Geographical structuring of Trypanosoma cruzi populations from Chilean Triatoma infestans triatomines and their genetic relationship with other Latino American counterparts

Abstract

In order to obtain more information about the population structure of Chilean Trypanosoma cruzi, and their genetic relationship with other Latino American counterparts, we performed the study of T. cruzi samples detected in the midgut content of Triatoma infestans insects from three endemic regions of Chile. The genetic characteristics of these samples were analysed using microsatellite markers and PCR conditions that allow the detection of predominant T. cruzi clones directly in triatomine midgut content. Population genetic analyses using the Fisher’s exact method, analysis of molecular variance (AMOVA) and the determination of F_ST showed that the northern T. cruzi population sample was genetically differentiated from the two southern population counterparts. Further analysis showed that the cause of this genetic differentiation was the asymmetrical distribution of TcIII T. cruzi predominant clones. Considering all triatomines from the three regions, the most frequent predominant lineages were TcIII (38%), followed by TcI (34%) and hybrid (8%). No TcII lineage was observed along the predominant T. cruzi clones. The best phylogenetic reconstruction using the shared allelic genetic distance was concordant with the population genetic analysis and tree topology previously described studying foreign samples. The correlation studies showed that the lineage TcIII from the III region was genetically differentiated from the other two, and this differentiation was correlated with geographical distance including Chilean and mainly Brazilian samples. It will be interesting to investigate whether this geographical structure may be related with different clinical manifestation of Chagas disease.

INTRODUCTION

Chagas disease is well recognized as the most serious human parasitic disease of the Americas in terms of its social and economic impact. Trypanosoma cruzi is the etiological agent of Chagas disease, transmitted by triatomine bugs; it is one of the most important vector-borne diseases in Latin America, infecting approximately 15 million people (WHO, 2007). Several reports have shown that T. cruzi is mainly diploid, with a clonal population structure with possible rare genetic exchange (Tibayrenc et al., 1986; Brisse et al., 2000; Machado and Ayala, 2001; Gaunt et al., 2003; Tibayrenc, 2003; de Freitas et al., 2006). T. cruzi has been classified into two major phylogenetic lineages, T. cruzi I (TCI) and T. cruzi II (TCII) (Souto et al., 1996; Momen, 1999). Later, based on multilocus enzyme electrophoresis and random amplified polymorphic DNA, TCII was further subdivided into five discrete phylogenetic clusters (Brisse et al., 2000). Today, the six T. cruzi lineages are named as: TcI, TcII, TcIII, TcIV, TcV and TcVI, respectively, which will be used in the present paper (Zingales et al., 2009). Several lines of evidence support the idea that TcV and TcVI lineages are hybrid, originated by a relatively recent genetic fusion between TcII and TcIII (Machado and Ayala, 2001; Westenberger et al., 2005). Later, using mitochondrial sequences and five microsatellite markers, it was shown that the two hybrids, TcV and TcVI lineages, are not separate clusters but rather make up a single hybrid lineage, which was confirmed recently by our group (de Freitas et al., 2006; Venegas et al., 2009a).

These phylogenetic analyses conducted in laboratory cultured T. cruzi strains and clones have provided some information related to the population structure of this parasite, such as the observation that the TcI lineage has been found in a vast geographical and ecological area of Latin America and has been associated both with sylvatic and domestic transmission cycles, as well as both with arboreal and terrestrial mammals (Brisse et al., 2000; Tibayrenc, 2003; Miles et al., 2009), whereas lineages TcII–TcVI exhibit a more restricted geographic and ecological distribution. Thus, the majority of T. cruzi clones of lineages TcV and TcVI were found in the southern cone of South America, while the TcIV lineage has been more frequently found in Venezuela and the Amazon Basin (Brisse et al., 2000; Miles et al., 2009). The TcIII lineage was more frequent in the Amazon Basin and in Paraguay and mainly associated with terrestrial mammals (Miles et al., 1977, 2009; Fernandes et al., 1998; Zingales et al., 1998; Yeo et al., 2005). Very interestingly, recently population genetic studies using a high-resolution microsatellite analysis, nuclear and mitochondrial genes have shown that the TcIII lineage is geographically differentiated and is strongly associated with a terrestrial ecotope (Marcili et al., et al., 2009; Llewellyn et al., 2009 a, b). However, it is worth mentioning that since the majority of these studies have been conducted in laboratory cultured parasites, stocks or isolates, they may not represent the real population structure and ecological distribution of natural populations of T. cruzi, due to selective pressure that the laboratory culture and mouse passages exert over the clones and mixtures of lineages which the T. cruzi samples contain (Deane et al., 1984; Macedo and Pena, 1998; Solari et al., 2001; Coronado et al., 2006).

Chagas disease is widespread in Chile, distributed in rural and peri-urban areas in the seven northern regions of the country. The principal and domiciliary vector of T. cruzi is Triatoma infestans, but the peridomestic and sylvatic triatomines Mepraia spinolai and Mepraia gajardoi also bear T. cruzi. The interruption of the domestic cycle of transmission of T. cruzi was completed by the Chilean Health Minister in 1999 (Lorca et al., 2001). Nevertheless, a permanent entomological vigilance and serological study of people at risk must be conducted to avoid the re-colonisation of domestic niches by new triatomines.

The main aims of the present manuscript were to determine whether Chilean T. cruzi populations are geographically structured and whether this phenomenon may be due to genetic differences among or within lineages. In addition, the genetic relationship of these natural T. cruzi populations with other Latino American counterparts, was investigated. We think that obtaining more insights about the organisation and the factors which originated the T. cruzi populations, is highly relevant since this information could be important in the knowledge about the epidemiology and pathogenesis of Chagas disease.

MATERIALS AND METHODS

Population of Triatomines

The triatomines were captured by the Program of Eradication of the Domiciliary Infestation of Triatoma infestans of the Ministry of Health during period 2005–2007. The sample size in the study corresponded to 15, 37 and 12 T. infestans from the Atacama (III) region (middle point among locations of Alto del Carmen and Tierra Amarilla, −27°27′S, −70°15′W), Valparaiso (V) region (mainly from Petorca location −32° 16′S, −70° 58′W) and Metropolitan (M) region (Calera de Tango location, −33°37′S, −70°47′W) regions of Chile (Table 1).

Table 1. Genetic characteristics of Trypanosoma cruzi samples isolated from Triatoma infestans insects captured by the Chilean Vector Control Program of Health Ministry determined by three microsatellite markers.

