Identification of six Trypanosoma cruzi phylogenetic lineages byrandom amplified polymorphic DNA and multilocus enzyme

electrophoresis

Sylvain Brisse *, Christian Barnabe, Michel Tibayrenc

Centre dEtudes sur le Polymorphisme des Microorganismes (CEPM,) UMR No. 9926 Centre National de la Recherche Scientifique/Institut de

Recherche pour le Developpement, IRD, 911 avenue Agropolis, BP 5045, 34 032 Montpellier Cedex 01, France

Received 20 September 1999; received in revised form 25 October 1999; accepted 25 October 1999

Abstract

Genetic characterisation of Trypanosoma cruzi variants is of foremost importance, due to considerable genetic and biologicalheterogeneity in the parasite populations. Two major phylogenetic lineages, each highly heterogeneous, have been previouslydescribed within this species. Here we characterised a geographically and ecologically diverse sample of stocks representative of

the breadth of the known clonal diversity of each major lineage, using random amplified polymorphic DNA with 20 primersand multilocus enzyme electrophoresis at 22 loci. Molecular hybridisation experiments were performed to control the homologyof randomly amplified DNA markers. Both sets of data were highly consistent and supported the existence of two major

lineages. Additionally, we found that lineage 2 appeared further partitioned into five sharply delineated phylogenetic clusters,each comprising one of the following reference strains: CanIII cl1 (Z3 reference), M5631 cl5, Esmeraldo cl3 (Z2 reference), CLBrener, and MN cl2. The two first clusters were found mainly in sylvatic environments, whereas the three latter were restrictedto domestic transmission cycles and were only collected South to the Amazon Basin. In contrast, lineage 1, which included

Miles Z1 reference strain X10 cl1, was not further subdivided and was encountered across the entire endemic area, in bothdomestic and sylvatic cycles. Thus, T. cruzi appeared to be subdivided into six discrete typing units, or DTUs, exhibiting distinctgeographic and ecological ranges. Reliable diagnostic markers for the two major lineages and the five smaller DTUs of lineage 2

are described, and correspondence with previous classifications of T. cruzi genotypes is given in order to help communication onT. cruzi phylogenetic diversity. # 2000 Australian Society for Parasitology Inc. Published by Elsevier Science Ltd. All rightsreserved.

Keywords: Chagas disease; Classification; Diagnostic markers; Epidemiology; Multilocus enzyme electrophoresis; Phylogeny; Random amplified

polymorphic DNA; Trypanosoma cruzi

1. Introduction

Trypanosoma cruzi, the protozoan parasite respon-

sible for Chagas disease in Latin America, shows con-

siderable genetic polymorphism. This was first

demonstrated by multilocus enzyme electrophoresis

(MLEE) [13] and subsequently confirmed by restric-

tion analysis of kinetoplast DNA [4], pulse field gel

electrophoresis [5], DNA fingerprinting [6], and ran-

dom amplified polymorphic DNA (RAPD) [7, 8].

Analysis of MLEE polymorphisms in population gen-

etics terms has shown that T. cruzi has a basically clo-

nal population structure, with genetic exchange being

rare enough to allow the propagation of clonal geno-

types over broad geographic regions and long periods

of time [2, 3, 9].

International Journal for Parasitology 30 (2000) 3544

0020-7519/00/$20.00 # 2000 Australian Society for Parasitology Inc. Published by Elsevier Science Ltd. All rights reserved.PII: S0020-7519(99 )00168 -X

www.elsevier.nl/locate/ijpara

* Corresponding author. Present address: Eijkman-Winkler

Institute, Academic Hospital Utrecht, AZU G04.614, Heidelberglaan

100, 3584 CX Utrecht, The Netherlands. Tel.: +31-30-2507625; fax:

+31-30-2541770

E-mail address: sbrisse@lab.azu.nl (S. Brisse)

Chagas disease is characterised by a wide spectrumof clinical outcomes, ranging in severity from asympto-matic infections to severe cardiac and digestive-tractdamage, and the infective genotypes might play an im-

portant role in this variation [10]. Some biologicalproperties, among which virulence in mice [11], sensi-tivity to antichagasic drugs [12] or transmissibilitythrough triatomine bugs [13, 14] are statistically corre-

Table 1

Origin and genetic characterisation of the 50 stocks under study

No. Stock Taxon Date Country Locality Host Clonal genotypea DTUb

1 M1117 T. cruzi marinkellei 1969 Brazil Abaetetuba, Para Phyllostomum hastatus n.d.

2 SC13 T. cruzi ? Colombia San Carlos, Antioquia Rhodnius pallescens n.d. 1

3 Davis2 T. cruzi 1983 Honduras Tegucigalpa Triatoma dimidiata 13 1

4 STP3.1 T. cruzi ? Colombia Coyaima, Tolima Rhodnius prolixus n.d. 1

5 OPS21 cl11 T. cruzi 1977 Venezuela Cojedes Human 19 1

6 Tehuentepec cl2 T. cruzi 1938 Mexico ? Triatominae 12 1

7 OPS4 T. cruzi 1976 Venezuela Carabobo Didelphis marsupialis 21 1

8 X10 cl1 T. cruzi

lated with genetic dierences. Moreover, dierences inthe degree of pathogenicity to humans of distinct T.cruzi genotypes have been reported [15].Both the evolutionary stability and the distinct bio-

logical properties of T. cruzi natural clones underlinethe importance of genetic characterisation and intra-specific classification based on phylogenetic analysis.Trypanosoma cruzi clonal genotypes appear to clus-

ter into two major phylogenetic lineages based onMLEE and RAPD data [8, 16] and on the polymorph-ism of rRNA and mini-exon genes revealed by PCRand sequencing analyses [17, 18].However, each lineage shows considerable hetero-

geneity, in terms of both genetic diversity and geo-graphic and ecological spectrum. The purpose of thisstudy was to investigate the phylogenetic diversitywithin each lineage, to determine if the clonal geno-types could be grouped into a limited number of dis-crete phylogenetic lineages, and to identify diagnosticmarkers for these lineages.

