Molecular Phylogenetics and Evolution 34 (2005) 616–624

www.elsevier.com/locate/ympev

Celebrity with a neglected taxonomy: molecular systematics of the medicinal leech (genus Hirudo)

Peter Trontelja,¤, Serge Y. Utevskyb

a Biotechnical Faculty, Department of Biology, University of Ljubljana, P.O. box 2995, SI-1001 Ljubljana, Sloveniab Department of Zoology and Animal Ecology, V.N. Karazin Kharkiv National University, Pl. Svobody 4, Kharkiv 61077, Ukraine

Received 2 August 2004; revised 29 September 2004Available online 15 December 2004

Abstract

The medicinal leech is the most famous representative of the Hirudinea. It is one of few invertebrates widely used in medicine andas a scientiWc model object. It has recently been given considerable conservation eVort. Despite all attention there is confusionregarding the taxonomic status of diVerent morphological forms, with many diVerent species described in the past, but only two gen-erally accepted at present. The results of the phylogenetic analysis of a nuclear (ITS2 + 5.8S rRNA) and two mitochondrial genesequences (12S rRNA, COI) suggest that the genus Hirudo is monophyletic. It consists, apart form the type Hirudo medicinalis andthe East Asian Hirudo nipponia, of three other, neglected species. All of them have already been described either as species or mor-phological variety, and can readily be identiWed by their coloration pattern. The type species is in weakly supported sister relationwith Hirudo sp. n. (described as variety orientalis) from Transcaucasia and Iran. Sister to them stands Hirudo verbana from south-eastern Europe and Turkey, which is nowadays predominantly bred in leech farms and used as ‘medicinal leech.’ The North AfricanHirudo troctina is the sister taxon to this group of Western Eurasian species, whereas the basal split is between H. nipponia and theWestern Palaearctic clade. 2004 Elsevier Inc. All rights reserved.

Keywords: Hirudo medicinalis; Hirudo nipponia; Hirudo troctina; Hirudo verbana; Hirudo orientalis; Phylogeny; Taxonomy

1. Introduction ing and medical applications. Ancient medicine consid-

The medicinal leech (Hirudo medicinalis Linnaeus,1758) is the most prominent representative of the classHirudinea, and one of the most renowned invertebrates,infamous for its ectoparasitic bloodsucking. It has beenwidely applied in neurophysiological and developmentalgenetic studies. Furthermore, the medicinal leech is aclassic laboratory object used for educational purposes.It may be regarded as one of the best-studied animals inrespect of its morphology, physiology, and behavior(Sawyer, 1986). The medicinal leech owes its emotionaland cultural connotations to its haematophagous feed-

* Corresponding author. Fax: +386 1 2573390.E-mail address: peter.trontelj@bf.uni-lj.si (P. Trontelj).

1055-7903/$ - see front matter 2004 Elsevier Inc. All rights reserved.doi:10.1016/j.ympev.2004.10.012

ered the leech as a panacea, Wrst mentioned in the 2ndcentury by the physician Galenus. The exploitation ofleeches reached its peak in the Wrst half of the 19th cen-tury. For example, 70 million leeches a year were caughtin Russia (Voskresensky, 1859), and incredible 100 mil-lion leeches were imported to France in 1850 (Herter,1937). The leeches were collected mainly in South Euro-pean countries and transported over a great distance(Lukin, 1976). In addition to its historical fame, themedicinal leech has experienced a recent renaissance inthe medical use of leeches both as a source of pharma-ceutical substances and directly, for example in plasticsurgery or dermatology (e.g., Sohn et al., 2001; Whitakeret al., 2004a,b).

Leech collecting for medicinal purposes in the 19thcentury, recent collecting pressure, and a general loss

P. Trontelj, S.Y. Utevsky / Molecular Phylogenetics and Evolution 34 (2005) 616–624 617

and pollution of wetland habitats have caused a dra-matic decline of H. medicinalis throughout its geograph-ical range (Elliott and Tullett, 1984; Sawyer, 1981).Particularly large numbers of leeches are still being har-vested in some Southeastern European countries andTurkey (Wells and Coombes, 1987). The internationalconcern for its conservation is reXected by internationalconventions and regulations. It has been included in theIUCN Invertebrate Red Data Book (Wells et al., 1983),added to Appendix II of the Convention on Interna-tional Trade in Endangered Species of Wild Fauna andFlora (CITES) in 1987, to the appendices of the Conven-tion on the Conservation of European Wildlife and Nat-ural Habitats (Berne Convention) in 1998, and listed onAnnex V of the EU Council directive 92/43/EEC on theconservation of natural habitats and of wild fauna andXora in 1992.

