Mem Inst Oswaldo Cruz, Rio de Janeiro, 113(1) January 2018
Trypanosoma cruzi strain TcIV infects raccoons from Illinois
1Southern Illinois University, Department of Zoology, Carbondale, IL, USA
2Southern Illinois University Carbondale, Department of Forestry, Carbondale, IL, USA
3Southern Illinois University Carbondale, Cooperative Wildlife Research Laboratory, Carbondale, IL, USA
4Western Kentucky University, Department of Biology, Bowling Green, KY, USA
BACKGROUND The northern limits of Trypanosoma cruzi across the territory of the United States remain unknown. The known vectors Triatoma sanguisuga and T. lecticularia find their northernmost limits in Illinois; yet, earlier screenings of those insects did not reveal the presence of the pathogen, which has not been reported in vectors or reservoir hosts in this state.
OBJECTIVES Five species of medium-sized mammals were screened for the presence of T. cruzi.
METHODS Genomic DNA was isolated from heart, spleen and skeletal muscle of bobcats (Lynx rufus, n = 60), raccoons (Procyon lotor, n = 37), nine-banded armadillos (Dasypus novemcinctus, n = 5), Virginia opossums (Didelphis virginiana, n = 3), and a red fox (Vulpes vulpes). Infections were detected targeting DNA from the kinetoplast DNA minicircle (kDNA) and satellite DNA (satDNA). The discrete typing unit (DTU) was determined by amplifying two gene regions: the Spliced Leader Intergenic Region (SL), via a multiplex polymerase chain reaction, and the 24S? ribosomal DNA via a heminested reaction. Resulting sequences were used to calculate their genetic distance against reference DTUs.
FINDINGS 18.9% of raccoons were positive for strain TcIV; the rest of mammals tested negative.
MAIN CONCLUSIONS These results confirm for the first time the presence of T. cruzi in wildlife from Illinois, suggesting that a sylvatic life cycle is likely to occur in the region. The analyses of sequences of SL suggest that amplicons resulting from a commonly used multiplex reaction may yield non-homologous fragments.
The etiologicalagent of American trypanosomiasis, Trypanosoma cruzi (Euglenozoa) commonlyinfects mammals as well as triatomine bugs throughout the tropical and subtropicalregions of the New World. In the United States of America, the presence of T.cruzi has been documented in both vectors and in wildlife screened for thepresence of the parasite in 16 states (Bi et al. 2010 2 , Brown et al. 2010 6 , Bernet al. 2011). Additionally, positive vectors have been identified in nine states,including Alabama, Arizona, California, Georgia, Louisiana, New Mexico, Tennesseeand Texas (Bern et al. 2011 1 ). Contrastingly, the distribution of T. cruziin wildlife has been reconstructed through the screening of mammals, chieflyraccoons (Procyon lotor) and Virginia opossums (Didelphis virginiana),from Arizona, Florida, Georgia, Kentucky, Louisiana, Maryland, Missouri, NorthCarolina, Oklahoma, South Carolina, Tennessee, Texas and Virginia (Bi et al.2010, Bern et al. 2011 1 , Rosypal et al. 2014 28 , Hodo et al. 2016 18 ). However, thesylvatic life cycle of T. cruzi has been documented only in southernand coastal states (Garcia et al. 2015 13 , Herrera et al. 2015 16 , Hodo et al. 2016 18 ).
T. cruzigrows chiefly via cellular fission, with occasional recombination. Yet, itsgenetic diversity across the Americas is relatively high, especially in areaswhere several competent vectors and mammalian hosts are involved (Brenièreet al. 2016). To add to this complexity, some strains show clear host preferences(Brenière et al. 2016 4 ). As a consequence, the species has been subdividedinto six universally recognized discrete typing units (DTUs), including TcIthrough TcVI, and a seventh, TcBat, that cycles through bats (Zingales et al.2009, 2012, Lima et al. 2015 20 , Hodo et al. 2016 18 ). In the United States, the strainsTcI, TcIV, and, less commonly, TcII have been documented in both mammals andtriatomines (Herrera et al. 2015 16 ). From these, TcI has been identified as thestrain inducing most of the autochthonous infections in humans in Louisianaand Texas (Dorn et al. 2007 9 , Garcia et al. 2015 13 , 2017), whereas TcIV has beendetected in raccoons in Kentucky (Bi et al. 2010 2 ) and south eastern states (Roelliget al. 2013).
