Sand fly fauna of Crete and the description of Phlebotomus (Adlerius) creticus n. sp. (Diptera: Psychodidae)

The Greek island of Crete is endemic for both visceral leishmaniasis (VL) and recently increasing cutaneous leishmaniasis (CL). This study summarizes published data on the sand fly fauna of Crete, the results of new sand fly samplings and the description of a new sand fly species. All published and recent samplings were carried out using CDC light traps, sticky traps or mouth aspirators. The specific status of Phlebotomus (Adlerius) creticus n. sp., was assessed by morphological analysis, cytochrome b (cytb) sequencing and MALDI-TOF protein profiling. Published data revealed the presence of 10 Phlebotomus spp. and 2 Sergentomyia spp. During presented field work, 608 specimens of 8 species of Phlebotomus and one species of Sergentomyia were collected. Both published data and present samplings revealed that the two most common and abundant species were Phlebotomus neglectus, a proven vector of Leishmania infantum causing VL, and Ph. similis, a suspected vector of L. tropica causing CL. In addition, the field surveys revealed the presence of a new species, Ph. (Adlerius) creticus n. sp. The identification of the newly described species is based on both molecular and morphological criteria, showing distinct characters of the male genitalia that differentiate it from related species of the subgenus Adlerius as well as species-specific sequence of cytb and protein spectra generated by MALDI-TOF mass spectrometry.

The main aims of this study were to review historical data on the sand fly fauna of Crete and describe a new Phlebotomus (Adlerius) species sampled during surveys carried out from 2014 to 2019. To confirm the status of the new species, in addition to morphological and genetic criteria, for the first time, protein profiling using MALDI-TOF mass spectrometry was also deployed.

Sand fly data
All published data recording sand fly presence in Crete were gathered. The literature search concerning the review follows the Prisma Journal Publishing protocol workflow [8]. PubMed, Web of Science, Google Scholar databases and web searches were screened from 1910 to 30 November 2019, using the following keywords: Phlebotominae, Phlebotomus, Crete, Sergentomyia, Greece, sand flies, CDC light traps, sticky traps, mouth aspirator, electric aspirator. Full text articles in English containing information on phlebotomine sand flies from Crete were selected. Other articles, including those in other languages, that contain desired information were also included, based on the cited databases knowledge of the authors. Samplings were carried out using CDC miniature light traps (John W. Hock Co., Gainesville, FL, USA), Sticky Traps (A4 paper coated with castor oil) or mouth aspirators. The sampling sites were mostly animal farms, houses, schools, churches, deserted houses, wells, caves and rural areas with different cultivations or wild vegetation.

Study areas and sampling
The new samplings were carried out in five study areas, two in Heraklion, two in Chania and one in Lasithi between 2014 and 2019 ( Table 1). The study areas were Fodele and Foinikia in Heraklion prefecture, Agia Roumeli and Botanical Gardens in Chania and Xerokampos in Lasithi. The CDC miniature light traps equipped with a fine net cage were used for all the samplings. In all sampling areas, 7-9 light traps were placed in different microhabitats for one to two days. The light traps were set 1.5 h before sunset and were collected 2 h after dawn. Specifically, in Fodele 9 light traps for 2 sampling nights in May 2019 were used, in Foinikia 9 traps for 1 night in August of 2018, in Agia Roumeli 7 traps for 2 sampling nights in May of 2014, and in Botanical Garden 9 traps for 2 sampling nights in August of 2019. In Xerokampos two samplings were carried out during April of 2014 and in May of 2019 using 9 traps for 2 sampling nights.

