Molecular phylogeny of Psychodopygina (Diptera, Psychodidae) supporting morphological systematics of this group of vectors of New World tegumentary leishmaniasis

New World sandflies are vectors of leishmaniasis, bartonellosis, and some arboviruses. A classification based on 88 morphological characters was proposed 27 years ago when the New World phlebotomines were organized into two tribes Hertigiini and Phlebotomini. The latter was structured into four subtribes (Brumptomyiina, Sergentomyiina, Lutzomyiina, and Psychodopygina) and 20 genera. The subtribe Psychodopygina, including most of the American vectors of tegumentary Leishmania comprises seven genera from which no molecular work has been produced to support this classification. Here, we carried out a molecular phylogeny based on combined sequences (1,334 bp) of two genes: partial 28S rDNA and mtDNA cytochrome b from 47 taxa belonging to the Psychodopygina. The Bayesian phylogenetic reconstruction agreed with the classification based on morphological characters, supporting the monophyly of the genera Psychodopygus and Psathyromyia, whereas Nyssomyia and Trichophoromyia seemed to be paraphyletic. The paraphylies of the two latter groups were exclusively caused by the doubtful position of the species Ny. richardwardi. Our molecular analysis provides additional support to adopt the morphologic classification of Psychodopygina.


Introduction
Leishmaniases are widespread vector-borne diseases caused by about 20 species [32] of the genus Leishmania, transmitted by the bite of Phlebotomine sandflies. They exhibit three main clinical forms: visceral (VL), cutaneous (CL), and mucocutaneous (MCL) leishmaniasis, with a global annual incidence estimated to be between 0.7 and 1 million new cases [9]. In the Americas, 15 species of Leishmania are known to affect humans [17]. Among these species, only L. infantum is responsible of VL in the New World, with an incidence estimated between 4,500 and 6,800 cases per year. The incidence of both CL and MCL is much higher: between 187,200 and 300,000 new cases yearly [9], mostly caused by L. braziliensis. Sandfly vectors have been identified for 12 of these 15 species, with the remaining three (L. venezuelensis, L. lindenbergi, and L. martiniquensis) not having proven vectors [2,14,29,32].
Due to long co-evolution [2,32], there is a strong link between the Leishmania species and the groups of sandflies transmitting them. In the Old World, L. infantum is mostly transmitted by sandflies belonging to the subgenus Larroussius, whereas L. major is transmitted by members of the subgenus Phlebotomus. In the New World, all the proven vectors of VL belong to Lutzomyiina, whereas most of the vectors of CL and MCL belong to Psychodopygina, especially the genera Psychodopygus and Nyssomyia, the main vectors of L. braziliensis ( Table 1).
Here we provide for the first time a molecular phylogeny of the Psychodopygina subtribe to compare it with the current morphological classification in use. This work allows for a better understanding of the evolutionary relationships within this group.

Sandfly collection
Sandflies were captured using CDC miniature light traps in Mexico, Nicaragua, French Guiana, Ecuador, Brazil, and Bolivia. A total of 48 taxa representing six genera of the Psychodopygina (Psychodopygus, Nyssomyia, Trichophoromyia, Psathyromyia, Bichromomyia, and Martinsmyia) and one species of the genus Evandromyia as an outgroup representative species of the Lutzomyiina subtribe were analyzed (Table 2).
Head, wings, and genitalia from specimens were cleared and mounted between a slide and cover slide [1]. The thorax related to each specimen was stored at À20°C or in 96% ethanol before DNA extraction. Some specimens were dried and stored at À20°C or room temperature waiting for processing. Routinely, two specimens were processed for each taxon, except for scarce species for which only one specimen was processed, and for Ny. richardwardi (four specimens processed).

DNA extraction, PCR amplification, and sequencing
DNA was extracted using a QIAmp Ò DNA Mini Kit (Qiagen, Hilden, Germany), following a classical protocol [12]. Polymerase chain reactions (PCR) were performed in a 50 lL volume using 5 lL of DNA extracted solution and 50 pmol of primers C3B-PDR/NIN-PDR [18], and C1'/D2 [41] described previously to amplify partial cytochrome b (cytb) from mitochondrial DNA and domains D1-D2 from ribosomal DNA 28S, respectively. PCR products were sequenced in both directions using the primers for DNA amplification. Sequences obtained are available in GenBank and accession numbers are indicated in Table 2.

