Produção Nacional
Revisado por pares
Phylogenetic and phenotypic relationships among Triatoma carcavalloi (Hemiptera: Reduviidae: Triatominae) and related species collected in domiciles in Rio Grande do Sul State, Brazil
2009; Wiley; Volume: 34; Issue: 2 Linguagem: Inglês
10.1111/j.1948-7134.2009.00023.x
ISSN1948-7134
AutoresCarlos Eduardo Almeida, Paula L. Marcet, Márcia Gumiel, Daniela M. Takiya, Margareth Cardozo-de-Almeida, Raquel S Pacheco, Catarina Macedo Lopes, Ellen M. Dotson, Jane Costa,
Tópico(s)Insect symbiosis and bacterial influences
ResumoTriatoma carcavalloi is considered a rare Chagas disease vector often collected inside domiciles in Rio Grande do Sul State. In this Brazilian state, T. carcavalloi has been collected in the same ecotope (rock piles) with two other species (T. rubrovaria and T. circummaculata), with which it also shares morphological characteristics. Previous morphological studies placed T. carcavalloi in the same species complex ("infestans complex") and subcomplex ("rubrovaria subcomplex") as T. rubrovaria, whereas T. circummaculata was placed in the "circummaculata complex." The phylogeny of a group composed of 16 species of triatomines was revaluated with the inclusion of T. carcavalloi by Bayesian analysis using mtDNA sequences of subunits 12S and 16S of the ribosomal RNA, and the cytochrome oxidase I (COI) genes. The phenotypic relationship among T. carcavalloi and related triatomines was also inferred from morphometrics. Phylogenetic results indicate that T. carcavalloi is a sister species of T. rubrovaria, and both were recovered as closely related to T. circummaculata. Morphometric studies confirmed the closeness among T. carcavalloi, T. rubrovaria, and T. circummaculata, prompting the placement of the latter species in the "infestans complex" and "rubrovaria subcomplex." Chagas disease is caused by a flagellated protozoan parasite, Trypanosoma cruzi, and is transmitted to humans mainly by blood-sucking triatomine true bugs. Approximately nine million Latin American people are at risk of contracting T. cruzi (Schofield et al. 2006). Vector transmission is still considered the main mode of infection in Brazil; therefore, most efforts against Chagas disease focus on the interruption of its natural vectorial transmission by controlling domiciliary vectors with pyrethroid insecticides. The National Health Foundation (Funasa) has employed intense insecticide spraying since the 1990s in Brazil directed at controlling Triatoma infestans, the triatomine responsible for most of Chagas disease cases in the past. The important accomplishment achieved by this program has been the interruption of T. cruzi transmission (Silveira and Vinhaes 1999, WHO 2002). After the successful vector control program for T. infestans, triatomines formerly considered sylvatic have begun threatening to transmit Chagas disease in several areas of Brazil. In the Rio Grande do Sul State (RGS), for example, data from the Chagas Disease Control Program indicated an increase in domiciliary and peridomiciliary invasion of T. rubrovaria. This species has become the most frequent triatomine species captured in this state (Almeida et al. 2000). Triatoma carcavalloi Jurberg, Rocha and Lent, 1998 is considered a rare species. Its description was based on six females captured by the technicians of the Brazilian Health Foundation in the municipalities of Santana do Livramento, Canguçu, Jaguarão, and Dom Feliciano in RGS. Since its description, little information has been provided regarding its phylogeny, ecology, vector competence, and bionomics. Almeida et al. (2002a) observed that three triatomine species (T. carcavalloi, T. rubrovaria, and T. circummaculata;Figure 1) occur in sympatry in the same ecotope (rock piles) in RGS. All of them have also been found inside domiciles in this state (Martins et al. 2003, Jurberg et al. 1998). Jurberg et al. (1998) mentioned that T. carcavalloi exhibits an intermediate size between T. rubrovaria and T. circummaculata and emphasized some particular characteristics used to distinguish T. carcavalloi from T. rubrovaria and T. circummaculata, such as the narrower anterior portion of pronotum and its smaller abdomen. Further diagnostic morphological characteristics of T. carcavalloi were provided by optical and scanning electron microscopy (Santos-Mallet et al. 2008), mainly regarding the number of trichobotria (more than 14) and the peculiar shape of the stridulatory sulcus and buccula. Triatoma carcavalloi (1), T. rubrovaria from Uruguay (2), and T. circummaculata (3). Morphological