Synthesis of the all-trans-retinal Chromophore of Retinal G Proteincoupled Receptor Opsin in Cultured Pigment Epithelial Cells

Transcription

1 JBC Papers in Press. Published on November 26, 2001 as Manuscript M Synthesis of the all-trans-retinal Chromophore of Retinal G Proteincoupled Receptor Opsin in Cultured Pigment Epithelial Cells Mao Yang and Henry K. W. Fong * Departments of Microbiology and Ophthalmology, Keck School of Medicine, University of Southern California, and Doheny Eye Institute, Los Angeles, California Short Title: RGR Opsin and all-trans-retinol Dehydrogenase in RPE * Address correspondence to this author at the Doheny Eye Institute, 1355 San Pablo Street, Los Angeles, CA Tel: ; Fax: ; hfong@hsc.usc.edu Abbreviations: RPE, retinal pigment epithelium; RGR, RPE retinal G protein-coupled receptor; LRAT, lecithin:retinol acyltransferase Copyright 2001 by The American Society for Biochemistry and Molecular Biology, Inc.

2 Yang and Fong Page 2 SUMMARY Light-dependent production of 11-cis-retinal by the retinal pigment epithelium (RPE) and normal regeneration of rhodopsin under photic conditions involves the RPE retinal G protein-coupled receptor (RGR) opsin. This microsomal opsin is bound to all-trans-retinal which, upon illumination, isomerizes stereospecifically to the 11-cis isomer. In this paper, we investigate the synthesis of the all-trans-retinal chromophore of RGR in cultured ARPEhRGR and freshly isolated bovine RPE cells. Exogenous [ 3 H]all-trans-retinol is incorporated into intact RPE cells and converted mainly into retinyl esters and all-trans-retinal. The intracellular processing of [ 3 H]all-trans-retinol results in physiological binding to RGR of a radiolabeled retinoid, identified as [ 3 H]alltrans-retinal. The ARPE-hRGR cells contain a membrane-bound NADPHdependent retinol dehydrogenase that reacts efficiently with all-trans-retinol, but not the 11-cis isomer. The NADPH-dependent all-trans-retinol dehydrogenase activity in isolated RPE microsomal membranes can be linked in vitro to specific binding of the chromophore to RGR. These findings provide confirmation that RGR opsin binds the chromophore, all-trans-retinal, in the dark. A novel all-trans-retinol dehydrogenase exists in the RPE and performs a critical function in chromophore biosynthesis.

3 Yang and Fong Page 3 INTRODUCTION The retinal pigment epithelial (RPE) cells are highly active in the metabolism of retinoids and are essential for the synthesis of the 11-cis-retinal chromophore of visual pigments (1, 2). Many specialized enzymes and retinoid-binding proteins are involved in the production of 11-cis-retinal from all-trans-retinol. Lecithin:retinol acyltransferase (LRAT) is among the most active enzymes in retinoid processing and acts early in the retinoid cycle by catalyzing the esterification of all-trans-retinol soon after uptake into the RPE cells (3-5). The retinyl esters are normally the predominant retinoids, even while the content and distribution of retinoids in the RPE may vary under different light and dark conditions (6-10). Other enzymes that affect the content and distribution of retinoids include retinyl ester hydrolase (11, 12), 11-cisretinol dehydrogenases (13-17), and a putative isomerohydrolase or isomerase (18-21). The retinoid-binding proteins in the RPE include cellular retinolbinding protein (CRBP) (22, 23), cellular retinaldehyde-binding protein (CRALBP) (24, 25) and a unique opsin, RPE retinal G protein-coupled receptor (RGR) (26, 27). RGR is a membrane-bound opsin with seven transmembrane domains and is expressed in Müller cells, as well as in the RPE (28, 29). It is closely related in amino acid sequence to invertebrate visual pigments and retinochrome, a photoisomerase that catalyzes the conversion of all-trans- to 11-cis-retinal in squid photoreceptors (30). The RGR opsin is bound in the dark to all-trans-retinal and has absorption maxima at ~469 and ~370 nm (31, 32). Illumination in vitro results in the stereospecific conversion of the bound alltrans-retinal to the 11-cis isomer. RGR is involved in the formation of 11-cis-

4 Yang and Fong Page 4 retinal in mice and is necessary for maintaining normal steady-state levels of both 11-cis-retinal and rhodopsin in a light-adapted eye (33). These results indicate that RGR functions to generate 11-cis-retinal in vivo and participates in a light-dependent visual cycle. Mutations in human RGR, which is located on chromosome 10q23 (34), are associated with cases of recessive and dominant retinitis pigmentosa (35). Although RGR is a major all-trans-retinal-binding protein, it is unclear how the all-trans-retinal chromophore is generated in RPE and Müller cells. Only low amounts of all-trans-retinal, if any, have been reported in the RPE (6-10), yet the RPE must be able to synthesize the chromophore of RGR. Indeed, the role of RGR in a photic visual cycle may require synthesis of the all-transretinal chromophore directly from all-trans-retinol. These considerations suggest that a novel all-trans-retinol dehydrogenase exists in the RPE to provide the chromophore of RGR. Recently, we demonstrated the uptake of exogenous all-trans-retinol into lentivirus-transduced ARPE-hRGR cells and subsequent incorporation of the retinoid into a bound ligand of RGR (36). In this paper, we confirm that all-trans-retinal is synthesized in RPE cells and becomes bound to RGR physiologically. We also present evidence for a novel all-trans-retinol dehydrogenase in both cultured ARPE-hRGR and isolated bovine RPE cells.

