Hosts and vectors of Trypanosoma cruzi discrete typing units in the Chagas disease endemic region of the Paraguayan Chaco

SUMMARY Active Trypanosoma cruzi transmission persists in the Gran Chaco region, which is considered hyperendemic for Chagas disease. Understanding domestic and sylvatic transmission cycles and therefore the relationship between vectors and mammalian hosts is crucial to designing and implementing improved effective control strategies. Here we describe the species of triatomine vectors and the sylvatic mammal reservoirs of T. cruzi, in different localities of the Paraguayan and Bolivian Chaco. We identify the T. cruzi genotypes discrete typing units (DTUs) and provide a map of their geographical distribution. A total of 1044 triatomines and 138 sylvatic mammals were captured. Five per cent of the triatomines were microscopically positive for T. cruzi (55 Triatoma infestans from Paraguay and one sylvatic Triatoma guasayana from Bolivia) and 17 animals (12·3%) comprising eight of 28 (28·5%) Dasypus novemcinctus, four of 27 (14·8%) Euphractus sexcinctus, three of 64 (4·7%) Chaetophractus spp. and two of 14 (14·3%) Didelphis albiventris. The most common DTU infecting domestic triatomine bugs was TcV (64%), followed by TcVI (28%), TcII (6·5%) and TcIII (1·5%). TcIII was overwhelmingly associated with armadillo species. We confirm the primary role of T. infestans in domestic transmission, armadillo species as the principal sylvatic hosts of TcIII, and consider the potential risk of TcIII as an agent of Chagas disease in the Chaco.


I N T R O D U C T I O N
Trypanosoma cruzi is the causative agent of Chagas disease, a neglected human protozoan disease that is estimated to affect approximately six million people, spanning 21 endemic Latin American countries, with 60-80 million at risk of infection (WHO, 2015). Trypanosoma cruzi is genetically heterogenous, infecting a large number of mammal species and transmitted by haematophagous triatomine insect vectors. Nomenclature is historically complicated, but T. cruzi is currently subdivided into six subspecific groups, referred to as genetic lineages or discrete typing units (DTUs) and designated TcI to TcVI (Zingales et al. 2012). A cohort of geographically disparate bat trypanosomes, provisionally designated as TcBat, has been shown to share phylogenetically close affiliations with TcI (Marcili et al. 2009a), although more detailed sampling is required to confirm this as a formal taxonomic group. In a recent review, analysing more than 400 sequences with two mitochondrial (CytB and COII) and one nuclear gene (Gpi), authors propose three significant reliable mitochondrial clades, named mtTcI, mtTcII and mtTcIII, instead of seven . Phyloepidemiology and host vector associations of T. cruzi are complex, but have been partially resolved (Yeo et al. 2005;Miles et al. 2009;Messenger et al. 2015;Brenière et al. 2016). TcI is widespread through the Americas. This is the major DTU found infecting Didelphis opossums in nature, which is believed to be its most ancestral host. TcI was reported predominating in domestic transmission cycles in northern countries of South America (for example, Colombia and Venezuela) and in Central America ). TcIII and TcIV primarily circulate in sylvatic transmission cycles, the former especially associated with armadillos (Yeo et al. 2005) and the latter with a variety of sylvatic mammal species ). TcIV is a secondary cause of Chagas disease in Venezuela. TcIII is rarely reported from domestic transmission cycles. In contrast, TcII, TcV and TVI predominate in domestic transmission cycles in Southern Cone countries of South America Messenger et al. 2015;Brenière et al. 2016). Remarkably, TcV and TcVI are known to be natural hybrids derived from genetic exchange between TcII and TcIII in recent evolutionary history and are at present strongly associated with domestic transmission cycles (Zingales et al. 2012;Brenière et al. 2016).
Trypanosoma cruzi infection is considered primarily a zoonosis and as such eradication is not possible. Effective control of Chagas disease is achieved by interrupting vectorial transmission, primarily through residual insecticide-spraying to reduce domestic infestation and also by serological surveillance and interruption of transmission by blood transfusion, organ donation and congenitally (WHO, 2015).
