Analysis of the Visual Cycle in Cellular Retinol-Binding Protein Type I (CRBPI) Knockout Mice

Abstract

purpose. To determine whether the visual cycle is affected in mice without a functional gene for cellular retinol-binding protein type I (CRBPI ^−/− mice).

methods. Visual-cycle retinoids and rhodopsin levels were analyzed in eyes of dark adapted (DA) CRBPI ^−/− and wild-type (wt) mice before and during recovery from a flash. The rate of dark adaptation was analyzed using electroretinography (ERG).

results. all-trans-Retinyl esters were reduced to approximately 33% of wt levels in DA CRBPI ^−/− mice. Recovery from a flash in wt mice produced transient accumulations of all-trans-retinal and all-trans-retinyl ester, as the pulse of retinoid produced by the flash traversed the visual cycle. In CRBPI ^−/− mice, all-trans-retinal accumulated transiently, as in wt mice. However, all-trans-retinol also accumulated transiently in the neural retina, and the transient increase in all-trans-retinyl ester of the wt was reduced. Rates of 11-cis-retinal and rhodopsin formation were comparable in wt and CRBPI ^−/− mice. Dark adaptation was delayed by a factor of approximately two.

conclusions. The accumulation of all-trans-retinol in neural retina, in the absence of CRBPI and the reduced amount of retinyl esters in the RPE suggest that the binding protein participates in a process that drives diffusion of all-trans-retinol from photoreceptor cells to RPE, perhaps by delivering vitamin A to lecithin-retinol acyltransferase (LRAT) for esterification. Because the perturbation occurred upstream of a slow step of the visual cycle, there was no major impairment of the rate of visual pigment regeneration.

The amount of vitamin A in the visual system is small compared with its turnover during daylight, and the retinoid chromophore must be recycled after bleaching to allow continued vision. ¹ The cycle of bleaching (photoisomerization) and regeneration of rhodopsin in rod photoreceptors proceeds through a complex pathway (Fig. 1) . Photoisomerization of the chromophore of rhodopsin converts 11-cis-retinal to all-trans-retinal, which is reduced to all-trans-retinol by reduced nicotinamide adenine dinucleotide phosphate (NADPH) and photoreceptor retinol dehydrogenase (prRDH) within the rod outer segment. The all-trans-retinol leaves the rod photoreceptor cells and diffuses to the retinal pigment epithelium (RPE), where it is sequentially esterified by lecithin-retinol acyltransferase (LRAT), converted to 11-cis-retinol by an isomerase, and oxidized to 11-cis-retinal by nicotinamide adenine dinucleotide (NAD) and one or more short-chain dehydrogenase reductases. The 11-cis-retinal then diffuses back into the rod photoreceptor cell, where it regenerates rhodopsin and completes the visual cycle (for reviews of the visual cycle see Refs. ^{1 2 3 4} ). Molecular details of the visual cycle remain, for the most part, very poorly characterized. In particular, the isomerase remains elusive, and the intercellular diffusion processes are not very well understood. 

The several retinoid-binding proteins that are present in retina offer an approach to understanding the visual cycle. The functions of these proteins have been inferred from in vitro studies, and the recent availability of knockout mice has allowed these hypotheses to be tested. Considerable insight into molecular aspects of the visual cycle has been derived from this approach. For example, deletion of the gene for cellular retinaldehyde-binding protein (CRALBP) produces mice with a substantial delay in visual pigment regeneration because of an impaired isomerase reaction, ⁵ suggesting that the binding protein is a functional component of this reaction. Deletion of the gene for interphotoreceptor retinoid-binding protein (IRBP), a protein thought to function in the intercellular diffusion of retinoids, results in mice with normal visual-cycle kinetics, ^{6 7} suggesting that the role of this protein in visual physiology should be reevaluated. Mice deficient in the gene for retinol-binding protein (RBP) were blind as young adults but eventually regained visual function when fed a vitamin A–replete diet, ⁸ indicating that vitamin A can be delivered to RPE in adults through carriers other than RBP. 

