INTRODUCTION
Trypanosomes are ubiquitous parasites of a large number of vertebrates and are usually transmitted by bloodsucking invertebrates. Arthropods (insects and ticks) are vectors of trypanosomes in mammals and birds, whereas leeches are vectors of trypanosomes in aquatic vertebrates. Leeches, flies and mosquitoes are implicated in the transmission of trypanosomes among amphibians and reptiles (Telford, Reference Thompson, Gibson, Plewniak, Jeanmougin and Higgins1995; Stevens et al. Reference Telford and Kreier2001; Hamilton et al. Reference Hamilton, Stevens, Gaunt, Gidley and Gibson2004, Reference Hamilton, Gibson and Stevens2007). Among the reptilian trypanosomes, more species have been reported in lizards than in snakes, chelonians or crocodilians. In general, trypanosomes of snakes appear to be rare, despite reports of infection in snakes from distinct families inhabiting terrestrial and aquatic environments around the world. Pathogenicity of snake trypanosomes is virtually unknown (Telford, Reference Thompson, Gibson, Plewniak, Jeanmougin and Higgins1995).
The first references to the occurrence of trypanosomes in snakes appeared at the beginning of the 20th century in various continents (Wenyon, Reference Jakes, O'donoghue and Adlard1909; Brumpt, Reference Brumpt1914), later followed by a few papers about snake trypanosomes in Brazil (Pessôa, Reference Pessôa1928; Arantes and Fonseca, Reference Arantes and Fonseca1931a; Fonseca, Reference Fonseca1935), Africa (Fantham and Porter, Reference Fantham and Porter1950, Reference Fantham and Porter1953) and North America (Ayala et al. Reference Ayala, Atkinson and Vakalis1983; Chia and Miller, Reference Chia and Miller1984). Most papers on snake trypanosomes are only occurrence reports or morphological descriptions. Growth of these trypanosomes could not be sustained for more than a few passages in culture. Culturing was only successful with the following trypanosomes: T. hydrae and T. yaegeri from the aquatic snakes Nerodia fasciata confluens and Agkistrodon piscivorus leucostoma, respectively (Ayala et al. Reference Ayala, Atkinson and Vakalis1983; Chia and Miller, Reference Chia and Miller1984) and T. butantanense from Ophis merremii (Arantes and Fonseca, Reference Arantes and Fonseca1931a) but no further studies of these trypanosomes were carried out.
Telford (Reference Thompson, Gibson, Plewniak, Jeanmougin and Higgins1995) registered 21 species of snake trypanosomes. Several species should be synonymies and at least 2 species of trypanosomes from Brazilian snakes (T. merremii and T. manguinhensis) (Arantes and Fonseca, Reference Arantes and Fonseca1931b) were left out of his revision list. Snake trypanosomes have only been found in the blood of their hosts, and tissue forms have never been described. Morphology of the blood forms allows distinction of 2 major groups. One is represented by large and wide trypomastigotes with few dividing forms in peripheral blood (50–100 μm long and up to 3–5 μm wide) and includes T. najae; T. clozeli and T. primeti (cited by Telford, Reference Thompson, Gibson, Plewniak, Jeanmougin and Higgins1995); T. brazili (Brumpt, Reference Brumpt1914; Brygoo, Reference Brygoo1965a); T. phylodriasi (Pessôa, Reference Pessôa1928); T. constrictor and T. salamantae (Pessôa and Fleury, Reference Rodrigues, Paiva, Campaner, Stevens, Noyes and Teixeira1969); T. merremii (Arantes and Fonseca, Reference Arantes and Fonseca1931b); T. cascavelli (Pessôa and De Biasi, Reference Pessôa and Fleury1972) and T. yaegeri (Ayala et al. Reference Ayala, Atkinson and Vakalis1983). The other group includes slender trypanosomes, shorter than 50 μm, with frequent dividing blood forms: T. erythrolampi (Wenyon, Reference Jakes, O'donoghue and Adlard1909), T. butantanense (Arantes and Fonseca, Reference Arantes and Fonseca1931a), T. mattogrossense (Fonseca, Reference Fonseca1935) and T. hydrae (Ayala et al. Reference Ayala, Atkinson and Vakalis1983; Chia and Miller, Reference Chia and Miller1984). Between the two groups lies T. haranti, which is short (31 μm) and wide (8–10 μm) (Brygoo, Reference Brygoo1965b).
