Population genomics of the immune evasion (var) genes of Plasmodium falciparum

Abstract

Var genes encode the major surface antigen (PfEMP1) of the blood stages of the human malaria parasite Plasmodium falciparum. Differential expression of up to 60 diverse var genes in each parasite genome underlies immune evasion. We compared the diversity of the DBLalpha domain of var genes sampled from 30 parasite isolates from a malaria endemic area of Papua New Guinea (PNG) and 59 from widespread geographic origins (global). Overall, we obtained over 8,000 quality-controlled DBLalpha sequences. Within our sampling frame, the global population had a total of 895 distinct DBLalpha "types" and negligible overlap among repertoires. This indicated that var gene diversity on a global scale is so immense that many genomes would need to be sequenced to capture its true extent. In contrast, we found a much lower diversity in PNG of 185 DBLalpha types, with an average of approximately 7% overlap among repertoires. While we identify marked geographic structuring, nearly 40% of types identified in PNG were also found in samples from different countries showing a cosmopolitan distribution for much of the diversity. We also present evidence to suggest that recombination plays a key role in maintaining the unprecedented levels of polymorphism found in these immune evasion genes. This population genomic framework provides a cost effective molecular epidemiological tool to rapidly explore the geographic diversity of var genes.

Figure 1. Cumulative Diversity Curves for DBLα from the Amele and Global Populations

The cumulative diversity of DBLα was determined by comparing each new DBLα repertoire to previous repertoire(s) using a 96% DNA sequence identity cut-off. We plotted the number of accumulating “types” (distinct sequences) against the number of “sequences” compared. For example, the first point on the plot will be the number of types found in both isolates A and B on the y-axis against the total number of sequences compared on the x-axis. The next point will be the number of types found in isolates A, B, and C against the total number of sequences compared. The plot shown here is the average curve of 1,000 permutations of the order that isolates were compared (i.e., isolates A, B, then C; or A, C, then B; or B, C, then A).

Figure 2. PTS among DBLα Repertoires from Amele and Global Populations

PTS was plotted for pairs of Amele DBLα repertoires, global DBLα repertoires, and Amele DBLα repertoires paired with global DBLα repertoires. The median (thick line), interquartile range (boxes), approximate 95% sampling interval (whiskers = 1.5 times the interquartile range), and suspected outliers (asterisks) are shown. The raw data for these plots can be found in Figure S5.

Figure 3. Diversity of Amele DBLα Sequences

Relationships among the var gene sequences from Amele were defined by (A) phylogenetic analysis of Amele DBLα types. A multiple alignment of amino acid sequences was constructed in Clustal W [70] with parameters set at gap opening = 5 and gap extension = 0.05, for all translated DBLα types (n = 177 after removal of those with frame shifts) in the Amele population. A neighbor-joining tree was constructed in MEGA version 3.0 [71] for Amele types using the p-distance algorithm, pairwise alignment, and 1,000 bootstrapping repetitions. The consensus tree is shown. (B) Analysis of the pairwise alignment quality scores for Amele DBLα amino acid sequences. Here, the frequency distribution of scores is shown. Note the approximately Normal distribution of scores for the cluster of scores to the left, i.e., the dark line indicates maximum likelihood estimate for the Normal distribution. The asterisk indicates a cluster of pairs with high quality scores, which correspond to the shared sequences among isolates (types).