Structure and Diversity of the Microbiome in Amazonian Sand Flies: Insights into Vector–Microbe Interactions

Abstract

This study uses high–throughput sequencing of the 16S rRNA gene and specific PCR to analyze the microbiome and identify secondary endosymbionts in sand flies from the Amazon region, important vectors of parasitic and viral diseases. Specimens of Psychodopygus, Trichophoromyia, Nyssomyia, Trichopygomyia and Brumptomyia were collected and analyzed. The results revealed that the richness, diversity, and composition of the microbiome are influenced by several factors, such as insect species specific composition, and insect sex. The core microbiome community was represented by 18 genera, with Novosphingobium, Cutibacterium, Methylobacterium and Staphylococcus being the most prevalent. The highest diversity at the genus level was observed in sand flies of epidemiological relevance as Psychodopygus and Nyssomyia, dominated by Novosphingobium (66.5%), Cutibacterium (29.4%) and Methylobacterium (20.4%), while in non–vectors such as Trichophoromyia, Delftia predominated (59.9%). Endosymbiont analysis showed a high prevalence of Cardinium (20%) and Wolbachia (33%), as well as the presence of Spiroplasma, Arsenophonus and Rickettsia. In addition, some bacterial genera related to the inhibition of parasite development, which have entomopathogenic activity and are involved in the degradation of insecticides were identified. Our results are relevant and contribute to the knowledge of the characterization of the microbiome and the endosymbionts in leishmaniasis vectors in the Amazon region and show promise for improving vector management, highlighting the importance of investigating their interaction with pathogens and their impact on vector biology.

Graphical Abstract

Supplementary Information

The online version contains supplementary material available at 10.1007/s00248-025-02663-5.

Introduction

Sand flies (Diptera: Psychodidae) are insects of public health interest due to the ability of females of some species to transmit various pathogens, including viruses such as Phlebovirus, Vesiculovirus, Orbivirus [1–3], and parasites of the genus Leishmania, causative agent of leishmaniasis, a relevant disease due to the diversity of its clinical manifestations and wide geographical distribution [4]. It is also relevant for the transmission of bacteria such as Bartonella bacilliformis causative agent of Carrion fever, mainly in Peru [5], which has motivated in the last five years the search for these bacteria in sand flies from Brazil, Mexico and Colombia [6–8].

Leishmaniasis in Colombia is a zoonosis triggered by the infection of several species of Leishmania, including in the subgenera Leishmania and Viannia such as L. (L.) amazonensis, L. (L.) infantum chagasi, L. (L.) mexicana, L. (V.) braziliensis, L. (V.) guyanensis, and L. (V.) panamensis [9, 10]. Most of these parasite species are mainly transmitted by females of the genera Lutzomyia, Nyssomyia, Pintomyia, Psychodopygus among others [11]. In the Amazon region, specifically in the departments of Amazonas and Caquetá [12], leishmaniasis has affected 1,676 people between 2020 and 25, where the most frequent clinical form is cutaneous leishmaniasis (CL) [13, 14]. In this region, the proven or suspected vectors involve individuals of the genera Psychodopygus and Nyssomyia [11, 12, 15], which, in addition to being highly anthropophilic, have been shown to have a remarkable plasticity and diversity of blood food sources, including preference for domestic species such as Bos taurus, Gallus gallus, and Sus scrofa [16–18], as well as wild species like Caniculus paca, Dasypus novemcinctus, Dasyprocta leporina, Saimiri macrodon, and Tamandua tetradactyla [16, 18, 19].

Particularly in this region, access to diagnosis and treatment of tropical diseases such as leishmaniasis, dengue and malaria represents a great challenge, especially for the communities farthest from urban centers [20], due to the high costs associated with travel and accommodation for those affected, as well as the lack of equipped health centers within the territories [21]. These limitations not only hinder the early detection of cases but also increase the risk of spreading these tropical diseases and contribute to under–registration, preventing an accurate assessment of their impact on the population. In addition, the Amazon rainforest suffers increasing environmental impacts that influence the ecology of tropical diseases, such as mining and its resulting pollutants, deforestation, chemical control of insect pests for crop management, and the direct influence of climate variability [22–25]. These activities transform the surrounding environment, impacting the quality of water, soil, and atmosphere, as well as non–target species of the surrounding fauna and flora, affecting the sustainability of the Amazonian ecosystems and the well–being of the different ethnic groups that inhabit the region [24, 26, 27]. The considerable functional implications of these changes may also significantly impact the increased interaction between sand flies and human and animal hosts, as well as feeding dynamics, changes in the life cycles of these insects and their behaviors, and even the composition and diversity of the bacterial communities of these vectors, which often depend on the conditions of the local or regional environment [28, 29].

The bacterial communities inherited and acquired by sand flies during the feeding process in the immature stage [30], or by adults during feeding on different sources of sugar and blood, the latter exclusively in the case of females [31–33], can generate symbiotic associations that not only influence their biology but also facilitate their adaptation and ubiquity in urbanized and altered environments, or even in the presence of insecticides used in vector control campaigns, favoring their survival in challenging conditions, influences vectorial competence and the ability of sand flies to transmit pathogens such as Leishmania, and even participate in the dispersal of potentially pathogenic bacteria or viruses in plants and animals [31, 34–36], representing a high-risk factor for transmission [23, 37, 38].

Currently, the study of the microbiome, the entire gene content of the microorganisms present in a sample, of insect vectors has generated new vector control strategies, specifically through the use of secondary endosymbionts, such as Wolbachia in Aedes aegypti that impact the replication of the dengue, Zika and Chikungunya arboviruses [39–41] or Microsporidia in Anopheles [42], which prevent the development of Plasmodium [43]. This has prompted numerous investigations aimed at understanding the current status of the diversity of these endosymbionts in sand flies and their potential role in the development and/or transmission of Leishmania. In sand flies from the Neotropics, the detection of Wolbachia has been recorded in approximately 25 species, mainly from Colombia, Mexico, and Brazil, including some vector species such as Lu. (Lut.) longipalpis, Lu. (Trl.) gomezi, Pi. (Pif.) evansi, Ps. davisi and Ps. panamensis [7, 44–50], as well as in some species not related to Leishmania transmission such as Brumptomyia hamata, Sciopemyia sordellii or Micropygomyia (Mic.) micropyga. The information available does not yet report precisely the interaction with Leishmania or its influence on aspects such as reproduction or fitness of sand flies, nor have Wolbachia strains been isolated, and the reported studies are at the level of estimation of infection rates and phylogenies [7, 44, 45, 48, 51]. Like Wolbachia, other endosymbionts that can induce reproductive phenotypes in their hosts have also been detected in sand flies. Among them, Arsenophonus, recently detected in Lu. (Lut.) longipalpis, Ps. panamensis, and Pi. (Pif.) evansi [52] can not only modify the reproductive system in other insects through mechanisms such as androcide [53], but it has also been observed that strains of this bacterium can influence susceptibility to insecticides [54].

Cardinium, on the other hand, was detected in Mi. (Mic.) cayennensis, Pi. (Pif.) evansi, Psathyromyia (Psa.) shannoni y Trichopygomyia triramula [45, 48], and cause cytoplasmic incompatibility (CI) in Encarsia pergandiella [55]. In the case of Rickettsia, which has also been recently recorded in Pi. (Pif.) pia [48], Lu. (Lut.) longipalpis [56] and Pa. (Xip.) aclydifera [49], it has been associated with CI phenotypes in Nisediocoris tenuis [57], and androcide in Adalia decempunctata L. [58]. While Spiroplasma, detected in Lu. (Lut.) longipalpis, Phlebotomus chinensis, and Ph. papatasi [51, 56, 59, 60], besides inducing CI, have been associated with a new mechanism called Male killing (MK), which causes embryonic death without sex distinction [61] and increases resistance to infection by bacterial and fungal pathogens in Drosophila melanogaster [62]. In recent years, reports of other bacterial species different from endosymbiotic bacteria show their capability to affect or favor vector competence or parasite development in sand flies [63, 64].

Studies on the microbiome of sand flies most frequently include adult stages, while characterization of microbial communities has been performed mainly in the gut and whole tissue to address the microbiome by culture–dependent and culture–independent methods [56, 63, 65, 66]. So far, studies conducted in Colombia have explored the composition of the microbiome of sand flies in the Andean and Caribbean regions [44, 45, 48, 52, 56, 66]. Therefore, it is essential to expand knowledge about the composition of the microbiome and the presence of secondary endosymbionts in sand flies from other regions of the country, including those where research is limited, such as the Colombian Amazon region, in order to assess their possible influence on pathogen transmission and explore complementary biological control strategies.