Sequence alignment and phylogenetic analyses
Sequences were edited and aligned using Muscle software [16] and then manually checked for compliance with the criteria: (1) minimizing the number of inferred mutations; (2) favoring substitutions over insertions and deletions; and (3) favoring transitions over transversions due to the higher probability of their occurrence.
The TVM+I+G model of molecular evolution was determined for both data partitions out of 88 possible models with JModeltest v0.1.0. [9,39] via the Akaike Information Criterion (AIC) [8]. The AIC values for these models in JModeltest were 7062.889520 for cytb and 11236.768300 for D12, both with a delta AIC of zero. Phylogenetic reconstruction was conducted using Bayesian (BA) inference in MrBayes [43]. Bayesian analyses on the combined two-genes dataset were performed using a mixed model with a partition by gene assigning independent models of evolution to each partition. All parameters were unlinked between partitions, except topology and branch lengths. Analyses consisted of two runs of four simultaneous Markov chains each for 10 million generations, sampling a tree every 1,000 generations and applying a 25% burn-in after checking for convergence using TRACER v1.4 [40] and AWTY [36] to confirm that the standard deviation of split frequencies approached zero. The resulting trees were kept, calculating posterior probabilities in a 50% majority-rule consensus tree. The species Evandromyia (Aldamyia) walkeri, a taxon belonging to the Lutzomyiina, was used as outgroup in our analyses because this subtribe is the sister group of Psychodopygina (Fig. 1).

Sequencing
Sequence fragments for D1 and D2 amplified segments, including the flanking conserved domains of rDNA28S were 687 to 690 base pairs (bp). Their alignment included 690 nucleotide positions: 535 constant, 153 variable but parsimony uninformative, and 86 were parsimony informative. The lengths of partial cytochrome b sequences obtained ranged from 501 to 545 bp. The final alignment dataset included 545 nucleotide positions: 288 were constant, 235 were variable but parsimony uninformative, and 176 were parsimony informative. No intraspecific variation was observed for both markers.

Phylogenetic reconstruction
The Bayesian analysis of the combined two-gene dataset (1,334 bp) recovered highly supported monophyletic groups with posterior probabilities (PP > 0.9) for most of the clades (Fig. 2). Some of these same clades were not recovered using both genes independently as nodal support was not strong (PP < 70) (Supplementary Figures S1 and S2).
The recovered phylogenetic reconstruction supports the monophyly of the genus Psathyromyia, sister to all other Psychodopygina members. The genus Psathyromyia was composed of three subgenera: Forattiniella, Psathyromyia, and Xiphopsathyromyia. The subgenera Forattiniella and Xiphopsathyromyia were found to be monophyletic, whereas the subgenus Psathyromyia appeared as paraphyletic.
The monophyly of the genus Martinsmyia was not confirmed because we processed only one species. Moreover, the genus Martinsmyia was composed of the groups alphabetica (here represented by Mt. brisolai) and gasparviannai from which no sample was analyzed. The genus Trichophoromyia was recovered as paraphyletic because it included one species belonging to the genus Nyssomyia (Ny. richardwardi) (Fig. 2), rendering the genus Nyssomyia also paraphyletic. The position of Ny. richardwardi varied depending on the selected marker. According to D1D2 28S rDNA, it was recovered within the genus Trichophoromyia, sister to Th. reburra. On the other hand, the analysis of cyt b placed this species sister to Mt. brisolai ( Supplementary Figures S1 and S2). If we excluded Ny. richardwardi from the combined dataset as well as from each dataset independently (cytb or rDNA), the genera Trichophoromyia and Nyssomyia were recovered monophyletic with high confidence (PP > 0.9).
The genus Psychodopygus was found to be monophyletic and divided into six groups. The group guyanensis, represented here by 4 of its 7 species (Ps. corossoniensis, Ps. luisleoni, Ps. francoisleponti, and Ps. geniculatus) and the group davisi (Ps. davisi and Ps. claustrei) appeared monophyletic with high nodal support (Fig. 2). The rest of the group had lower posterior probabilities, from which the group panamensis was found to be paraphyletic with Ps. hirsutus, Ps. ayrozai, Ps. nocticolus, and Ps. llanosmartinsi placed in one clade that excluded the congener Ps. panamensis. Furthermore, the groups chagasi, arthuri, and bispinosus were represented by only one species (Ps. squamiventris squamiventris, Ps. lloydi and Ps. bispinosus, respectively) and their position was less supported.

