﻿New psychropotid species (Echinodermata, Holothuroidea, Elasipodida) of the Western Pacific with phylogenetic analyses

﻿Abstract Holothurians of the family Psychropotidae are widely distributed but remain the least studied deep-sea holothurians. On an expedition to the Western Pacific, six psychropotid specimens were collected by the Jiaolong Human Operated Vehicle (HOV). Through morphological examination, four of them were identified as a new species, Benthodytesjiaolongisp. nov., which was characterized as having minute papillae, a narrow brim, and a terminal anus; and the ossicles were rods and primary crosses. The remaining two specimens were identified as Psychropotesverrucicaudatus Xiao, Gong, Kou & Li, 2019, first recorded at the Kyushu-Palau Ridge. The phylogenetic analysis showed that B.jiaolongisp. nov. and P.verrucicaudatus were embedded in the clades Benthodytes and Psycheotrephes, respectively, and that Benthodytes was paraphyletic. The new species clustered with Benthodytessanguinolenta and was separated from the clade containing the other Benthodytes species.


Introduction
Holothurians of the family Psychropotidae (Elasipodida) were first identified by Théel (1882) who defined four genera of deep-sea sea cucumbers discovered on the H.M.S. Challenger Expedition. Subsequently, Hérouard (1909) and Belyaev and Vinogradov (1969) erected Triconus Hérouard and Nectothuria Belyaev & Vinogradov, which were later regarded as synonyms of Psychropotes by Hansen (1975). Meanwhile, Euphronides Théel, 1882 was also accepted as a synonym of Psychropotes. Psychropotidae comprises three genera and 37 species. Hansen (1975) distinguished the three genera by the presence or absence of an unpaired dorsal appendage, the position of the anus, and the presence or absence of circum-oral (or post-oral) papillae. Although, taxonomists have long worked on this family, Psychropotidae are still the least studied deep-sea holothurians. Thus, the phylogenetic relationships within Psychropotidae remain unclear.
An expedition of the Jiaolong Human Operated Vehicle (HOV) concentrated on further increasing our understanding of the biodiversity, connectivity, and conservation value of the Western Pacific. During sampling, six specimens of Psychropotidae were collected from seamounts on the Kyushu-Palau Ridge and Weijia Guyot (Fig. 1). Based on an analysis of the external morphological characters and ossicles, we identified four specimens as a new species (Benthodytes jiaolongi sp. nov.) and the other two as new records of Psychropotes verrucicaudatus Xiao, Gong, Kou & Li, 2019.

