Is Rhodnius prolixus (Triatominae) invading houses in central Brazil?

Acta Tropica 107 (2008) 90–98

Contents lists available at ScienceDirect

Acta Tropica journal homepage: www.elsevier.com/locate/actatropica

Is Rhodnius prolixus (Triatominae) invading houses in central Brazil? Rodrigo Gurgel-Gonc¸alves a,b,∗ , Fernando Abad-Franch c , Jonatas B.C. Ferreira a,b , ´ Daniella B. Santana a , Cesar A. Cuba Cuba a ´ Laborat´ orio de Parasitologia M´edica e Biologia de Vetores, Faculdade de Medicina, Area de Patologia, Universidade de Bras´ılia, Asa Norte, CEP 70910-900, Distrito Federal, Brazil b Laborat´ orio de Zoologia, Universidade Cat´ olica de Bras´ılia, QS 07 Lote 01 EPTC Bloco M, sala 331, CEP 72030-170, Distrito Federal, Brazil c Instituto Leˆ onidas & Maria Deane, Fiocruz Amazˆ onia, Rua Teresina 476, Adrian´ apolis, CEP 69057-070, Manaus, Amazonas, Brazil a

a r t i c l e

i n f o

Article history: Available online 1 May 2008 Keywords: Rhodnius prolixus Rhodnius neglectus Geometric morphometrics Vector surveillance Chagas disease Brazil

a b s t r a c t Sylvatic triatomines of the genus Rhodnius commonly fly into houses in Latin America, maintaining the risk of Chagas disease transmission in spite of control efforts. In the recent past, adult bugs collected inside houses in central Brazil were identified as R. prolixus, a primary disease vector whose natural geographical range excludes this region. Three nearly sibling species (R. neglectus, R. nasutus, and R. robustus), secondary vectors with limited epidemiological significance, occur naturally south of the Brazilian Amazon. The specific status of Rhodnius specimens found inside houses in central Brazil is therefore an epidemiologically important (and still debated) issue. We used wing and head geometric morphometrics to investigate the taxonomic status of 230 adult specimens representing all four ‘R. prolixus group’ species (19 populations from palm trees, domiciles, and reference laboratory colonies). Discriminant analyses of shape variation allowed for an almost perfect reclassification of individuals to their putative species. Shape patterning revealed no consistent differences between most specimens collected inside houses in central Brazil and R. neglectus, and showed that R. robustus and R. neglectus occur sympatrically (and fly into houses) in southern Amazonia. Furthermore, all Brazilian specimens clearly differed from our reference R. prolixus population. Using geometric morphometrics, we confidently ascribed individual triatomines to their species within the problematic ‘R. prolixus group’, illustrating the potential value of this approach in entomological surveillance. Our results strongly support the idea that R. neglectus, and not R. prolixus, is the species invading houses in central Brazil. © 2008 Elsevier B.V. All rights reserved.

1. Introduction The genus Rhodnius is comprised of 16 recognized species; natural populations occupy arboreal ecotopes (preferentially palm trees) in ∼28 biogeographical provinces from Central America to southern Brazil (Abad-Franch and Monteiro, 2007). Populations of several Rhodnius species have adapted to artificial environments in different regions; one of them, R. prolixus, is a major domestic vector of human Chagas disease. Synanthropic triatomine populations have been targeted by successful control programs based on household spraying with residual insecticides (e.g., Dias, 2007). However, the presence of sylvatic vector populations near houses limits the effectiveness of such control interventions in many regions. Adult triatomines recurrently invade households, maintaining the risk

∗ Corresponding author. Fax: +55 61 273 3907. E-mail address: [email protected] (R. Gurgel-Gonc¸alves). 0001-706X/$ – see front matter © 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.actatropica.2008.04.020

of Chagas disease transmission and occasionally originating new domestic–peridomestic breeding colonies (Miles et al., 2003). Such dynamics are common across Amazonia, the Orinoco basin, Central America, and the drier ecoregions of central Brazil (Barretto et al., 1968; Silveira et al., 1984; Garcia-Zapata et al., 1985; Silva et al., 1992, 1999; Guilherme et al., 2001; Feliciangeli et al., 2003; ˜ et al., 2003; Sanchez-Martin et al., 2006; Oliveira and Silva, Romana 2007; Aguilar et al., 2007). A large fraction of these adventitious vectors belongs to the ‘Rhodnius prolixus group’, which includes R. prolixus, R. robustus, R. nasutus, and R. neglectus (Barrett, 1991). These four species are phenotypically almost sibling, and their natural distribution ranges (not always precisely defined) often overlap partially. Under these circumstances, it seems virtually unavoidable that some field-collected specimens be misidentified (Monteiro et al., 2001; Miles et al., 2003). R. neglectus is apparently restricted to the savanna-like central Brazilian Cerrado and the neighboring southern fringes of the Amazon, while R. nasutus occupies drier areas in the Caatinga of

