A multiplex PCR assay that separates Rhodnius prolixus from members of the Rhodnius robustus cryptic species complex (Hemiptera: Reduviidae)

Authors


Corresponding Author Fernando Monteiro, Laboratório de Doenças Parasitárias, Departamento de Medicina Tropical, Instituto Oswaldo Cruz, Avenida Brasil 4365, Rio de Janeiro, RJ, Brasil 21045-900. Tel./Fax: +55 21 2280 3740; E-mail: fam@ioc.fiocruz.br

Summary

Rhodnius prolixus is one of the most important primary vectors of human Chagas disease in Latin America. Its morphology is, however, identical to that of the members of the Rhodnius robustus cryptic species complex, which includes secondary vectors. The correct identification of these taxa with differential vector competence is, therefore, of great epidemiological relevance. We used the alignment of 26 mitochondrial cytochrome b haplotypes (663 bp) to select for PCR-amplifiable species-specific regions. We designed one forward primer on a region conserved across all haplotypes, and three reverse primers that anneal to species-specific regions and amplify fragments of different lengths for R. prolixus (285 bp) and for members of the two major R. robustus lineages: group I (349 bp) and groups II–IV (239 bp). These fragments were easily identifiable on regular 1.5% agarose gels. This multiplex PCR assay was successfully tested on 81 specimens from six Latin American countries, and used to determine the phylogeographic boundaries for each species. It is a simple, objective, and cost-effective assay. Its PCR-based nature makes it applicable to any insect developmental stage, as well as to dried specimens, and insect remains. It should be particularly useful in areas where representatives of these Rhodnius species occur in sympatry.

Abstract

Rhodnius prolixus est l'un des plus importants vecteurs primaires de la maladie de Chagas humaine en Amérique latine. Sa morphologie est cependant identique à celle des membres du complexe cryptique des espèces R. robustus qui comprend des vecteurs secondaires. L'identification correcte de ces taxa par la compétence différentielle du vecteur est dès lors de grande importance épidémiologique. Nous avons utilisé l'alignement de 26 séquences haplotypes du cytochrome b mitochondrial (663 pb) afin de sélectionner des régions spécifiques à l'espèce, amplifiables par PCR. Nous avons conçu une amorce directe sur une région conservée chez tous les haplotypes et trois amorces inverses s'hybridant à des régions spécifiques à l'espèce et amplifiant des fragments de différentes tailles pour R. prolixus (285 pb) et pour les membres des deux lignées principales de R. robustus: groupe I (349 pb) et groupes II à IV (239 pb). Ces fragments étaient facilement identifiables sur gel d'agarose classique à 1,5%. Cette analyse par PCR multiplex a été examinée avec succès sur 81 spécimens provenant de six pays latino-américains et utilisée pour déterminer les limites phylogéographiques pour chacune des espèces. Il s'agit d'une analyse simple, objective et rentable. Le fait d’être basée sur la PCR la rend applicable à n'importe quel stade de développement, aux spécimens séchés ou même à des restes d'insecte. Il devrait être particulièrement utile dans les secteurs où les représentants de ces espèces Rhodnius sévissent en sympatrie.

Abstract

Rhodnius prolixus es uno de los vectores primarios más importantes de la enfermedad de Chagas en Latinoamérica. Su morfología, sin embargo, es idéntica a la de otros miembros del complejo críptico de especies R.robustus, que incluye vectores secundarios. La identificación correcta de este taxón con competencia diferencial del vector es por lo tanto de una gran relevancia epidemiológica. Hemos utilizado la alineación de 26 haplotipos del citocromo b mitocondrial (663pb) para seleccionar regiones especie específicas amplificables por PCR. Diseñamos un cebador directo en una región conservada en todos los haplotipos, y tres cebadores reversos que anidaban con regiones especie-específicas y amplificaban fragmentos de diferentes tamaños para R. prolixus (285pb) y para miembros de los dos linajes principales de R. robustus: grupo I (349pb) y grupos II-IV (239pb). Estos fragmentos eran fácilmente identificables en geles de agarosa al 1.5%. Este ensayo de PCR multiplex fue utilizado con éxito en 81 especimenes de seis países Latinoamericanos, así como para determinar las fronteras filogeográficas de cada especie. Se trata de una prueba simple, objetiva y costoefectiva. La naturaleza del ensayo, basado en una PCR, la convierte en aplicable a cualquier estado de desarrollo de cualquier insecto, a especimenes secos e incluso a restos de insectos. Sería particularmente útil en áreas en las que los representantes de esta especie de Rhodnius están en simpatría. Utilizamos la alineación de 26 haplotipos del citocromo b (663pb) para seleccionar regiones especie-específicas amplificables mediante PCR.

