Taxonomy and natural history of Cavernocypris hokkaiensis sp. nov., the first ostracod reported from alpine streams in Japan

We describe the cypridoidean ostracod Cavernocypris hokkaiensis sp. nov. from riverbed sediments in an alpine stream at an elevation of ca. 1850 m in the Taisetsu Mountains, Hokkaido, Japan. This species differs from congeners in having (1) the outer surface of the carapace smooth, with sparse, tiny setae, but without shallow pits; (2) the carapace elongate rather than triangular in lateral view; (3) the antennula consisting of seven podomeres; (4) first palpal podomere of maxillula with five dorsodistal and one ventro-subdistal setae; (5) the fifth limb lacking setae b and d; and (6) the fifth limb lacking a vibratory plate. We provided the key to the Cavernocypris species. We determined partial sequences for the cytochrome c oxidase subunit I (COI; cox1) and 18S rRNA (18S) genes in C . hokkaiensis . Our sample contained only females, and we obtained a partial 16S rRNA sequence for the endosymbiotic bacterium Cardinium from C. hokkaiensis , indicating the possibility that this ostracod species reproduces parthenogenetically. Our field survey and observations of captive individuals suggested that C. hokkaiensis may be endemic to the Taisetsu Mountains, with a low population density, narrow distributional range, and slow maturation to sexual maturity.


Introduction
The genus Cavernocypis Hartmann, 1964, one of 20 genera in the subfamily Cypridopsinae Kaufmann, 1900(Savatenalinton 2018Meisch et al. 2019), is distinguished from the other 19 genera by the following combination of features (cf. Smith et al. 2017): (1) carapace elongate to triangular in lateral view, (2) left valve overlapping right valve along ventral margin, (3) surface of valves smooth or with small pits, (4) swimming setae of antennae very short, (5) distal segment of maxillular palp elongate, (6) terminal segment of seventh limb not separated, and (7) uropodal ramus flagellum-like and present only in females.
To date, six Cavernocypris species have been described from the Palearctic and Nearctic regions (Meisch et al. 2019;Külköylüoğlu 2020); in Japan, one species, C. cavernosa Smith, 2011, has been reported from Shiga and Shizuoka prefectures (Smith 2011;Tanaka et al. 2015). Cavernocypris species inhabit the interstitial environment of riverbed sediments, the littoral zone of mountain lakes, springs, and caves (Marmonier et al. 1989;Smith et al. 2017). Although there is little information on their ecology, Forester (1991) found C. wardi Marmonier, Meisch & Danielopol, 1989 only in cold-water habitats at ca. 0-14 °C and suggested that the species may be cryophilic.
Streams in the alpine vegetation zone between the treeline and the permanent snow line are cold and nutrientpoor (Niedrist and Füreder 2017). They arise from glacial melt, snowmelt, rain runoff, and groundwater springs and are highly environmentally heterogeneous (Hotaling et al. 2017). Many organisms inhabiting alpine streams are endemics (e.g. Muhlfeld et al. 2011), uniquely adapted to harsh conditions (Lencioni et al. 2009). Ostracods are often detected in ecological research in alpine streams (e.g., Suren 1993;Zbinden et al. 2008), but their taxonomy, particularly outside Europe, has not been well studied.
The Taisetsu Mountains in Daisetsuzan National Park are located in the center of Hokkaido, Japan, and consist of several gently sloping volcanic peaks in the 2000 m class.
Above the treeline at ca. 1400-1500 m elevation (Amagai et al. 2018), there are several aquatic features, including alpine streams, but except for insects (e.g. Konno 2003;Konno et al. 2003), the aquatic invertebrates have not yet been investigated.
Here we describe a new species of Cavernocypris from an alpine stream in the Taisetsu Mountains, the first record of an ostracod from cold alpine waters in Japan. We present nucleotide sequences for this species for parts of the mitochondrial cytochrome c oxidase subunit I (COI) and nuclear 18S rRNA (18S) genes and provide preliminary comments on its phylogenetic position based on 18S data. We also present information on its natural history based on a field survey and preliminary rearing results. Finally, we demonstrate with molecular data (part of the mitochondrial 16S rRNA gene, 16S) the infection of this species by Cardinium, a group of "reproduction-manipulating" endosymbiotic bacteria (Ma and Schwander 2017).

