Introduction

Infections by Toxoplasma gondii are cosmopolitan, affecting endothermal animals, including humans ( ^{DUBEY, 2010} ). In Brazil, studies have revealed seroprevalence of approximately 90% in humans, especially in the southern region of the country, and also that clinical illness is almost always associated with cases of immunosuppression ( ^{DUBEY, 2010} ). Recent studies regarding the population genetics of T. gondii worldwide have shown that, although reproduction is almost always asexual, the genetic variability of this infectious agent is much greater than previously expected ( ^{PENA et al., 2008} ; ^{SHWAB et al., 2014} ). This variability may be related to clinical conditions, which can range from unapparent infections to cases of encephalitis. Some of these conditions can end up causing the death of the host, even among immunocompetent individuals, whether these are humans or animals ( ^{DEMAR et al., 2008} ; ^{CARME et al., 2009} ; ^{DUBEY, 2010} ; ^{WENDTE et al., 2011} ). Studies have suggested that the existence of a wild cycle for T. gondii with sexual reproduction in wild felids, together with genetic changes due to clonal reproduction (asexual) in multiple species of wild animals, could be intrinsically related to the existence of greater genetic variability of the parasite in South America ( ^{CARME et al., 2009} ; ^{KHAN et al., 2011} ; ^{WENDTE et al., 2011} ).

^{Demar et al. (2008)} and ^{Carme et al. (2009)} suggested that this phenomenon may occur in Brazil, especially in areas bordering countries such as Guyana, French Guiana and Suriname, because of the continuous nature of the Amazon rainforest. Different genotypes of T. gondii found in isolates from Brazilian wild fauna reinforce this hypothesis ( ^{PENA et al., 2011} ; ^{CAÑÓN-FRANCO et al., 2013} ; ^{BARROS et al., 2014} ; ^{VITALIANO et al., 2014} ). However, studies on the genetics of T. gondii among the wild fauna of northeastern Brazil are still scarce.

In the state of Pernambuco, Brazil, ^{Pena et al. (2011)} carried out isolation and molecular characterization of T. gondii from a young male red-handed howler monkey (Alouatta belzebul) and from an adult male jaguarundi (Puma yagouaroundi) at the zoo of Parque Estadual de Dois Irmãos, located in Recife, Pernambuco. In other municipalities of this state, seroepidemiological surveys of infection by T. gondii in wild animals kept in captivity were conducted among neotropical felids ( ^{RAMOS SILVA et al., 2001} ), wild birds ( ^{LEITE et al., 2007} ), capuchin monkeys (Sapajus spp.) ( ^{FERREIRA et al., 2015} ), and marine manatees (Trichechus manatus) ( ^{ATTADEMO et al., 2016} ). Investigations on anti-T. gondii antibodies in free-ranging wild animals were conducted among cattle egrets (Bubulcus ibis) in Fernando de Noronha ( ^{COSTA et al., 2012} ), and among rodents and marsupials in the Atlantic Rainforest ( ^{SIQUEIRA et al., 2013} ). These studies demonstrated that infection by this parasite among wild animals presents wide geographical distribution in Pernambuco, which demands research on isolation and genetic characterization of T. gondii.

Considering the wide array of wild species with positive results in serodiagnostic investigations on T. gondii infection and the relatively low number of published scientific papers on genetic characterization of this parasite from wild animals in northeastern Brazil, the objective of the present study was to isolate and characterize the genotypes of T. gondii strains in free-ranging and captive wild mammals and birds in the state of Pernambuco.

Material and Methods

Between March 2014 and September 2015, fragments of the brain, heart, skeletal muscle and diaphragm of 105 wild animals (71 birds and 34 mammals) were collected ( Table 1 ). These specimens were either free-living (n=59) or kept in captivity (n=46) at the zoo and Conservation Unit of Parque Estadual de Dois Irmãos and from the Tangará wild animal screening center of Agência Estadual de Meio Ambiente de Pernambuco (CPRH). Sampling was performed according to convenience, after these animals had died due to different causes.