R. Gurgel-Gonc¸alves et al. / Acta Tropica 107 (2008) 90–98

northeastern Brazil. R. robustus comprises a complex of cryptic species that occur throughout the Amazon moist forests and the Orinoco plains. Finally, R. prolixus often infests houses in parts of Venezuela, Colombia, and Central America; sylvatic populations have only been confirmed to occur in palm trees of the ColombianVenezuelan Orinoco basin (Lent and Wygodzinsky, 1979; Dujardin et al., 1998; Abad-Franch and Monteiro, 2007; Guhl, 2007). R. prolixus has however been also reported from Brazil after collections made in areas clearly outside the currently accepted range of the species (Pinho et al., 1998; Schofield and Dujardin, 1999). Until the 1950s, Rhodnius specimens collected inside houses in central Brazil were all identified as R. prolixus. After Lent (1954) described R. neglectus, these bugs (collected from either palm ´ and S˜ao trees or houses in the states of Minas Gerais, Goias, Paulo) were reassigned to this new species. However, it was later claimed that R. prolixus could also be found in some areas of the Amazonia-Cerrado transition (Diotaiuti et al., 1984; Silveira et al., 1984; see also Pinho et al., 1998). Dujardin et al. (1991) tackled the issue by comparing allozyme profiles of different R. prolixus and R. neglectus populations. Results revealed no differences between laboratory-reared R. prolixus and field-caught bugs (identified as R. neglectus) from peridomestic environments ´ On the other hand, laboratory R. neglectus of Mamba´ı, Goias. specimens drawn from the FIOCRUZ reference colony had a distinct allozyme profile. The authors concluded that a species other than R. neglectus was invading houses in central Brazil, leaving open the possibility that it was R. prolixus (Dujardin et al., 1991). A relatively rich bibliography shows how both traditional (e.g., Dujardin et al., 1997, 1999a; Patterson et al., 2001) and geometric morphometric analyses (e.g., Mat´ıas et al., 2001; Villegas et al., 2002; Gumiel et al., 2003; Schachter-Broide et al., 2004; Dujardin et al., 2007; Feliciangeli et al., 2007) provide valuable tools for the study of taxonomically problematic groups within the Triatominae. In this work, we used geometric morphometric analyses for species-level identification of adult Rhodnius specimens collected within houses in Brazil. Several field-collected (from sylvatic and domestic environments) and laboratory bug samples were compared, in order to assess whether size and shape variation analyses

91

may provide suitable taxonomic markers for the members of the problematic ‘R. prolixus group’. 2. Materials and methods 2.1. Bug samples Two hundred and thirty adult Rhodnius specimens from 19 populations (five from palm trees, two from laboratory colonies, and 12 from households in five Brazilian states) were analyzed (Table 1). Sylvatic samples were collected from palm trees, mainly Mauritia flexuosa (see Gurgel-Gonc¸alves et al., 2003, 2004 for details), under permissions issued by the Brazilian Environmental Agency (IBAMA). Our R. neglectus reference sample was drawn from the colony at the Centro de Pesquisas Rene´ Rachou (FIOCRUZ Minas Gerais, Brazil), founded with specimens from the type locality. ´ Brazil) was also samOne sylvatic population from Mamba´ı (Goias, pled, allowing us to confidently reassess the results presented by Dujardin et al. (1991). Our R. prolixus reference sample came from a colony (Laboratorio de Investigaciones en Parasitolog´ıa Tropical, ´ Colombia) founded with domesUniversidad del Tolima, Ibague, tic specimens collected in the Magdalena Valley—a trans-Andean location where no other members of the ‘R. prolixus group’ occur (Abad-Franch and Monteiro, 2007). Domestic Brazilian specimens were collected inside or around houses during routine searches by workers of the Chagas Disease Control Program (CDCP) between 2006 and 2007. All bugs were preliminarily identified after Lent and Wygodzinsky (1979). Samples whose specific status was dubious (and we aimed at testing) were however treated as ‘problem specimens’ in our metric analyses. These samples included the Mamba´ı population and the specimens collected within houses by CDCP agents (Rhodnius sp. in Table 1). 2.2. Morphometrics Right hemelytra (forewings) were mounted between microscope slides and cover slips using a mounting resin (Entellan® ), and digitally scanned (1200 dpi). Six type I (venation intersection) and one type II landmarks (Bookstein, 1991) were digitized (Fig. 1).

Table 1 Origin, habitats, geographic coordinates and number of wings and heads of Rhodnius sp. populations used in geometric morphometric analyses Populations

Origine

Habitats

Coordinates

Wings

Heads

R. neglectusa R. neglectusa R. neglectusb R. robustusc R. nasutusb R. prolixusd Rhodnius sp. Rhodnius sp. Rhodnius sp. Rhodnius sp. Rhodnius sp. Rhodnius sp. Rhodnius sp. Rhodnius sp. Rhodnius sp. Rhodnius sp. Rhodnius sp. Rhodnius sp. Rhodnius sp.

Buritizal, SP, Brazil Taguatinga, TO, Brazil Uberaba, MG, Brazil ´ PA, Brazil Maraba, Sobral, CE, Brazil Tolima, Ibague, Colombia Mamba´ı, GO, Brazil Ituiutaba, MG, Brazil ´ Tocantinopolis, TO, Brazil Araguatins, TO, Brazil Taguatinga, TO, Brazil ´ Quirinopolis, GO, Brazil ´ GO, Brazil Cidade de Goias, Uruana, GO, Brazil Piracanjuba, GO, Brazil ´ MT, Brazil Cotriguac¸u, Ponte e Lacerda, MT, Brazil ´ Caceres, MT, Brazil Dourados, MS, Brazil

Palm trees (Mauritia flexuosa) Palm trees (M. flexuosa) Laboratory colony Palm trees (Attalea speciosa) Palm trees (Copernicia prunifera) Laboratory colony Palm trees (M. flexuosa) Domiciles Domiciles Domiciles Domiciles Domiciles Domiciles Domiciles Domiciles Domiciles Domiciles Domiciles Domiciles

20◦ 11 S, 47◦ 42 W 12◦ 23 S, 46◦ 24 W 19◦ 32 S, 48◦ 01 W 05◦ 16 S, 49◦ 50 W 03◦ 40 S, 40◦ 13 W 02◦ 59 N, 74◦ 29 W 14◦ 27 S, 46◦ 08 W 18◦ 57 S, 49◦ 35W 06◦ 19 S, 47◦ 24 W 05◦ 38 S, 48◦ 05 W 12◦ 23 S, 46◦ 23 W 18◦ 28 S, 50◦ 27 W 15◦ 54 S, 50◦ 08 W 15◦ 35 S, 49◦ 41 W 17◦ 18 S, 49◦ 01 W 09◦ 52 S, 58◦ 23 W 15◦ 13 S, 59◦ 21 W 16◦ 07 S, 57◦ 36 W 22◦ 13 S, 54◦ 47 W