RPE and Müller (glial) cells of the retina contain cellular retinol-binding protein type I (CRBPI). ^{9 10 11} CRBPI has been suggested, based on in vitro studies, to play a role in several reactions of vitamin A metabolism, including oxidation to retinaldehyde, a precursor of retinoic acid ^{12 13 14} ; esterification by LRAT ^{15 16 17} ; hydrolysis of retinyl esters ^{18 19} ; and cellular uptake. ^{17 20 21} Thus, it was surprising that mice with a targeted disruption of the CRBPI gene (CRBPI ^−/− mice) were anatomically normal, ²² indicating that retinoic acid production during development was not grossly perturbed. The amounts of hepatic retinyl esters were reduced, however, relative to wild-type (wt) mice, suggesting that CRBPI delivers vitamin A to LRAT for esterification. In addition, the turnover rate of vitamin A was six times faster than in the wt animals, perhaps explained by a reduced residence time of the vitamin in tissue retinyl ester pools. 

The phenotype of perturbed hepatic vitamin A metabolism suggested that the visual cycle may also be affected, because CRBPI and LRAT are both present in RPE, ^{9 23 24 25} and both have been postulated to play important roles in the regeneration of visual pigments. The results reported herein indicate that in CRBPI ^−/− mice the amount of retinyl esters in RPE is reduced to approximately 33% of wt amounts, that the diffusion of all-trans-retinol from photoreceptor cells to RPE is slower, and that resensitization of the visual system after bleaching is delayed approximately twofold. The slower rate of diffusion of all-trans-retinol from neural retina to RPE is probably due to the inefficient delivery of all-trans-retinol to LRAT in RPE. 

Materials and Methods

Animals

CRBPI ^+/+ and CRBPI ^−/− mice were derived from lines previously described. ²² Genotypes of the animals used were confirmed by polymerase chain reaction, as described. ²² All procedures with animals were approved by the University of Washington Animal Care Committee and were in accord with recommendations of the American Veterinary Medical Association Panel on Euthanasia ²⁶ and the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. 

Photobleaching and Regeneration Protocols

Mice were dark adapted for at least 24 hours before each experiment. Handheld mice were individually subjected to a single flash from a photographic flash unit (Sunpak Auto 433D Thyristor; ToCad America, Inc., Parsippany, NJ) at the maximum setting. Mice were killed by cervical dislocation at intervals after the flash in dim red light, and the eyes were removed, rinsed with sterile water, and frozen within 30 seconds. 

Rhodopsin Determination

All procedures involving visual pigments were performed in dim red light to minimize photoisomerization. Rhodopsin was solubilized from homogenized whole eyes, as described previously, ²⁷ and its concentration determined by the decrease in absorbance at 500 nm after bleaching, using an extinction coefficient of 42,000 l · mol/cm. 

Retinoid Extraction and Analysis

All procedures with retinoids were performed under red illumination to avoid photoisomerization and photodecomposition. Retinoids were extracted from thawed, homogenized, whole mouse eyes (four per analysis) and analyzed with a normal-phase HPLC column, as described previously, ^{5 28} except that ethyl all-trans-9-(4-methoxy-2,3,6-trimethylphenyl)-3,7-dimethyl-2,4,6,8-nonatetraenoate (TMMP) ²⁹ was used as the internal standard instead of retinyl acetate. 

