ABSTRACT
The eyes of some crustaceans store substantial amounts of retinyl esters, with most of the retinol in the 11-cis configuration. Earlier work in this laboratory suggested that in lobster and crayfish the mechanism of isomerization of retinol to the 11-cis form involves the hydrolysis of all-trans retinyl esters. Although this is the same process as that occurring in the vertebrate eye, it is different from the retinal photoisomerase reaction known in other arthropods, specifically diurnal insects (Hymenoptera and probably Diptera). Using homogenates of crayfish, we have tested this proposed mechanism by inhibiting retinyl ester synthetase activity in the presence of exogenous all-trans retinol. Inhibition of lecithin:retinol acyl transferase with 5 μmol l-1 retinyl bromoacetate or 2 mmol l-1 phenylmethylsulfonyl fluoride blocks the formation of both all-trans and 11-cis retinyl esters as well as 11-cis retinol, as shown by direct assay and by the decrease in counts derived from tritiated all-trans retinol. The similarity of this isomerization to the mechanism in vertebrate pigment epithelium is thus an interesting example of convergent evolution in the biochemistry of visual pigments, in which the pigments themselves (the opsins) are largely conserved across phyla.
Introduction
The means by which animals isomerize retinol or retinal to the 11-cis configuration for incorporation of the latter into their visual pigments is one of the most variable features of the visual pigment cycle. The first of these enzymes to be discovered was the retinal photoisomerase of cephalopod molluscs, an intrinsic membrane protein known as retinochrome (Hara and Hara, 1972; Ozaki et al. 1986; Hara, 1988). Certain diurnal insects also have a retinal isomerase, but in bees (Apis mellifera), where it is best known, it is a soluble protein (Schwemer et al. 1984; Smith and Goldsmith, 1991a,b). A retinal photoisomerase has also been reported in Limulus polyphemus (Smith et al. 1992), but nothing is known of its biochemistry.
A very different system is found in the pigment epithelium of vertebrate retinas, where isomerization takes place at the oxidative level of retinol (rather than retinal) by coupling the isomerization to the hydrolysis of all-trans retinyl palmitate (Bernstein et al. 1987; Law and Rando, 1988; Barry et al. 1989; Deigner et al. 1989; CaÑada et al. 1990; Livrea et al. 1990).
The diurnal insects that have a retinal photoisomerase do not have reserves of retinyl esters. Vertebrate pigment epithelium, however, characteristically stores retinol as retinyl esters. Similarly, the eyes of mantis shrimps contain several hundred pmoles of 11-cis retinyl esters (Goldsmith and Cronin, 1993), and in macruran decapods (Homarus, Procambarus), there are also substantial stores of 11-cis retinol, principally as esters of docosahexaenoate (C22:6) with somewhat lesser amounts of retinyl oleate (C18:1) (Suzuki et al. 1988; Srivastava et al. 1996). Moreover, homogenates of the eyes of both Homarus and Procambarus are capable of forming 11-cis retinyl esters from added all-trans retinyl docosahexaenoate, a reaction that takes place in the dark (Srivastava et al. 1996). These observations suggest that some crustaceans with nocturnal or crepuscular habits not only store 11-cis retinyl esters but may form 11-cis retinol by a mechanism very similar to that employed by the vertebrate eye.
Because in vertebrate pigment epithelium the energy for isomerization is obtained by coupling to the hydrolysis of retinyl esters, if all-trans retinol is provided as substrate it must first be esterified before it can be isomerized to 11-cis. A critical test of this mechanism is to inhibit the formation of retinyl esters (Trehan et al. 1990).
In vertebrate pigment epithelium, coenzyme A (CoA)-dependent synthesis of retinyl esters can be demonstrated in vitro (Saari and Bredberg, 1988), but retinyl esters are also formed by the transfer of the fatty acid from a phospholipid by the enzyme lecithin:retinol acyl transferase (LRAT) (Barry et al. 1989; Saari and Bredberg, 1989). In the present experiments, we find that inhibitors of LRAT activity sharply decrease the formation of 11-cis retinol and 11-cis retinyl esters from exogenous all-trans retinol.
Materials and methods
Preparation of homogenates
Crayfish (Procambarus clarkii Girard) (Carolina Biological Supply Company, Burlington, NC, USA) were kept in a shallow-water aquarium at room temperature (range 20–22 °C) and fed dry dog food for periods of up to 2 weeks prior to use. The animals were immobilized by placing them on crushed ice for 10–20 min, the eyestalks were removed by cutting at their base, the eyes were bisected, and the soft tissue was removed. The pigmented tissue was severed from the optic nerve and homogenized manually (Wheaton tissue grinder with hollow glass pestle) in 0.1 mol l-1 phosphate buffer (pH 7.4) containing 2 mmol l-1 dithiothreitol, using a volume of 0.5 ml per eye. This crude homogenate was used for the assays.
