Phylogeography and taxonomy of Trypanosoma brucei

Abstract

Background: Characterizing the evolutionary relationships and population structure of parasites can provide important insights into the epidemiology of human disease.

Methodology/principal findings: We examined 142 isolates of Trypanosoma brucei from all over sub-Saharan Africa using three distinct classes of genetic markers (kinetoplast CO1 sequence, nuclear SRA gene sequence, eight nuclear microsatellites) to clarify the evolutionary history of Trypanosoma brucei rhodesiense (Tbr) and T. b. gambiense (Tbg), the causative agents of human African trypanosomosis (sleeping sickness) in sub-Saharan Africa, and to examine the relationship between Tbr and the non-human infective parasite T. b. brucei (Tbb) in eastern and southern Africa. A Bayesian phylogeny and haplotype network based on CO1 sequences confirmed the taxonomic distinctness of Tbg group 1. Limited diversity combined with a wide geographical distribution suggested that this parasite has recently and rapidly colonized hosts across its current range. The more virulent Tbg group 2 exhibited diverse origins and was more closely allied with Tbb based on COI sequence and microsatellite genotypes. Four of five COI haplotypes obtained from Tbr were shared with isolates of Tbb, suggesting a close relationship between these taxa. Bayesian clustering of microsatellite genotypes confirmed this relationship and indicated that Tbr and Tbb isolates were often more closely related to each other than they were to other members of the same subspecies. Among isolates of Tbr for which data were available, we detected just two variants of the SRA gene responsible for human infectivity. These variants exhibited distinct geographical ranges, except in Tanzania, where both types co-occurred. Here, isolates possessing distinct SRA types were associated with identical COI haplotypes, but divergent microsatellite signatures.

Conclusions/significance: Our data provide strong evidence that Tbr is only a phenotypic variant of Tbb; while relevant from a medical perspective, Tbr is not a reproductively isolated taxon. The wide distribution of the SRA gene across diverse trypanosome genetic backgrounds suggests that a large amount of genetic diversity is potentially available with which human-infective trypanosomes may respond to selective forces such as those exerted by drugs.

Figure 1. Distribution of 142 Trypanosoma brucei isolates used.

Geographic origin of A) 87 Trypanosoma brucei isolates included in the phylogenetic analysis of partial CO1 sequences and B) 140 T. brucei isolates genotyped at 8 microsatellite loci for population genetic analysis. For each country a triplet of numbers or dashes indicates sample sizes for T. b. brucei (blue), T. b. gambiense (group 1 and group 2 inclusive; green), and T. b. rhodesiense (red).

Figure 2. Phylogenetic tree of 87 Trypanosoma brucei isolates.

50% majority rule consensus tree from the Bayesian analysis of 812 bp of kDNA (CO1) for Trypanosoma brucei and three congeneric outgroups. The frequency with which a particular haplotype was recovered from each of four taxa is indicated in parentheses (left to right: T. b. brucei (blue) / T. b. rhodesiense (red) / T. b. gambiense group 1 (dark green) / T. b. gambiense group 2 (light green)). Clade support values for each node are indicated by Bayesian posterior probability (top) and maximum parsimony bootstrap percentage (bottom). T. b. gambiense group 1 is represented only by haplotypes Hap8 and Hap9; all other T. b. gambiense are group 2. Letters A through C indicate the major clades identified.

Figure 3. Haplotype network.

Maximum parsimony haplotype networks showing genealogical relationships among Trypanosoma brucei kinetoplast haplotypes. Panel A highlights the relationships among lineages of T. b. brucei, T. b. rhodesiense, T. b. gambiense group 1 and T. b. gambiense group 2 (color-coded). Circles are sized proportional to the frequency with which a particular haplotype was recovered. Numbers in the circles correspond to haplotype ID. Empty circles indicate haplotypes that are inferred to exist but were not sampled. Red numbers next to haplotypes containing T. b. rhodesiense indicate the SRA types of the included T. b. rhodesiense isolates (1, SRA type 1; 2, SRA type 2; +, SRA type not known). The light blue boxes correspond to the clades defined in Figure 2. Panel B shows the geographic range of each haplotype.

Figure 4. Genetic structure of Trypanosoma brucei isolates.

Plots show Bayesian clustering of 140 Trypanosoma brucei genotypes based on 8 microsatellite loci and their association with kinetoplast haplotypes and the presence or absence of the SRA gene. Clustering of genotypes is shown for K = 3, 5 and 11 partitions (top three panels). The isolate code is indicated for each isolate. The geographical origin of isolates is indicated by a single letter (A, Angola; B, Botswana; C, Cameroon; D, Democratic Republic of Congo; E, Ethiopia; F, Burkina Faso; H, Chad; I, Ivory Coast; K, Kenya; L, Liberia; M, Mozambique; N, Congo Brazzaville; O, Somalia; Q, Equatorial Guinea; R, Central African Republic; S, Sudan; T, Tanzania; U, Uganda; Z, Zambia). The taxonomic assignment of isolates is indicated by color-coded bars across the fourth panel (T. b. brucei, blue; T. b. rhodesiense, red; T. b. gambiense group 1, dark green; T. b. gambiense group 2, light green). SRA type is indicated by number when known; otherwise just presence (+) or absence (−) of SRA is indicated. Kinetoplast haplotypes (squares, color coded by taxon), when available, are displayed in the bottom panel.

Figure 5. Genetic structure of Trypanosoma brucei isolates inferred from principal components analysis.

Principal component analysis score plot. Points representing individual genotypes are connected by a line to the centroid of an ellipse, which circumscribes a region encompassing 95% of the variance observed within five trypanosome taxa or subgroups identified by STRUCTURE analysis: Tbr (red), Tbb Cluster 2 (dark blue), Tbb non-Cluster 2 (light blue), Tbg group 1 (dark green), Tbg group 2 (light green). The first two principal components (PC1 and PC2) explain 31.2% and 8.5% of the total variance in the data, respectively. One sample of Tbg group 1 was omitted (b028) due to probable misclassification.