Sampling and morphological observations
The samples described in the present study were collected by the Jiaolong HOV at a depth of 2408-2602 m, from the Kyushu-Palau Ridge and Weijia Guyot. Before preservation, a Canon EOS 5DII camera (Canon Inc., Tokyo, Japan) was used to take photographs of the specimens on board the ship. Then, a piece of dorsal tissue was cut from all specimens and frozen at -20 °C for DNA extraction. Finally, the specimens were fixed in 10% seawater formalin or 99% alcohol and deposited at the Repository of Second Institute of Oceanography (RSIO). Sodium hypochlorite was used to dissolve body tissues (tentacles, dorsum, ventrum, brim, dorsal warts and gonads), and ossicles present in these tissues were rinsed five times with purified water. The ossicles were observed using a scanning electron microscope (TM 1000; Hitachi, Ltd., Tokyo, Japan).
For a more comprehensive phylogenetic analysis, we not only used the sequences of Psychropotidae obtained here, but also used mitochondrial sequences of Elpidiidae Théel, 1882 and two species of Stichopodidae Haeckel, 1886, an outgroup (Table 2). Twenty-five COI and 18 16S sequences were aligned using MAFFT 7 (Katoh and Standley 2013) using the E-INS-I strategy. Alignment gaps and missing data were represented as '-' and '?'. The 16S and COI alignments were concatenated (COI/16S = 687/578 bp), analyzed with Maximum likelihood (ML) and Bayesian inference (BI) algorithms. JModelTest 2.1.10 (Darriba et al. 2012) was used to find the best-fit model from 88 competing models using Akaike information criterion (AIC) calculations. In each case, GTR+I+G was the best-fit model for BI analyses. MrBayes 3.2 (Huelsenbeck and Ronquist 2001) was used to conduct BI analyses. Markov Chain Monte Carlo (MCMC) iterations were run for 1 000 000 generations with sampling every 100 generations. The first 25% of trees were discarded as burn-in, and the consensus trees were summarized in 75% majority-rule trees. RAxML GUI 1.5 (Silvestro and Michalak 2012;Stamatakis 2014) was used to perform the ML analysis with the GTR+GAMMA+I substitution model for 1000 bootstraps, as recommended by Miller et al. (2017).   2C); skin transparent, thin, soft, and gelatinous after fixing. Brim retracted less than 0.7 cm in width. Approximately nineteen pairs of dorsal papillae poorly developed, minute, closely placed in two bands along anterior dorsal radii. Another four single minute papillae on posterior dorsal edge. Approximately 28 pairs midventral tube feet arranged in two rows. Mouth ventral, with circum-oral papillae. Anus terminal, unguarded. Due to the contraction, tentacles could not be clearly observed. Few ossicles observed. Dorsal ossicles in the anterior body wall, consisting of primary crosses with spiny arms, and spinous rods (Fig. 4A-F). Rods approximately 400 µm long, arms of crosses approximately 200 µm long. Tentacles with rods, 400-500 µm long ( Fig. 4G-J). Other body parts devoid of ossicles.
Description of paratypes. RSIO3710601. Specimen approximately 22 cm in length, 5 cm wide at maximum point. Color red-violet in situ at the seabed ( Fig. 2A, B); pale violet at sea surface, with transparent skin; white color after preservation in 10% seawater formalin for 5 years. Paired dorsal papillae as present in holotype absent, minute papillae also not distinguished. Owing to long-term preservation, quantity of midventral tube feet could not be determined, but were arranged in two rows. Brim could not be distinguished. Mouth ventral, with circum-oral papillae, anus terminal. Eighteen tentacles retracted to stalk. Ossicles not observed.
RSIO590504.Specimen approximately 22 cm in length before preservation in 10% seawater formalin. Color red-violet on deck, skin transparent; white color after preservation. During sampling, a piece of sponge was stuck in the ROV pump sampler, and the specimen was damaged by the sponge, meaning that the tentacles could not be determined and the dorsal tips could not be distinguished. Quantity of midventral tube feet could not be determined. Mouth ventral, anus terminal. Ossicles not observed.
RSIO590506. Specimen approximately 13 cm in length before preservation in 99% alcohol and heavily damaged. Color red-violet at sea surface, skin transparent. The specimen was stained with sponge as was RSIO590504 and many external characters could not be distinguished. Mouth ventral, anus terminal. Few rods observed on dorsal region (Fig. 4A Remarks. Hansen (1975) revised the genus Benthodytes and proposed that this genus, except Benthodyets superba Koehler & Vaney, 1905, could be divided into two distinct groups based on the ossicles and external morphology.
Benthodytes typica was described by Théel in 1882 based on specimens collected by the Challenger Expedition. The original description indicated approximately eight, minute, retractile processes located on each of the dorsal ambulacra and unbranched spinose calcareous spicula scattered on the integument. Hansen (1975) re-examined B. typica and reported that the specimens showed considerable variation. Benthodytes papillifera Théel, 1882 was described based on 13 specimens taken from three Pacific Challenger stations. Théel (1882) described this species as being similar to B. sanguinolenta based on the tentacles and tube feet. Hansen (1975) re-examined specimens from each of the stations and proposed that the variation in B. papillifera represented the geographic variation of B. typica. In the original description of Benthodytes glutinosa Perrier, 1896, Perrier (1896 indicated that the differences from B. typica were the more elongated shape and the complete absence of dorsal papillae. Hansen (1975) considered this species to be a synonym of B. typica.
In general, the morphological features of B. typica can be summarized as follows: 3-7 pairs of minute papillae arranged on the dorsal surface and rods scattered on the The characteristics of B. sanguinolenta as described by Théel (1882) included the many minute retractile processes scattered on the dorsal surface; the form of calcareous deposits could not be distinguished. According to a re-examination by Hansen (1975), the dorsal minute papillae were arranged in two radial bands and the rods were only present on the midventral tube feet and tentacle discs of specimens from station 663. Rogacheva et al. (2009) recorded B. sanguinolenta and the main characteristics can be described as: minute dorsal papillae arranged in two bands or between the two bands; approximately 1-4 papillae placed in a band, narrowing to one or two papillae at the posterior end; ossicles were not found. The differences in the characteristics between the new species B. jiaolongi sp. nov. and B. sanguinolenta can be listed as follows: (1) Dorsal papillae of B. sanguinolenta are arranged in two bands, whereas those of the new species were arranged in two rows on the anterior dorsal ambulacra; (2) Ossicles of the new species were only present in the tentacles and in the dorsum. Rods are present in the tentacles and dorsum, and primary crosses are only present in the dorsum; whereas the rods are only present in tube feet and in the tentacles in B. sanguinolenta. Description. RSIO6018004. Specimen resembles a barbell after collection, approximately 20 cm in length before preservation 10% seawater formalin (Fig. 5C, D). Before preservation, height of appendage was approximately 50 mm, and width at base approximately 30 mm (Fig. 5C, D). Dorsal skin transparent with brownish red color on seabed and dark brown on deck. Warts covering dorsal skin and appendage; giant ossicles in warts visible (Fig. 5E, F). Approximately 30 pairs of degenerated tube feet arranged in two rows along middle of ventrum. Sixteen tentacles forming a circle. Brim broad and covered with warts on dorsum.

