Rhodnius pallescens microsatellite markers for population genetic analysis in Rhodnius ecuadoriensis: preliminary assessment

Rhodnius ecuadoriensis Lent & León (Hemiptera: Reduviidae) es el prinicipal vector de    la enfermedad de Chagas en Ecuador, donde la estructura genética de sus poblaciones es poco conocida. Nosotros probamos seis Repeticiones Cortas en Tamdem (RCT) de R. pallescens Barber en poblaciones selváticas y domésticas de R. ecuadoriensis. Dos microsatelites fueron monomórficos, dos dieron resultados ambiguos y dos fueron polimórficos (16 y 19 alelos) y fueron utilizados para análisis. Los resultados de las frecuencias alélicas, AMOVA y los pruebas Bayesianas para genética favorecen la teorí­a de la existencia de una sola población. Estos resultados preliminares sugieren que las poblaciones selváticas y domésticas d R. ecuadoriensis intercambian frecuentemente migrantes. Por consiguiente el control de la Enfermedad de Chagas requiere vigilancia entomológica continua en la costa del Ecuador.


INTRODUCTION
Chagas Disease (CD), caused by Trypanosoma cruzi and transmitted via the feces of infected triatomine bugs, is endemic in western Ecuador Rhodnius ecuadoriensis is the main vector in the coastal region (Grijalva et al., , 2012(Grijalva et al., , 2014, where it occupies both artificial and sylvatic ecotopesmainly Phytelephas aequatorialis palm trees and squirrel nests (Abad-Franch et al. 2001Grijalva et al. 2012Grijalva et al. , 2014Suárez-Dávalos et al. 2010). Quantitative phenotypic analyses suggest that wild populations are involved in the re-infestation of insecticide-treated households , Grijalva et al. 2014, which is common in central coastal Ecuador (Grijalva et al. 2011;2014). Yet, the population dynamics of this vector species remain largely unknown; in particular, powerful molecular genetics tools have never been used for the study of inter-population relationships in R. ecuadoriensis as they have in other important CD vectors (e.g., Fitzpatrick et al. 2008). Here we address R. ecuadoriensis population connectivity in central coastal Ecuador by testing the utility of Short Tandem Repeat (STR) markers developed for the closely-related species, R. pallescens (Harry et al. 1998, Abad-Franch et al. 2009).

MATERIAL AND METHODS
Fieldwork was conducted in 2009-2010 in the province of Manabí, where CD is endemic and insecticide-based vector control strategies perform poorly (Grijalva et al. 2011). Triatomines were manually collected in domestic-peridomestic structures and nearby sylvatic habitats in eight localities (Fig. 1).
All collections were conducted following protocols approved by the Institutional Review Boards of Catholic University of Ecuador and Ohio University. Overall, 270 specimens were studied, 150 collected in sylvatic habitats and 120 in artificial structures. Sampling included 100 individual ecotopes (1 to 16 bugs/ecotope, median=2, interquartile range 2-3), including squirrel nests (Sciurus stramineus), hen nests, houses, bird nests (Campylorrynchus fasciatus and one Synallaxis sp.), mouse and rat nests, guinea pig enclosures, and timber piles. Specimens included all stages except first-instar nymphs, with adults representing about 67 % of the sample and evenly distributed in relation to sex. Hence, we are confident that our sample fairly represents local population diversity. DNA was isolated and purified using DNeasy® extraction kits (Qiagen) following the manufacturer's protocol for animal tissue. Microsatellite amplification followed a protocol modified from Harry et al. (1998), using primers designed for six loci (L3, L9, L13, L25, L43, L47) of R. pallescens, a species phylogenetically close to R. ecuadoriensis (Abad-Franch et al. 2009). For PCR, we used the GoTaq® Colorless Master Mix kit (Promega), adding 6.4 μl of MgCl 2 25mM, 0.4pM of each fluorochrome-labeled primer, and 10ng/μl of DNA for a final reaction volume of 25μl. PCR was carried out in a PTC-100® thermal cycler (MJ Research), with a denaturation step (95 ºC for 5 min) followed by 40 cycles of denaturation (95 ºC, 30 sec), annealing (50 ºC, 30 sec) and extension (72 ºC, 30 sec), and a final extension step (72 ºC, 5 min). For loci L13 and L43, a nested PCR was carried out using the product of the first PCR asthe template and the protocol aboveexcept that no additional MgCl 2 was used.
We first screened the full microsatellite panel with 30 randomly selected bug samples to check for polymorphism over all loci. This scan revealed that two loci were monomorphic (L9 and L25) and two (L3 and L43) yielded ambiguous allele signatures that precluded confident scoring. These four loci were therefore disregarded in subsequent analyses, and only the two polymorphic markers yielding unambiguous alleles (L13 and L47) were used to analyze the full sample of triatomines. Genotyping was performed by capillary electrophoresis in an ABI 313xl with a standard G500 LIZ ladder (Applied Biosystems); .fsa files from capillary runs were then analyzed with Peak Scanner™ 1.0 for manual allele scoring. A total of 93 samples yielded no PCR product; results below are therefore based on 177 bugs, 82 from sylvatic and 95 from artificial ecotopes. Different analyses were conducted in Excel® spreadsheets, Arlequin 3.1 (Excoffier et al. 2005), and STRUCTURE 2.2 (Pritchard et al. 2000).