Aquatic leeches have been implicated as trypanosome vectors for hosts that live all or part of their lives in an aquatic environment, such as fishes and species of aquatic snakes, chelonians and anurans (Telford, Reference Thompson, Gibson, Plewniak, Jeanmougin and Higgins1995; Stevens et al. Reference Telford and Kreier2001; Hamilton et al. Reference Hamilton, Stevens, Gaunt, Gidley and Gibson2004; Ferreira et al. Reference Ferreira, Campaner, Viola, Takata, Takeda and Teixeira2007). While the role of leeches as vectors of species that parasitize aquatic snakes is well documented, the transmission of trypanosomes to terrestrial snakes is poorly understood. Brumpt (Reference Brumpt1914) and Brygoo (Reference Brygoo1965a) reported on the development of T. brazili in the leech Placobdella brasiliensis fed on the aquatic snake Helicops modestus. Chia and Miller (Reference Chia and Miller1984) reported the development of T. hydrae in the leech Placobdella parasitica. Pessôa (Reference Pessôa and De Biasi1968) described the development of T. hogei in the leech Haementeria lutzi when it fed upon an infected snake Rachidelus brazili. These infected leeches failed to transmit their trypanosomes to Helicops modestus, suggesting host-restriction (Pessôa, Reference Pessôa and De Biasi1968). Attempts to transmit snake trypanosomes through mosquitoes failed in all cases (Pessôa and De Biasi, Reference Pessôa and Fleury1972; De Biasi et al. Reference De Biasi, Pessôa, Puorto and Fernandes1975; Ayala et al. Reference Ayala, Atkinson and Vakalis1983). Trypanosomes of anurans and lizards can be transmitted by sand flies (Anderson and Ayala, Reference Anderson and Ayala1968; Ayala and McKay, Reference Ayala and McKay1971; Ayala, Reference Ayala1971; Telford, Reference Thompson, Gibson, Plewniak, Jeanmougin and Higgins1995). However, there are no data about the role of sand flies as vectors of snake trypanosomes.
The evolutionary history of reptilian trypanosomes is far from understood. Although only 6 species of reptilian trypanosomes have been included in phylogenetic analyses to date, the results revealed that these trypanosomes are highly divergent and polyphyletic. Trypanosomes from major reptilian groups (lizards, chelonians and crocodilians) were separated in distant clades. T. chelodina from tortoise nested in the fish–trypanosome clade (species transmitted by aquatic leeches), whereas T. grayi that infects crocodiles and is transmitted by the tsetse fly, was positioned closest to avian trypanosomes. Lizard (Order Squamata) trypanosomes of Varanidae (T. varani) and Phrynosomatidae (T. scelopori) clustered with a trypanosome from Gekkonidae, forming the lizard clade. T. therezieni from Chamaeleonidae clustered with anuran trypanosomes (Haag et al. Reference Haag, O'h Uigin and Overath1998; Stevens et al. Reference Telford and Kreier2001; Jakes et al. Reference Jakes, O'donoghue and Adlard2001; Hamilton et al. Reference Hamilton, Stevens, Gaunt, Gidley and Gibson2004, Reference Hamilton, Gibson and Stevens2007). According to data from these studies, positioning of lizard trypanosomes within Trypanosoma is still unresolved, varying according to genes, taxa and methods used for phylogenetic inferences.
Altogether, biological and molecular data on reptilian trypanosomes suggest complex, distinct relationships between these parasites and their vertebrate and invertebrate hosts. Moreover, the data indicate a complex evolutionary history of trypanosomes even when isolated from the same order (Squamata). Analysis of a large number of trypanosomes and their putative vectors is crucial to an understanding of the evolutionary history of reptilian trypanosomes. The taxonomy of snake trypanosomes also lacks consistency, and the validity of species assigned to these trypanosomes has long been questioned (Brygoo, Reference Brygoo1965b; Telford, Reference Thompson, Gibson, Plewniak, Jeanmougin and Higgins1995). In fact, all the existing species have been created according to morphology and host origin, which are insufficient taxonomic criteria for trypanosomatids in general (Wallace et al. Reference Wenyon1983).
In this study, we isolated and characterized snake trypanosomes with the aim of establishing a molecular basis for the identification of species using methods other than morphological and host restriction analysis. We also characterized a trypanosome from sand flies that is closely related to snake trypanosomes. SSU rDNA sequences were used to infer phylogenetic relationships among snake, lizard and sand fly trypanosomes and between these trypanosomes and species from other reptiles and vertebrates.
MATERIALS AND METHODS
Handling of snakes
In this paper we report on the presence of trypanosomes in snakes sent to the Butantan Institute, São Paulo, Brazil, the major centre for the production of snake anti-venom sera. Snakes were captured in various regions of Brazil between 2000 and 2005. After being anaesthetized with CO2, they were bled by tail or heart puncture using sodium citrate as anticoagulant. Infection of snakes was carried out by intravenous inoculation through the caudal vein. Prior to use, snakes for experimental infections were examined for trypanosome infection by repeated examination of blood samples by haemoculture and kept in separate cages. All procedures were performed according to the recommendations of Butantan Institute Ethics Committee.
Isolation and culture of snake trypanosomes
Blood samples from snakes were examined for the presence of trypanosomes by microhaematocrit centrifugation (MH). Regardless of the MH results, the blood samples were inoculated in tubes containing biphasic medium (BAB-LIT) consisting of 15% rabbit red blood cells mixed with 4% Blood Agar Base (BAB, DIFCO) overlaid with LIT (Liver Infusion Tryptose) medium supplemented with 10% FBS (Ferreira et al. Reference Ferreira, Campaner, Viola, Takata, Takeda and Teixeira2007). Because this medium does not support continuous trypanosome growth, primary cultures were transferred to a monolayer of Hi-5 insect cells (Trichoplusia ni) overlaid with Grace's medium supplemented with 10% FBS.
