Selective Utilization of Serum Vitamin A for Visual Pigment Synthesis

Goldfish (Carassius auratus) possess a pair of scotopic visual pigments: rhodopsin (opsin conjugated to vitamin A₁ aldehyde) and porphyropsin (opsin conjugated to vitamin A₂ aldehyde) (Tsin & Beatty, 1978, 1979; Tsin et al. 1981; Tsin & Santos, 1985; Tsin & Flores, 1986). (Vitamin A is used collectively to represent all vitamin A₁ and A₂ compounds. Vitamin A aldehyde = retinal; vitamin Ai alcohol = retinol; vitamin A₁ esters = re tinyl esters; vitamin A₂ aldehyde = 3,4-didehydroretinal; vitamin A₂ alcohol = 3,4-didehydrore tinol; vitamin A₂ esters = 3,4-didehydroretinyl esters.) The relative proportions of these two visual pigments in the goldfish neuroretina are dependent on temperature and light (intensity and photoperiod) and can be altered by exogenous thyroxine (Tsin & Beatty, 1978, 1979; Tsin et al. 1981; Tsin & Flores, 1986). The mechanism underlying these environmental influences on visual pigments is not known at present (Suzuki et al. 1985; Tsin et al. 1985; Tsin & Flores, 1986; Tsin, 1986a,b; Tsin & Chambers, 1988).

In studying the biochemical basis for changes in the proportions of these two visual pigments in the goldfish neuroretina, the vitamin A composition (vitamin A₁/A₂ ratio) in the retinal pigment epithelium (RPE; which is located adjacent to the retina) was measured by HPLC (Tsin et al. 1983, 1986b). These earlier studies show that the proportions of vitamin A in the RPE are not reflected in the proportions of the two scotopic visual pigments (rhodopsin and porphyropsin), suggesting that the stored vitamin A in the RPE does not determine the rhodopsin/porphyropsin ratio in the neuroretina.

An important issue yet to be resolved is the composition of vitamin A in the blood and its relationship to the visual pigments in the eye. Exactly how plasma vitamin A enters the RPE and participates in the visual cycle is not understood (Bok, 1985). However, it has been reported that vitamins A₁ and A₂ are both transported by serum retinol-binding protein (Shidoji & Muto, 1977) which binds with RPE cells to deliver vitamin A to the eye (Bok, 1985). Is it possible that the proportion of retinol and 3,4-didehydroretinol in the blood which supplies vitamin A to the eye for visual pigment formation determines the composition of visual pigments?

In the present study, HPLC and spectrophotometry were employed to measure vitamin A levels in the blood, RPE and liver as well as visual pigments in the retina. Our results suggest that the mechanism regulating changes in the proportion of the two visual pigments may not be related to a selective transport of vitamin A in the blood. Instead, it is possible that visual pigment changes are regulated by biochemical processes located in the neuroretina and the RPE.

Goldfish were purchased from Grassyfork Fisheries (Martinsville, Indiana; 10-12·5 cm; common and comets). Upon arrival, one group of goldfish was housed in a rectangular aquarium at 30°C under dim light (16h:8h L:D cycle; from a 7·5 W light bulb; see Tsin & Beatty, 1979, for the light intensity) to induce rhodopsin. Another group was subjected to a similar light regimen at 20°C to ensure pure porphyropsin. Animals were killed after 3 weeks of light and temperature treatments. During the experiment, animals were fed daily with TetraFin goldfish food (Echo Products, Ortonville, Michigan).

All procedures involving retinal tissue were carried out in a dark room using a dim red light. The eyes of an individual were removed following overnight dark adaptation. Retinae were dissected and combined in cold 4% alum. The preparation was centrifuged slowly, the supernatant buffer discarded, and the remaining tissue rinsed twice with cold distilled water. 900 μl of 2% digitonin in phosphate buffer (pH 7) was added to the retinal tissue which was homogenized in a glass-glass homogenizer. The mixture was then incubated at 4°C for 1h, followed by centrifugation at 12000g (SS34, for 30min). 100 μl of 0·1 moll⁻¹ hydroxylamine (neutralized with NaOH) was added to the supernatant before the absorbance was recorded (at 700-300 nm) in a spectrophotometer. The sample was then exposed to room light (bleached) for 10 min and a subsequent absorbance curve was recorded. The difference in the absorbance of the pigment extract before and after bleaching was then calculated by an Apple computer to derive a difference spectrum (see Fig. 1B). The amount of visual pigment present in the retinal extract and the rhodopsin/porphyropsin ratio were calculated from the difference spectrum according to a method described earlier (Munz & Beatty, 1965; Tsin & Beatty, 1978).

