Animals

Onchidium sp. (Gastropoda, Mollusca, 3–5 g body weight) were collected on the seashore of Tateyama (Chiba Prefecture) between June to August and maintained in a tank of aerated artificial sea-water (Tetra Marine) under natural light/dark and an additional 12hrlight/12hr-dark cycle by a submergible NS-fluorescent lamp (Nissei, 15W) at 20−23°C. The intensity of illumination on the water surface at the center of the tank was about 600–1000 lux but varied depending on the weather. Animals inhabit under the stones in the artificial seawater or clung to the corners of the tank except for feeding time.

HPLC analyses

Onchidium were dark-adapted overnight before use. SEs and DEs were excised under deep red light with the scissors under a Wild stereoscopic microscope and frozen immediately. Samples were maintained at −80°C in the dark until use. Light-adapted animals were picked up from the tank at noon, then, DEs were excised from the dorsal mantle under a ring-type fluorescent lamp (SZ-FLR, 10W, about 1500–2000 lux) of a binocular microscope (Olympus SZ 4045) and frozen, similar to the dark-adapted samples. Retinals were extracted and analyzed according to the method of Makino-Tasaka and Suzuki (1986). Seventeen dark-adapted SEs, 50 dark-adapted DEs and 50 light-adapted DEs were homogenized in 0.2 ml phosphate buffer (0.06 M, pH 6.8) and 0.1 ml 1M hydroxylamine (neutralized with NaOH). Methanol (1ml) was added to the homogenate to convert retinals to retinaloximes. Oximes were extracted three times with 1 ml dichloromethane and 2 ml n-hexane. Extracts were evaporated under reduced pressure, dissolved in 100 μl elution solvent and subjected to HPLC, using the JASCO system equipped with a Zorbax SIL column (2.1×250mm, Shimadzu). The elution solvent used was 7% ether in n-hexane containing 0.075% ethanol, at a flow rate of 0.6 ml/min. Absorbance at 360 nm was monitored with a UV monitor (JASCO UVIDEC 100-III). Peak areas were determined by integrating the absorbance at 360 nm with a Chromatopac E-1A integrator (Shimadzu). The 11-cis and all-trans-retinal isomers were quantified from peak areas and standard curves determined with authentic compounds.

Fluorescence histochemistry

SEs and DEs were isolated from animals dark-adapted overnight, embedded in Tissue Tek OCT compound (Sakura Finetechnical Co.) and frozen immediately by immersion in liquid nitrogen under red light. Cryosections of both eyes (10 μm thickness) were mounted onto the same 3-aminopropyltriethoxy-silan-coated glass slides and simultaneously treated. Sections were air-dried for 3 hr at room temperature. Some specimens were immersed in 0.2% NaBH₄ for 1 sec at 4°C to induce fluorescence by conversion of retinochrome to N-retinyl protein, and rinsed gently with water. Other specimens were treated with a 20% formaldehyde aqueous solution for 3 min at room temperature to induce rhodopsin fluorescence, immersed in methanol for 2 min and 0.2% NaBH₄ for 5 sec at 4°C, and rinsed with water after each step. Fluorescence was observed with an epifluorescence microscope (Nikon Optiphoto EFD) through a L-420 filter to exclude reflected light upon excitation. To excite samples, UV light (334-365 nm) from a 100W high-pressure mercury lamp was isolated by combining a UG-1 filter and DM-400 dichromatic mirror. Since the fluorescence from reduced retinal proteins faded very quickly, it was difficult to record data from Onchidium samples with a conventional camera attached to an epifluorescent microscope. For recording a drastic change in faint fluorescence, the image observed with an epifluorescent microscope was amplified using a highly sensitive image intensifier (Hamamatsu Photonics, V1366P) and recorded with a high-speed VTR system (nac, MHS-200). Still pictures were subsequently photographed from VTR. Each fluorescent micrograph (Figs. 4, 5, 7 and 8) was printed from the same film under similar conditions to compare fluorescence

I. HPLC analyses

Stalk eye: Chromophores of photopigments in the 17 dark-adapted SEs were extracted as retinaloximes. HPLC analyses revealed the presence of both 11-cis-retinal and all-trans-retinal, but no 3-dehydroretinal in SE (Fig. 1). The total amount of retinal was calculated as 6.90 pmol (average of 0.41 pmol/SE), of which 11-cis- retinal comprised 2.85 pmol (average of 0.17 pmol/SE) and all-trans-retinal, 4.05 pmol (average of 0.24 pmol/SE). The proportion of all-trans-retinal was greater (58.7%) than that of 11-cis-retinal (41.3%).