Sand fly morphological identification
The specimens were stored in vials containing absolute ethanol HPLC grade (Thermo Fischer Scientific, Gloucester, UK) at room temperature prior to mounting, except for the specimens later analyzed by MALDI-TOF protein profiling which were stored in 70% molecular grade ethanol at − 20 °C before further processing. The samples were divided into specimens mounted in toto and specimens processed for molecular biology. In the latter case, the head, thorax including wings and genitalia were removed and placed in a drop of ethanol before their processing similar to specimens mounted in toto, while the other parts of the body were kept in ethanol for molecular analysis. Soft tissues were lysed in a bath of KOH 10%, then bleached in Marc-André solution, and mounted between microscope slide and cover slide in Euparal ® for species identification after dehydration in graded ethanol series [9]. Some specimens were mounted immediately after clearing in Marc-André solution or in high viscosity CMCP-10 medium (Polysciences, Inc., Warrington, PA, USA). Morphological identification was performed under a BX61 microscope (Olympus, Japan). Measurements and counts were taken using the Stream Motion software (Olympus, Japan) and a video camera connected to the microscope. Identification was performed based on the keys available for Crete and adjacent regions [10][11][12][13][14].
The terminology adopted for the characters is the most recent one for phlebotomine sand flies [15]. The following measurements were made for the specimens of the new species for both sexes: flagellomeres 1, 2 and 3, labrum-epipharynx. For males, we also measured the lengths of the parameral sheath, the distance from the tubercle to the tip of the parameral sheath (indicated as the distance from the tip of the aedeagus to the subterminal tooth by Artemiev [10]), sperm pump, aedeagal ducts, gonocoxite, beginning and ending of the tuft of internal setae of the gonocoxite, number of internal setae of the gonocoxal tuft and area of the internal tuft of setae of the gonocoxite (Fig. 1).
The percentage of tuft length vs gonocoxite length was calculated as follows: tuft length × 100)/gonocoxite length. The median tuft position vs gonocoxite was calculated as follows: beginning of the tuft of internal setae of the gonocoxite + tuft length/2) × 100/gonocoxite length. Drawings were made using a camera lucida.

DNA extraction and cytochrome b mtDNA sequencing
Following morphological identification, DNA was extracted using DNeasy Blood & Tissue Kit (Qiagen, Hilden, Germany) from 16 individuals (males and females) of Ph. (Adlerius) creticus n. sp. collected during previous surveys at four localities in Crete as well as from two specimens of Ph. simici Nitzuescu, 1931 from Crete and three specimens of Ph. balcanicus Theodor, 1958 from Iran. cytb PCR assays were performed using primers and conditions described by Esseghir et al. [16] and sequenced both directions according to Sanger's method using the same primers as in the PCR assay.

Phylogenetic analyses
The optimal nucleotide substitution models were identified using PartitionFinder (PF) v.2.1.1 [25]. We ran PF two different times using the greedy search algorithm with linked branch lengths in calculations of likelihood scores under the Bayesian information criterion (BIC). The difference between these two runs was the restriction of candidate models to only those that are available in either MrBayes v.3.2.6 [26] or PhyML v.3.0 [27]. The models which included both gamma distribution (G) and invariable sites (I) were ignored [28].
Phylogenetic trees were constructed using Bayesian Inference (BI) and Maximum Likelihood (ML) methods. The BI analysis was performed in MrBayes v3.2.6 [26] with 4 runs and 8 chains per run for 10,000,000 generations, with a sampling frequency of 100. From the sampled trees, 25% were discarded as ‛burn-inʼ phase and therefore, a majority rule consensus tree relied on the remaining trees and posterior probabilities were calculated as the percentage of samples recovering any particular clade [29]. ML analysis was performed with PhyML v.3.0 [27] with nearest-neighbor-interchange search, bio-neighbor joining starting tree under the suggested models selected in PF. Bootstrap values were estimated by 1000 replicates [30]. Phlebotomus perfiliewi represented the outgroup in the phylogenetic analyses.

MALDI-TOF protein profiling
Samples that were subjected to MALDI-TOF MS analysis were processed as previously described [31]. It was demonstrated that mass spectrometry-based approach is not suitable for specimens collected by sticky traps [32]; therefore, only specimens collected by CDC light traps were included into the assay. Within one month after the collection in the field, specimens stored in 70% ethanol were dissected, heads and terminalia were mounted by CMCP-10 mounting medium (Polysciences) for morphological typing, rest of abdomens were stored for DNA isolation and cytb sequencing as described above and thoraxes were ground by a manual BioVortexer homogenizer (BioSpec, Bartlesville, USA) with sterile disposable pestles in 10 μl of 25% formic acid. Two µl of the homogenate were mixed with 2 µl of freshly prepared MALDI matrix, an aqueous 60% acetonitrile/0.3% TFA solution of sinapinic acid (30 mg/ml; Sigma-Aldrich, St. Louis, USA). One μl of this mixture were spotted directly onto a steel MALDI plate in duplicates. Protein mass spectra were measured using an Autoflex Speed MALDI-TOF spectrometer (Bruker Daltonics, Billerica, USA) in a mass range of 4-25 kDa and compared by FlexAnalysis 3.4 software. For cluster analysis and species identification, the protein profiles were processed using MALDI Biotyper 3.1 and searched against an in-house database that comprises reference spectra of 23 different sand fly species including following Adlerius species: Ph. arabicus Theodor, 1953, Ph. balcanicus, Ph. halepensis Theodor, 1958 and Ph. simici.

