The sea cucumber sample (Fig. 1) was collected by diving from the coastal waters of Wenchang in China in August 2022 and was transported alive to the laboratory in fresh seawater. The sample was identified based on morphological characteristics and DNA barcoding before alcohol preservation, and the individual was chosen for mitogenome sequencing. Muscle tissues were placed in 95% ethanol at -20 °C for further study.

Figure 1. 

Dorsal (A) and ventral (B) views of H. ocellata.

To verify the identity of the specimen, both optical microscopy and scanning electron microscopy were used to observe and analyze the types and sizes of ossicles in the body wall, tentacle, tube feet, and papillae. The ossicles were extracted using the method modified by Teo and Ng (2009). Three tissue pieces weighing about 1 g were collected from each target site and rinsed with distilled water. Then, 8% sodium hypochlorite was added to each sample and shocked to digest for approximately 3–5 min. After white particles appeared at the bottom of the tube, undigested tissue pieces were picked out and left to stand for 5 minutes. An appropriate amount of turbid liquid was absorbed with white particles and placed on a slide. We used the light microscope to preliminarily observe these samples and confirmed different types of ossicles. Then, the other turbid liquid was dropped on the conductive tape and stoved in a drying oven, followed by gold spray treatment with an ion sputtering apparatus. A scanning electron microscope (SEM; HITACHI, S-4800) was used to observe and photograph these samples. The size of the ossicles in the photographs was measured by the soft ruler and converted to actual lengths based on the scale bars. Ten individuals from each ossicle were randomly selected for measurement. The calcareous ring was observed and depicted with the stereoscopic microscope. The size of the calcareous ring in the photographs was measured by the soft ruler and converted to actual lengths based on the scale bars.

According to the manufacturer’s protocol, the total DNA of the sea cucumber was extracted using the TIANamp Marine Animal DNA Kit (TIANGEN, Beijing, China). The DNA library was sequenced by Origingene Co., Ltd. (Shanghai, China) using the Illumina NovaSeq^TM 6000 platform with an insert size of 300–500 bp. Approximately 9.09 GB of raw data from Holothuria ocellata were generated with 150-base-pair paired-end read lengths. The quality control and assessment of data were used by Cutadapt v1.16 and FASTQC v0.11.4, respectively (Martin 2011). The mitogenome was assembled by NOVOPLASTY v4.2 with default parameters (Dierckxsens et al. 2017).

The mitogenome of H. ocellata was annotated by the MITOS WEBSERVER (Donath et al. 2019) (http://mitos2.bioinf.uni-leipzig.de/index.py). The graphical circle map of the mitogenome was drawn using the online PROKSEE tool (Stothard et al. 2019) (https://proksee.ca/). The relative synonymous codon usage (RSCU) was conducted using CODONW v1.4.4 (http://codonw.sourceforge.net). The blank region of the mitogenome was defined as the control region (D-loop) and compared with the mitogenome of the reference species to determine the control region. The TRNASCAN-SE tool was used to identify the tRNA gene and predict the secondary structure diagram of tRNA (Chan and Lowe 2019). The composition skewness of each segment was calculated by the following formulas: AT-skew = (A − T) / (A + T); GC-skew = (G − C) / (G + C) (Perna and Kocher 1995).

Besides the sequence obtained in this study, mitogenomes of 35 previously sequenced sea cucumber species in the class Holothuroidea and 2 previously sequenced starfish species in the class Asteroidea (with the latter as the outgroup taxon) were used in the phylogenetic analyses. We extracted the nucleotide sequences of the 13 protein-coding genes from each mitogenome as the dataset to construct the phylogenetic tree. Sequences were extracted and concatenated by PHYLOSUITE v1.2.3 (Zhang et al. 2020; Xiang et al. 2023), and these sequences were aligned and refined through MAFFT v7.505 and MACSE v2.06 (Katoh and Standley 2013; Ranwez et al. 2018). Ambiguously aligned fragments of 13 alignments were removed in batches based on GBLOCKS with the default parameter settings (Talavera and Castresana 2007). MODELFINDER v2.2.0 was used to select the best-fit model through the BIC criterion (Kalyaanamoorthy et al. 2017). Maximum likelihood (ML) phylogenies were inferred by IQ-TREE v2.2.0 under the model automatically selected by the ‘Auto’ option for 20000 ultrafast bootstraps (Guindon et al. 2010; Minh et al. 2013; Nguyen et al. 2015). Bayesian inference (BI) phylogenies were inferred through MRBAYES v3.2.7a under the GTR+I+G+F model (2 parallel runs, 200000 generations), with the initial 25% of sampled data discarded as burn-in (Ronquist et al. 2012).

