Introduction

The genus Orobdella Oka, 1895 consists of terrestrial gastroporous leeches in East Asia (Sawyer 1986). The species diversity of Orobdella has been revised recently, and now this genus includes eight species (Nakano 2010, 2011a, b). Among these species, only one quadrannulate species, Orobdella kawakatsuorum Richardson, 1975, has been known from Hokkaido, Japan (Richardson 1975). This species was described based on the two specimens collected from Sapporo, and its holotype has been deposited at the National Museum of Nature and Science, Tokyo (NSMT). Orobdella kawakatsuorum is characterized especially by its possession of six annuli between gonopores and a simple tubular gastroporal duct.

Quadrannulate Orobdella specimens were recently obtained from various places in Hokkaido. Most of these specimens were identified as Orobdella kawakatsuorum. However, several specimens differ from not only Orobdella kawakatsuorum, but also the other quadrannulate species, Orobdella esulcata Nakano, 2010, Orobdella tsushimensis Nakano, 2011, and Orobdella whitmani Oka, 1895, in several characterisics. Therefore, they are described as a new species herein. In addition, an emended description of Orobdella kawakatsuorum is presented on the basis of its holotype and newly collected materials. The phylogenetic position of the new species is also estimated using mitochondrial COI, tRNA^Cys, tRNA^Met, 12S rDNA, tRNA^Val and 16S rDNA sequence data.

Materials and methods

For the taxonomic study, leeches were collected from Hokkaido, Japan (Fig. 1), under rocks along mountain or forest trails. Altitude and coordinates for localities were obtained using a Garmin eTrex GPS unit.

The preparation of the collected materials for the morphological and molecular analyses follows Nakano (2011b). Two measurements were taken: body length (BL) from the anterior margin of the oral sucker to the posterior margin of the caudal sucker, and maximum body width (BW). Examination, dissection, and drawing of the specimens were accomplished under stereoscopic microscopes with drawing tubes (Leica S6E, M125 and WILD HEERBRUGG TYP 308700).

The numbering convention is based on Moore (1927): body somites are denoted by Roman numerals and annuli in each somite are given alphanumeric designations.

For the molecular phylogenetic analyses, the sequence data of nine Orobdella species were newly obtained (Table 1). As outgroup, three Erpobdelliformes leeches, Gastrostomobdella monticola Moore, 1929 (Gastrostomobdellidae), Erpobdella japonica Pawłowski, 1962 (Erpobdellidae), and Mimobdella japonica Blanchard, 1897 (Salifidae), were included.

Voucher specimens used in this study have been deposited in the National Museum of Nature and Science, Tokyo (NSMT), the Universiti Malaysia Sarawak (UNIMAS), and the Zoological Collection of Kyoto University (KUZ).

PCR and DNA sequencing

Genomic DNA was extracted from botryoidal tissues preserved in 99% ethanol using a modification of the method in Okamoto et al. (2006). After digestion of botryoidal tissues with proteinase K (100 μg/ml) at 37°C for eight–ten hours, DNA was extracted two times with phenol and one time with 25:24:1 phenol/chloroform/isoamyl-alcohol, and precipitated in two volumes of 99% ethanol with one-tenth volume of 3.0 M sodium acetate (pH 5.2). Precipitated samples were dried and stored in TE buffer (10 mM Tris-HCl and 1 mM EDTA [pH 8.0]). Primer sets used in this study are listed in Table 2: for COI, LCO1490 and HCO2198 (Folmer et al. 1994), and LCO-in and HCO-out; for tRNA^Cys, tRNA^Met, 12S, tRNA^Val and 16S (abbreviated 12S), 12SA-out and 12SB-in, and 12SA-in and 12SB-out. All amplification reactions were performed in a GeneAmp PCR System 2700 (Applied Biosystems) or a MyCycler (Bi-Rad Laboratories) using an Ex Taq Polymerase Kit (Takara Bio Inc.). Reaction mixtures were heated to 94°C for 5 min, followed by 35 cycles of 94°C (10 s), 42.5°C (20 s), and 72°C (1 min 13 s for COI, and 1 min for 12S) and a final extension at 72°C for 6 min. The amplified DNA fragments were purified using polyethylene glycol (20% PEG 6000) precipitation.

All samples were sequenced in both directions. Sequencing reactions were performed using a BigDye Terminator v3.1 Cycle Sequencing Kit (Applied Biosystems). Each sequencing reaction mixture was incubated at 96°C for 2 min, followed by 40 cycles of 96°C (10 s), 50°C (5 s), and 60°C (45 s for COI, and 40 s for 12S). The products were collected by ethanol precipitation and sequenced on an ABI 3130xl Genetic Analyzer (Applied Biosystems). Obtained sequences were edited using DNA BASER (Heracle Biosoft S.R.L.). These sequence data were deposited in GenBank.

