Differentiation between Trypanosoma cruzi and Trypanosoma rangeli using heat-shock protein 70 polymorphisms

Authors


Abstract

Objective

Differential diagnosis of infection with Trypanosoma cruzi or T. rangeli is relevant for epidemiological studies and clinical practice as both species infect humans, but only T. cruzi causes Chagas' disease. Their common antigen determinants complicate the distinction between both species, while current PCR assays used for differentiation show some drawbacks. We developed and validated a generic PCR discriminating the species by restriction fragment length polymorphism (RFLP) analysis and a duplex PCR specifically amplifying a differently sized fragment of both species.

Methods

The assays are based upon a partial region of the heat-shock protein 70 gene (hsp70). The analytical sensitivity and specificity were determined for both PCRs. The assays were analytically evaluated on a panel of six T. cruzi, one T. cruzi marinkellei and four T. rangeli strains, various other infectious pathogens, a panel of spiked samples of T. cruzi, and artificially mixed infections of T. cruzi and T. rangeli. Finally, the tools were applied on 36 additional isolates of Trypanosoma species.

Results

The detection limit of the PCRs was between 0.05 and 0.5 parasite genomes, and 1–10 parasites spiked in 200 μl blood. In artificial mixtures, PCR–RFLP picked up both species in ratios up to 102 and duplex PCR up to 104. In the 36 isolates tested, both single and mixed infections were identified. All assays were shown to be specific.

Conclusion

Our PCRs show high potential for the differential diagnosis of T. cruzi and T. rangeli, which in view of their sensitivity can aid in the confirmation of infection with these parasites in vectors, reservoirs and clinical samples.

Abstract

Objectif

Le diagnostic différentiel de l'infection par Trypanosoma cruzi ou T. rangeli est nécessaire pour les études épidémiologiques et la pratique clinique car les deux espèces infectent les humains, mais seul T. cruzi provoque la maladie de Chagas. Leurs déterminants antigéniques communs compliquent la distinction entre les deux espèces, alors que les tests actuels de PCR, utilisés pour la différenciation montrent quelques inconvénients. Nous avons développé et validé une PCR générique permettant de discriminer les espèces par le polymorphisme de la taille des fragments de restriction (RFLP) et par un PCR duplex amplifiant spécifiquement un fragment de taille différente pour les deux espèces.

Méthodes

Les tests sont basés sur une région partielle du gène de la protéine heat-choc 70 (hsp70). La sensibilité et la spécificité analytique ont été déterminées pour les deux PCR. Les tests ont été évalués analytiquement sur un ensemble de 6 souches de T. cruzi, 1 souche de T. cruzi marinkellei et 4 souches de T. rangeli, ainsi que divers autres agents pathogènes infectieux, une série d’échantillons enrichis de T. cruzi et de mélanges artificiels d'infections par T. cruzi et T. rangeli. Enfin, les outils ont été appliqués sur 36 isolats supplémentaires d'espèces Trypanosoma.

Résultats

La limite de détection de la PCR était comprise entre 0,05 et 0,5 génomes de parasites, et de 1 à 10 parasites enrichis dans 200 μl de sang. Dans les mélanges artificiels, la PCR-RFLP a détecté les deux espèces dans des proportions allant jusqu’à 102 et la PCR duplex, dans des proportions allant jusqu’à 104. Sur les 36 isolats testés, les infections uniques et mixtes ont été identifiées. Tous les tests ont été trouvés spécifiques.

Conclusion

Nos PCR montrent un fort potentiel pour le diagnostic différentiel de T. cruzi et T. rangeli, qui en raison de leur sensibilité pourraient aider à la confirmation de l'infection par ces parasites chez les vecteurs, dans les réservoirs et dans les échantillons cliniques.

Abstract

Objetivo

El diagnóstico diferencial de la infección por Trypanosoma cruzi o T. rangeli es relevante para estudios epidemiológicos y la práctica clínica, puesto que ambas especies infectan a los humanos pero solo T. cruzi causa la enfermedad de Chagas. Sus determinantes antigénicos comunes complican la distinción entre ambas especies, mientras que los ensayos de PCR actualmente utilizados para diferenciarlos presentan problemas. Hemos desarrollado y validado una PCR genérica que discrimina entre las especies mediante un análisis de polimorfismos de longitud de fragmentos de restricción (RFLP) y una PCR-Dúplex que amplifica específicamente fragmentos de tamaños diferenciados en ambas especies.