Introduction

Human Chagas disease is transmitted mainly as a consequence of contact with Trypanosoma cruzi infected feces of domestic populations of triatomine bugs (Hemiptera: Reduviidae). As disease treatment remains impractical on a large scale and vaccines are unavailable, reduction in disease transmission has been achieved by targeting the vectors. However, despite the recent success of multinational initiatives launched with the purpose of controlling vector-borne transmission (as well as that by infected blood transfusion), recent estimates show that substantial efforts are still necessary: overall prevalence remains over 12 million, with 200 000 new cases/year in 15 Latin-American countries (Morel & Lazdins 2003).

Rhodnius prolixus is presently one of the most important Chagas disease vector in Latin America, particularly in Venezuela, Colombia, and Central America. Aiming at reducing disease transmission in those areas, the multinational Andean Pact and Central America initiatives were launched in 1997 (WHO 2002). However, as the planning and monitoring of modern control interventions must rely on accurate taxonomic judgments (Abad-Franch & Monteiro 2005), one of the current challenges that Chagas disease vector control still faces is the distinction of cryptic species and the devising of simple methods for their identification (Miles et al. 2003). This is important because vectorial capacity often varies greatly even among closely related taxa. Rhodnius prolixus and Rhodnius robustus are a good example of this particular problem as only R. prolixus is believed to be an efficient vector. However, morphological and genetic similarities have led to uncertainties on their taxonomy (summarized in Monteiro et al. 2001), which, in turn, have hampered the precise assessment of their epidemiological roles.

Recent molecular data have confirmed that not only is R. prolixus a valid taxon, but that R. robustus is actually a complex of four cryptic species (which were provisionally referred to as R. robustus I, II, III, and IV) that occur in northern South America (Monteiro et al. 2003).

Since, to date, there are no reliable morphological characters to discriminate among these closely related taxa, we describe a method based on a single multiplex PCR reaction, for distinguishing among R. prolixus, R. robustus I, and the group R. robustus II–IV. This approach has been successfully used in other medically important arthropod groups such as Cecidophyopsis mites (Lava Kumar et al 1999), Anopheles mosquitoes (Phuc et al. 2003, Kengne et al. 2003), and Cuterebra flies (Nöel et al. 2004).

Materials and methods

Samples used

Twenty-five field-collected insects were obtained from palm trees in Venezuela, French Guyana and Brazil, and four insects came from a colony originated with insects from Tocantins, Brazil (Table 1, Figure 1). Insect collection was performed by either felling trees or through the use of live-bait traps (Abad-Franch et al. 2000). Insects were identified to the ‘species complex’ level (i.e. R. prolixus plus the four R. robustus forms; Monteiro et al. 2003) by morphology (Lent & Wygodzinsky 1979), and further within the complex, genetically, as described below.

Table 1.   Information on the samples used in this study
SpeciesCollection siteGeographical CoordinatesN multiplexedN sequencedField/ colonyDate of ColectionHaplotype
  1. *Sequenced in this study (29 samples in total). **New haplotypes (v through A, accession numbers EF011723-EF011728, respectively). Other samples were previously sequenced (Monteiro et al. 2003).

