Animals

The ascidians were collected in temperate and tropical shallow waters in Japan and Italy (see Table 1 for collection sites). The specimens were packed in a plastic container filled with seawater and immediately brought to the laboratory.

Table 1

Tunic pH and occurrence of tunic bladder cells and tunic net cells in ascidians

pH measurement

The pH in the tunic was measured with the needle-tip pH micro-electrode (Orion Res. Inc., Massachusetts) or pH test paper (timol blue and phenol red; Advantec, Tokyo). The electrode was inserted into the tunic or the colonies, and the tunic specimens or the colonies were bruised on the pH test paper. The pH measurement could not conducted in some species that have a very thin tunic.

Microscopy

The unfixed pieces of tunic or colony were sectioned with a razor blade by hands. The sections (about 0.5 mm thick or less) were mounted with seawater and observed under an epifluorescence microscope equipped with Nomarski differential interference contrast optics. Some specimens were vitally stained with 10 μM LysoSensor Yellow/Blue DND60 (Molecular Probes Inc., Eugene, Oregon) in sea-water for 30 min or more. The specimens were observed with violet light (380–425 nm) excitation. This dye is a fluorescent pH indicator that accumulates in acidic organelles. The color of the fluorescence that is emitted is pH dependent–predominantly yellow in acidic organelles and blue in less acidic organelles. The intensity of the fluorescence is also pH dependent: the intensity is stronger in a more acidic environment, whereas no fluorescence is emitted in neutral or basic environments. In the species collected in Italy, Ohtsuchi (Miyagi, Japan), and Iriomote (Okinawa, Japan), neutral red diluted in seawater (ca. 0.01%) was used for the vital staining of acidic organelles instead of LysoSensor.

Tunic rounding assay

This assay was carried out to assess the contractility of the tunic in some species that have tunic net cells. The colonies were sectioned into about 0.5 mm thick with a razor blade, and the fragments of zooids in the tunic slices were removed with forceps. Then, the tunic slices were placed in a petri dish filled with seawater, and incubated at room temperature for 12–24 hr.

Acidity of the tunic

The tunic pH was measured with a pH test paper and a needle-type pH electrode. The results from these methods, however, include some errors due to the contamination of seawater and hemolymph of the materials. In this report, the tunic pH was classified into 3 levels based on the measured value; strongly acidic (SA: <pH 3), acidic (A: pH 3–6), neutral (N: pH 6–8.5). The tunic pH was never more basic than the pH of seawater (ca. pH 8.5), so far studied.

As shown in Table 1, the tunic contains strong acid (level SA) in the species of particular taxa, i.e., Didemnidae, Holozoinae, Diazoninae, Ascidiidae and Corellidae. Exceptionally, the tunic pH is almost neutral in Ascidia zara and Phallusia mammillata of Ascidiidae and Didemnum sp. of Didemnidae. The acidity is not so strong (level A) in some didemnids, i.e., Didemnum molle, D. moseleyi, D. proliferum, and Leptoclinides madara, when the pH was measured with the pH electrode. In the other taxa, the tunic pH is almost neutral (level N) and, in most cases, slightly more acidic than seawater.

Acid containers in the tunic

Vital staining with LysoSensor Yellow/Blue or neutral red demonstrated the distribution of the acid in the tunic. In the all species having acidic tunic, the tunic contains giant, highly vacuolated cells that are often over 50 μm in diameter (Fig 1A–5A). A large vacuole occupies a great portion of the cytoplasm, and a thin layer of cytoplasm surrounds the vacuole. LysoSensor evenly stains the vacuolar lumen and emits strong yellow fluorescence (Fig. 1B–5B). Based on these characteristics, the tunic cells are concluded to be the tunic bladder cells that have been described in some ascidiids and didemnids (Cf. Goodbody, 1974). Neutral red also evenly stains the vacuolar lumen of the tunic bladder cells.

Figs. 1–8. Pair-images of the tunic slice vitally stained with LysoSensor (A, Nomarski differential interference contrast; B, epifluorescence). The specimens in Fig. 1–7 contain the tunic bladder cells. Scale bars, 20 μm for Fig 1, 2, 3, and 6; 50 μm for Fig 4, 5, 7, and 8.

