Zooxanthellate, Sclerite-Free, and Pseudopinnuled Octocoral Hadaka nudidomus gen. nov. et sp. nov. (Anthozoa, Octocorallia) from Mesophotic Reefs of the Southern Ryukyus Islands

: Shallow water coral reefs are the most diverse marine ecosystems, but there is an immense gap in knowledge when it comes to understanding the diversity of the vast majority of marine biota in these ecosystems. This is especially true when it comes to understudied small and cryptic coral reef taxa in understudied ecosystems, such as mesophotic coral reef ecosystems (MCEs). MCEs were reported in Japan almost ﬁfty years ago, although only in recent years has there been an increase in research concerning the diversity of these reefs. In this study we describe the ﬁrst stoloniferous octocoral from MCEs, Hadaka nudidomus gen. nov. et sp. nov., from Iriomote and Okinawa Islands in the southern Ryukyus Islands. The species is zooxanthellate; both specimens host Cladocopium LaJeunesse & H.J.Jeong, 2018 (formerly Symbiodinium ‘Clade C’) and were collected from depths of ~33 to 40 m. Additionally, H. nudidomus gen. nov. et sp. nov. is both sclerite-free and lacks free pinnules, and both of these characteristics are typically diagnostic for octocorals. The discovery and morphology of H. nudidomus gen. nov. et sp. nov. indicate that we still know very little about stoloniferous octocoral diversity in MCEs, their genetic relationships with shallower reef species, and octocoral–symbiont associations. Continued research on these subjects will improve our understanding of octocoral diversity in both shallow and deeper reefs.


Introduction
Coral reefs make up only 0.2% of the earth's ocean but are estimated to harbor a quarter of all marine species [1,2] and are the most diverse marine ecosystems on the planet. Unfortunately, these diverse marine communities are also one of the most threatened [3][4][5][6]. The 'hotspot' concept, a term used to mark a relatively restricted geographic area accommodating exceptionally high concentrations of biodiversity and endemism [7][8][9] has highlighted the wealth of species that are at risk and how localized such areas of richness can be [10]. However, there are vast gaps in knowledge concerning the majority of marine biota [11,12], making the recognition of biodiversity geographic patterns and hotspots questionable [13,14], as priorities identified for one taxon may not reflect the diversity of other taxa [14,15]. This is especially true for understudied localities and environments, such as understudied coral reef ecosystems. air) and were preserved in 70-90% ethanol and subsamples in 95% ethanol. The current study is part of an ongoing survey of mesophotic and deep reef work. Vouchers and type material were deposited at the National Museum of Nature and Science (NSMT), Tokyo, Japan (Table 1). Both specimens were examined for the presence of sclerites by dissolving entire polyps and stolons in 4% hypochlorite (household bleach). Additionally, to visualize polyp tentacles and pseudopinnules, polyps were fixed in 20% formalin and embedded in methylene blue (1%).

DNA Extraction, Amplification, and Sequencing
DNA was extracted from polyps using a DNeasy Blood and Tissue kit (Qiagen, Tokyo, Japan). PCR amplification and sequencing were performed for four markers, of which three were mitochondrial (cytochrome c oxidase subunit I (COI), the MSH homologue mtMutS, and subunit ND6) and the fourth was the nuclear ribosomal marker (28S rDNA). Additionally, for Symbiodiniaceae, the nuclear internal transcribed spacer (ITS) region of ribosomal DNA was amplified. Protocols in [34] were followed and PCR products were treated with Exonuclease I and alkaline phosphate (shrimp) and sent for bidirectional sequencing on an ABI 3730XL (Fasmac, Kanagawa, Japan). Sequences were assembled and edited using Geneious R11 [40] and BioEdit [41]. COI, mtMutS, and ND6 were checked for introns, exons, and stop-codons in AliView [42].
Diversity 2019, 11, x FOR PEER REVIEW 3 of 14 and were preserved in 70-90% ethanol and subsamples in 95% ethanol. The current study is part of an ongoing survey of mesophotic and deep reef work. Vouchers and type material were deposited at the National Museum of Nature and Science (NSMT), Tokyo, Japan (Table 1). Both specimens were examined for the presence of sclerites by dissolving entire polyps and stolons in 4% hypochlorite (household bleach). Additionally, to visualize polyp tentacles and pseudopinnules, polyps were fixed in 20% formalin and embedded in methylene blue (1%).