Métodos

Las pruebas están basadas en una región parcial del gen de la proteína de choque térmico 70 (hsp70). Para ambas PCRs se determinaron la sensibilidad y especificidad analítica. Las pruebas fueron evaluadas de forma analítica en un panel de 6 cepas de T. cruzi, 1 de T. cruzi marinkellei y 4 de T. rangeli, otros patógenos infecciosos, un panel de muestras a las que se añadió T. cruzi, y mezclas artificiales de T. cruzi y T. rangeli. Finalmente las herramientas se aplicaron en 36 aislados adicionales de especies de Trypanosoma.

Resultado

El límite de detección de las PCRs estaba entre 0.05 y 0.5 genomas de parásito y 1 a 10 parásitos en 200 μl de sangre. En mezclas artificiales, la PCR-RFLP detectó ambas especies en una proporción de hasta 102 y la PCR-Duplex de hasta 104. En los 36 aislados evaluados se identificaron tanto las infecciones únicas como las mixtas. Todas las pruebas eran específicas.

Conclusión

Nuestras PCRs muestran un gran potencial para el diagnóstico diferencial de T. cruzi y T. rangeli, lo cual dado su nivel de sensibilidad podría ayudar a la confirmación de la infección con estos parásitos en vectores, reservorios y muestras clínicas.

Introduction

Chagas' disease, caused by the protozoan parasite Trypanosoma cruzi, affects 21 countries in Latin America with a burden of approximately 0.43 million disability-adjusted life years (DALYs) (World Health Organization 2010). An estimated eight to nine million people are infected and over 100 million are at risk. Yearly, there are close to 41 200 new reported cases of infection by T. cruzi and 12 500 deaths (World Health Organization 2007; Hotez et al. 2008; Schmunis & Yadon 2010). The illness has a variable clinical course that may include asymptomatic infection, acute disease or a chronic condition.

Trypanosoma cruzi is a heterogeneous taxon with multiple hosts, vectors and routes of infection and is currently classified into six discrete typing units (DTU): Tc I–VI. These DTUs are differently distributed in the endemic regions and transmission cycles and probably impact the clinical manifestation and severity of the disease differently (Zingales et al. 2012). Trypanosoma rangeli is the second most common American trypanosome that infects humans, but it is a non-pathogenic parasite. T. rangeli presents extensive genetic variability, and kDNA minicircle characteristics identified two groups named KP1(+) and KP1(−), according to the presence or absence of KP1 minicircles, respectively (Vallejo et al. 2002). This parasite shares morphological similarities and antigenic determinants with T. cruzi, making differential microscopic and serological diagnosis challenging. Both parasites occur in the same geographical locations and share insect vectors (Vallejo et al. 1999; Morales et al. 2002). Mixed infections with both parasite species have been reported in humans (Guhl et al. 1987; Coura et al. 1996; Chiurillo et al. 2003), vectors and reservoir animals (Vargas et al. 2000; Ramirez et al. 2002; Dias et al. 2007; Pavia et al. 2007; Thekisoe et al. 2010; Saldaña et al. 2012), necessitating accurate differential diagnosis (Thekisoe et al. 2010). The lack of an adequate specific diagnostic procedure and the absence of clinical manifestations have been responsible for the underestimation of human T. rangeli incidence (Coura et al. 1996; D' Alessandro & Saravia 1999; Vargas et al. 2000).

Several techniques such as resistance to complement lysis and reaction with monoclonal antibodies (Acosta et al. 1991), lectin agglutination lysis (de Miranda Santos & Pereira 1984), nuclear or kinetoplast DNA probes (Greig et al. 1990; Macedo et al. 1993; Vallejo et al. 1994), and the restriction patterns of kDNA (Vallejo et al. 1994) have been used for distinguishing both trypanosome species, but none is completely satisfactory when used alone, and all require propagation of the parasite (Morales et al. 2002). To solve these problems, different PCR methods and targets have been described for differentiation: kinetoplast DNA (Sturm et al. 1989; Dorn et al. 1999; Vallejo et al. 1999); the mini-exon gene (Murthy et al. 1992; Fernandes et al. 2001); a 195-bp nuclear repeated satellite DNA (Breniere et al. 1993); the 24S large subunit ribosomal RNA (Souto et al. 1999); a gene coding for a flagellar protein (Silber et al. 1997); cysteine proteinase (Tanaka 1997), DNA repetitive element P542 (Vargas et al. 2000); telomeric sequences (Chiurillo et al. 2003); a short interspersed repetitive element (SIRE) sequence inserted into the histone H2A gene; and the small nucleolar RNA-C11 gene (Pavia et al. 2007). However, these PCR assays show some disadvantages such as the amplification of similar bands both in T. cruzi and T. rangeli (Souto et al. 1999), the amplification of bands of different size related with the amplification of polymorphic regions (Murthy et al. 1992; Grisard et al. 1999; Fernandes et al. 2001) and a bias towards T. cruzi in the case of mixed T. cruzi and T. rangeli infections (Sturm et al. 1989; Dorn et al. 1999; Vallejo et al. 1999; Vargas et al. 2000). Therefore, it is necessary to develop new techniques as better options to specifically identify each of the parasite species in single and mixed infections.