For morphological analysis, glass-slide smears were prepared from the snakes' blood using either whole blood samples, buffy coats from MH capillary tubes or culture supernatants. All smears were fixed with methanol, stained with Giemsa and photographed with a digital camera.
Isolation and culture of trypanosomes from sand flies
The trypanosomes from sand flies used in the present study were isolated as previously described (Freitas et al. Reference Freitas, Naiff and Barrett2002). The sand flies were captured using CDC light traps suspended at ~1 m from the ground and from the bases of trees by aspiration of the trunk in the northern region of Rondonia State, Brazilian Amazonia. The isolate from the sand fly Viannamyia tuberculata was cultured as described above for snake isolates.
Experimental infection of snakes with trypanosomes
For experimental infections of snakes, cultured trypanosomes of Crotalus durissus were inoculated into the caudal vein of either the same species of snake or snakes of different species, namely, Bothrops moojeni and Oxyrhopus guibei. The trypanosome forms and number of inoculated parasites varied according to the experiment. Infection was assessed by MH analysis of blood samples collected by tail bleeding from day 15 to day 60–120 at intervals of 20–30 days. Parasitaemias were assessed by determining the number of flagellates in MH buffy coats smeared on glass slides and stained by Giemsa.
PCR amplification of SSU rDNA, sequencing and data analysis
Genomic DNA of cultured trypanosomes was extracted by the classical phenol-chloroform method. PCR amplifications of the V7–V8 SSU rDNA regions or whole SSU rDNA were carried out using the oligonucleotides and reaction conditions as described previously (Rodrigues et al. Reference Siddall and Desser2006; Ferreira et al. Reference Ferreira, Campaner, Viola, Takata, Takeda and Teixeira2007). The PCR products were automatically sequenced, and the sequences obtained were aligned using ClustalX (Thompson, Reference Vidal and Hedges1997). The resulting alignment was manually refined. Phylogenetic inferences were assessed by Parsimony (P) and Bayesian (B) methods. Parsimony analyses performed in PAUP* v4b10 via 100 replicates of random addition sequence followed by branch swap (RAS-TBR) and bootstrap analysis (100 replicates) was carried out as previously described (Ferreira et al. Reference Ferreira, Campaner, Viola, Takata, Takeda and Teixeira2007). Distance matrices were generated using uncorrected p-distance. Bayesian analysis was performed in MrBayes v3.1.2 (Huelsenbeck et al. Reference Huelsenbeck, Ronquist, Nielsen and Bollback2001). Tree searches employed GTR plus gamma and a proportion of invariable sites. To infer phylogenetic relationships, sequences of snake trypanosomes determined in this study (Table 1) were aligned with sequences of several other trypanosomes (Fig. 3A), which were generated by previous studies (Lukes et al. Reference Lukes, Jirku, Dolezel, Kral'Ova, Hollar and Maslov1997; Haag et al. Reference Haag, O'h Uigin and Overath1998; Stevens et al. Reference Telford and Kreier2001; Hamilton et al. Reference Hamilton, Stevens, Gaunt, Gidley and Gibson2004), recovered from GenBank (Accession number): (a) Trypanosomes from lizards, T. therezieni (AJ223571); T. sp Gecko (AJ620548); T. scelopori (U67182); T. varani V54 (AJ005279); crocodile, T. grayi (AJ005278), and tortoise, T. chelodina (AF297086); (b) trypanosomes from anurans, T. rotatorium (B2-II) (AJ009161); T. mega (AJ223567); T. fallisi (AF119806); T. chattoni (AF119807) plus Brazilian isolates from toads and frogs, T. sp 444 (EU021225); T. sp 406 (EU021236); T. sp 362 (EU021232); T. sp 316 (EU021226); T. sp 364 (EU021229); (c) Trypanosomes from fish, T. triglae (U39584); T. boissoni (U39580); T. cobitis (AJ009143); Sequences of trypanosomes from mammals and birds were also included: (a) Mammals, T. vivax (U22316); T. brucei rhodesiense (AJ009142); T. congolense (savannah) (AJ009146); T. simiae (AJ009162); T. theileri (AJ009164); T. cyclops (AJ131958); T. sp (ABF) of wallaby (AJ620564); T. microti (AJ009158); T. lewisi (AB242273); T. dionisii (AJ009152); T. cruzi (VINCH 89) (AJ009149); T. cruzi marinkellei (AJ009150); T. conorhini (AJ012411); T. rangeli (AJ009160); (b) Birds, T. bennetti (AJ223562); T. corvi (AY461665); T. avium (APO1) (AF416559); T. avium (SIM3) (AF416563). Sequences of trypanosomes from sand flies were also added to the aligned sequences: T. sp 103 (EU021237), T. sp 887 (EU021245) and T. sp 888 (EU021241). Other sequences included in the alignment were T. binneyi from platypus (AJ132351) and T. sp (K&A) from aquatic leech (AJ009167). The alignment used in this study is available from the authors upon request.