The retinal pigment epithelium/choroid (RPE) and liver were also removed and dehydrated using anhydrous sodium sulphate. Vitamin A in these tissues was extracted with acetone. Acetone was removed with a stream of nitrogen and the sample redissolved in 1ml of 0·5% dioxane in n-hexane. Each sample was introduced into a 5% water-deactivated alumina column (1 cm x 2 cm) from which vitamin A esters (eluting with 10 ml of 0·5% dioxane in n-hexane) and alcohols (eluting with 10 ml of 10% dioxane in n-hexane) were collected. These vitamin A fractions were then analysed isocratically by HPLC (Tsin et al. 1984, 1985). Vitamin A esters were saponified according to the method of Bridges & Alvarez (1982) before HPLC analysis. Vitamin A was quantified in HPLC by comparing the peak area to that of known standards.

Blood samples were removed from caudal vessels in a 2 ml syringe fitted with a 22 gauge needle and centrifuged for 1 min to separate cells from the plasma. 100 μl of water was added to 100 μl of plasma before 200μl of ethanol was added to denature plasma proteins, including retinol-binding protein. Vitamin A was extracted in three portions of 800μl of petroleum ether and the vitamin A extract was then purified on a 5% water-deactivated alumina column (similar to the method stated above) prior to HPLC analysis.

In agreement with earlier reports (Tsin & Beatty, 1978, 1979; Tsin et al. 1981; Tsin & Flores, 1986), the present study shows that goldfish responded to temperature treatment by altering the ratio of the two visual pigments in their neuroretinas. Upon arrival from the supplier, goldfish had pure porphyropsin in their eyes. After 3 weeks, the neuroretinas of those animals subjected to the lower temperature (20 °C) remained porphyropsin-rich (98·5% porphyropsin) whereas those subjected to the higher temperature (30°C) contained a higher proportion of rhodopsin (58·7% porphyropsin). Fig. 1 shows an example of goldfish photopigment extracted from the retina of fish held at 30 °C. The difference spectrum (Fig. 1B), which was derived from the absorbance spectra obtained before and after bleaching (Fig. 1A), was similar to that of a mixture of rhodopsin and Iporphyropsin with the absorbance maximum (510 nm) located between 499 nm (pure rhodopsin) and 522nm (pure porphyropsin) (Tsin & Beatty, 1978, 1979).

The HPLC elution of vitamin A extracted from the goldfish contained only two major components: all-ftwis retinol (vitamin A_x alcohol) and all-trans 3,4-didehydroretinol (vitamin A₂ alcohol). The identity of these vitamins A was confirmed by their absorbance spectra and by comparison of their elution characteristics with authentic retinol and 3,4-didehydroretinol. Fig. 2 shows typical HPLC elution profiles of retinols extracted from the blood, liver and RPE of goldfish held at 20 °C. In the normal phase adsorption chromatography employed in our study, retinol eluted from the HPLC within 5·7min whereas 3,4-didehydroretinol had a slightly longer retention time of 6·3 min (Fig. 2). The level of retinol exceeded that of 3,4-didehydroretinol in all blood samples (Fig. 2A; 20·7% 3,4-didehydroretinol), whereas the reverse was observed in liver and RPE samples (Fig. 2B,C).

Table 1 summarizes vitamin A₁ and A₂ compositions in the blood, liver and RPE. Under both light and temperature regimens, the goldfish possessed mainly vitamin Ai alcohol (30% and 33% as vitamin A₂) in the blood and no significant difference (Student’s t-test; P⩽0·05) was noted in the vitamin A composition between the two groups. In contrast, the percentage of vitamin A₂ (52–76%) in liver and RPE was significantly higher (by about 20–50%) than in the blood. Although the percentage of vitamin A₂ in the RPE appeared to be slightly higher than that in the liver in fish held at 20°C, this relationship was not noted in fish held at 30°C.

This study is one of the few on the vitamin A composition of fish blood. The level of total vitamin A in the goldfish blood estimated in our HPLC analysis (46pmol 100 μl^-1 of plasma or 13 μg 100 ml⁻¹) agrees well with published data for carp (8·6μg 100ml⁻¹; Field et al. 1943), trout (17·5 μg 100ml⁻¹; Field et al. 1943) and cod (3–12μg 100ml⁻¹; Plack & Woodhead, 1966). It is lower than the accepted value of blood retinol in humans (40 μg 100 ml⁻¹; Thompson et al. 1971; Neeld & Pearson, 1963; Drujan, 1971).