Sand fly data
The first published study of the sand fly fauna of Crete appeared in 1910 and since then, 16 more works were published, revealing the presence of 10 Phlebotomus spp. and 2 Sergentomyia spp. (Table 3).

Sand fly sampling in 2014-2019
Overall, 608 sand fly specimens were collected which corresponded to nine different species with the most common and abundant species being Se. minuta and Ph. neglectus (Table 4). In addition to species known in Crete, 151 specimens of Ph. (Adlerius), morphologically close to Ph. balcanicus and Ph. zulfagarensis were identified. These specimens were further analysed and their morphological and molecular identification is described below. Table 5 sums up the total number of specimens, of each species, caught in present samplings and previous publications. More than 30,000 specimens have been reported and about 63.27% of them were identified as the medically important species Ph. neglectus. Moreover, a significant proportion of the specimens were identified as Ph. similis (17.4%) and Ph. papatasi (9.49%). In all prefectures except Chania, Ph. neglectus (> 50%) and Ph. similis (~30%) were the most common and abundant species. In Chania, Ph. papatasi was the most common and abundant species which comprised almost the 60% of the sand fly specimens (Fig. 2).

Morphological typing of Ph. creticus n. sp
Fifty-four male individuals collected by CDC light traps were used for the morphological analysis, originating from three localities (8 from Toplou, 23 from Xerokampos and 23 from Fodele). The descriptive statistics for 22 characters are given in Table 6. Eleven females from Xerokampos were used for the description, descriptive statistics for 6 characters are given in Table 7. To exclude a presence of two distinct populations, the normality of the distributions of each of the morphological characters was tested by Shapiro-Wilk test (P > 0.05) and was found not to be significantly different from normal.

Sequence analysis
The cytb gene was successfully amplified and sequenced and the final dataset consisted of an alignment of 446 bp. The pairwise distances between species ranged from 0.044 to 0.188, while the closest related species to Ph. creticus n. sp. appeared to be Ph. balcanicus with a mean distance of 0.044. The mean distance within the Ph. creticus n. sp. samples was 0.008. The pairwise mean distances between and within species are provided in Table 8.

Phylogenetic analyses
The best-fit nucleotide substitution model for cytb was Hasegawa-Kishino-Yano (HKY) + I for all codon positions and for both MrBayes and PhyML. Both analyses led to phylogenetic trees with similar topologies and hence, only the consensus MrBayes tree including posterior probabilities and bootstrap values is presented here (Fig. 3). All species used in these analyses formed monophyletic clades and specifically, the specimens of Ph. creticus n. sp. formed a well-supported monophyletic clade. Phlebotomus creticus n. sp. appears to be more closely related to Ph. balcanicus than the other species included in the analyses.

MALDI-TOF protein profiling
In total, 28 specimens of five Phlebotomus species from Xerokampos (n = 26) and Fodele (n = 2) were analysed: Ph. creticus n. sp. (n = 12), Ph. killicki (n = 1), Ph. neglectus (n = 5), Ph. similis (n = 7) and Ph. simici (n = 3). Reproducible protein spectra with a high number of intense signals within the mass range of 4-25 kDa were generated for all analysed specimens. These protein profiles were species-specific and showed species-unique peaks that allowed reliable and conclusive differentiation of the analyzed specimens. Protein spectra of the specimens identified based on morphology as belonging to four known species were similar to the corresponding reference spectra of the respective species. Protein spectra of all Ph. creticus n. sp. specimens were identical and differed substantially from the spectra of other species as shown by a hierarchical cluster analysis (Fig. 4a).  13:547 Moreover, they also showed a number of specific peaks, not shared by protein spectra of four Adlerius species in the reference database (Fig. 4b).        Setae randomly implantated on tergites II to V. Presence of papillae on tergites III to VII. Genitalia. Presence of about 8 + 1 setae on tergite VIII. Lack of protuberance on tergite IX. Spermathecae incompletely segmented. Basal part of the ducts wide with thick walls. Those of the paratype Crete IT8 have been collapsed during the mounting. The measurements indicated are those of the paratype crete7 mounted in Marc-André to be observed, measured and drawn before final process and mounting. Length of the ducts: 600 µm (including 100 µm of the wide basal part and 500 µm of the narrow ducts); length of the body: 100 µm. Genital fork 192 µm long. Cerci rounded at their top, 178 µm long. No seta observed on the sternite X.