5 Yang and Fong Page 5 EXPERIMENTAL PROCEDURES Materials - all-trans-[11,12-3 H(N)]Retinol (50 Ci/mmol) was obtained from NEN Life Science Products. all-trans-retinol, all-trans-retinal and alltrans-retinyl palmitate standards were purchased from Sigma. 11-cis-Retinal was provided by Dr. Rosalie Crouch (Medical University of South Carolina, Charleston, SC). 11-cis-Retinol was prepared by the reduction of 11-cis-retinal in the presence of NaBH 4, as described previously (37). NADP and NAD were from Sigma. Organic solvents were HPLC grade. Dichloromethane and hexane were obtained from Fisher. Diethyl ether and methanol were from J. T. Baker Inc. Ethanol was from Gold Shield Chemical Company. Cell culture - ARPE-19 cells, a human RPE cell line, maintain many characteristics of normal RPE cells, but do not express detectable levels of RGR opsin (38). ARPE-hRGR cells, which stably express human RGR, were obtained by transduction of ARPE-19 cells with a recombinant lentivirus-human RGR vector (36). The ARPE-19 and ARPE-hRGR cell lines were cultured in DME/F12 (1:1) medium (Life Technologies) supplemented with 10% fetal bovine serum (FBS) (Hyclone) and 1% Glutamine-Penicillin-Streptomycin (Irvine Scientific) at 37 o C in a 5% CO 2 incubator. Cells at passage no were grown and maintained at confluency for 1 to 2 weeks before use in each experiment. Incubation of cultured cells with [ 3 H]all-trans-retinol and analysis of radiolabeled proteins - ARPE-hRGR and ARPE-19 cells were preincubated overnight with serum-free (retinol-free) RPMI1640 medium (Life Technologies) at 37 o C in 5% CO 2. The cells were then washed with RPMI1640

6 Yang and Fong Page 6 medium and incubated in the dark with a mixture of [ 3 H]all-trans-retinol (0.25 µci/ml, 50 Ci/mmol), 500 µg/ml fatty acid-free bovine serum albumin (BSA), and 0.5% sucrose in RPMI1640 medium. After incubation for various lengths of time at 37 o C in 5% CO 2, the cells were washed with phosphate-buffered saline (PBS), collected by scraping in 2 ml 67 mm sodium phosphate buffer, ph 6.7, and homogenized with a Dounce glass homogenizer. After the ph was adjusted to 8.0 with 1 N NaOH, 38 mg/ml NaBH 4 was added to the suspension. The membranes were centrifuged at 150,000 g for 1 h at 4 o C. [ 3 H]-Labeled proteins were analyzed by gel electrophoresis and fluorography. After separation by 12% SDS-PAGE, the proteins were fixed, and the gel was soaked in ENLIGHTNING Rapid Autoradiography Enhancer (NEN Life Science Products). The gel was dried and exposed to Kodak X-omat AR 5 autoradiographic film (Eastman Kodak Co.). The autoradiographic films were scanned, and relative band densities were determined using Scion Image 1.62c software (Scion Corp., Frederick, MD). Isolation of bovine RPE cells and incubation with [ 3 H]all-trans-retinol - Bovine eyes were obtained from a local abattoir. The RPE cells were isolated under ambient illumination 2-3 h after enucleation. The anterior segments, lens, vitreous and neural retina were removed, and the eyecups were washed twice with ice-cold PBS to remove retina debris. The RPE cells were scraped off gently with a metal spatula and collected in ice-cold PBS. The cells were then centrifuged at 800 g for 5 min at 4 o C. The pellet was washed once with icecold PBS and centrifuged again. The following procedures were performed in the dark. The RPE cells were resuspended in 10 ml DME/F12 (1:1) medium supplemented with [ 3 H]all-trans-retinol (1.0 µci/ml, 50 Ci/mmol), 500 µg/ml fatty acid-free BSA, and 0.5% sucrose and incubated in a culture flask for 3 h at

7 Yang and Fong Page 7 37 o C in 5% CO 2. After incubation, the cells were collected with a cell scraper, centrifuged and washed once with PBS. The cells were resuspended in 600 µl of PBS and homogenized with a Dounce glass homogenizer. The sample was treated with NaBH 4, as described above. The membranes were then centrifuged at 150,000 g for 1 h at 4 o C and analyzed by gel electrophoresis and fluorography to detect [ 3 H]-labeled proteins. Alternatively, part of the cell homogenate was saved for Western blot assay, or the retinoids were extracted in the presence of hydroxylamine, as described below. Extraction of retinoids from RPE cells or membranes - All procedures were performed under dim red light. Retinal isomers were extracted by the method of hydroxylamine derivatization, as described previously (39, 40). Cultured ARPE-hRGR, ARPE-19 or bovine RPE cells were washed with PBS and homogenized in 300 µl of PBS using a Potter-Elvehjem micro tissue grinder. The whole homogenate or membranes were mixed with 300 µl of methanol and then 30 µl of 2 M NH 2 OH. After the mixture was incubated at room temperature for 5 min, 300 µl of CH 2 Cl 2 was added and mixed by vortexing for 30 s. The organic and aqueous phases were separated by centrifuge at 14,000 g for 1 min. The aqueous phase was extracted twice more with CH 2 Cl 2. The pooled CH 2 Cl 2 solution was dried under nitrogen gas flow, dissolved in 600 µl of hexane, filtered through glass wool, dried again, and then stored at -80 o C for later analysis. HPLC Analysis of Retinoids - The isomers of retinaloximes were analyzed by HPLC, as described previously (41, 42). The extracted retinaloximes were dissolved in hexane and separated on a Resolve Silica column (3.9 x 150 mm, 5 µm) (Waters Corp.) using a Waters 2690 HPLC module. The running

