Comparison of sand fly trapping approaches for vector surveillance of Leishmania and Bartonella species in ecologically distinct, endemic regions of Peru

Background In Peru, the information regarding sand fly vectors of leishmaniasis and bartonellosis in the Amazon region is limited. In this study, we carried out sand fly collections in Peruvian lowland and highland jungle areas using different trap type configurations and screened them for Leishmania and Bartonella DNA. Methodology/Principal findings Phlebotomine sand flies were collected in Peruvian Amazon jungle and inter Andean regions using CDC light trap, UV and color LED traps, Mosquito Magnet trap, BG Sentinel trap, and a Shannon trap placed outside the houses. Leishmania spp. screening was performed by kDNA PCR and confirmed by a nested cytochrome B gene (cytB) PCR. Bartonella spp. screening was performed by ITS PCR and confirmed by citrate synthase gene (gltA). The PCR amplicons were sequenced to identify Leishmania and Bartonella species. UV and Blue LED traps collected the highest average number of sand flies per hour in low jungle; UV, Mosquito Magnet and Shannon traps in high jungle; and Mosquito Magnet in inter Andean region. Leishmania guyanensis in Lutzomyia carrerai carrerai and L. naiffi in Lu. hirsuta hirsuta were identified based on cytB sequencing. Bartonella spp. related to Bartonella bacilliformis in Lu. whitmani, Lu. nevesi, Lu. hirsuta hirsuta and Lu. sherlocki, and a Bartonella sp. related to Candidatus B. rondoniensis in Lu. nevesi and Lu. maranonensis were identified based on gltA gene sequencing. Conclusions/Significance UV, Blue LED, Mosquito Magnet and Shannon traps were more efficient than the BG-Sentinel, Green, and Red LED traps. This is the first report of L. naiffi and of two genotypes of Bartonella spp. related to B. bacilliformis and Candidatus B. rondoniensis infecting sand fly species from the Amazon region in Peru.


Introduction
Phlebotomine sand flies are small insects of crepuscular and nocturnal activity with weak flight capabilities, distributed in tropical and sub-tropical regions around the world. Female sand flies must feed on blood to complete oogenesis and lay eggs, and obtain blood meals from a large variety of vertebrates, including humans [1]. Due to this broad host range, some phlebotomine species may transmit pathogens to a vertebrate host, such as parasites of the genus Leishmania, bacteria of the genus Bartonella, and some arboviruses [2,3]. Phlebotomine sand flies have an important role in the epidemiology of these diseases; monitoring and control of sand flies are, therefore, of high priority [4].
The leishmaniases are a group of neglected tropical diseases endemic in 98 countries worldwide with 0.9 to 1.6 million new cases per year, between 20,000 and 30,000 deaths, and 350 million people at risk of infection [5]. In Peru, more than 4,000 new cases of cutaneous and mucocutaneous leishmaniasis are reported yearly; no visceral leishmaniasis cases have been reported to date [6]. Human bartonellosis, also called Carrion's disease or Oroya fever, is caused by Bartonella bacilliformis and is a severe, debilitating illness with a mortality rate of over 80% in untreated individuals [7]. The initial infection is characterized by fever and severe anemia; the secondary infections in this phase are associated with transitory immune system depression. After a variable time of acute infection, the presence of vascularized skin warts ("verruga peruana", which is considered the chronic presentation of Carrion's disease) is frequent [8]. This neglected disease is limited to Peru, Ecuador, and Colombia, with higher prevalence in Peru where new endemic areas have been recently reported [7,9]; no cases have been reported during the last two decades in the latter two countries. Other Bartonella species can objective of this study was to evaluate different vector trapping methods for their efficacy in capturing sand flies and to identify potential sand fly vector species of two neglected diseases, leishmaniasis and bartonellosis, in different ecological sites in Peruvian lowland and highland jungle areas and inter-Andean regions. We hypothesized that there would be a significant variation in the response of sand flies to different trap types according to study site and sand fly fauna composition. Moreover, we screened female sand flies captured for both Leishmania and Bartonella DNA detection to identify potential vector species. Given the high sand fly diversity in selected study sites, we assumed that different Leishmania and Bartonella species are circulating in geographical regions where their presence has not been previously reported, especially in the Amazon region.