Materials and methods

Collection permits

The specimens were collected under the framework permit for the collection of wild species for scientific research without commercial purposes, granted by the National Environmental Licensing Authority to the Universidad Nacional de Colombia (Resolution No. 0255 of March 14, 2014, article 3), in compliance with Decree No. 1376 of 2013 of the Ministry of Environment and Sustainable Development, and transported with mobility certificate No. 59453.

Study Area and Sand Flies Collection

The study area includes the municipalities of Florencia in the department of Caquetá and Leticia in the department of Amazonas (Fig. 1a). They are part of the Colombian Amazon region, an extensive area of humid tropical forest (Bh–T) [67], with relative humidity above 80% [68]. Caquetá has a mean annual temperature ranging from 24.8 to 25.9 °C and annual rainfall between 2,483 mm and 4,385 mm [68]. Amazonas has average temperatures ranging from 25.3 to 25.7 °C and an annual rainfall between 2,660 mm and 3,538 mm [68].

Fig. 1.

Sand flies collection sites. (a) Geographical location of Amazonas and Caquetá departments, Colombia, South America. (b) Jericó, Sebastopol, Santo Domingo, and Macagual, collection points in the municipality of Florencia, Caquetá. (c) Tanimboca and San Pedro de los Lagos, collection points in the municipality of Leticia, Amazonas. Source of layers [69–72] and Qgis, v. 3.34.12 [73]

Collections were conducted during August 2024 on the sidewalk of Jericó, Sebastopol, the Amazon Research Center, CIMAZ–Macagual, located to the northwest of the Florencia municipality (Fig. 1b), and in Santo Domingo, a rural settlement located to the west of the municipality. While in Leticia (Fig. 1c), the collection was conducted during November 2024 in San Pedro de los Lagos, an area inhabited by the Ticuna people approximately 6.5 km from the urban area. In all cases, sand flies were captured in peridomiciliary and extradomiciliary areas and gallery forests near primary forests. The collection was conducted during two to three consecutive days using 12 CDC (Centers for Disease Control and Prevention) light traps operated for 12 h (18:00 to 6:00). During the second day, a Shannon trap was also used for one hour (18:00 to 19:00) at the site with the highest capture of sand flies by CDC. Sample collection was complemented with the search for possible resting places using mouth aspirators and Prokopack [74] aspirators for one hour (18:00 to 19:00).

Taxonomic Identification

A total of 1,104 specimens were collected in Amazonas (N = 191; 17.3%) and Caquetá (N = 913; 82.7%), represented by 68.4% (N = 755) females and 31.6% (N = 349) males, which were stored at − 20 °C until identification by classical taxonomy was completed. For morphological identification, under sterile conditions, the head of each specimen was detached, and the last three abdominal segments were cut as described [75]. They were subsequently clarified and mounted on Euparal medium (Carl Roth GmbH + Co. KG, Germany) following the procedures described by Posada–López [76]. Identification was performed according to the classification proposed by Galati [77], Young and Duncan [78], and bibliographic records of sand flies from the Amazon region of Colombia and Brazil [15, 17, 76, 79–81]. The abbreviation of the genre and subgenre names follows that proposed by Marcondes [82].

DNA Extraction of Sand Flies

A total of 355 specimens corresponding to the thorax and proximal segments of 301 females and 54 males were selected for this study, representing 32.1% of the total abundance. Of these, 97 specimens (91 females, 6 males) were classified into 22 pools, consisting of two to six individuals of the same species and sex per sample, for total genomic DNA extraction using the Quick–DNA Tissue/Insect Miniprep kit (Zymo Research, Irvine, CA, USA), following the manufacturer’s instructions, and extraction controls included 70 µL aliquots of Core Community (Zymo-Research, Irvine, CA, USA) and ultrapure, autoclaved, irradiated Type I water.

In the remaining 258 samples, DNA was extracted using the high salt concentration protocol described by Porter et al. [83], which included a mechanical lysis step with micropestles, heat shock lysis, and protein precipitation with proteinase K (Thermo Fisher Scientific Baltics UAB). For this procedure, 90 females were grouped into 22 pools with two to six individuals of the same species per sample, while 168 (119 females, 49 males) were analyzed individually. In all cases, before processing, each specimen was carefully rinsed sequentially with 70 and 96% molecular grade ethanol and sterile ultrapure water. The DNA obtained was analyzed spectrophotometrically with an N60 nanophotometer (IMPLEN, Germany) and subsequently used for molecular confirmation of sandfly species by partial amplification of the cytochrome oxidase subunit I (COI) gene [84]. Once sand flies were confirmed by integrative taxonomy, microbiome characterization was performed from a batch of 44 samples (18 pools, 10 species), and endosymbiont detection was performed on all samples processed in this study (n = 209 samples, 30 species, 355 specimens).

Microbiome Composition by high–throughput Sequencing

For the analysis of the microbiome by next–generation sequencing (NGS), samples were selected based on their epidemiological relevance, considering the role of the species as proven or potential vectors of pathogens. In addition, priority was given to the representation of common species between the two departments studied, with a particular focus on females due to their importance in the transmission of Leishmania. In total, composition was analyzed of 44 specimens grouped in 18 pools, belonging to 11 species and five genera distributed as follows: Br. mesai (pool = 1; Specimens = 1), Ny. antunesi (pool = 3; Specimens = 8), Ps. ayrozai (pool = 3; Specimens = 4), Ps. davisi (pool = 4; Specimens = 21), Ps. carrerai thula (pool = 1; Specimens = 1), Ps. chagasi (pool = 1; Specimens = 1), Ps. paraensis (pool = 2; Specimens = 8), Th. howardi (pool = 1; Specimens = 3), Th. velezbernali (pool = 1; Specimens = 3) and Ty. witoto (pool = 1; Specimens = 1) (Table S1).

Bacterial DNA integrity and possible host DNA interference during amplification were assessed by specific PCR, partially amplifying the small subunit of the 16S rRNA gene with primers 27 F and 1492R, resulting in a fragment of ~ 1400 bp [85]. The PCR reaction contained 1X buffer, 2 mM of MgCl₂, 0.2 mM of dNTPs, 0.25 µM of each oligonucleotide (forward and reverse), 1U of Taq DNA polymerase (Excel taq, SMOBIO Technology, Hsinchu Science Park, Taiwan) 1 mg/mL of BSA (Scientific Inc., Waltham, MA, USA), 10–25 ng of genomic DNA and ultrapure water employing the thermal profile previously described by Espejo et al. [85]. Escherichia coli (ATCC) DNA was used as a positive control, and ultrapure water as a negative control. Amplicons were stained with EZ–Vision (Amresco, USA) and separated by electrophoresis on a 1.5% agarose gel for 50 min at 80 V. To compare the size of the bands, the 100 bp plus molecular weight marker (Thermo Fisher Scientific Baltics UAB) was used. Visualization of the products was performed on the Essential V6 transilluminator (UVITEC–Cambridge).

For targeted sequencing of the hypervariable region V3–V4 of the 16S rRNA gene, specific amplicons were prepared with primers 341 F and 806R were used to obtain a product of approximately 470 bp [86]. The PCR was optimized using Phusion High–Fidelity PCR Master Mix (New England, Biolabs) under standard conditions with 35 denaturation cycles and an annealing temperature of 55 °C. The amplified products were used for library construction compatible with Illumina platforms using the NEBNext Ultra™ II PCR–free DNA library preparation kit NEBNext Ultra™ II (New England, Biolabs). And sequenced on a HiSeq 2500 PE250 platform (Illumina), with 250–cycle reads in each paired–end direction.

Detection of Bacterial Endosymbionts by Specific PCR

To detect the presence of Arsenophonus, Cardinium, Microsporidia, and Spiroplasma in sand flies, markers targeting to the small subunit of the 16S rRNA gene were used, while for Wolbachia, the wsp gene encoding the major surface protein was partially amplified (Table S2). The procedure was performed on the total DNA of 209 samples, represented by 355 sand flies, including the pools that were considered for the analysis of the microbiome (Table S1).

For the detection of Arsenophonus and Cardinium, the primers ArsF1–ArsF2 and CLOF1–CLOR, respectively, were used (Table S2). Cloned fragments of the 16S rRNA gene from A. nasoniae [52], and Ca. cardinium [48] were included as positive controls. The reaction conditions and thermal profile followed those previously reported by Duron et al. [87]. The Microsporidia detection, the primers ss18f and ss1492r (Table S2), and a plasmid with a cloned fragment of Microsporidia MB from An. arabiensis [48] were used as a positive control. Thermal cycling conditions are described by Vossbrinck et al. [88]. For Spiroplasma, primers SPIRO1 and SPIRO2 (Table S2) were used, and a plasmid with a cloned fragment of the 16S rRNA gene from S. poulsoni (SPHY1) was included [52]. The reaction conditions and thermal profile followed those previously reported by Karatepe et al. [60]. Whereas the presence of Wolbachia was confirmed using primers wsp81F and wsp691R (Table S2). In this case, a plasmid with a cloned fragment of Wolbachia DNA (Supergroup A, strain wMel) [48] was used as a positive control. The reaction conditions and thermal profile were followed previously reported by Zhou et al. [89].