Despite the high conservation concern, medical atten-tion, and biological interest devoted to the medicinalleech, there is confusion regarding the taxonomic statusof diVerent morphological forms. Its taxonomic historywas much livelier than its monotonous present. In the19th century a number of species of Hirudo weredescribed on the basis of their external coloration.Moquin-Tandon (1827) summarized the taxonomicdiversity of the genus recognizing seven species of San-guisuga ( D Hirudo). Later on he changed his mind andconcluded that they are all varieties of the same species,H. medicinalis, with the single exception of the NorthAfrican ‘dragon leech’ Hirudo troctina Johnson, 1816;which he regarded as a separate species (Moquin-Tan-don, 1846). This view persisted until recently, whenNesemann and Neubert (1999) re-established the speciesstatus of the long forgotten south European taxonH. verbana Carena, 1820, but questioned the taxonomicdistinctiveness of H. troctina. Conversely, Hechtel andSawyer (2002) corroborate the distinction between theEuropean H. medicinalis and the North African H. troc-tina, sharing Moquin-Tandon’s (1846) opinion that allvarieties of medicinal leech in Europe represent the samespecies. The Russian leech literature recognizes threemorphological ‘forms’ of the medicinal leech (Lukin,1976; Shevkunova and Kristman, 1962; Stschegolew andFedorova, 1955; Utevsky et al., 1998): f. oYcinalis and f.serpentina, which correspond to the taxa H. verbana andH. medicinalis, respectively, and a third form fromTranscaucasia and Iran known as the ‘Persian medicinalleech’, Hirudo medicinalis orientalis. The taxonomic sta-tus of this leech has been uncertain.

Whitman (1886) added to this list Hirudo nipponiafrom Japan.

Faced with such a confusing taxonomy based entirelyon the pattern of coloration, one is inclined to acceptSawyer’s (1986) view that pigmentation variability issimply a general characteristic of H. medicinalis, the pig-mentation pattern corresponding to a considerable

extent to the geographical origin of the populations. Thepresent study is aimed at clarifying the status of the men-tioned morphological varieties of the medicinal leechand their relationships. It is based on DNA sequences ofa protein coding mitochondrial gene (COI), the mito-chondrial 12S rRNA gene, and a variable part of thenuclear ribosomal gene cluster.

2. Materials and methods

2.1. Samples

The present study is based on 18 medicinal leech spec-imens belonging to Wve taxa mentioned above. Theywere identiWed according to their dorsal and ventral col-oration pattern. Leeches were photographed from bothsides under an Olympus SZX9 stereo microscope withan Olympus DP10 digital camera. They were Wxed andstored in 96% ethanol at ¡20 °C until DNA extraction.Locality data and accession numbers are given in Table 1.The taxa medicinalis, orientalis, and verbana were col-lected from diVerent parts of their range. The taxa troc-tina and nipponia were obtained from captivity. Othersequences of H. medicinalis and its relatives from Gen-Bank were included to conWrm the monophyly of thegenus.

2.2. DNA extraction, ampliWcation, and sequencing

Small pieces (app. 5 £ 2 £ 2 mm) of skin and muscletissue were cut of the lateral part of the body. Care wastaken not to reach the digestive system which oftencontained blood from unknown host species. GenomicDNA was isolated using the GeneElute MammalianGenomic DNA minprep kit from Sigma–Aldrich (Stein-heim, Germany). The ribosomal internal transcribedspacer (ITS2) including a major part of the 5.8S rRNAgene was ampliWed using the primers ‘ITS3’ and ‘ITS4’described by White et al. (1990). PCR was performed byapplying 30 cycles of 45 s at 94 °C, 45 s at 48 °C, and 60 sat 72 °C, following a 3 min denaturation step at 94 °C.For PCR ampliWcation of the mitochondrial 12S rRNAgene the primers 5�-GCCAGCAGCCGCGGTTA-3�and 5�-CCTACTTTGTTACGACTTAT-3� weredesigned by comparing available invertebrate 12S rDNAsequences. Cycling conditions were as for the ITS2,except that annealing was performed at 46 °C. The COIgene fragment was ampliWed using primers LCO 1490and HCO 2198 from Folmer et al. (1994) by applying 34cycles of 45 s at 94 °C, 45 s at 48 °C, and 1 min at 72 °C,after an initial 3 min denaturation step at 94 °C. All PCRproducts were puriWed by elution from a 1.5% agarosegel, precipitated with 2.5 vol. of 100% ethanol, andwashed with 70% ethanol. They were sequenced usingCy5 labeled primers (with the same sequences as ampliW-

618 P. Trontelj, S.Y. Utevsky / Molecular Phylogenetics and Evolution 34 (2005) 616–624

cation primers), and the Thermo Sequenase PrimerCycle Sequencing Kit on an ALF Express II automatedsequencer (both from Amersham–Pharmacia Biotech,Uppsala, Sweden).

