Helobdella blinni sp. n. (Hirudinida, Glossiphoniidae) a new species inhabiting Montezuma Well, Arizona, USA

Abstract A new leech species Helobdella blinni sp. n., is described from Montezuma Well, an isolated travertine spring mound located in central Arizona, USA. In its native habitat, Helobdella blinni had been previously identified as Helobdella stagnalis (Linnaeus, 1758), which was later reclassified to Helobdella modesta (Verrill, 1872). Similar to the European Helobdella stagnalis and North American Helobdella modesta, Helobdella blinni has six pairs of testisacs, five pairs of smooth crop caecae, one lobed pair of posteriorly-directed crop caecae, one pair of eyes, a nuchal scute, and diffuse salivary glands. However, the pigmentation of this new species ranges from light to dark brown, unlike Helobdella modesta which tends to be light grey in color. Also, Helobdella modesta produces a clutch of 12-–35 pink eggs, whereas Helobdella blinni produces smaller clutches of white eggs (7–14, 0.5 ± 0.15 mm, N = 7) and consequently broods fewer young (1–14, 7 ± 3.3 mm, N = 97). Helobdella blinni are also able to breed year-round due to the constant warm water conditions in Montezuma Well. Their breeding season is not restricted by seasonal temperatures. These species are morphologically similar, however, comparing the COI mtDNA sequences of Helobdella blinni with sequences from nearby populations of Helobdella modesta and other Helobdella species from GenBank indicate that Helobdella blinni is genetically distinct from these other Helobdella populations.


Introduction
Montezuma Well is a collapsed travertine spring mound located 72 km south of Flagstaff in the Verde Valley of Northern Arizona (34. 6491°N,111.7522°W (DD)) ( Fig. 1A). The age of Montezuma Well is estimated to be ~11,000 years (Wagner and Blinn 1987). This location is thermally constant year-round (19-24˚C) and is continuously replenished by two vents located at the well bottom. Montezuma Well is 0.76 ha in area and approximately 20 m deep. Most of the shoreline drops off immediately into open water, except at the northeast corner where water drains through a shallow region called the "swallet" and empties into Wet Beaver Creek which is located east of Montezuma Well (Fig. 1A-B). The water within Montezuma Well has unique water chemistry, containing high levels of arsenic (>100μg/L) and dissolved CO 2 (>300mg/L) (Cole and Barry 1973).
Four leech species are known to inhabit Montezuma Well, including an endemic pelagic predator , the erpobdellid Motobdella montezuma (Davies et al. 1985), and three other glossiphoniid species currently identified as Helobdella papillata (Moore, 1952), H. elongata (Castle, 1900), and a species currently thought to be H. stagnalis (Linnaeus, 1758), all of which inhabit the swallet (Fig. 1B). These Montezuma Well leech populations are thought to have been isolated from other leech populations for as long as 11,000 years (Wagner and Blinn 1987).
In support of this hypothesis, Beresic-Perrins's (2010) description of brood size, parental behavior, and life history of the Montezuma Well population of H. stagnalis suggests that this leech is distinct from other known populations of H. stagnalis, a species originally described from Europe and which had until very recently been considered to be a widespread cosmopolitan leech species, inhabiting both Europe and North America.  addressed this problem by resurrecting the original species description for the North American leech, Helobdella modesta (Verrill, 1872) which had long been considered to be a synonym of the European H. stagnalis (Moore 1898). The molecular analysis by Moser et al. (2011) provided confirmation for the resurrection of H. modesta by . Even though the two species are morphologically indistinguishable (Verrill 1872, Moore 1898, Moore 1952, they differ genetically. Henceforth, we will refer to the North American H. stagnalis as H. modesta. Here, we compare key traits, both morphological and molecular, among members of the Montezuma Well Helobdella sp. population, several other nearby populations of H. modesta, and several other Helobdella species. Our molecular analysis includes the cytochrome c oxidase subunit I (COI) mitochondrial gene region to test the hypothesis that the Montezuma Well population of H. modesta is a distinct species and warrants a new species description. This region is known to be sufficiently variable to reveal interspecific differences and unlikely to suggest differences due to elevated mutation rates (Apakupakul et al. 1999).

Sampling
A total of 34 individuals of Helobdella sp. inhabiting Montezuma Well were collected from the underside of rocks in the swallet: five specimens were collected in June 2011 for molecular analysis and 29 were collected in June 2012 to assess morphological characteristics. For the molecular analysis, the leeches were preserved in 95% ethanol and others, for museum collections, were fixed with buffered formalin overnight and preserved in 70% ethanol. Additionally, a total of 10 specimens of H. c.f. modesta from Rio de Flag ponds near the Rio de Flag Waste Water Facility outflow in Flagstaff, Arizona (35.18418°N, 111.63294°W (DD)) and Oak Creek, AZ near the Cave Springs campground (34.9961°N, 111.7394°W (DD)) were collected for molecular analyses. These specimens were also fixed in 95% ethanol.