Triatomine			Minimal number of clonesa	Lineage of predominant T. cruzi cloneb
Code	Region	Location
	III
1	"	Alto del Carmen (AC)	4	TcIII
9	"	AC	2	TcI
10	"	AC	2	ND
44	"	AC	2	ND
55	"	Tierra Amarilla (TA)	2	TcI
86	"	TA	3	TcIII
106	"	TA	3	ND
107	"	TA	4	TcIII
108	"	TA	5	TcIII
109	"	TA	6	ND
110		TA	6	TcIII
111	"	TA	4	TcIII
112	"	TA	5	TcIII
8–120	"	TA	7	TcI
51–163	"	Mallorca	4	TcI
	V
23	"	Alto del Puerto	2	ND
101	"	Chincolco	3	TcIII
3	"	San Antonio	2	TcI
51	"	Pedernal (PED)	2	TcI
89	"	PED	2	TcIII
90	"	PED	4	TcIII
11	"	Petorca (PET)	2	ND
9	"	PET	2	TcI
24	"	PET	2	ND
25	"	PET	2	TcIII
27	"	PET	4	ND
43	"	PET	1	TcIII
56	"	PET	2	TcIII
59	"	PET	1	TcI
92	"	PET	3	Hybrid
102	"	PET	3	ND
103	"	PET	8	TcI
11–123	V	PET	1	ND
18–130	V	PET	7	TcI
31–143	V	PET	3	ND
58–170	V	PET	3	ND
60–172	V	PET	2	TcIII
61–173	V	PET	2	ND
63–175	V	PET	5	TcIII
64–176	V	PET	4	TcI
65–177	V	PET	2	TcIII
69–181	V	PET	4	Hybrid
72–184	V	PET	5	TcIII
75–187	V	PET	2	TcIII
76–188	V	PET	1	TcI
78–190	V	PET	1	TcIII
14	V	Sobrante (SO)	2	TcI
15	V	SO	2	TcI
16	V	SO	1	TcI
17	V	SO	2	TcI
64	V	SO	3	Hybrid
88–200	V	San Felipe	1	TcI
	M
68	M	Calera de Tango (CT)	1	TcI
69	M		2	TcI
2–114	M	CT	3	TcIII
33–145	M	CT	9	TcI
34–146	M	CT	4	TcIII
35–147	M	CT	3	Hybrid
37–149	M	CT	3	TcI
39–151	M	CT	4	TcIII
45–157	M	CT	5	TcIII
49–161	M	CT	2	TcIII
50–162	M	CT	5	TcI
2	M	Sanatorio	5	Hybrid

^aBased on the genotype of predominant T. cruzi clones and total number of alleles, as described (the section on ‘Materials and methods’).

^bLineage assignation was performed using three different methods, as described (the section on ‘Materials and methods’). The nomenclature is according to Zingales et al. (2009): TcI (DTU I), TcII (DTU IIb), TcIII (DTU IIc), TcIV (DTU IIa), TcV (DTU IId) and TcVI (DTU IIe). Hybrid corresponds to TcV and TcVI (de Freitas et al., 2006).

T. cruzi DNA Samples from T. infestans

Samples were obtained by the dissection of the complete digestive apparatus of each triatomine. These samples were boiled, centrifuged and the supernatants were used to obtain DNA for the PCR test, as described (Venegas et al., 2010).

PCR Assay

PCR was used to detect the presence of DNA of T. cruzi and was performed in duplicate with the samples of triatomines, according to a described protocol (Dorn et al., 1999). The kinetoplast primers used for the reaction of PCR were S35 and S36 and the nuclear primers were TcZ1 and TcZ2. The products of the PCR were electrophoresed in 2% agarose gel. Gels were stained with ethidium bromide (5 μg/ml) and photographed.

Analysis of Microsatellites

Three microsatellite loci were analysed (SCLE10, SCLE11 and MCLE01) with a modification of the technique described by Oliveira et al. (1998), using a second amplification with an aliquot of 1 μl from the first PCR and the same amplification conditions. The alleles were detected using primers stained with fluorescence (Oliveira et al., 1998). The amplification products were sent to the Roy J. Carver Biotechnology Center, University of Illinois, USA, for analysis by capillary electrophoresis and fluorescence detection with an automatic sequencer and an appropriate software program. The number of base pairs (bps) of each allele was determined using cloned alleles from each marker as controls. In this analysis, the minimal detectable peak height was set to 80 arbitrary fluorescence units (FUs).

Genotyping of the Predominant T. cruzi Clone in Each DNA Sample

The determination of the predominant T. cruzi clone genotype in samples containing multiple alleles per locus was conducted taking into account the distance and the height in FU of the highest two peaks, the major and second peaks, directly in the electropherogram, as described (Venegas et al., 2010). Briefly, when the height ratio between the second and the major peak was less than 0.33, it was concluded that the predominant clone genotype in this sample has only a single allele, and therefore, is homozygous. By contrast, if the height ratio between the second and the major peak was greater than or equal to 0.33, it was analysed further, locus by locus. For instance, Fig. 1a, shows the electropherogram of the sample 24 analysed by the SCLE10 microsatellite marker. Four alleles of 207, 217, 251 and 255 bp were detected. The major and the second peaks correspond to the alleles 207 and 251 with a height of 400 and 210 FU, respectively. The ratio of the heights of the second peak divided by the major one is 0.53, but the distance between them is 44 bp. This distance is beyond the 32 bp that separates two alleles of a heterozygous clone, according to the data previously found by de Freitas et al. (2006) and Venegas et al. (2009a). For this reason, when this happens, it is necessary to look for a third allele that is within the maximum range expected. In this sample, the allele 217 meets this requirement, whose heights ratio among the 217 and 207 alleles is 0.33, just the minimum ratio accepted for a heterozygous clone. The minimum ratio was decided based on those found in heterozygous clones previously described (de Freitas et al., 2006; Venegas et al., 2009a). So, the genotype of the predominant clone of this sample is 207/217, and the minimum number of clones detected in it is 2. In the triatomine sample 25, the major and second alleles are the 207 and 217, respectively, which are within the allowed distance for heterozygous T. cruzi clone (Fig. 1b). As well as the 24 sample, the minimal number of clones detected is 2.

Fig. 1.

Electropherogram of two midgut Triatoma infestans samples analysed by the SCLE10 microsatellite marker. The panels a and b correspond to the samples code 24 and 25 from the V region. The number below each peaks are the corresponding allele in base-pair length. The numbers beside each peak represent the height of each one in fluorescence units (FU). The genotype of predominant clones of both samples was estimated as 207/217 with two minimal clones (see the section on ‘Materials and method’).

A similar analysis was conducted using the SCLE11 and MCLE01 markers with the exception that the maximum distances of separation between the major and the second peaks were 16 and 20 bp, respectively. This approach was applied with corresponding SCLE10, SCLE11 and MCLE01 locus in each T. cruzi sample (Table 1).

The minimum number of T. cruzi clones in each triatomine sample was estimated based on the genotype of the predominant clone plus the total number of alleles found in each sample divided by 2. Thus, if the predominant clone is homozygous and the sample has three additional alleles (a total of four alleles), it is concluded that the MNC is 3, because an odd number of alleles is considered as another clone. By contrast, in the same sample, if the predominant clone is heterozygous, it is concluded that the MNC is 2, as occur for samples 24 and 25 (Fig. 1).