2.3. Sequence alignment

The length of the obtained COI fragment sequenceswas about 630 bp. They were unambiguously aligned byhand. The correctness of the alignment was veriWed atthe amino acid level. ITS2 and 12S rDNA sequencesvaried considerably in length, especially when outgroupspecies were included (502–564 and 482–505 bp, respec-tively). A preliminary CustalX (Thompson et al., 1997)alignment with default parameters revealed potentiallyambiguously aligned positions in both genes. A sensitiv-ity approach suggested by Gatesy et al. (1993) andimplemented in SOAP (Löytynoja and Milinkovitch,2001) was applied to identify positions that are sensitiveto modiWcations of Clustal gap opening and gap exten-sion penalties. Gap penalties ranging from 7 to 19 for

opening, and 3–9 for extension, both by steps of two,were explored. It turned out that the blocks of ambigu-ous alignment, even if only the 50% conservation crite-rion was applied, contained an important part ofphylogenetically informative positions. Eliminatingthem lead to poorly resolved trees. We therefore com-pared the preliminary trees obtained from diVerentalignments without excluded sites, for possible conXict intopologies. All 28 compared trees showed identicaltopologies, but they diVered in node support. Since therewas no conXict in phylogenetic signal among diVerentalignments, we decided not to exclude any positions. Wechose the alignment leading to the highest overall nodesupport as an indication of low character conXict. Basicproperties of the aligned sequences are shown in Table 2.

2.4. Phylogenetic analyses

The ITS2, 12S rDNA, and the COI sequence align-ment were phylogenetically analyzed as a combined, sin-gle dataset. Before combining the alignments, a partition

Table 1Leech taxa, their geographic origin, and DNA sequences included in the analysis

* The ISO country code: AZ, Azerbaijan; DE, Germany; HR, Croatia; IT, Italy; KR, Korea; MA, Morocco; MK, Former Yugoslav Republic ofMacedonia; NP, Nepal; SI, Slovenia; and UA, Ukraine.

Taxon Locality* GenBank Accession No.

ITS2 12S rDNA COI

Hirudo medicinalis Linnaeus, 1758 Podvinci pond, Ptuj, SI 1 AY763166 AY763159 AY763149Petajnci gravel pit; Murska Sobota, SI 2 AY763166 AY763158 AY763149Saale River, Rotherburg, Halle, DE (2 ex.) AY763166 AY763157 AY763148Pond in Gorelaya Dolina, Kharkiv Region, UA AY763166 AY763156 AY763148Pond near Vasyshcheve, Kharkiv Region, UA AY763166 AY763156 —

Hirudo sp. n. (orientalis variety sensu Shevkunova and Kristman, 1962)

Agdam District, AZ1Unknown locality in Azerbaijan, AZ2

AY763170 AY763163 AY763154AY763170 AY768704 AY763154

Hirudo verbana Carena, 1820 Pond Globobaj near Povir, Seqana, SI AY763167 AY763160 AY763151Veliko Panko blato, Island Pag, HR AY763167 AY763160 —Lecce, IT AY763167 AY763160 AY763150Piethen kaolin pit, Halle, DE AY763167 AY763160 —Lake Ohrid, Ohrid, MK (2 ex.) AY763167 AY763160 AY763150

Hirudo troctina Johnson, 1816 Bazaar in Marrakech, MA (2 ex.) AY763171 AY763164 AY763155Hirudo nipponia Whitman, 1886 Leech farm Hans Biomed, KR (2 ex.) AY763169 AY763162 AY763153Haemopis sanguisuga (Linnaeus, 1758) Lake Cerknica, Cerknica, SI AY763173 — —

River Pivka, Postojna, SI — AF099960n/a — — AF462021

Limnatis cf. nilotica Konavlje, Dubrovnik, HR AY763168 AY763161 AY763152Haemadipsa cf. zeylanica Chitwan, NP AY763172 AY763165 —Haemadipsa sylvestris Blanchard, 1894 n/a — — AF003266

Table 2Basic properties of the aligned Hirudo DNA sequences used in this study

OP/EP, gap opening/extension penalty applied to obtain alignment; # var, number of variable sites; # inf, number of parsimony informative sites.

Sequence alignment Length OP/EP # var # inf % A % C % G % T

ITS2 517 13/7 56 11 18 27 30 2512S rDNA 498 13/7 137 45 38 11 15 36COI all positions 631 — 167 88 30 16 15 39

1st position — — 39 14 29 16 25 302nd position — — 6 2 14 26 16 443rd position — — 122 72 45 6 5 44

ITS2 + 12S + COI 1646 — 360 144 28 18 20 34

P. Trontelj, S.Y. Utevsky / Molecular Phylogenetics and Evolution 34 (2005) 616–624 619

homogeneity test (Farris et al., 1995) as implemented inPAUP* (SwoVord, 2003) was performed to examinewhether they diVered signiWcantly in their phylogeneticsignal. Partitions were tested against one another, alltogether, and mitochondrial against nuclear, by branch-and-bound searches applied to 1000 replicates of ran-domly partitioned characters.

Phylogenetic reconstruction was carried out usingmaximum likelihood, Bayesian inference, and parsi-mony. A model of sequence evolution for the use withmaximum likelihood and Bayesian inference wasselected via hierarchical likelihood testing to test alterna-tive models of evolution, employing PAUP* and Model-test (Posada and Crandall, 1998). The preferred model ofnucleotide substitution was the ‘transversion model’with gamma distributed rate heterogeneity. The likeli-hood-estimated substitution rates were R(A–C) D 0.7801,R(A–G) D 4.4853, R(A–T) D 2.7268, R(C–G) D 0.8419, R(C–T) D4.4853, and R(G–T) D 1.0000. The base frequencies of thecombined data set were estimated at 0.2806 (A), 0.1884(C), 0.1856 (G), and 0.3453 (T). The estimated gammadistribution shape parameter � was 0.5175. Maximumlikelihood searches were performed using Win-PAUP*4b10 by applying 100 random addition cycles with TBRbranch swapping.