Fig. 1

High–performance liquid chromatogram of stalk eye extract (a) and standard oximes (b). (a) Seventeen dark-adapted stalk eyes (SE) contained 6.90 pmol retinal. The proportions of 11-cis- and all-trans-retinal were 41.3% and 58.7%, respectively. The average amount of 11-cis-retinal was 0.17 pmol, while that of all-trans-retinal was 0.24 pmol per SE. (b) Standard oximes (5 pmol each). 11₁, 11-cis-retinaloxime; 11₂, 11-cis-3-dehydroretinaloxime; All₁, all-trans-retinaloxime; All₂, all-trans-3-dehydroretinaloxime.

Dorsal eye: Chromophores of photopigments were extracted as retinaloximes from the 50 dark-adapted and 50 light-adapted DEs. HPLC analyses similarly revealed 11-cis-retinal and all-trans-retinal, but no 3-dehydroretinal in DE (Fig. 2). In light-adapted DE, the total amount of retinal was calculated as 2.05 pmol (average of 0.04 pmol/DE), of which 11-cis-retinal was 1.10 pmol (average of 0.022 pmol/DE) and all-trans-retinal was 0.95 pmol (average of 0.019 pmol/DE). The relative proportions of 11-cis-retinal and all-trans-retinal were 53.7% and 46.3%, respectively (Fig. 2a). In the dark-adapted DE, the total amount of retinal was 8.64 pmol (average of 0.17 pmol/DE), of which 11-cis-retinal comprised 5.18 pmol (average of 0.10 pmol/DE) and all-trans-retinal, 3.46 pmol (average of 0.07 pmol/DE). In the dark-adapted DE, the proportions of 11-cis-retinal was greater (59.9%) than that of all-trans-retinal (40.1%) (Fig. 2b).

Fig. 2

High–performance liquid chromatogram of light-adapted (a) and dark-adapted (b) dorsal eye extracts. (a) Fifty light-adapted dorsal eyes (DE) contained 2.05 pmol retinal. The proportions of 11-cis- and all-trans-retinal were 53.7% and 46.3%, respectively. The average amount of 11-cis-retinal was 0.022 pmol, while that of all-trans-retinal was 0.019 pmol per DE. (b) Fifty dark-adapted DEs contained 8.64 pmol retinal. The proportions of 11-cis- and all-trans-retinal were 59.9% and 40.1%, respectively. The average amount of 11-cis-retinal was 0.10 pmol, while that of all-trans-retinal was 0.07 pmol per DE. 11₁, 11-cis-retinaloxime; All₁, all-trans-retinaloxime.

In dark-adapted eyes, the amount of retinal of one DE was far less (two-fifth) than that of one SE. Furthermore, in light-adapted DE, the total amount of retinal was far less (one-fourth), with a slightly lower proportion of 11-cis to all-trans-retinal, than in dark-adapted DE (Figs. 2a, 2b).

II. Fluorescence histochemistry

Stalk eye: The SE is an ellipsoidal vesicle of about 200 μm in diameter, comprising a cornea, lens and retina (Figs. 3a, 4a, 5a). The fine structure of the SE has been described in detail elsewhere (Katagiri et al., 1995). The rhabdomeric-type photoreceptor cell extending well-developed microvilli in the SE is designated the ‘type 1 visual cell’. The retina is a pigmented columnar epithelium divided into villous, pigmented, somatic and neural layers (Figs. 3a, 4a, 5a). The villous layer consists of an enormous tuft of microvilli of type 1 visual cells. The dark pigmented layer corresponds to the distal heavily pigmented cytoplasm of visual cells and pigmented supportive cells, while the somatic layer contains the nuclei and cytoplasm of visual and pigmented supportive cells. In the neural layer, axons of visual cells converge to form the optic nerve.