Determination of Genetic Variation within Each Population

After the genotype of each predominant T. cruzi clone was determined in the different midgut samples of triatomines, the genetic parameters within each population were estimated (Table 2). Allele number (a) was determined using the MSA software (Dieringer and Scholötterer, 2003); the observed heterozygosis (Ho), expected heterozygosis (He) and Hardy–Weinberg equilibrium (HWE) were determined using the Guo and Thompson (1992) method and the program Arlequin (Excoffier et al., 2005).

Table 2. Microsatellite diversity of three Chilean T. cruzi population samples found in T. infestans midgut from the III, V and Metropolitan regions^a.

Population sample	Parameter	Locus
		SCLE10	SCLE11	MCLE01
III	n	12	12	12
	ab	10	10	10
	Hoc	0.000	0.083	0.384
	Hec	0.978	0.952	0.941
	HWEc	0.000*	0.000*	0.000*
V	N	35	32	33
	a	17	24	16
	Ho	0.241	0.414	0.393
	He	0.874	0.953	0.904
	HWE	0.000*	0.000*	0.000*
M	N	8	11	8
	a	7	15	11
	Ho	0.250	0.818	0.666
	He	0.766	0.9480	0.954
	HWE	0.000*	0.174	0.015*
Total number of different alleles observed		27	34	26

^a III, V and M: groups of midgut samples from 15, 37 and 12 Triatoma infestans captured in the III, V and Metropolitan Chilean regions. Samples with information on at least two loci were considered.

^bUsing MSA program (Dieringer and Scholötterer, 2003).

^cUsing Arlequin program (Excoffier et al., 2005).

n = individual number genotyped by locus; a = allele number; He = expected heterozygosis; Ho = observed heterozygosis; HWE = P value from the Hardy–Weinberg equilibrium test; asterisk = significant differences.

Analysis of Genetic Differentiation

This was conducted by three approaches: locus by locus Fisher method, multilocus analysis by the Fisher method across three loci and analysis of molecular variance (AMOVA). The genetic markers used in the three approaches were SCLE10, SCLE11 and MCLE01, described previously (Oliveira et al., 1998; de Freitas et al., 2006). The genotypic differentiation by the first and the second approach were analysed with the Genepop program using the Markov algorithm (Raymond and Rousset, 1995a, b). The AMOVA studies were performed with the Arlequin software using the genotype data determined from the predominant T. cruzi clone in each sample with unknown gametic phase, inferring the haplotypes from the distance matrix, using F_ST as genetic distance and a significant P value of genetic differences among population samples of 0.05 (Weir and Cockerham, 1984; Excoffier et al., 1992).

Lineage Determination of Predominant T. cruzi Clones

The nomenclature used in the present paper for the different lineages, according to the most recent agreement of several researchers, was: TcI (DTU I), TcII (DTU IIb), TcIII (DTU IIc), TcIV (DTU IIa), TcV (DTU IId) and TcVI (DTU IIe) (Brisse et al., 2000; Zingales et al., 2009). However, since that by several criteria, the lineages TcV and TcVI are hybrids, and by microsatellite markers, correspond to the same phylogenetic cluster, in the present paper, they were named as Hybrid (de Freitas et al., 2006).

In order to have a robust assignation of the different T. cruzi lineages, three different methods were used simultaneously with the Geneclass program: the method based on microsatellite lengths described by Goldstein et al. (1995), the method based on allelic frequencies described by Paetkau et al. (1995) and the Bayesian method described by Rannala and Mountain (1997). At least two coincidences in the same lineage with different methods were necessary to assign the final lineage of each predominant T. cruzi clone. Otherwise, the predominant clone was considered as not determined (ND, Table 1).

Correlation Analysis between Genetic and Geographical Distances

Correlation among genetic and geographical distances was estimated by the Mantel test using the Arlequin program and the linear genetic distance F_ST/(1−F_ST) (Mantel, 1967; Schneider et al., 2000).

Phylogenetic Reconstruction

In order to determine the genetic relationships between the Chilean T. cruzi sample populations and its foreign counterparts, a phylogram tree was constructed by the program PHYLIP 3.5c (Felsenstein, 1989) using the allele frequencies obtained with the SCLE10, SCLE11 and MCLE01 microsatellite loci. In order to determine the best phylogenetic reconstruction according to the published tree topology obtained with microsatellite markers (de Freita et al., 2006; Venegas et al., 2009a), three measures of genetic distances were tested. Two measurements based on the infinite allelic model, the chord distance Dc of Cavalli–Sforza and Edwards (1967) and the shared allele distance of Bowcock et al. (1994). In addition, genetic distance based on the stepwise mutation model of Goldstein et al. (1995) was also used. The ‘Neighbour-Joining’ algorithm (Saitou and Nei, 1987) was employed for tree construction and 1000 bootstrap iterations were used to test the confidence of the nodes (Felsenstein, 1989).

Published Data of Laboratory Cultured T. cruzi Clones and Strains Genotyped with Different Microsatellite Markers

This was taken from the published data of de Freitas et al. (2006) and Venegas et al. (2009a). The several T. cruzi clones and strains were grouped based on their geographic origin (Table 3) in which the majority corresponded to regions of Brazil, such as Amazon (AM) (−4°30′S, −64°41′W), Espiritu Santos (ES), Minas Gerais (−19°42′S, −44°1′W), Para (PA), Rio de Janeiro (RJ), Sao Paulo (SP), Goias (GO) (−14°53′S, −49°34′W) and Piauí (PI).

Table 3. Data previously published by de Freitas et al. (2006) and Venegas et al. (2009a) corresponding to T. cruzi clones and strains cultured in the laboratory and typed with five microsatellite markers. The groups analysed in the present manuscript are shown in the last column.