The samples thus collected were divided into two groups. Samples from thirty-two animals were used for bioassays in mice and those from 73 animals were used to make direct molecular diagnosis. For the bioassay, tissue samples (between 5 and 50 g depending on the size of the animal) were collected, pooled and processed in accordance with the technique described by ^{Dubey (1998)} . Aliquots of 1.0 mL of the final homogenate were inoculated subcutaneously into each of three Swiss mice, which were kept under observation for 45 days regarding the appearance of clinical signs of toxoplasmosis. At the end of the 45-day period, blood was drawn from the surviving mice for serological evaluation through the Modified Agglutination Test (MAT, cut-off titre ≥ 25) ( ^{DUBEY & DESMONTS, 1987} ). The mice were then euthanized and tissue imprints of their lungs and brains were examined to search for tachyzoites or tissue cysts to confirm infection by T. gondii ( ^{DUBEY, 1998} ).

Tissue samples collected for direct molecular diagnosis were individually macerated into a 0.85% saline solution. An aliquot of 300 µL of each homogenate was subjected to DNA extraction using a commercial kit (Qiagen® DNeasy Blood & Tissue, Dusseldorf, Germany), following the recommendations of the manufacturer. Toxoplasma gondii DNA in the samples was screened through PCR using primers TOX4 (CGCTGCAGGGAGGAAGACGAAAGTTG) and TOX5 (CGCTGCAGACACAGTGCATCTGGATT), targeting a non-coding 529bp DNA fragment that is repeated 200-300 times in the T. gondii genome ( ^{HOMAN et al., 2000} ).

The DNA extracted from the positive samples was then subjected to genotyping through the PCR-RFLP technique using 11 markers: SAG1, 3’5’SAG2 + alt.SAG2, SAG3, BTUB, GRA6, c22-8, c29-2, L358, PK1, Apico ( ^{SU et al., 2010} ) and CS3 ( ^{PENA et al., 2008} ). Samples with complete RFLP genotyping were also analyzed using 15 microsatellite markers (MS): TUB2, W35, TgMA, B18, B17, M33, IV.1 and X1.1 (typing markers); and N60, N82, AA, N61 N83, M48 and M102 (fingerprinting markers), following the protocols of ^{Ajzenberg et al. (2010)} , in order to increase genotyping precision. The results were analyzed using the software Genemapper® 4.1 (Applied Biosystems).

The present study was conducted while respecting biosafety and animal welfare norms under authorizations from SISBIO (No. 37855-1; Instituto Chico Mendes de Conservação da Biodiversidade, ICMBio); the Ethics Committee for the Use of Animals (CEUA) of Universidade Federal Rural de Pernambuco (UFRPE) (No. 109/2014); the Bioethics Committee of the School of Veterinary Medicine and Animal Science of Universidade de São Paulo (FMVZ-USP) (No. 1921061113); Parque Estadual de Dois Irmãos (PEDI) (No. 216/2014); and Agência Estadual de Meio Ambiente de Pernambuco (CPRH) (No. 010675/2014).

Results

Among the 32 samples animals that were used for the bioassay in mice, T. gondii was isolated from a free-ranging striated heron (Butorides striata) (3.1%) in the metropolitan region of Recife-PE. This isolate was named TgButstBrPE1, and ToxoDB-RFLP genotype #13 was identified in it. During the 45-day period of post-inoculation observation, none of the three mice inoculated in this group died or presented clinical signs of toxoplasmosis.

Among the 73 animals used for direct diagnosis by means of PCR, seven (9.6%) were positive for T. gondii. Heart was the positive tissue in six of these animals. Type BrIII genotype (ToxoDB-RFLP genotype #8) was diagnosed from a captive otter (Lontra longicaudis ) at the zoo of Parque Estadual de Dois Irmãos ( Table 2 ). This sample was identified as PS-TgLonloBrPE1. Table 3 lists the results from RFLP genotypic characterization of the T. gondii isolate from the free-ranging striated heron and from the primary sample of the captive otter.

The microsatellite analysis enabled genotyping of the isolate TgButstBrPE1, which presented the following sequence of alleles: TUB2 = 291, W35 = 242, TgMA = 205, B18 = 162, B17 = 342, M33 = 165, IV.1 = 278, XI.1 = 356, N60 = 142, N82 = 109, AA = 277, N61 = 87, N83 = 312, M48 = 213 and M102 = 164, which corresponds to Caribbean 1 type.

