Multilocus Sequence Typing schemes have proved to be a valuable tool in determining the genetic structure of pathogens and detection of inter-specific diversity . Even though, these approaches have been originally described for bacterial species, new MLST schemes are being deployed for eukaryotes such as Candida, Batrachochytrium dendrobatidis, Fusarium solani and Leishmania[25–28]. MLST schemes have been useful in order to resolve biological questions such as genetic structure, evolution of reproduction, recombination signals and evolutionary trends of pathogens [19, 29, 30]. Regarding the T. cruzi taxon, two different schemes have been developed for the discrimination of the six reported variants [9, 23]. Nevertheless, to date no robust epidemiological studies have emerged using these methodologies. Herein, we presented the first molecular epidemiology survey using nuclear MLST scheme in order to establish genetic structure and diversity at intra-lineage level (TcI).
The results of the genetic diversity parameters calculated, demonstrated tremendous variability among the 13 loci tested (Table 1). On average 4% of the full concatenated sequences (more than 5 kb) were catalogued as polymorphic sites and more than 150 different genotypes were found. This corroborates the divergent pattern displayed within this DTU. This is not novel, since using single loci such as spliced-leader intergenic miniexon gene, cytochrome b and multilocus strategies as microsatellites and mtMLST schemes have demonstrated a high degree of polymorphisms within TcI populations [15, 17, 18, 24]. Despite of observing a marked degree of polymorphisms, there are certain genes that were considered as non-variable such as RB19, GPI, LAP and TR. These genes are clearly conserved across the six DTUs [13, 31]. This could explain the lack of resolution when applied to our dataset. Likewise, the values of nucleotide diversity confirms that current nomenclature of the T. cruzi taxon is not out of order (on average 0.01517); however, at intra-lineage level this value is elevated which suggest the absence of stability at lineage level and not in accordance with the clonal propagation theory (PCE).
Reliable phylogenetic reconstruction of independent and concatenated gene fragments was carried out (Additional file 1: Figure S1; Figure 1). Most of the tree topologies suggest two clear monophyletic clades with robust bootstrap values. The clones associated to the domestic cycle of transmission (TcIDOM) and the clones associated to the peridomestic/sylvatic cycle of transmission. The trees were compared with the NJ trees obtained using the DSTs observing a total congruence among phylogenetic reconstruction using DNA alignments and cladistics clustering. The intriguing observation was the clear incongruence detected among the topologies of the single gene fragments. This is the case for GPX, GTP, STPP2, RB19, GPI, HMCOAR and TR vs. PDH, RHO1, LAP, SODB, LYT1 and GTP where the topologies do not coincide (Additional file 1: Figure S1). This incongruence at nuclear level may suggest recombination and/or chromosomal rearrangements among the dataset. This event of phylogeny incongruence has been reported in other organisms such as Leishmania, Schistosoma and Giardia duodenalis where genetic exchange has been suggested [32–34]. Despite of this pattern, many authors imply that T. cruzi displays clonal propagation with rare recombination events . This phylogeny incongruence may evidence recombination at a high rate among Colombian TcI populations; the recombination at intra-TcI level has already been reported in Ecuadorian populations and at mitochondrial level in Colombia where introgression has also been described [5, 19]. In addition, the chromosomal rearrangement not attributed to recombination but attributed to genomic reassortment is not a point to exclude. Recently Lima et al. have reported inter- and intra-strain karyotype heterogeneities suggesting that chromosomal rearrangements have occurred during the evolution of T. cruzi[36, 37]. When the genes that showed different topology were analyzed (GPX, GTP, GPI and TR), it was not rare to observe they may be under chromosomal rearrangements since they are located in four independent chromosomes (35, 12, 6 and 37) respectively .
The DSTs were employed to test evolutionary hypotheses such as linkage disequilibrium and possible founders of specific genotypes using LIAN and eBURST software. The results showed that population is at linkage equilibrium and the values of the association index support the idea of recombination. Likewise, this rate of recombination is supported upon the results provided by eBURST software. The linkage equilibrium is the best-fit measure of recombination among a dataset and widely used specifically in organisms where clonal propagation has been evinced [38–40]. Hence, the significance in the absence of linkage disequilibrium allows us to state that genetic exchange is a frequent mode of propagation among Colombian T. cruzi I natural populations. The foreseen absence of preponderant clonality has been reported at nuclear level using microsatellites in this dataset. Likewise, we observed a lack of congruence between nuclear and mitochondrial phylogenies (data not shown) demonstrating for the first time using mtMLST and nMLST schemes the presence of cryptic sexuality within T. cruzi. Additionally, the number of DSTs obtained was equal to the number of clones analyzed demonstrating the absence of lineage stability in space and time that is not in accordance with the PCE model.