DNA Extraction, Amplification, and Sequencing
DNA was extracted from polyps using a DNeasy Blood and Tissue kit (Qiagen, Tokyo, Japan). PCR amplification and sequencing were performed for four markers, of which three were mitochondrial (cytochrome c oxidase subunit I (COI), the MSH homologue mtMutS, and subunit ND6) and the fourth was the nuclear ribosomal marker (28S rDNA). Additionally, for Symbiodiniaceae, the nuclear internal transcribed spacer (ITS) region of ribosomal DNA was amplified. Protocols in [34] were followed and PCR products were treated with Exonuclease I and alkaline phosphate (shrimp) and sent for bidirectional sequencing on an ABI 3730XL (Fasmac, Kanagawa, Japan). Sequences were assembled and edited using Geneious R11 [40] and BioEdit [41]. COI, mtMutS, and ND6 were checked for introns, exons, and stop-codons in AliView [42].

Molecular Phylogenetic Analyses
Multiple sequence alignments were performed using MAFFT 7 [43] and coding markers were aligned using MACSE [44] under default parameters. The phylogenetic position of the collected specimens (n = 2) was determined by aligning the consensus sequences for markers 28S rDNA, COI, and mtMutS to a reference dataset of 124 octocoral genera, including Cornularia pabloi and Cornularia cornucopiae as outgroup (total n = 144), as used in Lau and Reimer [33]. This resulted in alignments of 887 bp for 28S rDNA, 717 bp for COI, and 714 bp for mtMutS, and a total concatenated three-marker dataset of 2318 bp. The separate markers were run in ML analyses, to check for contamination and congruency (Supplementary Materials Figures S1-S3).
A separate phylogenetic analysis was made to examine the lower level phylogenetic relationships of the collected mesophotic specimens, using a concatenated four-marker dataset. The concatenated four-marker dataset resulted in an alignment of 2670 bp (total n = 12). A total of seven reference species were included in the analysis, which clustered in nearby clades with the specimens in the three-marker dataset, including Rhodelinda sp. and Telesto sp. as outgroup. The four separate markers (28S rDNA, 787 bp; COI, 708 bp; mtMutS, 734 bp; ND6, 441 bp) were also run in ML analyses, to check for contamination and congruency (Supplementary Materials Figures S4-S7).
Additionally, ITS sequences from the two specimens were aligned with a total of 25 reference sequences (Cladocopium spp. and Durusdinium spp.), including Gerakladium sp. as outgroup. The resulting dataset comprised 641 bp and a total of 27 sequences and was run in ML analyses (Supplementary Materials Figure S8).
Alignments of the separate markers were concatenated using SequenceMatrix 1.8 [45]. ML analyses were run with RAX-ML 8 [46], using the GTRCAT model. The best ML tree was calculated using the -D parameter. A multi-parametric bootstrap search was performed, which automatically stopped based on the extended majority rule criterion. The Bayesian inference was performed with ExaBayes 1.5 [47] using the GTR substitution model. Four independent runs were run for 10,000,000 generations during which convergence (with a standard deviation of split frequencies < 2%) was reached. Bootstrap supports and posterior probabilities were depicted on the branches of the best ML tree using P4 [48]. The resulting trees were visualized in FigTree 1.4.2 [49]. Additionally, average distance estimations within species and within genera were computed using MEGA X [50] by analyzing pairwise measures of genetic distances (uncorrected P) among sequences (Supplementary Materials Tables S1-S3). Type species: Hadaka nudidomus sp. nov. by original designation and monotype. Diagnosis: Colony with polyps connected through flattened ribbon-like stolons, which are loosely attached to a hard substrate. Polyps retract fully into the calyx, which is cylindrical to conical in shape, narrowing at the base and does not retract fully into the stolon. Tentacles have a wide rachis with a protruding ridge and pseudopinnules of different lengths arranged on either side, giving the polyps feather shaped tentacles. No sclerites. Zooxanthellate.