For Bayesian inference, the program MrBayes 3.0b4(Huelsenbeck and Ronquist, 2001) was used. The chosensubstitution model was speciWed by setting six substitu-tion parameters and a gamma distributed rate. A Mar-kov chain–Monte Carlo search was run with four chainsfor 2 £ 106 generations, taking samples every 100 genera-tions. After determining the burn-in at 20,000–30,000generations by plotting the likelihood of the trees, theWrst 500 trees were discarded. From the resulting 19,500trees, one thousand trees were picked at random, and aposteriori probabilities for individual clades wereassessed based on their observed frequencies.

Parsimony searches were performed with Win-PAUP*4b10 using the branch-and-bound search option. To takeinto consideration the estimated transition bias andpotential saturation of the third codon position of theCOI gene, extra searches were run with a twofold weightassigned to transversions, the 3rd positions of the COIgene excluded, and both modiWcations combined.

Nonparametric bootstrapping (1000 replicates) wasused to obtain relative measures of the phylogenetic sig-nal supporting individual nodes in the maximum likeli-hood and parsimony analyses.

Within the likelihood framework, we tested the statis-tical support for the recovered clades employing theapproximately unbiased (AU) test proposed by Shimo-daira (2002). We tested the most likely tree against thetrees obtained by constraining the clades of interest to benon-monophyletic using the appropriate options inPAUP*. Each clade was tested separately, using the pro-gram CONSEL (Shimodaira and Hasegawa, 2001).

3. Results

All reported results are based on total molecular evi-dence resulting from the concatenation of one nuclearand two mitochondrial data sets (Table 2). Thisapproach seemed justiWed by the results of the partitionhomogeneity test performed for diVerent character parti-tions. Tests of nuclear vs. mitochondrial, and ITS2 vs.COI alignments revealed high congruence (P D 0.789and 1.000, respectively). Comparisons of ITS2 vs. 12S,12S vs. COI, and of all three datasets at once, revealedless homogeneity, but still failed to reject congruenceamong them (P D 0.110, 0.127, and 0.209, respectively).

The perhaps most important result of this study is theWnding that four diVerent groups, tentatively assigned todiVerent coloration types of the medicinal leech (Fig. 1),were recovered as distinct lineages. That there can be lit-tle doubt about their species status is demonstrated bythe diVerence in genetic variation between and withinlineages. The average uncorrected pairwise sequencedivergence within coloration types represented from twoor more localities (medicinalis, verbana, and orientalis)was 0.0011 § 0.0007, whereas the mean distance betweenthem was 0.0528 § 0.0061. The orientalis coloration typelacks a proper taxonomic description. For the purposeof this paper we therefore refer to it as Hirudo sp. n. TheNorth African taxon H. troctina merits a distinct taxo-nomic status even more evidently, since it is the sistertaxon of the former three species (Figs. 1 and 2). Tocompare the level of divergence between Hirudo specieswith interspeciWc distances within other leech genera, weemployed some published COI sequences (Borda andSiddall, 2004; Utevsky and Trontelj, 2004). For example,uncorrected COI distances between the four Hirudo spe-cies (0.086 § 0.008) are higher than between species dis-tances of the Wsh leech genus Piscicola (0.041 § 0.006),and lower than distances between Haemadipsa species(0.153 § 0.011).

Parsimony (Fig. 1), likelihood, and Bayesian (Fig. 2)searches resulted in trees with identical topologies. Themonophyly of the medicinal leech genus was stronglysupported by non-parametric bootstrapping, and sup-ported at P t 0.1 by Shimodaira’s (2002) approximatelyunbiased test. It is worth mentioning that Lee (2000)found that these kinds of clade signiWcance tests arerather severe, and that clades with P 6 0.1 are usuallystrongly corroborated. The Wrst, well supported andanticipated division within this clade was the splitbetween the East Asian species H. nipponia and thegroup of Western Palaearctic taxa. The branching orderwithin this group was less strongly supported, but it stillmakes sense from a biogeographic perspective, separat-ing the North African H. troctina from theEuropean + Near East group. For the latter group, reli-able distribution data are still too scarce to allow con-structing a precise biogeographic picture.