To determine whether the species belongs to the genus Bohadschia or to Holothuria, we selected the mitochondrial sequence fragments of all Bohadschia and some Holothuria from NCBI to perform the phylogenetic tree. 13 rrnL genes and 9 cox1 genes, including sequences from this study, were selected to build Neighbor-Joining (NJ) phylogenetic trees based on two genes, respectively. Sequences were aligned by SEQMAN from DNASTAR software (USA). The trees were constructed, and the distances between and within groups were determined by MEGA v5.0 (Tamura et al. 2011) with 1000 bootstrap replicates based on genetic distances calculated with the Kimura-2-parameter (K2P) model by the same software.

The specimen, in the state of natural extension, is 24.35 ± 0.3 cm long and 5 ± 0.2 cm wide. The body is cylindrical, tapering at both ends. The bivium presents convex (Fig. 1A), while the trivium presents flattened (Fig. 1B). The mouth is ventral, and the anus is located in a terminal. The dorsal color is grayish on both sides and yellowish-brown in the middle, interspersed with tan spots. The whole dorsal surface is covered with conic papillae, which are milky at the tip and brown at the base. Each papilla is usually surrounded by a narrow light-grey band and a white band from inside to outside, although the white band may be absent or non-significant in some papillae. The ventral has a lighter color and slightly flatter shape than the dorsal. The ventral papillae are arranged in 3–4 bands, concentrated in the middle of the trivium, showing a trend of spreading to the interambulacrum on both sides. Papillae on the sides of the ventral are prominent in size, arranged along the sides of the body to form the dorsoventral junction with a number of 41 or 42 on each side.

The body wall of the specimen has four major ossicle types, including multiple buttons, C-shaped ossicles, rods, and tables. Ventral/dorsal body walls have similar table and button types. The majority of the buttons are knobbed, yet the minority are smooth and nearly oval, usually with 3 pairs of holes, 30–68 µm in length and 30–40 µm in width (Fig. 2A, B). Some buttons are long ellipsoids with up to 4–7 pairs of holes, 63–88 µm in length, and 27–38 µm in width (Fig. 2C, D). The paired holes of most buttons are symmetrical and similar in size to each other. Tables are intact with nearly circular discs with smooth or wavy margins. There are basically two types of typical tables: 1) the disc is convex with a larger size of 81–136 µm in diameter, with dozens of holes in the circumferential margin; the spire is moderate in length, relatively thin, and terminating in a cluster of spines (Fig. 2E, F); 2) the disc is convex with a smaller size, 64–82 µm in diameter, with 8–12 holes in the circumferential margin; the spire is moderate in length, relatively thick, and terminating in a cluster of spines (Fig. 2G, H). The rods in the body wall are commonly composed of simple spicules, which are smooth and elongated, with a wide range of length variations, 131–496 µm long (Fig. 2I). The C-shaped ossicles in the body wall are commonly C-form, smooth, and elongated, with a wide range of length variations, 200–300 µm long (Fig. 2K). Tentacles with spiny rods, 232–376 µm long (Fig. 2P–R).

Figure 2. 

H. ocellata: scanning electron microscopy images of ossicles. A, B. Buttons with 3 pairs of holes; C, D. Buttons with 4–7 pairs of holes; E, F. Tables with larger discs; G, H. Tables with smaller discs; I, J. Rods in papillae, tube feet, or body wall; K. C-shaped ossicles in the body wall; L-O. Perforated plates of papillae, or tube feet; P–R. Rods in tentacles; S, T. Tack-like tables of papillae.

Ossicles of papillae and tube feet include rods, perforated plates, and tables. The morphology of rods is diverse and complex, with spines present or not and thicknesses different (Fig. 2I–J). Most perforated plates, measuring up to 200 µm in length, have perforations in the middle, and some plates are perforated throughout their whole length (Fig. 2L–O). The type of tables in papillae is basically consistent with those of the body wall or tube feet, except for the tack-like tables. The tack-like tables, only found in papillae, with concave bottoms, are few in number, standing up to 240 µm high, and have a tall attenuating spire with several cross-beams, resembling the shape of a nail (Fig. 2S, T).

