A new species of Hirudo (Annelida: Hirudinidae): historical biogeography of Eurasian medicinal leeches

Background: Species of Hirudo are used extensively for medicinal purposes, but are currently listed as endangered due to population declines from economic utilization and environmental pollution. In total, five species of Hirudo are currently described throughout Eurasia, with Turkey being one of the major exporters of medicinal leech, primarily H. verbana. Results: To define the distribution of Hirudo spp. within Turkey, we collected 18 individuals from six populations throughout the country. Morphological characters were scored after dorsal and ventral dissections, and Maximum Likelihood (ML) and Bayesian Inference (BI) analyses resolved phylogenetic relationships using mitochondrial cytochrome c oxidase subunit I (COI), 12S ribosomal RNA (rRNA), and nuclear 18S rRNA gene fragments. Our results identify a new species of medicinal leech, Hirudo sulukii n. sp, in Kara Lake of Adiyaman, Sülüklü Lake of Gaziantep and Segirkan wetland of Batman in Turkey. Phylogenetic divergence (e.g., 10–14 % at COI), its relatively small size, unique dorsal and ventral pigmentation patterns, and internal anatomy (e.g., small and pointed atrium, medium-sized epididymis, relatively long tubular and arc formed vagina) distinguish H. sulukii n. sp. from previously described Hirudo sp. Conclusions: By ML and BI analyses, H. sulukii n. sp. forms a basal evolutionary branch of Eurasian medicinal leeches. Phylogeographic interpretations of the genus identify a European Hirudo “explosion” during the upper Miocene followed by geological events (e.g., Zanclean flood, mountain building) that likely contributed to range restrictions and regional speciation of extant members of the clade.


Background
Hirudinid leeches are parasitic to a variety of vertebrates leading many to regard them with distaste, but their medicinal utility is well established. For centuries, Hirudo medicinalis and related species (e.g., H. verbana, H. troctina) were prescribed to treat virtually every human ailment from arthritis to yellow fever, most without efficacy. In 1830, during their peak usage, a Paris hospital employed more than five million medicinal leeches [30]. Consequently, populations of H. medicinalis in Central Europe were depleted, and non-sustainable collecting led to their extinction in many areas. Pollution and habitat drainage further added to their decline, forcing Europe to import medicinal leeches from the Ottoman Empire (Anatolia), North Africa and Russia [31] to meet demand. By the late 1900's, the advent of "modern" medicine drastically reduced clinical demand for leeches, allowing some threatened populations to rebound.
Leech therapy languished for most of the 20th century, considered "quackery" by mainstream medical practitioners [66], but the discovery of various bioactive compounds in leech saliva [27,39], and recognition of the leech's superior ability to relieve venous congestion (e.g., [58]), has led to renewed interest in clinical applications. Current fields of employment include reconstructive microsurgery, hypertension, and gangrene treatment [24]. In light of 19th century threats to medicinal leech populations as demand increased, considerable conservation steps were implemented to ensure their continued availability. Pursuant to these efforts, much confusion resulted regarding the taxonomic status of different morphological forms [18,28,56,65]. Phylogenetic analysis of nuclear and mitochondrial DNA sequences suggest that the genus Hirudo is monophyletic [60], and that species or morphological varieties can be readily identified by coloration patterns. Molecular studies have shown that European medicinal leeches, although usually marketed as H. medicinalis, comprise a complex of at least three species: H. orientalis, the commonly sold H. verbana and the relatively rare H. medicinalis [4,37,54,55,60]. Kutschera and Elliott [36] analyzed the behavior of adult H. medicinalis, but could not find differences with respect to its sister taxon H. verbana. Morphological and molecular data demonstrate that commercially available medicinal leeches are generally not H. medicinalis [35,56,60], but rather specimens belonging to the Eastern phylogroup H. verbana [61,62], which is predominantly bred in leech farms and used as a modern 'medicinal' stock.
Turkey is rich in wetlands and known to support at least two species of medicinal leech, H. medicinalis and H. verbana. Prior to~2000, it was believed that medicinal leeches from Turkey's wetlands were only H. medicinalis [21,31]. Molecular characterization of Turkish leeches was not performed until the turn of the century, however, and leeches from the Kızılırmak and Yesilirmak Deltas on the Black Sea coast, comprising the majority of leech specimens destined for export, have proven to be to H. verbana [4,51,55].
Mapped localities of all Hirudo species show extensive, belt-shaped ranges extending from east to west. To establish the distribution of Hirudo species in Turkey, one of the major exporters of medicinal leeches worldwide, we sampled broadly in three representative localities within the western, eastern and southeastern regions of Turkey. Our data identifies a new species for the genus, H. sulukii n. sp., that forms a basal evolutionary branch among European medicinal leeches and sheds light on the evolutionary history of the genus.