For some authors, the evidence of recombination cannot be pointed out just by detecting incongruence among phylogenies regardless of using nuclear and/or mitochondrial DNA . Nevertheless to deal with this issue we submitted the whole alignment (> 5 kb of nuclear DNA) for the detection of recombination signals using RDP. We detected recombinants using the MAXCHI algorithm finding three robust clones with strong signals of recombination (Figure 2). In all the recombinants, we were able to distinguish minor and major parents among the breaking points. Curiously, the breaking point for the three recombinants was detected in the GTP gene. The three recombinants were isolated from human infections in acute phase and among the parents, there is always one clone belonging to the sylvatic cycle of transmission (isolated from Rhodnius prolixus). This finding is paradoxically interesting in terms of detecting emerging genotypes within this DTU and also the enigmatic question regarding the place across the life cycle of T. cruzi where recombination occurs. These recombinants and their major and minor parents coincide with the finding of DSTs using eBURST. In this case, two events of recombination emerged obtaining a first ancestry recombination event between one clone from Triatoma maculata (TmPA1cl6) and one clone from Alouatta seniculus (YAS1cl2); both from an arboreal niche in the sylvatic cycle of transmission of Chagas disease. Moreover, the second event is figured out by one clone from R. prolixus (N5P14cl3) and one from human oral infection (LERcl14) giving rise to the specific genotype TcIDOM. This is of paramount importance because the recombination signals are always reported in at least one isolate from triatomine and might be implying that recombination may be taking place in the reduviidae insects and not within mammal reservoirs and/or hosts. Insect vectors have played a remarkable role in the detection of recombination in different parasitic protozoa; in this sense, recombination signals have been detected in vitro in Phlebotomus for the case of Leishmania donovani and Glossina for the case of T. brucei respectively [40, 42]. The recombination among parasitic protozoa plays an important role in the diversification of these microorganisms and has been detected in vitro and in vivo; and needs to be related to the severity of parasitic diseases.
The phylogenetic reconstruction using the DNA alignments and DSTs profiles based on the nMLST scheme allowed us to establish two robust monophyletic clades named TcIDOM and one associated to peridomestic/sylvatic clones. This grouping has been established since 2007 when the first subdivision based on spliced leader of the mini-exon gene was conducted . Other researchers continued these efforts but most of the associations were made upon the cycles of transmission within TcI [18, 43–45]. This subdivision may not be absolute because in some cases domiciliated insect vectors may cluster with sylvatic population that is not unlikely since they may carry populations of the sylvatic foci as natural predictors of parasite transmission dynamics. In this case, it is more accurate to name the genotypes associated to human infection as an emergent genotype. Robust coalescent Bayesian dating suggest that this genotype emerged approximately 23 000 ± 12 000 years ago and followed by population expansion, broadly corresponding with the earliest human migration into the Americas . Emergent genotypes have been reported in other parasites such as Plasmodium and Leishmania[46–48] as a strong signal of recombination. When a detailed analyses of the number of domestic and peridomestic/sylvatic genotypes was observed among the polymorphic genes (Figure 1). We could determine a high number of genotypes in the peridomestic/sylvatic clones as strong signals of cryptic sexuality but a low number of genotypes among TcIDOM clones. This likely suggest that T. cruzi exhibits a high frequency of genetic exchange with a later clonal expansion of specific genotypes that become stable in space and time. Therefore, TcIDOM has become and might be the sibling of multiple recombination events. In terms of parasite success, as the parasite interacts with its host, the parasite becomes less virulent and produces pathologies that are more often associated with long chronic disease; this is the case of TcIDOM that has been related with severe forms of end-stage chronic cardiomyopathy in Colombia and Argentina compared to sylvatic strains which support the premises herein mentioned .
In terms of evolutionary trends, we calculated the ratio of nonsynomymous to synonymous changes (Table 1). In these calculations, patterns of positive selection (STTP2, RHO1 and TR) and/or stabilizing selection were detected (GPX, HCOAR, PDH, GTP, SODA, SODB, LAP, GPI, LYT1 and RB19). When these genes were applied to the six DTUs within the T. cruzi taxon, it was possible to determine that most genes were under stabilizing selection and just a minority under positive selection . In terms of natural history of TcI the fact of detecting that most of the genes are under stabilizing selection implies that the emergence of genotypes within this DTU is possible and this genotype may be orthodoxically an event of recombination between genes that probably display positive selection favouring the appearance of alleles that increase the frequency of certain trait among the population .