Morphological examination
We documented number of eyes and their placement, color pattern, presence of papillae, number of and structure of gastric caecae, body size, presence of nuchal scute, gonopore placement, egg size and number, and number of offspring using a Nikon binocular dissecting microscope. We then deposited the examined materials in the Invertebrate Zoology collection at the Smithsonian Institution, National Museum of Natural History (USNM).

Molecular analysis
Whole DNA was extracted from the caudal suckers of the individual leeches using a Qiagen DNeasy Blood & Tissue Kit (Cat. No. 69504), with each sample incubated overnight in a water bath set at 54°C. Using Siddall and Borda's (2002) PCR method, the mitochondrial gene region, cytochrome c oxidase subunit I (COI) was amplified. The primers were LCO1490 5'-GGTCAACAAATCATAAAGATATTGG-3' and HCO2198 5'-TAAACTTCAGGGTGACCAAAAAATCA-3' (Folmer et al. 1994). The PCR product was purified through the use of the QIAquick PCR Purification Protocol (Cat. No. 28104), checked for PCR product using gel electrophoresis, and sequenced with an ABI Prism 3730 sequencer (Applied Biosystems). We imported the seven "cleanest" sequences and 71 comparative sequences (Table 1) (Kumar et al. 2016). We aligned the sequences automatically using MUSCLE (Edgar 2004) and then corrected the alignments by hand. We partitioned the data and performed the substitution model test by codon in Partitionfinder (Lanfear et al. 2012). The best substitution model test was General Time Reversal (GTR) +gamma which we used in our maximum-likelihood (ML) analysis (Lanave et al. 1984, Tavare 1986, Rodriguez et al. 1990). For ML analysis, we used RAxML v. 8 (Stamatakis 2014) and included 1,000 nonparametric bootstrap replicates. We used MrBayes for Bayesian inference analysis with ten million generations with a 25% burn-in and our support was assessed based on clade posterior probabilities (Ronquist and Huelsenbeck 2003). These analyses were conducted through CIPRES (Miller et al. 2010). We used PAUP* 4.0 (Swofford 2003) to construct parsimony phylogenies with 100 random additions. We performed the parsimony analysis twice, treating the deletions in the sequences as a 5 th state and then as missing data. We performed an uncorrected p-distance analysis to examine nucleotide differences between sequences with 1,000 replicates in MEGA7.0.18 (Kumar et al. 2016).
Etymology. We have named this new species, Helobdella blinni in honor of Dr. Dean W. Blinn for his dedication to natural history research at Montezuma Well. For over 20 years at Northern Arizona University, Dr. Blinn studied a wide range of organisms and their interactions at Montezuma Well including predator-prey interactions between Motobdella montezuma and the endemic amphipod, Hyalella montezuma Cole & Watkins, 1977. Description. External morphology. Length of specimens 11 to 22 mm (mean + SE 16.6 + 3.2 N=24) and width 3 to 8 mm (5.7 + 1.1 N=28) (Table 3, Figs 2, 3).  Individual color ranges from translucent with brown spots to dark brown (Fig. 4).
No dorsal papillae; one pair of eyes located at somite II (0.07 + 0.02 mm diameter, N = 11), distance between eyes 0.1 to 0.03 mm apart (N = 13). A scallop-shaped  nuchal scute is present on the dorsal side, length 0.293 to 0.432 mm (0.335 + 0.05 N=9) and width 0.27 to 0.386 mm (0.32 + 0.04 N=9). One annulus separates the female and male gonopores. The caudal sucker diameter averages 1.6 + 0.3 mm (N = 27). The eggs (diameter 0.5 + 0.15 mm, N = 28) are laid on the ventral side of the parent in soft-walled transparent cocoons (7-11 eggs per cocoon, N = 7). The mouth is located subterminally in the oral sucker (Figs 2, 3). Internal morphology. Average oral sucker diameter is 0.7 + 0.19 mm (N = 15), proboscis length is 3.5 + 1.1 mm (N = 17) (Table 3). Diffuse salivary glands are located near the anterior of the first pair of crop caecae. There are five pairs of smooth crop caecae and one lobed pair of posteriorly directed post caecae. Six pairs of compact testisacs are located in between each of the crop caecae. The intestine contains four pairs of caecae, with the first two pairs anteriorly directed and the other two pairs posteriorly directed. The intestine leads into an unraised anus located two annuli from the caudal sucker (Figure 3).
Development and growth. This species breeds year-round with peaks in spring and fall. Our specimens had an average of 7 to 11 white eggs (diameter 0.5 + 0.15 mm, N = 7) fixed to their ventral surface. Laboratory collections (2007-10) of H. blinni documented the eggs hatching 1 to 2 weeks after ovipositing (Beresic-Perrins 2010). Once hatched, the young attach to the ventral surface of the parent, allowing the parent to hunt for food and feed the young, occasionally feeding along with them. Prey consists of oligochaetes and other invertebrates. The average number of young per adult is 7 + 3.3 (N = 97) ranging from 1 to 14 offspring. The young remain attached to the parent for an additional four to five weeks after hatching. Once the juveniles leave the parent, they tend to aggregate together on rocks (Beresic-Perrins 2010).