Lineagesa	Strains	SCLE10	SCLE11	MCLE01	Originb	Reference	Code: population samples studied here
	Rb1	255/259	139/139	128/136	AM/Brazil	de Freitas	TcI-Br-A
	1502	255/255	141/143	136/136	AM/"	de Freitas	"
TcI	SE	287/297	143/151	136/140	AM/"	de Freitas	"
(Z/Z)	Rb2	251/255	139/141	136/142	AM/"	de Freitas	"
	1523	255/255	143/147	144/144	AM/"	de Freitas	"
	Cutia cl1	255-255	143/143	132/132	ES/Brazil	Venegas	TcI-Br-B
	1004	253/255	139/139	130/130	MG/"	de Freitas	"
	1006	253/255	139/141	130/130	MG/"	de Freitas	"
	1001	255/255	143/143	130/130	MG/"	de Freitas	"
	SilvioX10 cl1	235/275	153/153	130/130	PA/Brazil	de Freitas	"
	D7	245/253	143/143	134/142	RJ/Brazil	de Freitas	"
	Gamba cl1	253/255	143/143	130/130	SP/Brazil	de Freitas	"
	13379 cl7	239/251	139/143	132/132	Bolivia	Venegas	TcI-Br-C
	Colombiana	235/257	139/139	136/136	Colombia	de Freitas	"
	Col18/05	253/257	139/139	136/136	Colombia	de Freitas	"
	A83	253/253	141/141	134/138	FG	de Freitas	"
	A87	253/253	141/141	142/150	FG	de Freitas	"
TcII	GOCH	271/281	153/159	128/150	GO/Brazil	de Freitas	TcII-Br-G
(Y/Y)	580	279/279	149/149	130/134	GO/"	de Freitas	"
	183744	281/281	149/149	130/134	GO/"	de Freitas	"
	578	281/281	149/155	130/148	GO/"	de Freitas	"
	581	277/277	153/161	130/148	GO/"	de Freitas	"
	577	273/281	151/155	152/154	GO/"	de Freitas	"
	Esm cl3	289/289	151/159	126/138	MG/Brazil	Venegas	TcII-Br-M
	Mas 1 cl1	271/289	151/151	130/130	MG/"	Venegas	"
	Tu 18 cl11	289/289	149/163	124/132	MG/"	Venegas	"
	GMS	273/285	149/157	126/134	MG/"	de Freitas	"
	84	267/271	153/155	126/134	MG/"	de Freitas	"
	JSM	267/275	143/151	128/128	MG/"	de Freitas	"
	Tu18 cl11	255/285	149/163	128/130	MG/"	de Freitas	"
	1043	275/281	157/157	128/134	MG/"	de Freitas	"
	169/1	273/275	149/149	128/138	MG/"	de Freitas	"
	JAF	269/275	143/149	130/134	MG/"	de Freitas	"
	1005	275/275	153/155	130/156	MG/"	de Freitas	"
	209	269/275	153/155	132/132	MG/"	de Freitas	"
	Ig539	281/281	151/153	132/150	MG/"	de Freitas	"
	JG	273/275	145/149	134/136	MG/"	de Freitas	"
	JHF	273/273	149/149	134/136	MG/"	de Freitas	"
	207	245/251	151/155	134/142	MG/"	de Freitas	"
	239	271/275	151/157	134/150	MG/"	de Freitas	"
	84Ti	271/271	143/147	136/136	MG/"	de Freitas	"
	200pm	277/277	153/155	136/136	MG/"	de Freitas	"
	Be62	273/275	155/155	136/138	MG/"	de Freitas	"
	Gil	271/271	155/157	136/144	MG/"	de Freitas	"
	803	271/281	145/145	136/152	MG/"	de Freitas	"
	MPD	267/271	143/153	138/138	MG/"	de Freitas	"
	1931	275/275	141/153	140/140	MG/"	de Freitas	"
	1014	269/275	149/149	140/142	MG/"	de Freitas	"
	CPI95/94	277/277	149/155	124/132	PI/Brazil	de Freitas	TcII-Br-X
	OPS27/94	287/297	149/159	132/136	PI/"	de Freitas	"
	GLT593	271/285	151/151	136/136	RJ/"	de Freitas	"
	GLT564	275/275	151/151	136/136	RJ/"	de Freitas	"
	Y	273/273	153/153	130/130	SP/"	de Freitas	"
TcIII	M5631 cl5	245/245	155/155	134/134	PA/Brazil	Venegas	TcIII-Br
(X/X)	M6241 cl6	255/255	151/153	130/132	PA/"	Venegas	"
	3663	245/245	155/155	128/132	AM/"	de Freitas	"
	222	215/251	151/155	132/132	MG/"	de Freitas	"
	115	245/251	153/153	132/132	MG/"	de Freitas	"
	226	245/251	153/155	132/132	MG/"	de Freitas	"
	3869	249/249	153/157	132/132	AM/" “	de Freitas	"
	231	245/251	151/155	134/134	MG/"	de Freitas	"
Hybrid	167	237/275	153/153	130/130	MG/Brazil	de Freitas	H-Bra
TcV–TcVI	1022	233/275	153/153	130/134	MG/"	de Freitas	"
(X/Y)	182	237/275	153/157	130/146	MG/"	de Freitas	"
	CLBrener	237/275	153/159	132/148	RS/Brazil	Venegas	"
	MN cl2	255/289	149/149	130/132	IV/Chile	Venegas	H-Chi
	NR cl3	255/289	149/165	128/130	Salvador/Chile	Venegas	"
	Tula cl2	289/289	149/151	130/130	IV/Chile	de Freitas	"
	Tula 1954	237/273	153/159	132/132	IV/Chile	Venegas	"

^aEach lineage is composed of T. cruzi clones and strains isolated and cultured in the laboratory using different microsatellite markers (de Freitas et al., 2006; Venegas et al., 2009a). The nomenclature used corresponds to the recent agreement: TcI (DTU I), TcII (DTU IIb), TcIII (DTU IIc), TcIV (DTU IIa), TcV (DTU IId) and TcVI (DTU IIe) (Brisse et al., 2000; Zingales et al., 2009). Hybrid corresponds to TcV and TcVI (de Freitas et al., 2006).

^bBrazil regions: Amazon (AM) (−4°30′S, −64°41′W), Espiritu Santos (ES), Minas Gerais (−19°42′S, −44°1′W), Para (PA), Rio de Janeiro (RJ), Sao Paulo (SP), Goias (GO) (−14°53′S, −49°34′W) and Piauí (PI). France Guyana (FG).

RESULTS

Genetic Variation within Each T. cruzi Population

This study was performed in order to determine whether geographic origin has any impact on the genetic characteristics of T. cruzi populations. According to this goal, T. cruzi Chilean and foreign population samples were analysed. The Chilean population samples were from triatomine midguts collected in the III, V and Metropolitan regions (see the section on ‘Materials and methods’ and Table 1). The III region has a hyper-arid climate and is separated by several hundred kilometres of desert from the V and M regions. However, both the V and M regions have similar ecological conditions with a Mediterranean climate characterised by a moderate average temperature (20°C), occasional rain during the winter and close one from another. The sampling point of III region was located 560 and 707 km from the corresponding points in the V and M regions, respectively. In addition, the sampling points in these two last regions were located at 147 km.

The foreign T. cruzi population samples correspond to published data by de Freitas et al. (2006) and Venegas et al. (2009a). These published data are composed of T. cruzi clones and strains isolated and cultured in the laboratory, mainly from regions of Brazil. Most T. cruzi samples were from the Amazon (−4°30′S, −64°41′W), Goias (−14°53′S, −49°34′W) and Minas Gerais (−19°42′S, −44°1′W) (de Freitas et al., 2006) (Table 3).

