Animal collection

Animals were collected at ten sites in Toyama Bay, Japan during the deep-sea dives by the manned submersible Shinkai 2000 (JAMSTEC) (Table 1). In situ habitation and behavior of M. hians in response to disturbance by the submersible were observed and partly video-recorded. A total of 36 live specimens were collected by gentle grab with the submersible's manipulator at depths ranging from 592 m-978 m as summarized in the table. The specimens were fixed with 10% formalin-seawater or 2.5% glutaraldehyde-sea-water on board the support vessel.

Table 1

Light and electron microscopy

Some fixed tunic pieces were cut into slices approximately 0.5 mm thick with a razor blade and observed with a light microscope equipped with Nomarski differential interference contrast optics (DIC).

For the paraffin sections, the formalin-fixed specimens were dehydrated through an ethanol-xylol or butanol series and embedded in paraffin. Serial sections of 7–10 μm thick were stained with Delafield's hematoxylin and eosin.

For the resin sections, the glutar-fixed specimens were postfixed with 1% osmium tetroxide-0.1 M sodium cacodylate (pH 7.4) for 2 hr, dehydrated through an ethanol series, cleared with n-butyl glycidyl ether and embedded in low-viscosity epoxy resin. Sections of 1 μm thick were stained with toluidine blue for light microscopy. Thin sections were doubly stained and observed in a transmission electron microscope (JEM-1010, JEOL) at 80 kV.

General morphology and habitat

Megalodicopia hians is composed of a body that looks like a ripe pomegranate and a long peduncle protruding from the body (Fig. 1A). The body contains the large oral aperture and main organs of the ascidian such as the branchial basket, stomach, heart, and gonads. The oral aperture has huge anterior and posterior lobes. Each lobe has many parallel circular muscle fibers, and longitudinal muscles run across them. The muscle fibers are crowded at the oral aperture and make a broad muscular band at the margin of the lobes. These muscle fibers become more thinly spaced as they are apart from the lobe margin. Branchial basket is unfolded and has some indistinct transverse and longitudinal vessels. There are numerous oral tentacles of about 1mm long. The short endostyle, about 1 cm long, is present along the median line in the ventral side of the branchial basket. The unciliated branchial perforations (stigmata) are rectangular and 200–300 μm in diagonal. They are larger and more densely scattered in the anterior region of the branchial basket than in the posterior. A small atrial siphon opens at the antero-dorsal edge of the body. The gut forms a simple loop and is positioned on the right side of the body. The stomach is oval and irregularly folded on its surface. Gonads are in the gut-loop, and when the organism is alive, it is white in color. The ovary forms a spherical mass and it is situated anterior to the hemisphere-shaped testis. In the collected specimens, the length from the top of the body to the end of peduncle was found to be 5∼26 cm (Table 1). The lengths did not show any correlation with the depths of the collection sites. The tunic was found to be very thin, and its surface was often attached with hydrozoans. Some other solitary ascidians adhered to the bodies of the eight specimens collected from the same station at a depth of 700 m (1280 B-1c, 1d, 1e, 1g, 1i, 1j, 3d, 3e; 37°07.846′N, 137°09.680′E).

Fig. 1

The fixed specimens Megalodicopia hians Oka (A). Arrow indicates the oral aperture, and arrowhead indicates the atrial aperture. This ascidian is consisted of body and the peduncle. Scale bar, 2cm. Aggregation of M. hians on the slope (B and C). Animals face to the same direction (B, lower left corner, C, lower side).

M. hians was attached on a rock bed of the slope with the posterior end of the peduncle. The peduncle is provided with a pair of thick longitudinal muscular bands in the ventral part. These muscular bands begin from the posterior end of the body. The rock bed was covered with muddy sediment. At the collection sites, water temperature was approximately 0.3°C, dissolved oxygen was 4.6∼4.7 ml/L, and the velocity of the water current was 0.03∼0.32 m/s. Although the current direction at the M. hians aggregations was not precisely determined, up-slope currents at <850 m deep have previously been reported (Odamaki and Kikuchi, 1985). Most of the M. hians individuals oriented the wide-opened oral aperture to down-slope direction, i.e., in the counter-direction to the presumed upwelling currents (Fig. 1B, C). Tens of individuals attached to hard substrates such as a thinly sedimented rock bed, forming a patchy aggregation of ranging from approximately a few to several m². Patchy aggregations were scattered along the rock-exposure area of the slope, with several patches occurring over the migrated rocks on the soft sediment floor. The M. hians individuals showed quick (>5 sec) protection behavior during the submersible's approach to the animals. They closed their mouth apertures and exhibited a crouching action in response to sampling by the manipulator. It was likely that the organisms were reacting to the turbulence caused by the manipulator rather than the light and noise disturbance caused by the submersible.