Trypanosomes from snakes, lizards and sand fly nested in the clade lizard/snake positioned in the phylogenetic tree inferred in this study (Fig. 2)
(TryCC, Trypanosomatid Culture Collection, Department of Parasitology, ICB, USP, Brazil.)
1 Trypanosomes isolated in this study.
2 Ayala (Reference Ayala1970).
3 Minter-Goedbloed et al. (Reference Pessôa1993)
3,4 Hamilton et al. (Reference Hamilton, Stevens, Gaunt, Gidley and Gibson2004).
5 Putative vectors based on molecular and/or field evidence.
6 Experimentally confirmed vectors.
RESULTS
Occurrence and morphology of trypanosomes in the blood of the snakes investigated
We examined the blood of 459 snakes belonging to 37 different species in 4 families by microhaematocrit (MH) and haemoculture methods. The species were (a) Aniilidae – Anilius scytale (6 individuals examined); (b) Boidae – Boa constrictor (7), Corallus hortulanus (1); (c) Colubridae – Chironius bicarinatus (4), C. exoletus (12), C. fuscus (2), C. quadricarinatus (3), C. scurrulus (1), Dipsas catesbyi (1), Drymarchon corais (2), Echinanthera sp. (2), Leptodeira annulata (8), Leptophis ahaetulla (3), Liophis almadensis (2), L. atraventer (1), L. miliaris (1), L. poecilogyrus (4), L. reginae (3), L. typhlus (1), Mastigodryas boddaerti (1), Oxyrhopus trigeminus (1), O. clathratus (1), O. guibei (12), Philodryas olfersii (3), P. patagoniensis (4), Pseudoboa nigra (1), Rachidelus brazili (2), Sibynomorphus mikanii (1), Spilotes pullatus (9), Thamnodynastes rutilus (4), T. strigatus (1), Tomodon dorsatus (12), Tropidodryas sp. (2), Xenopholis undulatus (1); (d) Viperidae – Bothrops spp. (202), Crotalus durissus ssp. (134), Lachesis muta (4).
Microhaematocrit examination was positive for 7 Crotalus durissus (rattlesnake, cascavel) and for 4 Bothrops jararaca (lancehead, jararaca). Blood smears from rattlesnakes showed large and wide trypomastigotes (average 55 μm long by 8 μm wide) pointed at the extremities, with a large undulating membrane and usually a short free flagellum. The kinetoplast was very small and positioned adjacent to the outer body margin, distant from the large, ellipsoid and nearly central nucleus. Trypomastigotes were usually coiled in a circle with long overlapping extremities (Fig. 1A1) and dividing forms were not observed. No trypanosomes were found in lancehead blood smears, but microscopy analysis directly of the MH buffy coats revealed trypanosomes similar to those in the rattlesnake blood smears. Blood trypomastigotes from rattlesnakes examined in this study were indistinguishable from those of T. cascavelli first described by Pessôa and De Biasi (Reference Pessôa and Fleury1972) in the same snake species, but could easily be distinguished from those of T. butantanense from Ophis merremii shown in microphotographs of blood smears used for its first description in an earlier study (Arantes and Fonseca, Reference Arantes and Fonseca1931a) (Fig. 1A2).
Selected microphotographs (Giemsa-stained) of blood trypanosomes from naturally and experimentally infected snakes. (A1) Blood trypomastigotes of specimens of Crotallus durissus from which isolates 632, 631, 629 and 693 were obtained; (A2) blood trypomastigotes from archived slides used for the first description of Trypanosoma cascavelli from C. durissus and Trypanosoma butantanense from Ophis merremii. (B) Blood trypomastigotes of C. durissus and Oxyrhopus guibei snakes experimentally infected with cultured forms of isolate 425 from C. durissus. k, Kinetoplast; f, flagellum; n, nucleus.
Isolation, growth and morphology of culture forms of snake trypanosomes
Snake blood samples examined for the presence of trypanosomes by haemocultures revealed 7 positive cultures from C. durissus (Viperidae) from southeastern Brazil and 1 from Pseudoboa nigra (Colubridae) from central Brazil (Table 1). Positive cultures from C. durissus came from snakes with positive MH whereas the culture from P. nigra came from a specimen that was not examined by MH. Low parasitaemias may be responsible for negative results of blood smears and MH. On the other hand, haemocultures can be negative for MH-positive snakes. This was the case with B. jararaca (Viperidae), which although having a positive MH consistently failed to yield positive haemocultures suggesting different growth requirements. All cultures were stored in liquid nitrogen in our culture collection.