R. prolixus1 Orica, Francisco Morazan, Honduras15° 28′ N 86° 11′ W17field1999a
2 Las Palmas, Guatemala14° 26′ N 88° 35′ W16fieldJune 1995b
3 Tituque, Guatemala14° 40′ N 89° 23′ W-9fieldJune 1995b
3 Tuticopote, Guatemala14° 43′ N 89° 21′ W-3fieldJune 1995b
4 Modesto Loaiza, Coyaima, Colombia03° 46′ N 75° 12′ W-1colonySeptember 1996b
5 Ibague, Colombia04° 26′ N 75° 14′ W-1colonyFebruary 1995b
6 Pampanito, Trujillo, Venezuela09° 24′ N 70° 30′ W73colony1997c
6 Pampan, Trujillo, Venezuela09° 26′ N 70° 28′ W51colony1987c
7 Pampanito, Trujillo, Venezuela09° 50′ N 70° 30′ W62colony1960c
8 São José Tiznados, Guárico, Venezuela09° 23′ N 67° 33′ W83colony1988b
9 Ortiz, Guárico, Venezuela09° 37′ N 67° 17′ W-4fieldJuly 2001b
10 Cojedes, Venezuela09° 37′ N 68° 55′ W-1colony1995b
11 Barinas, Venezuela08° 37′ N 70° 12′ W54*field2004b
12 Portuguesa, Venezuela09° 18′ N 69° 27′ W-5*field2000u,x**
R. robustus I6 Pampanito, Trujillo, Venezuela09° 24′ N 70° 30′ W84colony1997d
6 Trujillo, Venezuela09° 22′ N 70° 25′ W32*field2004d
13 Candelaria, Trujillo, Venezuela08° 25′ N 65° 30′ W63colony1988e,f
R. robustus II14 Napo, Ecuador00° 25′ S 76° 55′ W-2colony-g
15 Carauarí, Amazonas, Brazil04° 54′ S 66° 54′ W54fieldFebruary 2000h,i,j
16 Porto Velho, Rondônia, Brazil08° 44′ S 63° 25′ W-1colony1985k
17 Apuí, Amazonas, Brazil07° 12′ S 59° 53′ W14colonySeptember 1996l,m
18 Monte Negro, Rondônia, Brazil10° 15′ S 63° 17′ W51*field2004v**
R. robustus III19 Itupiranga, Pará, Brazil04° 25′ S 49° 59′ W-1colony1984n
20 Purupuru, Amazonas, Brazil03° 17′ S 59° 40′ W34colonyDecember 1995n
21 Novo Repartimento, Pará, Brazil04° 20′ S 49° 50′ W35colonyAugust 1998o
33 Tocantinópolis, Tocantins, Brazil06° 22′ S 47° 25′ W-4*colony-w**
R. robustus IV22 Barcarena, Pará, Brazil01° 30′ S 48° 39′ W-5colony1996p
23 Balbina, Amazonas, Brazil00° 55′ S 59° 28′ W-1colonyNovember 1983r
24 UHE Paredão, Roraima, Brazil03° 02′ S 61° 27′ W55colonyMarch 1987s
25 Rio Mapuera, Pará, Brazil00° 00′ S 59° 07′ W-3colonyJune 1986t
12 Portuguesa, Venezuela09° 18′ N 69° 27′ W-5*field2000s
26 Cayenne, French Guiana04° 56′ N 52° 20′ W-1fieldMarch 2000q
26 Cayenne, French Guiana04° 56′ N 52° 20′ W22*fieldJanuary 2003q
26 Macouria, French Guiana04° 55′ N 52° 21′ W1-fieldOctober 2003-
27 Bélizon, French Guiana04° 15′ N 52° 39′ W11*fieldOctober 2003z**
28 Rémire, French Guiana04° 53′ N 52° 16′ W1-fieldAugust 2001-
28 Rémire, French Guiana04° 53′ N 52° 16′ W2-fieldSeptember 2003-
29 Camopi, French Guiana03° 09′ N 52° 20′ W1-fieldOctober 2002-
30 Kaw, French Guiana04° 28′ N 52° 02′ W1-fieldDecember 2003-
31 Mucajaí, Roraima, Brazil02° 27′ S 60° 55′ W-3*fieldNovember 2004A**
32 Manaus, Amazonas, Brazil03° 09′ S 60° 01′ W-2*fieldNovember 2004y**
Total  N = 81N′ = 113   
Figure 1.