observations led to the placement of T. carcavalloi in the same species complex ("infestans complex") and subcomplex ("rubrovaria subcomplex") of T. rubrovaria, whereas T. circummaculata was allocated into the "circummaculata complex." This last species complex is composed of T. circummaculata and T. limai, and it does not present any subcomplex (Dujardin et al. 2002). Such classification of these species into these groups, however, did not agree with the phylogenetic analysis based on mitochondrial genes (12S and 16S ribosomal RNA, and the cytochrome oxidase I), where T. rubrovaria and T. circummaculata appear as closely-related species (García and Powell 1998, García et al. 2001, Sainz et al. 2004). Additionally, these molecular analyses did not include T. carcavalloi as a result of its recent description (1998). The phylogenetic placement of T. carcavalloi in relation to related species (T. rubrovaria and T. circummaculata) was herein studied based on DNA sequences and morphometry. Therefore, these two approaches were combined to address questions regarding the taxonomic and phylogenic status of T. carcavalloi and its allies. Field specimens were used for all analyses. They were collected by manual, diurnal captures in the municipality of Encruzilhada do Sul, RGS, Brazil (30°32′38″S; 52°31′19″W; Figure 2) where savanna-like or steppe-like subtropical mixed prairies are the predominant vegetation type (IBAMA 2008). Hence, T. carcavalloi (n=1), T rubrovaria (n= 17), and T. circummaculata (n=44) were collected in 1998 and 1999. Fifteen more specimens of T. carcavalloi were collected in 2002 in Encruzilhada do Sul, together with T. rubrovaria and T. circummaculata specimens. During this fieldwork, Triatoma circummaculata was the most frequently collected (57%), followed by T. rubrovaria (29%), and finally T. carcavalloi (14%). These three species were collected in sympatry, moreover in the same ecotope: a sylvatic-like environment among rock piles where some of them were collected sharing the same shelter. These specimens of T. carcavalloi were collected in the same ecotope but in different collecting spots, distributed along a range of ≈3km apart. Species were identified according to Lent and Wygodzinsky (1979) and Jurberg et al. (1998). Grey squares mark the location where sympatric specimens of T. carcavalloi, T. rubrovaria, and T. circummaculata were captured in Rio Grande do Sul, Brazil, and the origin of specimens of T. rubrovaria from Uruguay of the Entomological Collection of Oswaldo Cruz Institute (Laranja's Collection). Six samples of T. carcavalloi were randomly taken from the pool of field specimens and stored at -20° C until processed. DNA was purified with the Wizard Genomic Purification Kit (Promega, Madison, WI) following the manufacturer's recommendations. Polymerase chain reaction (PCR) was performed using the primers defined in Table 1 for each gene, and PCR conditions were optimized for this species by a touchdown approach. For 12S and 16S, the same amplification conditions were used: denaturation at 94° C for 3′; 5 × (93° C for 1′; 54° C for 1′; 72° C for 2'); 5 × (93° C for 1′; 53° C for 1′; 72° C for 2′); 25 × (93° C for 1′; 52° C for 1′; 72° C for 2′) and final elongation at 72° C for 15′. For COI, the PCR conditions described above were used, but the annealing temperatures were 46°, 45° and 44° C for each cycle, respectively. Amplifications by PCR were carried out in an Eppendorf Master Gradient Thermocycler (Brinkman Instruments, Westbury, NY). Amplified products were visualized on a 1.5% agarose gel stained with ethidium bromide. Purified PCR products were sequenced using a BigDye Terminator Cycle Sequencing Ready Reaction Kit (Applied Biosystems, Foster City, CA), according to the manufacturer's recommendations. Both strands were sequenced in an automated DNA sequencer (Applied Biosystems model 3100). Consensus sequences were edited and assembled using the SEQMAN program (Applied Biosystems), aligned using ClustalW (Larkin et al. 2007) available through Mega3 (Kumar et al. 2004), and manually checked thereafter. Pairwise and multiple alignments were run based on the IUB DNA weight matrix and parameter costs were set to 15.0 for gap opening and 6.66 for gap extension. Fragments of mtDNA of the 12S and 16S ribosomal RNA, and the cytochrome oxidase I (COI) genes sequenced for T. carcavalloi were aligned with sequences of other species used in previous studies and available on GenBank (Table 2). Other species included representatives of "infestans complex," putatively related, T. circummaculata, and