Fig. 1

Trididemnum savignyi.

Fig. 2

Distaplia dubia.

Fig. 3

Rhopalaea sp.

Fig. 4

Ascidia ahodori.

Fig. 5

Diplosoma virens, a non-spiculate didemnid. There are many layers of the tunic bladder cells beneath the tunic surface (arrowhead).

Fig. 6

Didemnum moseleyi, a spiculate didemnid. There are only a few layers of the tunic bladder cells beneath the tunic surface (arrowhead).

Fig. 7

Ascidia zara. Fluorescence is not found in the bladders cells.

Fig. 8

Giant cells (hemocytes) of Corella sp.

In the acidic tunic, the tunic bladder cells are usually distributed in close formation near the tunic surface and they often form a layer that is packed with the tunic bladder cells (Fig. 5). The amount of the tunic bladder cells is relatively small in D. molle, D. moseleyi, D. proliferum, and L. madara. For instance, there are only one or two rows of the tunic bladder cells in D. moseleyi (Fig. 6). In these species, the acidity is relatively weak when the pH is measured with the pH electrode, and spicules are densely distributed throughout the tunic. In the neutral tunic of Didemnum sp., there are no tunic bladder cells, and the tunic is full of spicules. In the neutral tunic of Ascidia zara and Phallusia mammillata, there are many tunic bladder cells, but they do not contain acidic fluid (Fig. 7).

The tunic bladder cells were exclusively found in Didemnidae, Holozoinae, Diazoninae, and Ascidiidae. Although the tunic pH in Corella japonica is acidic (level SA), the vital staining with neutral red showed that there is no acid storage in the thin tunic. In Corella japonica and Corella sp. (Cf. minuta), large vacuolated cells are found in hemocoel, and they contain acid in the vacuoles (Fig. 8). These hemocytes are known as giant cells and their morphological characteristics are the same as those of the tunic bladder cells. The acidic pH in the tunic of C. japonica may be caused by the contamination of the contents of the giant cell. The giant cells are not found within the tunic in the two Corella species studied here.

Besides the tunic bladder cells, amoeboid cells contain round vesicles that emit blue or yellow fluorescence. The brightness of these fluorescence is much less than the fluorescence of the tunic bladder cells, indicating that the pH in the vesicles are less acidic than the pH in the vacuoles of the tunic bladder cells. Neutral red also stains some of the vesicles. This type of cells is found in the all species examined here, including non-acidic tunic, and thus, the vesicular contents have no effects on the tunic pH measured by the present method. Such fluorescent vesicles are also detected in epithelial cells of tunic vessels and some blood cells in the tunic vessels. No fluorescence was detected in acellular materials in the tunic; tunic cuticle, tunic matrix and hemolymph in the tunic vessels.

Tunic net cells and tunic rounding

Tunic net cells are exclusively found in all species of Polyclinidae and Polycitorinae, and some species of Didemnidae (Fig. 9). In spiculate didemnid species, since the spicules screen other structures in the tunic, I could not conclude whether they have tunic net cells or not. In non-spicu-late didemnid, Diplosoma species, tunic net cells extend the filopodia through a crowd of tunic bladder cells (Fig. 9C).

Fig. 9

Tunic net cells (arrows) in Aplidium pliciferum (A), Eudistoma gilboviride (B), Diplosoma virens (C). Scale bars, 20 μm.

Within 1-day incubation, the shrinkage of the tunic slices occurs in the species of Polyclinidae and Polycitorinae in which tunic net cells are distributed (Fig. 10 and 11). The extent of tunic shrinkage varies among species. For instance, the tunic slices of Polycitor proliferus are turned into spheres, whereas the tunic slices of Aplidium sp. (Cf. sagamiense) shrink in some extent but not become spheres. This assay was not performed in the species of Didemnidae, because the slicing of tunic is difficult due to the presence of spicules, and acid from the ruptured tunic bladder cells may damage the tunic net cells.

Figs. 10 and 11

Tunic shrinkage in Polycitor proliferus (Fig. 10) and Eudistoma parvum (Fig. 11). A, Tunic slices freshly prepared. B, Tunic slices have shrunk into spheres after the incubation in seawater for 20 hr. Scale bar, 1 mm.