For each endosymbiont, ultrapure water was included as a negative control, except for Wolbachia, where DNA from Ae. aegypti strain Rockefeller, donated by the Program for the Study and Control of Tropical Diseases (PECET), was used. Amplicons from each reaction were stained with EZ–Vision (Amresco, USA), and separated by electrophoresis on a 1.2% agarose gel with 1X TBE buffer for 50 min at 80 V. To compare the size of the bands, the 100 bp plus molecular weight marker (Thermo Fisher Scientific Baltics UAB) was used. Visualization of the products was performed, and PCR products of the expected size were subjected to DNA sequencing in both directions using Sanger methodology. Estimation of the minimum infection rate in positive samples was calculated using the following formula: Minimum infection rate (MIR) = Number of sand flies detected with the endosymbiont/Total specimens tested x 100 [90].

Bioinformatic and statistical analysis of bacterial communities and endosymbionts present in sand flies of the Amazon biome

Bacterial Community Composition and Diversity

To determine the frequencies of bacterial genera and species present in the samples, the DADA2 software package (version 1.16) was used [91]. Sequential processing included filtering by quality, assembly, and removal of potentially chimeric sequences. Taxonomic assignment of each unique amplicon sequence variant (ASV) was defined using the naive Bayesian classifier (RDP) with taxonomic reference Silva database, version 138.1 [92]. The sample–matched datasets were processed using the DADA2 workflow version 1.26.0, and procedures for analyzing the matched reads were evaluated with parameters considering the number of sequencing cycles and the average amplicon size. The specific parameters used for filtering were as follows: “filterAndTrim(fnFs, filtFs, fnRs, filtRs, compress = TRUE, truncQ = 2, truncLen = c(226,226), trimLeft = c(1,1), maxN = 0, maxEE = c(2,2), rm.phix = TRUE, matchIDs = TRUE, multithread = TRUE)”.

After taxonomic annotation and removal of non–relevant data, such as sequences from animal chloroplasts and mitochondria, an additional curation of the ASVs with the highest frequency in the samples that did not reach species–level annotation with SILVA 138.1 was performed. For this, a similarity search was performed in updated databases, including LPSN [93], EzBiocloud [94], and NCBI’s 16S rRNA gene database (BlastN) [95], using sequences from type strains of Bacteria and Archaea. Statistical analysis was performed using the online tool Microbiomeanalyst [96]. The relative abundance of profiles of bacterial species generated from the analysis in Microbiomeanalyst was downloaded to analyze the relationship between shared and unique bacterial species in the epidemiologically relevant sand flies, using a network performed in Cytoscape 3.10.3 software [97].

Plots of alpha diversity abundances considering Shannon, Chao1, and Simpson indices, and a heat map for the representation of the core microbiome were performed by analysis of variance (ANOVA) on the whole data set. Assessment of beta diversity was performed by Bray–Curtis dissimilarity distance, and differences were assessed by PERMANOVA. Principal coordinate analysis (PCoA) and non–metric multidimensional scaling (NMDS) plots were obtained at the genus and species level using the Bray–Curtis index, and correlation networks using Pearson’s correlation coefficient, considering a P–value threshold of 0.05. After obtaining relative abundances, an endosymbiont search was performed to correlate these data with endosymbiont–specific PCR results. The results were visualized in a Sankey plot of interactions, designed using the free online tool SankeyMATIC (www.sankeymatic.com). The dataset of raw reads obtained from paired ends of bacterial sequences obtained by amplification of the V3–V4 regions of the 16S rRNA gene are available in the NCBI Bioproject repository PRJNA1216679, BioSamples SAMN46440723–SAMN46440740.

Sequence Analysis of Endosymbiont Sequences Obtained by Specific PCR

The obtained endosymbiont sequences were edited and aligned using Finch TV software V 4.0 (Geospiza, Inc.) and MEGA (Molecular Evolutionary Genetics Analysis) X v11, respectively [98]. Sequence similarity analysis was performed by comparing sequences registered in the GenBank database with the basic local alignment tool (BlastN). The filtered reference sequences with quality parameters were downloaded, and multiple alignments were constructed with the sequences of this study using ClustalW [99] with predetermined parameters. The diversity and phylogenetic clustering of the detected haplotypes were analyzed using maximum likelihood analysis in IQ-TREE2 [100]. The nucleotide substitution model was identified using the ModelFinder module [101], according to Bayesian information criterion (BIC), and branch robustness was determined using the ultrafast bootstrap method [102] with 1000 replicates. And the number of haplotypes of the obtained sequences were determined using DnaSP 6.12.03 software [103]. The sequence of Candidatus amoebophilus asiaticus (AF366581.1) was used as the outgroup for the Cardinium dendrogram, while the Wolbachia sequence identified in Dirofilaria immitis (AJ252062.1) was the outgroup selected for the Wolbachia dendrogram. Representative samples of the sequences obtained were deposited in GenBank with the following accession codes: Cardinium: PV031533–PV031552, Wolbachia: PV974757–PV974771.

Results

Bacterial Composition of Psychodopygus Spp., Trichophoromyia Spp., Nyssomyia antunesi, Trichopygomyiawitoto and Brumptomyiamesai Sand Flies Collected in the Colombian Amazon Biome

Microbiome sequencing analysis generated a total of 2,307,796 quality–filtered reads, with an average of 128,210 reads per sample, ranging from 10,931 to 176,053 (Fig. S1a). From these reads, 1,728,340 clustered into seven ASVs at the phylum level, 1,452,965 clustered into 73 ASVs assigned at the genus level, and 1,452,421 clustered into 126 ASVs at the species level. Rarefaction analysis revealed that most samples reached a plateau (Fig. S1b) and that Good’s coverage for each sample ranged from 99.99% to 100%, indicating that the sequencing effort was sufficient to capture most of the diversity present in the samples.

It was observed that the microbiome associated with sand flies is dominated by three main bacterial groups representing 99.5% of the entire population: Proteobacteria (79.6%), Firmicutes (10.4%) and Actinobacteriota (9.4%) (Fig. 2). Specifically, it was found that in sand flies of the genus Psychodopygus (Fig. 2), the phylum Proteobacteria largely dominates in all species, with relative abundance values between 69.4 and 97.1% in species such as Ps. ayrozai and Ps. davisi, respectively (Fig. 2). Although the phyla Actinobacteriota and Firmicutes have a lower abundance, they have a high frequency among all species, ranging from 3.5 to 30.5%. These abundances are particularly notable in Ps. ayrozai (28.5% of Actinobacteriota) and Ps. davisi (10.2% of Firmicutes), while Fusobacteriota, Bacteroidota, and Deinococcota appear in very low proportions (< 1%) (Fig. 2). In the genus Nyssomyia (Fig. 2), the abundance of the phylum Proteobacteria ranged from 74.3 to 90.2%. Specifically, Ny. antunesi, the only species representative of this phlebotomine genus, shows differences in the abundance of the phyla Actinobacteriota and Firmicutes between individuals collected in two different localities of the department of Caquetá, i.e., phlebotomine specimens collected in Santo Domingo (P_22) have a higher representation of microbial communities of Firmicutes (22.2%), while specimens collected in Macagual (P_6) show a higher proportion of Actinobacteria (28.2%). Particularly, individuals from Macagual exhibited a frequency of Verrucomicrobiota (3.6%) higher than that detected in the rest of the individuals collected in Caquetá (< 1%), while in Amazonas, the highest representation of this phylum was detected in Ps. ayrozai (P_24).

Fig. 2.

Stacked bar chart of phylum abundance in total DNA samples of sandfly species collected in two departments of the Colombian Amazon biome. Localities in the department of Amazonas: SPL: San Pedro de los Lagos. Localities in the department of Caquetá: JR: Jericó, STD: Santo Domingo, MGC: Macagual. ♀: Female, ♂: Male

In Br. mesai, the microbial composition shows a clear dominance of Proteobacteria (65.7%), followed by Actinobacteria (26.9%) and Firmicutes (7.4%). Similarly, in the genera Trichophoromyia and Trichopygomyia, whose species were collected in extradomiciliary areas of San Pedro de los Lagos, Amazonas, a similar relative abundance of bacterial phyla is observed. Both sand fly genera are dominated by Proteobacteria with relative abundance values ranging between 78.2 and 78.4%, followed by Firmicutes (10.1 and 11.7%) and Actinobacteria (10.3 and 11.0%). Finally, in these insects, the phyla Verrucomicrobiota, Bacteroidota, and Deinococcota are present with a relative abundance of less than 1% (Fig. 2).