Fig. 3

Schematic drawings of stalk eye (a) and dorsal eye (b) of Onchidium. (a) Stalk eye is a closed vesicle, consisting of a cornea (C), lens (L) and retina (R). The retina is conveniently divided into villous (VL), pigmented (PL) somatic (SL) and neural (NL) layer. The PL corresponds to the distal cytoplasm of visual cells and pigmented supportive cells which contain abundant dark pigment granules. Incident light enters from the top, corneal side. M, microvilli of type 1 visual cell; ON, optic nerve. (b) Dorsal eye is an open vesicle. It comprises a pigment layer (PL), an inverted retina (R) and a lens (L) but lacks cornea. The retina is a single layer of ciliary-type photoreceptor cells of which the distal outer segments (OS) show a concentric lamellar structure and face the PL. Paranucleus (P) is located near the base of the OS. Axons extending from ciliary-type photoreceptor cells converge to the optic nerve (ON). The pear- shaped lens is located in the central portion of the eye. It is mostly composed of several large lens cells, which are rhabdomeric-type photoreceptors. The distal portion of the lens cell is composed of massive microvilli (M). Incident light enters from the microvilli of the lens. N, nucleus of lens cell.

Fig. 4

(a) Differential interference light micrograph of the stalk eye (SE) was taken from the same section as that in Fig. 4b–d after VTR recording. The pigmented layer (PL, lines) is located between the inner villous (*) and outer somatic layer (SL, ▴). To identify the location of fluorescence in the SE, the height of the PL is drawn with the lines and the outline of the SE is depicted with dots. CT, connective tissue. (b)–(d) Fluorescent micrographs of the stalk eye (SE) treated with sodium borohydride. (b) Immediately after irradiation, a yellow-green fluorescent band was observed in the SL (▴) under the PL. The lens (L) displayed strong autofluorescence. (c) The fluorescence image of the same specimen (Fig. 4b) at 10 sec after UV irradiation. Fluorescence in SL (▴) diminished rapidly. (d) After 30 sec irradiation, fluorescence derived from retinochrome in the SL (▴) disappeared.

Fig. 5

(a) Differential interference light micrograph of the stalk eye was taken from the same section as that in Fig. 5b-d after VTR recording. See abbreviations of Fig. 4a. (b)–(d) Fluorescence micrographs of the stalk eye (SE) treated with sodium borohydride after denaturation with formaldehyde and methanol. (b) Upon UV irradiation, yellow-green fluorescence of the retino-chrome was visualized in the SL (▴) under the PL. Additional fluorescence was observed in the VL (*) along the internal side of PL. Fluorescence in the SL was stronger than that in VL. The optic density of three circled areas, the villous layer (v), somatic layer (s) and connective tissue (c), were examined. The lens displayed intense autofluorescence. (c) At 10 seconds after irradiation, fluorescence of the retinochrome considerably faded in the SL (▴), but that of rhodopsin in the VL (*) disappeared. Autofluorescence of the lens was observed throughout. (d) At 30 seconds after irradiation, the fluorescence of the SL (▴) remained faintly, although endogeneous fluorescence of the lens and connective tissue remained.

Upon immersion in water or formaldehyde solution, autofluorescence was observed in the epidermis and its surface and in unknown cells in the connective tissue of the tentacle. The SE does not show autofluorescence except for the lens. The particularly strong autofluorescence of the lens interfered with the intrinsic fluorescence of the villous layer in the retina when it was treated with NaBH₄ with or without denaturation. When cryosections of the SE were treated with NaBH₄, yellow-green fluorescence was observed in the somatic layer, which faded rapidly and disappeared completely within 60 sec after UV irradiation (Figs. 4b–d). Upon NaBH₄ treatment after denaturation, additional fluorescence was observed in the villous layer (Fig. 5b). Fluorescence in the villous layer was weaker than that in the somatic layer. At 10 sec after irradiation, the yellow-green fluorescence in the villous layer had almost disappeared, but remained weakly in the somatic layer (Fig. 5c). Fluorescence in both layers faded considerably at 30 sec (Fig. 5d), disappearing completely within 60 sec after UV irradiation, although the endogeneous fluorescence of the lens and connective tissue was observed throughout.