We could see many small invertebrates such as crustaceans, chaetognaths, and so on, swimming around the lights of the submersible. They also swam around the ascidians. Many fragments of crustaceans and diatoms were found in the gut and feces of M. hians. We observed a crustacean caught in the lobe of the oral aperture of one specimen (1278 B-2b; Fig. 2A). The sizes of the fragments were various, and whole shells of crustaceans were sometimes found in the feces (Fig. 2B). Many diatoms were also found in the gut (Fig. 2C) and feces (Fig. 2D).

Fig. 2

Diet of M. hians. Small crustaceans found in the oral aperture (A) and feces (B). Diatoms found in the gut (C) and feces (D). Scale bars; (A) 1mm, (B) 200 μm, (C) 10 μm, (D) 200 μm.

Tunic morphology

The tunic of M. hians is gelatinous and semi transparent. The tunic matrix is overlaid with a thin layer of tunic cuticle (Fig. 3). Tunic cells were found throughout the tunic matrix, and many of them were distributed with bias just beneath the tunic cuticle. There were no blood vessels found in the tunic. The tunic cuticle was found to be approximately 50 nm thick or less and composed of electron-dense, fibrous materials (Fig. 4). The cuticular surface was found to be flat with no protuberant structures.

Fig. 3

A cross-section of the tunic (paraffin section stained with hematoxylin and eosin). Tunic cuticle (c) covers tunic matrix (tm). Arrow-heads indicate some tunic cells distributed in the tunic matrix. Scale bar, 50 μm.

Fig. 4

Electron micrograph of the tunic cuticle (c) covering tunic matrix (tm). Scale bar, 0.2 μm.

There appear to be several types of tunic cells; some spherical-shaped and others amoeboid-shaped (Fig. 5). Amoeboid-shaped cells were found to contain granules and vesicles (Fig. 6A). Some spherical cells have a large vacuole (Fig. 6B), and some contain metachromatic granules (Fig. 6C). Although electron micrographs showed granular and vesicular inclusion in the cytoplasm, the fine structures of organelles were not preserved well (Fig. 7).

Fig. 5

Tunic cells observed with DIC optics. Scale bar, 10 μm.

Fig. 6

Tunic cells in resin sections of the tunic (stained with toluidine blue). Arrowheads in C indicate the tunic cells containing metachromatic granules. Scale bar, 10 μm.

Fig. 7

Electron micrographs of tunic cells. An amoeboid-shaped cell with electron-dense granules and vesicles (A) and a spherical cell with a large vacuole (B). Scale bars, 1 μm.

Bacteria were occasionally found in the cytoplasm of tunic cells (arrow in Fig. 8). Because the bacteria appeared to be enclosed in the membrane, they were probably endocytized by the tunic cell. The bacteria were also found in the tunic matrix (Fig. 9). Many of them seemed to be distributed in the sub cuticular area. All bacteria both inside and outside the cells were of the same morphological type.

Fig. 8

Tunic cell containing a bacterium (arrow). n, nucleus of the tunic cell. Scale bar, 1 μm.

Fig. 9

Rod-shaped bacteria distributed in the tunic matrix. Scale bar, 0.2 μm.

Multicellular structures were often found beneath the tunic cuticle (Fig. 10), though these structures were not distributed in any other part of the tunic. A long multicellular cord was found to have several sac-shaped branchlets (arrowhead in Fig. 10), and the termini of the branchlets protruded out through the tunic cuticle (arrow in Fig. 11). The cord sometimes branched but connected with neither other cords nor any other structures. The histological sections showed that the cord consists of spindle-shaped cells forming a line and runs parallel to the tunic surface (Fig. 12, 13).

Fig. 10

Multicellular structure (arrows) in the tunic. Arrowheads indicate branchlets. Scale bar, 100 μm

Fig. 11

The terminus of the sac-shaped branchlet (arrow) protrudes outside the tunic cuticle. Scale bar, 50 μm.

Fig. 12

Multicellular cord in the tunic observed with DIC optics. Scale bar, 20 μm.

Fig. 13

Multicellular cord in the tunic matrix (tm) (paraffin section). c, cuticle. Scale bar, 20 μm.