Calcareous ring, with radial plates obviously larger than interradial ones; radial plates with anterior concave and obviously posterior bifurcation; interradial plates with cuspate anteriorly (Suppl. material 1).

The complete mitochondrial genome of the sea cucumber (GenBank No. OQ731944) is 15,797 bp in length (Fig. 3). The mitogenome contains 37 typical genes (13 PCGs, 22 tRNAs, and 2 rRNAs) and 2 control regions. Most mitochondrial genes are encoded on the H-strand, except for nad6 and 5 tRNA genes (trnS2, trnQ, trnA, trnV, and trnD) that are encoded on the L-strand (Table 1). The nucleotide composition of mitogenomes has a higher A+T bias of 59.01%, and all PCGs show negative GC-skew except for nad6 (Table 2).

Figure 3. 

Mitochondrial genome maps of H. ocellata. Genes encoded on the heavy or light strands are shown outside or inside the circular gene map, respectively.

The number of PCGs in mitogenomes is consistent with the general findings on the mitogenomes in Holothuriidae species. Twelve PCGs (cox1, cox2, cox3, atp8, atp6, nad1, nad2, nad3, nad4, nad4l, nad5, and cob) are coded on the heavy strand (H-strand), while the remaining one (nad5) is coded on the light strand (L-strand). All 13 PCGs collectively encode 3696 amino acids. All PCGs use the initiation codon ATG. The termination codons TAA and TAG are commonly observed, although the incomplete termination codon T is found in cox2 and nad4 in the mitogenome.

We calculated the relative synonymous codon usage (RSCU) of the mitogenome (Table 3, Fig. 4), and the results showed that the frequency of NNA and NNC (N represents A, T, C, G) is higher than NNT and NNG. The most frequent amino acids in the coding sequences of mitochondrial proteins are Leu1, Phe, and Ile (> 290) (Table 3). Moreover, the minimally used amino acid in the mitogenomes is Trp (< 30).

Figure 4. 

The relative synonymous codon usage (RSCU) in H. ocellata mitogenome. Codon families are labeled on the x-axis. The termination codon is not given.

Similar to the most holothuroids, the mitogenome of Holothuria ocellata has one rrnS (12S rRNA) and one rrnL (16S rRNA) gene. The rrnS gene is located between trnF and trnE, and the rrnL gene is located between nad2 and cox1. In the H. ocellata mitogenome, the A+T content of rRNA is 59.82%. The AT-skew of rRNA is strongly positive, whereas the GC-skew is slightly negative, indicating that the contents of A and C are higher than those of T and G in the rRNA, respectively. And there are 22 tRNA genes. The secondary clover-leaf structures of tRNA genes identified in the mitogenome are shown in Fig. 5. These tRNA genes vary in length from 57 bp to 75 bp. All predicted tRNAs display the typical clover-leaf secondary structure, except for trnS1.

Figure 5. 

Inferred secondary structures of the 22 tRNA genes of H. ocellata mitogenomes.

Phylogenetic relationships are constructed based on the sequences of 13 PCGs of 38 mitogenomes using BI and ML methods. The phylogenetic trees constructed by the two methods are consistent with high intermediate bootstrap values, and the topological structure of the trees is entirely the same (Fig. 6A). The results showed that 10 Holothuria sequences formed a monophyletic group in both ML and BI analyses. Moreover, H. ocellata formed a sister group with Holothuria (Theelothuria) spinifera Théel, 1886, and both divided with Bohadschia argus Jaeger, 1833.

Figure 6. 

A. Phylogenetic tree of 36 Holothuroidea sequences constructed by Bayesian inference (BI) and maximum likelihood (ML) methods based on concatenated sequences of 13 PCGs. Asteroidea species were used as the outgroup; B. Phylogenetic tree of 12 Holothuria and Bohadschia sequences constructed by Neighbor-Joining (NJ) methods based on concatenated sequences of the rrnL gene; C. Phylogenetic tree of 9 Holothuria and Bohadschia sequences constructed by Neighbor-Joining (NJ) methods based on concatenated sequences of the cox1 gene. The species in the red frame indicates the sequences generated in this study.