The intestinal contents of 15, 37 and 12 triatomines from the III, V and M regions positive by conventional PCR detection for T. cruzi DNA (Dorn et al., 1999), were analysed with the SCLE10, SCLE11 and MCLE01 microsatellite markers (Oliveira et al., 1998; de Freitas et al., 2006; Venegas et al., 2009a), as shown in Table 1. The genotype of the predominant T. cruzi clone was determined as described (Venegas et al., 2010) and the genetic parameters within each population were estimated (Table 2). As shown, the total number of alleles observed in the Chilean population samples for SCLE10, SCLE11 and MCLE01 markers were 27, 34 and 26, respectively, which means that the total number of possible haplotypes was 23 868. In other words, in theory, the present methodology using these three microsatellite markers has the potential resolving power to differentiate 23 868 T. cruzi clones. In addition, as expected for clonal populations almost all loci were not found in Hardy–Weinberg equilibrium, as was evidenced by the significant differences among observed and expected heterozygosis. Interestingly, thought this is the first glance, these genetic characteristics were more similar among the V and M regions than among III and V, or III and M regions. For instance, a clear difference in heterozygosis between the III and V regions and between the III and M regions was observed.

Interpopulation Analysis

In order to determine whether the three T. cruzi Chilean population samples from triatomine midgut were genetically differentiated, three analytical methods were used: a locus by locus and multilocus analysis using the Fisher’s exact test ((Raymond and Rousset, 1995a, b), and the AMOVA (Weir and Cockerham, 1984; Excoffier et al., 1992). With the first approach and the SCLE10 locus, a significant genetic differentiation was detected both between III and V and between III and M T. cruzi population samples (Table 4).

Table 4. Analysis of genotypic differentiation using the SCLE10, SCLE11 and MCLE01 loci for each Trypanosoma cruzi group pair.

Locus	Groups compareda		P valueb	SE
SCLE10c	III	V	0.00025*	0.00010
	III	M	0.02763*	0.00113
	V	M	0.65085	0.00432
SCLE11	III	V	0.29116	0.00542
	III	M	0.32907	0.00327
	V	M	0.44513	0.00487
MCLE01	III	V	0.12652	0.00313
	III	M	0.07590	0.00132
	V	M	0.06904	0.00211
Across three loci (Fisher method)
	III	V	0.000745*
	III	M	0.023990*
	V	M	0.251287

^aIII, V and MR: groups of midgut samples from 15, 37 and 12 Triatoma infestans captured in the third, fifth and Metropolitan regions, respectively. Samples with information on at least two loci were considered.

^bMarkov chain parameters: dememorisation = 10 000; Batches = 100; iterations per batch = 5000.

^cLocus by locus using the exact G test (see the section on ‘Materials and methods’).

*P<0.05.

With the second approach, the Fisher’s exact method across those three loci, a significant genetic differentiation was detected between the III and V as well as between III and M (Table 4). The third approach confirmed the result obtained with the second one, although the contribution to the variance among the populations samples was low (4.12%), the F_ST of 0.0412 was statistically significant (P = 0.018) (Table 5).

Table 5. Multilocus AMOVA analysis of predominant Trypanosoma cruzi genotypes found in the midgut of Triatoma infestans from the III, V and Metropolitan regions.

Source of variation	df	Sum of squares	Variance componentsb	Percentage of variation
Among populationsa	2	4.281	0.03560Va	4.12
Within populations	125	103.586	0.82869Vb	95.88
Total	127	107.867	0.86429
Fixation index	FST	0.04119

^aPopulations: III, V and M: groups of midgut samples from 15, 37 and 12 Triatoma infestans captured in the III, V and Metropolitan regions. Samples with information on at least two loci were considered. Molecular distance: number of different alleles (F_ST) (see the section on ‘Materials and methods’).

^bSignificance tests (1023 permutations):

V_a and F_ST: P(rand. value>obs. value) = 0.01760

P(rand. Value = obs. value) = 0.00000

P(rand. value> = obs. value) = 0.01760±0.00338.

Based on the previous results which showed that the T. cruzi population sample from the III region was genetically differentiated from its V and M counterparts, it was interesting to determine whether other biological characteristics could also be different among these T. cruzi population samples. Thus, we analysed the lineage and the minimal number of T. cruzi clones (MNC). In order to have a robust assignation of the lineages, three dissimilar methods were used (see the section on ‘Materials and methods’). The results showed that no significant differences were observed among the predominant T. cruzi lineages of the III, V and M regions (Table 6). Interestingly, the most frequent lineage in regions was TcIII, followed by TcI, while the TcII and Hybrid (TcV–TcVI) lineages were less frequent (Fig. 2). In addition, no significant statistical differences were detected among the minimal number of T. cruzi clones comparing the III, V and M regions (Fig. 3).

Table 6. Lineages of predominant Trypanosoma cruzi clones detected in the midgut of T. infestans triatomines from three Chilean regions (III, V and M).

Sample groupsa	No. (%) of T. cruzi clones found of each lineageb
	TcI	TcII	TcIII	H	NDd
III (nc = 15)	4 (26.7)	0.0	7 (46.7)	0.0	4 (26.7)
V (n = 37)	13 (35.1)	0.0	12 (32.4)	3 (8.1)	9 (24.3)
M (n = 12)	5 (41.7)	0.0	5 (41.7)	2 (16.7)	0.0
Total (n = 64)	22 (34.4)	0.0	24 (37.5)	5 (7.8)	13 (20.3)

Pair	P valuee
III/V	0.3478; 0.5	1	0.9326; 0.334	0.548	0.0312; 0.861
III/M	0.448	1	0.675; 0.795	0.188	0.106
V/M	0.683; 0.16	1	0.584	0.584	0.09

^aIII, V and M: groups of midgut samples from 15, 37 and 12 Triatoma infestans captured in the III, V and Metropolitan regions, respectively.

^bDetermined using the Geneclass program using simultaneously the methods of: Goldstein et al. (1995), Paetkau et al. (1995) and Rannala and Mountain (1997) (see the section on ‘Materials and methods’). The lineage nomenclature used corresponds to the recent agreement (Zingales et al., 2009).

^c Triatomine number.

^dND = not determined.

^eSignificant (*P<0.05). Determined by Fischer’s exact method (one value) or Chi-squared (two values).

Fig. 2.

Lineage frequency of predominant T. cruzi clones found in T. infestans midgut collected in the III, V and M regions. The lineage were determined using the microsatellite markers SCLE10, SCLE11 and MCLE01 using the Geneclass program by the methods of Goldstein et al. (1995), Paetkau et al. (1995) and Rannala and Mountain (1997) (see the section on ‘Materials and methods’). The lineage nomenclature used corresponds to the recent agreement (Zingales et al., 2009). Sample sizes were 15, 37 and 12 from III, V and M regions, respectively.

Fig. 3.

Frequency of minimum number of Trypanosoma cruzi clones detected in T. infestans midguts by three microsatellite markers in population samples from the III, V and M regions. The markers used were SCLE10, SCLE11 and MCLE01, previously described (Oliveira et al., 1998). The population samples included 15, 37 and 12 triatomine insects from the III, V and M regions, respectively (see the section on ‘Materials and methods’).

With the aim to conduct a deeper study to try to identify the possible source of the genetic differentiation among these three Chilean T. cruzi population samples, further analyses were performed with the AMOVA method and the determination of the F_ST index among all corresponding subpopulation pairs. Thus, the T. cruzi subpopulation samples within the TcI and TcIII lineages were analysed. The result showed that there was no significant differentiation among the TcI subpopulation samples from the III, V and M regions. The percentage of variation among samples was 7.35%, but this was not statistically significant (P = 0.235) (data not shown). This result was confirmed determining the corresponding F_ST among all subpopulation pairs, since all P values were superior to 0.05 (data not shown).