Spawn in the tunic

Sac-like structures of 5-6 mm in diameter were found in the tunic of the two specimens collected at a depth of 978 m (1279 B-1a; 37°09.648′N, 137°12.366′E) and 942 m (1279 B-1d; 37°09.678′N, 137°12.310′E), respectively. These structures were exclusively found in the right-side tunic of the ascidian body. One specimen (1279 B-1a) had six sacs, and the other (1279 B-1d) had four. These sacs were adjacent to one another, and contained many eggs / embryos (Fig. 14A, 14B). The sac was almost embedded in the tunic, except for the center of the upper side of the sac (Fig. 14B). The edge of the tunic covering the sac was circular (arrowheads in Fig. 14A). The eggs /embryos were white and spherical, approximately 200 μm in diameter (Fig. 14B, 14D). In the histological observation, the shell of this sac consisted of two layers (Fig. 14C). The outer layer was continuous with the tunic matrix, but it had no direct connection with the inner layer. The inner layer was well stained with hematoxylin. The center of the upper side of the sac was not covered by either the outer layer or the tunic. In the histological section of the eggs /embryos, a layer of ectoderm entirely covered the ellipsoidal yolk granules (Fig. 14E).

Fig. 14

Sac-like structure containing spawn in ascidian tunic. (A) The sac observed through the tunic surface. Arrowheads indicate the edge of the tunic covering. (B) Cross -section of the sac. Arrows indicate the periphery of the tunic. (C) Enlargement of the shell of the sac. (D) Eggs /embryos in the sac. (E) Histological section of an egg. Arrows indicate the ectodermal layer. il, inner layer; ls, lumen of the sac; ol, outer layer; tm, tunic matrix; yg, yolk granule. Scale bars; (A) 1mm, (B) 500 μm, (C) 100 μm, (D) 200 μm, (E) 50 μm.

Taxonomical Remarks

The morphologic characteristics of specimens collected in this study are consistent with the original description of Megalodicopia hians made by Oka (1918) and the record by Tokioka (1953). These records are based on specimens collected from the Japan Sea at a depth of 360 m and from Sagami Bay (Japan), which faces to the Pacific Ocean at 350 m, respectively. Sanamyan (1998) has observed the morphology of M. hians collected from the South Kurile Islands and redescribed the record of this species from the Weddell Sea, Antarctica by Kott (1969). The specimen collected from South Kurile Islands was “gradually tapering to the base of the peduncle. Peduncle not distinctly demarcated from body.” However, the bodies were obviously constricted at the junction with the peduncles in the specimens investigated here and in the specimens described by Oka (1918) and Tokioka (1953). Therefore, it might be possible that the specimen from the South Kurile Islands is not of the same species as the Japanese M. hians.

Food habit

Most of the ascidians are filter-feeders. Water enters the branchial basket from the branchial aperture, and the captured particulate materials are carried by the mucus towards the oesophagus. The water current is made by the ciliary movement of stigmata. However, because M. hians does not have ciliated stigmata, it is not able to produce water current by itself. In situ observation showed that most of the animals face their oral apertures in the same direction, keeping their apertures open. This finding may suggest that M. hians utilizes the water current for feeding. Because there is a water current (0.03∼0.32 m/s) in their habitat, this ascidian might catch relatively large materials drifting in the water currents and engulf them without distinction. Various sizes and kinds of animal fragments and diatoms were observed in the gut of this ascidian. These findings also support other evidence of the nonselective feeding of this species. Based on a previous study, Monniot and Monniot (1978) have also suggested the presence of weak filtration and a mixed diet such as one including fine organic elements and small crustaceans in octacnemid species based on the morphology of the branchial basket. In another abyssal group, the family Hexacrobylidae, the animals are thought to catch small invertebrate with their 6-lobed oral siphon (Cf. Monniot et al., 1975). The present species, however, does not have large tentacle-like structures, and such catching-behavior was not recorded in the present in situ observation.

Tunic morphology and parasitic inclusions

The tunic of M. hians is basically the same as those of other ascidians. In M. hians, the tunic is not a tough integument, because the tunic cuticle is very thin and the tunic matrix is gelatinous. If the animal needs to be protected from predators and/or infectious organisms, cellular or chemical defenses should be available in the tunic. The biased distribution of tunic cells around the tunic surface may indicate their defensive function. Although there are several types of tunic cells, the function of each type is uncertain. Some amoeboid-shaped tunic cells containing granules and vesicles are similar in morphology to tunic phagocytes in Aplidium yamazii (Hirose et al., 1994) and some other species. In the present study, the fine structures of the tunic cells were not well-preserved; some vacuoles appeared to have swelled, and the cell membranes were sometimes discontinuous. Deformation and puncture of tunic cells might have been caused by the rapid change in hydraulic pressure during animal collection.