Assay for isomerase activity
For substrate, [11,12-3H(N)]-all-trans retinol (Dupont NEN Research Products, Boston, MA, USA) was diluted to a specific activity of 8.88×106 Bq μmol-1 using unlabeled all-trans retinol (Sigma, St Louis, MO, USA). For each reaction mixture, 200–300 pmol of this retinol in hexane was added to 1 ml of homogenate, using 100 μl of bovine serum albumin (100 mg ml-1) as carrier. The reaction mixture was incubated for 0–120 min in the dark at 22 °C, and the reaction was stopped by the addition of an equal volume of chilled ethanol. In control tubes, the homogenate was added after the addition of ethanol and the tubes were kept for 2 h at either 0 or 22 °C. There was no difference in the levels of either retinol or retinyl esters between these controls. In previous experiments involving retinyl esters as substrates and (in the case of lobster) higher levels of endogenous esters, there was no consistent evidence for loss of retinoid during 2 h of incubation (Srivastava et al. 1996).
Retinol and its esters were extracted into hexane and separated by high-performance liquid chromatography (HPLC). 1 ml samples of the eluate were collected and radioactivity was measured in a liquid scintillation counter (Beckman Instruments LS 8000, Columbia, MD, USA). The counts that entered 11-cis retinyl esters were more than two orders of magnitude lower than the tritium label found in 11-cis retinol; reported changes of 20–40 cts min-1 are relative to base values of 100–200 cts min-1 in individual ester fractions.
High-performance liquid chromatography
Isomers of retinol and retinyl esters were separated using a normal-phase silica column (Rainin Microsorb, Woburn, MA, USA) and isocratic elution at a flow rate of 1 ml min-1. The mobile phase was 9 % dioxane in hexane for free retinol. Under these conditions, the esters elute close to the front; they were recovered as a single peak and separated using 0.35 % dioxane in hexane as the mobile phase. The absorbance of the eluate was measured at 328 nm and the system was quantified using peak integration software (Perkin Elmer Nelson Systems, Cupertino, CA, USA). The fluorescence of the eluate was also monitored. See Srivastava et al. (1996) for further details and for the identification of the fatty acid moieties of the retinyl esters.
Use of inhibitors
Isomerization of retinol was examined in the presence of known inhibitors of retinol esterification: all-trans α-retinyl bromoacetate (RBA), phenylmethylsulfonyl fluoride (PMSF) and progesterone. Retinol, PMSF and progesterone were obtained from Sigma (St Louis, MO, USA); RBA was synthesized by reacting retinol with bromoacetylbromide (Aldrich, Milwaukee, WI, USA) in the presence of pyridine (Gawinowicz and Goodman, 1982).
In these studies, 1 ml of eye homogenate was preincubated with several concentrations of inhibitor. RBA was preincubated with the homogenate for 10 min at 30 °C, while PMSF and progesterone were preincubated for 10 min and 6 min respectively at 25 °C, before the addition of 3H-labeled all-trans retinol. After 2 h of incubation, retinol and retinyl esters were extracted and the geometric isomers were assayed as described above.
Results
Formation of 11-cis retinol
When homogenized crayfish eyes were incubated at 22 °C with all-trans retinol (150–200 pmol eye-1), there was a several-fold increase in the concentration of the 11-cis isomer over 2 h (Fig. 1). By comparison, there was much smaller change in the level of 13-cis, the isomer most likely to form non-enzymatically from the all-trans substrate. Because there were traces of the 13-cis isomer in the unlabeled all-trans retinol with which the tritiated compound was diluted to form the stock solution, 1–2 pmol of 13-cis retinol had probably been added to the reaction mixture with the substrate. There was little increase in 13-cis retinol level during the 2 h incubation, which is why the 13-cis retinol that is recovered at 2 h has such a low specific activity relative to the 11-cis isomer.
The results of 24 experiments are summarized in Fig. 2, where the open bars indicate the amount of cis isomer present immediately after adding substrate and the filled bars the amount 2 h later. The 2–3 pmol eye-1 of 11-cis retinol initially present was endogenous retinol (Fig. 3A). Extracts of eyes contain less 13-cis retinol (<1 pmol eye-1), however, and the 3–4 pmol of 13-cis isomer was added with the much larger amount of all-trans substrate. The increases in levels of both 11- and 13-cis retinol in Fig. 2A are highly significant (P<0.001, paired t-tests).