The cytoplasmic 70-kDa heat-shock protein genes (hsp70), which are organised as a tandem array in a head-to-tail manner (Requena et al. 1988), have been exploited for Leishmania species identification in the Old and New World using PCR, followed by restriction fragment length polymorphism (RFLP) analysis, showing high sensitivity and specificity in clinical samples (Garcia et al. 2007; Fraga et al. 2012; Montalvo et al. 2012; Veland et al. 2012). Recently, Cuervo et al. (2013) proposed a PCR–RFLP based on this gene, using the SphI enzyme, for differentiating the two groups of T. rangeli KP1(+) and KP1(−).The goal of this study was to exploit the cytoplasmic hsp70 gene for differential diagnosis of T. cruzi and T. rangeli through the design of a PCR–RFLP and species-specific PCRs.

Materials and methods

Strains and isolates

Table 1 lists the species and geographical origin of the strains and isolates used in different steps of our study and representative of T. cruzi, T. cruzi marinkellei and T. rangeli. As DNA was obtained from different laboratories and collections, the reference species identification and genotype was based on various data such as multilocus sequence typing (MLST), multilocus enzyme electrophoresis (MLEE), multilocus microsatellite typing (MLMT), PCR and PCR–RFLP techniques. All DNAs were isolated from parasite cultures, and DNA was obtained from the different institutes acknowledged at the end of this article.

Table 1. Strains and isolates used in the study
SpeciesStrain nameCountryGenotypeaAccession numberEvaluation panelb (Figure 3)Spiked samplesb (Figure 5)Application panelb (Figure 6)
  1. a

    Trypanosoma cruzi genotype listed as DTU; T. rangeli listed as KP1 present or absent; nk, not known.

  2. b

    Isolates tested are indicated with v.

  3. c

    Sequences from strains retrieved from GenBank, and the remaining ones were determined in this study.

  4. d

    Accession numbers starting with XM are derived from a contemporary annotation of full-genome sequences and were also retrieved from the GenBank database.

Trypanosoma cruzi (= 27)SP104ChileI KC959988 v  
OPS21VenezuelaI KC959989   v
P209BoliviaI   v
CutiaBrazilI   v
X10/4BrazilI  v 
MASBrazilII KC959990   v
EsmBrazilII KC959991 v  
Tu18BoliviaII   v
IVVChileII   v
CBBChileII   v
PERUPeruIIM26595c   
M5631BrazilIII KC959992   v
M6241BrazilIII KC959993 vv 
CM17ColombiaIII KC959994   v
X109/2ParaguayIII   v
92122102RUSAIV KC959995   v
CanIIIBrazilIV KC959996 v  
TheisUSAIV   v
SO3BoliviaV KC959997 v  
MNChileV KC959998   v
Bug2148BrazilV   v
SC43BoliviaV   v
CL BrenerBrazilVIKC959999/XM_812644cd/XM_812645cd  v
TulaChileVI  vv
P63ParaguayVI KC960000 v  
YBrazilnk FJ222459 c   v
BRAZILBrazilnkX67716c   
       
Trypanosoma rangeli (= 25)RPLA4Venezuelank KC960001 v  
CAN ACVenezuelank KC960002 v  
LEM2946French Guyanank KC960003 v  
LEM2963French Guyanank KC960004 v  
B28V1A06P2C3Venezuelank KC960005   v
B28P1A06P1C1Venezuelank KC960006   v
LEM2947French Guyanank KC960007   v
R1271Colombiank KC960008    
Sjm30Boliviank KC960009    
TCI2307Ecuadornk   v
TET2408Ecuadornk   v
TET2414Ecuadornk   v
TET2417Ecuadornk   v
TPY2422Ecuadornk   v
PerijáColombiank   v
226Colombiank KC960010   v
H14HondurasKP1+EF108422c/ EF442767c/ EF442768c   
TreColombiaKP1−EF108420c/EF108421c/EF108419c   
TBJ2273Ecuadornk   v
TET2408Ecuadornk   v
TLG2355Ecuadornk   v
TRC2377Ecuadornk   v
R1625El SalvadorKP1+   v
SC58BrazilKP1−   v
TLG2343_cl3Ecuadornk   v
T. cruzi marikellei (= 2)M1909Venezuelank KC960011 v  
B3Brazilnk   v

