Molecular validation of anthropophilic Phlebotominae sandflies (Diptera: Psychodidae) in Central Panama

Six Phlebotominae sand fly species are incriminated as biological vectors of human pathogens in Panama, but molecular corroboration is still needed. We aim at confirming the identity of Phlebotominae species documented as anthropophilic in Panama. Adult sandflies were collected from August 2010 to February 2012 in Central Panama using CDC light traps. Species confirmation was accomplished through molecular barcodes and allied sequences from GenBank. A total of 53,366 sand fly specimens representing 18 species were collected. Five species were validated molecularly as single phylogenetic clusters, but Psychodopygus thula depicted two genetically divergent lineages, which may be indicative of cryptic speciation.

In Panama, six anthropophilic (i.e., man-biters) species of Phlebotominae sandflies (Order Diptera, Family Psychodidae) have been implicated as vectors of Leishmania parasites, the causing agent of American cutaneous leishmaniasis (ACL). (1,2,3,4,5,6,7) Some of these taxa are also suspected vectors of Phlebovirus pathogens to a broad range of animal hosts, including humans. (8,9,10,11) Lutzomyia gomezi, Lutzomyia sanguinaria, Nyssomyia ylephiletor, Nyssomyia trapidoi, Psychodopygus panamensis, and Psychodopygus thula are widespread across forested areas of Panama feeding on various animal species depending on habitat quality and host availability. (12) Earlier taxonomic work using molecular markers supported the specific status of all these taxa in Central Panama, except for Lu. gomezi, for which significant lineage divergence was suggested. (6) Authors hypothesized Lu. gomezi to be a cryptic species complex based on the results of phylogenetic analysis using partial DNA sequences of the mitochondrial Cytochrome C Oxidase Subunit One gene (CO1). (6) This finding was of great epidemiological significance because morphologically identified Lu. gomezi was found infected with Leishmania naiffi, in pristine site at a high infection rate. L. naiffi causes cutaneous leishmaniasis (CL) in South America, but had never been reported from the country of Panama before. (6) A subsequent molecular study, using samples of Lu. gomezi from across the entire country, found low levels of genetic differentiation among populations, thus rejecting the hypothesis of linages diversification in this taxon. (13) Although there seems to be an acceptable level of agreement between morphology and DNA barcodingbased taxonomy for Phlebotominae sand fly species in Panama, (6) there is still a need to validate species boundaries using samples from additional sites, particularly from Central Panama. Herein, we revisit the molecular identity of Phlebotominae sand fly species documented as man-biters (12) in Central Panama, including more locations than the ones used in earlier work. (6) In so doing, we surveyed for sand fly specimens repeatedly in three ecologically distinct areas to account for spatial changes in the community metrics and to look for Leishmania parasite infection in the most prevalent sand fly species.
The study was conducted in the lowland tropical rainforest ecosystem of Central Panama, a region formerly known as the Panama Canal Zone. Detailed information on the sampling area, trapping design and effort was published elsewhere. (14,15) Captured sandflies were separated from other insects under a stereoscope, labeled with a unique code, and initially identified using female morphological characters (16,17) and the taxonomic nomenclature by Galati. (18) Well-preserved specimens were subjected to molecular analysis using the Barcoding region (5' prime region of the CO1 gene) (http://www.barcodeoflife.org/). A total of 184 samples from 13 species, initially identified based on morphological characters, were randomly taken from the total collected and processed molecularly to validate their taxonomic identity. DNA extraction, polymerase chain reaction (PCR)-amplification and sequencing were done following standard protocols. (19) To determine whether our specimens were mistakenly classified or confused with other species within Phlebotominae, we employ Basic Local Alignments Search Tool (http://blast.ncbi.nlm.nih.gov/) to the allied nucleotide CO1 sequences of Phlebotominae in GenBank, includ-   (6) which were used here to validate species taxonomic designations. Brumptomyia galindoi (GU001735.1) was used as outgroup. 100% bootstrap values are shown in highly supported molecular clusters. Sequence codes represent the abbreviation of three sampling areas of central Panama: Barro Colorado Island (BCI), Achiote (ACH) and las Pavas (PVAS), plus the polymerase chain reaction (PCR) cycle code (See also maps of the study area previously published in Eastwood et al. (14) and Loaiza et al. (15) ).
ing those reported in Azpurua and collaborators. (6) We built a Neighbor-Joining (NJ) phylogenetic tree using all these CO1 sequences in MEGA v.5.1, (20) with Kimura 2 parameter (K 2 P) distances, and bootstrapped the topology with 1000 replicates to obtain branch support.