In Illinois, aMidwestern state located in the north central region of the United States, theparasite has been detected in humans who contracted the infection while residingor visiting endemic areas in Latin America (Bern et al. 2011 1 ). Relative to potentialvectors, museum records account for the presence of both Triatoma lecticulariaand T. sanguisuga in the state's territory (Fracker 1913 12 , Hagerty& McPherson 1999). Although these insects are known vectors of T. cruzielsewhere in the United States (Bern et al. 2011 1 ), early screenings failed atrevealing any infected individuals (Porter 1965 24 ). To that effect, the parasitehas yet to be detected in reservoirs or vectors native to Illinois.
Given the negativeresults obtained during previous parasitological surveys in the vector (Porter1965), and the relatively greater success in detecting infection in mammals,we concentrated our efforts at screening archived tissues collected from raccoons(Procyon lotor), bobcats (Lynx rufus) and other mammals for thepresence of T. cruzi. The purpose of this screening is to document itspresence and prevalence for the first time in this state. If successful, theseresults will facilitate the determination of the northernmost distribution limitof the etiological agent of Chagas disease. To date, the pathogen is expectedto occur in Illinois and other states in the American Midwest, yet this expectationis based solely on the distribution of the vectors.
Animal collection- Sixty bobcats, 37 raccoons, one red fox (Vulpes vulpes), five nine-bandedarmadillos (Dasypus novemcinctus) and three Virginia opossums were trappedor collected as road-kill in southern and south-central Illinois within 250miles of Jackson County, Illinois (Fig. 1). Bobcatswere collected between the summer of 2003 and 2012 from localities referredelsewhere (Hiestand et al. 2014 17 ), with four additional individuals collectedas road-kills between 2013-2015 in Jackson County. Thirty-three raccoons weretrapped using baited wire-cages (38 x 38 x 76 cm) in Jackson, Madison and Williamsoncounties, in Illinois, and four in Boone County, Missouri between 2013 and 2015.The organisms were humanely euthanized as described elsewhere (Boyles &Nielsen 2017). Upon necropsy, tissue samples from spleen, liver, muscle, andheart were stored at -80ºC.
Determinationof infection with T. cruzi - DNA was extracted from the heart, skeletalmuscle, and spleen of 60 bobcats, skeletal muscle of 37 raccoons and smoothmuscle of five armadillos, three opossums and a fox using a DNeasy Blood &Tissue Kit (QIAGEN, Valencia, CA). The presence of T. cruzi was determinedvia polymerase chain reaction (PCR) using specific primers designed for theamplification of the hypervariable region of kinetoplast DNA minicircle (kDNA)and the highly repetitive genomic satellite DNA (satDNA). The set of primersS35/S36 (Sturm et al. 1989 30 ) was used to amplify a fragment of about 330bp ofthe kDNA enforcing the following thermal profile 95C/5:00 (95C/1:00, 60C/1:00,72C/1:00) x 30 with a final extension at 72C/10:00. The primers Tcz1/Tcz2 (Moseret al. 1989) targeted a 188bp fragment of the satDNA and was amplified witha thermal profile consisting of 94C/5:00 (94C/0:40, 68C/1:00, 72C/1:00) x 40with a final extension at 72C/10:00. All reactions were performed on a totalvolume of 20 µL, including the primers, DNA template and the mix of reagentsincluded in the Taq DNA Polymerase kit (QIAGEN, Valencia, California). All reagentswere mixed under a negative pressure laminar flow hood away from area of DNAextraction following strict protocols to prevent and detect contamination. Everyset of reactions included a negative control of double distilled water. DNAextracted from T. cruzi strain TcIV collected from raccoons in Kentuckywas used as positive control to test all primers (Bi et al. 2010 2 ). Ampliconswere visualized on 2% agarose gels stained with ethidium bromide.