Specimen collection and maintenance
Leech specimens collected throughout Turkey (Kara Lake, Beyaz Cesme Marsh, Uluabat Lake, Segirkan wetland, Balik Lake, Sülüklü Lake) were transported to Fırat University, Fisheries Faculty (Elazig, Turkey) and maintained in separate 600 L fiberglass tanks based on collection location. Tank bottoms were elevated with peat soil 10 cm on one side to create a terrestrial to aquatic continuum. Leeches were fed one adult frog (e.g., Pelophylax ridibunda) blood meal per month (others have utilized mammalian blood), and typically survived 2+ years in the laboratory. Specimens were fixed in 70 % ethanol for molecular analysis and some were fixed with 10 % formaldehyde in PBS for dissection. External traits of live specimens were observed by stereomicroscopy. Preserved specimens were dissected dorsally and ventrally, with representative sketches of internal morphology derived directly from the type specimen.

DNA extraction
Tissue samples from live specimens were obtained by placing the leech in a 10 % ethanol sedating solution until it was unresponsive to touch. Approximately half of the caudal sucker was removed with a scalpel, and tissue cuttings were immediately processed using the E.Z.N.A.™ Tissue DNA kit (Omega Bio-Tek) following the manufacturer's instructions. Whenever possible, tissue from postmortem specimens was taken from the caudal sucker to avoid contamination from gut contents.

DNA sequence amplification of target genes
Nuclear 18S rRNA, mitochondrial 12S rRNA and partial cytochrome c oxidase subunit 1 (COI) DNA fragments were amplified from genomic DNA using the polymerase chain reaction (PCR). All 12S sequences were obtained under conditions described by Borda and Siddall [8]. PCR amplification protocols were conducted as described by Wirchansky and Shain [67] employing primers listed in Table 1. PCR products were purified using the Wizard SV Gel and PCR Clean-Up System kit (Promega, Inc.) according to the manufacturer's protocol.

DNA sequencing and editing
Purified PCR products were shipped to GeneWiz, Inc.
(South Plainfield, NJ) for Sanger DNA sequencing using an ABI 3730xl DNA analyzer. Each PCR product was sequenced in both directions using amplification primers, and sequence chromatograms were viewed and manually adjusted in ChromasPro (Technelysium, Queensland, Australia) or BioEdit [26]. Sequence alignments were made with MUSCLE [17] or CLUSTAL W [29,38]. Accession numbers for all CO1, 12S and 18S sequences are listed in Suppl. Data (Table 1).

Phylogeny
Maximum-likelihood (ML) analyses were performed for all DNA comparisons, using the pipeline sequence MEGA 7 [34] to align corresponding sequences from multiple individuals or homologous DNA across species, Gblocks [9] for alignment curation, PhyML [25] for tree building and TreeDyn [11] for tree drawing, as configured in the Phylogeny.fr platform [14]. The aLRT statistical test (approximation of the standard Likelihood Ratio Test; [3]) embedded in PhyML determined branch support values. Default settings were used for all parameters. Bayesian Inference (BI) analysis was performed on the combined data set (morphological parameters, 18S, 12S, COI in Nexus format) in MrBayes v. 3.2.1x64 [48,49]. Data were partitioned for 18S and 12S, and by codon position for COI. ModelTest [47] via FindModel was used to determine the optimal model of evolution for each gene under the Akaike Information Criterion (AIC; [46]). The general time reversible (GTR) model with a gamma distributed rate parameter was used for COI, 12S and 18S. Two analyses were run simultaneously with all parameter sets unlinked by partition for two million generations each, sampling every 100 generations, with a burn-in achieved by <50,000 generations. Setting the burn-in to 500,000 generations left a total of 7413 trees sampled for assessment of posterior probabilities. Gaps were treated as missing data, and default settings were used for all other parameters.

Results
Specimens of Hirudo were collected from multiple locations in Turkey ( Fig. 1; Tables 2 and 3). These localities are separated by 1312 km (Uluabat Lake to Kara Lake), 1306 km (Uluabat Lake to Beyaz Cesme Marsh) and 289 km (Kara Lake to Beyaz Cesme Marsh). Leeches were typically found in muddy bottoms, as well as underwater and in aquatic/terrestrial vegetation (typically reedbeds), with banks of water proving the most prevalent habitat.
Male reproductive apparatus notably large, with thick muscular penis sheath terminating in a bulbous prostate, located at ganglion in segment XI. Epididymis medium-sized, spherical, more than twice size of pearlescent-sheened ejaculatory bulb, tightly packed masses of ducting standing upright on either side of the atrium. Testisacs ovoid and larger than ovisacs, located posterior to ganglion in segment XIII. Female reproductive system relatively coiled tubing. The pearlescentsheened vagina long and upright, evenly bowed tube entering directly into ventral body wall. Oviducts a thin duct forming several coiled and covered with a thick layer of glandular tissue, bi-lobed ovaries. Ovisacs globular ovoid or small bean seed-shaped (Fig. 6).