17 23 11 34 30 32 19 6 9 3 1 1 4 1 1 5 1 1 7

30 23 8 45 30 37 21 4 8 3 1 1 4 1 1 5 1 1 6

a b c d e

Sylvatic populations. Colony from Centro de Pesquisas Rene´ Rachou, CPqRR-FIOCRUZ, Minas Gerais. Uberaba specimens were collected in the same locality of the holotype of R. neglectus. F1 and F2 generations. ´ Colombia. Colony from Laboratorio de Investigaciones en Parasitolog´ıa Tropical, Facultad de Ciencias, Universidad del Tolima, Ibague, ´ CE (Ceara), ´ GO (Goias), ´ MT (Mato Grosso) and MS (Mato Grosso do Sul). SP (S˜ao Paulo), TO (Tocantins), MG (Minas Gerais), PA (Para),

92

R. Gurgel-Gonc¸alves et al. / Acta Tropica 107 (2008) 90–98

Fig. 1. Right wing of Rhodnius neglectus with the seven landmarks used in morphometric analysis. Following Bookstein (1991), point 4 corresponds to a type II landmark, and the remain to type I landmark.

Heads were pinned on triangular cards and photographed with a digital camera (Sony® Ciber-shot 5.1 Mp) adapted to a NOVA® stereomicroscope (25× magnification). Eight head landmarks were used in morphometric analyses (Fig. 2). Landmark coordinates were recorded using tpsDig 1.18 (Rohlf, 1999a).

field-collected Brazilian populations of R. nasutus and R. robustus, and removing R. prolixus. Factorial maps were constructed to graphically show the distribution of specimens and populations in the shape space defined by the two discriminant factors; to improve clarity in the graphic output, convex polygons enclosing all specimens within each group were overlaid on the plots and individual dots removed (except for ‘problem specimens’ whose position we wanted to examine). Reclassification of specimens to their original putative groups was assessed by contingency table analysis and Kappa statistics (Landis and Koch, 1977). Finally, we measured the contribution of size to shape variation (allometry) using multiple regression of shape discriminant factors against CS values (wings and heads). DFA, ANOVA, MANOVA, multiple regression and Tukey tests were computed with Statistica® . Kappa statistics were computed with JMP® . 3. Results

2.3. Size and shape variation 3.1. Size variation We used “centroid size” (CS), an isometric size estimator derived from coordinate data (Bookstein, 1991), to analyze wing and head size variation. The CS value was extracted from the coordinate matrix of each individual structure using tpsRelw version 1.18 (Rohlf, 1999b) and log-transformed. Shape variables were obtained using the Generalized Procrustes Analysis (GPA) superimposition algorithm (Rohlf, 1996). First, a consensus configuration was derived from raw coordinate data using a least-squares iterative approach. “Partial warps” were then computed as deformations of each individual structure in relation to the consensus configuration. Both uniform and non-uniform deformation components were used in the analyses. Shape variables were computed and tested for variation using tpsRelw 1.18 (Rohlf, 1999b).

As expected, size sexual dimorphism was consistently observed. Female wings and heads were significantly larger than those of males. Wing size was largest in R. robustus and smallest in R. nasutus. Head size was markedly smaller in R. prolixus than in any other species/population analyzed here (Fig. 3). We scored statistically significant differences among wings (males: ANOVA F4,61 = 41.5; p < 0.01; females: ANOVA F4,65 = 43.3; p < 0.01) and mainly heads (males: ANOVA F4,76 = 102.3; p < 0.01; females: ANOVA F4,75 = 98.6; p < 0.01) across Rhodnius species/populations. The heads and wings of R. neglectus from Tocantins were larger on average than those from S˜ao Paulo, but the difference was not statistically significant (p = 0.10).

2.4. Statistical analyses

3.2. Shape variation and allometry

Size variation (wing and head CS values) among populations was explored by means of ANOVA and Tukey tests (alpha = 0.01). The homogeneity of shape variables across groups was tested by MANOVA; the same variables (partial warps and uniform components) were used as input for multivariate discriminant function analysis (DFA). In a first analysis, we explored the relationships between ‘problem specimens’ and our R. prolixus and R. neglectus reference populations. We used shape variables to derive a discriminant model under which R. prolixus, R. neglectus (sylvatic from Tocantins and S˜ao Paulo and the FIOCRUZ stock), the Mamba´ı population, and other ‘problem specimens’ collected in houses of five Brazilian states (see Table 1) could be classified. Because several ‘problem specimens’ could not be confidently ascribed to any of the reference groups, we ran a second discriminant analysis including

Factorial maps of both wing and head shape variation revealed significant differences between ‘problem specimens’ and our reference R. prolixus population (Fig. 4). ‘Problem specimens’ from ´ (including the Mamba´ı population), Minas Gerais, and Mato Goias Grosso do Sul were all indistinguishable from our R. neglectus reference populations. Some specimens from Tocantins and Mato Grosso were also identified as R. neglectus, while others significantly differed from both R. neglectus and R. prolixus. DFA-based correct reclassification scores were high for both head and wing shape variables. All R. prolixus were correctly reclassified, while Mamba´ı and R. neglectus specimens often shifted groups (Table 2). Head and wing shape differences between R. neglectus and R. prolixus were patent. R. neglectus heads were narrow and elongated when compared with those of R. prolixus, which had larger eyes. Wing shape changes largely involved landmarks 5 (subcostal vein) and 6 (median and cubital veins) (Fig. 4). Multiple regression of discriminant factors (shape variation) against CS revealed no significant allometric trend for wings (R2 = 0.004; p = 0.84), but a significant allometric content was observed in the head shape dataset (R2 = 0.45; p < 0.01). Factorial maps derived from the second discriminant analysis (comparing R. neglectus populations with R. robustus and R. nasutus) showed that R. neglectus and R. robustus are sympatric in the states of Tocantins and Mato Grosso (Fig. 5). R. neglectus was the ´ only species found within human dwellings in Minas Gerais, Goias, and Mato Grosso do Sul. In these analyses, correct reclassification scores were very high for both wings and heads (Table 3). Multiple regression revealed a significant allometric content in the first discriminant factor (CV1: p < 0.01). Size predicted 18% of

Fig. 2. Head of Rhodnius prolixus illustrating the eight landmarks used in morphometric analysis. Following Bookstein (1991), points 1–2 and 5–6 correspond to type I landmark.