Tunic bladder cells

The tunic bladder cell is a large tunic cell in which a large vacuole occupies the bulk of the cytoplasm. This type of tunic cells had been described in several ascidians of the families Ascidiidae and Didemnidae as bladder cells, Blasenzellen, cellules vésiculeuses, and kalymmocytes (see Mackie and Singla, 1987). In the present study, the tunic bladder cell was found in the almost all species belonging to Didemnidae, Holozoinae, Diazoninae, and Ascidiidae, and it was never found in the other groups. Because the tunic bladder cell usually contains acid in the large vacuole, its occurrence is closely related with the acidity of the tunic. The tunic pH is always acidic in the species having bladder cells, except for Ascidia zara and Phallusia mammillata in which the tunic bladder cell does not contain acid in the vacuoles. In the species that do not have the bladder cells, tunic pH is always neutral, except for the species of Corellidae. In the corellid species, there is no acid storage in the tunic, but hemocytes, called giant cells, contain acid. The acidity of the tunic is supposed to be caused by the contamination of the contents of the giant cell in these species.

The acidic tunic does not mean that the pH is acidic in the any areas of the intact tunic. Vital staining with acid-sensitive dyes indicate that the pH in tunic matrix seems to be neutral, and the vacuoles of the bladder cells seems to be the only acid storage site that makes the tunic pH extremely acidic. When the pH is measured with the test paper or pH electrode, the tunic is bruised or injured by the insertion of the electrode. The measured pH is, therefore, the pH of the mixture of the aqueous components in the tunic, the surface seawater, and the acid leaked from the bladder cells. The hemolymph may be also contaminated in some cases.

The tunic acid would be an effective irritant against the predators. Moreover, Stoecker (1980b) indicated that the acid also protects the ascidians from bio-fouling. Many tunic bladder cells are always concentrated beneath the tunic surface, and this distribution pattern is consistent with the defensive function of the tunic acid. In Phallusia nigra, the pH of the tunic surface is also acidic when it was measured with a flat-surfaced pH electrode (Hirose et al., 2001). The acid should have essential functions in some species, because the bladder cells have to spend much energy to produce and maintain the acid. The bladder cells do not contain the acid in two ascidiid species, Ascidia zara and Phallusia mammillata. They would have lost the ability to produce the acid, and it may not be necessary for these species to have tunic acid for survival. In some didemnid ascidians, there are only one or two layers of the bladder cells beneath the tunic surface, while the tunic is full of spicules and there is no space for the bladder cells. In these species, the tunic pH measured with the electrodes is less acidic than other didemnids. Since the spicules have a protective function, the presence of the numerous spicules may cancel out the loss of protection due to smaller amount of the tunic acid.

Stoecker (1980b) reported the tunic pH in 35 ascidian species. Although the occurrence of acidic tunic in the limited families agree with the present results in the main, there are some inconsistencies in the subfamilies Euherdomaninae and Polycitorinae; Stoecker (1980b) reported that the tunic of Pseudodistoma saxicavum (Euherdomaninae), Cystodytes dellechiajei (Polycitorinae), and some Eudistoma spp. (Polycitorinae) is acidic, whereas the tunic of Pseudodistoma kanoko (Euherdomaninae), Eudistoma species (Polycitorinae), and Polycitor proliferus (Polycitoridae) is neutral and the bladder cells are not distributed in the tunic in the present study (see Table 1). Since Stoecker did not describe the acid storage site in the tunic, it is unknown whether these acidic species have the tunic bladder cells containing acid.

Tunic net cells

The tunic net cells extend long filopodia that connect with the filopodia of other tunic net cells, and they form a cellular network in the tunic. This type of cells was referred as myocytes in Diplosoma species (Mackie and Singla, 1987) or elongated tunic cells in Aplidium yamazii (Hirose et al., 1994). In the present study, tunic net cells were found only in some limited groups of aplousobranchians; Polyclininae, Euherdomaninae, Didemnidae, and Polycitorinae.