In addition, female sandflies were pooled in groups of up to 50 individuals, according to species, trap location, and height, and tested for infection with Leishmania parasites. The DNA of pools was extracted using the Biosprint® 96 DNA Blood kit (Qiagen) DNA Blood Kit on a BioSprint® 96 extraction robotic platform (Qiagen). Pooled DNA was then used to amplify the minicircle kinetoplast DNA of Leishmania parasite using the PCR, primers, and cycling conditions reported in Montalvo and collaborators. (21) A second confirmatory PCR, targeting the entire length of the Leishmania Hsp70 gene (1,286 base pairs) was conducted following the protocol, primers, and PCR cycling conditions reported in Cardoso da Graça and collaborators. (21) DNA amplicons representatives of positive samples were subjected to Restriction Fragment Length Polymorphism (RFLP) technique (22) and identified by comparing of RFLP banding patterns with those published in Montalvo and collaborators. (23) Leishmania infection rate in sandflies was calculated overall and per species using maximum likelihood estimates (MLE) of pooled samples using an online calculation tool available at http:// www.cdc.gov/ westnile/resourcepages/mosqSurvSoft.html.
We generated valid DNA barcode sequences for 160 female Phlebotominae samples out of a total of 184 attempted (> 85% success rate). Failures were due to double peaks in the electropherograms recovered in multiple amplification cycles. We removed these samples from further analyses. 102 unique CO1 sequences or haplotypes were obtained from samples initially assigned to Phlebotominae based on morphology. All these haplotypes were unambiguously aligned and no insertions or deletions were found. The absence of pseudogenes was established by the lack of stop codons, low pairwise divergence and clear electropherograms. Individual length for these CO1 haplotypes ranged from 615 to 649 base pairs (bp), with a final alignment length of 610 bp (GenBank accession numbers MN257585-MN257605). Phlebotominae CO1 sequences formed 14 DNA barcode clusters in the NJ phylogenetic tree, of which some matched with 99% homology the barcode dataset of sand fly species previously reported from Barro Colorado Island (BCI) by Azpurua and collaborators. (6) These clusters were well-supported statistically, with the majority of haplotypes being found in the three sampling areas   (6) and therefore, they were validated as single evolutionary units or molecular species through phylogenetic analysis. However, Ps. thula (formerly known as Lutzomyia carrerai thula) comprised two moderately divergent molecular lineages, which had not been reported earlier. One of these lineages matched with 99% homology the barcode sequence of Ps. thula (GenBank accession GU001753.1) described by Azpurua and collaborators, (6) therefore we call this taxon Ps. thula sensu stricto. Another unidentified lineage in the same cluster showed more than 2% genetic distance from Ps. thula s.s., hence it may be a different taxonomic unit under incipient speciation (Figure). The NJ phylogenetic tree comprised two main clades where samples from the genera Lutzomyia and Pressatia grouped together in the most basal clade, and away from the other remaining genera in a second derived clade (Figure). Samples assigned to the following genera: Lutzomyia (Lu. gomezi, Lu. sanguinaria), Psychodopygus (Ps. panamensis, Ps. thula), Nyssomyia (Ny. trapidoi, Ny. ylephiletor), Pressatia (Pressatia camposi, Pressatia dysponeta), Psathyromyia (Psathyromyia carpenteri), Bichromomyia (Bichromomyia olmeca bicolor), Trichopygomyia (Trichopygomyia triramula), Micropygomyia (Micropygomyia trinidadensis) and Pintomyia (Pintomyia ovallesi) all clustered jointly with others from the same genera, and were separated from samples in other classes by roughly 6% to 10% genetic distances (Figure).