Discussion

Diagnoses of T. gondii infection obtained through isolation and molecular methods among wild birds in Brazil have been reported both from free-living wild specimens ( ^{GONDIM et al., 2010} ; ^{LIMA et al., 2011} ; ^{VILELA et al., 2011} ; ^{BARROS et al., 2014} ; ^{VITALIANO et al., 2014} ) and from captive specimens ( ^{VITALIANO et al., 2010} ). Nonetheless, research involving molecular diagnoses of T. gondii in northeastern Brazil is still scarce ( ^{GONDIM et al., 2010} ). In the case of wild mammals, isolation and molecular diagnosis seem to be more common, in relation to both captive and free-living specimens ( ^{SILVA et al., 2006} ; ^{TRUPPEL et al., 2010} ; ^{PENA et al., 2011} ; ^{CABRAL et al., 2013} ; ^{CAÑÓN-FRANCO et al., 2013} ; ^{VITALIANO et al., 2014} ).

The ToxoDB-RFLP genotype #13 that was identified in the present study had been previously described in isolates from chickens, goats, a red-handed howler monkey and a fox (Cerdocyon thous ) in northeastern Brazil ( ^{DUBEY et al., 2008} ; ^{FEITOSA et al., 2017} ; ^{RAGOZO et al., 2010} ; ^{PENA et al., 2011} ; ^{ALMEIDA et al., 2017} ). These findings corroborate the wide circulation of this genotype in this region. The present study provides the first description of a T. gondii isolate genotyped as #13 from a wild bird. This isolate caused no mortality in mice, thus corroborating other authors’ findings that ToxoDB-RFLP genotype #13 is associated with low virulence in mice ( ^{DUBEY et al., 2008} ; ^{RAGOZO et al., 2010} ; ^{PENA et al., 2011} ). Moreover, a microsatellite analysis (which presented greater resolution) revealed the same sequence of alleles (unpublished data), that were found in the isolate from a red-handed howler monkey, also from Recife, Pernambuco ( ^{PENA et al., 2011} ). This indicates that there is wide circulation of this clone in the city of Recife. The fact the striated heron examined here was rescued a few days before its death indicates that the infection was acquired in the wild. This suggests that the isolates from domesticated animals and from captive wild animals are genotypically similar to those isolated from free-ranging wild animals.

Regarding direct molecular diagnosis based on primary samples, ^{Dubey (2010)} reported that this technique was very effective for the diagnosis, but not so much for genotyping, due to the small amount of genetic material usually found. Molecular diagnosing of T. gondii using primary samples from wild animals has been described in capybaras ( ^{TRUPPEL et al., 2010} ), sparrows ( ^{VILELA et al., 2011} ), several species of small neotropical felids ( ^{CAÑÓN-FRANCO et al., 2013} ) and several species of wild birds and mammals ( ^{VITALIANO et al., 2014} ). ^{Cañón-Franco et al. (2013)} also conducted genotyping using primary samples, individually, obtaining 14.6% (63/433) of success, similar to what was obtained in the present study (14.3%).

Two blond capuchins (Sapajus flavius), two otters (L. longicaudis ), one spectacled owl (Pulsatrix perspicillata), one American kestrel (Falco sparverius) and one striated heron (B. striata ) were positive for T. gondii using PCR. These are the first reports of molecular diagnosis using primary samples in those species. In the case of the blond capuchins, a critically endangered species, identification of this agent is important, since these specimens were part of a reproduction program for ex situ conservation ( ^{LYNCH ALFARO et al., 2014} ; ^{FERREIRA et al., 2015} ; ^{IUCN, 2018} ).

The otter in which the T. gondii Type BrIII genotype was identified did not present any clinical sign of toxoplasmosis. The death of this specimen was associated with senility and multiple organ failure due to neoplastic metastases. Type BrIII is a common clonal lineage circulating in Brazilian territory, and it is found in different animal hosts and also in humans.

The two striated herons that were positive for T. gondii in the present study were individuals from two different districts in Recife. They had been living near water courses in densely populated areas. The facts that this species inhabits urban areas and has a diet composed of fish and other aquatic organisms ( ^{GWYNNE et al., 2010} ) suggest that the transmission route was probably ingestion of T. gondii oocysts from water contaminated by domestic sewage. These results indicated that this species is an intermediate host and can serve as an indicator for T. gondii infection among the synanthropic wild fauna of Pernambuco.

Conclusion

The genotyping information obtained from the samples of wild animals of the present study corroborated the findings that ToxoDB-RFLP genotype #13 has wide circulation in northeastern Brazil and that Type BrIII is a widely distributed clonal lineage in Brazil.