Remarks
Despite similarities between Hirudo sulukii n. sp. and other Hirudo species, the former can be distinguished from its closest relatives using internal and external features. Hirudo sulukii n. sp. differs from H. medicinalis and H. orientalis by the form of black spots on the dorsal, paramedian stripes of the body. Hirudo sulukii n. sp. has black, segmentally-arranged united ellipsoid and elongated spots, and dorsal lateral margins of body a pair of zigzagged black dorsolateral longitudinal stripes (Fig. 4a). The ventral coloration pattern of H. sulukii n. sp. has a variable number of irregular spots (Fig. 4b) [65]. Hechtel and Sawyer [28] considered external pigmentation to be not only the most useful, but also arguably the best character to distinguish species of Hirudo.
In this study we used the approach of Hechtel and Sawyer [28] and Utevsky and Trontelj [65] regarding the size of the epididymis in relation to the ejaculatory duct. The epididymes of Hirudo sulukii n. sp. (Fig. 6) and H. orientalis are medium-sized. In contrast, the epididymes of H. verbana are relatively small, whereas H. troctina and H. medicinalis have massive epididymes [65]. The vagina of Hirudo sulukii n. sp. is relatively long tubular and arc formed (Fig. 6), while in H. orientalis the vagina is tubular and evenly curved. The former two species do not show the central swelling and sharp folding typical for H. verbana. In H. medicinalis, the vagina can have two conditions: straight and tubular, or terminally curved [65]. Hirudo troctina has a bulbous vagina [28].
Moquin-Tandon [40] described at least five species of Hirudo including H. verbana and H. medicinalis, but later concluded that they were all varieties of the same leech species. The medicinal leech, H. sulukii n. sp., considered here was determined to be morphologically different than all species described by Moquin-Tandor [40,41].

Phylogenetic analyses
To determine the relationship of specimens to other Hirudo species, we subjected them to the comparative analysis of CO1 (cytochrome c oxidase subunit 1) and 12S rRNA from mitochondria, and nuclear 18S rRNA. Combined COI, 12S and 18S rRNA analysis contained 13 terminals with 1514 aligned characters. Maximum Likelihood of the combined data set yielded five equally parsimonious trees with 500-1000 steps ( Fig. 7; Additional file 1); concordant trees were generated independently with COI data (Fig. 8; Additional file 1). Collectively, H. sulukii n. sp., formed a basal branch among European medicinal leeches with strong bootstrap support, while resolution among H. medicinalis, H. orientalis and H. verbana lineages was ambiguous, as noted in previous studies [45,56]. Population structure was shallow among the collected specimens (<2 % divergence at CO1; Table 4), suggesting recent invasions into field sites sampled in the current study (see Fig. 1). The Asian species, H. nipponia, fell outside the Hirudo clade in combined sequence analyses (Fig. 7), suggesting a deep ancestral split with European species, and calling into question the designation of H. nipponia within the Hirudo phylogroup. Interestingly, H. nipponia was equidistant to European Hirudo species (~22-25 % at CO1), the latter of which were approximately equidistant to each other (i.e.,~10-14 % at CO1; Table 4). Inferring a divergence rate of~2 % per million years at the CO1 locus based on combined geological and molecular data within Oligochaeta [10,15,67], we estimate a lower Miocene split between lineages leading to H. nipponia and European Hirudo sp., and radiation of the latter species during the upper Miocene. Branch patterns of remaining species were consistent with those reported previously [45].