Study design
In July 2014, and January and February 2015, multiple trap configurations were tested for their efficacy in collecting phlebotomine sand flies in four sites located in different ecological regions where leishmaniasis and/or human bartonellosis are endemic. We morphologically identified female and male sand flies, then tested non-engorged female sand flies (no visible blood meal in the sand fly gut) for molecular detection of Bartonella and Leishmania DNA. Phylogenetic analyses were performed on DNA sequences obtained.

Study sites
We carried out sand fly collections in Peruvian lowland and highland jungle areas, and one inter-Andean Valley (Fig 1). All collections were performed in peri-and extradomiciliary sites. predominant. Caraz, Ancash, is a typical inter-Andean valley with dispersed housing in the rural area surrounded by crops (grasses, tubers, cereals) and livestock (sheep, cows, horses).

Sand fly collection
Trap locations within each village were selected based on the occurrence of human leishmaniasis and bartonellosis cases reported by the local Ministry of Health. A Latin square design was employed with traps being placed in the peridomicile and extradomicile (edge of the forest and inside the forest) environments and at least 50 meters apart to eliminate interference. No intradomiciliary sand fly collections were performed due some traps release CO 2 or emit UV light. This study involved attracting sand flies using luminous traps and odor baited traps that run overnight (1800-0600 hours) (Fig 2)

PLOS NEGLECTED TROPICAL DISEASES
incandescent lamp with blue, green, or red LED color bulbs [31,42]. All luminous traps were placed approximately one meter above ground level and operated overnight from 1800 to 0600 hours. The energy for light traps was provided by rechargeable 6 V gel batteries (Opalux, Lima-Peru).
Odor baited traps with chemical attractants. Two commercial traps were tested, Bio-Gents Sentinel Trap (BG; BioGents, Regensburg, Germany) with BG Lure as the attractant; and Mosquito Magnet (MM; Woodstream Coroporation, Lancaster, PA, USA) Trap Model Independence MM3200B which is battery-operated and runs on R-Octenol and propane gas, which is catalytically converted to produce CO 2 , heat and water vapor, mimicking a vertebrate host and attracting biting insects [27,31,34]. These traps operated overnight from 1800 to 0600 hours.
Shannon trap with protected human bait. This trap was used as a complementary method for sand fly collection in some places and when the weather conditions allowed. Insects were collected with Shannon trap from 1800 to 2100 hours by one operator protected with a mosquito jacket acting as human bait and using a manual aspirator and head lantern to manually catch sand flies [22,24]. The Shannon trap collections did not use volunteers and were instead performed by experienced and highly trained NAMRU-6 entomology researchers. NAMRU-6 Occupational Health & Safety regulations for entomological specimen field collection were followed accordingly.

Sand fly taxonomic identification
Sand fly specimens collected were counted and sorted by location, collection method, sex and collection date, stored in 1.5 mL microcentrifuge tubes with 70% ethanol, and transported to the Entomology Department at NAMRU-6 in Bellavista, Callao region. Specimens were identified morphologically using taxonomic keys and procedures developed by Young and Duncan (1994) [1] and Galati (2003) [43] with small modifications. Female specimens were processed with a modified protocol where the head and last two abdominal segments were separated and clarified with lactophenol (phenol:lactic acid 2:3) for two hours for taxonomic identification, while the rest of the body from each female sand fly was pooled (1-10 specimens per tube) in microcentrifuge tubes with 70% ethanol according to species, date, site and collection method, and stored at -20˚C [37,44]. Male specimens were clarified in 20% KOH for 12-24 hours followed by lactophenol for 2 hours. Voucher specimens were mounted permanently on Euparal.

