Cytochrome c oxidase subunit 1 gene as a DNA barcode for discriminating Trypanosoma cruzi DTUs and closely related species

Abstract

Background: The DNA barcoding system using the cytochrome c oxidase subunit 1 mitochondrial gene (cox1 or COI) is highly efficient for discriminating vertebrate and invertebrate species. In the present study, we examined the suitability of cox1 as a marker for Trypanosoma cruzi identification from other closely related species. Additionally, we combined the sequences of cox1 and the nuclear gene glucose-6-phosphate isomerase (GPI) to evaluate the occurrence of mitochondrial introgression and the presence of hybrid genotypes.

Methods: Sixty-two isolates of Trypanosoma spp. obtained from five of the six Brazilian biomes (Amazon Forest, Atlantic Forest, Caatinga, Cerrado and Pantanal) were sequenced for cox1 and GPI gene fragments. Phylogenetic trees were reconstructed using neighbor-joining, maximum likelihood, parsimony and Bayesian inference methods. Molecular species delimitation was evaluated through pairwise intraspecific and interspecific distances, Automatic Barcode Gap Discovery, single-rate Poisson Tree Processes and multi-rate Poisson Tree Processes.

Results: Both cox1 and GPI genes recognized and differentiated T. cruzi, Trypanosoma cruzi marinkellei, Trypanosoma dionisii and Trypanosoma rangeli. Cox1 discriminated Tcbat, TcI, TcII, TcIII and TcIV. Additionally, TcV and TcVI were identified as a single group. Cox1 also demonstrated diversity in the discrete typing units (DTUs) TcI, TcII and TcIII and in T. c. marinkellei and T. rangeli. Cox1 and GPI demonstrated TcI and TcII as the most genetically distant branches, and the position of the other T. cruzi DTUs differed according to the molecular marker. The tree reconstructed with concatenated cox1 and GPI sequences confirmed the separation of the subgenus Trypanosoma (Schizotrypanum) sp. and the T. cruzi DTUs TcI, TcII, TcIII and TcIV. The evaluation of single nucleotide polymorphisms (SNPs) was informative for DTU differentiation using both genes. In the cox1 analysis, one SNP differentiated heterozygous hybrids from TcIV sequences. In the GPI analysis one SNP discriminated Tcbat from TcI, while another SNP distinguished TcI from TcIII.

Conclusions: DNA barcoding using the cox1 gene is a reliable tool to distinguish T. cruzi from T. c. marinkellei, T. dionisii and T. rangeli and identify the main T. cruzi genotypes.

Keywords: Barcoding; Cytochrome c oxidase subunit 1; Discrete typing units; Glucose-6-phosphate isomerase; Subgenus Trypanosoma (Schizotrypanum); Trypanosoma cruzi.

Fig. 1

Designations of T. cruzi subpopulations through time and according to the method employed. The numbers represent dates of publication. The red star indicates current consensus of T. cruzi nomenclature and subdivisions

Fig. 2

Phylogenetic tree based on cox1 and the geographical origin of the isolates under study. a The cox1 gene differentiates T. cruzi DTUs TcI, TcII, TcIII and TcIV, Tcbat, T. c. marinkellei, T. dionisii and T. rangeli. The heterozygous hybrids TcV and TcVI cannot be differentiated and were placed into the same cluster as TcIV. The numbers at the nodes correspond to NJ, ML, MP and BI support values, respectively (only values >60 are shown). The scale-bar shows the number of nucleotide substitutions per site. The different haplotypes correspond to the diversity observed in cox1 sequences and are represented by colors in the tree. b The map represents the distribution of the characterized isolates among Brazilian biomes. Each colored circle indicates a different haplotype. Circle size represents the number of haplotypes. Abbreviations: Brazilian states: AC, Acre; BA, Bahia; CE, Ceara; ES, Espírito Santo; GO, Goiás; MS, Mato Grosso do Sul; PA, Pará; PI, Piauí; RJ, Rio de Janeiro; SC, Santa Catarina; TO, Tocantins

Fig. 3

Phylogenetic tree based on nuclear gene GPI. a Tree inferred with neighbor-joining method and Kimura-2-parameter model. GPI recognizes and differentiates T. cruzi DTUs TcI, TcII, TcIII and TcIV, T. c. marinkellei, T. dionisii and T. rangeli. One allele of TcV and TcVI sequences cluster with TcII, and the other allele clusters with TcIII. Tcbat was placed in the same cluster as TcI. b The tree inferred from maximum likelihood, parsimony and Bayesian inference. T. cruzi, T. c. marinkellei, T. dionisii and T. rangeli are clearly separated from each other. DTUs TcI and TcII are the most genetically divergent. Tcbat, TcI and TcIII fall in the same cluster. One allele of TcV and TcVI clustered with TcII, and the other allele clustered with the group comprising TcI, TcIII and Tcbat. The numbers at the nodes correspond to ML, MP and BI support values, respectively (only values >60 are shown). The scale-bar shows the number of nucleotide substitutions per site

Fig. 4

Phylogenetic tree based on the concatenation of cox1 and GPI sequences. The concatenated data show a clear separation between T. cruzi DTUs TcI, TcII, TcIII and TcIV, Tcbat, T. c. marinkellei, T. dionisii and T. rangeli. The heterozygous hybrids TcV and TcVI were not differentiated from each other. The numbers at the nodes correspond to NJ, ML, MP and BI support values, respectively (only values >60 are shown). The scale-bar shows the number of nucleotide substitutions per site