Phylogenetic analyses

COI sequences were aligned by eye since there were no indels. Mitochondrial 12S sequences were aligned using MAFFT X-INS-i (Hofacker et al. 2002, Katoh and Toh 2008, McCaskill 1990, Tabei et al. 2008) taking into account RNA secondary structure information, and then refined with GBLOCKS (Castresana 2000). The length of aligned sequences of COI was 1266 bp, and that of 12S was 718 bp. Prior to phylogenetic analyses, transition/transversion (ti/tv) rate ratios for each gene sequence was calculated using MEGA5 (Tamura et al. 2011) to test for saturation in base substitutions. It was confirmed that COI and 12S did not show any signs of saturation (ti/tv rate ration of COI was 1.02, and that of 12S was 1.07). Therefore, the concatenated sequences yielded a total of 1984 bp positions.

Phylogenetic tree were constructed using maximum likelihood (ML) and Bayesian inference (BI). Pairwise comparisons of Kimura-2 parameter (K2p) distance (Kimura 1980) were also calculated using MEGA5. ML phylogenies were calculated using TREEFINDER v October 2008 (Jobb et al. 2004) with the tool package PHYLOGEARS v 2.0 (Tanabe 2008), and then non-parametric bootstrapping (Felsenstein 1985) was conducted with 500 replicates. The best-fit models for each partition were selected using the Akaike Information Criterion (Akaike 1974) by using KAKUSAN4 (Tanabe 2011). For the 1st position of COI, the Tamura-Nei model (TN93) with gamma distribution (+G) and proportion of invariant sites (+I) was selected. The transversion model (TVM)+I was selected for the 2nd position, the transition model (TIM)+G for the 3rd position of COI, and the general time reversal model (GTR)+G for 12S. BI and Bayesian posterior probabilities (BPPs) were estimated using the MPI version of MRBAYES v 3.1.2 (Altekar et al. 2004, Huelsenbeck et al. 2001, Ronquist and Huelsenbeck 2003). The best-fit models for each partition were identified using the Bayesian Information Criterion (Schwarz 1978) also by using KAKUSAN4: for COI 1st position, GTR+G+I; the Felsenstein 1981 model (F81)+I for COI 2nd position; the Hasegawa-Kishino-Yano model (HKY85)+G for COI 3rd position; and GTR+G for 12S. Two independent runs for four Markov chains were conducted for 1.5 million generations and the tree was sampled every 100 generations. Based on checking the parameter estimates and convergence using TRACER v 1.5 (Rambaut and Drummond 2009), the first 5, 001 trees were discarded.

The nodes with bootstrap value (BS) higher than 70% were regarded as sufficiently resolved (Hillis and Bull 1993). Nodes with BPP higher than 95% were considered statistically significant (Leaché and Reeder 2002).

Results Taxonomy

Genus Orobdella Oka, 1895

Phylogenetic relationships

The ML tree with ln L = -12757.40 (Fig. 12) was nearly identical to the obtained BI tree (not shown). Monophyly of the genus Orobdella was well supported (BS = 99%, BPP = 100%). Two Orobdella species from Hokkaido, Orobdella koikei and Orobdella kawakatsuorum, formed a monophyletic group (BS = 100%, BPP = 100%). This clade was a sister taxon of the other Orobdella species. Monophyly of Orobdella koikei and Orobdella kawakatsuorum was well supported (Orobdella koikei: BS = 94%, BPP = 100%; Orobdella kawakatsuorum: BS = 97%, BPP = 100%).

The COI sequence divergence between Orobdella koikei and Orobdella kawakatsuorum was between 8.1–9.9% (mean = 9.0%). Intraspecific variation of COI sequences ranged between 4.8–8.1% (mean = 7.1%) in Orobdella koikei, and 0.30–4.9% (mean = 3.7%) in Orobdella kawakatsuorum. Sequence divergence of 12S between these two species was between 3.3–5.5% (mean = 4.3%). It was between 2.8–4.8% (mean = 3.9%) within Orobdella koikei, and 0–1.1% (mean = 0.71%) within Orobdella kawakatsuorum. Interspecific genetic distance of COI between Orobdella koikei and Orobdella kawakatsuorum had a significantly higher mean divergence as compared to the intraspecific variation of Orobdella koikei (t-test with unequal variance, p = 0.0096), but that of 12S did not have a significantly higher mean divergence (t-test with unequal variance, p = 0.22).