R. Gurgel-Gonc¸alves et al. / Acta Tropica 107 (2008) 90–98

93

´ RneSP (R. neglectus S˜ao Paulo), RneTO Fig. 3. Variation of wing and head centroid size (CS) among populations (males and females) of Rhodnius spp.: Rna (R. nasutus Ceara), (R. neglectus Tocantins), Rp (R. prolixus) and Rr (R. robustus). Boxes show mean CS values and standard errors (S.E.); standard deviations (S.D.) are shown as lines. Populations labeled with different letters (above boxes) were statistically different by Tukey tests (p < 0.01).

wing and 19% of head shape variation. Regarding venation conformational changes, the posterior apex of the forewing (landmark 4) was more distant from the end of the cubital vein (landmark 2) in R. robustus than in R. neglectus. The relative positions of wing landmarks 1 and 7 were important in the discrimination of these species. R. robustus had larger eyes than R. neglectus (Fig. 5).

4. Discussion 4.1. The use of metric traits for species diagnosis within the ‘R. prolixus group’ The use of morphometric traits in systematics relies on the assumption that variation of continuous phenotypic attributes is

94

R. Gurgel-Gonc¸alves et al. / Acta Tropica 107 (2008) 90–98

Fig. 4. Factorial maps in the plane of the two canonical factors of shape variation for wing and head (CV1 and CV2) presenting the distribution of Rhodnius neglectus (from Tocantins [TO], S˜ao Paulo [SP], and FIOCRUZ), R. prolixus and Rhodnius Mamba´ı populations. The symbols represent the origin of Rhodnius specimens captured in houses. The percentage contribution of each component to the total shape variation is shown on the axes in parentheses. Polygons enclose each group with different patterns. Drawings right the factorial maps are the consensus conformation of the head and wing shape (see landmarks in Figs. 1 and 2). The arrows indicate the differences in head and wing shape of R. neglectus and R. prolixus (see text for details).

the expression of underlying genetic diversity; the degree to which phenotypic traits differ between two groups is regarded as a proxy measure of evolutionary distance, with clear-cut differences often signaling mutual isolation (Sorensen and Foottit, 1992). Yet, for these principles to hold true the two main sources of variability (genetic and environmental) must be partitioned; the standard procedure within a strict morphometric framework entails the separate assessment of size (largely environmental) and shape (genetic) variation (Dujardin et al., 2002; Baylac et al., 2003). We used a geometric morphometric approach to explore the alpha systematics of ‘problematic’ Rhodnius specimens collected in central

Brazil, comparing them with reference populations of known specific status. Our results show that wing and head geometric traits can be used for accurate species-level identification of ‘R. prolixus group’ members, and shed light on the epidemiologically significant question of whether R. prolixus populations occur in Brazil (see below) (Diotaiuti et al., 1984; Silveira et al., 1984; Dujardin et al., 1991; Schofield and Dujardin, 1999). 4.1.1. Size variation Crude size disparity is often used in triatomine systematics. Species descriptions invariably refer to linear measurements,

R. Gurgel-Gonc¸alves et al. / Acta Tropica 107 (2008) 90–98

95

Table 2 ´ Brazil) wings and heads after discriminant analysis of shape variables Reclassification scores of Rhodnius neglectus, R. prolixus, and Rhodnius sp. from Mamba´ı (Goias, Groups

R. neglectus R. prolixus Rhodnius sp.a Kappab

Wings

Heads a

R. neglectus

R. prolixus

Rhodnius sp.

% correct reclassification

R. neglectus

R. prolixus

Rhodnius sp.a

% correct reclassification

36 0 4

0 32 0

4 0 15

90 100 79

50 0 14

0 37 0

4 0 7

92 100 33

0.86

0.85

R. = Rhodnius. a ´ Brazil) population (see text for details). Mamba´ı (Goias, b Kappa: measures the degree of agreement between observed and expected classification; it ranges from 0 to 1; a score between 0.80 and 1 is considered almost perfect agreement (Landis and Koch, 1977).