A typical case of quick decay of fluorescence derived from photopigments is shown in Fig. 6. Time courses of optical density corresponding to a change in fluorescence were analyzed at three spots in a sample of Fig. 5, the villous and somatic layers in SE and connective tissue, that was treated with NaBH₄ treatment after denaturation with formaldehyde and methanol. The lifetime of decay of fluorescence intensity in both villous and somatic layers differed from that in connective tissue. The optical density of fluorescence from the villous and somatic layers decreased to 30% (τ-13) and 35% (τ-14) after 15 sec, while that from connective tissue was maintained at 85% (τ-104) during exposure to light (Fig. 6).

Fig. 6

Time-course of the optical density that corresponds to fluorescent intensity in two portions of the stalk eye (SE) and connective tissue (CT). Three circled portions were examined; the villous (v) and somatic (s) layers, and the CT (c) in Fig. 5b. The lifetime of decay of fluorescence in the SE differs from that of CT. Fluorescence from the villous (○) and somatic (•) layers declined within 15 sec to about 35%, but that from CT (×) did not change considerably during irradiation. The time constants (τ) of the villous and somatic layers were 13 and 14 sec respectively, while that of CT was 104 sec.

Dorsal eye: The DE is spherical in shape (not a closed vesicle) and about 200 μm in diameter. The DE comprises a cup-shaped pigment layer, inverted retina and lens, but lacks cornea (Figs. 3b, 7a, 8a). The retina is composed of a single columnar epithelium of only ciliary-type photoreceptor cells. The distal portion (outer segment) of the ciliary-type photoreceptor cell shows a concentric lamellar structure and faces the pigmented cells of the pigment layer. The supra-nuclear region contains rough and smooth endoplasmic reticulum, several paranuclei (dense inclusions) and other organelles. A biconvex pear-shaped transparent lens is situated in the center of DE. It consists of several large lens cells. Each lens cell contains massive microvilli in the distal portion, numerous lamellar bodies and a network of endoplasmic reticulum in the proximal cytoplasm. Based on fine structure and photoresponse data, the lens cell is identified as a rhabdomeric-type photoreceptor cell (Katagiri et al., 1983, 1985, 1998).

Fig. 7

(a) Differential interference light micrograph of the dorsal eye corresponds to Figs. 7b-d. The dorsal eye (DE) consists of the pigmented layer (PL, lines), retina (R) and lens (L). To identify the location of fluorescence in the DE, the height of the PL is drawn with the lines and the outline of the lens is depicted with dots. The outer segment (*) of ciliary-type photoreceptor cell faces to the PL. An arrow shows para-nuclei of ciliary-type photoreceptor cell. CT, connective tissue; E, epidermis; L, lens cell; M, microvilli of lens cell; ON, optic nerve. (b)–(d) Fluorescence micrographs of the dorsal eye treated with sodium borohydride. (b) Immediately after irradiation, faint, obscure yellow-green fluorescence was observed in the entire dorsal eye, including the retina and lens. Strong spotty autofluorescence (arrow) was frequently detected in the supranuclear regions of ciliary-type photoreceptor cells in the retina. (c) After 10 sec and (d) 30 sec irradiation, almost all fluorescence disappeared in the DE, except for non-specific fluorescence in the epidermis (E).

Fig. 8

(a) Differential interference light micro-graph of the dorsal eye corresponds to Figs. 8b–d. See abbreviations in Fig.7a. (b)–(d) Fluorescence micrographs of the dorsal eye (DE) treated with sodium borohydride after denaturation with formaldehyde and methanol. (b) Upon UV irradiation, the yellow-green fluorescence of N-retinyl proteins was observed indistinctly in the DE. Specific fluorescence derived from rhodopsin did not appear in the retina or microvilli (M) of lens cells. (c) After 10 sec and (d) 30 sec irradiation, non-specific fluorescence remained in the DE. Diminishing of specific fluorescence derived from photopigments was not detected in the DE, although endogeneous fluorescence remained in the epidermis (E) and connective tissue.