Sand fly abundance and diversity analyses
Species abundance was calculated in Microsoft Excel 2013 using the Index of Species Abundance (ISA) [45] using the formula ISA = (a + R j )/k, where we established a rank for species abundance per site; "a" is the number of zero observations for each species in all sites multiplied by "c" which is the single largest rank in all the data set plus 1. The R j value corresponds to the sum of ranks for a given species in all the sites whereas "k" corresponds to the number of sites. The resulting ISA values were converted into the Standardized Index of Species Abundance (SISA) using the formula SISA = (c − ISA)/(c − 1). We quantified the number of sand flies captured per hour over each day of trapping and cumulatively over all trapping days. Normality of data was assessed by the Kolmogorov-Smirnov test in Minitab 17.1.0. Since the data did not meet the normality requirement, we used a non-parametric test for data analysis. A Kruskal-Wallis test was performed to assess whether sand fly capture rates per hour varied significantly between trap types for each site on separate days. Pairwise Wilcoxon signed-rank tests corrected for multiple comparisons using a Benjamini-Hochberg procedure were used to identify the trap type that accounted for significant differences [46]. The total sand fly species richness was quantified for each study site and trap type. Shannon-Wiener diversity indices of species evenness for each study site and trap type were calculated using the equation H ¼ À P n i¼1 p ln p, where "p" represents the proportion of each species "n" that was collected (∑p = 1). The Hutcheson t-test was used to assess differences in Shannon diversity between trap types at each site [47]. All analyses were carried out in PAST v3.12 and R v4.0.2.

Molecular screening of Leishmania and Bartonella DNA in sand flies
DNA extraction. Pools of non-engorged phlebotomine sand flies were air-dried to remove ethanol and ground with pellet-pestles in lysis buffer. Genomic DNA was extracted according to the manufacturer's protocol using the DNeasy Blood & Tissue kit (QIAGEN, Valencia, CA). The final DNA was eluted in 50 μL of elution buffer, and frozen at -30˚C. All PCR conditions for Leishmania and Bartonella screening are provided in S1 Text.
Molecular detection of Leishmania DNA. The initial screening for Leishmania was performed by PCR that targets Leishmania minicircle kinetoplast DNA (kDNA; approx. 700bp) conserved among species, using primers L.MC-1S and L.MC-1R as previously described [48] with some modifications on reagents concentration (S1 Text). Positive samples for kDNA PCR were confirmed by nested PCR targeting cytochrome b (cytB) [49,50]. Details of primers used are shown in Table 1. L. (V.) braziliensis and L. (Leishmania) infantum DNA were used as positive PCR controls whereas PCR mix without DNA was used as a negative control.
Molecular detection of Bartonella DNA. A PCR protocol that amplifies a segment between 400-600bp of the 16S-23S internal transcribed spacer (ITS) region was used for screening sand flies for Bartonella DNA [51]. Confirmation of positive samples was performed using a nested PCR protocol for gltA, the citrate synthase gene [52][53][54]. Both markers have been validated on a broad diversity of Bartonella species but have different advantages that work well in combination. ITS is highly sensitive for detection of Bartonella DNA, but sequence alignment and phylogenetic analysis are complicated by the many insertions and deletions present. In contrast, gltA is less sensitive but is more amenable to alignment and phylogenetic analysis. Furthermore, gltA is the most commonly used marker for molecular detection of Bartonella species, creating a large database for comparison [55]. Additional molecular markers, rpoB, the β subunit of bacterial RNA polymerase gene [56,57], and ftsZ, the filamenting temperature-sensitive mutant Z protein gene [58,59] were used to confirm ITS PCR and gltA PCR results (S1 Text). Details of primers used are shown in the Table 1. Bartonella bacilliformis DNA from culture (SANDI strain from Caraz, Ancash, Peru) [39] was used as a positive control and the PCR mix without DNA as a negative control. Both controls were included in each PCR run to evaluate the presence of appropriately sized amplicons and contamination, respectively. Only samples with PCR products that resulted in a correct amplicon size and a clear band were considered as positive for further DNA sequencing and phylogenetic analysis.