There has been remarkable progress in controlling Chagas in some regions of the Americas. However, the Gran Chaco region, which includes territories of Argentina, Bolivia and Paraguay, is currently considered one of the most difficult regions for effective control and remains highly endemic (Hotez, 2014a). The land area is vast and arid, with a low population density consisting of small widely dispersed communities of low socioeconomic status (Gürtler, 2009). More than 20 ethnic groups live in marginalized conditions with minimal access to health care provisions (Gracey and King, 2009;Hotez, 2014b). Indigenous communities of the Gran Chaco show consistently high seroprevalence of human T. cruzi infection, ranging from 12 to 83% with local variation (Canese and Brice, 1977;Rojas de Arias et al. 1993;Moretti et al. 2010;Samuels et al. 2013).
Although the success of vector control interventions in some areas of the Chaco has substantially reduced disease incidence, the main challenge is the long-term sustainability, and in particular, entomological surveillance. Reinfestation of treated dwellings, when the residual effect of insecticides decreases, is a common feature especially in areas with peridomestic vectors and/or reinvasion by secondary vectors from the sylvatic environment (Provecho et al. 2014;Gaspe et al. 2015). Other obstacles adversely affecting control in the Chaco include the low efficacy of pyrethroid insecticide spraying on often poorly constructed peridomestic structures in this region (Gürtler et al. 2007;Cécere et al. 2013). Of further concern, is the appearance of Triatoma infestans populations resistant to pyrethroid insecticides in localities of northern Argentina and southern Bolivia (Lardeux et al. 2010;Gurevitz et al. 2012). Sylvatic populations of T. infestans have been identified in the Chaco and pose a potential risk of reinvasion (Noireau et al. 1997a;Ceballos et al. 2011;Quisberth et al. 2011;Rolón et al. 2011) as do secondary vectors, including Triatoma sordida (Almeida et al. 2000;Damborsky et al. 2001;Feliciangeli et al. 2003;Lauricella et al. 2005); additionally sylvatic mammals are potential reservoirs of infection, and all of these factors may confound effective control. This study ascertains the different triatomine species present in the region, the mammal species infected and the associated circulating T. cruzi DTUs. Through a better understanding of the T. cruzi transmission dynamics we aim to improve control strategies.

Fieldwork
Fieldwork collections were performed from 2002 to 2008, with the objective of obtaining and genotyping new isolates of T. cruzi from triatomine bugs and sylvatic mammals. Isolates obtained in previous surveys (Yeo et al. 2005;Llewellyn et al. 2009, Rojas de Arias et al., manuscript in preparation) were also included to generate a more detailed picture of the distribution of T. cruzi DTUs in the Chaco region.

Study area
The study area encompasses the Paraguayan Chaco (western region), Bolivian Chaco (southern region) and also three further Paraguayan localities 250 km northeast of Asunción (San Pedro, San Alfredo and Aguapey), within the Department of San Pedro. Study areas are shown in Fig. 1. In total, data were acquired from 28 localities, 24 Paraguayan and four Bolivian. Of the 24 Paraguayan localities, 21 were situated in the Paraguayan Chaco spanning three departments (Boquerón, Presidente Hayes and Alto Paraguay). Within the Department of Boquerón 12 localities were studied: Betania, Campo Loro, Campo Nuevo, Campo Salado, Cesarea, Galilea, Jerico, Campo Alegre, Casuarina, Jotoisha, Tiberia and Samaria. Within the Department of Presidente Hayes a further eight localities were included: Cerrito, Estancia Salazar, 12 de Junio, 20 de abril, Campo Largo, 10 Leguas, Fischat and Jope. Lastly, the one locality from the third department was Don Anibal ranch. These aforementioned localities have been under epidemiological surveillance by the National Program Control of Chagas since 2001. Further details regarding localization, ethnic group and estimated population size of these communities is shown in Table A1 (Appendix A). The remaining three localities lie within the Department of San Pedro (Fig. 1), in the Southeast Chaco: San Pedro, San Alfredo and Aguapey. Localities from the Bolivian Chaco region, San Antonio, Mora, Cuatro Cañadas and Gutierrez were all from Santa Cruz Department.
