Integrative description of a new Dactylobiotus (Eutardigrada: Parachela) from Antarctica that reveals an intraspecific variation in tardigrade egg morphology

Tardigrades constitute one of the most important group in the challenging Antarctic terrestrial ecosystem. Living in various habitats, tardigrades play major roles as consumers and decomposers in the trophic networks of Antarctic terrestrial and freshwater environments; yet we still know little about their biodiversity. The Eutardigrada is a species rich class, for which the eggshell morphology is one of the key morphological characters. Tardigrade egg morphology shows a diverse appearance, and it is known that, despite rare, intraspecific variation is caused by seasonality, epigenetics, and external environmental conditions. Here we report Dactylobiotus ovimutans sp. nov. from King George Island, Antarctica. Interestingly, we observed a range of eggshell morphologies from the new species, although the population was cultured under controlled laboratory condition. Thus, seasonality, environmental conditions, and food source are eliminated, leaving an epigenetic factor as a main cause for variability in this case.

Description. (Figures 1-4, Supplementary Figure S2; measurement and statistics in Supplementary Table S1) Body colour white and slightly opaque (in live specimens) or transparent (in Hoyer's medium). Smooth cuticle without gibbosites, spines and pores. Dorso-lateral papillae also absent, but see remarks below ( Fig. 1). Eyes present in live specimens (Fig. 1c), and most specimens show eyes after mounting in Hoyer's medium. Ten peribuccal lamellae present (Fig. 2a,b) with the hook-shaped first band of teeth at the base (Fig. 2c). The first band of teeth (FBT) faint under Differential Interference Contrast (DIC) microscope, but clearly visible under SEM (Fig. 2a,b). The second band of teeth (SBT) larger caudally (Fig. 2d), followed by dorsal and ventral transverse crests. In most observed specimens, dorsomedian and ventromedian transverse crests with a pointed tip ( Fig. 2e,f, Supplementary Fig. S2). The dorsomedian crest slightly broader than the ventromedian crest (Fig. 2f,g). Dorsolateral and ventrolateral crests have similar size and several tips, not fragmented (Fig. 2f,g). Two oval perforated areas (porous areas 56 ) on the lateral sides posterior to transverse crests, respectively (Fig. 2g,h). The buccal tube (Fig. 3a), with slight antero-ventral bend (Fig. 3b). The asymmetric apophysis for the insertion of the stylet muscles (AISM) accentuated by the ventral lamina and shows the bulbous expansion with the deep invagination (or hook 53 ) (Fig. 3c). At the view from posterior to anterior direction, oval perforated area goes through the hole and into the mouth (Fig. 3d), one dorsal and one ventral to the stylet sheath (Fig. 3e). Two holes and two oval perforated areas in each lateral side of buccal crown, hence four holes and oval perforated areas in total. Piercing stylets of a living specimen 60 μm long with curved shape present (Fig. 3f). Stylet supports (Fig. 3g) inserted on the buccal tube at posterior position. The pharyngeal bar and three bi-lobed pharyngeal apophyses (Fig. 3h) at the posterior end of the buccal tube. Pharynx with two rod-shaped macroplacoids. The first macroplacoid with central constriction at the middle. Microplacoid and septulum absent (Fig. 3a,b,f,h). Claws of Dactylobiotus type (2-1-1-2) with a cuticular bar between the claw bases ( Fig. 4a,b). While claws on the first three pairs of legs sub-equal in length and proportion, claws on the last pairs clearly longer. Accessory points are present on the primary branches of all claws (Fig. 4c,d). Lunules absent. Eggs ( Figure 5; Supplementary Tables S2)  Large, white or slightly yellow, laid freely. Spherical or slightly oval. Transparent or slightly yellow after mounting in Hoyer's medium. Number of processes on the circumference varying from 20 to 37 (Fig. 5a-c). Process in the shape of a cone with circular bases (Fig. 5d), several concentric ring ridges on the surface with a single tip; though some are bi-or multi-furcated (Fig. 5e). Concentric ring ridges on the processes more or less pronounced depending on the level of 'inflation' (Fig. 5d-h). An irregular ring of fine pores encircling the processes, causing slight, uneven ridging of the eggshell between the processes (Fig. 5e). All the egg structures mentioned above are visible under both DIC and SEM observations. Molecular results (Supplementary Table S3) All sequences of Dactylobiotus ovimutans sp. nov. are deposited in GenBank under accession numbers: MT136805 (18S), MT136807 (28S), MT132332 (COI haplotype 1), MT132333 (COI haplotype 2).
