DNA barcoding does not separate South American Triatoma (Hemiptera: Reduviidae), Chagas Disease vectors.

BACKGROUND: DNA barcoding assumes that a biological entity is completely separated from its closest relatives by a barcoding gap, which means that intraspecific genetic distance (from COI sequences) should never be greater than interspecific distances. We investigated the applicability of this strategy in identifying species of the genus Triatoma from South America.
FINDINGS: We calculated intra and interspecific Kimura-2-parameter distances between species from the infestans, matogrossensis, sordida and rubrovaria subcomplexes. In every subcomplex examined we observed at least one intraspecific distance greater than interspecific distances.
CONCLUSIONS: Although DNA barcoding is a straightforward approach, it was not applicable for identifying Southern American Triatoma species, which may have diverged recently. Thus, caution should be taken in identifying vector species using this approach, especially in groups where accurate identification of taxa is fundamentally linked to public health issues.

Findings

DNA barcoding, as proposed by Hebert et al. [1] assumes that a biological entity is completely separated from its closest relatives by a barcoding gap [2], which means that intraspecific genetic distances (from COI sequences) are never greater than interspecific distances.

Triatoma Laporte (Hemiptera: Reduviidae) is the most diverse genus of Chagas Disease vectors, and accurate identification of species is imperative for the efficiency of vector control programs. The Triatoma genus is divided into species complexes and subcomplexes according to geographic distribution and morphological similarity [3].

Recently, Justi et al. [4] reported that the relationships between species assigned to South American Triatomasubcomplexes could not be untangled with the data in hand. We were then prompted to investigate whether DNA barcoding would be a useful tool for identifying the species within the infestans, matogrossensis, sordida and rubrovaria subcomplexes [3].

Kimura-2-parameter genetic distances [5] were calculated pairwise within each of the above mentionedsubcomplexes (Table 1) using the software MEGA v. 5 [6], and intra and interspecific distances were compared.