Leishmania and Bartonella DNA sequencing. The positive Leishmania and Bartonella PCR products were analyzed by electrophoresis on 2% agarose gel stained with GelRed (Biotium, Fremont, CA, USA). The PCR amplicons were purified using the Wizard SV Gel and PCR Clean-Up System (Promega, Madison, WI, USA) and sequenced along both DNA strands using the BigDye Terminator v3.1 Cycle Sequencing Kit (Applied Biosystems, Foster City, CA, USA) according to the manufacturer's instructions. Samples were sequenced at the Entomology Laboratory at NAMRU-6 (Lima, Peru) using an Applied Biosystems 3130 XL Genetic Analyzer sequencer. Sequences were confirmed as being Leishmania or Bartonella DNA using the NCBI Basic Local Alignment Search Tool (BLAST; https://blast.ncbi.nlm.nih.gov/Blast.cgi) prior to phylogenetic analysis.
Phylogenetic analysis of Leishmania DNA sequences. The alignment of cytB gene sequences obtained from sand flies was made using the ClustalW method, and sequences were analyzed with MEGA (Molecular Evolutionary Genetics Analysis) program version X [60]. The evolutionary distances were computed using the Kimura 2-parameter method [61]. Phylogenetic trees were constructed with the maximum likelihood method with Tamura-Nei model [62] and bootstrap support values for nodes were determined with 1,000 replicates [63]. Datasets of cytB Leishmania spp. reference sequences were obtained from GenBank. Phylogenetic analysis of Bartonella DNA sequences. Bartonella gltA sequences obtained from sand flies were aligned with reference sequences for named Bartonella species from Gen-Bank using MAFFT v7 [64]. The alignment was trimmed to a common length to eliminate poorly aligned positions using Phyutility v2.2 and Gblocks v0.91b [65,66]. Alignments were visually inspected for errors and manually corrected. A maximum likelihood tree was generated using RAxML v8 [67] using only the most closely matching reference sequences in the alignment. The GTRCAT model was used with 25 distinct rate categories and branch support was estimated with 1,000 bootstrap replicates.  Table). Lutzomyia davisi was the most abundant sand fly species in red LED trap (55.0%), Lu. auraensis was predominant in green LED and UV (70.5% and 29.6% respectively), Lu. whitmani in blue LED and CDC (34.6% and 24.6% respectively), and Lu. yucumensis was the predominant species in the sand fly collections with Shannon trap (55.7%).

Trap type evaluation for sand fly collection
The UV and blue LED traps collected the highest number of phlebotomine sand flies (53.4% and 21.3% respectively), while the green LED and red LED traps collected the lowest number (1.9% and 1.7% respectively) (S1 Table). To evaluate the effectiveness for sand fly collections of each trap type, the average of total sand fly specimens collected per hour per trap was calculated. Corrected for the total number of hours over two days of trapping, UV traps collected the greatest number of sand flies per hour (Fig 3). A Kruskal-Wallis test showed no significant differences in sand fly collections per trap type per day (chi-squared = 8.08, df = 5, p > 0.05), although there was large variation in sand fly trapping rates within each trap type over the two days of trapping (S1 Fig). Regarding species diversity index, UV traps collected a higher number of species richness (S2 Fig) and species evenness (Fig 4) than other trap types; the differences in Shannon diversity between UV traps and other traps at this site were all statistically significant (p < 0.05).
Tarapoto, San Martin. CDC, blue LED, UV, MM, BG-S, and Shannon trap were evaluated in three sites along Prolongación Alerta, Tarapoto district. A total of 1,335 phlebotomine sand flies were identified, 768 (57.5%) females and 567 (42.5%) males belonging to two genera, Lutzomyia (26 species) and Brumptomyia (2 species). Of them, 481 specimens were collected in the peridomicile and 854 inside the forest, but Kruskal-Wallis tests showed no significant differences in the sand fly species composition between both environments (p>0.05).  Table).