from overall body length to the dimensions of particular structures such as the head, pronotum, legs, or antennae segments (Lent and Wygodzinsky, 1979). Since Dujardin and his collaborators introduced modern morphometric techniques for the study of triatomines in the 1990s (Dujardin et al., 2002), the results of many investigations suggest that size variation largely mirrors environmental (not genetic) heterogeneity (e.g., Dujardin et al., 1997, 1998, 1999b; Schachter-Broide et al., 2004). Size-related traits are thus prone to either convergence (when two genetically distinct entities occupy similar environments) or divergence (when subsets of a genetically homogeneous metapopulation adapt to distinct microhabitats) in triatomines (Abad-Franch et al., 2003). Consequently, size variation has to be cautiously interpreted in the context of species-level diagnosis (in which a declaration on the genetic consequences of reproductive isolation is implicit), and a careful assessment of allometric trends is recommended (Dujardin et al., 2002). Our results provide several further examples of apparently erratic isometric size variation. For instance, R. prolixus is expected to be larger on average than R. neglectus, but our CS analyses revealed the opposite for most R. neglectus specimens. Similarly, R. robustus heads and wings are generally larger than those of R. neglectus, but some of the R. neglectus specimens from Tocantins we analyzed were similar in size to R. robustus. Our dataset also revealed a trend towards size-related metric divergence among R. neglectus geographical populations, a finding previously reported by Soares et al. (1999). Habitat-related size variation has also been detected among R. nasutus populations (Diotaiuti et al., 2005). 4.1.2. Shape variation and allometry Shape patterns tend to differ significantly among genetically distinct organisms. The differences can however be subtle enough to remain undetectable to the naked eye. This is generally the case when cryptic species go unrecognized in qualitative systematic investigations. Geometric morphometric analyses involve the explicit, quantitative assessment of otherwise hidden shape patterning among organisms; their power in dealing with complex systematic problems is now widely recognized (see Baylac et al., 2003 and references therein). Wing shape patterns are seldom used in (traditional) triatomine systematics, whereas several head characters are important taxonomic markers. Multivariate analyses of wing metric traits may however discriminate nearly sibling taxa and reveal fine-scale spatial structuring among populations of a single species (Mat´ıas et al., 2001; Villegas et al., 2002; Schachter-Broide et al., 2004; Feliciangeli et al., 2007). They even helped ascribe individual triatomine specimens to their parental lines (Dujardin et al., 2007). When applied to assign individual bugs to four nearly sibling Rhodnius species (R. prolixus, R. neglectus, R. nasutus, and R. robustus), our discriminant analyses of wings performed slightly

better than those of heads, but both yielded largely coherent results. Our first discriminant models revealed clear-cut, consistent differences between R. prolixus and R. neglectus (Fig. 4). Head variation was in agreement with the original description of R. neglectus by Lent (1954). Both wing and head shape comparisons allowed us to confidently conclude that most of our ‘problem specimens’ (including the Mamba´ı population) belong to the latter species; a few synanthropic bugs from Tocantins and Mato Grosso did not appear to resemble any of these two taxa. Our data also showed significant shape divergence between R. neglectus and R. robustus, which had more elongated heads and wings even when size was similar. The posterior apex of the forewing (landmark 4) was more distant from the end of the cubital vein (landmark 2) in R. robustus. Villegas et al. (2002) reported similar findings in their comparison of Venezuelan populations of R. prolixus and R. robustus, underscoring the key contribution of these two landmarks to wing shape changes in the ‘R. prolixus group’. Shape differences (involving the wing apex and the relative positions of ocelli and eyes) were also observed between R. neglectus and R. nasutus (details not shown). When comparing R. neglectus and R. prolixus specimens, our wing dataset revealed no allometric content, therefore suggesting that shape variation mainly reflects genetic variance; on the other hand, head shape differences were under a significant influence of size. These findings indicate that wing shape may be less sensitive to size changes than head shape is, but the mechanisms underlying this trend remain a matter of investigation. Overall, wing allometry components recorded in the R. neglectus–robustus–nasutus dataset were smaller than those described for R. prolixus and R. robustus by Villegas et al. (2002). 4.2. Which species of Rhodnius is invading houses in central Brazil? The ‘R. prolixus group’ is composed of four closely related, nearly sibling Rhodnius species with different epidemiological significance. R. prolixus is a major domestic vector north of the Amazon; R. neglectus often invades and colonizes artificial environments in the Brazilian Cerrado, while R. nasutus does so less frequently in the Caatinga; finally, the various lineages within the R. robustus complex invade (but do not colonize) households across the Amazon–Orinoco systems. For any given species within this group, presence records based on the identification of specimens after qualitative assessment of morphological-chromatic patterns will probably involve a certain proportion of misclassifications. Wrong records will in turn be used to define the species geographical range, giving rise to erroneous biogeographical inferences and mistaken estimations of epidemiological risk (Monteiro et al., 2001). The argument over the presence of R. prolixus (an extremely efficient disease vector) in Brazil suitably illustrates these difficulties

96

R. Gurgel-Gonc¸alves et al. / Acta Tropica 107 (2008) 90–98

Fig. 5. Factorial maps in the plane of the two canonical factors of shape variation for wing and head (CV1 and CV2) presenting the distribution of Rhodnius neglectus (from Tocantins [TO] and S˜ao Paulo [SP]), R. nasutus and R. robustus populations. The symbols represent the origin of Rhodnius specimens captured in houses. The percentage contribution of each factor to the total shape variation is shown on the axes in parentheses. Polygons enclose each group with different patterns. Drawings right the factorial maps are the consensus conformation of the head and wing shape (see landmarks in Figs. 1 and 2). The arrows indicate the differences in head and wing shape of R. neglectus and R. robsutus (see text for details).

(e.g., Dujardin et al., 1991). While laboratory-reared R. prolixus and peridomestic R. neglectus from Mamba´ı shared identical allozyme patterns, both differed from a R. neglectus reference colony. It was consequently suggested that a species other than R. neglectus, and closely related to R. prolixus, was invading houses in the study area (Dujardin et al., 1991). Our results, largely based on the comparison of field-collected material, strongly suggest that R. neglectus is the species invad-

ing and occasionally colonizing artificial environments in most of central Brazil. R. robustus and R. nasutus may also invade households in the humid Amazon and the Caatinga, respectively, but R. prolixus seems to be absent from the study area. We successfully confirmed these findings (clearly at odds with the conclusions of Dujardin et al. (1991)) by comparing mitochondrial cytochrome b DNA sequences (Monteiro et al., 2003; Pavan and Monteiro, 2007) from the same spec-

R. Gurgel-Gonc¸alves et al. / Acta Tropica 107 (2008) 90–98

97

Table 3 Reclassification scores of Rhodnius neglectus, R. robustus, and R. nasutus wings and heads after discriminant analysis of shape variables Groups

