On the diversity of malaria parasites in African apes and the origin of Plasmodium falciparum from Bonobos

Abstract

The origin of Plasmodium falciparum, the etiological agent of the most dangerous forms of human malaria, remains controversial. Although investigations of homologous parasites in African Apes are crucial to resolve this issue, studies have been restricted to a chimpanzee parasite related to P. falciparum, P. reichenowi, for which a single isolate was available until very recently. Using PCR amplification, we detected Plasmodium parasites in blood samples from 18 of 91 individuals of the genus Pan, including six chimpanzees (three Pan troglodytes troglodytes, three Pan t. schweinfurthii) and twelve bonobos (Pan paniscus). We obtained sequences of the parasites' mitochondrial genomes and/or from two nuclear genes from 14 samples. In addition to P. reichenowi, three other hitherto unknown lineages were found in the chimpanzees. One is related to P. vivax and two to P. falciparum that are likely to belong to distinct species. In the bonobos we found P. falciparum parasites whose mitochondrial genomes indicated that they were distinct from those present in humans, and another parasite lineage related to P. malariae. Phylogenetic analyses based on this diverse set of Plasmodium parasites in African Apes shed new light on the evolutionary history of P. falciparum. The data suggested that P. falciparum did not originate from P. reichenowi of chimpanzees (Pan troglodytes), but rather evolved in bonobos (Pan paniscus), from which it subsequently colonized humans by a host-switch. Finally, our data and that of others indicated that chimpanzees and bonobos maintain malaria parasites, to which humans are susceptible, a factor of some relevance to the renewed efforts to eradicate malaria.

Figure 1. Phylogenetic tree of Plasmodium based on mitochondrial genomes.

In the Bayesian phylogenetic tree presented, the values above branches are posterior probabilities expressed as percentages. Maximum likelihood and Bayesian methods lead to identical phylogenies. The names of the species that normally infect humans or chimpanzees are presented in bold. The sequence of the mitochondrial genomes derived from the Ape samples were named (also presented in bold) according to the country in which an Ape was sampled (DRC or UG, which stand for the Democratic Republic of Congo and Uganda, respectively), followed in parentheses by a single letter that indicates the particular Ape from which the sequence was obtained, and a number when two or more distinct sequence were obtained from the sample. Theses names were colour-coded according to the host species, indicated on the right, from which the sequences were derived (Pan t. troglodytes in blue; Pan t. schweinfurthii in red, and Pan paniscus in green). The Laverania clade is highlighted in yellow, and the branches carrying the sequences from the two novel lineages are labelled as the new species to which we propose they belong. The accession numbers of the sequences derived from the parasites found in chimpanzees and bonobos are provided in Table S1, and those of the other species are provided in the Methods.

Figure 2. Phylogenetic analyses of the Laverania group based on the dhfr-ts.

We report four dhfr-ts alleles, DRC (Sd1), DRC (Sd2) and DRC (Sd3) derived from the sample collected from one Ape (Shegue), and DRC (Id) derived from a sample from another Ape (Itaito). The DRC (Sd1) allele corresponds to the P. reichenowi sequence. In view of the similarity with the mitochondrial genome tree topology and the apparent lack of mixed species infection in the two animals from which sequences were obtained, we tentatively considered that DRC (Sd2) and DRC (Sd3) originate from P. billbrayi parasites, and DRC (Id) from P. billcollinsi (hence the quotation marks). Bayesian support for the nodes was inferred through a Monte Carlo Markov chain model as implemented in Mr. Bayes, with 10,000,000 generations after a “burn-in” of 3,000,000 generations. Sampling was performed every 100 generations. Mixing of the chains was properly checked after runs. Two phylogenies are presented for the gene encoding dhfr-ts. A. Phylogeny A (1789 bp), which included the P. falciparum (XM_001351443) and P. reichenowi (GQ369533, this study) dhfr-ts sequences and the four from parasites of Apes, reproduces the topology obtained from the mitochondrial genome. B. Phylogeny B (1690 bp aligned) uses rodent malarial parasites P. berghei and P. yoelii as outgroups, differs from the mitochondrial phylogeny by placing the root of the Laverania group within P. billbrayi alleles that are no longer monophyletic. We favour the phylogenetic hypothesis A over B since the latter is based on fewer base pairs and excludes an area with phylogenetic information among the Laverania species; such an area is not found in rodent or any other Plasmodium species so it is excluded from the phylogenetic analyses. Indeed, P. reichenowi (NC_002235 and DRC (Sd1)) cannot be clearly separated from P. falciparum indicating that rodent malarias may be too distant to serve as a reliable out-group for dhfr-ts.

Figure 3. Alignment of the msp2 block 3 sequences obtained from Pan troglodytes sp.

The predicted amino acid sequence of one member from each of the five msp2 block 3 allelic families uncovered from the Plasmodium parasites present in the chimpanzee samples. The alignment (Clusal V, DNASTAR Lasergene MegAlign version 7.2.1) comparisons were made against the only known P. reichenowi msp2 block 3 sequence , denoted “Pr” (Y14731). The msp2 block sequences obtained during our analysis were named according to the geographic origin of the samples “KNP” (Kibale National Park), followed by the sequence family (Pr for the P. reichenowi type in blue, and A to D for the others in black). Each distinct sequence found within each family was assigned a sequential number. The origins, names and accession numbers of all the msp2 block 3 sequences obtained in this study are provided in Table S1. In the alignment presented the representative sequences from the five allelic families that were included are: KNP-Pr (Prmsp2-A1, GU075719), KNP-A (msp2-KNP-A1, GU075722), KNP-B (msp2-KNP-B, GU075724), KNP-C (msp2-KNP-C1, GU075725) and KNP-D (msp2-KNP-D, GU075726). Stars (*) represent residue similarity and dashes (−) represent gaps.

Figure 4. Mitochondrial haplotype map for P. falciparum populations found in humans and in bonobos.

The mitochondrial genomes from parasite lines collected from humans and of the four obtained from parasites in bonobos, DRC (A), DRC (C), DRC (E) and DRC (L), were used to obtain the haplotype network presented. It was inferred under a median joining algorithm with posterior pruning using maximum parsimony criteria as implemented in Network 4.1.1.2 . The size of the circles is proportional to the haplotype frequency with each colour indicating which were derived from P. falciparum collected from bonobos, and the geographical origin of the sequences from P. falciparum collected from humans.