Genus
A giant cross with four arms visible in each wart. Arms 800-1000 µm in length, and maximum width between large arms approximately 500 µm. Arm flexion approximately 250 / 400 µm (Fig. 6A-D). Height of central rudimentary apophyses approximately 200-300 µm. Ventral ossicles divided into two types: primary cross with spiny arms (Fig. 8A, C) and cross with three arms (Fig. 7B), length of arm approximately 200 µm. Primary crosses with spinous arms in dorsum (Fig. 7D-F) and brim (Fig. 7H-J); arms up to 200 µm in length. Dorsal ossicles with spinous rod, 170 µm in length (Fig. 7G), and large primary crosses with spiny arms in brim (Fig. 7K). Tentacles with rods with irregular shape (Fig. 7L-R). Large rod with two apophyses at the end, approximately 900-1000 µm in length (Fig. 7L-M); small rod with apophyses in middle area was approximately 200 µm in length (Fig. 7N). Other rods with spiny arms, 500-800 µm in length (Fig. 7O-R Intraspecific variation. The specimens were clearly a new record for the South China Sea, as the species was previously known only from the Jiaolong seamount. The present specimens differed from those of Xiao et al. (2019) in external morphology and the ossicles. Due to the bad preservation, Xiao et al. (2019) could not observe the ossicle assemblage of the warts, which was possible in the specimens here under study.
The intraspecific differences can be listed as follows: (1) In the present specimens, the skin was transparent and the color was darker than that of the type specimen; (2) The width of the appendage at the base was also larger than that of the type specimen; (3) The length of the primary crossing arms distributed in the dorsum, ventrum, and brim was longer than that of the type specimen. Furthermore, the spinous rod of the dorsal ossicles was not present in the type specimen, and the ventral body wall of the specimens did not possess the tripartite ossicles of the type specimens; and (4) Most of the ossicles of the tentacles in our specimens were the same as those of type specimen, but longer.

Phylogenetic analyses
Owing to limited genetic sequences, the phylogenetic relationships of Elasipodida remains little studied. The new classification system of Elasipodida was constructed by Miller et al. (2017), whereby Deimatidae was separated from Elasipodida. The remaining families of Elasipodida included Elpidiidae, Laetmogonidae, Pelagothuriidae, and Psychropotidae, but their positions within Elasipodida remained unresolved.  used mitochondrial and nuclear genes to perform phylogenetic analyses of Elasipodida, especially the Psychropotidae, and the results showed that Benthodytes was a paraphyletic group of Psychropotidae based on analyses of the mitochondrial genes.
To obtain clearer phylogenetic relationships, we concatenated 25 COI and 18 16S sequences into a dataset to build ML and BI trees. Although the genetic sequences were limited, the topological structures of the ML and BI trees were mostly consistent with morphological classification. In addition, B. jiaolongi sp. nov. and P. verrucicaudatus were embedded in the clades of Benthodytes and Psycheotrephes, respectively (Fig. 8). The phylogenetic relationships of Psychropotidae clustered into four parts and were inconsistent with the traditional classification system (Hansen 1975). Benthodytes was divided into two clades in Psychropotidae and the new species was clustered with the clade of B. sanguinolenta. In addition, Psychropotes was a sister group to Psycheotrephes, and the clade of B. jiaolongi sp. nov. and B. sanguinolenta was a sister group to other Psychropotidae species. Psychropotes verrucicaudatus was not recovered in the clades of Psychropotes, but was clustered in the clade of Psycheotrephes exigua Théel, 1882, which meant that P. verrucicaudatus might belong to Psychrotrephes. Elpidiidae clustered into two clades: (1) Penigone clustered together into a supported group, but Peniagone diaphana Théel, 1882 was a sister group to other Peniagone species; (2) The other four genera of Elpidiidae clustered into a group, and Elpidia glacialis Théel, 1876 was distant from the other three genera. Protelpidia murrayi Théel, 1879 and Scotoplanes sp. TT 2017 were sister taxa, and Amperima robusta Théel, 1882 was sister to these genera.
Laetmogonidae was an obvious polyphyletic group, and Pannychia moseleyi Théel, 1882 was placed in the outmost clade of the other three families. Laetmogone wyvillethomsoni Théel, 1879 clustered with Elpidiidae and was sister to this clade; Benthogone abstrusa Sluiter, 1901 was clustered with Enypniastes eximia Sluiter, 1901, but the Bayesian posterior probabilities and bootstrap values of this clade were low.
Based on the morphological and phylogenetic analyses, B. jiaolongi sp. nov. can be identified as a new species closely related to B. sanguinolenta. In addition, our specimens provided a new record of P. verrucicaudatus in the Western Pacific, broadening its distribution. Our results support the hypothesis that Benthodytes is paraphyletic and that the clade of B. sanguinolenta and B. jiaolongi sp. nov. is separated from the other species of Benthodytes.