Wings R. neglectus

R. nasutus

R. robustus

% correct reclassification

R. neglectus

R. nasutus

R. robustus

% correct reclassification

R. neglectus R. nasutus R. robustus

39 30 0

1 0 0

0 0 34

97 100 100

52 0 0

1 28 1

1 2 40

96 93 96

Kappaa

Heads

0.98

0.95

R. = Rhodnius. a Kappa: measures the degree of agreement between observed and expected classification; it ranges from 0 to 1; a score between 0.80 and 1 is considered almost perfect agreement (Landis and Koch, 1977).

imens and populations; detailed results will be presented elsewhere. R. neglectus is a secondary vector potentially involved in human Chagas disease transmission throughout its wide distribution across the Brazilian Cerrado and the southern fringes of the Amazon. Household infestation (with adventitious bugs occasionally estab´ lishing breeding colonies) has been reported in the states of Goias, Minas Gerais, and S˜ao Paulo (Barretto et al., 1968; Garcia-Zapata et al., 1985; Silva et al., 1992, 1999). R. neglectus is currently the second most common triatomine species infesting artificial environments ´ Infested households were detected in 79% of in the state of Goias. the municipalities in the state, and over 4500 R. neglectus specimens (∼1% infected with Trypanosoma cruzi) were collected over 2 years (Oliveira and Silva, 2007). Extensive longitudinal surveillance systems capable of detecting and eliminating synanthropic R. neglectus populations are therefore needed across the range of the species. The development of reliable and practical methods for vector species discrimination is obviously crucial in the context of such disease control-surveillance programs, where decision-making is often based on taxonomic judgments (Monteiro et al., 2001).

5. Conclusions The correct taxonomic assignment of problem organisms is not only a cornerstone of biological research. It also has deep practical (and ethical) implications when the lives and welfare of human beings depend on the development of suitable strategies for the management of a given species—and not of another, superficially similar taxon. Disease vectors obviously enter this category, together with pathogens, crop pests or endangered keystone species. Most triatomine species can be confidently distinguished using external morphological-chromatic characters; some groups are however problematic, and a few taxa are essentially isomorphic (Lent and Wygodzinsky, 1979). Several epidemiologically important species pose such difficulties; those within the ‘R. prolixus group’ constitute a classic example. Various alternative taxonomic markers have already been tested for species-level identification of morphologically similar triatomines, but they tend to be expensive and laborious (Abad-Franch and Monteiro, 2005). Multivariate analysis of metric characters is probably too complex to become widely used in routine, decentralized entomological surveillance. However, it is not difficult to envision a reference system whereby control program managers could send ‘problem specimens’ (or digital pictures/scans of key structures) to a supporting laboratory where they could be compared to a reference morphometric dataset and thus ascribed to a specific taxon. We effectively mimicked this real-life situation by treating bugs collected by CDCP workers as ‘problem specimens’ in our analyses, and showed how this approach can be put into practice even with paradigmatically problematic triatomine species.

Acknowledgments This study benefited from international collaboration through ´ the ECLAT network. Special thanks are due to Fabio Oliveira Alves for fieldwork assistance. We also thank C.J. Schofield, J.P. Dujardin, M.A. Miles, and S. Catala´ for comments on earlier versions of the manuscript. L. Diotaiuti and F.B.S. Dias provided specimens from the R. neglectus FIOCRUZ colony, and the workers of the Brazilian Chagas Disease Control Program supplied field-collected ‘problem specimens’. Different parts of this research were funded by FAP-DF, FAPEAM and CNPq (Brazil), and by the WHO TDR Special Program.