8 Yang and Fong Page 8 buffer consisted of hexane supplemented with 8% diethyl ether and 0.33% ethanol and was pumped at a flow rate of 0.3 ml/min. Absorbance was measured at 360 nm and 320 nm with a Waters 2487 Dual Wavelength Absorbance Detector. The absorbance peaks were analyzed with the Millennium 32 Chromatography Manager software, version 3.20 (Waters Corp.). The HPLC column was calibrated before each run using all-trans- and 11-cis-retinaloxime and all-trans-retinol standards. Identification of the retinaloxime isomers was based on the retention times of the known retinaloxime products. The proportion of each isomer in the loading sample was determined from the total peak areas of both its syn- and anti-retinaloxime and was based on the following extinction coefficients (ε 360, in hexane): all-trans syn = 54,900, all-trans anti = 51,600, 11-cis syn = 35,000, 11-cis anti = 29,600, 13-cis syn = 49,000 and 13-cis anti = 52,100 (40, 41). The HPLC system was calibrated with pmol of syn-all-trans-retinaloxime standard. The alltrans-retinyl palmitate standard was eluted in running buffer composed of hexane supplemented with 8% diethyl ether and 0.33% ethanol, or hexane supplemented with 0.3% ethyl acetate. The [ 3 H]-labeled retinoids were separated as described above. Four drops of the HPLC eluate were collected manually per fraction and were mixed with 10 ml of ScintiVerse BD scintillation cocktail (Fisher Scientific). The amount of radioactivity was measured with a Beckman LS 6000IC counter. Identification of [ 3 H]-labeled retinoids was based on the retention times of known standards. Preparation of cell membranes - ARPE-hRGR and ARPE-19 cells were preincubated overnight in serum-free RPMI1640 medium at 37 o C in 5% CO 2. Total cell membranes were prepared from bovine RPE, cultured ARPE-hRGR

9 Yang and Fong Page 9 and ARPE-19 cells. The cells were homogenized in buffer containing 67 mm sodium phosphate, ph 6.6, and 250 mm sucrose. The homogenate was centrifuged at 300 g. Thereafter, the supernatant was centrifuged at 150,000 g for 1.5 h at 4 C. The membrane pellets were saved and stored at -80 C. Microsomal membranes were prepared as described previously (26, 36). Retinol dehydrogenase assay - The following procedures were performed in darkness or under dim red light. Microsomal membranes were washed and resuspended in buffer containing 50 mm Tris-HCl, ph 7.5, and 0.1% BSA. The reactions were initiated by addition of the membranes to 50 mm Tris-HCl, ph 7.5, 0.1% BSA, NADP or NAD, and 1 µm [ 3 H]all-trans-retinol (reaction vol. = 300 µl). The specific activity of [ 3 H]all-trans-retinol was made 5 Ci/mmol by dilution with unlabeled all-trans-retinol. Solutions of 10 mm NADP or NAD were made fresh and added to give the indicated final concentrations. After incubation at 37 o C for the specified amount of time, the reactions were terminated by addition of 300 µl of methanol and then 30 µl of 2 M NH 2 OH. The retinaloximes were extracted and analyzed by HPLC, as described previously. The HPLC fractions were collected, and the amount of radioactivity was determined. When nonradioactive substrates were used, retinol dehydrogenase activity was assayed with 5 µm exogenous all-trans-retinol or 11-cis-retinol in the presence of 200 µm NADP. Retinals were extracted by hydroxylamine derivatization and analyzed by HPLC. The extraction efficiency was monitored by the addition of [ 3 H]all-trans-retinol to the methanoldenatured samples. Under the experimental conditions, the reaction rate for production of all-trans-retinal was linear within the initial 10 min.

10 Yang and Fong Page 10 Cell-free synthesis and binding of all-trans-retinal to RGR in vitro - Microsomal membranes from bovine RPE, ARPE-hRGR and ARPE-19 cells were washed and resuspended in buffer containing 50 mm Tris-HCl, ph 7.5, and 0.1% BSA. The reactions were initiated in the dark by addition of the membranes to 50 mm Tris-HCl, ph 7.5, 0.1% BSA, 200 µm NADP or none, and 0.2 µm [ 3 H]all-trans-retinol (10 µci/ml, 50 Ci/mmol). After incubation at 37 0 C for 30 min, the membranes were sedimented by centrifugation at 150,000 g for 1 h at 4 o C. The pellet was washed, resuspended in 1 ml of PBS and mixed with 38 mg NaBH 4. The membranes were then centrifuged, resuspended in PBS containing 0.1% SDS and analyzed by gel electrophoresis and fluorography, as described previously. Western blot assay - The membranes were resuspended in PBS, and protein concentration was measured by the Bio-Rad protein assay (Bio-Rad Laboratories). The samples were separated by 12% SDS-PAGE and then electrotransferred to an Immobilon-P membrane (Millipore). An affinity-purified antipeptide antibody (28, 29), that is directed against the carboxyl terminus of bovine RGR, and the ECL detection reagents (Amersham) were used to detect RGR. Prestained protein standards (Life Technologies) were used for molecular weight markers. Irradiation of ARPE-hRGR cells - ARPE-hRGR cells were cultured in DME/F12 (1:1) medium containing 10% FBS. The cells were maintained overnight in culture flasks wrapped in aluminum foil and were then processed under dim red light. The cells were washed with PBS, removed with a cell scraper, centrifuged and resuspended in 300 µl of PBS. The suspended cells were transferred to a quartz cuvette and illuminated at room temperature for 5

11 Yang and Fong Page 11 min with 470-nm monochromatic light from an Oriel light source, Model (Oriel Corp.), equipped with a 150 watt Xenon arc lamp. An equivalent aliquot of ARPE-hRGR cells was kept in the dark as a control. Retinal isomers were extracted in the presence of hydroxylamine and and analyzed by HPLC, as described previously (39-42).