Materials and methods
Sampling was conducted at seven sites, including four streams (Stns 1-3, 7) fed by springs, a hot spring, and/ or snowmelt, two ponds (Stns 5, 6), and one waterfall basin (Stn. 4) ( Table 1, Fig. 1). Bottom sediment and water were placed in a bucket and stirred, and all but the sediment was filtered through a plankton net (mesh size 63 µm). This process was repeated several times at each site. Specimens were picked from the samples using a stereomicroscope (Olympus SZX9, Japan). Geographical coordinates and the elevation were obtained from GSI Maps (Geospatial Information Authority of Japan 2022). Water temperature was measured by using an O-274 thermometer (DRETEC, Japan).
Ostracods were fixed in 80% ethanol. The methods used for dissection, preparation of slides, light microscopy, scanning electron microscopy (SEM), and drawing were as described by Munakata et al. (2021). All material studied has been deposited in the Invertebrate Collection of the Hokkaido University Museum (ICHUM), Sapporo, under catalog numbers ICHUM-8247 to 8252.
An attempt was made to extract total DNA from the soft parts of three individuals by using a NucleoSpin Tissue XS Kit (Macherey-Nagel, Germany) following the manufacturer's protocol, but only one of the three extracts allowed successful PCR amplification. Primers used for the PCR amplification and sequencing of ostracod COI, ostracod 18S, and Cardinium 16S are listed in Munakata et al. (2021), except that CLO-f2 (GGTGCGTGGGCGGCTTATT) and CLO-r2 (AAAGGGTTTCGCTCGTTATAG) (Gotoh et al. 2007) were used instead of CLO-f1 and CLO-r1. PCR amplification conditions for COI with TaKaRa Ex Taq DNA polymerase (TaKaRa Bio, Japan), for 18S with KOD FX Neo (Toyobo Life Science, Japan), and for 16S with TaKaRa Ex Taq were as described by Munakata et al. (2021). All nucleotide sequences were determined by direct sequencing with a BigDye Terminator Kit ver. 3.1 and a 3730 DNA Analyzer (Life Technologies, USA). Fragments were concatenated using MEGA7 (Kumar et al. 2016). BLAST (Altschul et al. 1990) was used to search the International Nucleotide Sequence Database (INSD; International Nucleotide Sequence Database Collaboration 2022) for nucleotide sequences most similar to our sequences.  To explore the phylogenetic position of this species, a maximum likelihood (ML) phylogenetic tree was constructed based on the 18S dataset comprised of 66 ostracod sequences (one our sequence, and 64 cypridoidean and one pontocypridoidean (outgroup) sequences from INSD; 1547 positions in the aligned dataset; see Suppl. material 1-3: Table S1, Alignments S1, S2). The detailed method and result are shown in Suppl. material 4: File S1.
To obtain information on the life cycle, three nonadult individuals collected on 26 July 2021 were maintained singly in wells of a tissue culture plate filled with water collected from sampling site Stn. 1 and placed in a refrigerator at a temperature of 7 °C. Detritus collected from sampling site Stn. 1 was added to each well as a food source. Observations were made twice or more per month.