To test the analytical specificity of the PCRs, DNAs from clinical samples or cultures from various microorganisms and viruses were used: Trypanosoma brucei brucei, T. brucei gambiense, T. brucei rhodesiense, T. congolense, T. vivax, T. equiperdum, T. theileri, T. evansi, Leishmania (L.) infantum (chagasi), L. (L.) mexicana, L.(V.) braziliensis, L. (V.) naiffi, L. (V.) peruviana, L. (V.) guyanensis, L. (V.) lainsoni, Plasmodium falciparum, Schistosoma mansoni, Candida albicans, Candida parasilopsis, Cryptococcus neoformans, Haemophilus influenzae, Streptococcus pneumonia, Enterococcus, Mycobacterium tuberculosis, Mycobacterium habana, Staphylococcus epidermidis, Staphylococcus aureus, Pseudomonas aeroginosa, Escherichia coli, Neisseria meningitides, Epstein–Bar virus, herpes zoster virus, herpes simplex virus, and cytomegalovirus. These were obtained from various sources acknowledged at the end of the manuscript.

Trypanosoma cruzi epimastigotes of three strains (X10/4, Tula and M6241) of different DTUs were counted, and a 10-fold dilution series was prepared in human blood from a healthy individual, ranging from 106 to 1 parasite in 200 μl blood. Non-spiked blood was used as a negative control. Extraction of DNA was performed using the High Pure PCR Product Purification Kit (Roche Applied Science, Mannheim, Germany), whereby the DNA was eluted in 200 μl.

Primer design and in silico analysis of hsp70 coding sequences

Partial sequences of the hsp70 gene (Figure 1, Garcia et al. 2007) from different isolates of T. cruzi, T. rangeli and T. cruzi marinkellei (Table 1) were generated using the dideoxy nucleotide chemistry from the ABI PRISM BigDye™ Terminator cycle sequencing kit (PerkinElmer, Foster City, CA, USA). They were analysed on an ABI 3730 automated sequencer (PerkinElmer). The sequences were aligned with those previously published in GenBank (Table 1) using the software package MEGA5 (Molecular Evolutionary Genetic Analysis, version 5.05, Tamura et al. 2011, www.megasoftware.net). A phylogenetic network was constructed with the SplitsTree4 software (Huson 1998; Huson & Bryant 2006), using the Kimura 2-parameter model (Kimura 1980) and the neighbour-net algorithm (Bryant & Moulton 2004).

Figure 1.

Schematic representation of the Trypanosoma cruzi (GenBank accession number XM-812645) and T. rangeli (GenBank accession number EF108422) hsp70 genes encoding cytoplasmic HSP70, showing the annealing sites of the oligonucleotides in Table 2. White boxes represent the coding region of hsp70 gene. The sequenced fragment of 1380 nucleotides is shown on the top. The numbers below represent the nucleotide number of the coding region relative to the start codon. The coding regions are aligned, except for the 3′ end where both species have a different size. The PCR products (Table 2) are represented by grey-shaded boxes, indicating the size of the amplicon, for which no intraspecific variation was detected. The primers used in each PCR are indicated by horizontal arrows (not to scale) with the name and 5′–3′ orientation. Generic primers (for both species) are indicated with black horizontal arrows, and specific primers are indicated with grey horizontal arrows. The recognition sites of restriction enzymes used in RFLP analysis are indicated by vertical lines inside the white boxes: Sau3AI (discontinuous) and HaeIII (continuous).

One PCR primer set in a conserved region was selected for amplification of both species and subsequent differentiation through RFLP. Also PCR primer sets specific for T. cruzi and T. rangeli were selected for a duplex PCR (Figure 1, Table 2). In silico RFLP analysis and prediction of the resulting fragments was performed with the software packages MEGA5 and DNAMAN, version 4.02 (Lynnon Biosoft, Quebec, Canada), as well as with the online tools NEBcutter (tools.neb.com/NEBcutter2) and DistinctiEnz (www.bioinformatics.org/~docreza/cgi-bin/restriction/DistinctiEnz.pl).

Table 2. PCR primers and conditions
PCRPrimersaAnnealingPrimer Sequence 5′–3′Amplicon sizeReagent concentrationsbAmplification conditions
  1. a

    F = sense primer; R = antisense primer.