representatives of other Triatoma, Panstrongylus, and Rhodnius. No major incongruence was found among gene datasets based on the incongruence length difference (ILD, Farris et al. 1994) test conducted using PAUP* 4.0b10 (Swofford, 1998; COI vs 12S: p=0.21; COI vs 16S: p=0.23; 12S vs 16S: p=0.43), showing that the three mtDNA fragments studied here could be combined for maximum parsimony (MP) analyses. An hypothesis of phylogenetic relationships among Triatoma species was proposed based on the MP algorithm by heuristically searching for the best topology using 1,000 replicates of random addition sequence and tree-bisection-and-reconnection (TBR) branch swap. Bootstrap values were estimated with 1,000 pseudoreplicates. Additionally, another phylogenetic hypothesis was proposed based on Bayesian inference with the software MrBayes 3.1 (Huelsenbeck and Ronquist 2001) using a mixed-model approach. The hierarchical likelihood ratio test implemented in MrModeltest (Nylander 2004) was used to select the GTR+G model as the appropriate evolutionary model for each partition. Even though the same general model was selected for each partition, parameters were still left to vary independently for each one. The Bayesian analysis was conducted with two independent runs of four Markov chains for 1,000,000 generations from where 2,000 trees were sampled (sampled every 500th generation) and the average standard of split frequencies exhibited 0.005 at the end of the run. To summarize the information we discarded ≈25% of the samples obtained during the first generations. Rhodnius prolixus was used as the outgroup for all analyses. Bayesian posterior clade probabilities were calculated as the majority-rule consensus. The best evolutionary model indicated by MrModeltest was applied to obtain the corrected Maximum Likehood p-distances in order to make inferences on the genetic distances among taxa focused in this study. Because head shape is an important character to define interspecific relationships (Lent and Wygodzinsky 1979, Patterson et al. 2001), we applied a traditional morphometric technique based on measurements between landmarks. The same investigator took all measurements using a stereoscope microscope (Zeiss) with a 2.5 × objective and a 10 × ocular. Besides T. carcavalloi, T. circummaculata, and T. rubrovaria from the municipality of Encruzilhada do Sul, RGS, a second population of T. rubrovaria from Uruguay was added to the analysis. This T. rubrovaria population was taken from the "Laranja's Collection," deposited in the Entomological Collection of the Oswaldo Cruz Institute. Species used as outgroups (T. sordida [n=15] and T. guasayana [n=12]) were collected in the field in the states of Bahia, Brazil, and Santa Cruz, Bolivia (Table 3). Males and females of all species were included. For each specimen, ten measurements derived from the head were recorded: outer distance between eyes (OE), inner distance between eyes (IO), external distance between ocelli (EO), anteocular distance (AO), post-ocular distance excluding neck (PO), diameter of the eye (DE), antenniferous distance (AD), and first (R1), second (R2) and third (R3) segmental distances of the rostrum (Figure 3). Analysis of variation in log-transformed head values was performed by breaking down the total variability in size and shape based on the Mosimann proposal. An interspecific comparison approach was used to remove the isometric components of changes (Darroch and Mosimann 1985), where size is described as an attribute of a specimen and shape involves preservation among specimens of different sizes. Consequently, these variables were scaled for size and the principal components computed (PCA). The PCA was done on the covariance matrix derived from the shape variables. Data were analyzed using the computer program BAC, JMP (SAS Institute 1999). Scheme of morphometric measures taken on the head in a lateral (A) and dorsal view (B) of a Triatoma rubrovaria specimen. The arthropod fauna observed where T. carcavalloi was found was generally represented by T. rubrovaria, T. circummaculata, and the cockroach, Blaptica dubia. Field observations have not detected any ecological differentiation among the three triatomine species collected in the same ecotope. Those rock piles were characterized by loose and overlapping rocks in dry environments in the prairie. This sylvatic-like environment is herein called "ruderal" because it fits the description mentioned by Almeida et al. (2002a, 2008): an environment where the native vegetation was substituted by new pastures or short-lived