2. Materials and methods

2.1. Parasite stocks

In this paper, we will use the Linnaean triname T.cruzi marinkellei [19] to refer to those stocks isolatedfrom bats in South America that are morphologicallyidentical to T. cruzi, and will reserve the name of T.cruzi to the agent of Chagas disease. A panel of 49 T.cruzi stocks, numbered 250 in Table 1, was selectedto be representative of the overall clonal diversity ofeach major lineage observed among 440 T. cruzi stocksthat had been characterised at 22 enzymatic loci in ourlaboratory (Barnabe, unpublished). Trypanosoma cruzimarinkellei M1117 was included as an outgroup forphylogenetic analysis, according to its biochemicalcharacteristics [19, 20]. Eighteen stocks were cloned inour laboratory by micromanipulation and are indi-cated in Table 1 by the letters cl followed by a clonenumber. The stocks were cultured in liver infusiontryptose (LIT) medium containing 10% FCS and50 mg ml1 gentamycin.

2.2. Preparation of the samples

Stocks were harvested by centrifugation (2800 g,20 min, 48C) and washed in PBS (Na2HPO4 10 mM,NaH2PO4 10 mM, NaCl 150 mM, pH 7.2). Cells werelysed in an equal volume of hypotonic enzyme stabil-iser (EDTA 2 mM, dithiotreitol 2 mM, E-aminocaproicacid 2 mM), on ice for 20 min. The lysates were againcentrifuged (13 000 g, 10 min, 48C). The soluble frac-tion was stored at 708C until used in MLEE analysis,whereas the pellet of lysed cells was used for DNAextraction, according to the following protocol. Pellets

were resuspended in 400 ml TrisHCl 100 mM(pH 8.0), NaCl 400 mM EDTANa2 10 mM. SDS wasthen added to a final concentration of 1%, and thetubes were incubated for 2 h at 378C in the presence of100 mg ml1 RNase A, and then overnight at 558Cwith 200 mg ml1 proteinase K. After two phenolextractions, two chloroform:isoamyl alcohol (24:1)extractions and ethanol precipitation, the DNA wasresuspended in sterile double-distilled water. DNAconcentration was estimated by spectrophotometry at260 nm. DNA extracts were stored at 208C untilanalysis.

2.3. Protocol for isoenzyme analysis

Multilocus enzyme electrophoresis analysis was per-formed on cellulose acetate plates (HelenaLaboratories) according to Ref. [21], with slight modi-fications. The following 20 enzyme systems were used:aconitase (ACON, EC 4.2.1.3), alanine aminotransfer-ase (ALAT, EC 2.6.1.2), diaphorase (DIA, EC1.6.99.2), glyceraldehyde-3-phosphate dehydrogenase(GAPD, EC 1.2.1.12), glutamate dehydrogenaseNAD+ (GDH-NAD+, EC 1.4.1.2), glutamate dehy-drogenase NADP+ (GDH-NADP+ , EC 1.4.1.4), glu-tamate oxaloacetate transaminase (GOT, EC 2.6.1.1),glucose-6-phosphate dehydrogenase (G6PD, EC1.1.1.49), glucose-phosphate isomerase (GPI, EC5.3.1.9), isocitrate dehydrogenase (IDH, EC 1.1.1.42),leucine aminopeptidase (LAP, EC 3.4.11.1), malate de-hydrogenase NAD+ (MDH, EC 1.1.1.37), malate de-hydrogenase NADP+ or malic enzyme (ME, EC1.1.1.40), mannose-phosphate isomerase (MPI, EC5.3.1.8), nucleoside hydrolase (inosine) (NHi, EC2.4.2.-), peptidases 1 and 2 (substrates: L-leucyl-leu-cine-leucine and L-leucyl-L-alanine) (PEP1 and PEP2,respectively, EC 3.4.11 or 13.-), 6-phosphogluconatedehydrogenase (6PGD, EC 1.1.1.44), phosphogluco-mutase (PGM, EC 2.7.5.1), and superoxide dismutase(SOD, EC 1.15.1.1). These 20 enzyme systems corre-spond to 22 dierent genetic loci, since diaphorase andmalic enzyme exhibit the activity of two distinct locieach.

