Adaptations in energy metabolism and gene family expansions revealed by comparative transcriptomics of three Chagas disease triatomine vectors

Abstract

Background: Chagas disease is a parasitic infection caused by Trypanosoma cruzi. It is an important public health problem affecting around seven to eight million people in the Americas. A large number of hematophagous triatomine insect species, occupying diverse natural and human-modified ecological niches transmit this disease. Triatomines are long-living hemipterans that have evolved to explode different habitats to associate with their vertebrate hosts. Understanding the molecular basis of the extreme physiological conditions including starvation tolerance and longevity could provide insights for developing novel control strategies. We describe the normalized cDNA, full body transcriptome analysis of three main vectors in North, Central and South America, Triatoma pallidipennis, T. dimidiata and T. infestans.

Results: Two-thirds of the de novo assembled transcriptomes map to the Rhodnius prolixus genome and proteome. A Triatoma expansion of the calycin family and two types of protease inhibitors, pacifastins and cystatins were identified. A high number of transcriptionally active class I transposable elements was documented in T. infestans, compared with T. dimidiata and T. pallidipennis. Sequence identity in Triatoma-R. prolixus 1:1 orthologs revealed high sequence divergence in four enzymes participating in gluconeogenesis, glycogen synthesis and the pentose phosphate pathway, indicating high evolutionary rates of these genes. Also, molecular evidence suggesting positive selection was found for several genes of the oxidative phosphorylation I, III and V complexes.

Conclusions: Protease inhibitors and calycin-coding gene expansions provide insights into rapidly evolving processes of protease regulation and haematophagy. Higher evolutionary rates in enzymes that exert metabolic flux control towards anabolism and evidence for positive selection in oxidative phosphorylation complexes might represent genetic adaptations, possibly related to prolonged starvation, oxidative stress tolerance, longevity, and hematophagy and flight reduction. Overall, this work generated novel hypothesis related to biological adaptations to extreme physiological conditions and diverse ecological niches that sustain Chagas disease transmission.

Keywords: Chagas disease; Reduviid bugs; Transcriptome, metabolism, oxidative phosphorylation.

Fig. 1

Putative orthology search within Triatoma and R. prolixus. Bidirectional BLASTX and TBLASTN searches were performed between all four datasets, using the BLAST best reciprocal hit strategy to define putative orthologs. The numbers in the non-overlapping areas correspond to the total number of transcripts in each species. We identified 4054 1:1 orthologs in the four species. Whereas T. pallidipennis and T. dimidiata shared more than 10 thousand 1:1 orthologs, they shared around 7600 with T. infestans, recapitulating species phylogeny

Fig. 2

Structural characterization of the pacifastin family in the three - Triatoma species and R. prolixus. Based on InterPro annotations, transcripts of the non-redundant dataset (including singletons) were classified according to the presence of signal peptide, transmembrane regions and other InterPro domains. Absolute numbers for each type and species are shown in the matrix. Apart from the increase in gene number, particularly observed in T. pallidipennis and T. dimidiata (Table 4), we observed divergence and evolutionary innovation in secreted and membrane-bound pacifastins (red numbers). VWFC, von Willebrand Factor C domain; VWFD, von Willebrand Factor D domain. A detailed description of the pacifastin domain coding transcripts in Triatoma is found in Additional file 1

Fig. 3

Clustering analysis of the Calycin/lipocalin gene family with CLANS in the three-Triatoma and R. prolixus. 2D representation of the clusters obtained through the CLANS analysis. A large calycin expansion was identified in the three-Triatoma species, but more remarkable in T. infestans. Although such expansion involves typical salivary lipocalins (clades I, II, III, IV, V, VI and VII, and nitrophorins) two novel non-salivary clade were identified: the Intestinal clade (purple) and the FABP (pink). The colored dots in the graph shown indicate sequences included in each calycin group, while black dots represent divergent sequences that do not cluster within any predefined clade. A detailed description of the calycin coding transcripts in Triatoma is found in Additional file 2