620 P. Trontelj, S.Y. Utevsky / Molecular Phylogenetics and Evolution 34 (2005) 616–624

Applying a broader taxonomic sampling to the COIdata set did not change the ingroup relationships, orinterfere with its monophyly (Fig. 3). Moreover, fromonly about one third of the base pairs of the full com-bined dataset the same relationships were deduced, whichcan be viewed as an indication of high character congru-ence. The phylogenetic analysis of hirudiniform COI genesequences from GenBank showed that both sequencesfound under the name H. medicinalis [U74067 (Black etal., 1997), AF003272 (Siddall and Burreson, 1998)] aretaxonomically distinct from the Hirudo sequencesobtained in the present study (Fig. 3). It is interesting tonote that the divergence between Limnatis nilotica(AY425452) and our Limnatis cf. nilotica sequence iswithin the range of divergence between diVerent Hirudospecies. The high divergence between these two sequences

might suggest that, like in Hirudo, more than one speciescould be hiding under the same name.

4. Discussion

4.1. Taxonomic considerations

Organisms that are used as model in various Welds ofscience, or that are commercially exploited and bred, areoften taxonomically very clearly deWned. In this respectthe medicinal leech is quite unlike the fruit Xy, or thehouse mouse, where the growth of biological knowledgewent hand in hand with an increase in systematiceVort. Previous taxonomic work on the leech, mainlyfrom the early times of Linnean classiWcation, has

Fig. 1. Most parsimonious phylogenetic hypothesis of medicinal leeches deduced from combined nuclear and mitochondrial DNA sequences [treelength D 1223 steps, homoplasy index (HI) D 0.172, retention index (RI) D 0.553, rescaled consistency index (RC) D 0.458]. Typical, diagnostic dorsal(left) and ventral (right) patterns of the midbody region are shown for the species that were previously lumped under the name ‘H. medicinalis.’ Twovariants of the dorsal pattern are shown for Hirudo sp. n. (orientalis coloration type). Numbers on branches correspond to bootstrap values (in per-cent) resulting from four diVerent weightings: unweighted/twofold weighted transversions/third codon positions of COI gene sequences excluded/twofold weighted transversions plus third codon positions excluded. Specimens are labeled as in Table 1.

P. Trontelj, S.Y. Utevsky / Molecular Phylogenetics and Evolution 34 (2005) 616–624 621

were obtained from GenBank. Familial aYliation of species is abbreviated

s follows: He, Haemopidae; Hi, Hirudinidae; Hd, Haemadipsidae.

Fig. 2. Maximum likelihood tree of combined medicinal leech nuclear and mitochondrial DNA sequences. The model and parameters are describedin Section 2. Numbers on branches are, respectively, maximum likelihood bootstrap values, and Bayesian posterior probabilities. Boxes contain‘approximately unbiased test’ (Shimodaira, 2002) probability values for the monophyly of clades. Specimens are labeled as in Table 1.

Fig. 3. Bayesian 50% majority rule tree of medicinal leeches along with a wider assortment of hirudiniform species, based on mitochondrial COI genesequences. Numbers on branches are Bayesian posterior probabilities. Species names followed by GB accession numbers indicate that sequence data

a

622 P. Trontelj, S.Y. Utevsky / Molecular Phylogenetics and Evolution 34 (2005) 616–624

been overlooked or rejected by contemporary authors.Consequently, the molecular phylogenetic scrutiny didnot bring dramatic taxonomic changes and entirely newspecies. It merely helped to recognize some previouslydescribed taxa as distinct phyletic lineages. The pictureobtained looks so clear that one might wonder why theyall became synonymized with H. medicinalis. The asser-tion that all diVerent morphological forms or colorationtypes represent just variations of one and the same spe-cies (e.g., Lukin, 1976) cannot be justiWed by the resultsof the present molecular phylogenetic analyses. On thecontrary, there is a deep gap between the two levels ofvariation. Genetic variation between species exceeds thevariation within on average by a factor of 50 (see Section3). It should be noted, however, that this gap may even-tually become shallower when more individuals fromgeographically distant populations are analyzed. Thismay not be very likely for H. medicinalis and H. verbana,which have been sampled over a large part of theirranges, but the picture might change substantially whenother populations of H. troctina, Hirudo sp. n. (orientaliscoloration type), and H. nipponia are sampled.

Even though the coloration pattern of the individualleeches varied within a species (cf. variation of the dorsalpattern in Hirudo sp. n. on Fig. 1), it was always possibleto unequivocally call the species among several hundredsof examined specimens. Complete description and diVer-ential diagnosis for H. medicinalis and H. verbana can befound in Nesemann and Neubert (1999). Their status asdistinct biological species has been recently corroboratedin a study using random ampliWed polymorphic DNAvariation (Trontelj et al., 2004). A detailed description ofH. troctina was published by Hechtel and Sawyer (2002).However, their diagnosis distinguishing it from H. medic-inalis should be read with caution, since the authors actu-ally compare H. troctina with H. verbana from Turkey.Similarly, some new data on the biology and distributionof “H. medicinalis” in Turkey (Demirsoy et al., 2001;Kasparek et al., 2000) probably pertain to H. verbana.