A total of 53,366 specimens representing 18 species of Phlebotominae sandflies were gathered from three sampling areas of Central Panama. Comparison of Alfa diversity and richness metrics suggest that Phlebotomi-nae sand flies are more diverse and species rich at the ground level of BCI (i.e., A pristine site) (Table I), while its overall relative abundance does not vary significantly among sampling areas (Kruskall-Wallis p = 0.45) or between vertical strata (Mann-Whitney U = 37; p = 0.71). However, when the latter analysis was performed separately at each sampling site, Phlebotominae sand fly relative abundance differed significantly between vertical strata, being more numerous in the understory of BCI (Mann-Whitney U = 89; p = 0.05), but not in Achiote (ACH) (Mann-Whitney U = 146.5, p = 0.63) or in Las Pavas (PVAS) (Mann-Whitney U = 149.5, p = 0.70). The most prevalent species were Ps. panamensis (58.63%), Ny. trapidoi (14.78%), and Lu. gomezi (14.07%) in that order, followed by Lu. sanguinaria (3.34%) plus other 13 rare species. Ps. panamensis was equally prevalent in all three sampling areas, but the relative abundance of Ny. trapidoi and Lu. gomezi differed among sites; both species being more prevalent in ACH and PVAS, which are ecologically disturbed areas (Table II) The natural history of both anthropophilic and zoophilic Phlebotominae sand fly species and their roles as vectors of pathogens to humans are well acknowledged in Panama owing to more than 100 years of scientific research. (12) Nonetheless, in depth studies about the ecology, behavior and control of this medically important group of insects are still challenging due to high species diversity in the Neotropical region plus limitations to identify fresh samples accurately and rapidly. Taxonomic identification of Phlebotominae sand fly in Panama is largely based on adult morphological characters, but morphological keys are often incomplete, (16,17,18) and even senior taxonomists cannot sometimes distinguish among separate species, making it difficult to differentiate between vector and non-vectors. To date, only two studies have tested species boundaries in Phlebotominae sand fly with molecular approaches in Panama. (6,13) Azpurura and collaborators (6) generated molecular barcodes for 18 species of Lutzomyia from the forest understory of BCI, using the same sampling procedure used here. Authors suggested that Lu. gomezi and Dampfomyia vespertilionis were complexes of isomorphic species, while Ny. trapidoi, Ps. panamensis and Ps. thula were nominated as single molecular clusters. Our results based on more sites across Central Panama, including samples from the forest canopy of disturbed and undisturbed areas, contradict this finding and suggests that Lu. gomezi represents a single molecular entity, which also agrees with previous efforts to define the taxonomic status of Lu. gomezi across Panama. (13) In contrast, Ps. thula is likely two divergent lineages under incipient speciation not reported in earlier research. (6) Individuals from the two lineages of Ps. thula came from BCI, ACH and PVAS, which reinforces the hypothesis of lineage divergence for this taxon. To date, studies about the biology of Ps. thula are still incomplete in Panama. Larvae develop in decaying leaves (e.g., forest leaf-litter) of deeply shaded pristine forest environments, while adults use fallen tree trunks and green plants for oviposition, mating and also as diurnal resting sites. (24) Females of Ps. thula are most active at the ground level during the day, when the risk of ACL transmission to animals and humans likely increases by this species. (12,25) However, the role of Ps. thula as a vector of Leishmania (V) panamensis, the main parasite causing ACL in Panama, has still not been confirmed, (12,24,26) and it was not supported by our results either. Ny. trapidoi was the only species infected with Leishmania parasite in this study, with just one positive pool gathered from the understory of the disturbed site (ACH). Our results partially agree with findings by Azpurura and collaborators (6) where Ny. trapidoi was also found infected with a Leishmania parasite known as L. naiffi. However, the overall Leishmania infection rate in this study was extremely low in comparison to previous work, (6,7) and we could not corroborate the presence of L. naiffi in our positive pool either. Consequently, while our outcomes support the transmission involvement of Ny. trapidoi for human pathogens in forest environments of Central Panama, these findings must be validated in future studies.