12 Yang and Fong Page 12 RESULTS Uptake and binding of [ 3 H]retinoid to RGR in ARPE-hRGR cells We have demonstrated previously that precursor all-trans-retinol can be incorporated into a bound ligand of RGR in cultured ARPE-hRGR cells (36). ARPE-hRGR cells with high long-term expression of human RGR were incubated with [ 3 H]all-trans-retinol in serum-free RPMI1640 medium. The results indicated that a specific ~30-kDa protein bound [ 3 H]retinoid (Fig. 1A). The ~30-kDa protein band was not detected in control ARPE-19 cells that lack RGR (36). The binding of [ 3 H]retinoid to RGR in vivo was dependent on the time of incubation with [ 3 H]all-trans-retinol. Binding occurred within 20 min and reached the highest level by 2 h of incubation with [ 3 H]all-trans-retinol (Fig. 1B). Subsequently, the amount of [ 3 H]retinoid bound to RGR remained relatively steady for up to 6 h of incubation in the dark. Under the experimental conditions, no other membrane protein bound the radiolabeled retinoid. The detection of noncovalently bound [ 3 H]retinoid would not be expected in this assay. The results are consistent with the formation of a Schiff base linkage between [ 3 H]retinal and the ~30-kDa protein and subsequent reduction of the bond to a stable secondary amine in the presence of sodium borohydride. Synthesis of all-trans-retinal in cultured ARPE-hRGR cells The specific binding of [ 3 H]retinoid to RGR suggests that all-trans-retinal is produced in ARPE-hRGR cells as a chromophore for the RGR opsin. To verify that alltrans-retinal is synthesized in ARPE-hRGR cells, we extracted [ 3 H]-labeled retinals after incubation of the cells in the presence of [ 3 H]all-trans-retinol. The distribution of [ 3 H]retinoids extracted from the ARPE-hRGR cells included a

13 Yang and Fong Page 13 significant pool of retinyl esters (44%), all-trans-retinal (16%) and all-transretinol (40%) (Fig. 2, upper panel). The specific activity of [ 3 H]all-trans-retinal from the cells was ~51 Ci/mmol, as determined from the amount of radioactivity and the corresponding absorbance peak of all-trans-retinal in the HPLC chromatogram. This specific activity was virtually identical to that of the added exogenous [ 3 H]all-trans-retinol. The control ARPE-19 cells had a distinct profile of [ 3 H]retinoids and contained retinyl esters (66%), all-trans-retinal (3%) and all-trans-retinol (31%) (Fig. 2, lower panel). Neither type of cell contained significant amounts of 11-cis-retinal or 11-cis-retinol. The results indicate that the retinal isomer bound to RGR in the dark in ARPE-hRGR cells is solely alltrans-retinal. ARPE-hRGR cells cultured in growth medium also synthesize the alltrans-retinal chromophore from a precursor in serum (Fig. 3). The fresh culture medium supplemented with 10% FBS contains at least 30 pmol/ml of all-trans-retinol. all-trans-retinal was found in ARPE-hRGR cells (Fig. 3A), but not in control ARPE-19 cells that were maintained in serum-containing medium (Fig. 3B). When the ARPE-hRGR cells were preincubated in serumfree medium for 16 h, all-trans-retinal fell to an undetectable level (Fig. 3C). Irradiation of the ARPE-hRGR cells resulted in stereospecific isomerization of the endogenous all-trans-retinal to 11-cis-retinal (Fig. 4). The proportion of alltrans- to 11-cis-retinal in the ARPE-hRGR cells was 2.60 to 0 pmol in the dark and 1.12 to 0.55 pmol after illumination with 470-nm monochromatic light. Photoisomerization to 13-cis-retinal was not detected. all-trans-retinol dehydrogenase activity in ARPE-hRGR cells The production of the all-trans-retinal chromophore from all-trans-retinol requires

14 Yang and Fong Page 14 an oxidation reaction. The proposed enzymatic reaction may be catalyzed by a previously uncharacterized all-trans-retinol dehydrogenase in the RPE. We tested microsomal membranes from ARPE-hRGR cells for all-trans-retinol dehydrogenase activity in presence of nicotinamide dinucleotide cofactors. The membranes were capable of producing all-trans-retinal from all-trans-retinol in the presence of NADP but not NAD (Fig. 5A). At concentrations of NADP >8 µm, the synthesis of all-trans-retinal increased 5-fold in comparison to activity without cofactor or with NAD. The putative NADPH-dependent retinol dehydrogenase strongly preferred the all-trans isomer of retinol and did not react with 11-cis-retinol (Fig. 5B). The microsomal membranes from control ARPE-19 cells also contained all-trans-retinol dehydrogenase activity in the presence of NADP (results not shown). Uptake and binding of [ 3 H]retinoid to RGR in isolated bovine RPE cells Although the parental ARPE-19 cells maintain many characteristics of RPE cells, they are highly deficient in the expression of several RPE proteins, including RGR, and may not function normally. To demonstrate that the synthesis of all-trans-retinal and specific binding of the chromophore to RGR are physiological properties of normal RPE cells, we incubated freshly isolated bovine RPE cells with precursor [ 3 H]all-trans-retinol (Fig. 6). The uptake of [ 3 H]all-trans-retinol resulted in radiolabeling of a specific protein band that was equal in size to the RGR opsin (Fig. 6A). RGR was detectable by Western blot assay after the short-term incubation of the bovine RPE cells (Fig. 6B), but not after 5 days of culture in serum-containing medium (results not shown). After 3 h of incubation, the cells synthesized retinyl esters (62%), all-trans-retinal (23%), 11-cis-retinal (2%), and 11-cis-retinol (2%) (Fig. 6C). all-trans-retinol