While the Shannon trap was used only on the first two days of trapping in favorable weather conditions, it captured the highest number of sand flies per hour compared to the other traps (Fig 3). During permanent rainy conditions on the last three days of trapping, UV traps performed better than the other four traps (S3 Fig). Kruskal-Wallis tests showed no significant differences in sand fly collections per trap type per day on the first two days with favorable weather (chi-squared = 5.8, df = 5, p > 0.05), on the last three days with permanent rain (chisquared = 2.4, df = 3, p > 0.05), and across all five trapping days (chi-squared = 8.04, df = 5, p > 0.05). As with results from Madre de Dios, these results are explained by high variation in capture rates across trapping days within each trap type (S1 and S3 Figs). Shannon and BG-S traps captured a smaller number of sand fly species richness (S2 Fig) with significant lower species evenness compared to other trap types (p < 0.05; Fig 4).
San José de Lourdes, Cajamarca. Palmal and Nuevo Porvenir, two rural communities, were selected to evaluate MM, CDC, UV, blue LED, and BG-S traps. The rain was consistent  Table).
The MM trap captured the greatest number of sand flies per hour compared to other trap types (Fig 3); BG-S traps did not capture any sand flies at this site. The Kruskal-Wallis test showed significant differences in sand fly collections per trap type per day (chi-squared = 13.2, df = 4, p < 0.05). However, the Wilcoxon signed-rank test with Benjamini-Hochberg FDR failed to detect significant differences (p > 0.05). The CDC trap captured a lower species richness compared to other trap types (S2 Fig) with significantly lower Shannon diversity (p < 0.05; Fig 4).
Caraz, Ancash. Choquechaca community, located 10 km from Caraz city in the inter-Andean Santa River valley, was the study site selected for evaluating CDC, UV, blue LED, and MM traps for sand fly collection. A total of 1,229 phlebotomine sand flies collected in peridomicile were identified, 1,170 females (95.2%) and 59 males (4.8%) belonging to two species, Lu. verrucarum (99.3%) and Lu. peruensis (0.7%) (S4 Table). Since there were only two species captured (S2 Fig) and a high predominance of Lu. verrucarum, Shannon diversity index was not calculated for this study site. MM traps captured the highest average number of sand fly specimens per hour and were the most effective traps (861 specimens, 70.1% of the total sand fly specimens), followed by CDC (189, 15.4%), UV (77, 6.3%), and blue LED (73, 5.9%) (Fig 3). Due to variation across trapping days (S1 Fig), no significant differences in sand fly collections per trap type were found by the Kruskal-Wallis test (chi-squared = 0.76, df = 3, p > 0.05).    Table 2 and S5 Table). The estimated minimum infection rates using this molecular marker were 1.29% and 1.79% in Madre de Dios and San Martin sand flies, respectively. All analyzed sand fly pools from Cajamarca and Ancash were negative for Leishmania DNA.

Detection of Leishmania
Nested cytB PCR and DNA sequencing for Leishmania. Nested cytB PCR amplicons from positive pools were sequenced. One pool of Lu. carrerai carrerai collected with a UV trap from Madre de Dios was found infected with L. (V.) guyanensis; one pool of Lu. hirsuta hirsuta collected with a blue LED trap from San Martin were found infected with L. (V.) naiffi ( Table 2). The phylogenetic tree with cytB DNA sequences is shown in Fig 5. The estimated minimum infection rates using this molecular marker were 0.13% and 0.16% in Madre de Dios and San Martin sand flies, respectively.  Table). Lu. whitmani (1 pool) from Madre de Dios collected with UV trap; Lu. nevesi (12 pools), Lu. hirsuta hirsuta (2 pools) and Lu. sherlocki (1 pool) from San Martin collected with CDC, blue LED, UV, and MM traps; and Lu. maranonensis (1 pool) from Cajamarca collected with blue LED were found infected with Bartonella spp. by ITS PCR, and confirmed by nested gltA PCR, and additional molecular markers rpoB and ftsZ PCR (Table 3 and S6 Table). The minimum infection rate based on positive gltA PCR was 1.02%, 1.29% and 0.65% for Madre de Dios, San Martin, and Cajamarca respectively. No positive pools were found in non-engorged female sand flies from Ancash.