Cultures of trypanosomes from C. durissus (isolate 425) and P. nigra (isolate 1052) yielded abundant growth when co-cultivated over a monolayer of Hi-5 insect feeder cells. In culture, isolates from both snakes showed similar morphological features and growth behaviour. The morphology and developmental stages varied according to the culture phase (Fig. 2). After an almost quiescent phase of ~7 days, the number of flagellates increased markedly (log phase). In this phase there was a predominance of epimastigotes of different sizes and shapes, most of which were slender and had a sharp posterior extremity, a small dot-like kinetoplast positioned very close to the nucleus, a small undulating membrane and a long flagellum (Fig. 2). At the end of the log-phase, after ~10 days, the cultures displayed large epimastigotes with an enlarged anterior end, central nucleus and kinetoplast, conspicuous undulant membrane and a long free flagellum. Large trypomastigotes began to appear in this phase and predominated in the end of stationary phase (~15 days), when 2 trypomastigote forms were observed (Fig. 2). One form displayed a wide and roll-shaped body with a conspicuous and many-folded undulating membrane and the kinetoplast closer to the nucleus than to the posterior end. These wide forms were similar in shape but smaller than the blood trypomastigotes (Fig. 1A). The other trypomastigote form was small, slender, with a narrow undulating membrane and a small kinetoplast situated between the nucleus and the posterior end. These smallest forms, probably the metacyclic trypomastigotes, predominated in the end stage of cultures, and their number could be considerably increased when flagellates were transferred to Hi-5 cultures in Grace's medium containing only 2% FBS (Fig. 2).
Microphotographs (Giemsa-stained) selected to illustrate the development in culture of trypanosomes from snakes. Isolate 621 (T. cascavelli) from C. durissus; isolate 1052 from Pseudoboa nigra; and isolate 910 from sand fly. Sequential photographs represent the developmental forms observed in cultures of flagellates co-cultivated with Hi-5 insect feeder cells. (A) Epimastigotes (logarithmic phase, ~7 days); (B) enlarged epimastigotes (~10 days); (C) trypomastigotes and epimastigotes (stationary phase, ~15 days); (D) wide (w) and slender (s) trypomastigotes in the end stage of cultures (after 15 days). k, Kinetoplast; f, flagellum; n, nucleus.
Growth and morphology of culture forms of trypanosomes from sand flies
Trypanosome isolate 910 was recovered from the gut of a sand fly (Viannamyia tuberculata) captured in the Brazilian Amazonia that had a small number of flagellates in its anterior and posterior stomach. However, no red blood cells, which would indicate a recent blood meal, were observed in the fly's gut. This isolate exhibited the same growth and morphological features described above for snake trypanosomes (Fig. 2). This isolate was selected for this study because its V7–V8 SSU rDNA sequence was shown to be high, similar to corresponding sequences from snake trypanosomes. Other trypanosomes from sand flies included in this study differed in growth behaviour and morphology and clustered together with anuran trypanosomes in phylogenetic trees (Ferreira et al., manuscript in preparation).
Experimental infection of snakes and cross-infectivity of snake trypanosomes
Culture forms (4×107) of rattlesnake trypanosomes (isolate 621) were inoculated into specimens of laboratory-raised C. durissus, which still had trypanosomes in their blood 120 days later. The course of infection differed according to the percentage of trypomastigotes (wide or slender) inoculated. Parasitaemia was significantly higher (maximum 4×103 flagellates/ml) in 2 snakes inoculated with ~70% slender and 30% wide forms than in 2 snakes inoculated with ~30% slender and 70% wide forms (maximum parasitaemia 3·3×102 flagellates/ml). Snakes inoculated with epimastigotes did not become infected. Parasitaemias reached their highest levels in 2 weeks, remained high for 1 month and then gradually declined. The rattlesnakes with the highest parasitaemias became apathetic, asthenic and anorexic and died within 4 months. Giemsa-stained imprints of heart, spleen, bone marrow, brain and liver of these animals failed to reveal tissue trypanosomes.
For cross-infection experiments, cultured trypanosomes from rattlesnakes (isolate 425) were also inoculated (1×105 trypomastigotes, ~70% slender forms) in 1 snake of each of the following species: Bothrops moojeni (Viperidae), Oxyrhopus guibei (Colubridae) and C. durissus (Viperidae). All the infected snakes had large blood trypanosomes detectable by MH. The parasitaemias and course of infection varied according to the snake species. The number of blood parasites was always higher in rattlesnakes (maximum ~103 parasites/ml) than in O. guibei (maximum ~4×102 flagellates/ml). B. moojeni also became infected, but parasites were scarce and only detected by MH (maximum ~50 parasites/ml). Trypomastigotes found in blood from experimentally infected C. durissus and O. guibei were indistinguishable (Fig. 1B), but were clearly different from both wide and slender cultured trypomastigotes used for snake infections (Fig. 2). In addition, these forms were indistinguishable from those detected in the blood of naturally infected C. durissus (Fig. 1A).