  2. b

    Primer concentrations are listed per primer; the total MgCl2 concentration in the reaction mix is given (1.5 mm of which is included with the PCR buffer, as explained in 'Methods').

  3. c

    The annealing position of the primers is given relative to GenBank accession number XM_812645 (T. cruzi CL Brener).

  4. d

    The annealing position of the primers is given relative to GenBank accession number EF108422 (T. rangeli H14).

hsp70THSP70T_F703

1160–1180c

1157–1177d

GCAAGAGCAAGCAGACGGAGG470 bp

0.8 μm primers

3.5 mm MgCl2

1 cycle 95 °C 5 min

35 cycles 94 °C 40 s/65 °C 1 min/72 °C 1 min

1 cycle 72 °C 10 min

HSP70T_R1172

1629–1609c

1626–1606d

CTTTGCGTCAATGCGCTCGCG
hsp70TcHSP70T_F10481505–1528cGCAACCAGATTGTCATCACGAACG95 bp

0.8 μm primers

2.5 mm MgCl2

1 cycle 95 °C 5 min

35 cycles 94 °C 40 s/66 °C 1 min/72 °C 1 min

1 cycle 72 °C 10 min

HSP70T_R1142Tc1599–1578cGTCCTGCGMCTCGTACTTGGCA
hsp70TrHSP70T_F10481502–1525dGCAACCAGATTGTCATCACGAACG202 bp

1.0 μm primers

3.0 mm MgCl2

HSP70T_R1249Tr1703–1683dGCCTCCTCAATCTTGCCGCTC
hsp70TcTrHSP70T_F1048

1505–1528c

1502–1525d

GCAACCAGATTGTCATCACGAACG

95 bp

202 bp

1.0 μm primers

3.0 mm MgCl2

HSP70T_R1142Tc1599–1578cGTCCTGCGMCTCGTACTTGGCA
HSP70T_R1249Tr1703–1683dGCCTCCTCAATCTTGCCGCTC

PCRs

Four PCR-hsp70 protocols were performed (Figure 1, Table 2). The generic PCR-hsp70T was performed in 50 μl total volume, while the specific PCR-hsp70Tc, PCR-hsp70Tr and duplex PCR-hsp70TcTr were performed in 25 μl total volume, each containing 1× Q solution and 1× CoralLoad PCR buffer including 1.5 mm MgCl2, 200 μm of each deoxynucleoside triphosphate and 1 U HotStar Taq Plus DNA Polymerase (Qiagen, Hilden, Germany). The additionally added MgCl2 and primers varied as detailed in Table 2.

Various amounts of DNA were added to the PCRs, as mentioned with the respective results, and DNA concentrations were measured spectrophotometrically using the Nanodrop (ND1000, Thermo Scientific, Wilmington, DE, USA). Negative no-template controls were always included, along with a positive control consisting of 1 ng DNA from T. cruzi (SP104 strain) and T. rangeli (CAN AC strain). The amplification conditions for all PCRs are shown in Table 2. Thermal cycling was performed either in a PTC-150 (MJ Research, Waltham, MA, USA), in a MyCycler™ (Bio-Rad, Foster City, CA, USA) or in an iCycler (Bio-Rad). Analysis on a 2% agarose gel was used to verify the amplified product size and to check for non-specific amplification.

RFLP analysis

Restriction fragment length polymorphism analysis was carried out on amplicons from PCR-hsp70T. Digests were performed in a total of 10 μl containing 5 μl non-purified amplicon solution in 1× optimal buffer provided by the manufacturer, using 2 U HaeIII or 2 U Sau3AI (MBI Fermentas, St. Leon-Rot, Germany). Reaction mixtures were incubated 3 h at 37 °C. Subsequently, restriction patterns were analysed by electrophoresis on a 3% small fragment agarose gel (Gentaur, Brussels, Belgium) running at 3.5 V/cm for 3 h.

Results

A single fragment covering 1380 bp of the hsp70 coding region (Figure 1) was successfully PCR-amplified and sequenced for 24 strains of T. cruzi, T. cruzi marinkellei and T. rangeli. Table 1 lists the accession numbers of the newly determined sequences and 11 additional ones from GenBank. Figure 2 depicts the relationships between these sequences, which form a general reference framework for finding genetic polymorphisms to differentiate T. cruzi from T. rangeli. Figure 1 shows the three PCR fragments for which primers were developed: PCR-hsp70T (generic PCR for RFLP analysis), PCR-hsp70Tc and PCR-hsp70Tr (species-specific PCRs). The latter two were combined in a duplex assay PCR-hsp70TcTr. Primer HSP70T_R1142Tc (T. cruzi specific) has 1–2 mismatches with T. rangeli at the 3′ terminus, while primer HSP70T_R1249Tr (T. rangeli specific) has four mismatches with T. cruzi.