2.4. Random amplified polymorphic DNA protocol andhybridisations

One-hundred and twenty dierent decameric primerswere screened on a reference panel of 18 T. cruzistocks. The primers corresponded to the A, B, F, N, Rand U kits from Operon Technologies. The amplifica-tion reactions were performed as in Ref. [22], in a finalvolume of 60 ml containing 0.9 units Taq Polymerase(Boehringer), 100 mM each dNTP, 200 nM primer,1.5 mM MgCl2, 50 mM KCl, 10 mM TrisHCl, pH8.3, and 20 ng template DNA. Forty-five cycles (dena-turation: 1 min at 948C; annealing: 1 min at 368C;

S. Brisse et al. / International Journal for Parasitology 30 (2000) 3544 37

elongation: 2 min at 728C) were followed by a finalelongation step of 7 min at 728C. Amplification wasperformed in a PTC-100 thermocycler (MJ Research).Random amplified polymorphic DNA products wereanalysed by electrophoresis in 1.6% agarose gels inTAE buer (Trisacetate 40 mM, EDTA 1 mM),stained with ethidium bromide and visualised by ultra-violet light. The 21 primers that gave the most easilyreadable profiles were selected and used on the set of50 T. cruzi stocks in Table 1. One primer (N4) was dis-carded from the analysis because of lack of reproduci-bility. Table 2 gives the sequence of the 20 retainedprimers. Molecular hybridisation experiments wereperformed after transfer onto a nylon membrane(Hybond N+ , Amersham) according to Ref. [23]. TheRAPD fragments used as probes were eluted from thegel and labelled with [a-32P]dCTP using a RandomPriming kit (Boehringer) according to the suppliersrecommendations. Membranes were washed at inter-mediate stringency (0.1% SSC, 0.1% SDS at roomtemperature) before autoradiography.

2.5. Data analysis

Multilocus enzyme electrophoresis and RAPD datawere computed with the GENETICS software,designed in our laboratory and operated on SUNstations. Two genetic variability estimators were used.(i) The polymorphism index P= ni/n, where ni is thenumber of polymorphic loci and n is the total numberof loci, considering a RAPD pattern or a MLEE locuspolymorphic when the frequency of the most rep-

resented genotype is less than 95%. (ii) The mean gen-etic diversity [24], H=Shij/n, where n is the numberof loci (or RAPD primers) and hij=1Sxij 2, wherexij is the relative frequency of the ith pattern at the jthlocus (or RAPD primer).For both MLEE and RAPD data, the Jaccard

distance [25] was used to estimate the genetic dier-ences among the stocks. Each MLEE and RAPD gelband was coded with a number, starting with 1 for thefastest band in the case of MLEE, and 1 for the slow-est band in the case of RAPD. The distance was esti-mated after the following formula:

D 1 a=a b cwhere a=number of bands that are common to thetwo compared genotypes; b=number of bands presentin the first genotype and absent in the second;c=number of bands absent in the first genotype andpresent in the second.The unweighted pair-group method with arithmetic

averages (UPGMA) [26] and the neighbour joiningmethod [27] were used to cluster the genotypestogether according to their Jaccard distances, using theNEIGHBOR program of the PHYLIP package, ver-sion 3.5c [Felsenstein J. PHYLIP (Phylogeny InferencePackage) version 3.5c. Seattle, WA: Department ofGenetics, University of Washington, 1993]. Trees weredrawn with the TREEVIEW program, version 1.4 [28].Agreement between MLEE and RAPD genetic dis-

tances was estimated using the g test of linkagedisequilibrium [29], which corresponds to a correlationanalysis based on the non-parametric test ofMantel [30]. Briefly, this test relies on a Monte Carlosimulation with 104 iterations, which randomly per-mutes the dierent cells of one of the distancematrices. In contrast to the classical correlation test,this randomisation procedure does not make anyassumptions about the number of degrees of freedom.

3. Results

3.1. Random amplified polymorphic DNA polymorphismin T. cruzi

The 50 trypanosome stocks in Table 1 were analysedwith the 20 RAPD primers listed in Table 2. All pri-mers gave clear and reproducible amplification pat-terns, as was verified by comparison with the resultsobtained previously for a subset of 18 stocks screenedfor 120 dierent primers.The 49 T. cruzi stocks could be ordered into six

groups based on RAPD profiles obtained with primerF5 (Fig. 1) and primers A15, R16, U7 and U14 (notshown). The stocks within a given group exhibited noor few pattern dierences, whereas clear distinction

Table 2

Random amplified polymorphic DNA primers used in the present

study

Name Sequence

A10 GTGATCGCAT

A15 TTCCGAACCC

B2 TGATCCCTGG

B11 GTAGACCCGT

B15 GGAGGGTGTT

B17 AGGGAACGAG

B19 ACCCCCGAAG

F1 ACGGATCCTG

F4 GGTGATCAGG

F5 CCGAATTCCC

F8 GGGATATCGG

F13 GGCTGCAGAA

F15 CCAGTACTCC

N9 TGCCGGCTTG

N19 GTCCGTACTG

R5 GACCTAGTGG

R16 CTCTGCGCGT

U7 CCTGCTCATC

U11 AGACCCAGAG

U14 TGGGTCCCTC

S. Brisse et al. / International Journal for Parasitology 30 (2000) 354438

was achieved among groups by highly discriminatoryRAPD fragments, such as fragments F5-c to F5-g(Fig. 1).Genetic relatedness among groups was indicated by

common RAPD fragments, such as fragment F5-ebetween groups 2b, 2d and 2e. With some primers,stocks belonging to distinct groups even demonstratedthe same overall RAPD patterns, as is illustrated withprimer N9 for groups 2c, 2d and 2e (Fig. 2). However,such fragments would be suitable for deducing phylo-genetic relationships only if they would really corre-spond to homologous genetic characters. In order toinvestigate this question, hybridisation experimentswere performed. Figs. 1 and 2 show the resultsobtained using as a probe fragments F5-e and N9-b,respectively, from stock Tulahuen cl2. The resultsdemonstrated that fragments F5-e are homologous ingroups 2b, 2d and 2e, and that fragments N9-b fromgroups 2a, 2c, 2d and 2e are homologous. Similarly,homology was verified for the six other RAPD frag-ments tested: A10-e, B11-e, F13-e, N19-a, U7-e, andU14-a (data not shown).