Fig. 4

High transposable element (TE) transcriptional activity in T. infestans. a Absolute number of TEs classified according to type, and b) species of origin for the three - Triatoma species. For T. infestans only, c) DNA sequence identity (%) for each TE best-match according to TE type, and d) according to TE taxonomic origin. e) DNA sequence identity (%) in DNA TE best-match according to taxonomic origin. f) DNA sequence identity in LTRs best-match according to taxonomic origin. g) DNA sequence identity in R. prolixus best-match TEs according to TE type. h) DNA sequence identity in bird best-match TEs according to TE type. The increase in the absolute number TE in T. infestans is due to putative bird-derived non-LTR TEs. Higher sequence conservation suggests recent horizontal transfer; possibly form chicken, an important source of blood for T. infestans. A Kruskal-Wallis test with Dunn’s multiple comparison test was performed (*** p < 0.001). A full description of TEs analysis is in Additional file 3

Fig. 5

Structural divergence between the three - Triatoma 1:1 orthologs (BLAST best reciprocal hit to R. prolixus). a Distribution of GO (biological process) term frequency (y axis) according to Z value (the number of standard deviations from the median mean), as a measure of protein divergence (x axis). Negative Z values indicate higher conservation, whereas positive Z values indicate higher divergence. For instance, GO:0044710 (“single organism metabolic process”) has the same Z value in the three species, but is less divergent than GO:0044419 (“interspecies interaction between organisms”), with different although high Z values in the three species. b Biological process Z values for level 3 Gene ontology terms GO’s (rows). The Z value was numerically ordered according to T. pallidipennis and hierarchically clustered (Pearson’s correlation) according to species (columns). Blue tones reflect higher protein sequence % identity (i.e. more conserved), whereas yellow tones reflect lower % identity (more diverged). Significant higher protein sequence divergence were found in “single organism metabolic processes”, “interspecies interaction between organisms”, and oxidoreductase activity. c Molecular function Z values. Gene ontology terms level 3 GO’s. Kruskal-Wallis one Way ANOVA. Dunn’s correction for multiple testing. *** p < 0.001; ** p < 0.01. Non statistical significance in some conserved and divergent GO classes was probably because the lack of power due to low numbers of genes in those classes. Data corresponding to this figure, including % identity mean for each GO term and cellular compartment terms is in Additional file 5

Fig. 6

Schematic representation of metabolic enzymes belonging to the most divergent group (protein sequence identity) in the context of carbohydrate metabolism. Enzymatic mediators (italics) that were found in the divergent group in at least one Triatoma species (< 86% protein identity with putative R. prolixus orthologs, Rprc3.2_mapped dataset) are shown in green. Mediators that were not in the divergent group are in black, whereas those not found are labeled in grey. The glycolytic pathway is shown in orange. Gluconeogenesis is shown in purple, the oxidative phase of the pentose phosphate pathway is shown in red. The non-oxidative phase of PPP is shown in blue. Metabolites are shown in non italics gray. A full description of metabolic enzyme coding genes is described in Additional file 8

Fig. 7

Distribution of normalized dN-dS in reduviid oxidative phosphorylation complex subunits. Codon-by-codon Maximum Likelihood analysis was calculated for each ortholog using the HyPhy package [42]. The normalized dN-dS for each variant position was plotted according to increasing median value from left to right. Positive values indicate likelihood of selection. a Complex I. Core mitochondrial subunits (light blue), core nuclear-encoded subunits (dark blue), accessory A subunits (purple), accessory B subunits (pink); b Complex III. Respiratory subunits (light blue), core subunit (purple), low molecular weight subunits (pink); c Complex IV. Mitochondria-encoded subunits (purple); and d) ATPase complex. F1 ATPase subunits (yellow), Nuclear-encoded F0-ATPase subunits (orange) and mitochondria-encoded F0 subunits (red). Significant differences (one-way ANOVA test corrected for multiple comparison) in dN-dS distribution are shown under black bar. Red arrows indicate subunits participate in the oligomerization of respiratory supercomplexes in mammals [77]. A full description of coding genes involved in oxidative phosphorylation is described in Additional file 9