The Wfth species of the genus has never been recog-nized as such. It can most readily be identiWed by thebright green coloration of the dorsum, segmentallyarranged pairs of black quadrangular dots on itsparamedian dorsal stripes, and similarly arranged, butless regular light-colored markings on the predomi-nantly black venter. Stschegolew and Fedorova (1955)treated it as a variety of the form serpentina ( D H. medic-inalis). Among practitioners it was known as the Persianor Georgian leech, also mentioned by Lukin (1976).Shevkunova and Kristman (1962) Wrst applied the trino-mial H. m. orientalis for this supposed variety. However,article 15.2. of the International Code of ZoologicalNomenclature (International Commission on Zoologi-cal Nomenclature, 1999) speciWcally treats names ofvarieties or forms published after 1960 as infrasubspe-ciWc, and as such as not regulated by the Code. Thus, the

name ‘orientalis’ cannot be applied for the new speciesbecause it would not meet the Code’s criteria of avail-ability. We will formally describe Hirudo sp. n., corre-sponding to the variety H. m. orientalis sensuShevkunova and Kristman (1962), in a forthcoming tax-onomic paper.

4.2. Phylogenetic and biogeographic relationships

The genus Hirudo emerged as monophylum, both inthe analysis of the full, concatenated dataset, and withCOI sequences alone but with an extended taxonomicsampling. Conversely, Borda and Siddall (2004) foundthe genus to be paraphyletic. This discrepancy couldresult from the diVerent COI sequence used for H.medicinalis (see Section 3 and Fig. 3). It has been pointedout to us during the reviewing process that the source forboth H. medicinalis COI sequences found in GenBankwas the same—Carolina Biological Supply—and thatthe animals looked like H. medicinalis. The puzzle couldbest be solved by re-examining the specimens fromwhich the mysterious COI sequences were obtained.

In the present study, the relatively high support formonophyly is at the same time support for the sisterrelation between the Western Palaearctic clade and theEast Asian H. nipponia. One might wonder why theradiation in the West was not paralleled by at leastsome diversiWcation in the East. The answer to thisquestion possibly lies in the sampling imbalancebetween the eastern and the western part of the genus’range. Traditionally, H. nipponia has not been as widelyused as European medicinal leeches, nor has it beengiven as much taxonomic attention. It is possible that“H. nipponia” is just a taxonomic oversimpliWcation asit turned out to be the case with H. medicinalis. In orderto explore this possibility, more geographically sepa-rated populations from the extensive range of H. nippo-nia need to be sampled.

Somewhat paradoxical is the striking diVerencebetween the relatively deep divergence between, and thegenetic uniformity within species. Even in H. medicinalisand H. verbana that were sampled from localities morethan 2000 km apart and spanning a large part of theirknown range, genetic distance does not exceed 0.005substitutions per site. H. verbana populations fromdiVerent potential glacial refugia in Southern Italy andon the Southern Balkans (ancient Lake Ohrid) haveidentical DNA sequences. This might be an indication ofa rapid postglacial dispersion from genetically impover-ished refugial populations. The ability of medicinalleeches to stay attached to mobile mammal hosts for along time is in favor of such an explanation. Clearly,more populations and more individuals will have to beanalyzed, perhaps in combination with a more variablegenetic marker, to draw conclusions about phylogeogra-phy and demographic history.

P. Trontelj, S.Y. Utevsky / Molecular Phylogenetics and Evolution 34 (2005) 616–624 623

4.3. Implications for conservation

A new systematic understanding of a taxon of interna-tional conservation concern usually requires at least somere-consideration of conservation rules and priorities. Itwould be premature to say that all constitutive species of“H. medicinalis s. lato” merit the same conservationactions, since very little is known about their distribution,speciWc habitat, and other ecological requirements. Eventhe conservation status of the medicinal leech s. lato hasonly been established for a few countries (e.g., Elliott andTullett, 1984; Kasparek et al., 2000).

The distribution of diVerent taxa can be inferred indi-rectly by what has been said about the distribution of col-oration types or forms. According to the literature, H.medicinalis occurs in Western and Central Europe (Nese-mann and Neubert, 1999), the Ukraine (Lukin, 1976),and Lithuania (Zapkuvenè, 1972). H. verbana is knownfrom the Eastern Mediterranean region and the Balkans(Nesemann and Neubert, 1999), as well as Moldova,Ukraine, the Krasnodar Territory (Russia), and Armenia(Lukin, 1976; Utevsky et al., 1998). H. troctina is knownfrom Morocco, Tunisia, and Algeria (Hechtel and Saw-yer, 2002), whereas Hirudo sp. n. (orientalis colorationtype) was found in Georgia, Armenia, Azerbaijan, andIran (Lukin, 1976; Stschegolew and Fedorova, 1955;Utevsky et al., 1998). It is unknown to what degree thecurrent distribution has been inXuenced by range lossesand by introduction of animals to previously uninhabitedareas. The population structure has in some areas pre-sumably undergone substantial anthropogenic changes.