In control sections that were immersed in water or form-aldehyde solution, autofluorescence was observed in the epidermis, its surface and in cuboidal cells in the connective tissue of the dorsal mantle, which did not disappear with reducing treatment. When non-specific fluorescence was evident in the surrounding connective tissue, weak, obscure fluorescence overspread diffusely the DE (not shown). The paranuclei of the ciliary-type photoreceptor cell in the retina emitted strong autofluorescence that faded slightly in the control sections and rapidly after NaBH₄ treatment (Fig. 7b–7d) and was not observed after methanol treatment (Fig. 8b) upon UV irradiation.

Fluorescence in DE after treatment with NaBH₄ with or without denaturation was observed in samples that showed less autofluorescence in connective tissue. Upon NaBH₄ treatment, unclear fluorescence was observed in indefinite and vague areas in individual DE at various intensities and different fading times. Faint fluorescence observed in the retina and the distal microvilli of the lens faded within 30 sec (Figs. 7b–7d). Upon NaBH₄ treatment after denaturation with formaldehyde and methanol, vague fluorescence was observed (Fig. 8b). However, additional fluorescence was either not present or very faint in both the lens and retina. Non-specific fluorescence persisted in the epidermis and connective tissue around the DE (Figs. 7c–d, 8c–8d). The drastic fading of fluorescence within 30sec under UV irradiation characteristic of reduced photopigments was not detected in the DE, despite recording with a high-speed VTR system (Figs. 8c–8d).

I. Chromophores in the stalk and dorsal eyes of Onchidium

HPLC analyses in the present study revealed 11-cis-and all-trans-retinal in the SE of dark-adapted Onchidium and the amount of all-trans-retinal was higher than that of 11-cis-retinal. The all-trans-retinal is considered to be a chromophore of retinochrome (Hara and Hara, 1973; Ozaki et al., 1983, 1986). The rhabdomeric-type photoreceptor cells of gastropoda eyes, including Onchidium SE, contain abundantly photic vesicles in the cytoplasm (Eakin, 1972, 1990; Kataoka, 1975, Herman and Strumwasser, 1984; Katagiri et al, 1995). As photic vesicles contain retino-chrome (Hara and Hara, 1973; Ozaki et al., 1986), the all trans-retinal detected in SE by the present HPLC analyses is possibly derived from the retinochrome in photic vesicles of type 1 visual cells in the retina. Furthermore, 11-cis-retinal in the dark-adapted SE may be the chromophore of rhodopsin of type 1 visual cells of which the microvilli occupy almost all the villous layer (Katagiri et al., 1995). Numerous SEs were required for HPLC analyses due to their minute size. However, we could not obtain sufficient amounts of SE, and therefore could not examine the chromophores of the light-adapted SE. In the eye of the marine gastropoda, Conomulex, 11-cis- and all-trans-retinal were also detected by HPLC analyses (Ozaki et al., 1986). The 11-cis-retinal was dominant in extracts from the microvillar fraction, while all-trans-retinal was prevalent in extracts from the photic vesicle-containing fraction. These results indicate that the photopigment is rhodopsin in the former fraction and retino-chrome in the latter. The amount of all-trans-retinal in the Conomulex eye is several times greater than that of 11-cis-retinal. The observed ratio of retinal isomers in chromophores of Onchidium SE is similar to that in Conomulex eye (Ozaki et al., 1986).

On the other hand, in the dark-adapted Onchidium, the total amount of retinal in one SE was 2.4 times higher than that in one DE. A pair of SEs is situated in the tentacles. In contrast, more than 50 DEs are distributed on the dorsal mantle surface (Katagiri et al., 1985). Therefore, DE appears to be superior in terms of total numbers and consequently, total amount of retinal in an individual Onchidium. However, the SE is regarded as the principal eye within the multiple photoreceptive system (Hisano et al., 1972, Katagiri et al., 1985, 1990) on the basis of the presence of sufficient retinal per SE, existence of a rhodopsin–retinochrome system (Katagiri et al., 2001) and the typical structure of the gastropoda eye (Eakin, 1972, 1990; Kataoka, 1975; Jacklet and Colquhoun, 1983; Herman and Strumwasser, 1984; Katagiri et al., 1995).