Table 1

K2p-distances between species of the subcomplexes studied

Subcomplex	GenBank	Number	Geographic Origin
infestans					1	2	3	4	5
	KC249330	1	Chaco Tita, Cochabamba, Bolivia	T. delpontei 53
	KC249346	2	Chaco Tita, Cochabamba, Bolivia	T. infestans 44	0.021
	KC249349	3	Cotapachi, Cochabamba, Bolivia	T. infestans 58	0.025	0.018
	KC249352	4	Mataral, Cochabamba, Bolivia	T. infestans 60	0.025	0.018	0.005
	KC249354	5	Ilicuni, Cochabamba, Bolivia	T. infestans 63	0.021	0.016	0.000	0.006
	KC249355	6	Montevideo, Uruguai	T. infestans 69	0.072	0.061	0.064	0.069	0.103
matogrossensis					7	8	9	10	11	12
	KC249327,KC249328	7	Posse, GO, Brazil	T. costalimai 35
	KC249329	8	Chiquitania, Cochabamba, Bolivia	T. costalimai 42	0.154
	KC249360	9	São Gabriel D'oeste, MS, Brazil	T. matogrossensis 192	0.134	0.138
	KC249361	10	Bahia, Brazil	T. matogrossensis 31	0.151	0.152	0.047
	KC249391	11	Pantanal, MT, Brazil	T. vandae 28	0.156	0.151	0.047	0.040
	KC249392	12	Rio Verde do MatoGrosso, MT, Brazil	T. vandae 73	0.138	0.146	0.005	0.046	0.045
	KC249393,KC249394	13	Rondonópolis, MT, Brazil	T. vandae 74	0.158	0.150	0.048	0.059	0.007	0.052
rubrovaria					14	15	16	17	18	19	20	21	22	23	24
	KC249322	14	São Gerônimo, RS, Brazil	T. carcavalloi 78
	KC249323	15	Caçapava do Sul, RS, Brazil	T. circummaculata 120	0.039
	KC249324	16	Sítio Faxina, Piratini, RS, Brazil	T. circummaculata 121	0.029	0.025
	KC249325	17	Sítio Faxina, Piratini, RS, Brazil	T. circummaculata 122	0.017	0.039	0.033
	KC249356	18	Nova Petrópolis, RS, Brazil	T. klugi 75	0.018	0.037	0.031	0.017
	KC249369	19	Sítio Faxina, Piratini, RS, Brazil	T.rubrovaria 123	0.055	0.023	0.029	0.055	0.057
	KC249370	20	Sítio venda da Lagoa, Canguçu, RS, Brazil	T.rubrovaria 134	0.065	0.052	0.065	0.065	0.061	0.070
	KC249372	21	SítioFaxina, Pinheiro Machado, RS, Brazil	T.rubrovaria 136	0.042	0.019	0.027	0.036	0.036	0.031	0.035
	KC249373	22	Sítiovenda da Lagoa, Canguçu, RS, Brazil	T.rubrovaria 140	0.038	0.020	0.019	0.043	0.040	0.029	0.032	0.012
	KC249374	23	Canguçu, RS, Brazil	T.rubrovaria 156	0.039	0.020	0.019	0.045	0.042	0.029	0.032	0.012	0.000
	KC249375	24	Caçapava do Sul, RS, Brazil	T.rubrovaria 76	0.021	0.029	0.021	0.016	0.016	0.033	0.074	0.034	0.038	0.038
	KC249376	25	Quevedos, RS, Brazil	T.rubrovaria 77	0.029	0.030	0.043	0.022	0.029	0.046	0.065	0.031	0.046	0.048	0.026
sordida					26	27	28	29	30	31	32	33	34
	KC249338	26	Rivadaria, Argentina	T. garciabesi 89
	KC249342	27	Santa Cruz, Bolívia	T. guasayana 55	0.077
	KC249343	28	Santa Cruz, Bolívia	T. guasayana 82	0.065	0.056
	KC249379,KC249380	29	Romerillo, Cochabamba, Bolivia	T. sordida 46	0.029	0.060	0.060
	KC249381,KC249382	30	Romerillo, Cochabamba, Bolivia	T. sordida 47	0.030	0.061	0.061	0.000
	KC249383	31	La Paz, Bolívia	T. sordida 83	0.081	0.013	0.063	0.066	0.066
	KC249384	32	Pantanal, MS, Brazil	T. sordida 85	0.069	0.012	0.062	0.065	0.065	0.025
	KC249385	33	Santa Cruz, Bolívia	T. sordida 86	0.043	0.082	0.074	0.035	0.035	0.073	0.082
	KC249387	34	San Miguel Corrientes, Argentina	T. sordida 88	0.061	0.058	0.063	0.070	0.071	0.058	0.055	0.052
	KC249388	35	Poconé, MT, Brazil	T. sordida 90	0.069	0.017	0.058	0.075	0.075	0.031	0.011	0.078	0.051

Highlighted distances deviate from the DNA barcoding premis that intraspecific distances are smaller than interspecific distances.

In all subcomplexes we observed at least one intraspecific distance greater than interspecific distances (Table 1). To be considered appropriate to identify species within a group, intraspecific distances must always be greater than interspecific ones [2], and therefore DNA barcoding is not accurate for the species-level identification of South American Triatoma. Moreover, the method fails to account for hybridization events, which are naturally observed in Triatoma [7,8], and introgression, which is frequent in nuclear DNA [9]. These considerations argue that Hebert et al.’s [1] proposal of cataloguing biodiversity based only on DNA barcoding may severely underestimate it.

Besides that, as highlighted by Dujardin et al. [10], the morphological changes observed in closely related “species”, or “lineages” as we prefer to call them, may have led taxonomists to rush into describing subspecies or species, even genera. Molecular phylogenetic studies are in their infancy in unravelling the evolution of Triatominae, and a comprehensive molecular phylogeny, including more than one specimen for most lineages, was published only in 2014 [4], although several analyses were conducted focusing on small species groups. Taken together, these statements make it clear that further investigations of Triatominae evolution are long overdue, preferably integrating morphological, molecular and ecological data.

Lineage evolution has not occurred, but it is happening now. Concerning lineages designated in the infestans complex (including the subcomplexes studied here), separation is much clearer in terms of morphology than in molecular systematics. In cases where lineages have not reached reciprocal monophyly, defining taxonomic entities is not a straightforward issue [11]. Therefore caution is necessary, especially in a group where accurate identification of taxa is fundamentally linked to public health issues.

Conclusions

Although DNA barcoding is a straightforward approach, it was not applicable for identifying Southern American Triatoma species, which may have diverged recently. Thus, caution should be taken in identifying vector species using this approach, especially in groups where accurate identification of taxa is fundamentally linked to public health issues.