Indigenous communities consist of nomadic hunters, gatherers and fishermen with some groups of sedentary farmers from distinct linguistic groups (Rojas de Arias, 2003). Dwellings are typically of low quality, walls constructed of wattle and brick, adobe or palm trunk, soil floor with straw, and palm leaves or tin as roofing material. Villages are located in peripheral areas surrounding Mennonite towns. Domestic animals consist primarily of dogs and chickens. Further to the Northeast and in the Bolivian Chaco land use is predominantly for cattle farming and agriculture. To the east of the Paraguayan Chaco, the Department of San Pedro is characterized by more abundant vegetation and intensive agricultural land use.

Collection of triatomine bugs and sylvatic mammals
Each locality was visited only once. Most of the collection of domiciliary and peridomiciliary triatomines was made by manual searches (1·0 person hour per house) (Gürtler et al. 1998). Domestic areas inspected included walls, interior side of roofs, furniture and bedding. Peridomestic searches encompassed chicken coups, wood piles and other peridomestic structures (Gurevitz et al. 2011). In Campo Loro, Estancia Salazar and Fischat localities domestic bugs were obtained by the collection of local inhabitants. Sylvatic triatomines were collected by the use of live-bait Noireau traps (Noireau et al. 2000), manual dissection of natural ecotopes (bird nests, fallen trees and scrub) and light traps (Vazquez-Prokopec et al. 2006). Light traps were left overnight and checked in the morning. Captured bugs were placed in labelled containers, the developmental stage noted and identified to the level species by trained personel according to Lent and Wygodzinsky (1979). The captured nymphs were raised in laboratory conditions until they reached the adult stage to confirm the species. Sylvatic mammals were captured by the use of collapsible 'live-traps' including Sherman (H.B. Sherman Trap, Inc., Tallahassee, FL) and Tomahawk (Tomahawk Live Trap Co., WI), baited with a mixture of peanut butter, ripe banana and oats. Ten traps per night were set at approximately 10 m intervals on animal trails, near burrow entrances, in dense scrub or close to fallen trees. Traps were left in situ for 3 days, set at sunset and examined at dawn, where applicable local hunters were hired to collect live mammals. All specimens captured were sexed, identified to the species level (Neris et al. 2002), and released unharmed after processing (see below).

Isolation and characterization of trypanosomes
Trypanosomes were obtained from animals via xenodiagnosis and from triatomines by haemoculture and xenoculture as previously described (Miles, 1993). Mammals were first anaesthetized by intramuascular injection using ketamine (Holliday-Scott ® , 50-80 mg kg −1 body weight). Animalhandling procedures were in accordance with the American Society of Mammalogists (Sikes and Gannon, 2011). To excluded the presence of mixed infection biological clones of Trypanosoma cruzi were obtained by direct culture of infected triatomine feces, onto solid medium agar plates, as described previously by Yeo et al. (2007), five clones of each isolate were expanded in culture and DNA extracted [DNeasy kits (QIAGEN™)].
Genetic characterization of DTUs was undertaken using a combination of amplicon profiles from four different polymerase chain reactions (PCR) details of which are shown in Table 1. Genetic targets were the D7 divergent domain of the 24Sα rRNA (Souto et al. 1996), the size variable domain of 18S rRNA sequence (Brisse et al. 2000), the non-transcribed spacer of the mini-exon gene (Souto et al. 1996) and the PCR-restriction fragment length polymorphism (PCR-RFLP) of the intergenic region of the heatshock protein 60 (HSP60) gene ).