3.2. Clustering analysis

The RAPD data were used to estimate the phyloge-netic relationships among the stocks. Only the RAPDfragments showing high intensity were selected foranalysis. After amplification with the 20 primers, 234RAPD fragments were scored, among which 232 werepolymorphic. In total, 42 dierent genotypes were dis-

criminated, with two pairs (Cuica cl1/SO34 cl4, andSO50/CBB cl3) and two groups of four stocks (MNcl2/JSR6/NR cl3/BMS and CL Brener/Tulahuen cl2/Tula Yale/Guateque) showing identical profiles withall primers. The polymorphism index was P=1, andthe genetic diversity was H=0.80.The dendrogram obtained using the neighbour join-

ing method is presented in Fig. 3, and similar resultswere obtained using the UPGMA method (notshown). In agreement with its taxonomicdistinction [19], the bat trypanosome T. cruzi marinkel-lei M1117 was placed in external position. Trypano-soma cruzi stocks clustered into two well-separatedbranches. The first major branch was made of stocksbelonging to lineage 1 [16], in particular stock X10 cl1,reference of the zymodeme group Z1 of Miles [1]. Theremaining stocks formed the second major lineage ofT. cruzi [16] and were clearly clustered into the fivesubdivisions 2a2e. The stocks within each clusterappeared genetically much more closely relatedbetween them than to stocks in other clusters, as illus-trated in the dendrogram by the long inter-clusterbranch lengths, as compared with those within theclusters.The inclusion of reference stocks in the analysis

made it possible to compare the clusters identifiedherein with previous descriptions, as indicated inTable 3. Cluster 2a included stock CanIII cl1, refer-ence for zymodeme Z3 of Miles [1]. Stock Esmeraldocl3, reference for zymodeme Z2 of Miles [1], fell intocluster 2b, and stocks M6241 cl6 and M5631 cl5, pre-

Fig. 1. Random amplified polymorphic DNA (RAPD) analysis with primer F5. (A) Random amplified polymorphic DNA profiles obtained for

48 T. cruzi stocks (lanes 249) and T. cruzi marinkellei stock M1117 (lane 1). The numbers above the lanes correspond to the stock numbers in

Table 1. Lanes marked with M correspond to phage lambda DNA digested with BstEII. Five diagnostic RAPD fragments (see text and Fig. 3)

are indicated by a letter on both sides of the gel photograph. The distribution of the fragments across the stocks is as follows: F5-c is observed

in clusters 2b and 2d; F5-d in lineage 1 and stock M1117; F5-e in clusters 2b, 2d and 2e; F5-f in stock DogT (lane 23); F5-g in clusters 2c, 2d

and 2e and, less intensely amplified, in the remaining stocks. (B). Autoradiogram of the gels after transfer onto a nylon membrane and hybridis-

ation with fragment F5-e from stock Tulahuen cl2 (lane 47) used as a radioactive probe, showing the homology of F5-e fragments across stocks

of clusters 2b, 2d and 2e.

S. Brisse et al. / International Journal for Parasitology 30 (2000) 3544 39

viously described as Z3 with Z1 ASATcharacteristics [31], were included in cluster 2c. CloneCL Brener, the reference strain for the T. cruziGenome Project, was included in cluster 2e, whereascluster 2d was represented by stocks previouslydescribed as Bolivian Z2 [32].

3.3. Multilocus enzyme electrophoresis analysis

Multilocus enzyme electrophoresis analysis was per-formed at 22 isoenzyme loci on all 50 stocks, exceptstock Ep255 (not shown). When possible, according tothe patterns observed, an allelic interpretation ofMLEE variability was performed, based on the hy-pothesis that T. cruzi is a diploid organism [2, 3, 9]. Inagreement with previous results [3], the present set ofstocks exhibited considerable polymorphism. Forty-sixdistinct genotypes were found, with one pair of clones(BMS, Kundera) and a triplet (NR cl3, 92.80 cl1,SC43 cl1) sharing identical genotypes. The polymorph-ism index was P=0.90, and the mean genetic diversitywas H=0.55.After clustering analysis (Fig. 3), the stocks were as-

sociated into the six same clusters and two majorlineages as in RAPD analysis. The correlation betweenRAPD and MLEE genetic distances was highly signifi-cant as revealed by the Mantel test (r=0.89,P

of the major lineages and of the five additional subdi-visions of lineage 2. Three kinds of characterisationmarkers could be distinguished. First, some individualRAPD fragments or MLEE electromorphs wereobserved in all or almost all stocks of a given subdivi-sion of the dendrogram, and only in these stocks.Second, some specific RAPD and MLEE profiles wereobserved, even if each fragment or electromorph wasindividually amplified in other clusters. Third, we alsofound some RAPD and MLEE characters that wereshared specifically among two or more clusters,although these clusters were not branched together inthe dendrograms. When considered in combination,these characters are also very useful for the identifi-cation of the clusters. Examples of each kind ofcharacterisation markers are given in the legend ofFig. 3.