Genetic diversity among snake trypanosomes assessed by sequence polymorphisms of the variable V7–V8 region of SSU rDNA
Sequences of the V7–V8 region of SSU rDNA were determined for the 8 snake-trypanosome isolates studied here. Comparison of aligned sequences revealed high sequence similarity among the 7 isolates from C. durissus (average of ~98% shared similarity). Only isolates 630 and 632 showed slight differences. In contrast, the P. nigra isolate diverged significantly from those from C. durissus (~93% similarity), and the resulting dendrogram clearly separated the cluster formed by the isolates from C. durissus from P. nigra isolate 1052 (Fig. 3B). The sequence of the trypanosome isolated from the sand fly V. tuberculata had an average of ~94% similarity with C. durissus trypanosome sequences, and smaller similarity (90%) with the P. nigra isolate. The highest similarities with sequences from snake trypanosomes were shared by 3 lizard trypanosomes, namely, T. varani, T. scelopori and T. sp gecko (average ~70%). T. varani was included in the dendrograms as an outgroup of snake trypanosomes (Fig. 3B).
Phylogenetic analyses based on ribosomal sequences inferred by Parsimony. (A) Unrooted phylogeny inferred using SSU rDNA sequences of 44 trypanosomes, including sequences of 2 isolates from snakes (●), 4 from lizards (■), 4 from sand flies (▲) and 8 from anurans (★). (B) Dendrogram based on V7-V8 SSU rDNA sequences from the following isolates nested in the snake clade: 5 isolates of T. cascavelli from C. durissus; T. sp 630 and T. sp 632 from C. durissus; T. sp 1052 from P. nigra; and T. sp 910 from sand fly. T. varani from lizard was used as outgroup for the trypanosomes of the snake clade. The numbers at the nodes (n) correspond to Parsimony percentage bootstrap values derived from 100 replicates.
Phylogenetic relationships among trypanosomes from snakes, other reptilians, sand flies and other vertebrates inferred from whole sequences of SSU rDNA
The phylogenetic relationships among trypanosomes from snakes and several species of Trypanosoma from other hosts was inferred using whole sequences of SSU rDNA from 1 isolate from C. durissus (isolate 425) and from the isolate 1052 from P. nigra. These sequences were aligned with all available sequences of reptilian trypanosomes and with selected sequences of trypanosomes from mammals, amphibians, fishes and birds. Phylogenies inferred using Parsimony (Fig. 3A) and Bayesian (data not shown) methods generated trees with very similar branching patterns, showing all major clades of trypanosomes. In both analyses, a clade was formed by snake and sand fly trypanosomes (snake clade, 100% bootstrap), which was always positioned closest to a clade formed by the lizard trypanosomes T. varani, T. scelopori and T. sp gecko (lizard clade, 100% bootstrap) (Fig. 3A).
The sand fly trypanosome (isolate 910) characterized in this study was tightly positioned together with the snake trypanosomes and analysis of whole SSU rDNA sequences confirmed that it is closest to C. durissus isolates (~98% similarity) than to the isolate 1052 from P. nigra (~93%). Three trypanosome isolates from sand flies (103, 887 and 888) clustered with anuran trypanosomes. Despite significant internal divergence, the clade formed by lizard and snake trypanosomes (lizard/snake clade) was well supported (100% bootstrap) and separated by a large genetic distance from all other clades. T. therezieni from chameleon clustered with anuran trypanosomes, and was the only squamate trypanosome positioned outside the lizard clade. The aquatic clade, constituted by leech-transmitted trypanosomes of fishes, tortoises and platypuses, was separated by a high genetic distance from the lizard/snake clade (Fig. 3A). Considering that previous phylogenetic analysis demonstrated that SSU rDNA gene sequences alone were not reliable markers for inferring deep level trypanosome phylogeny (Hamilton et al. Reference Hamilton, Stevens, Gaunt, Gidley and Gibson2004, Reference Hamilton, Gibson and Stevens2007), we constructed an unrooted phylogeny to reveal the relationships among trypanosomes from snakes, other reptilians, sand flies and other vertebrates (Fig. 3A).
DISCUSSION
Despite several reports of trypanosomes in terrestrial and aquatic snakes not a single snake trypanosome has been molecularly characterized and, with the exception of the present study, snake trypanosomes have not been included in any phylogenetic studies.
The results of our surveys of snake blood showed a trypanosome prevalence of 5·6% for C. durissus and 2% for B. jararaca, with an overall prevalence of 2·33% in a total of 459 snakes from 37 species examined. In general, previous studies revealed low prevalence and parasitaemias of snake trypanosomes: De Biasi et al. (Reference De Biasi, Pessôa, Puorto and Fernandes1975) had to examine 146 Ophis merremii in order to find an infected one; Pessôa and Fleury (Reference Rodrigues, Paiva, Campaner, Stevens, Noyes and Teixeira1969) described 1 snake infected by T. constrictor and 2 infected by T. salamantae among more than 700 specimens of Boidae examined; Pessôa and De Biasi (Reference Pessôa and Fleury1972) had to examine 300 rattlesnakes to find 4 infected with T. cascavelli. In contrast, 30% of Rachidelus brazili were found infected with T. hogei (Pessôa, Reference Pessôa and De Biasi1968). Also exceptional was the finding of snakes displaying moderate and high parasitaemias, such as Ophis merremii infected with T. butantanense (Arantes and Fonseca, Reference Arantes and Fonseca1931a) and the aquatic snakes Nerodia fasciata confluens and Agkistrodon piscivorus leucostoma infected with T. hydrae and T. yaegeri, respectively (Ayala et al. Reference Ayala, Atkinson and Vakalis1983).