References Abad-Franch, F., Monteiro, F.A., Patterson, J.S., Aguilar, H.M., Beard, C.B., Miles, M.A., 2003. Population phenotypic plasticity linked to ecological adaptations in Triatominae. Rev. Inst. Med. Trop. S˜ao Paulo 45 (Suppl. 13), 199–200. Abad-Franch, F., Monteiro, F.A., 2005. Molecular research and the control of Chagas disease vectors. An. Acad. Bras. Ciˆencias 77, 437–454. Abad-Franch, F., Monteiro, F.A., 2007. Biogeography and evolution of Amazonian triatomines (Heteroptera: Reduviidae): implications for Chagas disease surveillance in humid forest ecoregions. Mem. Inst. Oswaldo Cruz 102 (Suppl. 1), 57–69. Aguilar, H.M., Abad-Franch, F., Dias, J.C.P., Junqueira, A.C.V., Coura, J.R., 2007. Chagas disease in the Amazon region. Mem. Inst. Oswaldo Cruz 102 (Suppl. 1), 47–55. Barrett, T.V., 1991. Advances in triatomine bug ecology in relation to Chagas disease. Adv. Dis. Vector Res. 8, 143–176. Barretto, M.P., Siqueira, A.F., Ferriolli, F.F., Carvalheiro, J.R., 1968. Estudos sobre reser´ vatorios e vetores do Trypanosoma cruzi. XXIII. Observac¸o˜ es sobre criadouros do ´ Rhodnius neglectus Lent, 1954 em biotopos artificiais (Hemiptera, Reduviidae). Rev. Inst. Med. Trop. S˜ao Paulo 10, 163–170. Baylac, M., Villemant, C., Simbolotti, G., 2003. Combining geometric morphometrics with pattern recognition for the investigation of species complexes. Biol. J. Linn. Soc. 80, 89–98. Bookstein, F.L., 1991. Morphometric Tools for Landmark Data: Geometry and Biology. Cambridge University Press, Cambridge, UK, p. 435. Dias, J.C.P., 2007. Southern Cone Initiative for the elimination of Triatoma infestans and the interruption of transfusional Chagas disease. Historical aspects, present situation, and perspectives. Mem. Inst. Oswaldo Cruz 102 (Suppl. 1), 11–18. ˚ Diotaiuti, L., Silveira, A.C., Elias, M., 1984. Sobre o encontro de Rhodnius prolixus Stal 1859 em Macaubeiras. Rev. Bras. Malariol. D. Trop. 36, 11–14. Diotaiuti, L., Lorenzo, M.G., Dias, F.B.S., Bezerra, C.M., Garcia, M.H., Paula, A.S., 2005. ´ ˚ 1859 Influˆencia da especie de palmeira sobre o tamanho de Rhodnius nasutus Stal, ´ ´ Brasil. Rev. Soc. Bras. Med. Trop. provenientes de carnaubas ou babac¸us do Ceara, 38 (Suppl. 1), 44. Dujardin, J.P., Garcia-Zapata, M.T., Jurberg, J., Roelants, P., Cardozo, L., Panzera, F., Dias, J.C.P., Schofield, C.J., 1991. Which species of Rhodnius is invading houses in Brazil? Trans. R. Soc. Trop. Med. Hyg. 85, 679–680. ´ Dujardin, J.P., Bermudez, H., Schofield, C.J., 1997. The use of morphometrics in entomological surveillance of silvatic foci of Triatoma infestans in Bolivia. Acta Trop. 66, 145–153. ˜ ´ Dujardin, J.P., Munoz, M., Chavez, T., Ponce, C., Moreno, J., 1998. The origin of Rhodnius prolixus in Central America. Med. Vet. Entomol. 12, 113–115. ´ Dujardin, J.P., Steindel, M., Chavez, T., Ponce, C., Moreno, J., Schofield, C.J., 1999a. Changes in the sexual dimorphism of Triatominae in the transition from natural to artificial habitats. Mem. Inst. Oswaldo Cruz 94, 565–569. Dujardin, J.P., Panzera, F., Schofield, C.J., 1999b. Triatominae as a model of morphological plasticity under ecological pressure. Mem. Inst. Oswaldo Cruz 94 (Suppl. 1), 223–228. Dujardin, J.P., Schofield, C.J., Panzera, F., 2002. Los vectores de la enfermedad de ´ ´ ´ ´ ampliada Chagas. Ivestigaciones taxonomicas, biologicas y geneticas, Version ´ y actualizada, N.S. 25, Academie Royale des Sciences d’Outre-Mer, Classe des ´ Sciences naturelles et medicales, Brussels, Belgium, pp. 144–151.

98

R. Gurgel-Gonc¸alves et al. / Acta Tropica 107 (2008) 90–98

Dujardin, J.P., Beard, C.B., Ryckman, R., 2007. The relevance of wing geometry in entomological surveillance of Triatominae, vectors of Chagas disease. Infect. Genet. Evol. 7, 161–167. Feliciangeli, M.D., Campbell-Lendrum, D., Martinez, C., Gonzalez, D., Coleman, P., Davies, C.R., 2003. Chagas disease control in Venezuela: lessons for the Andean region and beyond. Trends Parasitol. 19, 44–49. Feliciangeli, M.D., Sanchez-Martin, M., Marrero, R., Davies, C., Dujardin, J.P., 2007. Morphometric evidence for a possible role of Rhodnius prolixus from palm trees in house re-infestation in the State of Barinas (Venezuela). Acta Trop. 101, 169–177. Garcia-Zapata, M.T., Virgens, D., Soares, V.A., Bosworth, A., Marsden, P.D., 1985. House ´ invasion by secondary Triatominae species in Mamba´ı. Goias-Brazil. Rev. Soc. Bras. Med. Trop. 18, 199–201. Guhl, F., 2007. Chagas disease in Andean countries. Mem. Inst. Oswaldo Cruz 102 (Suppl. 1), 29–37. ´ S.M., 2001. Secondary Guilherme, A.L., Pavanelli, G.C., Silva, S.V., Costa, A.L., de Araujo, triatomine species in dwellings and other nearby structures in municipalities ´ Brazil. Rev. Panam. under epidemiological surveillance in the state of Parana. ´ Salud Publ. 9, 385–392. ´ S., Noireau, F., Rojas de Arias, A., Garc´ıa, A., Dujardin, Gumiel, M., Catala, J.P., 2003. Wing geometry in Triatoma infestans (Klug) and T. melanosoma Mart´ınez, Olmedo and Carcavallo (Hemiptera Reduviidae). Syst. Entomol. 28, 173–179. Gurgel-Gonc¸alves, R., Palma, A.R.T., Menezes, M.N.A., Leite, R.N., Cuba, C.A.C., 2003. Sampling Rhodnius neglectus (Triatominae) in Mauritia flexuosa palm trees (Arecaceae): a field study in the Brazilian Savanna. Med. Vet. Entomol. 17, 347–349. ˜ C.A., Cuba, C.A.C., 2004. Gurgel-Gonc¸alves, R., Duarte, M.A., Ramalho, E.D., Romana, Distribuic¸a˜ o espacial de populac¸o˜ es de Triatominae (Hemiptera, Reduviidae) em ´ palmeiras da especie Mauritia flexuosa no Distrito Federal, Brasil. Rev. Soc. Bras. Med. Trop. 37, 241–247. Landis, J.R., Koch, G.G., 1977. The measurement of observer agreement for categorical data. Biometrics 33, 159–174. ´ ˚ com descric¸a˜ o de uma nova Lent, H., 1954. Comentarios sobre o gˆenero Rhodnius Stal ´ especie do Brasil (Hemiptera Reduviidae). Rev. Bras. Biol. 14, 237–247. Lent, H., Wygodzinsky, P., 1979. Revision of the Triatominae (Hemiptera, Reduviidae), and their significance as vectors of Chagas disease. Bull. Am. Mus. Nat. Hist. 163, 520–529. Mat´ıas, A., de la Riva, J.X., Torrez, M., Dujardin, J.P., 2001. Rhodnius robustus in Bolivia identified by its wings. Mem. Inst. Oswaldo Cruz 96, 947–950. Miles, M.A., Feliciangeli, M.D., Rojas de Arias, A., 2003. American trypanosomiasis (Chagas’ disease) and the role of molecular epidemiology in guiding control strategies. BMJ 326, 1444–1448. Monteiro, F.A., Escalante, A.A., Beard, C.B., 2001. Molecular tools and triatomine systematics: a public health perspective. Trends Parasitol. 17, 344–347. ´ Monteiro, F.A., Barrett, T.V., Fitzpatrick, S., Cordon-Rosales, C., Feliciangeli, M.D., Beard, C.B., 2003. Molecular phylogeography of the Amazonian Chagas disease vectors Rhodnius prolixus and R. robustus. Mol. Ecol. 12, 997–1006. ´ ´ Oliveira, A.W.S., Silva, I.G., 2007. Distribuic¸a˜ o geografica e indicadores entomologicos ´ ´ Rev. Soc. Bras. de triatom´ıneos sinantropicos capturados no Estado de Goias. Med. Trop. 40, 204–208. Patterson, J.S., Schofield, C.J., Dujardin, J.P., Miles, M.A., 2001. Population morphometric analysis of the tropicopolitan bug Triatoma rubrofasciata and relationships