 Geographic distribution of the five cytochrome b haplotype clades of Rhodnius prolixus and R. robustus I–IV. Arrowheads indicate putative geographical origins of the two insect samples used by Larrouse (1927) to describe R. robustus (French Guiana and the mouth of the Tefé river in the Brazilian Amazon). The dotted line describes the range of R. prolixus.

DNA extraction, PCR, and sequencing

Each sample (one or two triatomine legs) was placed in an eppendorf, dipped in liquid nitrogen until freezing, and ground to a powder with Pellet Pestles® (Kimble/Kontes). DNA was extracted following Collins et al. (1987) and PCR-amplified with primers CYTB7432F and CYTB7433R (Monteiro et al. 2003). PCR reactions were performed in a PTC-100 Mini-cycler (MJ Research) programmed for one denaturation step at 96 °C for 5 min, followed by 35 cycles at 94 °C for 30 s, 48 °C for 45 s, and 72 °C for 45 s, and a final 10-minute extension step at 72 °C. Purification of PCR products was performed with the GFXTM PCR and gel band purification kit (GE Healthcare), and both fragment strands were subjected to fluorescent dye-terminator cycle sequencing reactions (BigDye® Terminator v3.1 Cycle Sequencing Kit, Applied Biosystems) using the same primers above, and run on an ABI 3730 automated sequencer. DNAStar SeqMan program (DNAStar, Inc.) was used to edit forward and reverse sequences and generate a consensus sequence for each sample. Consensus sequences were manually aligned to a 20 haplotype alignment earlier described (Monteiro et al. 2003) in order to generate the final alignment used for the multiplex primer design.

Multiplex primer design

The strategy employed for the design of the multiplex primers was to search the R. prolixus and R. robustus (I–IV) final haplotype alignment for species-specific regions suitable for the amplification of fragments with different sizes for each of these five species. In order to guarantee specificity, primers should have at least two mismatches (with respect to the other species) at or near the 3′ end. Candidate primer sequences were tested for self-dimerization and hairpin formation using the online program Oligo Analyzer 3.0 (Integrated DNA Technologies, Inc.; http://www.idtdna.com).

Results

DNA sequencing

Six new 663 bp cytochrome b (cyt b) haplotypes (GenBank accession numbers EF011723 – EF011728) were obtained from 29 insect samples and manually aligned to 20 haplotypes earlier described (Monteiro et al. 2003; accession numbers AF421339-AF421343 and EF011708-EF011722; Table 1).

Multiplex primer design and testing

This 26 R. prolixus and R. robustus (I–IV) haplotype alignment was searched for species-specific regions suitable for the design of primers that would amplify fragments with different sizes for these five genetic groups. Several combinations of forward and reverse primers were designed and tested. Fifty microlitre PCR reactions contained 5 μL 10× buffer (100 mM Tris–HCl, pH 8.3, 500 mM KCl), 0.25 mM dNTPs, 15 mM MgCl2, 120 ng of each primer, 2.5 U Taq, and 20–40 ng DNA template, and annealing temperature varied according to the combination of the primers being tested. Ten microlitre of PCR reaction products were run on 1.5% agarose gels containing ethidium bromide and checked for the presence of clear and reproducible diagnostic single-band results for each genetic group. The elected combination was made of a forward (‘universal’) primer that anneals to a region conserved among all haplotypes, and three reverse primers anneal to species-specific regions (Table 2) and amplify fragments of different sizes for R. prolixus (285 bp), R. robustus I (349 bp), and R. robustus II–IV (239 bp)(Figures 2 and 3). The monophyletic R. robustus groups II, III, and IV had to be lumped as there were not enough nucleotide differences to allow for their separation. Each reverse primer has two mismatches, when compared with other two groups, among the first seven 3′-end nucleotides (Figure 2). The thermal cycle profile was optimized for the following conditions: 96 °C for 5 min; then 35 cycles of 94 °C for 30 s, 52 °C for 30 s, 72 °C for 45 s, and a final extension at 72 °C for 10 min. This multiplex assay was used to amplify 33 specimens of R. prolixus collected from Honduras, Guatemala, Colombia and Venezuela; 17 specimens of R. robustus I from Venezuela, and 31 specimens of R. robustus II–IV from Ecuador, Brazil, and French Guiana, totaling 81 insects (Table 1; Figure 3).