The genus (supraspecific) and between species (interspecific) taxonomic level of the sand flies of the Amazon biome reflect significant differences in the profile and relative abundance of the bacterial genera (Fig. 3). Although Psychodopygus shows a similar profile among the species Ps. ayrozai, Ps. carrerai thula, Ps. chagasi and Ps. davisi, they present interspecific differences mainly associated with the relative abundance of Novosphingobium (11.2–66.5%), Cutibacterium (1–29.4%) and Methylobacterium (3.2–20.4%), while the other bacterial genera have an equal distribution (Fig. 3). These bacterial genera differences within the Psychodopygus genus increase when comparing to the profile of Ps. carrerai thula and Ps. paraensis, which have a high number of Wolbachia readings, with 43.1 and 84.7%, respectively. Additionally, Ps. paraensis has a high relative abundance of Staphylococcus (31.5%).

Fig. 3.

Stacked bar chart of the relative abundance of common bacterial genus present in sandfly species collected in two departments of the Colombian Amazon biome. Localities in the department of Amazonas: SPL: San Pedro de los Lagos. Localities in the department of Caquetá: JR: Jericó, STD: Santo Domingo, MGC: Macagual. ♀: Female, ♂: Male

At the intraspecific level, there are also differences in the composition of the microbiome in sand flies. A clear example is evidenced in Ps. davisi (samples P_2 – P_3), collected in San Pedro de los Lagos, which presents a different profile, influenced by the presence of Pantoea (76.6%) and Escherichia–Shigella (13.7%). This behavior of bacteriome variability at the intraspecific level is also evident in Ny. antunesi, specifically in the group of specimens collected in Macagual (P_6), whose profile is associated with Pseudomonas (39.6%) and Delftia (12.8%), differs significantly from the abundances of the specimens collected in Santo Domingo (P_22), Pseudomonas (2.0%) and Delftia (0.02%).

A similar behavior is also evident in Ps. paraensis, indicating that sand flies from Caquetá have a high prevalence of Wolbachia (84.7%), while in specimens from Amazonas, 47.5% of the bacterial composition is dominated by other genera like Staphylococcus (31.5%) and Cutibacterium (16.1%). Particularly, it was observed that Ps. paraensis individuals from Caquetá show a lower diversity compared to those from Amazonas, where a relative abundance of 42.8% is associated with other bacterial genera. In Ps. ayrozai, sample P_23 from Santo Domingo, Caquetá is dominated by Novosphingobium (17.1%), Cutibacterium (15.8%), and Pantoea (15.7%), while specimens collected in San Pedro de los Lagos, Amazonas (P_24), show a predominance of Cutibacterium (24.4%) and Novosphingobium (16.0%), excluding Pantoea.

The species within the genus Trichophoromyia, collected in Amazonas, present similarity in both abundance and diversity, of Delftia being the most abundant bacterial genus in Th. howardi (57.0%) and Th. velezbernali (59.1%), while the genus Trichopygomyia, also collected in Amazonas, presents a more equal distribution of bacterial genera where these do not exceed 30% (Fig. 3).

The core microbial community in sand flies from the Amazon biome is represented by 18 ASVs (Fig. 4) and shows a high prevalence (50 to 90%) and detection (1 to 19.3%) of Novosphingobium, Cutibacterium, Methylobacterium, and Staphylococcus. Other ASVs, such as Acinetobacter and Pseudomonas, had a prevalence of 50% and detection between 1 and 1.6%. The remaining ASVs presented prevalences below 40%, including Wolbachia and Pantoea determinants in a few species, such as Ps. davisi and Ps. paraensis.

Fig. 4.

Core community of all total DNA samples of sandfly species collected in two departments of the Colombian Amazon biome, where the highest prevalence at the genus level is shown in red and the lowest in blue

A total of 73 ASVs were identified at the genus level of sand flies, indicating that Ny. antunesi presented the highest bacterial richness (n = 68), followed by Ps. ayrozai (n = 63) and Ps. davisi (n = 56). The rest of the sand flies presented a bacterial richness between 39 and 43 ASVs at the genera level. Focusing exclusively on phlebotomine species recognized as either confirmed or potential vectors of importance, unique and shared ASVs that are part of the core community were detected. Bacillus, Cutibacterium, Corynebacterium, Enterobacter, Methylorubrum, Pseudomonas and Staphylococcus were found to be common in the five phlebotomine species selected: Ny. antunesi, Ps. ayrozai, Ps. chagasi, Ps. davisi, and Ps. paraensis (Fig. 5). Additionally, it was possible to detect an ASV associated with Bartonella sp. in females of Ps. paraensis collected in San Pedro de los Lagos, Amazonas, showing a prevalence of 1.5%, where at least one of the five specimens contained in the pool was positive for the bacterium, highlighting the importance of this species as a vector and its potential role in the transmission of different diseases in the Amazon region.

Fig. 5.

Network analysis showing unique (white circles) and shared (gray circles) bacterial genera of sand flies with epidemiological relevance in the Colombian Amazon biome: Ny. antunesi (green line), Ps. ayrozai (red line), Ps. chagasi (black line), Ps. davisi (yellow line), and Ps. paraensis (blue line)

Alpha and Beta Diversity Analyses

The results of the microbial alpha diversity analysis at the genus level indicate that the differences in richness in terms of relative abundance and dominant taxa were different among the phlebotomine species analyzed and significantly more evident at the sex level, being higher in males than in females in all the estimated indices (Fig. 6).

Fig. 6.

Alpha and beta diversity of bacterial communities in sandfly species from peridomestic environments of Amazonas and Caquetá. (a) Observed richness (p–value: 0.041333, [T–test] statistic: –2.392). (b) Chao1 index (p–value: 0.041333, [T–test] statistic: –2.392). (c) Shannon index (p–value: 0.0071423, [T–test] statistic: 3.08411). (d) Simpson index (p–value: 0.014396, [T–test] statistic: –2.8364). (e) The Non–metric Multidimensional Scaling Analysis (NMDS) of the differences in the microbial composition according to sex, NMDS (PERMANOVA F–value: 1.5897; R–R–squared: 0.090377; p–value: 0.094, Stress = 0.15061). (f) Principal Coordinates Analysis (PCoA) plot generated using Bray–Curtis Index (PERMANOVA F–value: 1.5897; R–R–squared: 0.090377; p–value: 0.095)

In this study, the observed richness presented statistically significant differences between females and males, as well as by phlebotomine species (Chao1: p–value: 0.041333, [T–test] statistic: − 2.392), (Fig. 6a). The same behavior was presented with the Chao 1 index (Fig. 6b), (p–value: 0.041333, [T–test] statistic: − 2.392), females presented values lower than 35 while males present higher richness, except in the species Br. mesai where the richness is very low (lower than 10), suggesting that males present higher microbiome richness, which varies according to the species. The Shannon index (Fig. 6c) (p–value: 0.0071423, [T–test] statistic: 3.08411) corroborates that, in addition to presenting a greater bacterial richness, males have a more evenly distributed microbial community, with values greater than 1.7. In females, Shannon values are more variable, with some less than 1 and others greater than 2.0; this variability is observed in both males and females, suggesting that the structure of the microbiome is influenced by the sex and species of the sand flies. Simpson’s index (Fig. 6d) also showed significant differences between sexes (p–value: 0.014396, [T–test] statistic: − 2.8364). For most species, the values were greater than 0.6, indicating high dominance within the bacterial community.

However, beta diversity analyses, represented by nonmetric multidimensional scaling (NMDS), showed no statistically significant differences in the overall composition of the microbiome between sexes (F value = 1.5897; R² = 0.0904; p-value = 0.094; stress = 0.1506) (Fig. 6e) or according to sample origin (F-value = 1.6077; R² = 0.0913; p-value = 0.087; NMDS stress = 0.1413). This pattern was confirmed by principal coordinate analysis (PCoA), in which both sex (weighted UniFrac distance, PERMANOVA: F value = 1.5897; R² = 0.0904; p-value = 0.095) and sample origin (F-value = 1.3926; R² = 0.0801; p-value = 0.177) (Fig. 6f) showed trends but did not reach statistical significance.