Class
Remarks: Hadaka gen. nov. et sp. nov. shows gross resemblance to Hanabira Lau, Stokvis, Imahara & Reimer, 2019 in having a similar polyp shape with feather or petal shaped tentacles and fused pinnules, which can still be distinguished by shallow furrows. Hadaka gen. nov. et sp. nov. differs from Hanabira in having no sclerites in any part of the colony and having a protruding ridge on the upper side of the tentacle. Genetically, Hadaka gen. nov. is well-supported and positioned in a different phylogenetic clade from Hanabira. The closest sister taxa of Hadaka gen. nov. is Acrossota Bourne, 1914, which is also sclerite-free, but morphologically very different; Acrossota lacks pinnules completely.
http://zoobank.org/39430672-5ADA-4EFF-9F5A-B4076B6B90C0 Description: Holotype colony consists of 15 polyps with flattened ribbon-like stolons encrusting a sponge. Polyps can be seen individually or clustered in groups and are spaced apart irregularly, 3 mm to 2 cm in between polyps and clusters. Stolons are 0.5 mm at their narrowest and 1 mm at their widest point. Polyps retract fully into the calyx (~1.8 mm wide and~3.55 mm in length), which is cylindrical to conical shaped, narrowing at the base, and does not retract fully into the stolon. Expanded polyps are~4-5 mm diameter in life. Tentacles have a wide rachis with a protruding ridge on the upper side and long pseudopinnules arranged on either side (~24-26 pseudo-pairs), giving the polyps feather shaped tentacles. When stained with methylene blue, the outline of the tentacles can be observed. Structures of the pinnule axis are visible; however, the notches that distinguish the pseudopinnules are not observed in the contour of the tentacle (Figure 2d). No sclerites were found in any parts of the specimens. Polyps are brown in life and yellowish-white in ethanol (Figure 2c). Zooxanthellate.
Morphological variation: There is a difference in color between the polyps of the holotype (NSMT-Co 1681) and paratype (NSMT-Co 1682); the polyps of the holotype are brown with a white oral disc and base of the tentacles and the polyps of the paratype are whitish yellow with a bright blue oral disc (Figure 2a,b).
Distribution: Southwestern Japan, southern Ryukyus Islands, around northern Okinawa Island, and inside the bay of western Iriomote Island in the East China Sea. Specimens were collected from depths of~33-40 m.
Remarks: The polyps of paratype NSMT-Co 1682 were all used for DNA extraction and sclerite examination, as they were initially thought to be a Hanabira yukibana specimen; three fragments of rock with stolon remain. The holotype colony (NSMT-Co 1681) was attached to sponge tissue, but this epibiont is not obligate, as the paratype was attached to rock.
Habitat: The holotype (NSMT-Co 1681) was found attached to sponge on a large piece of coral rubble (>15 cm) lying on a mixed small rubble/soft sediment bottom. The paratype (NSMT-Co 1682) Diversity 2019, 11, 176 6 of 13 was found on consolidated hard carbonate bottom. Both colonies were on the upward-facing side of the bottom.

Molecular Results
This study added a total of six sequences of Hadaka nudidomus gen. nov. et sp. nov. to the public reference database GenBank and no barcodes were available before. For the family Symbiodiniaceae, two Cladocopium spp. sequences were added. The phylogenies resulting from the ML analyses of the separate markers (COI, 28S rDNA, mtMutS, ND6) were highly congruent with those from the concatenated alignments for both the three-and four-marker datasets (Supplementary Materials Figures S1-S7). ML and Bayesian analyses for the concatenated datasets yielded almost identical tree phylogenies (Supplementary Materials Figure S9  Habitat: The holotype (NSMT-Co 1681) was found attached to sponge on a large piece of coral rubble (>15 cm) lying on a mixed small rubble/soft sediment bottom. The paratype (NSMT-Co 1682) was found on consolidated hard carbonate bottom. Both colonies were on the upward-facing side of the bottom.