Triatomines
A total of 1044 triatomine bugs were included in the current study, 1037 from Paraguay and seven from the Bolivian Chaco. Triatoma infestans (n = 715) was found in both domestic and peridomestic environments (n = 245 in domestic; n = 470 in peridomestic) and T. sordida (n = 203) only in peridomestic environments (Table 2). Adults, fourth-and fifthinstar nymphs were collected from both species in both areas. In Betania, Campo Salado, Galilea both species shared the same niche in chicken coops. In the localities of Campo Nuevo, Cesarea and Samaria only T. sordida was present, while in Estancia Salazar only T. infestans was found. Although T. infestans was collected from both peridomestic and domestic areas, only domestic specimens were microscopically positive for T. cruzi. Thus, 55 (5·4%) of T. infestans were positive, including adults (n = 41), fourth-instar (n = 10) and fifthinstar (n = 4) nymphs. Positive triatomines were from the localities of Jerico (n = 31), Galilea (n = 2), Betania (n = 1), Campo Loro (n = 5), Estancia Salazar (n = 8), Jope (n = 5) and Fischat (n = 3). In the sylvatic area, 115 of adults (n = 107) and nymphs (n = 8) of Triatoma guasayana, three adults of T. sordida, and one female of Triatoma platensis were captured from the localities of Campo Loro, Don Anibal ranch and Betania respectively and they were microscopically negative. Triatomines obtained from Bolivian Chaco by the cooperation of local inhabitants included seven  Brisse et al. (2000) reported bands of 125 bp in one strain (Saimiri 3) and of 130 bp for three strains of North American origin. Band of 117 bp was reported by Kawashita et al. (2001). c According to Brisse et al. (2000). They reported product a low intensity band of 300 bp in two strains (M6241 cl6 and M5631 cl5). d According to Brisse et al. (2000). e According to Westenberger et al. (2005) and . f Differentiates between TcII and TcVI. Nidia Acosta and others sylvatic adult bugs, T. sordida (one specimen) and T. guasayana (six specimens). One T. guasayana specimen from the locality of Mora was positive.

Trypanosome isolates
A total of 166 T. cruzi isolates were included in this study, consisting of 63 new field isolates and a further 103 obtained from previous collections, as described below. Table 2 summarizes the origins of the isolates.
Sixty-three isolates were genotyped to DTU; ten from eight triatomines and two Didelphis spp., could not be maintained in culture and were excluded. The isolates from previous collections (Table 2) included 12 that had been genotyped, six from domestic T. infestans and six from sylvatic armadillos from the Chaco region of Paraguay (Yeo et al. 2005). Eightythree isolates originated from domestic T. infestans (Rojas de Arias et al., in preparation). Eight isolates were from sylvatic armadillos in Bolivia ).

Characterization of trypanosome isolates
Amplicon sizes obtained with new trypanosome isolates were as expected for the corresponding DTU, according to previous surveys (Yeo et al. 2005;). Examples for PCR-RLFP HSP60/EcoRV that differentiate between TcIII, TcII and TcVI are shown in Fig. 2. Domestic T. infestans (n = 136) were infected with TcII, TcIII, TcV and TcVI. The most common DTU was TcV (64%), followed by TcVI (28%), TcII (6·6%) and TcIII (1·5%). Twenty-nine sylvatic isolates examined from different armadillo species and one T. guasayana (Bolivia) showed an amplicon profile corresponding to DTU TcIII. Five biological clones from each one of the 63 new field isolates were genotyped to detected mixed infections, but none were found.
TcIII was predominant among sylvatic isolates and widespread in different armadillo species, including D. novemcinctus (n = 20), E. sexcinctus (n = 6) and Chaetophractus spp. (n = 3) from both Paraguay and Bolivia. TcIII was rarely observed in the domestic environment, being found in only two T. infestans specimens from the Jope locality of the Paraguayan Chaco. A single sylvatic specimen of T. guasayana from Mora locality (Bolivia) also harboured TcIII.