Table 2.   PCR primers for the diagnostic multiplex assay
Primer sequence (5′–3′)FunctionSize of PCR product
  1. R = A or G.

TTTGCTCTTCACTTCCTC‘Universal’ (forward) 
GGRATAAAGTTTTCTGGATCR. robustus II–IV (reverse)239
ATATCATTCTGGCTGGATAR. prolixus (reverse)285
CCATTGCAGCAACCCCCR. robustus I (reverse)349
Figure 2.

 Alignment of cytochrome b consensus sequences (400 bp shown of the 663 bp original alignment) for Rhodnius prolixus (consensus for haplotypes a-c, u, and x) and each of the four cryptic R. robustus taxa (consensus for haplotypes d-f; g-m, and v; n, o, and w; p-t, y, z, and A; for R. robustus I, II, III, and IV, respectively). The first sequence (‘consensus’) is the consensus among all five groups. Rectangles outline group-specific areas selected for the design of the primers used in the multiplex-PCR assay, and arrows indicate the 3′ ends of each such areas. A dot (·) indicates identity with the consensus nucleotide; (R) indicates that either an A or a G can be found for that species at that site; (Y) codes for C or T; (M) for A or C.

Figure 3.

 Size-discriminating multiplex PCR products run on an ethidium bromide-stained 1.5% agarose gel. Lanes 1–3 =R. robustus II–IV; 4–6 = R. prolixus; 7–9 = R. robustus I. Lanes M are 100 bp DNA ladders, and lane N is a negative control.

Discussion

Whether R. prolixus and R. robustus are different taxa has long been a matter of controversy probably caused by a combination of three factors: morphologic similarity, loose diagnosis, and sampling. Rhodnius robustus description was based on two female specimens (one from French Guyana and the other from the mouth of the Tefé River in Brazil), considered to be larger and darker than a reference series of R. prolixus (Larrousse 1927). However, although this description was hardly diagnostic (Barrett 1996), the validity of the taxon was acknowledged by Lent & Jurberg (1969) and Lent & Wygodzinsky (1979), who provided additional characters such as shape differences of the median process of the pygophore and basal plate struts of the aedeagus (two male genitalia structures), and colour differences on the hind tibia of later stage nymphs. Thus, considering the geographic origin of the insects used by Larousse (1927), his description of R. robustus possibly stood on two of the four cryptic species it is now known to conceal (Figure 1). Lent & Jurberg (1969) probably faced the same difficulties as (although they made little use of Larousse's characters), again from their geographic origins, their samples are likely to be mixtures of R. robustus forms.

The examination of triatomine male genitalia can be technically challenging and some structures are simply too variable to be used for diagnostic purposes (Harry 1993). Thus, the identification of R. prolixus has relied solely on chromatic differences only observable on the hind tibia of 4th and 5th stage nymphs. In fact, as a consequence of these difficulties, the observation that the smaller R. prolixus is often found building up large colonies in houses, whereas all R. robustus seem to be entirely sylvatic, has led to the combined use of ‘smaller size and domesticity’vs.‘larger size and non-domesticity’ as a practical (and largely subjective) means to separate R. prolixus from R. robustus s.l. The need to objectively discriminate between them led to the employment of a variety of alternative approaches such as random amplification of polymorphic DNA (Garcia et al. 1998, Feliciangeli et al. 2002), electrophoresis of salivary heme proteins (Soares et al. 1998) and morphometry (Villegas et al. 2002). These approaches, although effective, were not tested on all R. robustus groups, and lack the simplicity required for a diagnostic assay.