Using the NJ method, the unrooted phylogenetic trees of the rrnL and cox1 genes both indicated that Holothuria and Bohadschia species form a monophyletic group, respectively (Fig. 6B, C). Holothuria and Bohadschia sequences are observably divided from each other. Besides, the tree of rrnL showed the Holothuria ocellata sequence obtained in this study clusters with the H. ocellata sequence from NCBI. And the tree of cox1 showed the H. ocellata sequence is divided with Bohadschia ocellata sequences from NCBI.

vs. B. ocellata Jaeger, 1833

The controversy regarding the nomenclature of two sea cucumber species (H. ocellata / B. ocellata) stems from the blurry original nomenclators. Jaeger (1833) first found a sea cucumber species and classified it into the subgenus Holothuria, tribes Bohadschia, the latter being subordinate to the former, and named this species “ocellata.” With the continuous progress of sea cucumber taxonomy, both Holothuria and Bohadschia have evolved into independent and valid genera (WoRMs 2023b, c). Due to the lack of an initial description and the continuous change of classification status, subsequent scholars had different opinions and controversies about this sea cucumber species. Some scholars continued to use H. ocellata as the name of the species they observed (Semper 1868; Pearson 1913; Heding 1939; Clark 1946; Rowe 1969; Liao 1980; Liao 1997; Teo and Ng 2009; Kamarudin et al. 2010). Conversely, some scholars used B. ocellata as the name of the species they found (Kim et al. 2013; Patantis et al. 2019; Javanmardi et al. 2020).

In this study, based on morphological characters and mitochondrial genome sequences, we believe that the two current Latin names, “Holothuria ocellata” and “Bohadschia ocellata,” represent two different species, respectively. Firstly, compared with the descriptions and figures of B. ocellata provided by Purcell et al. (2023), these two species have obviously different ossicles and spots: 1) The types of ossicles in H. ocellata include buttons and tables, while B. ocellata has none of these; 2) Compared to H. ocellata, the spots of B. ocellata are irregular and differ greatly in size or shape. Finally, after identifying this classification problem between H. ocellata and B. ocellata, we subsequently selected the mitochondrial sequences of all Bohadschia and some Holothuria from NCBI and performed two phylogenetic trees based on rrnL and cox1 genes. The result showed a clear separation between the sequences of H. ocellata and B. ocellata. Furthermore, the phylogenetic tree based on the sequences of 13 PCGs also proved that H. ocellata does not belong to the genus Bohadschia. For the above reasons, rather than synonyms, it is reasonable to consider H. ocellata and B. ocellata as two different species.

vs. H. kurti Ludwig, 1891

Due to the similar external morphology between H. kurti and a small specimen of H. ocellata, some scholars have regarded them as the same species (Liao and Clark 1995; Liao 1997). The sea cucumber H. kurti was first found by Sluiter (1889) and initially named “Holothuria lamperti Sluiter, 1889” (Ludwig 1891). Because of a naming duplication with “Holothuria lamperti Ludwig, 1886,” Ludwig (1891) subsequently renamed the sea cucumber species as “Holothuria kurti Ludwig, 1891.” In 1901, Sluiter described the ossicles of H. kurti for the first time and identified a table with cross-shaped chassis (synallactid-type table) as the characteristic of this species. Pearson (1913) compared the ossicles between H. kurti and H. ocellata, utilizing the differences in the types of tables as the basis for distinguishing the two species. Liao (1980, 1997) suggested that H. kurti might be the juvenile of H. ocellata, as he observed that the juveniles of H. ocellata had a similar synallactid-type table. In the meanwhile, he also found some large synallactid-type tables in the adults of H. ocellata. In previous studies, some sea cucumber species were also found to have their ossicles of adult individuals that were quite different from juvenile individuals (Gosliner et al. 1990; Massin et al. 2000; Soliman et al. 2019), which seemed to verify the credibility of Liao’s conclusion. However, Samyn and Vandenspiegel (2016) disagreed with the conclusion; they collected some specimens of H. kurti and found that these samples were over 7 cm long and had mature gonads; they thought that H. kurti and H. ocellata were two different species. Besides, it is indeed true that scholars did not find tack-like tables in the papillae of H. kurti, which were found in the sample we collected in this study (Fig. 2O, P). For the above reasons, we agree with the view of Samyn and Vandenspiegel (2016). It is reasonable to consider that H. ocellata and H. kurti are different species. We also hope that more molecular biological data on H. kurti will be available in the future to verify and support this conclusion.