15 Yang and Fong Page 15 (7%) and a small amount of 13-cis-retinal (3%) were also extracted from the cells. Cell-free synthesis and binding of all-trans-retinal to RGR in vitro Since the accumulation of all-trans-retinal in ARPE-hRGR and ARPE- 19 cells was dependent on the presence of RGR (Figs. 2 and 3), the synthesis or stability of all-trans-retinal may be connected with its binding to the opsin. To investigate chromophore synthesis and binding to RGR in an in vitro system, we incubated isolated RPE membranes with [ 3 H]all-trans-retinol in the presence of NADP and analyzed the binding of [ 3 H]retinoid to RGR. The results indicated that [ 3 H]all-trans-retinal was synthesized in membranes and bound directly to RGR with high specificity (Fig. 7). Nontransduced ARPE-19 cells did not contain the radiolabeled ~30-kDa protein band (Fig. 7A). The binding of [ 3 H]retinoid to RGR in bovine RPE microsomes was stimulated 3.1 fold by addition of NADP (Fig. 7B). The results indicate that all-trans-retinal, which is synthesized by the membrane-associated all-trans-retinol dehydrogenase, can be channeled to the binding site of RGR in the cell-free system.

16 Yang and Fong Page 16 DISCUSSION Little is known about the synthesis of all-trans-retinal as a chromophore for the RGR opsin in RPE cells. The endogenous all-trans-retinal bound to RGR may be synthesized directly from all-trans-retinol that is generated upon phototransduction, or from serum all-trans-retinol. This reaction would require a novel all-trans-retinol dehydrogenase in the RPE. In this paper, we demonstrate further evidence for the oxidation of precursor all-trans-retinol in RPE cells and its physiological incorporation into the chromophore of RGR in the dark. After its uptake into RPE cells, all-trans-retinol is converted rapidly into retinyl esters. The esterification of retinol is catalyzed by LRAT, the activity of which is higher than is known for most other visual cycle enzymes. In addition to esterification, our results indicate that retinol is oxidized efficiently and that significant amounts of all-trans-retinal accumulate in ARPE-hRGR and bovine RPE cells in the absence of light. Despite the high LRAT activity, the ratio of all-trans-retinal to retinyl esters formed after 3 h of incubation with precursor all-trans-retinol is 1:2.8 and 1:2.7 in ARPE-hRGR and bovine RPE cells, respectively. Since the specific activity of [ 3 H]all-trans-retinal in the ARPE-hRGR cells was close to that of the precursor [ 3 H]all-trans-retinol, alltrans-retinol can be converted directly into all-trans-retinal without entering the pool of endogenous retinyl esters. The results suggest that retinol esterification by LRAT and oxidation by an all-trans-retinol dehydrogenase represent an early and important bifurcation point in the processing of retinol after uptake into RPE cells. The esterification and oxidation are both critical reactions for continuance of the visual cycle in that they catalyze formation of

17 Yang and Fong Page 17 the substrate for a putative isomerohydrolase (18-20) and the chromophore for RGR, respectively. The difference in steady-state levels of all-trans-retinal in ARPE-hRGR and ARPE-19 cells indicates that the accumulation of all-trans-retinal is highly dependent on the presence of RGR. Although the membranes from ARPE-19 and other cells contain constitutive NADPH-dependent all-trans-retinol dehydrogenase activity and are capable of synthesizing all-trans-retinal from all-trans-retinol, all-trans-retinal generally does not accumulate in these cells (43, 44). Normal RPE and cultured ARPE-hRGR cells have uniquely elevated levels of all-trans-retinal, which may be stabilized by covalent binding to RGR. The sequestration of all-trans-retinal to RGR would block the retinoid from reversible reduction, oxidation to retinoic acid, formation of non-specific Schiff bases, and possible generation of harmful bis-retinoid, N-retinylidene-Nretinylethanolamine compounds within the RPE (45). Failure to detect alltrans-retinal in RPE may result from inefficient extraction methods or the loss of RGR expression in cultured cells. The irradiation of ARPE-hRGR cells resulted in stereospecific isomerization of all-trans-retinal to 11-cis-retinal, indicating a functional recombinant RGR and physiological activity of the endogenous all-trans-retinal. With the ARPE-hRGR cells, the 11-cis retinoid metabolic pathways that lie downstream of the RGR opsin can be analyzed under a variety of culture and lighting conditions. Our finding of all-trans-retinal synthesis and accumulation in RPE cells is not inconsistent with data from previous studies. Timmers et al. have demonstrated a slow constant increase in the synthesis of all-trans-retinal in isolated bovine RPE cells (46). Others have reported NADPH-dependent all-

18 Yang and Fong Page 18 trans-retinol dehydrogenase activity in RPE membranes. As early as 1975, Zimmerman et al. noted that RPE microsomes contain a relatively high amount of all-trans-retinol dehydrogenase activity in membrane vesicles that fractionate separately from contaminant membranes with the photoreceptor all-trans-retinol dehydrogenase (14, 15). These interesting findings further support the notion that the all-trans-retinal chromophore is synthesized by a novel all-trans-retinol dehydrogenase in the RPE. Nevertheless, the putative RPE all-trans-retinol dehydrogenase has not been identified unequivocally. Further characterization of the all-trans-retinol dehydrogenase in ARPE-hRGR and isolated RPE cells is required. Our results indicate that the retinol dehydrogenase in ARPE-hRGR cells is membrane-bound, prefers NADP as the cofactor in oxidation, and has high substrate stereospecificity for all-transretinol versus the 11-cis isomer. We hypothesize that this enzyme is responsible for chromophore synthesis in ARPE-hRGR cells. To investigate the coupling between chromophore synthesis and its binding to RGR, we examined the reactions in vitro. In a cell-free membrane system, RGR had apparent and specific access to the newly synthesized alltrans-retinal. The microsomal membranes of both the human ARPE-hRGR cell line and bovine RPE cells were sufficient for the synthesis of all-transretinal and its channeling to the binding site of RGR. The results indicate that both cells contain a membrane-bound all-trans-retinol dehydrogenase and may have a highly conserved system of providing the chromophore of RGR. The binding of [ 3 H]all-trans-retinal to RGR was enhanced in the presence of NADP, although RGR was also radiolabeled in the absence of added cofactor. Like preparations of various dehydrogenases (47), it is likely that the washed membranes still contained a significant amount of endogenous enzyme-bound