Differential diagnosis
In males, number of ascoids usually 2/f1-f3, 1/f4-f13, a range of 54-85 setae on the gonocoxal internal tuft equally distributed between the proximal and distal halves of the gonocoxite. At Fodele, Ph. creticus n. sp. was captured in four CDC light traps placed near or in cave entrances, facing southeast and south. The vegetation around was rich. Mainly chicken, rats and lizards dwell around the traps. No specimens of Ph. creticus n. sp. were collected by sticky traps. At Xerokampos, the species was captured in two traps placed at the entrance of two shallow caves facing southeast, at 200 m distance from the shore. The vegetation around the caves was composed of phrygana, biotope typical for island of Crete. At Agia Roumeli the species was captured in two traps placed at the entrance of two limestone rock caves, one facing east and the other facing south, in a separate beach located west of Agia Roumeli, just a few meters from the sea. The vegetation above the caves was composed from sparse phrygana. In Toplou Monastery, it was trapped in caves along the wall just before the Monastery.

Discussion
The island of Crete is an important and long-time active region of leishmaniases transmission and still provides ample numbers of human cases annually. As sand flies are the only proven vectors of these diseases in the Mediterranean basin, there is a need for sustained monitoring of the sand fly fauna and its role in Leishmania transmission. While studied for a long time, there are surprising new findings regarding the presence of sand fly species as documented by a recent description of Phlebotomus (Transphlebotomus) killicki Dvorak, Votypka & Volf, 2015 in the localities at the southern coast of the island [49]. The results of the presented study reveal that there is still more to discover.
Published data as well as new results from the present sampling (2014-2019) show that the most widespread and abundant species in Crete is Ph. neglectus, accounting for more than 60% of the recorded sand fly specimens. It is a proven vector of L. infantum in Greece, several other Balkan countries and the western part of Turkey [1]. Its abundance in all studied areas of Crete apparently contributes to the geographical distribution of VL cases throughout the island and as other species of the subgenus Larroussius are markedly scarce, we may conclude that Ph. neglectus remains a sole vector of the disease. The second most common and abundant species on the island is Phlebotomus similis. It is morphologically similar and phylogenetically closely related to Ph. sergenti [51,52] and is regarded as a suspected vector of L. tropica in regions where Ph. sergenti is not present, including Crete [2]. The fact that Ph. similis was found in all foci of human CL further fosters the considerations of this species as a yet unproven vector of L. tropica in the island. Phlebotomus papatasi is a proven vector of L. major in the Middle East and Northern Africa and it is known to be responsible for the phlebovirus infections in the Old World [53,54] including Crete [55]. However, L. major, the causative agent of zoonotic cutaneous leishmaniasis, does not circulate in Crete, probably due to absence of suitable reservoir species (gerbils). Vectorial competence of Sergentomyia spp. to mammalian diseases is still controversial and experimental studies that would conclusively test it are scarce [56] despite growing circumstantial evidence that suggest incrimination of some species in the transmission of human leishmaniasis and phleboviruses [57]. Of two Sergentomyia species previously recorded from Crete, only Se. minuta was found in this study. Bloodmeal analyses of population in southern Portugal recently demonstrated that this widely distributed Mediterranean species may be at least partially anthropophilic and the detection of Leishmania DNA in specimens from the same area emphasizes the need of further studies about the role of Sergentomyia species in leishmaniases transmission cycles [58].
Adlerius specimens collected at different sites throughout the island during 2014-2019 did not meet the criteria of any known species and exhibited unique morphological characters. Moreover, obtained sequences of cytb, when compared with sequences of species within the subgenus Adlerius available in public databases, were substantially different. That led us to the conclusion that the collected specimens represent a new species. The identity of this new species is based on a detailed morphological analysis of decisive characters that is further supported by sequencing of cytb gene, a widely used genetic marker, and for the first time in a description of a new sand fly species, also by comparison of speciesspecific protein spectra acquired by MALDI-TOF mass spectrometry.
According to Artemiev [10], the subgenus Adlerius includes 17 described and two undescribed species. While the females mostly appear to be undistinguishable, the main characters to identify the males of this subgenus are the antennal formula, used in all available identification keys [10,11,59] and the number of setae of the internal tuft of the gonocoxite as well as the position of this tuft on the gonocoxite.