We found the same haplotypes for 18S rRNA (1599 bp) and 28S rRNA (533 bp) in two analysed individuals. There are two haplotypes of two bp differences in the partial sequences of COI (673 bp) of 21 individuals and 19 eggs (Supplementary Figure S3). Seven individuals and four eggs belong to haplotype 1, and fourteen individuals and fifteen eggs belong to haplotype 2. The sequence reported for a Dactylobiotus sp. (GenBank accession number EF632526) 64 (collected from moss at Lake Jubany (ca. 62°14ʹ S; 58°40ʹ W), c. 4 Km from Lake CZO), was identical to Dac. ovimutans sp. nov. haplotype 1 (Supplementary Figure S3). The analysis of p-distances for COI involved four nucleotide sequences. There was a total of 518-bp positions in the final dataset. The p-distances for COI between Dac. ovimutans sp. nov. haplotypes was 0.001%; haplotype 1, haplotype 2, Dactylobiotus sp. Remarks. Species which belong to the genus Dactylobiotus are characterized in having the same number of macroplacoids, similar shape of AISM and bulbous expansion, and cuticular bars between claws, but differ mostly in the eggshell morphology 38,39,[65][66][67][68][69] . Nevertheless, some morphological traits may have been overlooked in previous, particularly early, descriptions. For example, FBT is faint or invisible under the Phase-Contrast microscope (PCM) or DIC microscope, thus the presence or absence of this character in other Dactylobiotus species needs to be verified. We also noted that measurements of most of other Dactylobiotus species were based on only few specimens, which hinders understanding the variation of taxonomically important traits. However, the morphometric characters available for these species have enable us to define the new species of Dac. ovimutans sp. nov.
The presence of dorso-lateral papillae between the third and fourth limbs is an important morphological character that distinguishes three species from others within the genus, i.e. Dac. dispar, Dac. selenicus Bertolani, 1982 57 and Dac. parthenogeneticus. In the new species, a dorso-lateral papillae-like structure was observed in only one individual SEM specimen among 9 SEM and 61 DIC specimens. We therefore believe the presence of dorso-lateral papillae in Dac. ovimutans sp. nov. is doubtful. More individuals with a dorso-lateral papillae-like structure need to be studied in order to reveal whether the structure is an artefact or represents different species.
-Differential diagnosis   66 in Dac. ambiguus), with a slightly higher percent ratio of the length of the secondary branch to the primary branch (br) value of claw IV (35.14-51.60 vs. ca. 33.99 in Dac. ambiguus). The closely packed egg processes of Dac. ambiguus produce a strongly polygonal base, while the larger distance between the processes of Dac. ovimutans sp. nov. give a circular to weakly hexagonal base. The outline of the egg processes of Dac. ovimutans sp. nov. is more triangular with a relatively large tip, while that of Dac. ambiguus was described as oval and acuminate 38 .
Dactylobiotus caldarellai (measurement data 72 ), recorded from Punta Arenas, Chile, is similar to the new species in having well-developed oral armature, although it is not clear whether the FBT at the base of the www.nature.com/scientificreports www.nature.com/scientificreports/ peribuccal lamellae is present 50 . Differences are noted between the pt value of stylet support insertion point (SSIP) and a shape of claws: slightly lower pt value of SSIP (67.  Table S4).