Molecular Results
This study added a total of six sequences of Hadaka nudidomus gen. nov. et sp. nov. to the public reference database GenBank and no barcodes were available before. For the family Symbiodiniaceae, two Cladocopium spp. sequences were added. The phylogenies resulting from the ML analyses of the separate markers (COI, 28S rDNA, mtMutS, ND6) were highly congruent with those from the concatenated alignments for both the three-and four-marker datasets (Supplementary Materials Figures S1-7). ML and Bayesian analyses for the concatenated datasets yielded almost identical tree phylogenies (Supplementary Materials Figure S9    Hadaka gen. nov. was sister to the remaining clade in the three-marker analyses; however, in the four-marker analyses, Phenganax was sister to Hadaka gen. nov. and Acrossota. Nonetheless, in both phylogenies, the two specimens of Hadaka gen. nov. formed a completely-supported clade. Additionally, genetic distances gave further support to phylogenetic affinities and morphological features justifying the establishment of a new genus. Between-genus distances (Hadaka compared to Acrossota and Phenganax) for COI were 2.52-2.54% and 7.69-13.33% for mtMutS, which are well above the intergeneric range for octocorals [51]. Hadaka gen. nov. was sister to the remaining clade in the three-marker analyses; however, in the four-marker analyses, Phenganax was sister to Hadaka gen. nov. and Acrossota. Nonetheless, in both phylogenies, the two specimens of Hadaka gen. nov. formed a completely-supported clade.
Additionally, genetic distances gave further support to phylogenetic affinities and morphological features justifying the establishment of a new genus. Between-genus distances (Hadaka compared to Acrossota and Phenganax) for COI were 2.52-2.54% and 7.69-13.33% for mtMutS, which are well above the intergeneric range for octocorals [51].   . Phylogenetic reconstruction using a four-marker concatenated dataset (28S rDNA + COI + mtMutS + ND6) among Hadaka nudidomus gen. nov. et sp. nov., closest sister species Phenganax spp. and outgroup specimens Rhodelinda sp. and Telesto sp. (total n = 12). The best maximum likelihood tree is shown, with values at branches representing bootstrap probabilities in percentages (top/left) and Bayesian posterior probabilities (bottom/right). In situ photographs are shown for the two octocoral species that are sclerite-free and lack free pinnules, Hadaka nudidomus gen. nov. et sp. nov. and Acrossota amboinensis. Photograph credit: in situ image RMNH Coel. 40798 Acrossota amboinensis, by Daniel Knop (modified from [52]; reproduced with permission from copyright holder).