A significant differentiation was observed among TcIII subpopulations, as is observed by the statistically significant variation and F_ST among T. cruzi subpopulation samples, 10.41 and 0.104, respectively (P = 0.012, Table 7). It is worth mentioning that only the F_ST values among the III and V and the III and M samples were statistically significant, among samples V and M F_ST was not significant (Table 8).

Table 7. Multilocus AMOVA analysis of TcIII predominant T. cruzi genotypes found in the midgut of T. infestans from the III, V and Metropolitan regions.

Source of variation	df	Sum of squares	Variance componentsb	Percentage of variation
Among populationsa	2	4.311	0.09162Va	10.41
Within populations	45	35.501	0.78892Vb	89.59
Total	47	39.812	0.88054
Fixation index	FST	0.10406

^aPopulations III, V and M: subgroups of TcIII predominant T. cruzi clones found in the midgut of T. infestans captured in the III, V and Metropolitan regions; each subgroup have 7, 12 and 5 individuals, respectively. Samples with information on at least two loci were considered. Molecular distance: number of different alleles (F_ST) (see the section on ‘Materials and methods’).

Significance tests (1023 permutations):

V_a and F_ST: P(rand. value>obs. value) = 0.01173

P(rand. value = obs. value) = 0.00000

P(rand. value> = obs. value) = 0.01173±0.00389.

Table 8. Genetic distances F_ST and corresponding P values among TcIII Trypanosoma cruzi population samples from the III, V and M Chilean regions.

FST
Regionsa	III	V	MR
III	0.00000
V	0.14036	0.00000
MR	0.14829	0.02028	0.00000
FST P valuesb
III	…
V	0.00879*	…
M	0.01074*	0.58008	…

^aPopulation samples: III, V and M: predominant T. cruzi clones found in the midgut of 7, 12 and 5 Triatoma infestans captured in the third, fifth and Metropolitan Chilean regions, respectively. Molecular distance: number of different alleles (F_ST) (see the section on ‘Materials and methods’).

^bSignificant values with 1023 permutations (*P<0.05).

Genetic Relationship of Chilean T. cruzi Populations with Their Counterparts from Some Other Latin American Regions

In order to determine whether the Chilean T. cruzi population samples were genetically differentiated from their counterparts in other countries, AMOVA analyses and their corresponding F_ST distances among pairs of subpopulations were performed. To obtain this goal, the first step was to determine whether the T. cruzi population samples available from the literature could be genetically differentiated within each lineage. Therefore, T. cruzi clones and strains from the published data of de Freitas et al. (2006) and Venegas et al. (2009a) were separated by geographical origin. Three groups within the TcI (TcI-A, TcI-B and TcI-C), three groups within TcII (TcII-Br-G, TcII-Br-M and TcII-Br-X), one group of TcIII (TcIII-Br) and two groups of TcV–TcVI (H-Bra and H-Chi) were produced (Table 3). The results showed that within TcI as well as the TcV–TcVI lineages, T. cruzi subpopulation samples are genetically differentiated, as is evidenced by their statistically significant F_ST P values (P<0.05) (Tables 9 and 10, respectively). In contrast, no significant differentiation was detected among the TcII T. cruzi subpopulation samples (data not shown).

Table 9. Multilocus AMOVA analysis of TcI T. cruzi subpopulations from Brazil.

Source of variation	df	Sum of squares	Variance componentsb	Percentage of variation
Among populationsa	2	6.847	0.21309Va	16.98
Within populations	31	32.300	1.04194Vb	83.02
Total	33	39.147	1.25503
Fixation index	FST	0.16979

^aPopulations: TcI Brazilian subpopulations TcI-Br-A, TcI-Br-B and TcI-Br-C (Table 3). Distance: number of different alleles (F_ST) (see the section on ‘Materials and methods’).

Significance tests (1023 permutations):

V_a and F_ST: P(rand. value>obs. value) = 0.00196

P(rand. value = obs. value) = 0.00000

P(rand. value> = obs. value) = 0.00196±0.00136.

Table 10. Multilocus AMOVA analysis of TcV–TcVI T. cruzi subpopulations from Brazil and Chile.

Source of variation	df	Sum of squares	Variance componentsb	Percentage of variation
Among populationsa	1	2.938	0.23772Va	18.67
Within populations	14	14.500	1.03571Vb	81.33
Total	15	17.438	1.27344
Fixation index	FST	0.18668

^aPopulations: TcV–TcVI T. cruzi subpopulations H-Bra and H-Chi (Table 3). Distance: number of different alleles (F_ST) (see the section on ‘Materials and methods’).

Significance tests (1023 permutations):

V_a and F_ST: P(rand. value>obs. value) = 0.00000

P(rand. value = obs. value) = 0.02444

P(rand. value> = obs. value) = 0.02444±0.00495.

AMOVA analysis and the corresponding F_ST determinations within the TcI lineage showed that the Chilean subpopulation from the III region is not differentiated from any Brazilian counterpart, since in all pairs of comparisons the P values were superior to 0.05 (data not shown). On the contrary, both the V and M Chilean T. cruzi subpopulation samples are differentiated from the Brazilian TcI-Br-B sample (P<0.05).

The AMOVA and F_ST determinations of the Chilean and Brazilian TcIII subpopulation samples showed that in all pairs of comparisons, the Chilean samples are significantly differentiated from their Brazilian counterparts, since the P values of F_ST for each subpopulation pair is much less than 0.05 (Table 11).

Table 11. Genetic distances F_ST and corresponding P values among TcIII Trypanosoma cruzi population samples from three Chilean regions and one Brazilian region.

FST
Regionsa	III	V	M	Brazil
III	0.00000
V	0.14036	0.00000
M	0.14829	0.02028	0.00000
Brazil	0.17918	0.17148	0.16412	0.00000
FST P valuesb
III	…
V	0.00684*	…
M	0.01758*	0.60254	…
Brazil	0.00195*	0.00000*	0.00195*	…

^aPopulation samples: III, V and M: predominant T. cruzi clones found in 7, 12 and 5 Triatoma infestans captured in the III, V and Metropolitan Chilean regions. A fourth population sample of eight Brazilian T. cruzi clones mainly from the Minas Gerais region (MG) was included. Molecular distance: number of different alleles (F_ST) (see the section on ‘Materials and methods’).

^bSignificant values with 1023 permutations (P<0.05).