Four triatomine species were found in the current study: T. infestans (n = 715), T. sordida (n = 203), T. guasayana (n = 115) and T. platensis (n = 1). Triatoma infestans was the only species found in the domestic environment and there was no evidence of domiciliation by T. guasayana and T. sordida. The latter two species were found in peridomestic and/or sylvatic areas. A single specimen of sylvatic T. platensis was captured. This presence of T. infestans in human dwellings (245 specimens) and peridomestic habitats (470 specimens) confirms Samples from sylvatic and domestic hosts in Paraguay and Bolivia: A: products without digestion, B: products after of digestion by EcoRV. Lanes: 1 and 12 contain hyperladder 4; 2: negative control; 3: TcIII from sylvatic Triatoma guasayana in Bolivia; 4-5: TcII from domestic T. infestans in Paraguay; 6-7-8: TcIII, TcII, TcVI reference strains, respectively; 9-10: TcIII from domestic T. infestans in Paraguay; 11: TcVI from domestic T. infestans in Paraguay. that it is the primary vector of human T. cruzi infection in the Paraguayan Chaco (Rojas de Arias et al. 1990Arias et al. , 1993Zelada et al. 1998;Acevedo et al. 2002). The overall prevalence of T. cruzi infection in captured triatomines was 5·4%, and all those infected in Paraguay were domestic T. infestans. One explanation is peridomestic T. infestans feeding on avian blood in chicken coops, and birds cannot maintain T. cruzi infection (Miles et al. 2003). Considering only domestic T. infestans the infection rate is 22·4%, representing a substantial risk of transmission to both humans and domestic animals. These findings accord with previous studies in the Paraguayan Chaco, which report human seroprevalence ranging from 12 to 83%, and house infestation rates from 26 to 100% (Canese and Brice, 1977;Chapman et al. 1984;Rojas de Arias et al. 1993Rojas de Arias, 2003). Thus, the Chaco region is the most highly endemic for Chagas disease in Paraguay, primarily among native Amerindians of the low socioeconomic status.
In peridomestic environments, two triatomines species were found (T. infestans and T. sordida). Triatoma infestans in peridomestic areas poses a potential risk of re-infestation of buildings if surveillance measures are not continuous. Although peridomestic areas are also insecticide sprayed as part of the national control programme, the residual effect is lower than in the domicile because of exposure to climatic conditions. As a consequence this makes peridomestic bugs more difficult to control, requiring that spraying and surveillance to be more frequent. Triatoma sordida is widely distributed throughout Central Brazil, Eastern and Central Bolivia, the Chaco region of Paraguay and northwestern Argentina where it occurs primarily in the sylvatic environment (Lent and Wygodzinsky, 1979;Diotaiuti et al. 1995). In the Bolivian Chaco, two putative cryptic species belonging to T. sordida complex, named groups 1 and 2, respectively, were recognized circulating in sympatry, using multilocus enzyme electrophoresis ). This species shows great capacity of adaptation to peridomestic sites, especially in association with chickens (Macchiaverna et al. 2015). In the current study, all specimens (n = 203) were negative for flagellates probably because their principal food sources are avian. A primary domiciliation by T. sordida in the Chaco region was described in localities of Velasco Province, Department of Santa Cruz (Bolivia), where 16·2% of bugs were found infected by T. cruzi, although the probability of transmission to humans was considered low (Noireau et al. 1997b;Brenière et al. 1998). Three species of sylvatic triatomine were found: T. guasayana (n = 115), T. sordida (n = 3) and T. platensis (n = 1). Their typical ecotopes were fallen trees and dense shrubs, where they were captured using Noireau traps. Triatoma guasayana was an active flyer seeking out potential hosts, with most flight activity occurring just after sunset. Both T. guasayana and T. sordida have been implicated as sylvatic vectors of T. cruzi in parts of the dry Chaco region (Wisnivesky-Colli et al. 1997;Vezzani et al. 2001). In Paraguayan localities, both species were frequently observed near and around households, especially the adults, which have a great capacity for flight (Yeo et al. 2005;Rolón et al. 2011). Because of these characteristics and their ability to colonize man made structures, they are candidate as secondary vectors. All our sylvatic T. guasayana and T. sordida captured in Paraguay were not infected with T. cruzi, apart from a single sylvatic Bolivian T. guasayana. However, infected T. guasayana and T. sordida have previously been reported in the Argentinean and Bolivian Chaco (Noireau et al. 2000;Bar et al. 2002;Lauricella et al. 2005;Ceballos et al. 2009) with average infection rates of 13·3 and 9·1%, respectively. In the Argentinean Chaco, sylvatic T. sordida have been reported with high infection rates (38·5%; Bar and Wisnivesky-Colli, 2001;Bar et al. 2002).