The corresponding increase in tritium activity associated with the two cis isomers is shown in Fig. 2B. That the all-trans substrate is the source of the 11-cis retinol is confirmed by a highly significant increase in counts in the latter (P<0.001). There was a much smaller increase in activity of the 13-cis isomer.
Formation of 11-cis retinyl esters
Crayfish eyes contain retinyl esters of (principally) two fatty acids (Fig. 3B), docosahexaenoate (C22:6) and oleate (C18:1) (Srivastava et al. 1996). The amounts are several times larger than the amounts of free retinol present (Fig. 4, open bars), and in both esters the retinyl moiety is largely in the 11-cis configuration, with smaller amounts of the all-trans form. Following incubation with all-trans retinol, there are small increases in levels of the 11-cis retinyl esters (Fig. 4) and significant increases in activity in all four isomers (Fig. 5).
Inhibition of isomerization by retinyl bromoacetate (RBA)
Fig. 6A shows the increase in 11-cis retinol formed from all-trans retinol substrate after 2 h as a function of the concentration of RBA present. Less 11-cis retinol appeared as the concentration of RBA increased. Fig. 6B shows that 5 μmol l-1 RBA significantly decreased (P<0.01, paired t-test for six experiments) the amount of tritium label that appeared in 11-cis retinol.
Fig. 7 shows the counts in 11- and 13-cis retinol after 2 h of incubation at three concentrations of RBA. The two horizontal lines (at <1000 cts min-1) show the average activity associated with each isomer at the start of incubation and immediately following the addition of substrate. As expected from the data of Fig. 6, increasing the RBA concentration diminished the appearance of counts in 11-cis retinol but had no effect on the much smaller activity that accumulated in 13-cis retinol during a 2 h incubation.
RBA also inhibited the formation of 11-cis retinyl esters. Fig. 8A shows the decrease in post-incubation levels of C22:6 and C18:1 esters with increasing RBA concentration. The pre-incubation levels are shown for comparison by the horizontal lines. In the presence of 2 and 5 μmol l-1 RBA, the 11-cis C22:6 ester level fell below the 14 pmol eye-1 originally present, signifying a net hydrolysis of ester.
Net accumulations of counts in cis and trans retinyl esters are shown in Fig. 8B. RBA (5 μmol l-1) caused highly significant decreases in both cases. As outlined above, the all-trans retinyl esters are the presumptive precursors for 11-cis retinol.
Effects on isomerization of other agents
Phenylmethylsulfonyl fluoride (PMSF) is an inhibitor of LRAT but not of CoA-dependent esterification of retinol (Ong et al. 1988; Randolph et al. 1991). In homogenates of crayfish eyes, PMSF inhibited the formation of 11-cis retinol from all-trans retinol substrate (Fig. 9A). The formation of 11-cis retinyl esters also appeared to be inhibited (Fig. 9B).
Progesterone is an inhibitor of acetyl CoA:retinol acyltransferase (Ross, 1982; Yost et al. 1988). In the present study, progesterone had no significant effect on the isomerization of exogenous retinol (data not shown).
Discussion
In the hypothesized mechanism of isomerization of retinoids, all-trans retinyl esters should form as intermediates in the isomerization process. The relatively small amounts of all-trans retinyl ester do not increase significantly in the presence of endogenous all-trans retinol, whereas the larger amounts of 11-cis retinyl esters do increase (Fig. 4). All-trans retinyl esters do form, however, as shown by their increasing levels of radioactivity (Fig. 5). The simplest interpretation is that, during the 2 h period of incubation, retinol is passing through small and relatively constant pools of all-trans retinyl ester, consistent with an isomerization mechanism for retinol that employs all-trans retinyl esters as an intermediate.
Inhibition of LRAT blocked the accumulation of tritium in all-trans retinyl esters (Fig. 8B) as well as the formation of both 11-cis retinol and 11-cis retinyl esters. This is consistent both with the proposed mechanism of isomerization coupled to the hydrolysis of all-trans retinyl esters and with the corollary conclusion that LRAT provides the major path for the formation of all-trans retinyl esters in crayfish eyes. LRAT may also form 11-cis esters, but the present data cannot confirm this directly because blocking the formation of 11-cis retinol should prevent the formation of 11-cis retinyl esters by any path. Less extensive experiments with progesterone, an inhibitor of CoA-dependent esterification, provided no evidence for activity of this alternative pathway.