The contractility of the sliced tunic was assessed in the species having net cells, except for the didemnids. In the all species so far examined, the tunic slices always shrank during the 12–24 hr incubation in seawater, and the tunic slices became spheres in some species. The shrinkage of the sliced tunic is thought to be promoted by the contraction of the network of the tunic net cells in Aplidium yamazii (Hirose and Ishii, 1995), and it seems to be concerned with the wound healing of the tunic (Hirose et al., 1997). The comprehensive occurrence of tunic shrinkage suggests that the contractility promoting tunic shrinkage is a fundamental function of the tunic net cells. Because the extent of shrinkage was different among the species and among the specimens, the amount and distribution pattern of the net cells may be different from the species to species. Mackie and Singla (1987) indicated that the cellular network in the tunic is involved in the conduction of impulses in the tunic in Diplosoma species, and it is possible that the net cells in other species also have the same function.

Phylogenetic considerations

Tunic bladder cells and tunic net cells are found in the particular groups of ascidians, as described above. The present study could not cover the four families of the class Ascidiacea: Agneziidae, Plurellidae, Octacnemidae, and Hexacrobylidae. However, the character-state (presence or absence of these cell types) would be informative to discuss ascidian phylogeny, because of the general stability of the distribution within the families or subfamilies. The class Ascidiacea consists of the two orders, Enterogona and Pleurogona, and this system is supported by the 18S rDNA molecular phylogeny (Wada et al., 1992; Wada, 1998).

Whereas tunic bladder cells and tunic net cells are found in some members of Enterogona that includes the suborder Aplousobranchia and Phlebobranchia, they are not found in Pleurogona so far studied here. Therefore, I suppose that neither tunic bladder cells nor tunic net cells would be distributed in the tunic of the common ancestors of Enterogona ascidians and Pleurogona ascidians, and these cell types are probably synapomorphs in some lineages of ascidians.

In the suborder Phlebobranchia, the tunic bladder cells are found in Diazoninae and Ascidiidae, they are not found in Cioninae, Corellidae, Perophoridae, and there are no data in Agneziidae, Octacnemidae, and Plurellidae. This indicates that Diazoninae and Ascidiidae may have a close relationship. The corellid species have giant cells as hemocytes. The giant cells are very similar in morphology to the tunic bladder cells and they also contain strong acid in their large vacuoles. Because some species of Ascidiidae are known to have the giant cells (Wuchiyama and Michibata, 1995), the family Corellidae may be related to Diazoninae and Ascidiidae. The tunic bladder cells and the giant cells may share a common precursor cells in hemocytes, while the homology of these cells should be carefully verified. Although the hemocytes named “giant cells” are also described in pyurid species (Cf. Dan-Shokawa et al., 1995), the pyurid's giant cells are much different in morphology from the giant cells in Ascidiidae and Corellidae; they do not have large vacuoles containing strong acid and their cytoplasm is full of numerous small vesicles about 0.2–0.6 μm in diameter (Sawada et al., 1991; Ohtake et al., 1994). Therefore, the pyurid's giant cells would have few relationships with the ascidiid and corellid's giant cells and the tunic bladder cells described here.

The character state distribution is more complex in the suborder Aplousobranchia. The tunic bladder cells are found in Holozoinae of Polycitoridae and Didemnidae, whereas the tunic net cells are found in Polyclininae of Polyclinidae, Euherdomaninae of Polyclinidae, Polycitorinae of Polyclinidae and Didemnidae. Neither the bladder cells nor the net cells are found in Clavelininae of Polycitoridae. As a result, the character-states of the three subfamilies of the family Polycitoridae are different from one another. These results may suggest that the suborder Aplousobranchia may be a polyphyletic group, and the classification at the family level would be problematic. The presence of the tunic bladder cells in Holozoinae and Didemnidae may indicate the close phylo-genetic relationship with some phlebobranchian groups bearing the bladder cells (and the giant cells). In Didemnidae, both of the two types of tunic cells are distributed, and this makes the phylogenetic discussion more complex. There may be two probable explanations for this condition; 1) the tunic net cells were acquired in the lineage that had the tunic bladder cells, and then the bladder cells were lost in this lineage except for in Didemnidae, or 2) the tunic bladder cells and tunic net cells were acquired in the different lineages, and the didemnid species acquired the tunic net cells (or the tunic bladder cells) independently. As described above, some spiculate didemnids contain fewer amounts of the bladder cells than other bladder cell-bearing species (Fig. 6), and the bladder cells might be lost the ability to produce the acid in Ascidia zara and Phallusia mammillata (Fig. 7). Moreover, spiculate didemnids may not have the tunic net cells. These observations may support the occurrence of the loss of these tunic cells in some lineages. However, I suppose that the loss of these cell types did not occur so frequently, because the character-state distribution (presence or absence of the bladder cells and net cells) is stable within families or subfamilies.