cropland. Therefore, domesticated animals (e.g., caprines, equines, and bovines) were frequently observed in this environment, but wild homeothermic animals were seldom seen (including their feces) among rock piles where triatomines were collected. Fragments of 310, 450, and 468 base pairs for the 12S, 16S, and COI genes of T. carcavalloi were obtained, respectively. After alignment, the ends of sequences were cropped according to the shortest one, and analyses were conducted with 314, 410, and 401 base pairs for these genes (including gaps), following the order above. The six specimens of T. carcavalloi analyzed here exhibited only one haplotype for all three genes studied. The proportions of base frequencies for the three genes sequenced for T. carcavalloi were for: (1) COI, A=0.29, C=0.26, G=0.17, T=0.28; (2) 12S, A=0.30, C=0.10, G=0.20, T=0.40; and (3) 16S, A=0.29, C=0.08, G=0.22, T=0.41. These proportions of each gene fragment of T. carcavalloi did not significantly differ from those of other triatomine species included in this study (Chi-square test of homogeneity of base frequencies across taxa: p=1.00). No regions of ambiguous alignment were detected. Considering the three sympatric species, similar patterns of p-distances were exhibited for the 16S and COI. For these genes, T. rubrovaria vs T. carcavalloi exhibited the lowest p-distances (0.012 for the 16S and 0.025 for the COI), whereas p-distances for T. rubrovaria vs T. circummaculata were higher (0.017 for 16S and 0.035 for COI). Only the 12S gene showed the lowest p-distance for T. rubrovaria vs T. circummaculata (0.007). For this last gene, T. rubrovaria vs T. carcavalloi showed a p-distance at 0.010, while T. circummaculata vs T. carcavalloi showed the highest p-distance (0.016). Considering the outgroups (T. guasayana and T. sordida) vs all sympatric species herein focused, T. guasayana exhibited lower values of p-distances (<0.0.030 for 12S, <0.025 for 16S, and <0.090 for COI; for all) than the ones exhibited by T. sordida (<0.0.052 for 12S, <0.072 for 16S, and <0.115 for COI for all). The latter species exhibited lower values of p-distances with T. rubrovaria (0.0.042 for 12S, 0.060 for 16S, and <0.112 for COI) than with T. guasayana (0.045 for 12S, 0.072 for 16S, and <0.120 for COI). The resulting Bayesian consensus phylogram of the combined (12S+16S+COI) mixed-model data is shown in Figure 4. T. carcavalloi was recovered as sister-species to T. rubrovaria, and this clade a sister to T. circummaculata, forming the "rubrovaria subcomplex," supported by 99/68 of Bayesian posterior probability/ parsimony boostrap. The maximum parsimony analysis of the combined genes recovered a single most parsimonious tree (L=986, CI=0.53, RI=0.50) with the same topology as the Bayesian analysis for the sympatric species. The range of changes per site supported the taxonomic status for T. carcavalloi as a putative species. Triatoma guasyana was consistently placed as a sister species to members of "rubrovaria subcomplex" supported by 99/72 of Bayesian posterior probability/ parsimony boostrap, respectively. Bayesian Inference consensus of mixed-model analysis of five partitions of mitochondrial sequences of 12S rDNA (314bp), 16S rDNA (410bp), and cytochrome oxidase I (401bp, 3-codon partition). The values above the nodes indicate the Bayesian posterior credibility of clades whereas below are the parsimony bootstrap percentages. Accession code and references with details about the species are shown in the Table 1. Classification follows Dujardin et al. (2002), with the exception of T. circummaculata (marked with asterisk), which is herein placed in the "rubrovaria subcomplex." The isometric size variable revealed that females were larger than males for all characters in species groups, as shown by Dujardin et al. (2002) for other triatomines. Triatoma circummaculata had the smallest values for overall variants, whereas T. rubrovaria specimens had the largest (Figure 5). However, intraspecific variation (even between sexes) was diluted in a way that specific groupings were all well defined when both principal components are considered. Boxplot of the centroid size based on ten variables of the head of Triatoma carcavalloi (Tcar), T. circummaculata (Tcir), T. guasayana (Tgua), T. rubrovaria from Brazil (Trubr) and Uruguay (Truy); and T. sordida (Tsor). A principal component analysis was performed including the three sympatric species, T. circummaculata, T. carcavalloi, and T. rubrovaria, and two outgrous (T. guasayana and T. sordida) to find the linear combinations of the shape