Here we did not find T. infestans in the sylvatic ecotope. However, sylvatic 'dark morph' T. infestans have been reported in the Chaco region of Bolivia, in nests of Myiopsitta monachus (parrot), in bromeliads and hollows of live trees in several localities (Noireau et al. 1997a, b;Brenière et al. 2012;Waleckx et al. 2012), in the Argentinean Chaco (Ceballos et al. 2009 and Chile (Bacigalupo et al. 2006). Prevalence of T. cruzi infection is markedly lower in such 'dark morph' forms from the Chaco region with prevalence of between 2·5 and 12·5% (Noireau et al. 2000;Brenière et al. 2012;Waleckx et al. 2012) or zero (Ceballos et al. 2009 probably due that avian blood is the more often source of food. A few surveys have previously reported putative sylvatic populations of T. infestans in the Paraguayan Chaco, although they were also presumed attributable to dispersed peridomiciliary populations (Velázquez and González, 1959;Usinger et al. 1966;Yeo et al. 2005). More recently putative sylvatic colonies were discovered using a trained dog , and these bugs were found 3 km from infested houses. It is significant that this species is capable of surviving in sylvatic ecotopes in at least in some regions of the Paraguayan Chaco. Further research is needed to establish the risk of reinvasion from such sylvatic populations of T. infestans.

Mammals.
Eight species of mammals belonging to five different orders were captured in the study area. The overall prevalence of infection by T. cruzi was 12·3% (17/138), although this percentage varied according to the genus. Armadillos were the most common species captured in both regions of Paraguay (Chaco and San Pedro Departments) and from different localities in Bolivia. In two recent surveys performed in the humid Argentinean Chaco, marsupials and rodents together with armadillos were the most frequently captured species Orozco et al. 2013). The scarce number of marsupials captured and none for rodents in our study is probably related to the environment, since most of our successful collections were from the dry zone of the Chaco. The highest rate of infection was observed in D. novemcinctus (28·5%) followed by E. sexcinctus (14·8%) and Chaetophractus spp. (4·7%). In previous surveys in the same area, infection in armadillos ranged from 3 to 63%, with the highest prevalence in the Dasypus (Yeo et al. 2005;Llewellyn et al. 2009). Dasypus novemcinctus and the other armadillo especies are commonly hunted by the inhabitants of rural communities for food or for handicraft products, and they may be kept alive for several days before being used. Thus, infected armadillos pose a risk for bringing sylvatic T. cruzi into the domestic habitat. The triatomine vectors involved in sylvatic transmission cycles in the Chaco region remain uncertain. Members of the genus Panstrongylus were reported associated with armadillo burrows in Brazil (Grisard et al. 2000), Venezuela ) and Argentina . Our finding of one infected sylvatic T. guasayana in fallen trees in Bolivia could suggest some role in sylvatic transmission. The omnivorous behaviour of some mammal species also may contribute to their acquisition of infection (Rabinovich et al. 2001). The prevalence of T. cruzi infection in Chaetophractus spp. was lower than the other armadillos. Although this species construct their own burrows, they are nomadic and rarely use the same burrow twice, and thus unlikely to become infested with triatomines. Three other mammal species: T. tetradactyla (anteater), Dasyprocta spp. (agutí) and N. nasua (coati) from Bolivian localities were not infected with T. cruzi. Natural infection of T. tetradactyla by T. cruzi has been reported in Brazil (Miles et al. 1981;Bento et al. 1992;Fernandes et al. 1999) and Colombia (Ramírez et al. 2011). In addition, anteaters and coati are the known hosts of T. rangeli (Miles et al. 1983;Dereure et al. 2001).
Two D. albiventris of 14 captured (14·2%) from the Department of San Pedro were infected with T. cruzi, although isolates were not genotyped. These marsupials are usually found in humid areas, so the dry expanse of some Chaco zones may not be suitable for them. They are frequently observed in close proximity to human populations, and high T. cruzi infection rates have been found in Brazil (21·9 and 45·2% prevalence;Grisard et al. 2000) and in the humid Chaco of Argentina (36 and 38% prevalence;Alvarado-Otegui et al. 2012;Orozco et al. 2013). In San Pedro Department, T. cruzi has also been found in the terrestrial opossum Monodelphis domestica (Yeo et al. 2005). Further studies are needed to understand fully the role of marsupials in transmission of T. cruzi in Paraguay.