Molecular analysis
A Bayesian inference phylogenetic tree of the COI sequence data is presented in Figure 5. We include the posterior probabilities and maximum-likelihood branch supports >50. The Arizona populations of Helobdella c.f. modesta formed a sister clade to Helobdella blinni sp. n., supported by both the Bayesian and parsimony analyses. The results of the uncorrected p-distance analysis revealed a difference of 13.3% (233 nucleotides included) between the two groups (  (Table 6).
When we aligned all 78 sequences, there were four, ten-codon deletions within all of the Arizona sequences and H. atli León-Regagnon 2005, Oceguera-Figueroa et al. 2010). When we performed the parsimony analysis, we included deletions as a 5 th state in our first analysis and in our second, we treated the deletions as missing data. In the resulting 5 th state tree, the two Arizona species remained sister taxa (100% support), but included in the clade was H. atli (100% and 58% support). The missing data tree placed H. blinni ancestral to H. modesta (Washington), H. modesta (Ohio), H. stagnalis (UK), and H. c.f. modesta with 100% branch support (Fig. 5).

Discussion
Helobdella blinni sp. n. has morphological and life-history traits similar to other Helobdella species, including possession of a nuchal scute, diffuse salivary glands, six pairs of testisacs, and extended parental care for the young (6-7 weeks; Tables 4, 5). Helobdella blinni, H. bowermani (Moser et al. 2013), H. octatestisaca (Lai et al. 2009), and H. c.f. modesta each have five pairs of smooth crop caecae as opposed to six pairs of lobed crop caecae in H. californica (Kutschera 2011) and H. papillornata   Figure 5. Bayesian Inference phylogenetic tree with 25% burn-in and support was assessed based on clade posterior probabilities tree. We included COI sequences from 31 species of Helobdella (family Glossiphoniidae). The Arizona populations are from Oak Creek (OC), Rio de Flag (RDF), and Montezuma Well (MW). Our outgroup included Cystobranchus salmositicus (Meyer, 1946), Gonimosobdella klemmi (Williams & Burreson, 2005), Myzobdella lugubris (Leidy, 1851), and Ozobranchus margoi (Davies, 1978). The shaded branches are the Arizona sample sequences. Branch labels include the Bayesian / ML probability. The blue nodes are supported by Bayesian Inference, Maximum-Likelihood, and parsimony analyses. The yellow nodes are supported by Bayesian Inference and Maximum-Likelihood analyses. The green nodes are supported by Bayesian Inference and parsimony analyses. The red nodes are supported by Bayesian Inference analysis only.
H. c.f. modesta, H. californica, H. temiscoensis, H. atli, and H. bowermani possess six pairs of testisacs, whereas H. papillornata has five pairs and H. octatestisaca has four pairs. Helobdella blinni also has a larger proboscis than the other Helobdella species (mean + SE, H. blinni 3.5 mm + 1.1, N=17, H. californica mean = 0.7 mm, H. papillornata mean= 2 mm). Furthermore, breeding periods also differ between H. blinni and the other Helobdella species (Tables 4, 5).   Bennike (1943) Helobdella blinni, unlike the other Helobdella species discussed here, breeds yearround, living in the thermally stable environment of Montezuma Well, with constant (19-24˚C) year-round temperatures (Table 5). Monthly samples have individuals carrying cocoons every month of the year, with peak seasons in the spring and fall, a situation quite different than that for other Helobdella species, which have seasonallyconstrained reproductive cycles, with egg-laying and brooding beginning in the spring and ending in the fall every year (Table 5). In addition to breeding year-round, H. blinni produces smaller broods (7-14 young) when compared to H. modesta and H. stagnalis (12-35 young) (Tables 4, 5) and has white eggs, unlike the characteristically pink eggs of H. modesta, H. californica, and H. papillornata (Table 4). The external pigmentation of H. blinni also tends to be dark brown, whereas most other Helobdella species are grey/brown in color (Fig. 4). Helobdella blinni are slightly longer (body length 11-22 mm, 16.6 + 3.2, N=24) than H. c.f. modesta (8-12 mm) and H. californica (10-18 mm), but slightly shorter in length than H. papillornata (15-40 mm) (   Fig. 5). Based on morphological, life-history, and molecular differences, we propose the Helobdella sp. leeches found at Montezuma Well should be considered a new species, likely the result of allopatric isolation. This concept supports our hypothesis that the leech species inhabiting Montezuma Well may have become isolated from other populations as far back as 11,000 years ago (Wagner and Blinn 1987). Helobdella blinni sp. n. can be considered a distinct species found in Montezuma Well and may also turn out to be endemic to the area. Further sampling and analyses are needed in order to verify endemism.
Although currently classified as Helobdella c.f. modesta, the Arizona populations from the Rio de Flag and Oak Creek may be an additional undescribed species based on our molecular analysis. Our next step is to investigate these populations more closely, comparing them to other local populations, including White Horse Lake and J.D. Dam Lake, AZ which also contain H. modesta.