Figure 2.

Complete phylogenetic network of the Trypanosoma hsp70 sequences in Table 1. Bootstrap support of the branches was inferred from 2000 replicates and is shown in percentages only for the branches leading to the three main groups. The scale is shown in the top left corner, in substitutions per nucleotide.

After determining the optimal conditions of each PCR (data not shown), the analytical specificity and sensitivity were tested in a panel of 11 strains representing the studied Trypanosoma species (Table 1 and Figure 3). For all T. cruzi, T. cruzi marinkellei and T. rangeli species, the amplicon size of PCR-hsp70T agreed with the in silico analysis (470 bp, Table 2, Figure 3a). The PCR was able to detect between 10 fg (0.05 parasite genomes) and 100 fg (0.5 parasite genomes, the amount used in Figure 3) of DNA. Nucleotide polymorphisms affecting restriction endonuclease recognition sites for differentiation between T. cruzi and T. rangeli were identified in silico. The enzyme Sau3AI produced restriction fragments of 340/91/39 bp for T. cruzi and T. cruzi marinkellei, and 222/118/109/21 bp for T. rangeli, as confirmed in Figure 3b. No intraspecies diversity was observed in the Sau3AI restriction digests. Restriction endonuclease HaeIII produced fragments of 354/94/22 bp for T. cruzi and T. cruzi marinkellei, and 307/94/47/22 or 307/69/50/44 bp for T. rangeli, as confirmed in Figure 3c. No intraspecies diversity was observed at the level of HaeIII restriction digests among T. cruzi. Both enzymes thus allow differentiation between the two species, from as little as 100 fg parasite DNA.

Figure 3.

PCR-hsp70T, RFLP results and PCR-hsp70TcTr evaluated in a panel of Trypanosoma cruzi, T. cruzi marinkellei (Tcm) and T. rangeli (Table 1). The species, strain name and T. cruzi discrete typing unit (DTU) of each strain are shown on the top. On the right, RFLP patterns are depicted for T. vivax (Tvi), T. congolense (Tco), Leishmania (Viannia) lainsoni (Lla) and L. (V.) guyanensis (Lgu), whereby these gels are aligned with the gels on the left, and only the region containing the T. cruzi and T. rangeli fragments is shown. All T. cruzi, T. cruzi marinkellei and T. rangeli products were amplified from 100 fg DNA, while the other products were amplified from 10 ng DNA. (a) Agarose gel analysis (2%) of PCR-hsp70T. (b) PCR-RFLP patterns on 3% agarose gel obtained with the restriction endonuclease Sau3AI. (c) PCR–RFLP patterns on 3% agarose gel obtained with the restriction endonuclease HaeIII. (d) Agarose gel analysis (2%) of PCR-hsp70TcTr. MM: Molecular weight marker GeneRuler™ 100 bp DNA ladder (MBI Fermentas), with band sizes indicated on the left in base pairs. Size indication of the PCR and RFLP fragments are depicted on the right of the gels in base pairs.

PCR-hsp70Tc was designed to specifically amplify a 95-bp fragment from the T. cruzi genome. This fragment was obtained from all T. cruzi of the Figure 3 panel, using 10–100 fg DNA, while none of the T. rangeli amplified even from 10 ng DNA. Conversely, the PCR-hsp70Tr 202-bp fragment was only amplified from 10 to 100 fg of T. rangeli DNA, but not from up to 10 ng T. cruzi DNA (results not shown). After establishing the optimal conditions for both PCRs, the three primers were combined in the duplex PCR-hsp70TcTr. Figure 3d shows the successful amplification of the 95- and 202-bp fragments from 100 fg DNA in the entire panel and from T. cruzi and T. rangeli, respectively.

With regard to the analytical specificity, PCR-hsp70T amplified the fragment of 470 bp in other Trypanosoma species (T. congolense and T. vivax), Leishmania (Viannia) guyanensis and L. (V.) lainsoni, from 10 ng genomic DNA. In T. congolense, an additional fragment of 1100 bp was obtained. In each of these cases, the amplicons could be distinguished from T. cruzi and T. rangeli DNA by RFLP analysis with both enzymes (Figure 3b,c). No PCR-hsp70T amplification was observed from any of the other DNAs from Leishmania, Trypanosoma, other parasites, bacteria or viruses mentioned in the 'Materials and methods' section. As for the duplex PCR-hsp70TcTr, this proved specific as none of the DNAs of the specificity panel were amplified.