Discussion
Maximum Likelihood and Baysian Inference analyses yielded trees with concordant topologies and strong support for H. sulukii as a basal branch of the European medicinal leeches. Relationships between H. medicinalis, H. verbana and H orientalis were less conclusive, consistent with confusion regarding their morphological identification [45,56]. The relatively small size of H. sulukii, unique dorsal and ventral pigmentation patterns, and internal anatomy (e.g., small and pointed atrium, medium-sized epididymis, relatively long tubular and arc formed vagina) are distinguishing features of this previously undescribed leech. Note that H. sulukii has thus far been collected only from relatively high elevation field sites (i.e., Kara Lake-Adiyaman 1233 m, Sülüklü Lake-Gaziantep 877 m, and Segirkan wetland-Batman 525 m), and its small size in comparison with other Hirudo species may reflect an adaptation to this environment (e.g., reduced foraging season/food supply), as suggested for other annelid species (e.g., [15]). Previously, only two medicinal leeches were thought to occur in Turkey, H. verbana and H. medicinalis, while a total of five are currently described throughout Eurasia. The range of H. verbana occurs to the south of H. medicinalis in an almost parapatric fashion with little overlap [5,32,42,43,51]. The former is subdivided into  Tables 2 and 3 for  specimen descriptions an Eastern (southern Ukraine, North Caucasus, Turkey and Uzbekistan) and Western phylogroup (Balkans and Italy) that do not overlap, suggesting distinct postglacial colonization from separate refugia [61,64]. Easternmost records are from Samarqand Province in Uzbekistan [61,64,65], resulting in an east-to-west extent of~4600 km. Leeches supplied by commercial facilities belong to the Eastern phylogroup, originating mostly from Turkey and the Krasnodar Territory in Russia, two leading areas of leech export.
Hirudo medicinalis is distributed from Britain and southern Norway to the southern Urals and probably as far as the Altai Mountains, occupying the deciduous arboreal zone [6,12,16,21,22,31,43,51,52,59,63,68]. Hirudo orientalis is associated with mountainous areas in the sub-boreal eremial zone and occurs in Transcaucasian countries, Iran and Central Asia, while H. troctina has been found in northwestern Africa and Spain in the Mediterranean zone [64]. Hirudo verbana and H. medicinalis have recently experienced range expansions while H. orientalis has remained geographically isolated within arid and alpine areas of Central Asia and Transcaucasia [61].
By molecular clock inference using divergence estimates at the CO1 locus [10,15,67], our data suggest a deep, ancestral split between European and Asian (i.e., the lineage leading to H. nipponia) medicinal leeches somewhere in the lower Miocene, followed by an "explosion" of Hirudo species upon their putative arrival to the European continent during the upper Miocene, 5-10 mya (Fig. 9). The possible misclassification of H. nipponia does not affect this evolutionary scenario since it represents a basal, sister branch to the European Hirudo phylogroup (see Fig. 8). This evolutionary timeline is supported by tree topologies and relative genetic distance of European Hirudo species to each other at the COI locus (i.e., 10-14 % divergence; see Table 4). The time frame of these events suggest the presence of an open habitat corresponding with, for example, formation of Levantine land bridges, which may have facilitated mammalian-based, passive dispersal of an ancestral Hirudo archetype throughout Europe. Thereafter, tectonic activity at the onset of the Pliocene~5.3 mya broke the land bridge between Morocco and Spain causing the Zanclean Flood that filled the Mediterranean basin, and in combination with mountain building throughout the European continent [7], appears to have restricted panmixia among extant Hirudo lineages, leading in part to their speciation and current geographic ranges. For instance, concurrent with the closing of the Tethys Sea by continental drift of the African and Arabian plates, mountain building events occurred in Southern Turkey forming the Taurus Mountain chain [13]. At present, the Species of Hirudo have had broad applications in medicine, ranging from reconstructive surgeries (e.g., facial, finger reattachment, ear flap) to anticoagulants/ analgesics secreted from salivary glands [2,24]. Thus the discovery of a new Hirudo species, particularly a basal member of this phylogroup, has considerable value in the context of medical potential. Specifically, natural variants of known bioactive factors (e.g., hirudin, antistasin, etc.) are logical candidates to explore for their potentially enhanced or novel pharmaceutical properties. The current study has prompted a more systematic survey of Hirudo throughout Turkey and surrounding regions with the collective aims of refining the evolutionary history of the genus, facilitating conservation efforts, and identifying species that may expand the repertoire of medicinal applications for this important Hirudinid genus.

Conclusions
By phylogenetic and morphological criteria, specimens collected from Kara Lake of Adiyaman, Sülüklü Lake of Gaziantep and Segirkan wetland of Batman in Turkey comprise a new species, Hirudo sulukii. Geographic isolation by the Taurus Mountain chain has likely contained H. sulukii within the regional sampling area. By ML and BI analyses, H. sulukii n. sp. forms a basal evolutionary branch of Eurasian medicinal leeches, preceded by a deeper ancestral split with the Asian medicinal leech. H. nipponia. Phylogeographic interpretations of the genus identify a European Hirudo "explosion" during the upper Miocene followed by geological events (e.g., Zanclean flood, mountain building) that likely contributed to range restrictions and regional speciation of extant members of the European clade.