3.5. Geographic and ecological distribution of the T.cruzi groups

Although based on a rather limited number ofstocks, clear dierences appeared in the geographicdistribution of the six clusters (Table 1). While isolatesof lineage 1 were found across the whole endemic area,the five lineage 2 clusters appeared geographicallymore restricted. Clusters 2b, 2d and 2e were foundonly in the Southern regions of the endemic area. Theonly exception was the Colombian stock Guateque ofcluster 2e. Inversely, cluster 2a was mainly representedin the Amazon Basin and North up to the USA,which was confirmed by the MLEE analysis of 20 ad-ditional stocks (Barnabe, unpublished). Cluster 2c wasfound both within the Amazon Basin and in Paraguay.Our data also suggested an ecological distinction

among the clusters, as indicated by the hosts or vectors

from which they were isolated (Table 1). Lineage 1was isolated in humans and in the domiciliary vectorTriatoma infestans, as well as in sylvatic environments,e.g. in Didelphis marsupialis. Clusters 2a and 2c werealmost exclusively found in sylvatic environments,whereas clusters 2b, 2d and 2e were found exclusivelyin humans and in domestic triatomines.These geographical and ecological characteristics of

the clusters were confirmed when considering the set of440 T. cruzi stocks analysed by MLEE in our labora-tory (Barnabe, unpublished).

4. Discussion

4.1. Description of six phylogenetic lineages within T.cruzi

We describe a sharp clustering of the present set ofstocks into six genetic clusters, five of them being as-sociated into the formerly described lineage 2 [16].Although distinction between the reference strains ofeach cluster was recognised earlier [1, 3, 22, 31, 32], thisis the first study showing the existence of sharply deli-neated phylogenetic units within lineage 2.The use of RAPD data in phylogenetic analysis may

not be appropriate, because some amplified fragmentscorresponding to non-homologous genomic sequencesin dierent stocks could have, by chance, the sameelectrophoretic mobility [33]. Although it was not poss-ible to test the homology of all fragments included inthe analysis, the fact that all eight hybridisation con-trol experiments were positive suggests that RAPDfragments can be used confidently in phylogeneticanalysis of T. cruzi clones. The present study is thefirst where such controls are performed for this parasite.

Table 3

Correspondence between the Trypanosoma cruzi discrete typing units (DTUs) described here and previous descriptions

Present study Previous descriptionsa Representative stocks

A B C D Ec Fc G

DTU1 Lineage 1 125 Z1 Lineage 2 Z IV Z2, Z5, Z7, Z10, Z12 14, 68 X10 cl1, Cuica cl1

DTU2a Lineage 2 2629 Z3 Lineage 1 5, 9, 10 CanIII cl1

DTU2b Lineage 2 3034 Z2 (Brazilian Z2d) Lineage 1 Z I Z4, Z11 5, 9, 10 TU18 cl2, Esmeraldo cl3

DTU2c Lineage 2 3537 Z3/Z1 ASATe Lineage 1, Lineage 3b 5, 9, 10 M5631 cl5; M6241 cl6

DTU2d Lineage 2 38, 39 Z2 (Bolivian Z2d) Lineage 1 Z1, Z3, Z9 5, 9, 10 MN cl2, SC43 cl1

DTU2e Lineage 2 4043 Lineage 1 Z III 5, 9, 10 CL Brener, Tulahuen cl2

aColumn A, Tibayrenc [16]; Column B, Tibayrenc et al. [3]; Column C, Miles et al. [1]; Column D, Souto et al. [17]; Column E, Romanha et

al. [42]; Column F, De Luca DOro et al. [43]; Column G, Higo et al. [44].bLineage 3 was proposed by Fernandes et al. [34].cThe correspondence between ZII [42] and Z6/Z8 [43] on the one hand, and the stocks surveyed here on the other hand, was impossible to

establish.dBrazilian and Bolivian Z2 were distinguished in Tibayrenc and Miles [32].eZ3/Z1 ASAT was described in Miles et al. [31]. Note that lineage nomenclature is inverted since the rRNA and mini-exon lineages 1 and 2 [17]

correspond to the RAPD and MLEE lineages 2 and 1 [16], respectively.

S. Brisse et al. / International Journal for Parasitology 30 (2000) 3544 41

Multilocus enzyme electrophoresis analysis providedfurther support to the subdivision of lineage 2, as itpartitioned the stocks into the same clusters as theRAPD analysis. This was true even if the delineationappeared less sharp in the MLEE dendrogram, whichcan be explained by the higher Jaccard distancesobserved within the clusters (Fig. 3).The view that T. cruzi is partitioned into six phylo-

genetic clusters must be considered as a basis forfuture descriptive work, rather than a fixed frame.Indeed, analysis of additional stocks will lead to identi-fication of new clusters, or new subdivisions within thepresent clusters, as already suggested by the divergentstocks DogT, SC13 and Davis2 within their respectiveclusters.Our results, based on a much broader number of

stocks and genetic markers, support the subdivision ofT. cruzi clonal diversity into two major lineages pre-

viously proposed on the basis of RAPD and MLEEdata [8, 16] and of the dimorphisms of the 24Sa rRNAgene and mini-exon intergenic region [17, 18]. Recently,mini-exon analysis of additional stocks that belong toclusters 2a and/or 2c of the present study, havesuggested the existence of a third lineage of T.cruzi [34]. Whether this challenging result is due toevolutionary rate dierences between markers or to thefact that mini-exon gene phylogeny does not accuratelyreflect the strain phylogeny estimated by RAPD andMLEE analyses, possibly due to genetic exchange,deserves further investigation.Tibayrenc [16] proposed that a fundamental distinc-

tion among micro-organisms lies between structuredand non-structured species, the first category beingcharacterised by the existence of discrete evolutionarylineages (discrete typing units or DTUs [35]) that canbe specifically identified by diagnostic characters, or