Discussion
In the three-and four-marker phylogenies, there was disparity in the position of Hadaka nudidomus gen. nov. et sp. nov. and Acrossota amboinensis. It remains unresolved how these genera and genus Phenganax are related to one another. A possible explanation could be that there is no sufficient signal in the sequences of both Hadaka nudidomus gen. nov. et sp. nov. and Acrossota amboinenesis due to the fact that the closest relatives for these genera are yet to be discovered.
Morphologically, there was a difference between the coloration of the polyps of the holotype and paratype found at Iriomote and Okinawa Islands, respectively. The differences in coloration suggested that perhaps the specimens hosted different members of Symbiodiniaceae. However, both specimens hosted genus Cladocopium and thus no biogeographical distinction in Symbiodiniaceae was observed. Members of Cladocopium spp. are known to be adapted to a wide range of temperatures and irradiances [53], which would be expected from MCEs, where irradiances are not only subject to seasonal variations but are already reduced.
Hadaka nudidomus gen. nov. et sp. nov is the first zooxanthellate stoloniferous octocoral described from mesophotic depths. Only one other zooxanthellate octocoral, an alcyoniid species, Sinularia mesophotica Benayahu, McFadden, Shoham & van Ofwegen, 2017, has been explicitly described from mesophotic depths [54]. However, it was not further specified which genus or species of Symbiodiniaceae was hosted by S. mesophotica and therefore, we cannot yet hypothesize differences of zooxanthellae hosted by octocorals from MCEs.
Nonetheless, another recent study has shown that there are geographical differences in the genera of Symbiodiniaceae in Hanabira yukibana Lau, Stokvis, Imahara & Reimer, 2019 from shallow coral reefs, as specimens found from Okinawa Island hosted Cladocopium while Durusdinium LaJeunesse, 2018 was hosted in specimens from Iriomote Island. However, in this previous study, similar to the current study, no consistently different patterns of polyp coloration related to symbiont associations were observed [34]. To this end, finer-scale examinations of Symbiodiniaceae using faster-evolving DNA markers [55] may reveal patterns yet unseen.
Hadaka nudidomus gen. nov. et sp. nov. is the second species within Octocorallia after Acrossota amboinensis that has no sclerites in any part of the colony and also no free pinnules; both species are taxonomically placed within family Clavulariidae. Acrossota amboinensis differs from Hadaka nudidomus gen. nov. et sp. nov. in colony form, polyp morphology, and habitat ( Figure 4); A. amboinensis does not have pseudopinnules, but instead lacks pinnules completely and has, so far, not been found at mesophotic depths. When comparing A. amboinensis to Phenganax spp. there are also distinct morphological differences; all Phenganax species have free pinnules and have completely different polyp and tentacle shapes. Nonetheless, the genera Acrossota and Phenganax are phylogenetically closely related. In a recent study, the phylogenetic topology for these genera was different from that generated in the current study [33], in which Acrossota is placed basally to all Phenganax species. As a result of the unresolved phylogenetic location of Acrossota within Clavulariidae, while it is clear these three genera are distinct, it remains unclear how Hadaka and Phenganax are related to Acrossota.
Moreover, it can be concluded that several octocoral species lack both sclerites and free pinnules, and thus, that such features are not completely rare in octocorals, which raises important implications for the definition of subclass Octocorallia, as sclerite characterization and the presence of pinnated tentacles are two of the major diagnostic features of the group [52,56,57].
It is clear that more species diversity data from many marine regions are needed before we can state with certainty that the southern Ryukyus harbor high levels of stoloniferous octocoral diversity and endemism, but at least it can be said that this region potentially harbors many undiscovered species, not only in shallow coral reefs [33,34], but also among the many unexplored MCEs in this region.
Recent studies have shown that MCEs harbor distinct and independent biological communities when compared to shallower reefs [21]. MCEs are not only affected by anthropogenic and natural impacts as are shallow reefs but have seldom been the focus of specific conservation efforts [21,58].
Thus, researchers have only begun to scratch the surface of what we know about mesophotic marine life [21,58,59], including information on stoloniferous octocoral diversity and octocoral-zooxanthellae relationships in MCEs. The discovery of Hadaka nudidomus gen. nov. et sp. nov. and other recent discoveries [54,57] emphasize the need for continued studies on MCE octocoral diversity, as undescribed species may disappear before we have the opportunity to discover and study them [21].
Supplementary Materials: The following are available online at http://www.mdpi.com/1424-2818/11/10/176/s1, Figure S1: Maximum Likelihood phylogeny reconstruction of 28S rDNA gene region of Hadaka nudidomus gen. nov. et sp. nov. from Okinawa and Iriomote Islands (Japan) and octocoral references from 123 genera, including outgroup Cornularia spp., Figure S2: Maximum Likelihood phylogeny reconstruction of COI gene region of Hadaka nudidomus gen. nov. et sp. nov. from Okinawa and Iriomote Islands (Japan) and octocoral references from 123 genera, including outgroup Cornularia spp., Figure S3 Table S1: Number of base differences per site from averaging over all sequence pairs between stoloniferous octocoral genera (Hadaka gen. nov., Phenganax, Acrossota) is shown (p expressed as percentage) for COI and mtMutS gene regions. Standard error estimates (S.E.) are shown above the diagonal. Analysis involved 9 and 6 nucleotide sequences for COI and mtMutS, respectively. All positions containing gaps and missing data were eliminated. There were totals of 708 and 734 positions in the final dataset for COI and mtMutS, respectively. Evolutionary analyses were conducted in MEGA X (Kumar et al. 2018)., Table S2: Number of base differences per site from averaging over all sequence pairs between stoloniferous octocoral taxa (Hadaka nudidomus gen. nov. sp. nov., Phenganax spp., Acrossota amboinensis) is shown (p expressed as percentage) for gene regions COI and mtMutS. Standard error estimates (S.E.) are shown above the diagonal. Analysis involved 7 nucleotide sequences for both COI and mtMutS. All positions containing gaps and missing data were eliminated. There were totals of 717 and 881 positions in the final dataset for COI and mtMutS, respectively. Evolutionary analyses were conducted in MEGA X (Kumar et al. 2018)., Table S3: Estimates of average evolutionary divergence over sequence pairs within stoloniferous octocoral genera (Hadaka gen. nov., Phenganax, Acrossota, Rhodelinda, Telesto) for gene regions COI and mtMutS. The numbers of base differences per site from averaging over all sequence pairs within each group (d) are shown (p expressed as percentage). Standard error estimates (S.E.) are shown in the second column and were obtained by a bootstrap procedure (1000 replicates). Analyses involved 9 and 6 nucleotide sequences for COI and mtMutS, respectively. All positions containing gaps and missing data were eliminated. There were totals of 708 and 734 positions in the final dataset for COI and mtMutS, respectively. Evolutionary analyses were conducted in MEGA X (Kumar et al. 2018).
Funding: This research received no external funding.