Monophyly of seven Orobdella species distributed in areas south of Hokkaido was recovered (BS = 95%, BPP = 100%). Monophyly of Orobdella esulcata + Orobdella shimadae + Orobdella dolichopharynx received strong support (BS = 95%, BPP = 100%), and that of Orobdella jimai + Orobdella octonaria was also recovered (BS = 92%, BPP = 100%).

Discussion

Orobdella koikei differs from the four other quadrannulate species of the genus, Orobdella esulcata, Orobdella kawakatsuorum, Orobdella tsushimensis, and Orobdella whitmani, in the following combination of characteristics (Table 3): 1) dorsal surface brown ; 2) IV uniannulate; 3) male gonopore at XI b6; 4) gastropore nad female gonopore at XIII a1; 5) gonopores separated by 1/2 + 4 + 1/2 annuli; 6) XXV triannulate; 7) gastroporal duct, tubular, but bulbous at junctions with gastropore and crop; 8) epididymides in XVII to XIX; and 9) atrial cornua ovate. Orobdella koikei is easily distinguished from Orobdella dolichopharynx Nakano, 2011, Orobdella ijimai Oka, 1895, Orobdella shimadae Nakano, 2011, and Orobdella octonaria Oka, 1895, in having mid-body somites that are quadrannulate; they are sexannulate in Orobdella dolichopharynx, Orobdella ijimai, and Orobdella shimadae, and octannulate in Orobdella octonaria.

The phylogenetic tree showed that the clade, which includes Orobdella species in Hokkaido, was a sister taxon of the other Orobdella species. This result is also recovered by the other phylogenetic analyses based on nucleic 18S and 28S sequences (Nakano et al. in press). According to the phylogenetic analyses, several characteristics are considered to have evolved in parallel. Each of Orobdella kawakatsuorum and Orobdella esulcata possess a tubular gastroporal duct (Nakano 2010). However, these two species are phylogenetically distant. The mid-body somite annulation of Orobdella leeches does not indicate phylogenetic relationships either. In the genus Orobdella, it is clear that the quadrannulate mid-body somite is a plesiomorphic character. Sexannulate mid-body somite, which Orobdella dolichopharynx, Orobdella ijimai, and Orobdella shimadae possess, evolved in parallel. Orobdella kawakatsuorum, and two species from Ryukyu Archipelago, Japan, Orobdella dolichopharynx and Orobdella shimadae, possess rudimentary male atrial cornua (Nakano 2011b). However, undeveloped male atrial cornua do not indicate any phylogenetic relationships between Orobdella kawakatsuorum and Ryukyu Orobdella species. These characters are not useful for estimating phylogenetic relationships in the genus Orobdella, although they are suitable for the species level classification.

Species delimitation in leeches based on genetic analyses, especially using COI DNA-barcode locus, has been discussed in many papers (see DeSalle et al. 2005 for review). The average sequence divergence of COI between Orobdella koikei and Orobdella kawakatsuorum was 9.0%, and that of 12S was 4.3%. Interspecific genetic divergence of COI between these two species showed a significantly higher value than that of the intraspecific variation, although the intraspecific genetic divergence of 12S sequences in Orobdella koikei (2.8–4.8%) was overlapped to a large extent with the interspecific divergence of 12S (3.3–5.5%). Thus, only the genetic distance of COI (mean = 9.0%) can be used as an indicator for deciding whether leeches are distinct species or not in the genus Orobdella, since Orobdella koikei and Orobdella kawakatsuorum are distributed syntopically at Sounkyo (Fig. 1).

Gilyarov et al. (1969) reported quadrannulate Orobdella species from Primorsky Krai, Russia, as Orobdella whitmani. Although they did not describe the detailed internal anatomy of the specimen, the photograph of the ventral surface of their specimen (fig. 1 in their paper) clearly shows that the male gonopore opened in the furrow of XI/XII and the female gonopore was at XIII a1. Thus, the number of annuli between the gonopores was 4 + 1/2. This characteristic is not identical to those of the other known quadrannulate Orobdella species. There is a strong possibility that the quadrannulate Orobdella species distributed in Primorsky Krai is an undescribed species. Primorsky Krai is located at the same latitude as Hokkaido, Japan. Clarifying the taxonomic status and phylogenetic position of Orobdella in Primorsky Krai will help to reveal the species diversity and the evolutionary history of the genus Orobdella.

Key to the known species of the genus Orobdella