Bartonella DNA sequencing. Amplicons from all positive pools by ITS and gltA were sequenced and confirmed as Bartonella DNA using BLAST. ITS sequences varied in length between 461-590 bp and the closest matching Bartonella species on GenBank were B. bacilliformis (average sequence identity = 91.5%) and B. ancashensis (average sequence identity = 88.1%). The presence of insertions and deletions in the alignment of ITS sequences complicates phylogenetic analyses, so phylogenetic trees were constructed only for the gltA gene. The DNA sequences of gltA had lengths between 319-366 bp and were classified into three lineages (Fig 6). Lineage 1 comprised Lu. whitmani (1 pool) from Madre de Dios, and Lu. nevesi (11 pools) and Lu. sherlocki (1 pool) from San Martin, which were found infected with genotypes most closely related to B. bacilliformis (average sequence identity = 89.6%). Lineage 2 comprised two pools of Lu. hirsuta hirsuta from San Martin and were also related to B. bacilliformis (average sequence identity = 89.6%). The third lineage comprised of one pool of Lu. nevesi from San Martin and one pool of Lu. maranonensis from Cajamarca. These species were found infected with a genotype most closely related to Candidatus B. rondoniensis (average sequence identity = 92%).

Discussion
This is the first study in Peru to evaluate the efficacy of different light and odor-baited trap types for phlebotomine sand fly collection in different ecological regions, and is the first study to determine the natural infection of Leishmania spp. and Bartonella spp. in multiple sand flies in different endemic areas outside of the Andean region. Both findings are unique in Peru and will help to better understand the behavior of phlebotomine sand fly vectors.
Knowledge about insect color vision suggests that there is a high sensitivity to light with short wavelengths that correspond to the ultraviolet spectrum in most insect families [68]. However, due to the higher diversity of visual conditions in which insects operate in nature and the higher diversity of habitats they have colonized, the sensitivity and response to spectral light at different intensities and wavelengths can vary across sites and among species, as has been reported previously [42,68,69]. In the Peruvian southeastern low jungle area (Madre de Dios region), where previous research has reported a great diversity and abundance of phlebotomine sand flies (>60 species), we have observed that the sand fly response of phlebotomine sand flies to light is in decreasing order at this site: UV > blue LED > incandescent light (CDC) > green LED > red LED (S1 Fig). The finding of a higher response to UV light in sand fly species from the Peruvian southeastern region is comparable to results obtained in Iraq and Italy [4,70]. The response of phlebotomine sand flies to green and red LED in this study was poor (1.7% of the total sand flies collected) in contrast to results obtained in Egypt [31], where the red LED was the most efficient trap type for catching Phlebotomus papatasi and results in the Brazilian rainforest [71,72], where they found that the green LED trap was the most effective in comparison to blue LED and CDC control (38-42% of the total sand flies collected).
Except for Shannon traps and protected human bait, traps that release CO 2 and other chemical attractants are very effective for collecting phlebotomine sand flies and other blood sucking insects as has been observed in Israel and Egypt [27,32], where different commercial models of MM releasing CO 2 only successfully captured P. papatasi and other sand fly species. Most studies with MM traps have been for surveillance of mosquito vectors of malaria and arboviruses [73,74] and experimental vector control programs [75]. For Aedes and other mosquito species, MM trap efficacy is increased when in addition to CO 2 the trap is supplemented with octenol [76]. However, when MM was compared with other traps, their efficacy was variable. Human Landing Collection (HLC) was more efficient for collection of Anopheles species (68.13%) compared with MM (31.88%) [34,77]; otherwise, MM collected more abundant specimens of Culicidae (59.94%) than CDC light traps (40.06%) [74]. In our study, the MM was the most efficient sand fly collection method in high jungle areas of Cajamarca region, when under unfavorable weather conditions with permanent rain, more than 80% of phlebotomine sand flies were collected with this trap type (Fig 3). Similar effectiveness of MM for sand fly collections was observed in inter-Andean valleys of Caraz, Ancash (Fig 3).