variables that explain the greatest variance in the data (Figure 6a). Significant differences between groups were shown with PCA by plotting the first and second principal component factors on a factorial map. Figure 6a shows that the first two derived principal components exhibit 93% of the total shape variability in the groups. The first principal component clearly separates T. guasayana from the remaining species, while the second principal component separates T. sordida from the remaining sympatric taxa, T. circummaculata and the very closely associated group composed of T. rubrovaria (Uruguay and Brazil), and T. carcavalloi. (a) Factorial map for the species studied: Triatoma carcavalloi, T. rubrovaria from Uruguay and Brazil, and T. circummaculata; including outgroups: T. guasayana and T. sordida, representing 93% of the total variation. This analysis derived from the first (CP1 = 84%) and the second Principal Component (CP2 = 9%). (b) Factorial map derived from CP1 (78%) and CP2 (9%), using only the three species herein. See Tables 3 and 4 for details about the sampling used. To get a more detailed separation among the three sympatric species, a second analysis was conducted using only T. circummaculata, T. carcavalloi, and T. rubrovaria species (Figure 6b). The first two derived principal components accounted for 87% of the total shape variability in the groups. The first component was able to significantly discriminate the three species. Triatoma. carcavalloi was closer to both populations of T. rubrovaria (from Brazil and Uruguay) than to T. circummaculata (Figure 5b). This result is a reflection of overall morphological differences from the head variables, e.g., the anteocular distance and width of the anteclypeus, being distinctly shorter for T. circummaculata if compared to T. rubrovaria and T. carcavalloi. Furthermore, the third rostral segment was the most discriminant variable separating T. circummaculata by being longer than those of its allies; while the outer distance between the eyes, the external distance between ocelli, and the diameter of the eye were longer for T. rubrovaria than for T. carcavalloi (Table 4). Since T. carcavalloi's description, no study has been performed to elucidate the phylogenic, ecological, bionomic, or epidemiological aspects of this species. We described the ecotope where these species were collected and performed an analysis using molecular (mitochondrial DNA) and morphometric markers. The phylogenetic position of T. carcavalloi suggests that this species is unambiguously related to the other two sympatric species (T. rubrovaria and T. circummaculata) collected in the same location in the state of RGS and ecotope (rock piles) in prairies from Southern Brazil. All analyses recognized T. carcavalloi as a consistent species, and the relationships among its cohabitant allies (T. rubrovaria and T. circummaculata) were congruent among the different approaches used. Morphological characteristics led Jurberg et al. (1998) to consider T. carcavalloi to be a closely related species to T. rubrovaria. Dujardin et al. (2002) presented a review of the triatomine complexes and subcomplexes based on morphological characters, where T. circummaculata was placed out of the "infestans complex."Dujardin et al. (2002) had already stated that this grouping has not been corroborated by previous molecular studies where T. circummaculata was found more closely related to members of the "infestans complex" (García and Powell 1998, García et al. 2001). Our molecular data set reinforces the previous findings based on DNA sequences, with the added evidence from our morphometrical analysis suggesting that T. circummaculata should be included in the "infestans complex" and not placed in a separate complex. Both approaches used here further suggest that T. circummaculata should be placed more specifically in the "rubrovaria subcomplex." Usinger et al. (1966) and Schofield (1994) mentioned that T. guasayana and T. sordida were closely related and morphologically very similar. Similarly, Dujardin et al. (2002) placed these species, as well as T. garciabesi and T. patagonica, in the "sordida subcomplex" of the "infestans complex." Despite the fact that T. guasayana and T. sordida are very morphologically similar and overlap in geographic distribution and ecotopes (mainly bromeliads and hollow trees) as observed by Noireau and Dujardin (2001) in the Bolivian Chaco, they can be readily separated by isoenzymatic analysis (Noireau et al. 1998). On the other hand, our resulting morphometric analysis did not