Assignationto a DTU - The DTU or genotype of T. cruzi within the positive sampleswas determined by amplifying and sequencing two different genetic markers: theintergenic region of the spliced leader intergenic region -SL- (also known asmini-exon intergenic region), and the D7 domain of the 24S?ribosomal DNA -24S?rDNA-. The SL was amplified by means of a multiplex PCR with five differentprimers: Tc1, Tc2, Tc3, Tr, and Me (Fernandes et al. 2001 10 ). From this set, primerME binds to the most conserved region of SL whereas the rest of the primersbind at upstream regions, resulting in bands of different size that are usedto identify the DTUs. If positive, primer Tc1 would yield a fragment of about200bp and allow the determination of DTU Tc1; Tc2 would yield a fragment ofabout 250bp and it would amplify DNA of DTUs TcII, TcV, and TcVI; Tc3 wouldyield a fragment 150bps and allow the determination of DTUs TcIII and TcIV.Finally, Tr would yield a fragment of 100bp of Trypanosoma rangeli. Thesereactions were completed in a total volume of 20 µL per reaction usingthe five primers, 40 ng template DNA mixed with the reagents included in theTaq DNA Polymerase kit (QIAGEN, Valencia, California). Every set of reactionswas carried out using stringent controls and they were mixed under a hood dedicatedto this process. Thermal profile for this multiplex reaction consisted of 94C/5:00(94C/0:30, 55C/0:30, 72C/0:30) x 35 with a final extension step of 72C/7:00.
The 24S?rDNA was amplified via two reactions consisting of a first reaction targetinga 300bp fragment using primers D75-D76 (Briones et al. 1999 5 ), using a thermalprofile described elsewhere (Marcet et al. 2006 22 ). An aliquote of 1 µLof this solution was used in a subsequent heminested reaction, targeting a 145bpfragment using primers D71 and D76 (Souto et al. 1996 29 ) and the thermal profiledescribed above. Amplicons were visualized on 2% agarose gels stained with ethidiumbromide.
Positive PCR productswere cleaned using exonuclease I-shrimp phosphatase, Exo SAP-IT (GE Healthcare,Cleveland, Ohio) to remove excess of nucleotides following manufacturer recommendations.Sequencing was conducted in both directions using 0.75 µL of TerminatorBigDye 3.2 (BigDye™Chemistry Perkin-Elmer Applied Biosystems, Norwalk, Connecticut), 3.0 µL5X BigDye Buffer, 9.25 µL molecular grade water, 1.0 µL DNA template,and 1.0 µL of primer (either Me or Tc3 at 3.2 pM) for a final volume of15 µL. The thermal profile consisted of 96C/1:00 (96C/0:15, 50C/0:10,60C/4:00) x 40. Products were purified with the aid of Sephadex columns (GEHealthcare, Buckinghampshire, UK), treated with 10 µL highly deionizedformamide (Hi-Di, The Gel Company, San Francisco, CA), and direct sequencedin a 3130XL Genetic Analyzer (Applied Biosystems, Grand Island, NY). Productswere directly sequenced in an ABI 3130xl gene sequencer in the ConservationGenetics Laboratory of Southern Illinois University (Carbondale, Illinois).Resulting sequences were uploaded to universal repositories (TableI).
A matrix was assembledwith known strains available in GenBank (Fernandes et al. 1998 11 , Sturm et al.2003, Cribb et al. 2004 7 , Herrera et al. 2015 16 ). Sequences were aligned usingClustal Omega (https://www.ebi.ac.uk/Tools/msa/clustalo/).This matrix was used to calculate genetic distance between sequences generatedin this study against reference sequences of TcIV, including AY367124 (isolatedfrom a raccoon in Georgia) and AY367123 (isolated from a patient from Brazil)(Sturm et al. 2003 31 ). This calculation was made using homologous sequences andeliminating all ambiguous base pairs in PAUP* Vers. 4.0a152 (Swofford 2003 32 ).Following alignment, the matrix was trimmed to include only homologous sequences(Fig. 2). This matrix includes 23 taxa and 153nucleotides and it is available in the permanent repository OpenSIUC (http://opensiuc.lib.siu.edu/zool_data/13/).The best-fitting substitution model calculated for homologous regions was determinedusing JModelTest 2 (Darriba et al. 8 2012), which selected the Jukes-Cantor asthe best fitting model via Akaike information criterion. Posterior probabilitiesof branches were reconstructed using MrBayes v. 3.2 (Ronquist et al. 2012 27 ).The Bayesian analyses were run under the following conditions; 4 chains, 3 runs,and 10,000,000 generations. Each chain was sampled every 1,000 generations.