vs. H. hamata Pearson, 1913

The samples obtained in this study were similar in external form and ossicle types to the species found by Teo and Ng (2009), which they thought these samples were H. ocellata. However, some authors suggested that Teo and Ng’s (2009) samples should be H. hamata (Aydin et al. 2019; Moazzam and Moazzam 2020). We do not agree with the point of Aydin et al. (2019) because: 1) H. hamata, as re-described by Aydin et al. (2019), has lateral papillae ± 20 in number on each side. However, H. ocellata, no matter which samples obtained from the current study or those of Teo and Ng (2009), exhibits approximately ± 40 lateral papillae on each side; 2) Tables of the body wall in H. hamata mainly have a spiny disc, whereas H. ocellata mainly has tables with a disc of smooth or wavy margin (Fig. 2E–H). Besides, Pearson (1913) first described the characteristics of both H. hamata and H. ocellata simultaneously. According to Pearson’s records, H. ocellata had the Cuvierian tubules, while H. hamata did not. This discrepancy further certifies that these two sea cucumbers are different species.

The complete mitochondrial genome of H. ocellata has a similar A-T content with the other holothuroid mitogenomes analyzed in the previous study (Utzeri et al. 2020; Sun et al. 2021; Li et al. 2022; Ma et al. 2022). The comparison of the gene order showed that H. ocellata has the same gene position on mtDNA as that of all other Holothuria, suggesting a common ancestral condition and a closer evolutionary relationship. The taxonomic controversy about H. ocellata centers on its subgenus: Jaeger (1833) initially thought it belonged to subgenus Holothuria, and some authors considered it as subgenus Metriatyla Rowe, 1969 (Rowe 1969; Cannon and Silver 1986). Moreover, some scholars have proposed that it should be classified into the subgenus Theelothuria based on the further study of ossicles (Liao 1997; Lane et al. 2000). The result of phylogenetic analysis indicated that H. ocellata preferentially forms a sister group with H. spinifera before clusters with Holothuria scabra Jaeger, 1833 (Fig. 6A). Besides, the result of phylogenetic analysis also indicated that H. ocellata preferentially forms a sister group with H. spinifera before clusters with Holothuria tubulosa Gmelin, 1791 (Fig. 6C). That means, compared to H. scabra and H. tubulosa, the member of subgenus Metriatyla and subgenus Holothuria, H. ocellata is more closely related to H. spinifera, which belongs to subgenus Theelothuria. In early studies, some scholars also found that H. ocellata is closely related to the species of subgenus Theelothuria: Théel (1886) considered H. ocellata to be doubtlessly nearly allied to H. squamifera; Pearson (1913) also found that H. ocellata and H. spinifera were exceedingly similar in morphology; and either H. squamifera or H. spinifera, both sea cucumber species, belong to subgenus Theelothuria. Because of that, it is reasonable to believe that H. ocellata should be classified within the subgenus Theelothuria.

Due to the lack of original descriptions and the difficulties in retrieving ancient references, identifying whether B. ocellata or H. ocellata refers to the same species discovered by Jaeger (1833) presents a challenge. Fortunately, Samyn and Vandenspiegel (2016) examined the holotype of Bohadschia ocellata Jaeger, 1833, and concluded that this is a valid Bohadschia species. In addition, the validity of the classification status of B. ocellata was certified by WoRMS. Because of that, we can assume that the species discovered by Jaeger was Bohadschia ocellata. Semper (1868) found a sea cucumber species and recorded it as H. ocellata, only citing Jaeger’s descriptions without further explanation. Théel (1886), the third recorder of H. ocellata, not only gave a detailed description but also made hand-drawn pictures of appearance and ossicles. After comparing the description and the hand-drawn images of the species, we found them to be largely consistent with the samples obtained in this study. We speculated that Théel’s (1886) record should be considered the actual first discovery of H. ocellata. According to the relevant provisions of the International Code of Zoological Nomenclature and considering the current classification of H. ocellata, we propose that it should be given a new formal Latin name. However, due to the absence of type specimens found by Théel (1886), it is not appropriate to give this species a new name directly by ourselves. Because of that, the full Latin name of this species should tentatively be “Holothuria ocellata Jaeger, 1833 sensu Théel 1886,” which means “specimens belonging to the species that Théel 1886 called Holothuria ocellata” and distinguishes it from “Bohadschia ocellata Jaeger, 1833.”