Host-vector of T. cruzi genotypes in the Paraguayan Chaco.
Ours is the most comprehensive survey of T. cruzi genotypes in the Paraguayan Chaco region, providing new insight into the transmission dynamics and dispersion among domestic and sylvatic cycles.
TcII, TcIII, TcV and TcVI were circulating in the region, with the hybrids TcV and TcVI being most frequently found, supporting earlier observations (Yeo et al. 2005;Lauthier et al. 2012;Maffey et al. 2012;Pérez et al. 2013). TcV and TcVI were predominant and the most dispersed, and found solely infecting T. infestans in the domestic cycle, also in agreement with previous surveys (Chapman et al. 1984;Acosta et al. 2001;Yeo et al. 2005). TcV and TcVI were reported in domestic T. infestans in the Bolivian Chaco (Pérez et al. 2013), in domestic and peridomestic triatomines (T. infestans and T. sordida) and domestic dogs and cats in the Argentinean Chaco (Maffey et al. 2012;Enriquez et al. 2013). Thus, TcV and TcVI constitute the largest current threat to human health, and have been associated with severe chronic manifestations of Chagas disease in the southern Cone countries (Corrales et al. 2009;Cura et al. 2012;Vicco et al. 2012;Lucero et al. 2016). TcV and TcVI are infrequently reported in sylvatic cycles: TcV has been observed in one sylvatic D. novemcinctus and one E. sexcinctus in Paraguay (Yeo et al. 2005), in a rodent (Octodontomys spp.), three opossums, two ferrets and one skunk in Argentina (de Luca D'oro et al. 1993;Montamat et al. 1992), and in two sylvatic triatomines (Triatoma spp.) from the Bolivian Chaco (M. Llewellyn, unpublished data). There is one record of TcVI in a D. marsupialis in the Northeast La Paz (the Jungas and Alto Beni regions) in Bolivia (Valette et al. 1988). It has been suggested that the domestic predominance of TcV and TcVI may be due to their recent anthropogenic origin and rapid clonal dissemination with T. infestans and human migration (Lewis et al. 2011). The occurrence of sylvatic TcV and TcVI in other regions, such as the Atlantic forest, remains to be fully explored.
TcII was found only in domestic T. infestans, in agreement with previous surveys in the Paraguayan Chaco, where it is also associated within human infections (Acosta et al. 2001;Yeo et al. 2005), although in lower frequency than the TcV and TcVI hybrids. TcII has been detected in single triatomines carrying mixed infection with TcVI (Yeo et al. 2007), and the presence of TcII may have been underestimated as discriminatory markers have not been applied. Like TcV and TcVI, TcII rarely been reported in sylvatic cycles, although this may reflect limited research. Recently, TcII was reported infecting one sylvatic T. infestans in the Bolivian Chaco . Likewise, this DTU has been reported from one monkey  and one E. sexcinctus, in Paraguay (Yeo et al. 2005) and from sylvatic mammals in Brazil (Fernandes et al. 1999;Bhattacharyya et al. 2015;Lisboa et al. 2015). Sylvatic TcII reservoirs are of particular interest as it is considered to be ancient (Westenberger et al. 2005;de Freitas et al. 2006).