To mimic mixed infections, several ratios of DNA quantities from both species were tested (Figure 4). The fragment of 470 bp was amplified from each mixture by PCR-hsp70T (Figure 4a). Agarose gel analysis of the digested PCR products showed that T. cruzi and T. rangeli could be detected in a ratio of up to 1:100 in the presence of one another (Figure 4b,c), even though the interpretation of the Sau3AI patterns is hampered by the presence of spurious bands at the extreme ratios shown in the left- and rightmost lanes of the gel. PCR-hsp70TcTr was able to amplify ratios of up to 1:10 000 (T. cruzi:T. rangeli) and 1:100 (T. rangeli:T. cruzi), respectively (Figure 4d).

Figure 4.

Evaluation of artificially mixed ratios of Trypanosoma cruzi (Esm strain) and T. rangeli (CAN AC strain) (Table 1). The amounts of each DNA added in the PCRs are shown on the top, Tc = T. cruzi; Tr = T. rangeli. a–d as in Figure 3. MM: Molecular weight marker GeneRuler™ 100 bp DNA ladder (MBI Fermentas), and band sizes are indicated on the left in base pairs. Size indication of the PCR and RFLP fragments is depicted on the right of each gel in base pairs.

The detection limit of PCR-hsp70T and PCR-hsp70TcTr was evaluated on blood samples spiked with decreasing numbers of living T. cruzi epimastigotes (Figure 5). The lower detection limit of PCR-hsp70T in three T. cruzi strains representing different DTUs was 1–10 parasites per 200 μl blood (Figure 5a). The digestion patterns could be interpreted down to 10 parasites per 200 μl, as the amplicons from lower amounts were too faint for reliable analysis (data not shown). The detection limit of PCR-hsp70TcTr was one T. cruzi parasite per 200 μl blood (Figure 5b).

Figure 5.

Sensitivity of PCR-hsp70T (a) and PCR-hsp70TcTr (b) evaluated in three panels of spiked samples of Trypanosoma cruzi on a 2% agarose gel. The T. cruzi strains and discrete typing unit (DTU) are indicated on the top (Table 1). The parasite quantity spiked in 200 μl of blood is shown on the top of each lane, whereby some amounts were tested in triplicate. MM: Molecular weight marker GeneRuler™ 100 bp DNA ladder (MBI Fermentas), and band sizes are indicated on the left in base pairs.

Finally, PCR-hsp70T and PCR-hsp70TcTr were applied on 10 ng DNA from 36 isolates from which the species is known (Table 1), and Figure 6 shows a selection of results. Of the 17 T. rangeli strains evaluated, only one strain (LEM2947) showed the HaeIII RFLP pattern of 307/69/50/44 bp, and the rest produced fragments of 307/94/47/22 bp. Both RFLP and PCR-hsp70TcTr identified the correct species in each case, except for two isolates of T. rangeli: TBJ2273 and TRC2377. In these, all methods identified DNA of both species, and hence, these are by our method scored as a mixed T. cruzi–T. rangeli infection.

Figure 6.

Application of PCR-hsp70T and its RFLP analysis, and PCR-hsp70TcTr on DNA from cultured parasite isolates. The isolate names (Table 1) are shown on the top. (a–d) as in Figure 3. MM: Molecular weight marker TrackIt™ 100 bp ladder (Invitrogen, USA), and band sizes are indicated on the left in base pairs. Size indication of the PCR and RFLP fragments is depicted on the right of each gel in base pairs. Tc = T. cruzi, Tr = T. rangeli.

Discussion

Precise diagnosis of infection with T. cruzi or T. rangeli in naturally infected vectors, reservoirs and humans is relevant for epidemiological studies. Also, considering that inaccurate serological tests can lead to misdiagnosis, unnecessary chemotherapy (frequently accompanied by side effects) and psychological trauma to the patients and their families (Chiurillo et al. 2003), we sought to develop a molecular test for detection and differentiation between these two species.

Previous PCR studies describing the detection and differentiation between T. cruzi and T. rangeli were designed either with one set of primers amplifying a differently sized fragment for each species (Murthy et al. 1992; Silber et al. 1997; Vallejo et al. 1999; Fernandes et al. 2001), or with species-specific primers (Vargas et al. 2000; Pavia et al. 2007; Ortiz et al. 2009), or as a duplex PCR (Souto et al. 1999; Chiurillo et al. 2003). The majority of these studies evaluated the analytical sensitivity using only one strain per species, and the analytical specificity or parasite mixtures were not tested.

