Ungulate malaria parasites

Abstract

Haemosporida parasites of even-toed ungulates are diverse and globally distributed, but since their discovery in 1913 their characterization has relied exclusively on microscopy-based descriptions. In order to bring molecular approaches to bear on the identity and evolutionary relationships of ungulate malaria parasites, we conducted Plasmodium cytb-specific nested PCR surveys using blood from water buffalo in Vietnam and Thailand, and goats in Zambia. We found that Plasmodium is readily detectable from water buffalo in these countries, indicating that buffalo Plasmodium is distributed in a wider region than India, which is the only area in which buffalo Plasmodium has been reported. Two types (I and II) of Plasmodium sequences were identified from water buffalo and a third type (III) was isolated from goat. Morphology of the parasite was confirmed in Giemsa-reagent stained blood smears for the Type I sample. Complete mitochondrial DNA sequences were isolated and used to infer a phylogeny in which ungulate malaria parasites form a monophyletic clade within the Haemosporida, and branch prior to the clade containing bird, lizard and other mammalian Plasmodium. Thus it is likely that host switching of Plasmodium from birds to mammals occurred multiple times, with a switch to ungulates independently from other mammalian Plasmodium.

Figure 1. Bright field microscope images of blood smears stained with Giemsa reagent.

Predicted P. bubalis parasites observed in (a), blood sample from a sick water buffalo in the Chachoengsao Province of Thailand in 2008; and (b), blood sample from a water buffalo in the Mukdahan Province of Thailand in 2015. Type-specific PCR of DNA isolated from the blood of this buffalo indicated that the parasite is Type I.

Figure 2. Phylogenetic relationships of ungulate malaria parasites within the Haemosporida.

Ungulate malaria parasites group together and branch before diversification of malaria parasites that infect bird, lizard, rodent, monkey, apes and human. The tree was inferred by the maximum likelihood (ML) method based on the GTR + I + G model, by using 36 whole mitochondria genome sequences as a concatenated nucleotide sequence data. Bootstrap values (BV) by ML with 100 replicates and Bayesian posterior probability (BPP) by BI are indicated for each internal branch. Collapsed clades are indicated, such as “Vivax and monkey Plasmodium” clade, and their compositions are described in the Methods section. The node at the split of the ungulate clade is indicated by an open circle and the node at base of the bird, lizard, and main mammalian clade is indicated by a closed circle. The length for the substitutions/site (0.05) is indicated.

Figure 3. Phylogenetic relationships of ungulate Plasmodium within the Haemosporida using concatenated coxI, cytb, and clpc nucleotide sequences (codon model).

Model analysis using ModelOMatic indicated that a codon model (AIC score 46566.5) is superior to the nucleotide and amino acid model (AIC score 50700.8 and 65635.5, respectively). Thus we analyzed the phylogenetic relationship using a codon model (GY + I + G + F). The maximum likelihood (ML) tree was inferred using concatenated coxI, cytb, and clpc gene sequences from 52 Plasmodium and related parasites (Supplementary Table S1). Gaps existing in equal to or greater than 50% of the sequences were excluded from the analysis. 0.5 indicates substitutions/site. Bootstrap values (BV) by ML with 1,000 replicates and Bayesian posterior probability (BPP) are indicated for each internal branch. The “fast bootstrap” method implemented in IQ-TREE was applied for the ML bootstrap analysis based on the codon model. The compositions of the collapsed clades are indicated in Supplementary Table S1. The BPP value between the ungulate Plasmodium clade and the Polychromophilus clade did not reach a value of 1.00. ML tree was examined (shaded inset) by regrafting the ungulate Plasmodium clade to the 6 nodes shown on the original ML tree, indicated by circles labeled with italicized red numbers. Values are shown for ΔL, log-likelihood difference to most likely ML tree in the set; SH, Shimodaira-Hasegawa test; WSH, weighted SH test; p, p-value. Tests were performed with 10,000 resamplings using the RELL method. All tests rejected the hypotheses that ungulate Plasmodium and bird, lizard, or other mammalian Plasmodium are monophyletic (asterisks, p < 0.05). However, all tests failed to clarify the relationship between the ungulate Plasmodium and Polychromophilus clades, and thus this relationship warrants further investigation.