Bacterial Community co-occurrence Network Analysis

Pattern analysis using Pearson’s correlation coefficient (r) allowed the identification of relationships between bacterial genera according to the sex of sand flies. We specifically analyzed the most abundant genera, including those of biotechnological relevance that could influence the regulation of the relative abundance of bacterial communities. In particular, the genus Elizabethkingia (Fig. S3) showed a high number of significant positive interactions, especially with Halomonas (r = 0.70, FDR = 0.041), Bradyrhizobium (r = 0.62, FDR = 0. 087), and Nesterenkonia (r = 0.52, FDR = 0.164), with p–values between 0.0011 and 0.028 (Fig. S4a, Table S3), where in turn, Methylobacterium_Methylorubrum interacts positively with these genera (Fig. S4b). In addition, Elizabethkingia presented a significant positive correlation with Brevibacterium (r = 0.63, p = 0.0049, FDR = 0.087), which positively correlates with Pantoea. Likewise, Elizabethkingia showed statistically significant positive associations with Thermus (r = 0.61, p = 0.007, FDR = 0.087), Brevibacillus (r = 0.55, p = 0.018, FDR = 0. 164), Massilia (r = 0.52, p = 0.026, FDR = 0.164) and Tepidimonas (r = 0.52, p = 0.027, FDR = 0.164), which are also positively related to Serratia (Fig. S3). It is worth noting the strong correlation as well of Serratia with Paenibacillus (r = 0.98, p = 2.16 × 10–08, FDR = 8.12 × 10–07) and Massilia (r = 0.77, p = 0.0002, FDR = 0.004) (Fig. S4c, Table S4).

On the other hand, Delftia (Fig. S3, Fig. S4d) showed positive correlations with Pelomonas, Saccharopolyspora, Ralstonia, and Faecalibacterium. However, it showed a non-significant negative correlation with Methylobacterium_Methylorubrum (r = 0.041, p = 0.87, FDR = 0.92). The latter genus presented a strong positive correlation with Novosphingobium (r = 0.88, p = 1.21 × 10–02, FDR = 4.4 × 10–01) (Fig. S3). Finally, the female–predominant genus Arsenophonus (Fig. S3) presented positive correlations mainly with Pantoea, Pelomonas, Saccharopolyspora, Ralstonia and Faecalibacterium (r = 0.54–0.67; p = 0.002–0.02; FDR = 0.08–0.26) (Fig. S4e), which were also predominantly associated with females. Although Arsenophonus presented negative associations mainly with Tepimonas (r = − 0.29) and Elizabethkingia (r = − 0.29), these were not statistically significant (p–value = 0.25, FDR = 0.86).

Detection of Endosymbionts Present in Sand Flies

From the sequencing of the V3–V4 regions of the 16S rRNA gene, reads associated with the endosymbionts Arsenophonus, Cardinium, Rickettsia, Spiroplasma, and Wolbachia were detected in nine of the 10 phlebotomine species available for this analysis (Table S5). The high number of Cardinium–associated reads (120,776; 99.9%) in Th. velezbernali females is highlighted, suggesting a high prevalence of this endosymbiont in the species. While in Ps. paraensis, Wolbachia predominated (138,197 reads; 63.1%), followed by Rickettsia (80,403 reads; 36.7%), Arsenophonus and Cardinium were detected in smaller proportions. On the other hand, in Ps. ayrozai, a high number of Wolbachia readings were found in both females (156,108, 53.0%) and males (125,918; 69.5%). In Ps. carrerai thula and Ty. witoto, Wolbachia was detected only in males, while in females of Ny. antunesi, Rickettsia was the only endosymbiont identified (58,587 reads; 42.1%).

Endosymbiont analysis by specific PCR allowed the detection of Cardinium and Wolbachia in 10 of the 30 species tested. Of the 355 samples evaluated, 21 (5.91%) amplified bands of the expected size for Cardinium were found (Fig. S2a), whereas 23 (6.4%) samples showed amplification for Wolbachia (Fig. S2b). The species with the highest detection of endosymbionts by this specific PCR approach was Th. cellulana (Table S6), with a minimum Wolbachia infection rate of 2.2% and with a higher prevalence in females. In addition, this same species presented the highest percentage of Cardinium infection in females (MIR: 3.66%), in addition to being detected in males (MIR: 0.28%). Ps. paraensis presented the highest percentage of Wolbachia detection (MIR: 2.3%), which was detected in 100% of the sand fly specimens analyzed, corresponding only to males, while in other species like Ps. ayrozai, Wolbachia was detected in both females (MIR: 0.6%) and males (0.6%). Although Br. mesai, Th. velezbernali, and Evandromyia sp., as well as Ny. antunesi, Ps. carrerai thula, Ps. panamensis, Sc. sordellii, Ty. witoto and Lutzomyia sp. presented minimum infection rates lower than 0.3%; the detection of the Cardinium and Wolbachia endosymbionts was evidenced in 100% of the specimens analyzed.

Additionally, molecular analysis integrating NGS and specific PCR strategies revealed a high prevalence of secondary endosymbionts in sand flies from the Amazon biome (Fig. 7). Specifically, the presence of Wolbachia in Ty. witoto, Ps. ayrozai, Ps. carrerai thula, Ps. chagasi, and Ps. paraensis, was confirmed by both techniques (Blue lines detected by specific PCR and brown lines by NGS), as was Cardinium in the Ps. chagasi and Th. velezbernali (Purple lines detected by specific PCR and yellow lines by NGS. However, endosymbiont readings associated with Arsenophonus and Spiroplasma, not detected by specific PCR using the molecular markers employed in this study, were found. Finally, it was evidenced that Ps. paraensis, Ps. ayrozai, and Ps. davisi showed greater diversity of endosymbionts.

Fig. 7.

Sankey plot showing interactions between endosymbionts detected by next–generation sequencing (brown, green, pink, and yellow lines) and specific PCR (blue and purple lines), and phlebotomine hosts collected in the Colombian Amazon biome. The thickness of the lines is proportional to the number of endosymbionts–positive specimens, while light colors represent females and dark colors represent males

Phylogenetic Analysis of Wolbachia and Cardinium Sequences Detected in Sand Flies

In total, 21 Cardinium sequences were obtained from PCR analysis of sand flies from both departments. After manual editing and end trimming, the edited consensus sequences were between 371 and 435 bp in length. When comparing the sequences obtained with those available in GenBank, a percentage of similarity was found that ranged from 99.6% to 100%. In the 21 sequences analyzed, seven haplotypes and a haplotypic diversity of 0.5842 were identified. The highest genetic diversity of Cardinium was observed within haplotype 2, which was represented by 13 sequences, 12 detected in Th. cellulana and one detected in Th. velezbernali. The evolutionary tree was traced on 86 sequences, including those available in GenBank. In total, 68 sites (4.7%) were constant, while 421 sites were informative for parsimony. A total of 479 distinct site patterns were identified in the alignment, which were used in the phylogenetic reconstruction, including the highest log likelihood (− 5,107.8381), the ModelFinder module selected the K2P + R2 model as the best nucleotide substitution model, according to the Bayesian information criterion (BIC = 11257.0207).

The Cardinium sequences detected in the sand flies of the Colombian Amazon are distributed in groups A, and C (Fig. 8). The Cardinium haplotypes of group A, mostly have proximity and are associated with endosymbionts previously reported in hosts such as Oppiella nova, Mi. (Mic.) cayennensis and Microzetorchestes emery (Fig. 8a); however, the formation of a new subgroup was evidenced, close to those previously reported in Mi. (Mic.) cayennensis but very well differentiated, which we named Cardinium Endosymbiont of Psychodopygus (shaded blue), This subgroup includes the haplotypes detected in Ps. paraensis and Ps. chagasi. In group C, groups a Cardinium haplotype associated with Ps. chagasi and mainly related to Candidatus Cardinium hertigii and haplotypes reported for Mi. (Mic.) cayennensis and Pa. (Psa.) shannoni. (Fig. 8b).

Fig. 8.

Phylogenetic position based on the 16S rRNA gene of the endosymbiont Cardinium detected in sand flies from the Amazon biome (bold letters), using the maximum likelihood method: (a) Sequences associated with group A. (b) Sequences associated with group C

For the Wolbachia endosymbiont, 22 sequences associated with seven phlebotomine species were obtained. After manual editing and end trimming, the sequences were between 454 and 495 bp in length. When compared with those available in GenBank, a similarity percentage between 90.4 and 100% was found. In the 22 sequences obtained, 13 haplotypes were identified, with a haplotypic diversity of 0.9429. The highest genetic diversity of Wolbachia occurred within Th. cellulana and Ps. paraensis, where the 11 sequences detected in the two species represented four haplotypes. The evolutionary tree was traced on 110 sequences, including those available in GenBank. In total, 68 sites (13.74%) were constant, while 275 sites were informative for parsimony. A total of 442 distinct site patterns were identified in the alignment, which were used in the phylogenetic reconstruction, including the highest log likelihood (− 6,412.8829), the ModelFinder module selected the TPM3 + F + R2 model as the best nucleotide substitution model, according to the Bayesian information criterion (BIC = 14017.0410).