Very little is known about host specificity in snake trypanosomes. In the few attempts at cross-infection that have been carried out, T. constrictor of Boa constrictor failed to infect Helicops modestus and T. salamantae of Epicatres cenchria failed to infect Boa constrictor (Pessôa and Fleury, Reference Rodrigues, Paiva, Campaner, Stevens, Noyes and Teixeira1969). In contrast, T. butantanense infected 9 snake species (Arantes and Fonseca, Reference Arantes and Fonseca1931a), and T. hydrae infected many species of colubrids (Ayala et al. Reference Ayala, Atkinson and Vakalis1983). Our results show that the trypanosome from the viperid C. durissus is able to infect the viperid B. jararaca and the colubrid O. guibei. These observations, although suggesting a lack of host restriction, do not allow any generalizations about what happens in field conditions. The fact that the 7 isolates from C. durissus captured in different locations at different times are similar to each other favours host-specificity. Cross-infection data from this and previous studies (Arantes and Fonseca, Reference Arantes and Fonseca1931a; Ayala et al. Reference Ayala, Atkinson and Vakalis1983) indicate that classification of snake trypanosomes based on the criterion of host origin is unacceptable.
Our results showed that the trypanosomes recovered either from the blood of C. durissus or O. guibei experimentally infected with a trypanosome isolate from C. durissus displayed identical morphology. In addition, the trypomastigotes from C. durissus examined in the present study are identical to those described for T. cascavelli (Pessôa and De Biasi, Reference Pessôa and Fleury1972). The fact that these trypanosomes come from the same host species does not prove that they are of the same species, as neither host origin nor morphology provides sufficient taxonomic criteria. Nevertheless, based on the morphological similarity of blood trypomastigotes from field-captured C. durissus in this study and those in the study of Pessôa and De Biasi (Reference Pessôa and Fleury1972) and on the lack of significant morphological and genetic polymorphism among the C. durissus isolates, we are keeping the name T. cascavelli for the isolates from this snake characterized in this study. The sand fly isolate (910) remained as Trypanosoma sp. Although the trypanosome from P. nigra (Trypanosoma sp 1052) characterized in this study is the first described from this species and was clearly separated by morphological and molecular data from C. durissus isolates, further data are required before it can be described as a new species.
Like trypanosomes from fishes, tortoises and some species of anurans, trypanosomes from aquatic snakes can be transmitted by leeches (Pessôa and Fleury, Reference Rodrigues, Paiva, Campaner, Stevens, Noyes and Teixeira1969; Chia and Miller, Reference Chia and Miller1984). Nevertheless, although also found in the Amazon region, the snakes C. durissus and P. nigra, which naturally harbour the trypanosomes characterized in this study, inhabit mainly savannah-like habitats, where leeches are not commonly found. The supposed role of ticks in the transmission of snake trypanosomes has never been investigated. Despite reports of the tick Amblyoma rotundatum in C. durissus (Dantas-Torres et al. Reference Dantas-Torres, Oliveira-Filho, Souza and Sá2005), we were unable to make ticks feed on rattlesnakes and trypanosomes did not develop in blood-ingurgitated A. rotundatum fed on the colubrid O. guibei infected with T. cascavelli (data not shown).
If leeches and ticks are excluded, haematophagous insects remain the only candidate vectors for terrestrial snake trypanosomes. Among them, sand flies stand out because they are vectors of trypanosomes among terrestrial lizards and anurans (Anderson and Ayala, Reference Anderson and Ayala1968; Ayala and McKay, Reference Ayala and McKay1971; Minter-Goedbloed et al. Reference Pessôa1993; Telford, Reference Thompson, Gibson, Plewniak, Jeanmougin and Higgins1995; Ferreira et al., manuscript in preparation). Accordingly, we characterized a new trypanosome from the sand fly V. tuberculata that was positioned closest to snake trypanosomes, suggesting that it is a trypanosome of another snake species. V. tuberculata has a widespread distribution and is commonly found in distinct ecotopes at the lower level of tree trunks (Freitas et al. Reference Freitas, Naiff and Barrett2002; Lainson et al. Reference Lainson, Ishikawa and Silveira2002). This is the first evidence of the role of sand flies in the transmission of trypanosomes among snakes.
The phylogenetic analyses of reptilian trypanosomes done in this study confirmed both their polyphyly and distribution in distant branches, with some clades suggesting a long association of trypanosomes with their vertebrate hosts, such as those related to Squamata and Crocodylia, which are separated by a large genetic distance, compatible with the Early Jurassic split among these major reptilian clades. The lizard/snake clade was formed only by trypanosomes from hosts of Squamata, including species of infraorders Serpentes, Iguania, Gekkota and Anguimorpha (Douglas et al. Reference Douglas, Janke and Arnason2006). The trypanosomes from snakes here examined appear to be transmitted by sand flies, as indicated by the tight positioning of a trypanosome from a sand fly within the clade of snake trypanosomes. Corroborating the role of sand flies as vectors, all the lizard trypanosomes nested in the lizard/snake clade are sand fly-transmitted (Ayala, Reference Ayala1970; Ayala and McKay, Reference Ayala and McKay1971; Christensen and Telford, Reference Christensen and Telford1972; Minter-Goedbloed et al. Reference Pessôa1993). However, sand flies also are proven vectors for anuran trypanosomes (Ayala, Reference Ayala1971; Ferreira et al., manuscript in preparation).