Field survey and observation of captive individuals
Among seven sampling sites, Cavernocypris ostracods were collected from only two sites in the Hokkai-sawa Stream (Stns 1 and 2) ( Fig. 1B, C). Ostracod density at both sites was low, with fewer than 10 individuals per 500 ml of filtered residue. No male individuals were detected.
Three captive non-adult individuals have remained alive and active for more than five months. The body length (LV-L) of each individual was 0.47, 0.48, 0.39 mm. No molts have been observed to date (the latest observation was on 7 January 2022).
Etymology. The epithet hokkaiensis is an adjective referring to the type locality, Hokkai-sawa Stream in Japan.
New Japanese name. Shibare-doukutsu-kaimijinko, referring to the habitat having low water temperature. Shibare is derived from the Japanese verb shibare-ru (freeze), in a Hokkaido dialect; Doukutsu-kaimijinko is the Japanese name for Cavernocypris (Tanaka et al. 2015).
Description Third podomere with dorsodistal seta reaching distal edge of fourth podomere and ventrodistal seta extending beyond middle of fourth podomere. Fourth podomere with two dorsodistal setae of unequal length (longer one extending to distal edge of seventh podomere) and two ventrodistal setae reaching distal edge of sixth podomere. Fifth podomere with two dorsodistal setae of unequal length (longer one as long as podomeres 2-7) and two shorter ventrodistal setae. Sixth podomere with four outer distal setae as long as podomeres 1-7 and shorter inner distal seta. Seventh podomere with three distal setae of unequal length and aesthetasc y a (ca. 60% length of longest seta).
An2 (Fig. 3D) with five podomeres. First podomere (coxa) with two ventral setae. Second podomere (basis) with ventrosubdistal seta reaching distal edge of third podomere. Exopodite with one long and two unequal short setae. Third (first endopodal) podomere with six inner subdistal short natatory setae extending slightly beyond distal edge of third podomere, ventrodistal plumed seta reaching distal edge of fifth podomere, and mid-ventral aesthetasc Y. Fourth podomere undivided, with two mid-dorsal setae, dorsosubdistal setae z 1-3 extending beyond middle of claws G 2,3 , mid-ventral plumed setae t 1-4 of unequal length, and distal claws G 1-3 of nearly equal length. Fifth podomere with bifurcate aesthetasc y 3 (longer part half the length of claw G M ) and claws G m,M ; G m ca. 70% length of G M ; G M reaching tip of claws G 1-3 . Md (Fig. 3E, F) with coxa, palp comprising four (one basal and three endopodal) podomeres, and vibratory plate. Coxa with several distal teeth and one subdistal plumed seta. First podomere (basis) with one ventrodistal seta, ventrodistal setae S 1,2 , and ventrodistal short seta α; setae S 1,2 subequal in length, bearing row of long setules. Vibratory plate (exopodite) with four rays. Second (first endopodal) podomere with three dorsodistal setae of unequal length (longest reaching tip of claws on fourth podomere), four mid-ventral long plumed setae (not extending beyond tip of claws on fourth podomere), and mid-ventral plumed short seta ß (shorter than half the length of mid-ventral plumed setae). Third podomere with four dorsosubdistal setae and two ventrosubdistal setae; inner distal region with plumed seta γ and two plumed setae. Fourth podomere with two distal setae and four distal claws. Mx (Fig. 3G) with palp comprising two podomeres, three endites, and vibratory plate (not illustrated). First palpal podomere with five dorsodistal setae of unequal length and one ventro-subdistal seta. Second palpal podomere not spatula-like, but rectangular, with two distal setae and three distal claws. First endite with two ventroproximal setae and ca. nine distal setae. Second endite with ca. eight distal setae. Third endite with two distal serrated spines and six distal setae. L5 ( Fig. 3H) with protopod and palp; vibratory plate absent. Protopod with two setae a and at least nine distal plumed setae of unequal length; setae b, c, and d absent. Palp with distal plumed setae h 1-3 . L6 ( Fig. 3I) with six podomeres. First and second podomeres (protopod) with seta d 2 . Third (first endopodal) podomere with ventrodistal plumed seta e reaching middle of fifth podomere. Fourth podomere with ventrodistal plumed seta f reaching beyond distal edge of fifth podomere. Fifth podomere with ventrodistal plumed seta g. Sixth podomere with dorsodistal seta h 3 , ventrodistal plumed seta h 1 , and distal curved claw h 2 . L7 ( Fig. 3J) with four podomeres; third and fourth podomeres fused to form pincer organ. First podomere (protopod) with plumed setae d 1,2,p . Second (first endopodal) podomere with ventrodistal plumed seta e not reaching middle of fused podomeres 3 and 4. Fused third and fourth podomeres with mid-ventral plumed seta f not reaching tip of L7, subdistal long plumed seta h 3 , distal hook-like seta h 2 , and subdistal tiny seta h 1 . UR (Fig. 3K) strongly reduced. Proximal part longer than wide, with one seta. Distal part flagellar in shape.
Rake organ (Fig. 3L) with stout rod and ca. eight blunt distal teeth.
Genital hooks on female copulatory organ present (not illustrated).