Fig. 3. Neighbour joining dendrograms based on the analysis of 20 RAPD primers (left) and 22 isoenzymatic loci (right) showing the genetic re-

lationships between 49 T. cruzi stocks and T. cruzi marinkellei stock M1117. The scale indicates the Jaccard distance along the branches. Six gen-

etic clusters, or Discrete Typing Units (DTUs, [37]), were distinguished and their names are given in the central column between the

dendrograms. DTU 1 corresponds to the first major lineage of T. cruzi, while the second major lineage is subdivided into DTUs 2ae.Diagnostic

RAPD fragments and isoenzymatic patterns, which were specifically observed in the stocks of a given branch of the dendrograms, are indicated

at the corresponding nodes. For example, fragment N9-d was observed in all stocks of lineage 2 (Fig. 2). Specific RAPD profiles are indicated

between brackets. For example, DTU 2d was identified by its specific F5 overall pattern, even if each fragment was individually amplified in

other DTUs (Fig. 1). Some characters that were shared between two or more groups, although these groups were not branched together in the

dendrograms, are indicated between parentheses. For example, fragment F5-e was observed specifically in DTUs 2b, 2d and 2e (Fig. 1). The

stars (*) in the MLEE dendrogram indicate that the corresponding diagnostic character suers one exception, which is given on the correspond-

ing branch. The homology of fragments A10-e, F5-e, F13-e, N9-b, N19-a, U7-e, and U14-e was controlled by hybridisation experiments.

S. Brisse et al. / International Journal for Parasitology 30 (2000) 354442

tags [35]. Our results strongly suggest that T. cruzi is atypical example of a highly structured species. Usingthe terminology proposed, T. cruzi is subdivided intotwo main DTUs, of which the second one is itself sub-divided into five additional DTUs. All these DTUs canbe identified by specific MLEE and RAPD characters.The present study suggests that multiprimer RAPDanalysis could be employed to identify infra-specificgenetic structure in other microbial species, either pro-karyotic or eukaryotic.Dierent evolutionary models could account for the

origin and dierentiation of the six DTUs [16, 36]. Thepreliminary results obtained here suggest that clonalevolution is the main parameter responsible for T.cruzi genetic diversity, as formerly proposed [3, 9].Indeed, the strong correlation found between RAPDand MLEE genetic distances within several DTUs indi-cates that linkage disequilibrium (non-random associ-ation of genotypes occurring at dierent loci) ispresent within those subdivisions too. The lack of cor-relation observed within DTUs 2b, 2c, 2d and 2e iseasily explained by a statistical type-II error, i.e. a lackof power of the test due to insucient data [16].Besides, the peculiarities observed for DTUs 2d and 2esuggest that they could correspond to clonal lineageshaving an origin by recombination between parentalstocks belonging to DTUs 2b and 2c, as was notedpreviously for some stocks falling in DTU 2e [22]. Ahybrid origin of DTU 2e could also account for itsdierent positioning in the RAPD and MLEE dendro-grams. Thus, genetic exchange could play a role, evenif rarely, in the generation of T. cruzi genetic diversity,as was proposed by others [37, 38].

4.2. Epidemiological and taxonomic significance

A widely accepted descriptive framework is sorelyneeded in T. cruzi genetic variability studies, sinceuntil now, comparison between publications was ren-dered unnecessarily dicult by the use of conflictingdescriptive systems (Table 3). The subdivisionsdescribed here are defined on a clear phylogeneticbasis, are each identified by numerous specific diagnos-tic markers, and comprise the major genetic variantsdescribed to date. Therefore, we feel that the six DTUsrepresent a useful description of the known clonaldiversity of T. cruzi. In future studies, it is advisable toinclude reference stocks representing each of the sixDTUs, for example those indicated in Table 3, whichare available upon request.The fact that the five DTUs of lineage 2 show

defined geographical and ecological characteristics con-trasts with the heterogeneity, in this respect, of thewhole lineage 2, and underlines the epidemiologicalrelevance of a distinction between the five additionalDTUs. Future development of molecular diagnostic of

the DTUs will be needed to improve the tools pre-sently available for Chagas disease epidemiology. Forexample, PCR amplification of minicircle DNA, oneof the most sensitive characterisation assays presentlyavailable, can only identify cluster 2d on the one hand,and a subset of lineage 1 on the other hand [3941].Two widespread Argentinian zymodemes have been

suggested to have distinct pathogenic properties inhumans [15]. As can be deduced from the comparisonof their Gpi profiles, which have been shown herein torepresent reliable diagnostic markers, the most virulentof the Argentinian zymodemes probably belong tolineage 1, whereas the less virulent one probably corre-sponds to DTU 2d (Table 3). In keeping with theseobservations, experimental studies have shown thatstocks of lineage 1 have a higher virulence in mice andtransmissibility through triatomine vectors, and alower sensitivity to antichagasic drugs, than stocksthat can be attributed to DTUs 2b and 2d [1114].However, the clonal genotypes could have distinct bio-medical properties even within a given cluster, and thebiological significance of the DTUs remains to bedetermined. Therefore, we feel no need to give themany specific or subspecific taxonomic status based onbiname or triname Linnaean nomenclature.