Bootstrap supportfor the branches was estimated enforcing the Jukes-Cantor model, with 1,000replicates in in PAUP* Vers. 4.0a152 (Swofford 2003 32 ).
Ethics -All methods were approved by the Institutional Animal Care and Use Committeeof Southern Illinois University Carbondale (Assurance number A-3078-01, protocolnumbers 11-042, 13-054 and 14-060).
Prevalence ofT. cruzi in tissue samples - Seven out of 37 (18.9%) raccoons were positivefor infection with T. cruzi, whereas no bobcat, armadillo, fox or opossumtested positive for infection. Five out of the seven infected raccoons werecollected in Jackson County, Illinois, while the remaining individuals werecollected from Williamson County, Illinois and Boone County, Missouri. All positivewere detected by amplification of a 330bp fragment of kDNA, using primers S35/S36(Fig. 3A). No amplicons resulted from the use of the primerset targeting satDNA.
Genotyping ofT. cruzi in tissue samples - The size of the amplified SL was approximately150bp (Fig. 3B), which is expected for DTU TcIII and TcIV.Further, the size of the amplified 24S?rDNA fragment following the heminested reaction is about 145bp (Fig.3C). Upon alignment and analysis of the SL sequences for phylogenetic signalwe obtained the tree shown on Fig. 4. In thistree, the sequences resulting from the screening of wildlife in Illinois andMissouri, as well as the positive controls from Kentucky, formed a clade withreference sequences for TcIV, chiefly AY367124 and AY367123.
The sequenced ampliconsisolated from raccoons in Illinois and Missouri include two polymorphisms notseen in reference AY367124. Yet, when the sequences from Illinois, Kentuckyand Missouri are pooled together, the average intraspecific genetic distanceis 0%; these pooled samples show an average distance of 0% when compared toreference AY367124 (isolated from a raccoon in Georgia, USA). The pooled sequenceshad an average genetic distance of 5% when compared against the reference sequenceAY367123, also considered TcIV and isolated from a human being in Brazil. Furthermore,the pooled sequences average 10% genetic distance when compared against theCL-Brener isolate identified as the TcVI reference. Finally, the pooled sequencesaverage a genetic distance of 19% when compared against the reference for strainTcIII (AF050521), isolated from an armadillo in the Brazilian Amazon (TableII).
We document forthe first time the presence of T. cruzi in wild raccoons fromIllinois. From the seven positive samples, it was possible to sequence and determinesix as DTU TcIV. Thus, our results demonstrate that T. cruzi is presentin Illinois as a fairly homogeneous strain, and it cycles in raccoons. Thisfinding is consistent with the available genotypic characterizations of T.cruzi in the United States, which indicates that all DTUs with the exceptionof TcIII, are present in the country; with TcIV predominantly causing infectionsin raccoons (Bern et al. 2011 1 , Roellig et al. 2013 26 , Garcia et al. 2017 14 ).