International conservation legislation (IUCN,CITES, Berne convention, EU Habitat Directive) dealsexclusively with H. medicinalis, carrying several poten-tial pitfalls. The perhaps greatest threat comes form thefact that most leeches imported form SoutheasternEurope and Turkey are H. verbana, as are the majorityof leeches in leech farms. Accidentally releasing theseanimals in Central or Western Europe, or even re-stock-ing natural habitats would mean the introduction of analien species. Nothing is known about the competitivebehavior between H. medicinalis and H. verbana, andreports on possible cross-breeding are ambiguous. Forthis and other reasons it is advisable to implement thetaxonomic changes to international conservation rulesas soon as possible. With this paper we hope to havecontributed to a wider appreciation of the speciWc subdi-vision of the medicinal leech, and therewith to a moreeVective conservation of each particular species.

Acknowledgments

We thank Matjaq Bedjanib (Slovenska Bistrica, Slove-nia), Dario Ferreri (Lecce, Italy), Cene Finer (Univ.Ljubljana), Clemens Grosser (Leipzig, Germany), Ke

Won Kang (Hans Biomed, Korea), Klemen Koselj(Eberhard-Karls Univ. Tübingen), and Vladimir Samoy-lov (Kharkiv, Ukraine) for furnishing and collectingleeches. We are further grateful to Joqica Murko-BuliT(Univ. Ljubljana) for her help in the laboratory, CeneFiner for photographing the animals, to Boris Sket(Univ. Ljubljana) for his guidance through leech taxon-omy, to Ljerka Lah for her valuable comments on lan-guage and style, and to two referees for theirconstructive comments on an earlier version of the man-uscript. This work was supported by the Slovenian Min-istry of Science, Education and Sport, and by the INTASGrant YSF No 01/2-0062.

References

Black, M.B., Halanych, K.M., Maas, P.A.Y., Hoeh, W.R., Hashimoto,J., Desbruyères, D., Lutz, R.A., Vrijenhoek, R.C., 1997. Molecularsystematics of vestimentiferan tubeworms from hydrothermal ventsand cold-water seeps. Mar. Biol. 130, 141–149.

Borda, E., Siddall, M.E., 2004. Arhynchobdellida (Annelida: Oligocha-eta: Hirudinida): phylogenetic relationships and evolution. Mol.Phylogenet. Evol. 30, 213–225.

Demirsoy, A., Kasparek, M., Akbulut, A., Durmuş, Y., Akbu-lut, N.E., Çalşkan, M., 2001. Phenology of the medicinalleech, Hirudo medicinalis L., in north-western Turkey. Hydrobiolo-gia 462, 19–24.

Elliott, J.M., Tullett, P.A., 1984. The status of the medicinal leech Hir-udo medicinalis in Europe and especially in the British Isles. Biol.Conserv. 29, 15–26.

Farris, J.S., Källersjö, M., Kluge, A.G., Bult, C., 1995. Testing signiW-cance of incongruence. Cladistics 10, 315–319.

Folmer, O.M, Black., M., Hoeh, R., Lutz, R., Vrijehoek, R., 1994. DNAprimers for ampliWcation of mitochondrial cytochrome c oxidasesubunit I from diverse metazoan invertebrates. Mol. Mar. Biol. Bio-technol. 5, 304–313.

Gatesy, J., DeSalle, R., Wheeler, W., 1993. Alignment-ambiguousnucleotide sites and the exclusion of systematic data. Mol. Phyloge-net. Evol. 2, 152–157.

Hechtel, F.O.P., Sawyer, R.T., 2002. Toward a taxonomic revision ofthe medicinal leech Hirudo medicinalis Linnaeus, 1758 (Hirudinea:Hirudinidae): re-description of Hirudo troctina Johnston, 1816from North Africa. J. Nat. Hist. 36, 1269–1289.

Herter, K., 1937. Die Ökologie der Hirudineen. In: Bronn, H.G. (Ed.),Klassen und Ordnungen des Tierreichs, 4, Bd., 3. Abt., 4. Buch, Teil2. Leipzig, pp. 123–319.

Huelsenbeck, J.P., Ronquist, F., 2001. MrBayes: Bayesian inference ofphylogeny. Bioinformatics 17, 754–755.

Kasparek, M., Demirsoy, A., Akbulut, A., Akbulut, N.E.,Çalşkan, M., Durmuç, Y., 2000. Distribution and status ofthe medicinal leech (Hirudo medicinalis L.) in Turkey. Hydrobiolo-gia 441, 37–44.

International Commission on Zoological Nomenclature, 1999. Inter-national Code of Zoological Nomenclature. International Trust forZoological Nomenclature, London.

Lee, M.S.Y., 2000. Tree robustness and clade signiWcance. Syst. Biol. 49,829–836.

Löytynoja, A., Milinkovitch, M., 2001. SOAP, cleaning multiple align-ments from unstable blocks. Bioinformatics 17, 573–574.

Lukin, E.I., 1976. Piyavki [Leeches]. In: Fauna SSSR. Academy of Sci-ences of the USSR, vol. 1. Nauka, Leningrad.

Moquin-Tandon, A., 1827. Monographie de la Famille des Hirudineés.Maison de Commerce, Montpellier.

624 P. Trontelj, S.Y. Utevsky / Molecular Phylogenetics and Evolution 34 (2005) 616–624

Moquin-Tandon, A., 1846. Monographie de la Famille des Hirudineés.J.-B. Bailliére, Paris.