We successfully applied PCR-hsp70T followed by digestion with Sau3AI and HaeIII, and PCR-hsp70TcTr to differentiate T. cruzi from T. rangeli. The analytical sensitivity obtained for both PCRs was similar to other PCR assays using as targets kDNA, flagellar protein F29, repetitive DNA element P542, H2A, telomeric sequences and cathepsin L-like proteases (Silber et al. 1997; Vallejo et al. 1999; Vargas et al. 2000; Chiurillo et al. 2003; Pavia et al. 2007; Ortiz et al. 2009). Our PCRs showed higher analytical sensitivity than reported PCRs based on nuclear repeated satellite DNA, the small nuclear RNA C11 gene and the 24S large subunit ribosomal RNA (Breniere et al. 1993; Souto et al. 1999; Pavia et al. 2007), and the specific detection of T. cruzi and T. rangeli based on the 18S ribosomal RNA (rRNA) and the small nucleolar RNA (snoRNA) genes using loop-mediated isothermal amplification (LAMP) (Thekisoe et al. 2010). Only Tanaka (1997) described the use of RFLP for the differentiation between T. cruzi and T. rangeli, based on cysteine proteinase. The detection limit corresponded to a single parasite cell, but mixtures and intraspecies variability were not evaluated. Our RFLPs were applicable on artificially mixed parasite DNAs and on all isolates tested despite their various geographical origins, two parameters often not studied in the above-mentioned reports.

Mixed infections with T. cruzi and T. rangeli have been reported in humans, vectors and reservoir animals (Guhl et al. 1987; Coura et al. 1996; Vargas et al. 2000; Ramirez et al. 2002; Chiurillo et al. 2003; Dias et al. 2007; Pavia et al. 2007; Thekisoe et al. 2010; Saldaña et al. 2012). Two isolates were detected by our methods as a mixed T. cruzi–T. rangeli infection. They were obtained from Rhodnius ecuadoriensis collected in Ecuador and previously typed as T. rangeli based on minicircle kDNA with S35/S36 primers (PCR-kDNA) (Vallejo et al. 1999). Rhodnius ecuadoriensis is considered one of the domestic and sylvatic vectors of T. cruzi in Ecuador (Vallejo et al. 2009), and natural infection with T. rangeli has been reported before (Guhl & Vallejo 2003).

Given the analytical sensitivity and specificity of the RFLP and duplex PCR approach, both methods are adequate for differentiation between T. cruzi and T. rangeli. PCR-hsp70TcTr has two advantages over RFLP: it needs only gel analysis following PCR, and its sensitivity for identifying mixed infections is higher, which is likely related to the fact that the specific primers do not compete for the same template. Hence, it seems the method of choice for analysing clinical samples. A disadvantage of any specific PCR is the danger of cross-reactivity with the other species in case the DNA amounts would be very high, even though we did not detect any problem up to 10 ng DNA, a quantity not expected in clinical samples. RFLP does not suffer from this latter drawback and is the method of choice for analysing highly concentrated culture DNA. Sau3AI is about 10 times the price of HaeIII and tends to produce more spurious bands at high DNA concentrations, probably from incompletely digested products. HaeIII, on the other hand, shows some variability in T. rangeli, and in our opinion, a combination of both enzymes is the best option. This would also ensure timely detection of amplicons from other Trypanosoma or Leishmania species, even though such cross-reactions seem exceptional. Our methods are promising candidates for testing on human, animal and environmental samples in clinical and epidemiological studies, but also in some limited-resource settings, where they can be used to the benefit of the patients and populations at risk of Chagas' disease.

Acknowledgements

The authors would like to thank all colleagues and institutes who kindly donated the Trypanosoma reference strains and DNA, among whom S. Deborggraeve, P. Büscher (Institute of Tropical Medicine, Antwerp, Belgium); F. Guhl (Universidad de los Andes, Bogota, Colombia); M. D. Lewis and M. A. Miles (London School of Hygiene and Tropical Medicine, London, United Kingdom); H. J. Carrasco (Instituto de Medicina Tropical, Caracas, Venezuela); and J. A. Costales and M. J. Grijalva (Centro de Investigación en Enfermedades Infecciosas (CIEI), Pontificia Universidad Católica del Ecuador, Ecuador and Tropical Disease Institute (TDI), Ohio University, USA). We also thank the Molecular Biology Laboratory (Parasitology, Virology and Bacteriology), IPK, Havana, Cuba. This work was funded by the third framework agreement of the Belgian Directorate General for Development with ITM Antwerp.

Ancillary