T. cruziwas detected in two of 62 bobcats sampled in Georgia (Brown et al. 2010 6 ), andwas not detected in any bobcats in our study. Contrastingly, infections in raccoonsappear to be common across the southern half of the United States. The differencein the prevalence of T. cruzi in bobcats compared to raccoons, may bea result of the diet preferences of the former. Bobcats only opportunisticallyconsume insects, whereas a raccoon's diet may consist of up to 40% insects,depending on the season (Llewellyn & Uhler 1952). The consumption of insectsmay increase the risk of contracting the pathogen, which was shown experimentally.Exposed raccoons contracted the pathogen after being fed with infected triatominebugs; in contrast, raccoons did not acquire the infection upon consumption ofmeat carrying the parasite (Roellig et al. 2009 25 ). Thus the oral transmissionmay only occur after raccoons consume the metacyclic trypomastigote stage ofT. cruzi when it is present in a triatomine bug or its feces. As bobcatsdo not usually consume insects, their exposure via the oral route of transmissionwould be greatly reduced. However, it must be considered that the detectionof T. cruzi in bobcats from Georgia was achieved by antibody testing,which may be more sensitive to the detection of current or past exposure tothe pathogen.
Infections in opossumsand armadillos in Illinois are yet to be documented. These mammals are frequentlyinfected with the pathogen in southern localities; perhaps, a greater samplesize may help detecting the prevalence of this parasite in these mammals. Itshould be considered that the prevalence for T. cruzi in surveys of wildcaught opossums varies from 8 to 60% (Bern et al. 2011 1 ).
In our attemptsto reconstruct the phylogeny of the strains in the United States, we discoveredthat several sequences used in the phylogenetic reconstruction of strains basedon SL are not homologous. All of these sequences do belong to the SL, yet theydo not amplify the same region of the gene. In some cases, the resulting sequencesappear to go either upstream or downstream relative to the region between 12,300to 12,400nt of reference T. cruzi 6949 strain CL-Brener (Fig.2). However, the sequences KM376441 and KM376442, identified as TcIV elsewhereappear to amplify a region not homologous with the rest of the sequences usedin this and other reconstructions. Nevertheless, although their use in phylogeneticreconstruction may not be recommended, the set of primers used in the multiplexreaction may remain an option to identify the involved strain in archived hosttissues, (i. e, DNA from parasites not grown in culture or isolates), providedtheir sequences are compared against known references featuring the entire SL,and a second set of primers targeting a different region such as 24S?rDNA is used (Zingales et al. 2009 33 , 2012).
Further study mayinclude screening triatomine bugs within Midwestern states for the presenceof this parasite, which would bring conclusive evidence of occurrence of thesylvatic cycle in the region. It would also help in better determining the riskof vector-borne transmission of T. cruzi in Illinois and other Midwesternstates. As noted by Roellig et al. (2013), there is considerable genetic variabilityin TcIV across southeastern states, which suggest that this strain has beenestablished in the region for a long time. Thus, it becomes important to comparethe genetic diversity of those parasites against those in Midwestern states,which may constitute the northern most limits of distribution of natural populationsof the pathogen. Any signals of a recent population expansion would be evidentin the form of low genetic diversity, which would be expected near or in theparasite's distribution limits. Determining the diversity, prevalence and geographicdistribution of T. cruzi in the United States is key to determine areasof risks of vector-borne transmission within the US. Finally, it would alsobe advisable to expand the sample size of armadillos screened for the presenceof T. cruzi and use more sensitive methods, including quantitative PCR(qPCR), hemoculture and antibody-based assays to identify population prevalencewith a greater accuracy. Finally, nine-banded armadillos, are becoming an ubiquitouspresence in Illinois (Hofmann 2009 19 ); the increased abundance of this insectivoremay afford an opportunity to track changes in dynamics and distribution of thisparasite in the country as well as the expansion of a strain associated withthem into new territories.
To Dr Rodrigo Carramiñanaand Dr Ed Heist (Southern Illinois University) for providing support for professionaldevelopment to CV and accessing and using of the sequencing facilities of theConservation Genetics Laboratory at SIU; and Dr Kurt Neubig (Southern IllinoisUniversity) who facilitated reagents for completing part of the sequencing.
FAJ and CV conceivedand designed the study; EZ, EB, CKN and CD completed field work and archivedtissues; CD isolated and culture Trypanosoma cruzi from raccoons collectedin Kentucky; CV, EZ and FAJ performed experimental work; CV and FAJ analysedsequences; CV and FAJ prepared a draft of the manuscript. All authors read andapproved the final version.