Herein, we provide biological, morphological and molecular phylogenetic evidence supporting a new lineage of trypanosomes from snakes and sand flies closely related to the lizard clade forming an assemblage of squamate trypanosomes, which we called the lizard/snake clade. This clade includes 2 trypanosomes from Brazilian snakes: T. cascavelli from C. durissus and a new unnamed species (T. sp 1052) from P. nigra. The lizard trypanosomes clustered into this clade were T. scelopori from the USA, and T. varani and T. sp gecko from Senegal, West Africa (Ayala, Reference Ayala1970; Hamilton et al. Reference Hamilton, Stevens, Gaunt, Gidley and Gibson2004). T. scelopori is from Phrynosomatidae and T. varani from Varanidae. These lizard families are respectively from the closely related Iguania and Anguimorpha infraorders, which are phylogenetically distant from the Serpentes (Douglas et al. Reference Douglas, Janke and Arnason2006). Therefore, the considerable genetic distance separating snake and lizard clades suggests some patterns of host-parasite association in the evolutionary history of squamate trypanosomes. However, all data corroborate more complex evolutionary histories for these trypanosomes. Accordingly, T. therezieni from the African chameleon, despite also belonging to the Iguania infraorder (Douglas et al. Reference Douglas, Janke and Arnason2006), was distantly positioned within the clade of anuran trypanosomes in this and in previous studies (Haag et al. Reference Haag, O'h Uigin and Overath1998; Stevens et al. Reference Telford and Kreier2001; Hamilton et al. Reference Hamilton, Stevens, Gaunt, Gidley and Gibson2004, Reference Hamilton, Gibson and Stevens2007).
Clades harbouring reptilian trypanosomes also suggested an association with vectors and ecotopes. Although from the distant infraorder Gekkota, T. sp gecko nested within the lizard clade closest to T. varani. Interestingly, these trypanosomes were isolated from hosts of the same African region that share the same ecotope and sand fly vectors. There is clear evidence that sand flies are vectors of trypanosomes for lizards (Ayala, Reference Ayala1970; Ayala and McKay, Reference Ayala and McKay1971; Christensen and Telford, Reference Christensen and Telford1972; Minter-Goedbloed et al. Reference Pessôa1993). Data from this study suggested that sand flies are vectors of trypanosomes for terrestrial snakes of savannah-like habitats. T. chelodina from aquatic turtle was nested in the clade that comprises fish trypanosomes transmitted by aquatic leeches (Siddal and Desser, Reference Stevens, Noyes, Schofield and Gibson1992; Jakes et al. Reference Jakes, O'donoghue and Adlard2001). Taken together, results presented in this and in previous studies suggest that host switching, probably mediated by vectors, appears to play an important role in the evolutionary history of reptilian trypanosomes in general (Maia da Silva et al. 2007; Ferreira et al. Reference Ferreira, Campaner, Viola, Takata, Takeda and Teixeira2007; Hamilton et al. Reference Hamilton, Gibson and Stevens2007).
Trypanosomes of terrestrial snakes, which appear to be transmitted by sand flies, are morphologically distinguishable from species infecting aquatic snakes transmitted by aquatic leeches (Brygoo, Reference Brygoo1965a, Reference Brygoob; Chia and Miller, Reference Chia and Miller1984). Phylogenetic analysis of trypanosomes from aquatic and terrestrial snakes can contribute to a better understanding of the evolutionary relationships of these trypanosomes with their vertebrate and invertebrate hosts. Furthermore, an understanding of the relationships among snake trypanosomes and those from other squamates could provide information helpful to evaluate the evolutionary scenarios hypothesized for the Serpentes, namely, that snakes evolved on land directly from burrowing lizards or that they evolved from extinct marine reptiles (Vidal and Hedges, Reference Wallace, Camargo, McGhee and Roitman2005; Douglas et al. Reference Douglas, Janke and Arnason2006).
We are indebted to several collaborators for their help in the snake handling. We are grateful to Dr Darci M. B. Battesti for collaboration in the experiments with ticks. We thank Dr Miguel T. Rodrigues and Dr Jeffrey J. Shaw for their critical comments on the manuscript and continued encouragement. This work was supported by grants from the Brazilian agencies CNPq and FAPESP. R. A. F. and T. V. B. were financed by Furnas Centrais Elétricas SA, by agreement with INPA. A. C. Rodrigues is a post-doctoral fellow sponsored by CNPq. L. B. Viola and R. C. Ferreira were recipients of scholarships from CNPq and FAPESP, respectively.