Discussion
To our knowledge, there is only one morphological phylogeny study, apart from the studies by Galati [22], that focus on the subgenus Lutzomyia in New World sandflies [38]. However, a large number of molecular studies related to     Next, Cohnstaedt et al. [10] based its phylogeny on cytb marker and 10 species of Pintomyia (Pifanomyia) belonging to the groups verrucarum, serrana, and townsendi, produced a nonsurprising phylogeny considering the analyzed species were closely related ones. Psychodopygus was raised for the first time to the genus level by Forattini [20,21] which at that time, comprised the greatest number of the species now included in the genera Bichromomyia, Psychodopygus, Nyssomyia, Trichophoromyia and Psathyromyia, partim of Psychodopygina by Galati [22]. Our study is strongly in agreement with Galati's hypotheses (Fig. 3). However, Forattini considered Viannamyia as a  distinct genus, but Galati 1981's classification [24] based on the morphology and implantation level of the larvae antennae, included the Viannamyia subgenus in Psychodopygus. We were not able to test this hypothesis given the lack of samples from Viannamyia. On the other hand, Ready et al. [42] considered only Psychodopygus as a distinct genus from Lutzomyia, arguing that the particular spermathecae of the females (with imbricated triangular rings) would justify its differentiation. Our phylogenetic reconstructions strongly support the monophyly of the genus Psychodopygus sensu Ready et al. [42] and Galati [22] as shown by both the concatenated dataset (ribosomal and mitochondrial, Fig 2.) and by each marker (Supplementary figures). The groups (formerly called series in the literature), guyanensis and davisi are monophyletic except for the group panamensis. This paraphyly is caused by the position of the group davisi separating Ps. panamensis from the other members of the group. The genus Nyssomyia, considered monophyletic by Galati [23] appears paraphyletic considering the position of Ny. richardwardi grouped within the genus Trichophoromyia. If we exclude the latter species of Nyssomyia, it could be considered as a clade subdivided into two well-supported monophyletic groups: one clade including the related species Ny. trapidoi and Ny. ylephiletor and a second clade including Ny. yuilli pajoti, Ny. umbratilis and Ny. antunesi grouped with the closely related species Ny. whitmani, Ny. intermedia and Ny. neivai. Additionally, introgression has been described for Ny. whitmani and Ny. intermedia [31,33,34] and both male and female hybrids between Ny. intermedia and Ny. neivai were observed [26]. We confirm the phylogenetic position of Ny. richardwardi was not due to a sequencing or laboratory error because four specimens were processed and sequenced at different times obtaining 100% identical data each time. Following the morphological identification keys of Galati [23][24][25], Ny. richardwardi and Ny. shawi form a cohesive group within the genus Nyssomyia with some morphologic characters shared across the genera Nyssomyia and Trichophoromyia (Table 4). For example, both species present simple setae on all flagellomeres, and in the males, the internal spine of the gonostyle is implanted close to its base. These characteristics are present in Trichophoromyia but not in Nyssomyia. However, the absence of Newstead sensilla on the palpomere II and spermathecae with less than 20 rings are present in Nyssomyia but not in Trichophoromyia. Further, both species exhibit cerci longer than wide in females from Nyssomyia but different from those of Trichophoromyia. In the morphological analysis, the clade constituted by both genera Nyssomyia and Trichophoromyia share more synapomorphies with Psychodopygus than with Bichromomyia in Galati's cladogram (Fig. 1); however, in our phylogenetic tree (Fig. 2) the genus Bichromomyia appears as the sister group of Nyssomyia.
The genus Psathyromyia was found to be monophyletic as well as the subgenera Forattiniella and Xiphopsathyromyia. However, the position of Pa. dasymera as sister to all other Psathyromyia species, including the members of the groups shannoni and lanei, renders the genus Psathyromyia paraphyletic.
Unfortunately, it was not possible to include samples belonging to the genus Viannamyia in our analysis. As a consequence, the phylogenetic position of Viannamyia within the Psychodopygina subtribe remains unsolved. In the future, exploring the relationships between the New World Leishmania agents of tegumentary leishmaniasis belonging to the genera Leishmania, Viannia, and Mundinia [5] and their vectors will prove to be essential in understanding specific vector associations and disease ecology. We suggest expanding on these phylogenetic explorations to clarify the taxonomic assessment of these vectors, strongly related to human and animal health.
Acknowledgements. The authors wish to thank François Le Pont, María Lorena Mejía, and Moises Gualapuro for their valuable support in the fieldwork and the provision of some specimens processed in the present work. This work was supported by Secretaria Nacional de Ciencia y Tecnología (SENESCYT) and Spanish collaboration funded by the Red de Investigación

Supplementary material
The supplementary material of this article is available at https://www.parasite-journal.org/10.1051/parasite/2023018/ olm. Figure S1: Phylogenetic reconstruction of the Psychodopygina subtribe from Bayesian analysis based on sequences of the D1 and D2 domains of 28SrDNA, the numbers in the nodes representing posterior probabilities. Figure S2: Phylogenetic reconstruction of the Psychodopygina subtribe from Bayesian analysis based on sequences of the cytochrome b gene of mtDNA, the numbers in the nodes representing posterior probabilities.