Dactylobiotus luci (measurement data 65 ), reported from Uganda, is similar to the new species in having well-developed oral cavity armature, but apparently lacks the FBT at the base of the peribuccal lamellae. The eggs of the new species have broader process base width (10.54-21.00 µm vs. 5.2-7.1 µm in Dac. luci).
Dactylobiotus dervizi (measurement data 67 ), recorded from the Komandorskiye Islands, Russia, is similar to the new species in having well-developed oral armature, but apparently without the FBT (Figures 7 & 8 67 ). Eyes are absent in Dac. dervizi while they are present in the new species. The eggs of the new species have broader process base width (10.54-21.00 µm vs. ca. 9.0 µm in Dac. dervizi). Eggshell has a discrete ring of pores around the process in the new species and random pores between the processes in Dac. dervizi.

Intraspecific variation in the egg morphology and its implications. During the culturing of
Dactylobiotus ovimutans sp. nov., we noticed a significant amount of morphological variation in the eggs, which were easily recognized in the number of processes around the circumference of the egg. The variable egg morphology is independent of the two different COI haplotypes ( Supplementary Fig. S5). While there was no significant difference in egg diameter, the egg processes differed in size and morphology. We found that eggs with fewer processes had triangular processes with concave wall (Fig. 5a), and eggs with more numerous processes had semi-triangular to hemispheric processes with convex wall (Fig. 5b,c). As the number of egg processes increased, the size of the processes become smaller. There were also some embryonated eggs in a pre-deposition state, i.e. with the processes still uninflated and flattened, or partially inflated (Fig. 5f-i). The causes preventing inflation of the processes was not clear. In an experiment with Dactylobiotus sp. from Signy Island lakes to explore desiccation, McInnes et al. 75 did observe collapsed processes under desiccation but flattened side to side, not concertinaed as in the pre-deposition state.
The cause of the difference in eggshell morphology remains elusive. It has been suggested that the dormant and active eggs may show disparate size and morphology (e.g. refs. [83][84][85] ). However, the variable egg types of Dac. ovimutans sp. nov. developed and hatched under the same laboratory condition, so this possibility seems unlikely. Alternatively, different eggs could be formed by seasonality. It has been shown in Bertolanius nebulosus (Dastych, 1983) 86 that through the year long life of an adult, eggs produced in winter had a more enlarged tips of the processes than those produced in summer 41 . However, Dac. ovimutans sp. nov. showed that the variable types of egg were laid under the same, stable laboratory condition (temperature, light and food), excluding the possibility of seasonality. For now, the only possibility that cannot be ruled out is a differential gene expression (epigenetic effect), as proposed by Stec et al. 42 .
the eggs of Dac. ovimutans sp. nov. from other studies. Pilato and Binda 50 suggested that Dac. caldarellai, which they described from Chile, was conspecific with the Dac. ambiguus described by Dastych 34 from KGI, and therefore the eggs would be the same. However, while Dac. caldarellai has fragmented ventrolateral crests 50 , Dastych's Dac. ambiguus from KGI has concatenate ventrolateral crests 34 . In this study, we have observed that the adult morphology of Dac. ovimutans sp. nov. is distinct from both Dac. caldarellai and the specimen from Canada which Pilato & Binda 50 considered to be Dac. ambiguus (original description of Dac. ambiguus 38 is insufficient for comparison). Dastych provided measurements for a single adult Dac. ambiguus reported from KGI 34 , which falls within the measurement range for Dac. ovimutans sp. nov., but the data is too limited to confirm the species. Both Dac. ovimutans sp. nov. and Dastych's KGI Dac. ambiguus 34 have concatenate ventrolateral crests, and the second band of teeth decreasing in size towards mouth opening. However, FBT was not observed in Dastych's Dac. ambiguus 34 . The egg morphology of Dastych's KGI Dac. ambiguus 34 is comparable to that of Dac. ovimutans sp. nov., in having similar size, space between processes and an irregular ring of fine pores encircling the processes (narrow wreath of minute spots around the bases of processes, as described by Dastych 34 ). In contrast, the egg morphology of the original Dac. ambiguus 38 differs because the processes are packed without space. These facts imply that the eggs from Dastych's KGI samples 34 might belong to neither Dac. caldarellai nor Dac. ambiguus, but to Dac. ovimutans sp. nov.
