Apicoplast phylogeny reveals the position of Plasmodium vivax basal to the Asian primate malaria parasite clade

Abstract

The malaria parasite species, Plasmodium vivax infects not only humans, but also African apes. Human specific P. vivax has evolved from a single ancestor that originated from a parasite of African apes. Although previous studies have proposed phylogenetic trees positioning P. vivax (the common ancestor of human and African ape P. vivax) within the assemblages of Asian primate parasites, its position has not yet been robustly confirmed. We determined nearly complete apicoplast genome sequences from seven Asian primate parasites, Plasmodium cynomolgi (strains Ceylonensis and Berok), P. knowlesi P. fragile, P. fieldi, P. simiovale, P. hylobati, P. inui, and an African primate parasite, P. gonderi, that infects African guenon. Phylogenetic relationships of the Plasmodium species were analyzed using newly and previously determined apicoplast genome sequences. Multigene maximum likelihood analysis of 30 protein coding genes did not position P. vivax within the Asian primate parasite clade but positioned it basal to the clade, after the branching of an African guenon parasite, P. gonderi. The result does not contradict with the emerging notion that P. vivax phylogenetically originated from Africa. The result is also supported by phylogenetic analyses performed using massive nuclear genome data of seven primate Plasmodium species.

Figure 1

Maximum likelihood tree of Plasmodium species. Unambiguously aligned positions of 30 protein coding genes were concatenated, and the resulting 6,937 amino acid and 20,811 nucleotide positions were used for the tree inference. RAxML 7.2.8 program with GTR +Γ model was used for the both amino acid and DNA datasets. For DNA data set, 3 codon positions were partitioned and applied for the program. Tree portion highlighted in (A) was enlarged and shown in (B). Bootstrap analyses were performed for 1000 replicates, and bootstrap values are shown for each internal branch with probabilities assessed by PhyloBayes.

Figure 2

Model test for phylogeny. Eighteen Plasmodium species were classified into 5 groups and an out group, and exhaustive analyses of the 105 trees with three nucleotide substitution model (Model A to C) were applied for the analyses by using PAML 4.8. Akaike Information criterion (AIC) values were calculated to evaluate the most appropriate model among the three. RELL bootstrap probabilities were shown for internal branches of the tree inferred by the codon +Γ model (Model C).

Figure 3

Impact of the substitution model for bootstrap proportion value and the maximum likelihood value of phylogenetic tree. Maximum likelihood trees were inferred using RAxML 7.2.8 program with several amino acid substitution models and 6,937 amino acid positions. Maximum likelihood value (-ln) and bootstrap probability value (BP) on the internal branch, which shows clade P. vivax with Asian primate malaria parasites, were plotted for each analysis, and Akaike Information criterion (AIC) values were calculated to evaluate the most appropriate model.

Figure 4

Tree inference of seven Plasmodium species using nuclear genome encoded genes orthologous to genes on P vivax chromosome 1 to chromosome 4. (A) Maximum Likelihood tree. Unambiguously aligned positions from 627 protein coding genes were concatenated and the resulting 330,500 amino acid and the first and second codon positions of 661,000 nucleotide positions were used for the tree inference. RAxML 7.2.8 program with GTR +Γ model was used for both the amino acid and DNA datasets. For DNA data set, the first and second codon positions were partitioned and applied for the program. (B) Bootstrap probability value (BP) on the internal branches, a to d, shown in (A). Bootstrap analysis was performed for 100 and 1000 replicates for amino acid and DNA dataset, respectively.