predicted and novel protein-protein interactions. As well as identifying new interactions for further biological studies, this library will be useful in identifying P. vivax proteins with vaccine potential, and studying P. vivax malaria pathogenesis and immunity.

Trial registration: ClinicalTrials.gov NCT00663546.

Fig 1. Western blot analysis confirms expression of 34/37 P. vivax recombinant proteins.

Biotinylated proteins were resolved by SDS-PAGE under reducing conditions, blotted, and probed using streptavidin-HRP. All proteins contain a ~25-kDa rat Cd4d3+4 tag. (*) indicates proteins that were run with the right ladder; all others were run with the left or both ladders.

Fig 2. Multiple P. vivax recombinant proteins are immunoreactive and contain conformational epitopes.

The immunoreactivity of 34 biotinylated P. vivax recombinant proteins was assessed using diluted (1:600) plasma pools from 14 Cambodian vivax malaria patients (blue bars) and five American malaria-naïve individuals (green bars). The immunoreactivity of heat-treated proteins was assessed in parallel using the Cambodian plasma pool (red bars); reduced responses indicate the presence of heat-labile conformational epitopes. The immunoreactivity of highly reactive proteins (*) was assessed using more-diluted (1:1000) plasma pools. Absorbance (A) at 405 nm was measured at various times, but only the mean value nearest to 1.0 for each antigen is shown. Negative control (–ve) was rat Cd4d3+d4 tag. Bar charts show mean ± SD; n = 3.

Fig 3. AVEXIS reveals novel interactions involving P. vivax recombinant proteins.

(A) Heat map of the initial P. vivax intra-library AVEXIS, with the intensity of absorbance (A) values at 485 nm and positive putative interactions in red: P12-P41 (bait-prey and prey-bait orientations), P12-PVX_110945 (bait-prey orientation), and MSP3.10-MSP7.1 (bait-prey orientation). (*) indicates baits with low protein levels (< 0.5 μg/ml after concentrating) and preys with activity below the threshold required by the assay. Positive control (+ve) is the P. falciparum P12-P41 interaction. Negative controls (–ve) are rat Cd4d3+d4 tag in (A-C). (B) P12-P41 interaction within and between P. vivax (Pv) and P. falciparum (Pf) proteins by AVEXIS. An interaction between Pv P12-Pv P41 and Pv P12-Pf P41 in both bait and prey orientations. A, absorbance in (B-C). Bar chart shows mean with range; n = 2 in (B-C). (*) indicates n = 1. (C) Replicated P. vivax intra-library AVEXIS using re-synthesized PVX_110945 and MSP7.1 bait proteins, confirming the P12-PVX_110945 interactions in both orientations, and the MSP3.10-MSP7.1 interaction in only the bait-prey orientation.

Fig 4. Quantification of the P. vivax P12-P41 interaction affinity by surface plasmon resonance.

(A) Recombinant, his-tagged P. vivax P12 and P. vivax P41, each eluted as a monodisperse peak after SEC which resolved as a single band of the expected size by SDS-PAGE (insets). (B, C, D) Increasing concentrations of analyte protein were injected over immobilized biotinylated ligand protein. Reference-subtracted binding data were plotted as a binding curve and the equilibrium dissociation constant was calculated using R _eq = CR _max/(C+K _D). Experiments included analyte-ligand combinations P. vivax P12-P41 (B), P. vivax P41-P12 (C), and P. vivax P12-P. falciparum P41 (D). Lower concentrations failed to reach equilibrium, which resulted in an overestimated K _D. SEC, size-exclusion chromatography.

Fig 5. Surface plasmon resonance confirms the P. vivax MSP3.10-MSP7.1 interaction.

(A) Recombinant, his-tagged P. vivax MSP7.1 eluted as a main peak with a small shoulder at higher masses than expected after SEC, likely due to oligomerization, and with a main band of the expected size by SDS-PAGE (inset). (B) Increasing concentrations of P. vivax MSP7.1 were injected over immobilized biotinylated P. vivax MSP3.10. Relatively high-affinity binding was observed, although none of the concentrations used reached equilibrium (inset). The binding did not fit a 1:1 model (red dashed line). The increase in response units at the start of the dissociation phase at the higher analyte concentrations of P. vivax MSP7.1 is likely an artefactual buffer effect. SEC, size-exclusion chromatography.