The ascidian tunic usually contains rod-shaped bacteria of one or two morphological types, and the tunic of M. hians also contains bacteria that are mostly distributed beneath the cuticle. Bacteria are also found in the vesicles of the amoeboid-shaped cells, and the bacteria would be endocytized by tunic cells. It is unknown whether the bacteria are symbionts for the ascidians.

The multi-cellular structure in the subcuticular area seems to be hyphae of fungi. Although this is the first time we have found fungi-like structures in the tunic matrix, the occurrence of algal and fungal filaments have been reported in some ascidians (Lambert et al., 1996).

Sac-like structures contained spawn were found in the tunic of M. hians. The shell of the sac was found to consist of two layers. The outer layer was continuous to the tunic matrix and seemed to be a tunic cuticle that was restored after spawning. No direct connection was detected between the inner layer and tissue of ascidian, and, as such, the inner layer might be derived from the spawner. Although there is no doubt that this sac and spawn are not derived from the ascidian, the spawner is uncertain. This is the first report regarding the occurrence of spawn in the ascidian tunic to our best knowledge. In the bathyal fauna, because there are few materials such as seaweed and rocks protruding from the rock bed covered by muddy sediment, the animal might utilize the ascidian as a substratum for spawning.

Phylogenetic consideration based on tunic morphology

It is generally agreed that Octacnemidae is a member of the suborder Phlebobranchia (Cf. Tokioka, 1953; Monniot et al., 1991; Kott, 1985), but its phylogenetic relationship with other phlebobranchian families remains controversial. For example, Tokioka (1953) puts M. hians in “Aberrant forms” with some other genera. He states that these genera undoubtedly belong to Phlebobranchia but that the relationships among them and with other families are uncertain. While most members of Phlebobranchia have the gut-loop on the left side of their body, the gut-loop is situated on the right side of the body in the family Corellidae and Octacnemidae families, possibly suggesting that Octacnemidae is closely related to Corellidae (Kott, 1985). On the other hand, Millar (1956; 1959; 1966) have pointed that the octacnemid species has evolved from an ancestral form of ascidians like Ciona based on observation of Octacnemus bythius, and that octacnemids might be related to the Cionidae.

Here, we provide a brief phylogenetic discussion of M. hians based on the tunic morphology. The fine structures of the tunic cuticle were investigated in 116 species covering the all families and subfamilies of the class Ascidiacea except for Octacnemidae, Plurellidae, and Dimeatidae, which are all bathyal or abyssal species (Cf. Hirose et al., 1997). The tunic cuticles of some ascidians have minute protrusions, and there is a general tendency for the protrusions occurs in limited families. Of the 27 phlebobranchians, cuticular protrusions were exclusively found in Pterygascidia longa (Cionidae) and Adagnesia vesiculiphora (Agneziidae). In M. hians, the tunic cuticle has a flat surface and no protrusions, and this character-state is the same as that of most other phlebobranchians species.

The presence or absence of the two types of tunic cells (tunic bladder cells and tunic net cells) was surveyed in 65 ascidian species covering 11 families (Hirose, 2001). As for phlebobranchians, the survey did not cover Agneziidae, Octacnemidae, Plurellidae, and Dimeatidae. Of the 14 phlebobranchian species, tunic bladder cells were exclusively distributed in Diazoninae and Ascidiidae, while tunic net cells were not found in any phlebobranchians. The tunic of M. hians has neither tunic bladder cells nor tunic net cells. This character-state may suggest some affinity to Cioninae, Corellidae, or Perophoridae, although 3 other phlebobranchian families have not yet been examined. Some corellids and ascidiids are known to have giant cells that are very large hemocytes containing sulfuric acid (Wright, 1981). The hemocytes found in the branchial wall of M. hians did not include giant cells (data not shown).

In the present report we first described the histological and fine structures of the tunic in Octacnemidae. Here, we could deal with only one species, M. hians, and it is uncertain whether the present observations can represent all octacnemids. Moreover, other bathyal or abyssal families might have independently originated from other lineages of ascidians. Although deep-sea ascidians have been poorly studied because of their limited availability, a further accumulation of their data is necessary to more fully disclose details of their life, phylogeny, and environmental adaptation.