Phlebotomus creticus n. sp. exhibits a range of 54-85 setae on the gonocoxal internal tuft (mean: 69). This count excludes the identification of Ph. balcanicus (92-130 setae after Artemiev [10] and more than 100 in the original description [reference number]), Ph. comatus (126-220 setae) and Ph. kyreniae (30-40 setae). There is an overlap related to this number of setae for Ph. creticus n. sp. and Ph. angustus (35-69 setae), Ph. salangensis (40-85 setae) and Ph. zulfagarensis (66-72 setae). However, these three species exhibit group of setae which is completely or at 90% in the proximal half of the gonocoxite [10] whereas that of Ph. creticus n. sp. is equally distributed between the proximal and distal halves of the gonocoxite ( Table 6, Fig. 1). Moreover, all three species are known to be distributed in Asia, in regions very distant from Crete. In contrast, in Ph. balcanicus the hair group is located mainly on the basal half of the coxite [12]. Consequently, regarding the typological systematics, there is substantial evidence to consider Ph. creticus as a distinct species.
Phlebotomus creticus n. sp. was recorded at different localities at both northern and southern coast of Crete divided by a mountain range of a considerable height, which further supports a long presence rather than a recent introduction. However, it is probably not a common species. Phlebotomus creticus n. sp. was not caught on sticky traps placed in the openings of holes on natural or manmade walls and ground. We assume that it preferentially rests inside shallow caves in limestone rocks. In such specific biotopes it could be a dominant species as demonstrated in Xerokampos. Its feeding preferences should be investigated, although wildlife hosts may be expected (mice, birds and possibly lizards). Beside this newly described species, Ph. simici of the same subgenus that had been reported from Crete in the past, was still recorded in this study, occurring even sympatrically at the type-locality Xerokampos. We may speculate that in previous entomological surveys, some of the specimens identified as Ph. simici or unidentified Adlerius sp. may be attributed to the newly described Ph. creticus n. sp.
The results of the molecular analyses support the description of a new species. Phylogenetic analyses of the cytochrome b gene, a mitochondrial marker widely used in phylogenetic studies of many insect groups including sand flies, strongly grouped all analyzed specimens of Ph. creticus n. sp. in a distinct monophyletic clade. This genetic marker was chosen as it provides the best coverage of the species within the subgenus Adlerius. Unfortunately, no sequences of any genetic marker are so far available for some species of the subgenus, including the three species with overlapping numbers of setae on the gonocoxal internal tuft. Genetic distance values obtained for Ph. creticus n. sp. and other compared Adlerius species are comparable with the distances recorded previously for other sand fly species, as shown by studies of morphologically-as well as genetically distinct species of the subgenera Larroussius and Phlebotomus [60] or Madaphlebotomus [9]. Moreover, MALDI-TOF protein profiling demonstrated that all processed specimens of Ph. creticus n. sp. produced unique, reproducible and species-specific profiles that clearly differentiate them from other species outside and within the subgenus Adlerius. This method of mass spectrometry has recently become a popular tool for species identification of various organisms including arthropod vectors as it is simple, rapid and cost-effective [61]. Here, we demonstrated for the first time that it can also successfully complement traditional morphological approach and established DNA-based molecular taxonomy in the process of revealing yet unrecognized sand fly species.
Our recent findings urge the need for a revision of the subgenus Adlerius using both morphological and molecular approaches. Species boundaries are not well defined and the vicariance of this group probably occurred recently as for other groups of phlebotomine sand flies such as Phlebotomus, Larroussius or Paraphlebotomus [16,18,62].

Conclusions
In this study we present a review of sand fly species recorded in the past and at present in Crete, an island with ongoing transmission of two Leishmania species due to the presence of competent sand fly vectors. The importance of this research is highlighted by the geographical position of the island and the current possibility of accidental introduction of more Leishmania species due to human migration and other activities. According to our findings, 10 Phlebotomus spp. and 2 Sergentomyia spp. were recorded, with the most common and abundant species being Ph. neglectus. We may assume that the findings of Ph. mascittii reported prior to the description of Ph. killicki may be attributed to the latter species. We identify and describe a new species Phlebotomus (Adlerius) creticus n. sp. from various localities in Crete. Its identification is based on morphological characters of the male genitalia that particularly differentiate it from related species of the subgenus Adlerius. The identity of the newly described species was confirmed by two molecular approaches (MALDI-TOF protein profiling and cytb sequence analysis). As there is no data on the vectorial competence and capacity of this new species, its potential role in the autochtonous transmission cycles of Leishmania shall be further studied.