The multiplex PCR assay presented here is the first simple and objective method able to discriminate among R. prolixus, R. robustus I, and the monophyletic group composed of R. robustus II–IV. Its broad-scale applicability is guaranteed by the demonstration of its efficacy on samples from a vast geographic area (Figures 1 and 3). It is also cost-effective, since no DNA sequencing is required (cf. Monteiro et al. 2003), and applicable to any insect developmental stage, dried specimens or even to insect remains. This method can now be used to accurately identify these closely related and morphologically similar species, thus providing a solid basis for future work aiming at their morphological distinction, as, for example, the promising approach of Villegas et al. (2002) based on wing shape divergence.

Rhodnius prolixus has long been used as a model for insect physiology and biochemistry, and many research groups worldwide maintain laboratory colonies of this species. Because such colonies are often boosted with the addition of new insect samples, they suffer the risk of becoming contaminated with R. robustus. The method described here can be used in the certification of the identity of such colonies.

Our sequencing results show that R. prolixus, R. robustus I, and R. robustus IV coexist in Venezuela (Figure 1). Because the species of the R. robustus complex play only a secondary role in Chagas disease transmission when compared to that of R. prolixus, it is important that these taxa are accurately identified so that vector-specific control actions can be implemented (Feliciangeli et al. 2003, Añez et al. 2004, Sanchez-Martin et al. 2006).

This should make this multiplex assay particularly useful in Venezuela even though it does not discriminate between R. robustus II, III, and IV, because, according to current knowledge on their geographic distribution, forms II and III are absent from that country (Figure 1). A study that would help identify, characterise, and determine the distribution of sylvatic populations of R. prolixus in Venezuela would be greatly beneficial as it would allow for the mapping of areas (e.g. villages) under high risk of becoming colonized (or re-colonized in case of insecticide-treated areas) by sylvatic R. prolixus.

Although of smaller epidemiological relevance, the vectorial capabilities of the R. robustus forms deserve a closer examination as suggested by differences on the light-attractiveness response of different ‘populations’ (Tonn et al. 1976, Miles et al. 1983). In fact, it is believed that sylvatic R. robustus s.l. might enter human habitations attracted by artificial light and transmit Chagas disease in Western Venezuela (Feliciangeli et al. 2002).

DNA amplification of the closely related Rhodnius neglectus by the multiplex primers produces the 239 bp R. robustus II–IV band (not shown). Rhodnius neglectus, however, occurs in drier habitats in Central Brazil and thus should not cause identification problems. The sole exception we are aware of is a borderline area in the state of Tocantins (no. 33 on Figure 1), where the distributions of R. neglectus and R. robustus II seem to overlap (L. Diotaiuti, personal communication), and therefore sequencing will be required for diagnosis. Also, it should be kept in mind that this is a mitochondrial-based assay, and being so will lose its effectiveness in putative hybrid zones where mtDNA introgression between R. prolixus and R. robustus species might occur.

The existence of R. prolixus and of R. robustus complex has been acknowledged in a recent checklist of valid Triatominae species (Galvão et al. 2003), and is further supported by ribosomal DNA (ITS-1 and 2) sequence data (Lazoski & Monteiro, unpublished observation). Galvão and collaborators (2003) also give an update of the geographical distribution of R. prolixus and R. robustus s.l., based on published literature. We believe, however, that some of these reports rest upon misdiagnosed specimens, and recommend the phylogeographic groups presented in Figure 1 as a starting point for further investigation on the distribution of these species.

This methodology could be applied to other cryptic species complexes such as Triatoma dimidiata and Phyllosoma for which some mtDNA sequence data are already available (Martinez et al. 2006, Pfeiler et al. 2006). Nevertheless, one should keep in mind that for such an assay to be of broad applicability, the sequence alignment used for primer design should derive from a comprehensive taxon sampling over most of the species’ distribution.

Acknowledgements

To C. Aznar, M. M. Teixeira, L. Diotaiuti, S. Fitzpatrik, F. Abad-Franch, and M. A. Miles for kindly providing specimens. The comments of F. Abad-Franch and of three anonymous referees helped improve the original manuscript. We thank B. Eschenazi for improving Figure 1 resolution, and the PDTIS-FIOCRUZ DNA sequencing core for running the sequencing reactions. This work was supported by the Brazilian Research Council, CNPq, and benefited from international cooperation through the ECLAT research network.

Ancillary