Our results of the trap type evaluation conducted in different ecological regions with distinct sand fly species composition indicate that phlebotomine sand fly behavior against different colors of light or chemical attractants is variable between study sites. In low jungle regions, UV and blue LED were more efficient and collected a higher diversity of sand fly species, whereas in high jungle and inter-Andean regions, MM and UV traps were more effective. In general, in our study we have observed that traps that use short wavelength light collect a higher diversity of species (Fig 4), while the traps that use chemical attractants as CO 2 and R-Octenol, are more selective to capture anthropophilic sand fly species, so it is recommended to use a combination of both trap types (traps based on light emission combined with traps that use chemical attractants) in entomological surveillance of potential sand fly vectors of leishmaniasis and bartonellosis.
It is known that leishmaniasis has been endemic in the Peruvian southeastern region for a long time [78], but the potential vectors have not been characterized until recent years. In communities located near the triple border between Peru, Brazil and Bolivia, 16 Lutzomyia species were found naturally infected with Leishmania parasites [37,44]. In the current study, we found Lu. carrerai carrerai naturally infected with L. (V.) guyanensis in Iberia, Madre de Dios. This Leishmania species has a wide distribution in the Amazon jungle, and in Peru it has been isolated from patients with cutaneous leishmaniasis from high and low jungle regions [79,80]. At the same study site, Lu. whitmani was found naturally infected with a genotype of Bartonella spp. (lineage 1) most closely related to B. bacilliformis based on gltA gene DNA sequences. There are no reports of autochthonous Carrion's disease transmission in Iberia or any other sites in Madre de Dios region. Recently was reported the detection of B. bacilliformis in ticks collected from wild mammals in San Lorenzo [81], a small town located along the interoceanic highway in Madre de Dios, using a real time PCR protocol, but these results are doubtful because they were not supported by DNA sequencing and phylogenetic analysis [82,83]. Currently, all efforts to find B. bacilliformis in other mammalian reservoirs other than humans have failed, as well as any incrimination of other arthropod vectors besides Lutzomyia species [14,[82][83][84][85].
In San Martin, northern Amazonian region, where leishmaniasis has higher prevalence and is an important risk area for transmission of this disease [6,86], we found Lu. hirsuta hirsuta naturally infected with L. (V.) naiffi in a rural area located only 5 kilometers from Tarapoto city, along the road to Cordillera Escalera. This is the first report of L.  [80]. Leishmania naiffi causes cutaneous leishmaniasis and is widespread in the Brazilian Amazonian region [87] and has also been found infecting human patients as well as Lu. tortura in the northern Amazonian region of Ecuador [50]. Our results suggest that L. (V.) naiffi has a wide distribution in the Amazon region and could be responsible for leishmaniasis cases in San Martin department, where L. (V.) guyanensis has also been detected in human patients [79,80]. Further studies on human populations are necessary to confirm L. (V.) naiffi transmission in Peru. In our study sites in San Martin we also found Lu. nevesi, Lu. hirsuta hirsuta and Lu. sherlocki naturally infected with two Bartonella genotypes related to B. bacilliformis based on gltA gene sequencing (lineages 1 and 2), and a third genotype of Bartonella spp. related to Candidatus B. rondoniensis naturally infecting Lu. nevesi (lineage 3). This is the first report of detection of Bartonella DNA in Lutzomyia species from the northern Amazonian region of San Martin. Leishmaniasis cases have a high prevalence in San Martin and are reported yearly by the Ministry of Health, but information about human bartonellosis in this region is very limited, with only a few cases reported during the last years but without evidence of autochthonous transmission [88]. Interestingly, we found that all sand fly species infected with Bartonella and Leishmania in Madre de Dios and San Martin regions had previously been reported as potential leishmaniasis vectors in the Amazon region of Peru and other countries [3,44].