In experiments involving feeding vitamin A₂ aldehyde to laboratory rats, John et al. (1966) reported that the level of vitamin A₂ in the rat blood increased in proportion to the level of vitamin A₂ esters in the liver. However, there is no evidence that the blood level of vitamin A₂ in the goldfish in the present study was influenced by ingested vitamin A₂ aldehyde, since the diet did not contain any vitamin A₂ (unpublished observations). Plack & Woodhead (1966) showed that cod had approximately 20-40% vitamin A₂ aldehyde in their blood and they estimated that the A₂ form of the vitamin constituted about 15% of total vitamin A alcohols and vitamin A esters. However, in a comparative study on vitamin A in the blood of non-mammalian vertebrates, Shidoji & Muto (1977) showed that only cyclostomes, such as the lampreys, contain vitamin A esters in their blood, and there was no evidence that other species (from teleosts such as carp and yellowtail to mammals) contain any forms of vitamin A other than retinol and 3,4-didehydroretinol. These studies therefore pose unresolved questions on vitamin A transport in fishes. In this study, we have shown that vitamin A₂ constituted approximately 30% of the vitamin A alcohols in goldfish blood, and this proportion does not change with temperature. In goldfish blood, we have not investigated any vitamin A components other than retinol and 3,4-didehydroretinol.

In mammals, vitamin A mobilization from the liver and its delivery to the target tissue via the blood are highly regulated processes (Goodman & Blaner, 1984). Vitamin A esters in the liver are first hydrolysed by hepatic retinyl ester hydrolase (Tsin et al. 1986a), and vitamin A alcohol is bound to serum retinol-binding protein (Goodman, 1984) and then transported in the circulation to the basal surface of the RPE where the vitamin A is released (Bok, 1985). Upon entry into the RPE, the vitamin A is esterified (Berman et al. 1980; Tsin, 1986a,b) and stored mainly as vitamin A esters (Lai et al. 1985), which later undergo hydrolysis (Blaner et al. 1987; Tsin & Lam, 1986) to supply vitamin A for visual pigment synthesis.

In lower vertebrates, where vitamin Ax and A₂ co-exist, events associated with the supply of vitamin Aj and A₂ to form visual pigments are far less well understood. The present study shows that vitamin A₁ and A₂ are both found in the fiver and the blood, yet the proportion of vitamin A₁ to A₂ differs significantly between these compartments (see Table 1). It is unlikely that the carrier protein in the blood selectively favours the transport of retinol rather than 3,4-didehydroretinol because both vitamins have a similar affinity for the serum retinol-binding protein (Shidoji & Muto, 1971). One possibility is that a selective hydrolysis of vitamin A esters may take place in the liver, resulting in the release of a higher proportion of vitamin Ax. Alternatively, if the eye were an important site of vitamin A₂ synthesis (Naito & Wilt, 1962; Tsin et al. 1985) or utilization (Shantz et al. 1946), then we would expect to find a significantly higher proportion of vitamin A₂ in the RPE (Table 1). It is clear, however, that the percentage of vitamin A₂ in different body compartments (liver, blood, RPE) does not reflect the photopigment composition in the retina (Table 1). This suggests that the formation of visual pigment in the retina is not dependent on a simple equimolar transfer of available circulating vitamin A to serve as chromophores for opsin. This suggestion is further substantiated by the observations (Yoshikami et al. 1969; Pearlman & Crescitelli, 1971; John et al. 1966) that feeding rats with vitamin A₂ and increasing the vitamin A₂ level in the blood did not lead to the formation of porphyropsin in the retina. However, Suzuki & Miyata (1988), using mice injected with vitamin A₂, found high proportions of vitamin A₂ in the liver, blood and neuroretina.

In conclusion, the present study clearly shows that environmental light and temperature exerted a significant influence on the proportions of rhodopsin and porphyropsin in the goldfish neuroretina. Moreover, temperature did not change the visual pigment ratios via an alteration of the relative availability of vitamin A₁ and A₂ in the blood. Other body compartments such as liver and retinal pigment epithelium (RPE) also stored vitamin A₁ and A₂ in proportions that seem unrelated to the scotopic visual pigment ratios in the eye. The mechanism of visual pigment changes therefore may lie within the tissue where visual pigments reside, the retina itself. Additional work is currently under way to explore further this process of photopigment conversion in the vertebrate neuroretina (Suzuki et al. 1985; Tsin & Chambers, 1988).

This research was supported by grants from the NIH (EY06438), NSF (8203064), UTSA (14-15330) and the San Antonio Area Foundation (the Norma Friedrich Ward Trust). We thank Hoffman La-Roche for the gift sample of vitamin A₂ and Mr Douglas Malsbury and Karl Rodriguez for technical assistance.