A striking predominance of TcIII was apparent in sylvatic isolates. Twenty-nine sylvatic armadillos from Paraguay (both regions) and Bolivia, one sylvatic T. guasayana (from Bolivia) and two domestic T. infestans (from Paraguay) harboured TcIII. TcIII is frequently and widely found in sylvatic habitats with armadillos, particularly the genus Dasypus (Yeo et al. 2005;Llewellyn et al. 2009;Morocoima et al. 2012). Armadillos infected with TcIII were also reported in Colombia (Saravia et al. 1987), Venezuela Morocoima et al. 2012), Bolivia ), Brazil (Lisboa et al. 2009;Marcili et al. 2009b) and Argentina Orozco et al. 2013). In San Pedro (Paraguay), armadillos and one specimen of M. domestica (opossum) were infected previously with TcIII (Yeo et al. 2005). One sylvatic T. guasayana from the Bolivian Chaco carried TcIII, presumably acquired by feeding on an armadillo; this is the first report of TcIII in T. guasayana in Bolivia. Triatoma guasayana is frequently found near houses, attracted by light and CO 2 , may therefore introduce TcIII into the domestic cycle. This DTU has also been isolated from terrestrial sylvatic triatomines collected, such as P. geniculatus, T. rubrovaria, T. brasiliensis and T. vitticeps in Brazil (Póvoa et al. 1984;Martins et al. 2008;Santos-Mallet et al. 2008;Lisboa et al. 2009) and from Panstrongylus spp. associated with a burrow of D. novemcinctus in Venezuela ).
Two domestic T. infestans from the Chaco region of Paraguay harboured TcIII. Previously in the same region TcIII isolates were obtained from domestic dogs (Chapman et al. 1984) and from sylvatic armadillos (Yeo et al. 2005), suggesting leaky separation between domestic and sylvatic cycles. Dogs are commonly used for hunting of armadillos in the Chaco region, and dogs may thus introduce TcIII into the domestic transmission cycles, but TcIII has so far not been isolated from human cases of Chagas disease in the Chaco region. In contrast, in the eastern region of Paraguay (Cordillera and Paraguarí Departments), using amplification products of the 24Sα rRNA, mini-exon and hybridization, TcIII was reported in human cases and domestic T. infestans (del Puerto et al. 2010) as well as from domestic and peridomestic T. sordida from Concepción Department (Sánchez et al.

2012)
; this interesting and surprising finding merits follow-up analyses.
Notably, TcI was absent from this survey. TcI has predominantly been found associated with arboreal marsupials, especially the Didelphis opossum throughout the Americas but also with rodents and other sylvatic mammals (Yeo et al. 2005;Messenger et al. 2015). Records of the presence and distribution of TcI in Paraguay are scarce. It has been identified in the direct analysis of feces of domestic T. infestans from the Chaco and eastern region (Cura et al. 2010), in samples from domestic T. sordida in Concepción Department (Sánchez et al. 2012), and in one human case from the Chaco region in a mixed infection with TcII (Risso et al. 2011). Unfortunately, isolates were not obtained from the two infected opossums from San Pedro Department. TcIV was not found among our many Chaco region isolates, but it has been reported by direct analysis of the intestinal contents of domestic and peridomestic T. sordida captured in Concepción Department (eastern region) (Sánchez et al. 2012), although additional analyses are required to confirm this observation.
Biological clones analysed in this study produced similar profiles to the original isolates with the combination of PCR techniques used in this study. The cloning technique on solid media has proven to be useful for discriminating mixed infections in T. cruzi reservoirs (Yeo et al. 2007), especially when a variety of DTUs are circulating sympatrically in the same area.
In summary, the distribution and the high prevalence of TcII, TcV and TcVI in domestic transmission cycles shows the remarkable diversity of T. cruzi in the Chaco region of Paraguay. In eight localities more than one T. cruzi DTU was present in the domestic transmission cycle showing the great capacity of T. infestans in indigenous communities to harbour a variety of T. cruzi populations. Furthermore, there is increasing evidence of interaction between domestic and sylvatic transmission cycles. Especially, TcIII in the Jope locality was found in both transmission cycles, suggesting introduction of TcIII into the domestic cycle. TcIII is known to be highly virulent in mice (Morocoima et al. 2012) and may therefore prove to be an agent of severe human Chagas disease. The abundance and aggressive nature of T. guasayana also carries a risk, if it should adapt to colonization of human dwellings.
The data generated here provide a regional baseline for future research and an indication of potential risks for human health. High-resolution analyses, including comparative genomics, will give further insight into T. cruzi transmission dynamics, interactions between sylvatic and transmission and molecular genetics, to inform the much needed improved control of Chagas disease in the Gran Chaco region.