Diversity of T cell epitopes in Plasmodium falciparum circumsporozoite protein likely due to protein-protein interactions

Abstract

Circumsporozoite protein (CS) is a leading vaccine antigen for falciparum malaria, but is highly polymorphic in natural parasite populations. The factors driving this diversity are unclear, but non-random assortment of the T cell epitopes TH2 and TH3 has been observed in a Kenyan parasite population. The recent publication of the crystal structure of the variable C terminal region of the protein allows the assessment of the impact of diversity on protein structure and T cell epitope assortment. Using data from the Gambia (55 isolates) and Malawi (235 isolates), we evaluated the patterns of diversity within and between epitopes in these two distantly-separated populations. Only non-synonymous mutations were observed with the vast majority in both populations at similar frequencies suggesting strong selection on this region. A non-random pattern of T cell epitope assortment was seen in Malawi and in the Gambia, but structural analysis indicates no intramolecular spatial interactions. Using the information from these parasite populations, structural analysis reveals that polymorphic amino acids within TH2 and TH3 colocalize to one side of the protein, surround, but do not involve, the hydrophobic pocket in CS, and predominately involve charge switches. In addition, free energy analysis suggests residues forming and behind the novel pocket within CS are tightly constrained and well conserved in all alleles. In addition, free energy analysis shows polymorphic residues tend to be populated by energetically unfavorable amino acids. In combination, these findings suggest the diversity of T cell epitopes in CS may be primarily an evolutionary response to intermolecular interactions at the surface of the protein potentially counteracting antibody-mediated immune recognition or evolving host receptor diversity.

Figure 1. WebLogo of Amino Acid Sequence of Circumsporozoite Protein from Malawi and the Gambia.

Panel A and Panel B are the Weblogos for Malawi and the Gambia, respectively. In Panel A, the TH2 region (blue) and TH3 region (pink) are underlined. The TH2 epitope maps almost exclusively to the α-helix, while the TH3 epitope maps to the flap. The polymorphic residues and types of amino acids that populate these sites appear to be conserved between two geographically disparate African parasite populations. Bits represent the information content, which is a relative measurement of sequence conservation, with higher values representing conservation and lower values consistent with sequence diversity at a position.

Figure 2. Nucleotide Diversity of pfcsp from Malawi and the Gambia.

The nucleotide diversity (π) was determined for the sequenced region using a sliding window (size: 50 bp, slide: 25 bp). Two peaks in diversity are seen corresponding to the TH2 (bp 897–972) and TH3 (bp 1046–1091) epitopes. The peak level of diversity and pattern of diversity was similar between the two populations.

Figure 3. TH2 and TH3 Pairings from Malawi and the Gambia.

(a) This figure shows all TH2-x/TH3-y haplotype pairings comparing observed and expected. Those that are that are statistically over and under represented based upon our contingency table analysis (p≤0.00009 for Malawi and p≤0.003 for the Gambia) are colored blue. (b) The data shown are for those pairings in Malawi (blue) and the Gambia (green) that are either observed >5 times in our data, or those predicted to occur >5 times based upon our contingency analysis. Each pairing is represented a unique symbol. Of note, three pairings (TH2-1/TH3-1, TH2-6/TH3-1, and TH2-3/TH3-2) were over represented in both populations. In both figures, the diagonal line represents if there were non-random association of pairings based on predicted and observed values. Points above the line represent pairing over represented in the population, while those represented below the line are those under represented in the population. A complete list of these significant pairings is provided in Table S2 and S3.

Figure 4. Amino Acid Distribution in TH2 and TH3.

The distributions of amino acids in the TH2 (Panel A) and TH3 (Panel B) epitopes are shown. The reference 3d7 amino acid sequences are shown on the X axis. Above the amino acid is a column representing the frequency of amino acids seen in isolates from Malawi (left half of the column) and the Gambia (right half of the column). The relative frequency of each amino acid is similar between the two geographically-distant African populations.

Figure 5. Location of TH2 and TH3 epitopes in the Structure of Circumsporozoite Protein.

The relative location of TH2 (blue) and TH3 (pink) are shown within the protein structure.

Figure 6. Sites of Polymorphism Mapped to the CS structure.

The variable amino acids within TH2 and TH3 are shown in Frontal (Panel A), Back (Panel B) and Pocket (Panel C) views. Panel C includes solvent as spheres. Based upon the sequence logo for Malawi, the information content (IC) in bits binned at increments of 0.25 was mapped to the crystal structures via a color scheme indicating the magnitude. The information content is a relative measurement of sequence conservation, with higher values being more highly conserved and lower values having more diversity at that sequence position. This value is based upon the R_seq, which is determined by the difference between the maximum possible entropy and the entropy observed in the distribution at that location . The maximum sequence conservation per site is dependent on the number of distinct symbols possible (20 for amino acids) and is therefore 4.32 bits for protein sequences. In this figure, red residues (IC<2.75) are the most highly mutable, followed by yellow (IC <3.0). The majority of residues that are highly variable face the external matrix and are not associated with the back of the molecule or included within the pocket.

Figure 7. Significantly energetically-constrained amino acid positions identified by MUMBO Analysis.

The 5 amino acids identified by MUMBO as having constrained ΔΔG mutational profiles relative to other identical amino acids within the crystal structure are color coded: ASN340 (red), Gly341 (orange) Ile342 (yellow) Glu343 (green), and Ser332 (cyan), are shown with respect to the TH2 and TH3 domains’ surface area (colored as in Figure 4). These residues cluster behind the conserved hydrophobic pocket and were identified because there mutational profile differed on average by 2 standard deviations from all other identical residues within the crystal structure.

Figure 8. Calculated ΔΔG of observed polymorphic amino mutations from the ancestral amino acid residue compared relative to median of all possible mutations at each position.

Free energy changes of polymorphisms in TH2 and TH3 are shown relative to the median change from all 19 substitutions from the predicted ancestral allele determined from Plasmodium sp. phylogeny. Mutations that have higher energy than the median are shown in red, while those with lower energy are shown in blue. Positive values represent increases in free energy and thermodynamic instability while negative values represent decline in free energy and greater stability. Neutral sequence where energetics have no effect would be expected to occur 50/50 above and below the median, while conservation of intramolecular function would be expected to minimize entropy and lead to lower energy states. Intermolecular interactions can lead to selection for less favorable states which are significantly enriched in the observed polymorphisms (17 increased vs 5 decreased, p = 0.00845).