The haplotypes detected in sand flies from the Colombian Amazon region were classified within supergroups A and B (Fig. 9), which include strains and haplotypes previously associated with arthropods. In supergroup A, haplotypes from Br. mesai and Th. cellulana were grouped, which showed a relationship with the wMel strain, previously identified in D. melanogaster (Bootstrap: 99%) (Fig. 9a). Within this same supergroup, we evidenced the formation of a new Wolbachia group and potentially a new strain, composed of sequences detected in species of the genus Psychodopygus, which we refer to as the wPsy group. Within this group there are associated sequences from the hosts Ps. ayrozai, Ps. paraensis and Ps. carrerai thula (blue shade). However, a clear divergence was identified with the Wolbachia sequence of wWito (pink shade) suggesting that they correspond to new and different haplotypes. Additionally, in this supergroup A, sequences from Lu. (Tri.) sherlocki were located close to sequences from the wWhi strain previously reported in Pa. (Psa.) shannoni. Supergroup B, haplotypes identified in Ps. ayrozai (n = 2) clustered with the wTac1 strain previously reported in Tachinid sp. (Fig. 9b). Haplotypes detected in Th. cellulana and Ny. antunesi solidly formed a new group (gray shade), which we refer to as wNys. Wolbachia sequences detected in Th. cellulana (n = 2) and Ps. panamensis (n = 1) showed similarity and closeness to the wAlbB and wPseu haplotypes previously detected in Ae. albopictus. Finally, other Wolbachia haplotypes were detected in Th. cellulana, and another specimen categorized as Lutzomyia sp. showed associations with Brumptomyia strains WbBrmei117 and Belonocema wTre4, respectively.

Fig. 9.

Phylogenetic position based on the wsp gene of the endosymbiont Wolbachia detected in sand flies from the Amazon biome (bold letters), using the maximum likelihood method. (a) Sequences associated with supergroup A. (b) Sequences associated with supergroup B

Discussion

Although the exploration of the microbiome components in sand flies vectors of leishmaniasis has provided insights about their possible implications in the transmission of pathogens, to date, only a fraction less than 20% of the estimated diversity of sandfly species at different locations in the Americas has been studied. The study of the microbiota communities in sand flies from the Amazon biome revealed that Proteobacteria, Actinobacteria, and Firmicutes are the dominant phyla, while at the genus level Novosphingobium, Cutibacterium, Methylorubrum, and Staphylococcus, are the most frequent. A remarkable finding was the high prevalence and diversity of secondary endosymbionts of interest for vector control in wild populations, which allowed us to identify the presence of a new subgroup of Cardinium and a new group of Wolbachia associated with supergroup A in sand flies of the genus Psychodopygus. Additionally, a new haplotype was identified in Ty. witoto and Th. cellulana, and a group associated with supergroup B in the genera Nyssomyia and Trichophoromyia. Other endosymbionts found were Arsenophonus, Rickettsia, and Spiroplasma. With the above description, this study provides significant information on the phylogeny and infection levels of these endosymbionts in wild sand flies.

This study demonstrated that host species harbor a distinctive microbiota, with greater similarity observed in phylogenetically close hosts, i.e., at inter- and intraspecific taxonomic levels, compared to hosts at supraspecific or less related levels. Also, geographical origin tends to influence the relative abundance of bacterial communities. This is consistent with previous findings that host-associated microbiota are closely related to host phylosymbiosis, and host ecology [33, 104, 105]. In addition, genera with antimicrobial potential were identified, such as Serratia, Bacillus, Delftia, Elizabethkingia, Haemophilus, Enterobacter, Pseudomonas, and Ralstonia, some of them involved in the inhibition of parasite development [63, 64, 106, 107], with activity as entomopathogens [108–112] and associated with the degradation of insecticides [112].

Our results show that Proteobacteria, Firmicutes, and Actinobacteria were the predominant phyla in the samples studied. These findings are consistent with the data from a meta-analysis of sand flies associated bacteria, from the Andean and Caribbean regions [48, 56]. However, a notable difference in the relative abundance of these phyla was observed. In the Colombian Amazon, Actinobacteria are more dominant than Firmicutes, while in the Andean and Caribbean region, Firmicutes is the second most abundant phylum after Proteobacteria [48, 56]. In the case of the Brazilian Amazon, there are also distinct compositions of the bacterial communities of specific host species, for example, Lu. (Lut.) longipalpis is dominated by Proteobacteria and Firmicutes, while in Lu. (Lut.) cruzi, Firmicutes, Proteobacteria, and Bacteroidetes predominate [113] and similarities with the study conducted in the host, Lu. (Hel.) ayacuchensis collected in Ecuador and Peru, where the microbiome was also dominated by Actinobacteria, Proteobacteria, and Firmicutes [114]. Both Ecuador and Peru, as well as the Colombian Amazon, present a biome characterized by a warm and humid climate, high biodiversity, and extensive tropical forests. This coincidence suggests a possible pattern associated with similar environmental conditions in these regions. In addition to the predominant phyla, the microbial composition of Amazonian sand flies includes, in lower abundance, the phyla Verrucomicrobiota, Bacteroidota, and Deinococcota; however, the percentage abundance of these is higher than previously recorded for New World [35, 48, 56] and Old World [65, 115] sandfly species.

At the genus level, a statistically significant difference was also observed in the composition and relative abundance compared to that described in sandfly species from the Andean and Caribbean regions (Fig. 10), in the Amazon biome the bacterial genera found in sand flies are mainly represented by Novosphingobium, Cutibacterium, and Methylorubrum, contrasting with the Andean and Caribbean regions, were Aeromonas, Ralstonia and Cutibacterium prevail [48, 56]. Likewise, sand flies from Ecuador and Peru exhibit a prevalence of Ochrobactrum, Corynebacterium, and Cutibacterium [114]. According to available information, Novosphingobium has not been previously documented as the genus with the highest abundance in sand flies. However, it was found to predominate in all analyzed host species, with less representation in the genus Trichophoromyia. Its ability to degrade organic pollutants, including PAHs like phenanthrene [116], 2,4-dichlorophenoxyacetic acid [117] or glyphosate [118], suggesting a particular importance of insecticides, herbicides and petroleum derivatives metabolism in our study site and there are feasible selective exposures in the region with domestic, agricultural and oil industry activities releasing/applying such compounds.

Fig. 10.

Principal coordinates analysis (PCoA) based on Bray-Curtis distances comparing the microbiome detected in sand flies from the Amazon region by sequencing the V3-V4 region of the 16S rRNA gene with previously reported data on sand flies from the Andean [48, 56] and Caribbean [48] regions of Colombia (PERMANOVA: pseudo-F = 3.45, p = 0.0002)

Cutibacterium was the second most abundant genus, detected in Andean region species like Br. hamata and Pi. (Pif.) pia, with a low infection rate of less than 0.01% [48]. This genus has also been associated with laboratory-reared Lu. (Lut.) longipalpis, infected with Leishmania [33]. In contrast, the results obtained in this study resemble those reported by Tabbabi et al., who reported Cutibacterium as the third most abundant genus in Lu. (Hel.) ayacuchensis collected in Ecuador and Perú [114]. Cutibacterium is part of the cutaneous microbiota and shows specific tropism for the moist and sebaceous regions of human skin [119]. Although there is limited knowledge about the biotechnological potential of this genus in insects, its presence may be related to specific feeding and digestion mechanisms [51]. Methylobacterium was the third most abundant genus in the bacteriome of sand flies in the Colombian Amazon biome, although it was absent in Trichophoromyia. Recently, its prevalence was documented in sand flies of Lu. (Lut.) longipalpis, Pi. (Pif.) evansi and Ev. (Ald.) dubitans from Andean and Caribbean region of Colombia region respectively [45, 48, 56]. Likewise, Methylobacterium was the most abundant genus in Ph. kandelakii–M and Ph. perfiliewi–M, exceeding 70% of the relative abundance at the genus level. In contrast, in Ph. kandelakii–B and Ph. major–B the abundance was less than 5%, while in Ph. alexandri–B and Ph. alexandri–M it did not exceed 0.5% [115]. At the family level, Methylobacteriaceae has been reported in the midguts of fed and gravid females of Ny. intermedia with low representation, in a leishmaniasis endemic area in Brazil [120]. It was also detected in the midgut microbiota of laboratory–reared Lu. (Lut.) longipalpis on the third- and sixth-day post-infection with L. (L.) infantum [121].