RESULTS
Overall, 16 alleles were identified for L13 and nine for L43; in both cases, allele frequencies were similar in wild and synanthropic populations, with only one L47 allele found at higher frequency among wild bugs ( Fig. 2 and Table). Allele frequency (f) ratios, estimated as f(wild)/f(synanthropic), for 11 shared alleles on two microsatellite loci (L13 and L47) in Rhodnius ecuadoriensis populations. Point estimates (solid circles) and 95% confidence intervals (CIs, between short horizontal lines) are shown, with allele codes given on the x-axis. The asterisk indicates the only allele (L47-4) for which we found evidence of a higher frequency in one of the populations (wild); for the rest, the fact that CIs include 1 (horizontal dotted line) indicates that allele frequencies are not significantly different at the 5% level.
Gene diversity was slightly higher in the wild (0,799 ± 0,023) than in the synanthropic population (0,783 ± 0,021). An analysis of molecular variance (AMOVA) suggested that an overwhelming proportion of genetic variation (98,7 %) lies among individuals within populations. Both AMOVA and an exact test of population differentiation suggest that allele frequencies in both populations are Taken together, these results all suggest that both wild and synanthropic populations belong to a single meta-population occurring across the study region. To further test this possibility, we conducted a Bayesian test of population subdivision using STRUCTURE 2.2. The analysis estimates and compares the likelihood of the data over a set of possible numbers of subdivisions (K); we run 20 iterations for each K value between 1 and 9, without a priori assignment of individuals to putative clusters. ΔK was then calculated following Evanno et al. (2005): the most likely number of subdivisions corresponds to the K value that maximizes ΔK. In our case, as shown in Fig. 3, there was strong support for the existence of a single genetic cluster (K=1).

DISCUSSION
We have presented the first population-level analysis of microsatellite loci in R. ecuadoriensis, a locally important CD vector that frequently reinvades and re-colonizes insecticide-treated households. The present study is a first step towards more detailed appraisals, which should include further polymorphic loci and the design of species-specific primer pairs (e.g., Fitzpatrick et al. 2008). The development of specific markers for R. ecuadoriensis is needed to improve our understanding of the vector's population dynamics, especially as wild populations seem to play a key role in CD transmission (Grijalva et al. 2014).
Even if still preliminary, the results of our analyses are suggestive of a single meta-population scenario in which R. ecuadoriensis migrants frequently move between sylvatic and artificial ecotopesat a frequency high enough to homogenize both gene pools. This agrees with ecological and quantitative phenotypic assessments showing both frequent reinfestation of treated households (Grijalva et al., 2011(Grijalva et al., , 2014) and a lack of morphological or morphometric differentiation of wild and synanthropic specimens ).
In line with previous findings (Grijalva et al. 2011), our data thus suggest that longitudinal surveillance will be a key requirement of long-term CD control in central coastal Ecuador. Entomological surveillance usually performs better when the community takes on an active role in reporting infestation (Abad-Franch et al. 2011); however, a timely, professional response of vector control services is also needed, and this should include not only insecticide application but also environmental management of peridomestic ecotopes prone to harbor R. ecuadoriensis breeding coloniesparticularlypalm trees and chicken coops in our study setting (Abad-Franch et al. 2005, Grijalva et al. 2011.

CONCLUSIONS
• Only two of six microsatellites developed for R. pallescens showed polymorphic loci in R. ecuadoriensis populations.

•
Results of our analyses are suggesting only one genetic cluster is present in populations of R. ecuadoriensis in the province of Manabi, where migrants appear to move between sylvatic and artificial ecotopes.