Acknowledgements

This work was supported by a concerted actionACCSV7 from the French Ministry of Research, aGroupement dEtudes et de Recherches sur lesGenomes (GREG) grant, a Groupement de Recherche(GDR) CNRS/ French Army, and a WHO TDR grantno. 880190. We thank Pr. H. Theis for providing theDogT stock.

References

[1] Miles MA, Souza A, Povoa M, Shaw JJ, Lainson R, Toye PJ.

Isoenzymic heterogeneity of Trypanosoma cruzi in the first auto-

chthonous patients with Chagas disease in Amazonian Brazil.

Nature 1978;272:81921.

[2] Tibayrenc M, Cariou ML, Solignac M, Carlier Y. Arguments

genetiques contre lexistence dune sexualite actuelle chez

Trypanosoma cruzi; implications taxonomiques. C R Acad Sci,

Paris 1981;293:2079.

[3] Tibayrenc M, Ward P, Moya A, Ayala FJ. Natural populations

of Trypanosoma cruzi, the agent of Chagas disease, have a com-

plex multiclonal structure. Proc Natl Acad Sci USA

1986;83:1159.

[4] Morel C, Chiari E, Plessmann Camargo E, Marrei DM,

Romanha AJ, Simpson L. Strains and clones of Trypanosoma

cruzi can be characterized by pattern of restriction endonuclease

products of kinetoplast DNA. Proc Natl Acad Sci USA

1980;77:68104.

[5] Henriksson J, Aslund L, Macina RA et al. Chromosomal local-

ization of seven cloned antigen genes provides evidence of

diploidy and further demonstration of karyotype variability in

Trypanosoma cruzi. Mol Biochem Parasitol 1990;42:2134.

S. Brisse et al. / International Journal for Parasitology 30 (2000) 3544 43

[6] Macedo AM, Melo MN, Gomes RF, Pena SD. DNA finger-

prints: a tool for identification and determination of the re-

lationships between species and strains of Leishmania. Mol

Biochem Parasitol 1992;53:6370.

[7] Steindel M, Dias Neto E, de Menezes C, Romanha AJ,

Simpson AJG. Random amplified polymorphic DNA analysis

of Trypanosoma cruzi strains. Mol Biochem Parasitol

1993;60:7180.

[8] Tibayrenc M, Neubauer K, Barnabe C, Guerrini F, Skarecky

D, Ayala FJ. Genetic characterization of six parasitic protozoa:

parity between random-primer DNA typing and multilocus

enzyme electrophoresis. Proc Natl Acad Sci USA 1993;90:13359.

[9] Tibayrenc M, Ayala FJ. Isozyme variability in Trypanosoma

cruzi, the agent of Chagas disease: genetical, taxonomical, and

epidemiological significance. Evolution 1988;42:27792.

[10] Miles MA, Povoa MM, Prata A, Cedillos RA, De Souza AA,

Macedo V. Do radically dissimilar Trypanosoma cruzi strains

(zymodemes) cause Venezuelan and Brazilian forms of Chagas

disease?. Lancet 1981;i:133840.

[11] Laurent JP, Barnabe C, Quesney V, Noel S, Tibayrenc M.

Impact of clonal evolution on the biological diversity of

Trypanosoma cruzi. Parasitology 1997;114:2138.

[12] Revollo S, Oury B, Laurent JP et al. Trypanosoma cruzi: impact

of clonal evolution of the parasite on its biological and medical

properties. Exp Parasitol 1998;89:309.

[13] De Lana M, Pinto A, Barnabe C, Quesney V, Noel S,

Tibayrenc M. Trypanosoma cruzi: compared vectorial transmis-

sibility of three major clonal genotypes by Triatoma infestans.

Exp Parasitol 1998;90:205.

[14] Pinto A, de Lana M, Bastrenta B et al. Compared vectorial

transmissibility of pure and mixed clonal genotypes of

Trypanosoma cruzi in Triatoma infestans. Parasitol Res

1998;84:34853.

[15] Montamat EE, De Luca DOro GM, Gallerano RH, Sosa R,

Blanco A. Characterization of Trypanosoma cruzi populations

by zymodemes: correlation with clinical picture. Am J Trop

Med Hyg 1996;55:6258.

[16] Tibayrenc M. Population genetics of parasitic protozoa and

other microorganisms. Adv Parasitol 1995;36:47115.

[17] Souto RP, Fernandes O, Macedo AM, Campbell DA, Zingales B.

DNA markers define two major phylogenetic lineages of

Trypanosoma cruzi. Mol Biochem Parasitol 1996;83:14152.

[18] Nunes LR, De Carvalho MRC, Buck GA. Trypanosoma cruzi

strains partition into two groups based on the structure and

function of the spliced leader RNA and rRNA gene promoters.

Mol Biochem Parasitol 1997;86:21124.

[19] Baker JR, Miles MA, Godfrey DG, Barrett TV. Biochemical

characterization of some species of Trypanosoma

(Schizotrypanum) from bats (Microchiroptera). Am J Trop Med

Hyg 1978;27:48391.