The Bartonella DNA genotype related to Candidatus B. rondoniensis (lineage 3) was also found naturally infecting Lu. maranonensis in San Jose de Lourdes, a high jungle area in Cajamarca region. This sand fly species was recently found infected with B. bacilliformis in an endemic area of Carrion's disease in Cutervo province (Cajamarca central region) and has been incriminated as a potential vector of this disease [41]. Lutzomyia maranonensis and Lu. robusta are the most abundant phlebotomine sand flies in the Peruvian northeastern region and southern Ecuador (Zamora Chinchipe province), and it has been suggested that these sand fly species could be the vectors of Carrion's disease and leishmaniasis in the Peru-Ecuador border regions [19,89,90]. Our results and those of Ulloa et al. [41] show that Lu. maranonensis is capable of harboring different Bartonella species. Similar results were reported by Villaseca et al.
[38] who found Lu. peruensis naturally infected with B. bacilliformis and B. taylori in the Urubamba valley, Cuzco region, Peru, where an important outbreak of acute Carrion's disease occurred during 1997-1998 without previous history of this disease in this region [91]. Candidatus B. rondoniensis was first found in triatomine bugs vectors of Chagas disease, Eratyrus mucronatus from French Guiana [92]. In Peru, there are no reports of Bartonella species infecting triatomine bugs, but different species of Triatominae are distributed across the country, especially in the northern region (San Martin, Cajamarca and Amazonas regions) where Chagas disease is endemic [93][94][95][96]. Further studies are needed to determine if triatomine bugs in Peru are also infected with Bartonella spp.
Regarding sand fly species found naturally infected with Bartonella spp. in this study, the majority of them are recognized as potential vectors of leishmaniasis, including Lu. whitmani, Lu. nevesi and Lu. sherlocki [44], whereas Lu. maranonensis is a suspected vector of leishmaniasis along the Peru-Ecuador border [90]. Our detection of Bartonella DNA related to B. bacilliformis in sand flies was unexpected. Due the high diversity of Bartonella species in different vertebrate hosts such as bats, rodents, cats, dogs, non-human primates and other wild vertebrates [91,[97][98][99][100][101], and arthropod vectors such as ticks, fleas, body lice, biting flies and triatomine bugs [11,12,102,103], it is unclear which hosts are reservoirs for B. bacilliformis-like lineages 1 and 2 or Candidatus B. rondoniensis-like lineage 3. While the lineages were most closely related to B. bacilliformis and Candidatus B. rondoniensis, they are still quite distant from these species and each other in absolute genetic relatedness (<95% sequence similarity) and may thus represent novel Bartonella species [104]. However, additional markers beyond gltA must be used to clarify the phylogenetic associations of these lineages with other Bartonella species [55,83,104]. Our data and that from Laroche et al. [92] with the description of Candidatus B. rondoniensis indicate that there is an underappreciated diversity of Bartonella lineages present in South American sand flies and other insects that may give us deeper insight into the origins of B. bacilliformis. Finally, the role of these Bartonella species in human disease is unknown. Further studies are needed to determine if these Bartonella genotypes detected in sand flies from low jungle (Madre de Dios) and high jungle (San Martin and Cajamarca) regions have the ability to cause human disease, or only infect vertebrate hosts.
According to our trap type evaluation in endemic areas of leishmaniasis and bartonellosis in Peru, the effectiveness of different traps for collecting sand flies vary according to ecological characteristics, climatic conditions and sand fly fauna composition across each site. Therefore, we recommend combining different traps to efficiently survey sand fly diversity across multiple habitats and maximize the surveillance of pathogens they may carry. Regarding pathogen detection in sand flies, this is the first report of L. (V.) naiffi in Peru and of detection of Bartonella spp. genotypes distantly related to B. bacilliformis and Candidatus B. rondoniensis infecting sand fly species from the Amazon jungle. Our results underscore the need for increased studies of the ecological aspects of Amazonian sand fly species, and to increase Carrion's disease surveillance in non-endemic areas, especially in the Amazon jungle where the risk of transmission for this disease can be higher due the capacity of sand fly species to transmit different human and newly emerging pathogens.

Disclaimer
The views expressed in this article are those of the author and do not necessarily reflect the official policy or position of the Department of the Navy, Department of Defense, nor the U.S. Government.  (1): S166-S171.