Nesemann, H., Neubert, E., 1999. Annelida: Clitellata: Branchiobdell-ida, Acanthobdellea, Hirudinea. In: Süßwasserfauna von Mitteleu-ropa, 6/2. Spektrum Akademischer Verlag, Heidelberg, Berlin.

Posada, D., Crandall, K.A., 1998. Modeltest: testing the model of DNAsubstitution. Bioinformatics 14, 817–818.

Sawyer, R.T., 1981. Why we need to save the medicinal leech. Oryx 16,165–168.

Sawyer, R.T., 1986. Leech Biology and Behaviour. Oxford UniversityPress, Oxford.

Shevkunova, E.A., Kristman, V., 1962. Piyavki [Leeches]. Bol’shayameditsinskaya entsiklopediya 24, 794–800.

Siddall, M.E., Burreson, E.M., 1998. Phylogeny of leeches (Hirudinea)based on mitochondrial cytochrome c oxidase subunit I. Mol.Phylogenet. Evol. 9, 277–285.

Shimodaira, H., 2002. An approximately unbiased test of phylogenetictree selection. Syst. Biol. 51, 492–508.

Shimodaira, H., Hasegawa, M., 2001. CONSEL: for assessing the conW-

dence of phylogenetic tree selection. Bioinformatics 17, 1246–1247.Sohn, J.H., Kang, H.A., Rao, K.J., Kim, C.H., Choi, E.S., Chung, B.H.,

Rhee, S.K., 2001. Current status of the anticoagulant hirudin: itsbiotechnological production and clinical practice. Appl. Microbiol.Biotechnol. 57, 606–613.

Stschegolew, G.G., Fedorova, M.S., 1955. Meditsinskaya piyavka iyeyoprimenenie [The Medicinal Leech and its Application]. Medgiz,Moscow.

SwoVord, D.L., 2003. PAUP*. Phylogenetic Analysis Using Parsimony (*and Other Methods). Version 4. Sinauer Associates, Sunderland, MA.

Thompson, J.D., Gibson, T.J., Plewniak, F., Jeanmougin, F., Higgins,D.G., 1997. The ClustalX windows interface: Xexible strategies formultiple sequence alignment aided by quality analysis tools.Nucleic Acids Res. 24, 4876–4882.

Trontelj, P., Sotler, M., Verovnik, R., 2004. Genetic diVerentiationbetween two species of the medicinal leech, Hirudo medicinalis and

the neglected H. verbana, based on random ampliWed polymorphicDNA. Parasitol. Res 94, 118–124.

Utevsky, S.Y., Utevsky, A.Y., Utevskaya, O.M., 1998. Nakhodkaaptechnoy meditsinskoy piyavki Hirudo medicinalis f. oYcinalis vUkraine. [A Wnd of the oYcinal medicinal leech Hirudo medicinalisf. oYcinalis in Ukraine]. Vestnik zoologii 32, 37.

Utevsky, S.Y., Trontelj, P., 2004. Phylogenetic relationships of Wshleeches (Hirudinea, Piscicolidae) based on mitochondrial DNAsequences and morphological data. Zool. Scripta, 375–385.

Voskresensky, A.E., 1859. MonograWya vrachebnykhpiyavok. [Mono-graph of Medicinal Leeches]. S. Petersburg.

Wells, S., Coombes, W., 1987. The status of and trade in the medicinalleech. TraYc Bull. 8, 64–69.

Wells, S.M., Pyle, R.M., Collins, N.M., 1983. The IUCN InvertebrateRed Data Book. IUCN. The World Conservation Union, Cam-bridge, UK.

Whitaker, I.S., Izadi, D., Oliver, D.W., Monteath, G., Butler, P.E.,2004a. Hirudo medicinalis and the plastic surgeon. Brit. J. Plast.Surg. 57, 348–353.

Whitaker, I.S., Rao, J., Izadi, D., Butler, P.E., 2004b. Hirudo medici-nalis: ancient origins of, and trends in the use of medicinalleeches throughout history. Br. J. Oral Maxillofac. Surg. 42,133–137.

White, T.J., Bruns, T., Lee, S., Taylor, J., 1990. AmpliWcation and directsequencing of fungal ribosomal RNA genes for phylogenetics. In:Innis, M.A., Gelfand, G.H., Sninsky, J.J., White, T.J. (Eds.), PCRProtocols. Academic Press, San Diego, pp. 315–332.

Whitman, C.O., 1886. The Leeches of Japan. Q. J. Microsc. Sci. 26, 317–416.

Zapkuvenè, D., 1972. Razvedenie i vyrashchivanie meditsinskikh piya-vok v laboratornykh usloviyakh (1. Razvedenie Hirudo medicinalisf. serpentina i H. medicinalis f. oYcinalis). [Breeding and growing ofmedical leeches under laboratory conditions (1. Breeding of Hirudomedicinalis f. serpentina and H. medicinalis f. oYcinalis)]. LietuvosTSR Moksl akademijos darbai, C serija 3, 71–76.du