Genomes of cryptic chimpanzee Plasmodium species reveal key evolutionary events leading to human malaria

Abstract

African apes harbour at least six Plasmodium species of the subgenus Laverania, one of which gave rise to human Plasmodium falciparum. Here we use a selective amplification strategy to sequence the genome of chimpanzee parasites classified as Plasmodium reichenowi and Plasmodium gaboni based on the subgenomic fragments. Genome-wide analyses show that these parasites indeed represent distinct species, with no evidence of cross-species mating. Both P. reichenowi and P. gaboni are 10-fold more diverse than P. falciparum, indicating a very recent origin of the human parasite. We also find a remarkable Laverania-specific expansion of a multigene family involved in erythrocyte remodelling, and show that a short region on chromosome 4, which encodes two essential invasion genes, was horizontally transferred into a recent P. falciparum ancestor. Our results validate the selective amplification strategy for characterizing cryptic pathogen species, and reveal evolutionary events that likely predisposed the precursor of P. falciparum to colonize humans.

Figure 1. SWGA of Plasmodium parasites.

(a,b) Selection of SWGA primer sets. (a) The average distance (kb) between the 10,000 most frequent parasite motifs (colour coded by length) is plotted for both the P. falciparum (Pf3D7) and human (GRCh37) genomes. The red box highlights motifs that are spaced (on average) <50,000 bp apart in the P. falciparum, but >500,000 bp apart in the human genome. (b) Average distances between the sequence motifs shown in a, but plotted for the P. reichenowi (PrCDC) and chimpanzee (Pan_troglodytes-2.1.4) genomes. Red dots indicate all motifs that fall within the red box in a, with circles and stars denoting those selected for SWGA primer sets 6A and 8A, respectively (Supplementary Fig. 1). (c) Validation of the SWGA primer sets. Human genomic DNA spiked with known quantities of P. falciparum DNA (5–0.001%) was subjected to consecutive rounds of SWGA, using primer set 6A in the first and primer set 8A in the second round. The number of total base pairs (in millions) sequenced is shown in relation to the per cent coverage of the P. falciparum (Pf3D7) genome for five parasite concentrations. DNA mixtures were subjected to two independent amplifications, with individual and combined results shown in blue and red, respectively (the expected genome coverage without SWGA is shown in green). (d) MiSeq read depth for all 14 chromosomes across the Pf3D7 genome shown for one representative amplification at each of five parasite concentrations (individual chromosomes are drawn to scale as indicated on top). For each parasite/human DNA mixture, the percentage of P. falciparum and estimated number of genome copies are indicated. An expanded view of coverage across chromosome 9 is shown in Supplementary Fig. 2.

Figure 2. Sequence diversity within three Laverania species.

The average pairwise nucleotide sequence diversity is shown for 3,111 orthologous core genes at fourfold degenerate sites for 12 geographically diverse strains of P. falciparum (black), two strains of P. reichenowi (red) and two strains of P. gaboni (blue). For P. falciparum field isolates, diversity information was obtained from SNP data (only data sets representing single-parasite strains were used for analysis; see the ‘Methods' section for detail).

Figure 3. Expansion and diversification of the FIKK multigene family in the Laverania subgenus.

(a) Phylogeny of FIKK genes from P. falciparum (green), P. reichenowi (red), P. gaboni (blue) and non-Laverania species (black). FIKK genes are labelled according to their P. falciparum orthologues, with the new P. gaboni gene designated FIKK9.15. Open symbols indicate pseudogenes in all members of the species (FIKK7.2 is intact in some strains of P. falciparum). The tree was inferred using maximum likelihood methods using an alignment of first and second codon positions. Internal branches with bootstrap support of <70% are collapsed. Scale bar, 0.05 substitutions per site. (b) Expression profiles of P. falciparum FIKK genes. Previously published microarray data from 21 clonal P. falciparum strains were used to calculate the expression levels of 19 FIKK genes at different time points during the intra-erythrocytic lifecycle (data for the remaining FIKK gene were not available). Colours represent mRNA expression levels relative to a reference pool, with red, black and green indicating higher, equal and lower expression levels than the reference pool, respectively. Genes were arranged to illustrate the sequential nature of FIKK gene expression.

Figure 4. Genome-wide comparison of Laverania inter-species distances.

Pairwise inter-species distances (substitutions per site) were calculated for all genes (dots) for which orthologs were identified in all three Laverania species (n=4,500), and the P. reichenowi–P. gaboni (x-axis) and P. falciparum–P. gaboni (y-axis) distances were plotted. Four genes that exhibit an unusually high P. reichenowi–P. gaboni and an unusually low P. falciparum–P. gaboni distance are highlighted in red.

Figure 5. Horizontal transfer of two essential invasion genes.

(a) Identification of an 8 kb transferred segment on chromosome 4. Inter-species distances (colour coded) are shown for syntenic orthologs of P. falciparum, P. reichenowi and P. gaboni. Four genes, including the essential invasion genes CyRPA and RH5, exhibit an unusually high P. falciparum–P. reichenowi (yellow) and an unusually low P. falciparum–P. gaboni (aqua) distance, respectively. Genes are ordered by chromosomal location. Since RH4 is absent from P. gaboni (see Supplementary Fig. 4), only the P. falciparum–P. reichenowi distance is shown. (b,c) Phylogenetic relationships of Laverania RH5 and EBA165 sequences, revealing an unexpectedly close relationship between the P. praefalciparum/P. falciparum lineage and the P. adleri lineage in RH5. Laverania parasites are colour coded according to their host species (chimpanzee, red; gorilla, green; human, black). Trees were inferred by maximum likelihood methods. Numbers at internal nodes represent bootstrap support values (only numbers >80% are shown). Scale bar, 0.02 substitutions per site (additional phylogenies are shown in Supplementary Fig. 9). (d) Schematic diagram of the horizontal transfer region on chromosome 4 in P. reichenowi CDC (top), P. falciparum 3D7 (middle) and P. gaboni SY75 (bottom). Genes are shown in blue; grey bars and broken lines indicate genome gaps (numbering is relative to the first base of chromosome 4). A green line indicates the location of the transferred segment in P. falciparum, with the broken part indicating that the 3′-break point is unclear. The region surrounding the 5′-break point is highlighted by a red box and shown in greater detail in Supplementary Fig. 10.