identify these two species as being highly similar (Figure 6a). Furthermore, T. guasayana exhibited p-distance values that indicated that it was more closely related to members of "rubrovaria subcomplex" than to T. sordida. In the present phylogenetic analysis, the monophyly of the "sordida subcomplex" was not corroborated, because T. guasayana was consistently identified as a sister to members of "rubrovaria subcomplex" with high nodal support by both Bayesian and MP analyses (Figure 4). Whether morphological similarities between T. sordida and T. guasayna are a result of ecological pressure remains a question. Morphometric analysis also demonstrated that T. rubrovaria from Brazil and Uruguay exhibited a degree of population differentiation that agrees with the intraspecific variation shown by Pacheco et al. (2003, 2007) for populations from these countries using molecular markers (RAPD analyses and rDNA intergenic-spacer sequences). We understand that phenotypic variability might be disturbed by various factors, such as ecotope, host, temperature, and others. However, the morphometric results for intraspecific and interspecific levels exhibited considerable agreement with genetic data in strongly supporting members of the "rubrovaria subcomplex" in relation to the outgroups chosen. We believe that morphometrics could better reflect the genetic relationships if we had included more specimens, improved the robustness of the results by using the geometric morphometrics approach, and increased the number of variables for the analysis. Crossing experiments have been useful to enable inferences on the genetic relationships and reproductive compatibilities among taxa (Costa et al. 2003), thus we suggest reciprocal laboratorial crossing experiments among T. rubrovaria, T. carcavalloi, and T. circummaculata in order to better understand the relationships among these three cohabitant triatomines. All three triatomine species, T. carcavalloi, T. rubrovaria, and T. circummaculata are restricted to the savanna-like or steppe-like subtropical mixed prairies in southern Brazil, northeastern Argentina, and Uruguay, with its closest species, T. guasayana, restricted to the Bolivian and Argentinian Chaco. During this study, all three focal species were found sharing a shelter in a specific locality, therefore apparently occupying the same ecological niche in a disturbed environment. No hybrids were detected through morphological analysis; hence the event that drove the genetic differentiation remains a question. However we believe that they have recently converged to the habitat after speciation due to environmental changes caused by human activities. According to Almeida et al. (2002a, 2008), human activities can create the ruderal environment, which can produce propitious situations for the installation of some triatomines, since the predator fauna might be disturbed, and the blood food resource for triatomines can be improved by increasing domesticated animals, such as sheep chicken, goats, pigs, and cows. Four other triatomines (T. brasiliensis, T. sordida, T. pseudomaculata, and Rhodnius neglectus) have already been found sharing the ecotope (peridomicile) as a result of human activities (Marchon-Silva et al. 1998). Therefore, we suggest looking for these three species where they occur in undisturbed environments. Phylogenetic analysis suggested that T. carcavalloi is the sister species of T. rubrovaria, which was corroborated by a morphometric approach. We emphasize that T. rubrovaria presents bionomic characteristics of a good vector of T. cruzi (Almeida et al. 2000, 2002a,b, 2003, 2005) and can maintain high rates (25%) of natural infection by that parasite (Martins et al. 2003, 2008). According to Jurberg et al. (1998), T. carcavalloi can also invade human domiciles. Therefore, we recommend monitoring the domiciliary invasion by T. carcavalloi, and increasing our understanding of this species through biological, ecological, and genetic studies, as well as its interaction with T. cruzi. This research was supported by the National Council for Scientific and Technological Development (CNPq), the Strategic Program of Research in Health (PAPES 3/Oswaldo Cruz Foundation), and the Brazilian Coordination for the Improvement of Higher Education Personnel (CAPES). We acknowledge the technicians of Funasa for their essential help in the field; anonymous referees, Karen Haag, Marli M. Lima, and L. Lynnette Dornak for kindly and carefully reviewing the manuscript, and Paula Constancia Pinto Aderne Gomes for improving the pictures.
Referência(s)