Phylogenetic Analysis of Chilean and Foreign T. cruzi Subpopulation Samples

To estimate the phylogenetic relationships among the Chilean and foreign T. cruzi subpopulation samples, phylogenetic trees were constructed based on four different genetic distances; the Cavalli–Sforza and Edwards (1967) chord distance, the Nei et al. (1983) distance, the Bowcock et al. (1994) distance based on the proportion of shared alleles and the Goldstein et al. (1995) distance based on the stepwise mutation model. According to the known T. cruzi phylogeny, the best result was obtained with the Bowcock method (Fig. 4), which is characterized by the main monophyletic clusters TcI and TcII located at the extremes of the tree. In addition, as indicated in the literature (Westenberger et al., 2005; de Freitas et al., 2006; Venegas et al., 2009a), the hybrid lineage (TcV–TcVI) is located between the nodes of TcIII and TcII. In agreement with the results of AMOVA and F_ST determinations, the three TcI Chilean T. cruzi subpopulation samples are located in the TcI monophyletic node, while the TcIII Chilean subpopulation samples are located in two separate clusters: on the one hand, the TcIIICh3 sample forms an internal node with the single sample from Brazil TcIII-Br, whereas the TcIIICh5 and TcIIIChM Chilean samples form another cluster with the unknown ND-V sample. Interestingly, the hybrid sample including T. cruzi clones from the V and M regions (H-Ch5M) is grouped together with the Chilean laboratory cultured T. cruzi clones belonging to TcV–TcVI lineage (H-Chi).

Fig. 4.

Unrooted phylogenetic tree of the T. cruzi population samples from the midgut of triatomines from Chilean and Brazilian regions. The phylogenetic reconstruction was performed using the allele sharing distance method (Bowcock et al., 1994) and the SCLE10, SCLE11 and MCLE01 microsatellite markers (Oliveira et al., 1998). The T. cruzi population samples from triatomine midgut were directly genotyped and correspond to groups of insects collected in the III (TcI-Ch3 and TcIII-Ch3), V (TcI-Ch5, TcIII-Ch5 and ND-Ch5) and Metropolitan (TcI-ChM and TcIII-ChM) Chilean regions. The lineage nomenclature corresponds to the recent agreement (Zingales et al., 2009); they were assigned based on three simultaneous methods and correspond to the following T. cruzi population samples: TcI lineage (TcI-Ch3, TcI-Ch5 and TcI-ChM) and TcIII lineage (TcIII-Ch3, TcIII-Ch5 and TcIII-ChM). H-Ch5M corresponds to a T. cruzi population sample of the TcV–TcVI lineage from triatomine midguts collected in the V and Metropolitan Chilean regions. The laboratory cultured T. cruzi population samples correspond to published data of de Freita et al. (2006) and Venegas et al. (2009a) and are the following: TcI from Brazil Amazon (TcI-Br-A), TcI mainly from Brazil-Minas Gerais (TcI-Br-B), TcI from Bolivia, Colombia and France Guyana (TcI-C), TcII from Brazil-Goias (TcII-Br-G), TcII from Brazil-Minas Gerais (TcII-Br-M), TcII from several Brazilian regions (TcII-Br-X), TcIII mainly from Minas Gerais (TcIII-Br), TcV–TcVI from Minas Gerais (H-Bra) and TcV–TcVI from Chile (H-Chi) (see Table 3 and the section on ‘Materials and methods’ for more details).

Correlation among Chilean and Brazilian TcIII Subpopulations

In order to determine whether the geographic distances have played an important role in the differentiation of these T. cruzi subpopulation samples, we performed the Mantel test (Mantel, 1967). A correlation coefficient of 0.71 was obtained with a significant P value of 0.038.

Table 12 shows the matrices of lineal genetic distance and geographical distance of the Chilean and Brazilian TcIII T. cruzi subpulations.

Table 12. Matrix of Slatkin linearized F_ST genetic distance [F_ST/(1−F_ST)] and geographic distances of TcIII Trypanosoma cruzi population samples from three Chilean regions and one Brazilian region.

FST/(1−FST)a
Regionsb	III	V	M	Brazil
III	0.00000
V	0.16328	0.00000
M	0.17411	0.02070	0.00000
Brazil	0.21829	0.20697	0.19635	0.00000
Geographic distance (km)b
Regions
III	0
V	560	0
M	707	147	0
Brazil	4056	4616	4763	0

^aReference: Slatkin et al. (1995).

^bPopulation samples: III, V and M: predominant T. cruzi clones found in 7, 12 and 5 Triatoma infestans captured in the III, V and Metropolitan Chilean regions, respectively. A fourth population sample of eight Brazilian T. cruzi clones mainly from the Minas Gerais region (MG) was included. The geographic distances were determined as the smallest distance between two regions with the Google heart online map (see the section on ‘Materials and methods’).

DISCUSSION

The evidence presented in this study strongly suggests that the Chilean T. cruzi populations from the midgut of domestic T. infestans are geographically differentiated. The following evidences support this conclusion.

First, the significant differences detected with the SCLE10 marker among the T. cruzi populations from the III and V regions as well as between the III and Metropolitan regions. It is very interesting that using only a single locus, it is possible to detect significant genotypic differentiation, although methods based on a single locus have less power of resolution than those based on multiple loci. One possible explication that SCLE10 could detect this differentiation may be the high polymorphism we detected, as is illustrated by its 27 total alleles. Considering that T. cruzi is meanly a diploid organism (Tibayrenc et al., 1986; Oliveira et al., 1998; Brisse et al., 2000), this means that using this single SCLE10 marker, in theory, it is possible to distinguish n(n−1)/2 = 27(27−1)/2 = 351 different combinations or genotypes, which may explain why it was possible to detect subtle genetic differences between these two Chilean T. cruzi population samples.

Second, the significant genetic differences detected by the multilocus Fisher’s exact method, using the microsatellite markers SCLE10, SCLE11 and MCLE01 between the III and V regions, as well as between the III and M regions of T. cruzi population samples. This result confirms the above evidence using a more powerful multilocus method. Considering that the total alleles observed with each microsatellite marker (27, 34 and 26 with SCLE10, SCLE11 and MCLE01, respectively), the total number of different haplotypes would be 27×34×26 = 23 868, in absence of linkage disequilibrium, which could explain the resolving power to detected genetic differences among these T. cruzi populations, apparently much higher than the resolving power of isoenzymes, traditionally used in the study of T. cruzi populations (Tibayrenc et al., 1986; Tibayrenc et al., 1990; Brisse et al., 2000).

Third, the AMOVA of haplotype distribution among the T. cruzi population samples from the III and V regions, and from III and M regions. This result reinforces that the genetic differentiation between the III and the V and M regions most probably is real and not due to a fortuitous event or an experimental artefact, because the AMOVA and the Fisher’s exact methods are based on dissimilar principles: the first on haplotype variance and the second on genotype frequences (Raymond and Rousset, 1995a, b).