Although the phylogenetic relationships among families and subfamilies are still uncertain, the genus Ciona is generally regarded as the most primitive among extant ascidians (Cf. Millar, 1966; Kott, 1969). Ciona is the only genus of the subfamily Cioninae, and neither the tunic bladder cells nor the tunic net cells are found in the two Ciona species. On the contrary, the tunic bladder cells are found in Rhopalaea sp. belonging to the subfamily Diazoninae that is the only sister subfamily of Cioninae. The species of Clavelininae, in which neither the bladder cells nor the net cells are distributed, are generally regarded to have primitive characteristics among aplousobranchian species. On the other hand, the didemnid species that have both of the two cell types are thought to be an advanced group (Kott, 1969).

The phylogenetic analyses based on the 18S rDNA sequence indicated that the Enterogona ascidians are more closely relate to the members of the class Thaliacea (salps, doliolids, and pyrosomas) than to the Pleurogona ascidians including stolidobranchians (Wada, 1998; Swalla et al., 2000). However, these analyses do not include any data of the species of the suborder Aplousobranchia. The occurrence of the tunic bladder cells spreads over the Enterogona, and this may support the close relationships between the phleboranchians and aplousobranchians. It should be also noted that the tunic net cells are found in the all of three genera in Pyrosomata (Thaliacea) but not in salps and doliolids (Hirose et al., 1999; Hirose et al., unpublished). It would be possible that the pyrosomas share a common ancestor with some aplousobranchians having the tunic net cells.

As for the taxa so far examined in this study, I temporarily propose two possible trees of ascidian phylogeny on the basis of the presence or absence of the tunic bladder cells and the tunic net cells (Fig. 12). These trees also include following assumptions; 1) the families or the subfamilies are monophyletic, 2) the common ancestor of ascidians has neither tunic net cells nor tunic bladder cells, 3) the lineage of stolidobranchians was branched before the occurrence of diversification in Enterogona (Cf. Wada, 1998; Swalla et al., 2000), 4) some aplousobranchians share a common ancestor with colonial species of Diazoninae (Cf. Kott, 1969). In one tree, the net cell is supposed to be acquired in the lineage that had acquired the bladder cells (Fig. 12A), and the two types of tunic cells are supposed to be acquired in the different lineages in the other tree (Fig. 12B). In either tree, both of the suborders Phlebobranchia and Aplousobranchia are not monophyletic, suggesting the necessity of reconsideration about the phylogenetic relationships among the families/sub-families in the order Enterogona. In addition, pyrosomas may be closely related with some groups of aplousobranchians that have the tunic net cells.

Fig. 12

Two possible trees of ascidian families and subfamilies studied here on the basis of the presence or absence of the tunic bladder cells and the tunic net cells with some assumptions (see text). The tree A supposes that the two types of tunic cells were acquired in the same lineage, and the tree B supposes that the two types were acquired in the different lineages. Asterisk, the presence of the tunic bladder cells; #, the presence of the tunic net cells.

These phylogenetic trees include the many assumptions, and the data set of the tunic cells, on which the trees are based, does not cover the four families. Moreover, the observation in only one species represents the character distribution in some taxa, such as Diazoninae, Perophoridae, Holozoinae, and Euherdomaninae. These phylogenetic trees, however, can point out the problems of the traditional phylo-genetic system. Since the molecular phylogeny suggests that thaliaceans are closely related to Enterogona (Wada, 1998; Swalla et al., 2000), the phylogeny in this order is very important to understand the evolutionary course of the subphylum Tunicata and to estimate the ancestral forms of the chordates. As well as the traditional approach, molecular phylogenetic study is surely needed especially in aplousobranchians, for better understanding of the tunicate phylogeny.