The present study is the first significant report showing that the Onchidium DE contains 11-cis- and all-trans-retinal as well as SE. The DE comprises two different types of photoreceptor cells, specifically, ciliary-type in the retina and rhabdomeric-type in the lens. Since it is very difficult to separate the retina and lens in the DE, HPLC analysis was performed on all photopigments derived from both ciliaryand rhabdomeric-type photoreceptor cells. It is proposed that as the retina is the essential receptive site in DE, the HPLC data mainly represent the chromophores of photopigments in ciliary-type photoreceptor cells. The proportion of 11-cis-retinal (59.5%) to total retinal in dark-adapted DE was characteristically greater than that in dark-adapted SE (41.3%). Light-adapted DE contained a decreased proportion of 11-cis-retinal to all-trans-retinal compared to dark-adapted DE, suggesting the recycling of retinals in the DE. Notably, total retinal in light-adapted DE was far less (one-fourth) than that in dark-adapted DE. This decreased amount of total retinal could be explained on the assumption that predominant ciliary-type photoreceptor cells in DE contain vertebrate-type visual pigments. In vertebrate eyes containing ciliary-type photoreceptors, all-trans-retinal is released from the protein opsin during light adaptation, which is converted into retinol and subsequently stored as retinylesters in pigment epithelial cells (Saari, 1994). It is possible that all-trans-retinal in the ciliary-type photoreceptor cells of light-adapted DE is released from protein and stored as retinylesters in adjacent pigmented cells that directly face the outer segments of ciliary-type cells. At the same time, retinylester might be stored in the paranuclei of ciliary-type photoreceptor cells, similar to the oil droplets reported as the storage site of retinylester in crayfish retina (Suzuki et al., 1988). Unfortunately, our HPLC analyses did not encompass data on retinylesters, since these are eluted at the solvent front and therefore not quantified with the present experimental conditions (Suzuki et al., 1988). Retinals detected in the light-adapted DE may be derived mainly from the photopigments (rhodopsin and retinochrome) of rhabdomeric-type lens cells in the DE. Rhabdomeric-type lens cells have massive microvilli, and contain no photic vesicles but abundant endoplasmic reticulum, which is the storage site of retinochrome (Katagiri et al., 1998). The photic vesicle is a special form of endoplasmic reticulum (Whittle, 1976). Photic vesicle and endoplasmic reticulum were impregnated by prolonged osmification that demonstrates the presence of lipids, such as vitamin A (Eakin and Brandenburger, 1970; Pourcho and Bernstein, 1975; Katagiri 1984, 1998). In the Onchidium DE, several coordinating systems for photopigment recycling may be present, such as those in ciliary-type and rhabdomeric-type photoreceptor cells. The rhodopsin-retinochrome system was reported in the SE (Katagiri et al., 2001). Since 11-cis and all-trans-retinal were detected in both SE and DE, the rhodopsin-retino-chrome system may be present in the DE similarly.

II. Fluorescence histochemistry

The localization of retinal protein in the DE was examined by fluorescence histochemistry and compared with that of SE. Fluorescence histochemistry is a convenient and valuable method for detecting photopigments, since the method only requires treatment with reducing agent and not antibodies (Ozaki et al., 1983). Fluorescence of N-retinyl protein that was reduced by borane dimethylamine was reported formerly in the snail eye (Eakin and Brandenburger, 1978). Ozaki et al. (1983, 1986) and Hara et al. (1992) demonstrated the localization of rhodopsin and retinochrome, based on differences in their reactivity to NaBH₄ in several invertebrate eyes. In Onchidium SE, fluorescence derived from rhodopsin and retinochrome was observed in villous and somatic layers of the retina. This fluorescence faded rapidly, in concurrence with earlier observations on the retinas of several gastropods (Ozaki et al., 1983, 1986). The stronger fluorescence in the somatic layer of Onchidium SE than that in the villous layer confirmed that the amount of retinochrome is greater than rhodopsin, as shown by HPLC analyses.