McInnes 47 recognized two types of eggs of Dac. ambiguus from Signy Island, Antarctica and noted that the first type had many processes in which the process bases were connected, while the second type had fewer unconnected processes. The overall morphology of the second type of eggs is reminiscent of the eggs of Dac. ovimutans sp. nov., such as egg size, processes with several concentric ring ridges on the surfaces, space between the processes and fine pore rings encircling the processes (see: Supplementary Fig. S6, Supplementary Table S5). Therefore, the second type eggs of Dac. ambiguus from Signy Island could in fact be the eggs of Dac. ovimutans sp. nov. Taking into account high dispersal abilities of tardigrades by wind 87-89 as well as birds 35 www.nature.com/scientificreports www.nature.com/scientificreports/ of resting forms being transported between Antarctic islands is plausible. Hence, the similarity of morphology of egg type described for Signy Island Dac. ambiguus indicates that these specimens should be considered as Dactylobiotus cf. ovimutans sp. nov, but further analysis is required for confirmation.

Diet of Dac. ovimutans sp. nov.
Unlike other cultures of Dactylobiotus species that were given a herbivorous diet (e.g. Dac. dispar 91 and Dac. parthenogeneticus 92 ), Dac. ovimutans sp. nov. fed on rotifers from the Lake CZO. Under the culture feeding on rotifers only, Dac. ovimutans sp. nov. were vigorous and laid many eggs with varying morphology. We recorded Dac. ovimutans sp. nov. feeding on rotifers (Supplementary Movie S1) and noted that after ingesting rotifers the gut of Dac. ovimutans sp. nov. turned a distinctive shade of yellow. However, in the absence of rotifers and cultured on sediments containing organic material from lake CZO, the gut of Dac. ovimutans sp. nov. turned green or brown, suggesting ingestion of algae or cyanobacteria. Dactylobiotus feeding on a nematode Plectus sp. was observed from a culture of Signy Island lake sediments with a mix of native species (McInnes pers. obs.). Therefore, we conclude that Dac. ovimutans sp. nov. is omnivorous.

Summary. A new species Dactylobiotus ovimutans from King George Island, Antarctica is distinguished from
Dac. ambiguus and Dac. caldarellai in having different claw morphology. The presence of Dac. ambiguus and Dac. caldarellai on KGI and other regions of the maritime Antarctic should be considered dubious, and/or requires further detailed, integrated taxonomic analysis. Culturing of Dac. ovimutans sp. nov. revealed a significant variation in the eggshell morphology, which was recognized by differences in the number and size of the processes. The eggs with fewer processes had larger and concave process profiles, while those with more numerous processes were smaller and convex process profile. Some eggs retained uninflated processes as in the pre-deposition state. Since all eggs were laid in the same stable laboratory condition, such variation is considered to be caused by epigenetic effects, and not being subject to different temperature, food source and seasonality.

Sample & Specimens. The KOPRI (Korea Polar Research Institute) ecology team near the King Sejong
Station, KGI, Antarctica (62°14′24.140″S, 58°44′36.571″W), collected benthic sediment samples from the Lake CZO. The sediment contained organic material, i.e. algae and moss. Samples were brought to KOPRI (Incheon, Korea), stored in 4 °C for several months, and eight tardigrades were extracted using a stereomicroscope (Leica M205C). Individuals were cultured separately at 11 °C on a 1.5% bacto-agar plate with Volvic ® water, with food source provided as rotifers and algae collected from the King George Island moss samples.