19 Yang and Fong Page 19 NADP that allowed background synthesis of [ 3 H]all-trans-retinal. The physical relationship between the RPE all-trans-retinol dehydrogenase and RGR is unknown. The all-trans-retinal chromophore of RGR may be synthesized also from precursor β,β-carotene by the oxidative cleavage activity of an enzyme, β,βcarotene-15,15'-dioxygenase (Bcdo) that has been found in the RPE (48, 49). The oxidative cleavage of β,β-carotene by Bcdo would directly generate all-transretinal under dark or photic conditions. The all-trans-retinal synthesized from β,β-carotene may then bind to RGR physiologically. In humans, β,β-carotene and vitamin A are available to the RPE at plasma levels of and µg/ml, respectively (50). On the other hand, only all-trans-retinol is transported from the photoreceptors to the RPE as an intermediate in the visual cycle. Consequently, the RPE all-trans-retinol dehydrogenase would be required to process all-trans-retinol rapidly in a continuous photic visual cycle. Like rhodopsin, the RGR opsin relies on retinol dehydrogenases for the processing of its retinal chromophore in biochemical pathways that lie upstream and downstream of photoisomerization. In contrast to the two-cell rhodopsin system, the all-trans-retinal chromophore of RGR is synthesized by a proximal retinol dehydrogenase within membranes of the RPE itself. After irradiation of RGR, the bound 11-cis-retinal is dissociated and converted to the alcohol by 11-cis-retinol dehydrogenase (51). RGR and the 11-cis-retinol dehydrogenase co-purify consistently and may be tightly associated in a protein complex. The evidence for functional interaction of all-trans-retinol dehydrogenase, RGR opsin, and 11-cis-retinol dehydrogenase suggests a model for the flow of retinoids in the photic visual cycle (Fig. 8). In this model, all-

20 Yang and Fong Page 20 trans-retinal is converted to 11-cis-retinal by rapid photoisomerization, and the overall rate of conversion of all-trans retinoids to the 11-cis isomer is limited by binding kinetics and the enzymatic reactions catalyzed by the retinol dehydrogenases. The ARPE-hRGR and isolated RPE cells provide a promising approach to compare and analyze the biochemistry and kinetics of retinoid processing in the RGR system.

21 Yang and Fong Page 21 ACKNOWLEDGMENT The authors thank Daiwei Shen and Pu Chen for technical assistance. This work was supported, in part, by National Institutes of Health grants EY03040 and EY08364.

22 Yang and Fong Page 22 REFERENCES 1. Saari, J. C. (2000) Investig. Ophthalmol. Vis. Sci. 41, McBee, J. K., Palczewski, K., Baehr, W., and Pepperberg, D. R. (2001) in Progress in Retinal and Eye Research (Osborne, N. N., and Chader, G. J. eds) Vol. 20, pp , Elsevier Science Ltd., Great Britain 3. Saari, J. C., and Bredberg, D. L. (1988) J. Biol. Chem. 263, Saari, J. C., and Bredberg, D. L. (1989) J. Biol. Chem. 264, Shi, Y. Q., Hubacek, I., and Rando, R. R. (1993) Biochemistry 32, Zimmerman, W. F. (1974) Vision Res. 14, Bridges, D. D. (1976) Exp. Eye Res. 22, Saari, J. C., Garwin, G. G., Van Hooser, J. P., and Palczewski, K. (1998) Vision Res. 38, Palczewski, K., Van Hooser, J. P., Garwin, G. G., Chen, J., Liou, G. I., and Saari, J. C. (1999) Biochemistry 38, Qtaishat, N. M., Okajima, T.-I. L., Li, S., Naash, M. I., and Pepperberg, K. R. (1999) Investig. Ophthalmol. Vis. Sci. 40, Blaner, W. S., Das, S. R., Gouras, P., and Flood, M. T. (1987) J. Biol. Chem. 262, Mata, N. L., Tsin, A. T., and Chambers, J. P. (1992) J. Biol. Chem. 267, Lion, F., Rotmans, J. P., Daemen, F. J. M., and Bonting, S. L. (1975) Biochim. Biophys. Acta 384,

23 Yang and Fong Page Zimmerman, W. F., Lion, F., Daemen, F. J. M., and Bonting, S. L. (1975) Exp. Eye Res. 21, Zimmerman, W. F. (1976) Exp. Eye Res. 23, Chai, X., Zhai, Y., and Napoli, J. L. (1997) J. Biol. Chem. 272, Jang, G. F., McBee, J. K., Alekseev, A. M., Haeseleer, F., and Palczewski, K. (2000) J. Biol. Chem. 275, Law, W. C., and Rando, R. R. (1988) Biochemistry 27, Deigner, P. S., Law, W. C., Canada F. J., and Rando, R. R. (1989) Science 244, Canada, F. J., Law, W. C., Rando, R. R., Yamamoto, T., Derguini, F., and Nakanishi, K. (1990) Biochemistry 29, McBee, J. K., Kuksa, V., Alvarez, R., de Lera, A. R., Prezhdo, O., Haeseleer, F., Sokal, I., and Palczewski, K. (2000) Biochemistry 39, Bok, D., Ong, D. E., and Chytil, F. (1984) Investig. Ophthalmol. Vis. Sci. 25, Eisenfeld, A. J., Bunt-Milam, A. H., and Saari, J. C. (1985) Exp. Eye Res. 41, Saari, J. C., Bredberg, L., and Garwin, G. G. (1982) J. Biol. Chem. 257, Bunt-Milam, A. H., and Saari, J. C. (1983) J. Cell Biol. 97, Shen, D., Jiang, M., Hao, W., Tao, L., Salazar, M., and Fong H. K. W. (1994) Biochemistry 33,