Fourth, the determination of the genetic distance F_ST among the T. cruzi population samples from III, V and M regions. The fact that the F_ST values comparing III and V, and III and M T. cruzi population samples were statistically significant reinforced the first observations which showed that the T. cruzi population from the III region is genetically differentiated from its V and M counterparts. It is worth mentioning that important different geographic characteristics exist between the two triatomine population niches formed by the III region and V–M. They are separated by about 700 km and have dissimilar geographic, climatic and demographic conditions (Apt et al., 1987). These different ecological niches also are associated with different peridomestic mammals as possible reservoirs (Rozas et al., 2007). We think that these ecological and geographical factors of T. cruzi have favoured the conditions for population structuring, originating two dissimilar genetic T. cruzi subpopulations. On the other hand, the V and M regions are separated by only 150 km and they have similar climate and a high rate of human migration; for these reasons, it is very probable that these two regions may really constitute a single ecotope and in consequence, have a single T. cruzi ‘deme’ or subpopulation.

In order to determine whether the genetic differentiation detected between the T. cruzi population samples III and V or between III and M is manifested in another genetic characteristic, we studied the distribution of clones in these samples. Unexpectedly, no significant difference in distribution was observed. However, it is worth recalling the high percentage of multiclonal T. cruzi samples observed in all populations, illustrated by the presence of more than one T. cruzi clone in each triatomine midgut sample. This evidence is concordant with the previous results in showing that in the triatomine intestine the T. cruzi parasites have a more favourable environment for growth and development than in the highly selective environment of the mammalian host (Oliveira et al., 1998; Coronado et al., 2006; Venegas et al., 2010).

At first glance, analysing simultaneously the distribution of the different T. cruzi lineages among the III, V and M regions, no statistically significant dissimilarities were detected. However, it is important to mention that the lineage distributions compared here correspond to lineages of predominant T. cruzi clones; therefore, we cannot discard that the distribution of minor clones may have greater differences among these T. cruzi populations samples or that there may exist subpopulations within each lineages with asymmetric genetic distribution. To determine whether this latter possibility could be occurring in the populations studied here, we proceeded to perform AMOVA studies and determinations of F_ST between subgroups or subpopulations consisting of a single lineage. The results obtained showed that statistically significant dissimilar haplotype distributions and F_ST values were found comparing TcIII subpopulation samples either between the III and V regions or between the III and M regions. However, no significant differences were found comparing the same subpopulation pairs only containing TcI individuals. This result not only confirms again the fact that the III region T. cruzi population is genetically different from its V and M region counterparts, but also strongly suggests that the III population of T. cruzi is a discrete entity, physically separated from the other two southern regions, and that this genetic differentiation is only due to the asymmetrical distribution of TcIII T. cruzi clones. To our knowledge, this is the first time that in Chile, it has been shown by population genetics methods that T. cruzi populations really are structured, and not only between lineages; we also demonstrated differential distributions within the same lineage. It is true that in other previous studies, analysing kDNA with restriction enzymes or/and with minicycle probes, some dissimilar distributions of lineages among different localities were observed. Nevertheless, in none of these studies, statistical analyses were conducted to support their observations, and they were mainly based on the analysis of laboratory cultured parasites (Sanchez et al., 1993; Solari et al., 2001; Torres et al., 2004). In this regard, there is a very interesting recent report from Brazil which used several microsatellite markers to show that between different regions, there is subtle but statistical significant genetic differentiation among TcIII T. cruzi populations and that this phylogeographic differentiation could have some impact in clinical manifestations of Chagas disease (Llewelyn et al., 2009a, b).

The very low frequency of the hybrid lineage TcV–TcVI detected in all the population samples studied here (7.8%) was similar to that found in a previous study of our group analysing T. cruzi samples from chronic chagasic patients (Venegas et al., 2010). However, this low percentage of the hybrid lineage is markedly different than published reports of T. cruzi stocks from triatomines, where 21–51% of stocks were characterized as TcIId, corresponding to the old Z2bol nomenclature and the current hybrid TcV lineage (Sanchez et al., 1993; Venegas et al., 1997; Torres et al., 2004; Zingales et al., 2009). Some of the causes which could explain this dissimilarity may be linked to the fact that here we are only typing the predominant T. cruzi clone found in each triatomine sample; therefore, minor T. cruzi clones such as the hybrid ones may not be detected. In addition, in the present paper, it was used as a direct method to type T. cruzi clones without previous isolating and culturing, as it was done in the cited literature (Sanchez et al., 1993; Venegas et al., 1997; Torres et al., 2004). Another explanation to this dissimilarity could be related with the differences among the nature of molecular makers (Junqueira et al., 2005; Coronado et al., 2006; Tillerias et al., 2006; Rozas et al., 2007; Venegas et al., 2009b).

The geographic distances separating these host insects may be a relevant factor involved in genetic differentiation of Chilean T. cruzi populations from triatomine midguts. The evidence which supports this suggestion is the correlation study using the Mantel test (Mantel, 1967; Schneider et al., 2000). Significant statistical correlation (correlation coefficient r = 0.71 and P = 0.038) was found between the Slatkin linearized F_ST genetic distance [F_ST/(1–F_ST)] and the geographic distances of the III, V, M and Brazil TcIII subpopulation samples. This result is concordant with recent studies analysing T. cruzi TcIII isolates from the north and centre of Latin America with a large panel of 49 microsatellite markers and sequences of the glucose-6-phophate isomerase locus (Llewellyn et al., 2009a). In this study, a significant statistical correlation was found between genetic and geographical distances among T. cruzi isolates from four localities of Latin America made up of population samples from the north (Brazil, Venezuela and Colombia), north of Bolivia, south of Bolivia and Paraguay.

With the aim to further obtain more information about the genetic relationship among the studied Chilean and foreign population samples, we conducted phylogenetic studies using different genetic distances. The best result, based on the expected topology and the high polymorphism of the three microsatellite markers used, was found using the infinite allelic model as genetic distance which corresponded to the shared genetic distance (Dc) and Neighbour-Joining method (Cavalli–Sforza and Edwards, 1967; Saitou and Nei, 1987). The tree topology was very similar to the described topology using five microsatellite markers, characterized by the most divergent TcI and TcII lineages at the each end of tree and the hybrid lineage node surrounded by the nodes of TCII and TCIII (de Freitas et al., 2006; Venegas et al., 2009a). The tree also reflects, shown by population genetics studies, the more homogenous nature of the TcI lineages from the III, V and M regions, because all these subpopulation samples were grouped in the same node. However, as expected, the TcIII from the III region was grouped in a distinct node in relation to the node in which the subpopulations in V and M regions were clustered. This phylogenetic construction strongly supports and validates the above population genetic results, giving more evidence that these results are probably not due to a biased sample or experimental error.

CONCLUSIONS

Taken together, the evidence shown in the present paper strongly supports the idea that the T. cruzi populations really are geographically structured in Chile and in other localities of Latin America. Second, in Chile, this geographical differentiation would be due to a dissimilar distribution of TcIII T. cruzi clones, at least in some regions. In addition, it was found that the genetic distance of TcIII populations is correlated with geographical distances among Chilean and foreign populations. The fact that the T. cruzi populations are geographical structured not only among lineages but even within one lineage could have relevant impact on the epidemiological and pathogenic characteristic of Chagas disease. These interesting aspects should be further investigated.