[20] Taylor AER, Edwards YH, Smith V et al. Trypanosoma

(Schizotrypanum) species from insectivorous bats

(Microchiroptera): characterization by polypeptide profiles. Syst

Parasitol 1982;4:15568.

[21] Ben Abderrazak S, Guerrini F, Mathieu-Daude F et al.

Isoenzyme electrophoresis for parasite characterization.

Methods Mol Biol 1993;21:36182.

[22] Brisse S, Barnabe C, Banuls AL, Sidibe I, Noel S, Tibayrenc M.

A phylogenetic analysis of the Trypanosoma cruzi genome pro-

ject CL Brener reference strain by multilocus enzyme electro-

phoresis and random amplified polymorphic DNA

fingerprinting. Mol Biochem Parasitol 1998;92:25363.

[23] Cuny G, Veas F, Roizes G. Pocket blotting: a method for trans-

ferring nucleic acids onto nylon membranes. Anal Biochem

1991;193:458.

[24] Nei M. Molecular evolutionary genetics. New York: Columbia

University Press, 1987.

[25] Jaccard P. Nouvelles recherches sur la distribution florale. Bull

Soc Vaudoise Sci Nat 1908;44:22370.

[26] Sneath PHA, Sokal RR. Numerical taxonomy. The principle

and practice of numerical classification. San Francisco:

Freeman, 1973.

[27] Saitou N, Nei M. The Neighbor-Joining Method: a new method

for reconstructing phylogenetic trees. Mol Biol Evol 1987;4:406

25.

[28] Page RDM. TreeView: an application to display phylogenetic

trees on personal computers. Comput Appl Biosci 1996;12:357

8.

[29] Tibayrenc M, Kjellberg F, Ayala FJ. A clonal theory of para-

sitic protozoa: the population structure of Entamoeba, Giardia,

Leishmania, Naegleria, Plasmodium, Trichomonas and

Trypanosoma, and its medical and taxonomical consequences.

Proc Natl Acad Sci USA 1990;87:24148.

[30] Mantel N. The detection of disease clustering and a generalized

regression approach. Cancer Res 1967;27:20920.

[31] Miles MA, Povoa M, De Souza AA, Lainson R, Shaw JJ,

Ketteridge DS. Chagas disease in the Amazon Basin: II. The

distribution of Trypanosoma cruzi zymodemes 1 and 3 in Para

state, north Brazil. Trans R Soc Trop Med Hyg 1981;71:217

25.

[32] Tibayrenc M, Miles MA. A genetic comparison between

Brazilian and Bolivian zymodemes of Trypanosoma cruzi. Trans

R Soc Trop Med Hyg 1983;77:7683.

[33] van de Zande L, Bijlsma R. Limitations of the RAPD technique

in phylogeny reconstruction in Drosophila. J Evol Biol

1995;8:64556.

[34] Fernandes O, Sturm NR, Derre R, Campbell DA. The mini-

exon gene: a genetic marker for zymodeme III of Trypanosoma

cruzi. Mol Biochem Parasitol 1998;95:12933.

[35] Tibayrenc M. Genetic epidemiology of parasitic protozoa and

other infectious agents: the need for an integrated approach. Int

J Parasitol 1998;28:85104.

[36] Maynard Smith J, Smith NH, ORourke M, Spratt BG. How

clonal are bacteria?. Proc Natl Acad Sci USA 1993;90:43848.

[37] Bogliolo AR, Lauria-Pires L, Gibson WC. Polymorphisms in

Trypanosoma cruzi: evidence of genetic recombination. Acta

Trop 1996;61:3140.

[38] Carrasco HJ, Frame IA, Valente SA, Miles MA. Genetic

exchange as a possible source of genomic diversity in sylvatic

populations of Trypanosoma cruzi. Am J Trop Med Hyg

1996;54:41824.

[39] Brenie`re SF, Bosseno MF, Revollo S, Rivera MT, Carlier Y,

Tibayrenc M. Direct identification of Trypanosoma cruzi natural

clones in vectors and mammalian hosts by polymerase chain

reaction. Am J Trop Med Hyg 1992;46:33541.

[40] Britto C, Cardoso MA, Vanni CMM et al. Polymerase chain

reaction detection of Trypanosoma cruzi in human blood

samples as a tool for diagnosis and treatment evaluation.

Parasitology 1995;110:2417.

[41] Brenie`re SF, Bosseno MF, Telleria J et al. Dierent behaviour

of two Trypanosoma cruzi major clones: transmission and circu-

lation in young Bolivian children. Exp Parasitol 1998;98:28595.

[42] Romanha AJ, Da Silva Pereira AA, Chiari E, Kilgour V.

Isoenzyme patterns of cultured Trypanosoma cruzi: changes

after prolonged subculture. Comp Biochem Parasitol

1979;62B:13942.

[43] de Luca DOro GM, Gardenal CM, Perret B, Crisci JV,

Montamat EE. Genetic structure of Trypanosoma cruzi popu-

lations from Argentina estimated from enzyme polymorphism.

Parasitology 1993;107:40510.

[44] Higo H, Yanagi T, Matta V et al. Genetic structure of

Trypanosoma cruzi in Central America and its comparison with

South American strains. Int J Parasitol 1997;27:136974.

S. Brisse et al. / International Journal for Parasitology 30 (2000) 354444