For light microscopic observation, tardigrades were prepared using thermal relaxation by incubating live individuals at 60 °C for 30 min (following ref. 93 ), and mounted on Higgins-Shirayama slide (HS-slide) 94 in Hoyer's medium. After drying seven days at 60 °C, the slides were sealed with nail polish, and examined under a Differential Interference Contrast (DIC) microscope (Carl Zeiss Axio Imager 2), with the camera AxioCam HRc. Additional tardigrades were prepared for SEM using thermal relaxation and the method in ref. 95 . SEM observations were made using a JEOL JSM-6610 and Field Emission SEM JSM-7200F, at KOPRI for KGI specimens. SEM observations of eggs from Signy Island have been conducted in British Antarctic Survey (BAS) following ref. 75 .
Twenty-one adult tardigrades and nineteen eggs, from the cultured animals, were selected randomly and the identities were checked under a stereo microscope. Then we extracted DNA from these individuals and eggs using commercial kits (TIANamp Micro DNA Kit and QIAamp DNA Micro Kit).
For classical taxonomic observations, we examined 61 individuals and 59 egg specimens under DIC, and 9 adult and 14 egg specimens under SEM. All specimens from this study were registered with ATNS (Antarctic Tardigrade Name of Specimen) numbers and stored in KOPRI.
Morphometrics. Character selection and measurements followed Binda & Pilato 72 . All measurements are in micrometres (μm). Characters were measured when the specimens were presented in the correct orientation on the slide. Body length was measured from the anterior tip to the end of the body, excluding the leg IV, and pt index is the percent ratio of the length of a character to the length of buccal tube 96 . In the claw measurements, the br index is the percent ratio of the length of the secondary branch to the length of the primary branch 56 . Allometric measurements were calculated with the exponent (b) and the Y-intercept (a*) of the regression of Thorpe's normalized characters versus body size 97 . Most of the morphological terminology follows Ramazzotti & Maucci 98 . However, several characters have been more recent revisions 53,99,100 , i.e. 'the first band of teeth' , 'the second band of teeth' , 'oval perforated area' , 'apophysis for the insertion of the stylet muscles' , and the claw's 'cuticular bar' . Molecular data. DNA was extracted from two individuals using TIANamp Micro DNA kit, and nineteen individuals and nineteen eggs using QIAamp DNA Micro Kit. We acquired three DNA partial genes from two individuals: 18S rRNA, 28S rRNA, and COI, and acquired COI partial gene sequences from nineteen individuals and nineteen eggs. The PCR mixture was prepared in a total volume 50 μl with Takara EmeraldAmp ® GT PCR Master Mix 25 μl, 0.5 μl of each primer, 2 μl of DNA template and 22 μl of ddH 2 O. The primers and PCR programmes are in Supplementary Table S6 & S7. PCR products were sent to the commercial company for sequencing (Macrogen, Korea). To calculate molecular distances for the COI fragments, two sequences were obtained from GenBank for Dactylobiotus (GenBank accession number EF632526 64 , AY598771 58 ). Sequences were processed in BioEdit ver.7.2.5 101 . Pairwise distances between nucleotide sequences were calculated using a distance model for all codon positions as implemented in MEGA X 102 . p-distance calculations for all positions containing gaps and missing data were eliminated. www.nature.com/scientificreports www.nature.com/scientificreports/ ZooBank LSIDs (Life Science Identifiers) can be resolved and the associated information viewed through any standard web browser by appending the LSID to the prefix 'http://zoobank.org/' . The LSID for this publication is: urn:lsid:zoobank.org:pub:362EE4A0-A6F2-4F78-AF65-EA87028EA1DD.

Data availability
The datasets generated during and/or analysed during the current study are available from the corresponding author on reasonable request.