24 Yang and Fong Page Tao, L., Shen, D., Pandey, S., Hao, W., Rich K. A., and Fong, H. K. W. (1998) Mol. Vis. 4, 25 ( 28. Jiang, M., Pandey, S., and Fong, H. K. W. (1993) Investig. Ophthalmol. Vis. Sci. 34, Pandey, S., Blanks, J. C., Spee, C., Jiang, M., and Fong, H. K. W. (1994) Exp. Eye Res. 58, Ozaki, K., Hara, R., Hara, T., and Kakitani, T. (1983) Biophys. J. 44, Hao, W., and Fong, H. K. W. (1996) Biochemistry 35, Hao, W., and Fong, H. K. W. (1999) J. Biol. Chem. 274, Chen, P., Hao., W., Rife, L., Wang, X. P., Shen, D., Chen, J., Ogden, T., Van Boemel G. B., Wu, L., Yang, M., and Fong, H. K. W. (2001) Nat. Genet. 28, Chen, X.-N., Korenberg, J. R., Jiang, M., Shen, D., and Fong, H. K. W. (1996) Hum. Genet. 97, Morimura, H., Saindelle-Ribeaudeau, F., Berson, E. L., and Dryja, T. P. (1999) Nat. Genet. 23, Yang, M., Wang, X.-G., Stout, J. T., Chen, P., Hjelmeland, L. M., Appukuttan, B., and Fong, H. K. W. (2000) Mol. Vis. 6, ( 37. Landers, G. M. (1990) Methods Enzymol. 189, Dunn, K. C., Aotaki-Keen, A. E., Putkey, F. R., and Hjelmeland, L. M. (1996) Exp. Eye Res. 62, Groenendijk, G. W. T., De Grip, W. J., and Daemen, F. J. M. (1979) Anal. Biochem. 99,

25 Yang and Fong Page Groenendijk, G. W. T., De Grip, W. J., and Daemen, F. J. M. (1980) Biochim. Biophys. Acta 617, Ozaki, K., Terakita, A., Hara, R., and Hara, T. (1986) Vision Res. 26, Hao, W., Chen, P., and Fong, H. K. W. (2000) Methods Enzymol. 316, Williams, J. B., Pramanik, B. C., and Napoli, J. L. (1984) J. Lipid Res. 25, McCormick, A. M., and Napoli, J. L. (1982) J. Biol. Chem. 257, Parish, C. A., Hashimoto, M., Nakanishi, K., Dillon, J. and Sparrow, J. (1998) Proc. Natl. Acad. Sci. 95, Timmers, A. M., van Groningen-Luyben, D. A., and de Grip, W. J. (1991) Exp. Eye Res. 52, Kato, T., Berger, S. J., Carter, J. A., and Lowry, O. H. (1973) Anal. Biochem. 53, Redmond, T. M., Gentleman, S., Duncan, T., Yu, S., Wiggert, B., Gantt, E., and Cunningham, F. X. Jr. (2001) J. Biol. Chem. 276, Yan, W., Jang, G. F., Haeseleer, F., Esumi, N., Chang, J., Kerrigan, M., Campochiaro, M., Campochiaro, P., Palczewski, K., and Zack, D. J. (2001) Genomics 72, Schünemann, H. J., Grant, B. J. B., Freudenheim, J. L., Muti, P., Browne, R. W., Drake, J. A., Klocke, R. A., and Trevisan, M. (2001) Am. J. Respir. Crit. Care Med. 163, Chen, P., Lee, T. D., and Fong, H. K. W. (2001) J. Biol. Chem. 276,

26 Yang and Fong Page 26 FIGURE LEGENDS FIG. 1. Specific binding of [ 3 H]retinoid to RGR in cultured ARPE-hRGR cells. (A) Incorporation of precursor [ 3 H]all-trans-retinol into the chromophore of RGR. ARPE-hRGR cells were preincubated overnight in serum-free RPMI1640 medium and then incubated with [ 3 H]all-trans-retinol (0.25 µci/ml, 50 Ci/mmol) in RPMI1640 for various lengths of time, as indicated. Total membrane proteins (50 µg per lane) were prepared and analyzed by gel electrophoresis and fluorography. The autoradiographic film was exposed for 15 days. (B) Time course of the binding of [ 3 H]retinoid to RGR. ARPE-hRGR cells were incubated with [ 3 H]all-trans-retinol for various lengths of time, as described in A. The protein bands were analyzed by scanning the autoradiographic films and using Scion Image 1.62c software to determine the relative band density. The results from two separate experiments were averaged and normalized with respect to band density at 1 h. FIG. 2. [ 3 H]Retinoids extracted from cultured ARPE-hRGR and ARPE-19 cells. ARPE-hRGR and control ARPE-19 cells were preincubated overnight in serumfree RPMI1640 medium and subsequently, incubated with [ 3 H]all-trans-retinol (0.25 µci/ml, 50 Ci/mmol) in RPMI1640 for 3 h. The [ 3 H]retinoids were then extracted from the cells in the presence of hydroxylamine and separated by HPLC. Four drops of eluate were collected per fraction from 4 to 27.5 min. RE, retinyl ester; t-ral, syn all-trans-retinaloxime; t-rol, all-trans-retinol. FIG. 3. Endogenous all-trans-retinal in cultured ARPE-hRGR cells. ARPEhRGR and control ARPE-19 cells were cultured in serum-containing or serumfree DME/F12 (1:1) medium. all-trans-retinal was extracted by hydroxylamine