The SE from dark-adapted Onchidium contained sufficient retinal (0.41 pmol/SE) to analyze emitted fluorescence after NaBH₄ treatment. However, in the DE (0.17 pmol/DE) from dark-adapted animals, it is difficult to identify the presence or absence of photopigment-specific fluorescence. In SE, the localization of photopigments was confirmed by both immunohistochemistry and fluorescence histochemistry (Katagiri et al., 1995, 2001). Therefore, SE may be employed as a standard positive control in current fluorescence histochemistry analyses. We simultaneously treated DE and SE mounted on the same glass slide using similar procedures to examine the localization of fluorescence in DE. In SE, the fluorescence from rhodopsin and retino-chrome was observed in specific regions after reducing treatment with or without denaturation, whereas in DE, faint diffuse fluorescence was observed and did not always correspond to the photoreceptive sites, such as the outer segments of ciliary-type photoreceptor cells and microvilli of the lens cells. Ohkuma and Tsuda (2000) developed a time-resolved fluorescence difference imaging method to visualize retinal proteins in tissues containing low quantities of retinal proteins but strong endogenous fluorescence. Octopus retina and ocellus of ascidian larva were treated according to the method of Ozaki et al. (1983). The subtracted fluorescence of N-retinyl protein before and after UV irradiation effectively distinguished the real change in fluorescence of bleached photopigments in these specimens (Ohkuma and Tsuda, 2000). Using the time-resolved difference fluorescence imaging method, real fluorescence changes in Onchidium SE were visualized in the villous and somatic layers, although the fluorescence of the connective tissue did not change (Katagiri et al., 2000). However, even this method was unsuccessful in visualizing fluorescence in specific regions of DE (Katagiri et al., unpublished data), similar to the present study.

It is very difficult to explain clearly why the fluorescence histochemistry technique was unsuccessful in visualizing the fluorescence of photopigments in the DE. There are a number of conceivable reasons for the unexpected obscure fluorescence of DE. Particularly, the amount of photopigment in the DE may be too low for fluorescence microscopy as shown by the present HPLC analysis. Non-specific unknown fluorescence found in the dorsal mantle tissue may cover the intrinsic fluorescence derived from photopigments. It is presumed that the nature of photopigments in DE may be different from that of SE. While fluorescence derived from rhodopsin was distinctly recognized in the rhabdomeric-type visual cells of the SE, it was hardly observed in ciliary-type photoreceptor cells in the DE. In DE, rhodopsin may be reduced only by NaBH₄ treatment without denaturation and may therefore diminish rapidly. It is presumed that the fluorescence in the lens cell of DE that appears after NaBH₄ treatment (Fig. 7b) is partially derived from rhodopsin. It should be noted that the chromophore of rhodopsin in both photoreceptors of the DE may be lost by formaldehyde/methanol treatment. On the other hand, the spot-like autofluorescence in the retina of DE was observed in the para-nuclei of ciliary-type photoreceptor cells. This specific autofluorescence disappeared after methanol treatment, indicating that the autofluorescence derived from methanol soluble substance such as retinyl-ester. It is known that retinyl-ester emits the fluorescence in the tissue without NaBH₄ treatment, although rhodopsin and retinochrome emit the yellow-green fluorescence after treatment with NaBH₄ with or without denaturation, respectively (Ozaki et al. 1983). The spot-like autofluorescence in Fig. 7b is possibly due to retinyl-ester. The paranucleus is a membrane-bound heterogeneous structure including an electron dense portion, and is one of the presumed storage sites of retinylester. Upon NaBH₄ treatment, fluorescence in the spot-like structure partly decreased. The ciliary-type photoreceptor cells contain a small number of endoplasmic reticulum distributed around the paranucleus. Therefore, we suppose that the retinochrome in the endoplasmic reticulum participates in fading fluorescence.

The present study demonstrates that both SE and DE of Onchidium have a common photopigment system, a rhodopsin-retinochrome system, containing the 11-cis- and the all-trans-retinal as chromophore. However, there are several important differences between the two eyes, specifically in terms of photoreceptor cell type, photoresponse, spectral sensitivity (Katagiri et al., 1985) and the appearance of fluorescence derived from photopigments. Photopigments in photoreceptor cells that comprise the Onchidium multiple photoreceptive system may differ in chemical nature.