The present study also reveals significant differences in the composition of bacterial genera at the supraspecific level in the taxonomy of sand flies. A clear example includes the sand flies of Brumptomyia, represented mainly by bacterial genera such as Novosphingobium and Cutibacterium, with a relative abundance not exceeding 40%. In contrast, in the host Trichophoromyia, the genus Delftia shows a relative abundance exceeding 55%. In addition, in Trichophoromyia, a high presence of Pseudomonas was evident, while in Brumptomyia, bacterial abundances of no more than 1% were observed. Exploration of the microbiome in the genus Trichophoromyia using next-generation sequencing has not been previously addressed. Vivero–Gomez et al. reported that Br. hamata from the Andean region is mainly represented by Aeromonas (26.4%) and Wolbachia, with Ralstonia and Acinetobacter. They also noted differences in microbial composition based on the species’ geographical origins [48]. In particular, the genera Nyssomyia and Psychodopygus present variable relative abundances and differ from those of Brumptomyia, Trichophoromyia, and Trichopygomyia, affirming the variability at the genus taxonomic level in sand flies. The diversity of the microbiota in Psychodopygus from the Caribbean region contrasts with the findings of the present study, where the predominant genera include Corynebacterium, Staphylococcus, Stenotrophomonas, and Acinetobacter, representing 46.1% of the bacterial community [48]. This variability in bacterial composition could also be influenced by environmental, geographical, and ecological factors. These findings support the need for further studies to better understand the dynamics of the microbiome across phlebotomine species and populations, and the possible influence on competition to transmit different Leishmania species on a regional scale.

Additionally, it was possible to detect bacterial genera mainly related to the inhibition of pathogen development and the reduction of vectorial capacity, in other insects, such as Delftia detected in specimens of Ps. ayrozai in this study, specifically, D. tsuruhatensis has been associated with the inhibition in the formation of P. falciparum and P. berghei oocysts in the intestine of An. gambiae and An. stephensi [106]. Another relevant genus is Serratia, common in Ps. ayrozai and Ps. davisi. It is relevant because species such as S. marcescens favors the dissemination of DENV viral particles in Ae. aegypti [122], exhibits larvicidal activity against An. dirus [108] and affects the survival, growth, and fecundity of Mythimna separata [109]. Other genera detected in insects of epidemiological relevance were Bacillus and Haemophilus, relevant due to the lethal effect recorded in some species such as H. parainfluenzae and B. thuringiensis against L. (L.) major in Ph. papatasi [64].

The present study contributes to the diversity and phylogenetics of secondary endosymbionts associated with the Amazon region of Colombia, generating new knowledge and perspectives for the potential monitoring and design of biological vector control strategies. The endosymbionts were detected by NGS and specific PCR with different molecular markers previously validated in the literature. Among them, Arsenophonus, Cardinium, Spiroplasma, Rickettsia, and Wolbachia are biotechnologically relevant for their ability to affect vector competence, reproduction, and fitness of some insects [53–55, 57, 61, 62, 123]. Wolbachia infection in specimens of Br. mesai, Lu. (Trl.) sherlocki, Ny. antunesi, Ps. ayrozai, Ps. carrerai thula, Ps. paraensis, Th. cellulana, and Ty. witoto had not been reported to our knowledge. At the phylogenetic level, the Wolbachia sequences obtained in Ps. paraensis and Br. mesai showed high similarity to wMel sequences of the previously described supergroup A D. melanogaster [89]. Although Wolbachia infection had not been recorded in Br. mesai, its presence had been evidenced in species of the same genus, such as Br. hamata [48]. Wolbachia has been documented to influence fecundity, fertility, and longevity [123–125], and in addition to affecting vector competence in Aedes [41], it plays a significant role in inhibiting fungal pathogenesis in D. melanogaster [126].

Additionally, in Lu. (Trl.) sherlocki, Wolbachia sequences were mainly associated with the wWhi strain previously detected in Pa. shannoni in Colombia and Brazil [46]. Wolbachia haplotypes derived from Ps. paraensis, Ps. carrerai thula, and Ps. ayrozai and forming the wPsy group located in supergroup A (new in Wolbachia phylogeny) represent the most diverse and abundant in the sand flies of the Amazonian biome. The wPsy group is close to wStv MI strains, whose hosts are related to D. sturtevanti where it has demonstrated its impact on reproduction by inducing CI [127, 128], recently, wStv was detected in Ny. whitmani [129]. Although the effect on phlebotomine fitness is not yet elucidated, its increasing detection in sand flies highlights the need for further studies to identify the phenotypes it induces in these host species. The supergroup B strains detected in Ps. panamensis and Th. cellulana is related to the wPseu and wAlbB strains previously detected in Ae. pseudoalpoictus [130] and Ae. albopictus [89]. In this supergroup, the formation of the new wNys group was detected in Ny. antunesi and Th. cellulana is highlighted. The wAlbB strain, a natural endosymbiont of Ae. albopictus, has been associated with the reduction of DENV–2 viral individuals in Ae. albopictus populations in China [131]. In vitro, studies have shown decreased replication of bluetongue virus and epizootic hemorrhagic fever virus in cell lines derived from Culicoides sonorensis and C. jones [132]; likewise, wAlbB has been shown to modulate the vectorial capacity and reduce the population density of Ae. aegypti in Australia through the expression of the CI phenotype in mosquitoes [133, 134].

By specific PCR using the WSP marker, Wolbachia was detected in 33% of the sandfly species tested. This result is slightly higher than the percentage previously documented by Ono et al., who stated that 27% of sandfly species from Colombia, Brazil, India, Israel, and Italy were found to be infected with Wolbachia [46]. It is also higher than previously reported in Panamá, where it was found that 15% of the species analyzed in Isla Barro were infected with Wolbachia [135]. However, it is lower than those recorded in sand flies from Colombia (28 and 42.8%) [44, 48] and Mexico (50%) [50]. On the other hand, the MIR of Wolbachia in sand flies from the Amazon region ranged from 0.28 (Br. mesai, Ny. antunesi, Ps. carrerai thula, Ps. panamensis, Ty. witoto and Lutzomyia sp. to 2.25% (Ps. paraensis and Th. cellulana). The Wolbachia infection observed in Ps. panamensis (0.28%) is higher than that recorded by Vivero–Gomez et al. in specimens collected in the Caribbean region (0.17%) [48], but lower than the detection rate observed in specimens of the same species in Chiapas, Mexico (4.5%) [7]. In general, detection rates in sand flies from the Colombian Amazon region were higher than those reported by Vivero et al. in sand flies from the Andean and Caribbean region, where the calculated MIR showed fluctuations of 0.17 and 1.56% [48]. However, they are lower than the minimum detection rates reported in sand flies from Panama (Lu. vespertilionis = 4.33%, Ny. trapidoi = 6.29%) and Mexico (Pa. (Psa.) shannoni = 17.22, Ps. panamensis = 1.66%, Da. (Cor.) beltrani = 0.55%) [50, 135].

Recent studies have documented that although Wolbachia is abundant in insects, it does not always impact the diversity of bacterial species or vice versa, and may depend more on other intrinsic host factors such as genetics, immune response, internal microbial interactions, and developmental stages, among other factors [112]. However, at abundances higher than 80% Wolbachia may impact the microbiome profile in sand flies and may have repercussions in the development and transmission of Leishmania, although more studies are needed [136]. Additionally, the behavior of the prevalence or infection rates detected in the different phlebotomine species could be influenced by the heat waves that the Colombian Amazon region has suffered since the high temperatures above the average (27 °C) can reduce the fixation, density, and positivity of some Wolbachia strains [137]. Also, the variation in the density of endosymbionts may be associated with the age of the adults or the conditions experienced in the previous generation [138], which is not estimated during the processing of sand flies.

In addition, this research provides new information on the natural infection of Cardinium and the phylogenetic position of different sequences in groups A and C, obtained from the DNA of Pa. (For.) aragaoi, Ps. ayrozai, Ps. chagasi, Ps. paraensis, Th. cellulana, Th. velezbernali and Evandromyia sp. Most of the sequences are phylogenetically located in group A, whose members have been detected in various hosts including insects (Diptera, Hymenoptera, Hemiptera), mites, opiliones, and spiders [139]. Our sequences are mainly associated with haplotypes previously detected in Microzetorchestes emery [140]. However, phylogenetic analysis revealed that some sequences present genetic differences concerning previously reported strains, which allowed us to identify the formation of the new subgroup, Cardinium Endosymbiont of Psychodopygus, present up to the time of this publication, in the species Ps. paraensis and Ps. chagasi from the Colombian Amazon region.