with Old World species of Triatoma: evidence of new World ancestry. Med. Vet. Entomol. 15, 443–451. Pavan, M.G., Monteiro, F.A., 2007. A multiplex PCR assay that separates Rhodnius prolixus from members of the Rhodnius robustus cryptic species complex (Hemiptera: Reduviidae). Trop. Med. Int. Health 12, 751–758. Pinho, A.P., Gonc¸alves, T.C.M., Mangia, R.H., Russell, N.S.N., Jansen, A.M., 1998. The ˚ 1859, naturally infected by Trypanosoma cruzi occurence of Rhodnius prolixus Stal in the state of Rio de Janeiro, Brazil (Hemiptera, Reduviidae, Triatominae). Mem. Inst. Oswaldo Cruz 93, 141–143. ˜ C.A., Brunstein, D., Collin-Delavaud, A., Souza, O., Ortega-Barr´ıa, E., 2003. Romana, Public policies of development in Latin America and Chagas disease. Lancet 362 (9383), 579. Rohlf, F.J., 1996. Morphometric spaces, shape components and the effects of linear transformations. In: Marcus, L.F., Corti, M., Loy, A., Naylor, G.J.P., Slice, D. (Eds.), Advances in Morphometrics. NATO ASI, Series A Life Sciences. Plenum Publication, New York, pp. 117–129. Rohlf, F.J., 1999a. tpsDig, version 1.18. Department of Ecology and Evolution, State University of New York, Stony Brook, NY, http://life.bio.sunysb.edu/morph/. Rohlf, F.J., 1999b. tpsRelw, version 1.18. Department of Ecology and Evolution, State University of New York, Stony Brook, NY, http://life.bio.sunysb.edu/morph/. Sanchez-Martin, M.J., Feliciangeli, M.D., Campbell-Lendrum, D., Davies, C.R., 2006. Could the Chagas disease elimination programme in Venezuela be compromised by reinvasion of houses by sylvatic Rhodnius prolixus bug populations? Trop. Med. Int. Health 11, 1585–1593. ¨ Schachter-Broide, J., Dujardin, J.P., Kitron, U., Gurtler, R.E., 2004. Spatial structuring of Triatoma infestans (Hemiptera Reduviidae) populations from northwestern Argentina using wing geometric morphometry. J. Med. Entomol. 41, 349–643. Schofield, C.J., Dujardin, J.P., 1999. Theories on the evolution of Rhodnius. Actual. Biol. 21, 183–197. Silva, I.G., Silva, J.L., Silva, H.H.G., Camargo, M.F., Moura, A.F., Elias, M., Santos, A.H., 1992. Distribuic¸a˜ o dos vetores da Tripanossom´ıase Americana capturados no ´ no per´ıodo de 1984/88. An. Soc. Entoambiente domiciliar no estado de Goias, mol. Bras. 21, 139–154. ´ Silva, R.A., Bonifacio, P.R., Wanderley, D.M.V., 1999. Doenc¸a de Chagas no estado de S˜ao Paulo: comparac¸a˜ o entre pesquisa ativa de triatom´ıneos em domic´ılios e ´ ´ notificac¸a˜ o de sua presenc¸a pela populac¸a˜ o em area sob vigilˆancia entomologica. Rev. Soc. Bras. Med. Trop. 32, 653–659. Silveira, A.C., Feitosa, V.R., Borges, R., 1984. Distribuic¸a˜ o de triatom´ıneos capturados no ambiente domiciliar, no per´ıodo de 1975/83, Brasil. Rev. Bras. Malariol. D. Trop. 39, 15–312. Soares, R.P.P., Barbosa, S.E., Dujardin, J.P., Schofield, C.J., Siqueira, A.M., Diotaiuti, L., 1999. Characterization of Rhodnius neglectus from two regions of Brazil using isoenzymes, genitalia morphology and morphometry. Mem. Inst. Oswaldo Cruz 94, 161–166. Sorensen, J.T., Foottit, R.G., 1992. The evolutionary quantitative genetic rationales for the use of ordination analyses in systematics: phylogenetic implications. In: Foottit, R.G., Sorensen, J.T. (Eds.), Ordination in the Study of Morphology, Evolution and Systematics of Insects: Applications and quantitative Genetic Rationales. Elsevier, New York, NY, pp. 29–53. Villegas, J., Feliciangeli, M.D., Dujardin, J.P., 2002. Wing shape divergence between ´ Rhodnius prolixus from Cojedes (Venezuela) and Rhodnius robustus from Merida (Venezuela). Infect. Genet. Evol. 2, 121–128.