Morphological comparison
Cavernocypris hokkaiensis sp. nov. resembles C. cavernosa and C. danielopoli Smith & Kamiya, 2017 in lacking setae b and d on L5, but differs from them in that (1) the outer surface of the carapace is smooth, with sparse, tiny setae, but without shallow pits (pits present in C. cavernosa and C. danielopoli); (2) the carapace is elongate rather than triangular in lateral view (triangular in C. danielopoli); (3) first palpal podomere of Mx has five dorsodistal and one ventro-subdistal setae (only five dorsodistal setae present in C. danielopoli; not described in C. cavernosa); and (4) L5 lacks the vibratory plate (vibratory plate present in C. cavernosa and C. danielopoli). Character states in all congeners are summarized in Table 2.

Reproductive mode
Our sample comprised only females, indicating that C. hokkaiensis may be parthenogenetic. The endosymbiotic bacterium Cardinium has previously been detected (e.g., by means of 16S sequences) in nonmarine ostracods engaged in parthenogenetic or mixed reproduction, and infection with Cardinium might be a causative factor in the parthenogenetic reproductive mode (Schön and Martens 2019). Our study is the first to detect Cardinium in a species of Cavernocypris, implying that C. hokkaiensis may be parthenogenetic. It should be noted that male individuals have likewise not been reported among the congeners C. cavernosa, C. danielopoli, C. wardi, and some populations of C. subterranea (Wolf, 1920) (Marmonier et al. 1989;Külköylüoğlu and Vinyard 1998;Smith 2011;Smith et al. 2017).

Ecology, distribution, and life cycle
The results of our field survey suggest that C. hokkaiensis is distributed in an extremely narrow area, only in Hokkaisawa Stream. It was not found at three sites distant from Hokkai-sawa Stream (Stns 5, 6, and 7). Its absence at two sites in Akaishi Stream (Stns 3 and 4), which Hokkai-sawa Stream joins, may be related to environmental differences between two streams. Hokkai-sawa Stream is fed by spring water and snowmelt, and is thus cold (1.4-6.8 °C measured in the summer season; Table 1; Konno et al. 2003) and slightly acidic (pH 6.2; Konno et al. 2003). In contrast, Akaishi Stream is fed by water from the Yudokuonsen hot spring in the Ohachidaira Caldera, which has high water temperature (48 °C) and high H 2 S content (245 mg/l) (Uzumasa et al. 1959), and measurements taken at Stn. 3 indicate that Akaishi Stream is warmer (13 °C) and more acidic (pH 2.8-3.3) than Hokkai-sawa Stream (Table 1; Konno et al. 2003). No pH data were available for our Stn. 4, but the water temperature was similar to that at Stn. 3 (13 °C; Table 1). Based on samples collected from sites almost identical to our Stns 1 and 3 in Hokkai-sawa and Akaishi Streams, Konno et al. (2003) found no lotic aquatic insects in common between the two streams. The warmer, more acidic condition of Akaishi Stream may be an unsuitable habitat for C. hokkaiensis.
Our rearing experiment, though we could prepare only three live individuals, provided preliminary data about the life history of C. hokkaiensis. We observed no molting by captive C. hokkaiensis individuals for more than five months at 7 °C. Ostracod life cycles typically comprise eight non-adult and one adult instars, i.e., ostracods molt eight times before becoming sexually mature adults. Instars are not uniform in duration, but tend to become longer with successive instars (e.g., Heip 1976;Liberto et al. 2014). Although we could not determine the true instar for three captive individuals, their estimated instar was A-1 (0.47-and 0.48-mm individuals) or A-2 (0.39-mm individual) if this species follows Brooks's rule (Brooks 1886;Watabe and Kaesler 2004). Our result could be an artefact from the artificial conditions during culturing, but it may also be a natural phenomenon, and we could speculate that this species may require more than a year to become sexually mature.
Our field survey and observation of captive individuals may indicate that C. hokkaiensis is an endemic species adapted to the harsh alpine environment of the Taisetsu Mountains, with a low population density, narrow distributional range, and potentially slow maturation. If this is the case, then habitat loss and fragmentation due to anthropogenic activities, or a decrease in snowfall and snowfields due to climate change, could lead to a rapid population decline of this species. Additional ecological and biological information is necessary to confirm whether C. hokkaiensis is a narrow endemic, and to design an informed conservation strategy.
Key to the Cavernocypris species (modified after Külköylüoğlu (2020)