In particular, the only Cardinium sequence detected from Th. velezbernali was associated with the haplotype related to Mi. (Mic.) cayennensis recently found on the Caribbean coast of Colombia. Considering that Th. velezbernali a recently described species [79], elucidating its relationship with Cardinium could provide information to understand the possible host expansion of this endosymbiont. Finally, the phylogenetic proximity of a sequence detected in Ps. chagasi with recently described Mi. (Mic.) cayennensis and Pa. (Psa.) shannoni [48], allowed grouping them within group C, which mainly includes sequences detected in the genus Culicoides [139]. The variability in infection rates detected in this study by specific PCR using the 16S rRNA gene marker is similar to that reported in previous studies. In Pi. (Pif.) evansi, an infection rate of 0.68% was detected [45], whereas, more recently, the presence of Cardinium was identified in Ty. triramula, Pa. (Psa.) shannoni, and Mi. (Mic.) cayennensis with natural infection rates of 0.34, 1.03, and 3.8%, respectively [48]. The effects of natural infection by Cardinium on biology in sand flies have not yet been characterized. However, it has been documented in other insects to induce feminization, parthenogenesis, and IC, in addition to negative effects on host fitness [141–145]. It has also been documented as a precursor in the reduction of microbiome diversity in Nilaparvata lugens [146] and was recently implicated as a key microorganism in thermotolerance in Bemisia tabaci of the Mediterranean [147], a relevant finding, as climate change could play a crucial role in the adaptation and expansion of vector species, such as sand flies.

In this study, the detection of other endosymbionts was accomplished exclusively through next–generation sequencing. Among these, Arsenophonus was identified in Ps. paraensis, Ps. davisi, and Th. howardi. Previous research has associated Arsenophonus with increased susceptibility to insecticides of N. lugens [148]. It has also been identified in Pi. (Pif.) evansi was treated with antibiotics and infected with L. (L.) infantum [35], and was recently detected in Lu. (Lut.) longipalpis, Ps. panamensis, and Pi. (Pif.) evansi collected from the Andean and Caribbean regions of Colombia [52, 56]. Another endosymbiont detected by NGS is Spiroplasma, which was found in Ph. chinensis [59] and has also been recently detected in Lu. (Lut.) longipalpis [56]. Although its role in sand flies is unclear, the detection of Spiroplasma only in females of Phlebotomus sp. suggests it may contribute to an androcide phenotype in these sand flies [60]. The analysis of the microbiome of sand flies from the Amazon biome also allowed us to determine the natural infection of two genera of veterinary relevance. Rickettsia can have negative effects, including male embryonic lethality, increased susceptibility to insecticides, reduced fecundity, and longevity among other phenotypes exerted on some arthropods [149, 150], and even some strains are related to tick–borne spotted fever, tick–borne typhus, and mite–borne rickettsiosis [151]. In our study, this endosymbiont was detected only in females of Ps. paraensis, Ps. davisi and Ny. antunesi and was previously reported in Ph. chinensis [59], Pa. (Xip.) aclydifera [49], Pi. (Pif.) pia [48] and Lu. (Lut.) longipalpis [56].

The detection of 200 reads of the 16S rRNA gene associated with Bartonella sp. in a dataset of Ps. paraensis females stands out (0.11%), considering that this genus includes species such as B. bacilliformis, the etiological agent of Carrion’s disease, endemic to different regions of America, including Peru and Ecuador, where Pi. (Pif.) verrucarum, Lu. (Hel.) noguchii and Lu. (Hel.) peruensis participates in its transmission [152, 153]. Although the presence of this bacterial genus had not been recorded in sand flies from the Colombian territory, recent studies using quantitative real–time PCR (qPCR) and specific PCR, based on the internal transcribed spacer (ITS) of the 16–23 S rRNA gene, have detected its natural infection in sand flies from the Brazilian Amazon, including Ps. paraensis collected in the state of Pará [6], as well as in species from the Peruvian Amazon such as Ny. whitmani and Pi. (Pif.) nevesi [154]. The high occurrence rates of this bacterial genus in sand flies from Brazil and Peru (< 8.6%) and the low number of reads detected in Colombian Amazonian sand flies could indicate that its influence in this region is minimal. However, it is necessary to expand research on this bacterial genus in sand flies from bordering areas in the Amazon to understand its possible role in the transmission of this bacterium.

Interestingly, the detection of Arsenophonus, Bartonella, Spiroplasma, and Rickettsia by NGS could be related to their low density in the sand flies of the Amazon biome, which would explain the negativity obtained with the conventional methods used in this study. However, a factor that cannot be overlooked is the impact of the high temperatures that have affected the Amazon region in recent years [155], as they could negatively influence the dynamics of abundance and prevalence of endosymbionts when hosts have been exposed to thermal stress [156], also altering microbial diversity and functions and affecting host fitness [156, 157].

Conclusion

The search for complementary strategies to mitigate the incidence of leishmaniasis has led to the investigation of the microbiome of sand flies. This research aims to enhance our understanding of microorganism-host interactions and to identify those bacteria that influence the biology of these insects and their capacity to transmit pathogens. This study revealed that, under the evaluated conditions, the bacterial community of sand flies in the Colombian Amazon is represented by the genera Novosphingobium, Cutibacterium, Methylorubrum, and Staphylococcus. The composition, diversity, and relative abundance of the microbiota are influenced mainly by the taxonomic level (supra, inter, and intraspecific). In addition, new records of natural infection by secondary endosymbionts such as Arsenophonus, Cardinium, Rickettsia, Spiroplasma, and Wolbachia are included. The study also identified the presence of bacterial genera such as Serratia, Haemophilus, and Elizabethkingia, recognized for their high biotechnological potential. These bacteria can influence the development of pathogens, including parasites and viruses, as well as affect vector competence in diverse insects, highlighting the importance of studying the microbiome of sand flies in different regions of Colombia, such as the Amazon. This is crucial due to the role these genera may play in pathogen transmission and their impact on human health. Finally, the findings on the high detection rates of Cardinium and Wolbachia by specific PCR methods provide new information on the identity and phylogenetic position of haplotypes present in Amazonian sand flies. This emphasizes the need to further explore their effect on the biology of these insects, including potential influence on disease transmission and the implications for insect fitness and reproductive phenotypes.

Supplementary Information

Below is the link to the electronic supplementary material.

Author Contributions

K.C.T, Conceptualization, Methodology, Specimen Identification, Sample processing, Software, Sequence analysis, Validation, Formal Analysis, Investigation, Data Curation, Writing – Original Draft Preparation, Writing – Review & Editing, Visualization; R.V.G, Conceptualization, Methodology, Specimen Identification, Sample processing, Software, Sequence analysis, Validation, Formal Analysis, Investigation, Resources, Data Curation, Writing – Original Draft Preparation, Writing – Review & Editing, Visualization, Supervision, Project Administration, Funding Acquisition; D.D.G, Methodology, Writing – Review & Editing; H.J. Software, Sequence analysis, Validation, Data Curation, Writing – Review & Editing; G.C.R, Resources, Writing – Review & Editing, Project Administration, Funding Acquisition; C.X.M.H, Conceptualization, Methodology, Validation, Resources, Investigation, Writing – Review & Editing, Supervision, Project Administration, Funding Acquisition. All authors read and approved the final manuscript.K.C.T and R.VG. contributed equally to this work, “co-first authors.”

Funding

This study received financial support from project Hermes 57545 of the Universidad Nacional de Colombia, awarded to C.X.M.H, K.C.T received financial support from the Scholarship Program of Ministerio de Ciencia, Tecnología e Innovación, Call 15, for Human Capital Development in the context of the Bicentennial and the 2021–2022 Biennial Plan. The authors, G.C.R, R.V.G, H.J, and D.D.G, received no specific funding for this work. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Data Availability

Sequence data associated with the 16S rRNA gene amplicon reads are available in the SRA database under the following codes: SAMN46440723, SAMN46440724, SAMN46440725, SAMN46440726, SAMN46440727, SAMN46440728, SAMN46440729, SAMN46440730, SAMN46440731, SAMN46440732, SAMN46440733, SAMN46440732, SAMN46440733, SAMN46440734, SAMN46440735, SAMN46440736, SAMN46440737, SAMN46440738, SAMN46440739, SAMN46440740.Sequence data associated with the Cardinium are available in the GenBank database as follows: PV031533, PV031534, PV031535, PV031536, PV031537, PV031538, PV031539, PV031540, PV031541, PV031542, PV031543, PV031544, PV031545, PV031546, PV031547, PV031548, PV031549, PV031550, PV031551, PV031552.Sequence data associated with the Wolbachia are available in the GenBank database as follows: PV974757, PV974758, PV974759, PV974760, PV974761, PV974762, PV974763, PV974764, PV974765, PV974766, PV974767, PV974768, PV974769, PV974770, PV974771. 

Declarations

Competing interests

The authors declare no competing interests.