27 Yang and Fong Page 27 derivatization and separated by HPLC. (A) ARPE-hRGR and (B) control ARPE- 19 cells cultured in serum-containing medium. (C) ARPE-hRGR cells cultured in serum-containing medium and subsequently, incubated in serum-free medium for 16 h. t-ral, syn all-trans-retinaloxime. FIG. 4. Photoisomerization of all-trans-retinal in ARPE-hRGR cells. ARPEhRGR cells (~1.5 x 10 7 ) were cultured in the dark in serum-containing (retinolcontaining) medium. The cells were suspended in PBS and either kept in the dark (upper panel) or illuminated with 470-nm monochromatic light at 410 lux (lower panel). The retinal isomers were extracted by hydroxylamine derivatization and analyzed by HPLC. 11-cis, 11-cis-retinal; all, all-trans-retinal (in the form of syn-retinaloximes). FIG. 5. all-trans-retinol dehydrogenase activity in ARPE-hRGR microsomal membranes. (A) all-trans-retinol dehydrogenase activity in the presence of NADP or NAD. ARPE-hRGR membranes (0.2 mg protein/ml) were incubated for 10 min with [ 3 H]all-trans-retinol in the presence of various concentrations of NADP or NAD. The reaction products were extracted by hydroxylamine derivatization and analyzed by HPLC. The results are expressed as the percentage of precursor [ 3 H]all-trans-retinol converted to [ 3 H]all-trans-retinal. (B) NADPH-dependent retinol dehydrogenase activity with all-trans-retinol or 11-cis-retinol substrate. ARPE-hRGR membranes (0.5 mg protein/ml) were incubated with 5 µm exogenous all-trans-retinol or 11-cis-retinol in the presence of 200 µm NADP. Retinals were extracted by hydroxylamine derivatization and analyzed by HPLC. The extraction efficiency was monitored by the addition of [ 3 H]all-trans-retinol as an internal standard.

28 Yang and Fong Page 28 FIG. 6. Incorporation of precursor [ 3 H]all-trans-retinol into the chromophore of RGR in bovine RPE cells. (A) [ 3 H]retinoid-bound proteins in bovine RPE membranes. Total membrane proteins (240 µg) were prepared and analyzed by gel electrophoresis and fluorography. The autoradiographic film was exposed for 22 days. (B) Immunoblot analysis of membrane proteins (40 µg) indicating immunoreactivity to an anti-bovine RGR antibody. (C) [ 3 H]Retinoids from bovine RPE cells isolated from six eyes and incubated with [ 3 H]all-trans-retinol for 3 h. The [ 3 H]retinoids were extracted in the presence of hydroxylamine and separated by HPLC. Four drops of eluate were collected per fraction from 4 to 29.5 min. RE, retinyl ester; t-ral, all-trans-retinaloxime; c-ral, 11-cisretinaloxime; t-rol, all-trans-retinol; c-rol, 11-cis-retinol. FIG. 7. Linkage of synthesis and binding of all-trans-retinal to RGR in vitro. (A) ARPE-19 and (lane 1) ARPE-hRGR (lane 2) microsomal membranes (88 µg protein) incubated with [ 3 H]all-trans-retinol in the presence of 200 µm NADP. (B) Synthesis and binding of all-trans-retinal to bovine RGR. Bovine microsomal membranes (54 µg protein) were washed 3 times and incubated with [ 3 H]all-trans-retinol in the absence (lane 1) or in the presence (lane 2) of 200 µm NADP. The autoradiographic films were exposed for 13 and 15 days in A and B, respectively. FIG. 8. Proposed model of the photic visual cycle and interaction of retinol dehydrogenases with the chromophore of RGR. A light-dependent pathway of the visual cycle and regeneration of rhodopsin is dependent on RGR. In this model, all-trans-retinol from the photoreceptors is converted to alltrans-retinal by an all-trans-retinol dehydrogenase in the RPE. The all-transretinal accumulates in microsomal membranes by binding to RGR and is

29 Yang and Fong Page 29 isomerized stereospecifically to 11-cis-retinal in the presence of light. The bound 11-cis-retinal dissociates efficiently upon reduction to 11-cis-retinol by 11-cis-retinol dehydrogenase or may be transported directly to photoreceptors for recombination with the apoprotein of rhodopsin. The availability of free chromophore binding sites on RGR may be a rate-limiting step in the accumulation or turnover of all-trans-retinal. A photoactivated RGR may also stimulate a G protein regulatory pathway (G*) that leads to higher synthesis of 11-cis-retinoids. The thermal isomerization of all-trans-retinol to 11-cis-retinol follows two proposed alternative pathways that involve an I) isomerohydrolase (18-20) or II) carbocation intermediate (2, 21). t-ral, alltrans-retinal; t-rol, all-trans-retinol; t-re, all-trans-retinyl ester; c-ral, 11- cis-retinal; c-rol, 11-cis-retinol; c-re, 11-cis-retinyl ester; prrdh, photoreceptor retinol dehydrogenase; LRAT, lecithin:retinol acyltransferase; REH, retinyl ester hydrolase; t-rdh; all-trans-retinol dehydrogenase; c-rdh, 11-cis-retinol dehydrogenase.

30 A 20 min 1 h 3 h 60 kda

31

32

33 A B C

34

35

36 A B C 50 kda 37 syn anti 25 3

37 A B kda 37 25

38 RGR t-ral Light RGR c-ral t-rdh t-rol G* t-ral REH Carbocation intermediate LRAT c-rol t-re c-re Isomerohydrolase c-rdh c-rdh c-ral RPE t-rol prrdh t-ral t-ral Rhodopsin Light c-ral Rhodopsin c-ral Photoreceptor