Although crayfish eyes contain much more retinyl ester than free retinol, roughly equivalent amounts of 11-cis retinol and 11-cis retinyl ester were formed from exogenous all-trans retinol (compare Figs 2A and 4). This result is consistent with the hypothesized pathway, in which all-trans retinol must be isomerized to the 11-cis configuration before it is esterified for storage. There is a complication, however. In the presence of RBA (2–5 μmol l-1), there was a net loss of C22:6 retinyl ester (Fig. 8A). This indicates that in these homogenates the level of ester was determined both by its rate of formation from exogenous retinol and by a hydrolysis that is probably prevented in vivo.
When retinol is added to homogenates of the eye as all-trans retinyl docosahexaenoate, 11-cis retinyl esters of both C22:6 and C18:1 are formed (Srivastava et al. 1996). The isomerization presumably occurs in association with hydrolysis, and the 11-cis esters should form secondarily. In the present work, the substrate was all-trans retinol, which was neither esterified nor isomerized when the transesterification enzyme (LRAT) was inhibited.
Approximately 15 pmol per eye of 11-cis retinoid (alcohol plus esters) are made in 2 h under the conditions of these experiments. This quantity is approximately 15 % of the retinal present in the eye. Although the reaction may be more efficient in vivo, there is little reason to believe that much faster rates are necessary. First, there are endogenous stores of 11-cis retinyl esters amounting to approximately 30 pmol per eye in these experiments, but the levels increase when the animals are kept cold (Suzuki et al. 1988) or in darkness (R. Srivastava and T. H. Goldsmith, unpublished data). Second, in arthropod metarhodopsin, all-trans retinal is not exchanged for the 11-cis form to regenerate visual pigment. Either metarhodopsin is photoisomerized to rhodopsin or the entire rhodopsin molecule is replaced in a process that involves de novo synthesis of opsin (reviewed by Schwemer, 1986). Hafner and Bok (1977) measured the appearance of tritiated leucine in crayfish rhabdoms and showed that opsin replacement takes place over a time span of hours or days. Similarly, Cronin and Goldsmith (1984) found that, following intense adapting exposures, recovery of crayfish visual pigment required several days in darkness.
In the insects that have been studied (see references in the Introduction), retinal is isomerized by a mechanism that uses the energy of absorbed light. What adaptive processes might be responsible for two such different isomerization processes in arthropods? The soluble photoisomerase of insects occurs in strongly diurnal species (Apis mellifera, Calliphora erythrocephala), and in Apis retinoid reserves consist of retinal bound to the photoisomerase. Retinol is also present, but it remains bound to protein and does not become esterified in significant amounts. This is an isomerization system that works well where there is ample light, but it may serve poorly where the ambient light level is not dependable. Crayfish and lobster have superposition optics, which is an anatomical adaptation for low light conditions and where a photoisomerase may be an inefficient system for regenerating chromophore. Their alternative of coupling isomerization to hydrolysis of esters ultimately derives its energy from metabolism and, as it operates in darkness, may well be more suitable for nocturnal or crepuscular conditions or for life in murky waters. A useful feature of retinyl esters is that they are readily stored in lipid vesicles.
There is, however, another difference to consider in making this comparison: the arthropods known to have the retinal photoisomerase are insects, whereas the species depending on ester hydrolysis are crustaceans. There are, of course, crustaceans with apposition eyes and more diurnal habits, and there are nocturnal insects with superposition eyes. A wider comparative search should reveal whether, within the arthropods, these two systems for retinoid isomerization reflect phylogenetic inertia or physiological adaptations to ecological conditions.
Cronin and Goldsmith (1984) found that, in crayfish, visual pigment recovery time was shortened if the last adapting light flash was blue rather than orange. Whether this signifies the presence of a supplementary photoisomerase, however, is not clear. We have been unable to find evidence for a photoisomerase in vitro. In flies, metarhodopsin is removed from the membrane faster than rhodopsin and, as a consequence, opsin is removed faster in the presence of green light (which does not activate the photoisomerase) than in darkness (Schwemer, 1986). Although this observation cannot account for the effect of blue light on the rate at which crayfish rhabdoms are restored, it indicates that the relative amounts of rhodopsin and metarhodopsin in the photoreceptor membranes can affect the rate of recovery. This is clearly a matter that requires further exploration.
ACKNOWLEDGEMENTS
This work was supported by grants EY00222 and EY00785 from the National Eye Institute.