ERG Analysis

ERG analysis was performed essentially as described earlier. ^{5 30} The following protocol was used to determine dark adaptation time courses after complete bleaching. Mice were anesthetized, and electrodes were applied and tested with dim flashes for proper positioning. The eye was then exposed, through a fiber optic cable, to illumination (1 mW/cm²) from a halogen source for 5 minutes, a level of illumination calculated to bleach more than 99% of the rhodopsin in the retina. Responses to a 4.5 μJ/cm² flash of white light, which produced 22,000 photoisomerizations per rod, were then recorded at 10-minute intervals. Amplitudes of the a-wave were measured at 7 ms after the flash and averaged from data taken from at least three mice. Mice were maintained in an anesthetized state throughout by boosts with 25% of the initial dose of anesthetic delivered every 20 minutes, beginning 45 minutes into the experiment. The electrode was removed and replaced with fresh methyl cellulose every 20 minutes. 

Statistical Analysis

Rate constants were obtained for the decay of all-trans-retinal and for the generation of 11-cis-retinal from first-order rate plots. The statistical significance of the difference in rate constants was determined by Student’s t-test and a 95% confidence level. The lines shown in Figure 5 were fit to the data by linear regression analysis. 

Results

Visual-Cycle Activity in CRBPI ^−/− Mice

We analyzed visual-cycle retinoids in CRBPI ^−/− and wt mice during recovery from a flash to determine whether the absence of CRBPI affects the flow of retinoids. HPLC traces obtained 15 minutes after a flash are shown in Figure 2 . Retinyl esters were reduced in CRBPI ^−/− mice, and all-trans-retinol was elevated, relative to levels in wt control animals. A more detailed analysis of visual-cycle retinoid intermediates is shown in Figure 3 . The flash used in these experiments in wt mice bleached approximately 35% of the rhodopsin, which was reflected in a corresponding decrease in the amount of 11-cis-retinal in the retinoid profile (Fig. 3 , top). The levels of all-trans-retinal and all-trans-retinyl esters increased transiently during recovery from the flash, as the bolus of retinoid released by photoisomerization made its way around the visual cycle. The concentrations of other visual-cycle retinoids were not appreciably altered during the recovery, indicating that reactions leading to their utilization are rapid, relative to the reduction of all-trans-retinal and to the isomerization of all-trans- to 11-cis-retinoid, as noted earlier. ⁵ The distribution of retinoids in dark-adapted CRBPI ^−/− mice was similar to that in the wt control mice, except that retinyl esters were reduced to approximately one third the wt amount (Fig. 3 , bottom). During recovery from a flash, both all-trans-retinal and all-trans-retinyl esters accumulated transiently, as observed in the wt. However, all-trans-retinol also increased transiently, suggesting that a new slow step had been introduced into the visual cycle. 

The first-order rate constant for decay of all-trans-retinal in wt mice (k = −0.069 minutes) was significantly different from that observed in CRBPI ^−/− mice (k = −0.050 minutes) at a confidence level of 95%. In addition, the amount of all-trans-retinal remaining 120 minutes after the flash was significantly higher in CRBPI ^−/− mice. Although the recovery of 11-cis-retinal in the CRBPI ^−/− mice appeared convex to the eye (Fig. 3 , bottom), the overall rates of synthesis were comparable in CRBPI ^−/− and wt mice (discussed later). 

Site of Accumulation of all-trans-Retinol

The information in Figure 3 shows that all-trans-retinol accumulated transiently in eyes from CRBPI ^−/− mice during recovery from a flash, but because whole eyes were analyzed, it does not indicate whether accumulation occurred in neural retina or the RPE. To resolve this question, eyes were removed from mice before (dark-adapted) and 30 minutes after a flash and dissected into neural retina and RPE-choroid fractions. Retinoids were extracted and analyzed by HPLC. In eyes from dark-adapted CRBPI ^−/− mice, retinyl esters were exclusively associated with the RPE-choroid fraction (results not shown), demonstrating that the dissection was complete. After a flash, all-trans-retinol accumulated largely in neural retina and only to a minor extent in RPE-choroid (Fig. 4) . We presume the accumulation of all-trans-retinol in neural retina was in rod photoreceptor cell outer segments, because the enzyme responsible for its production (prRDH) is localized there. ³¹  

Rates of Rhodopsin Regeneration

We examined the rates of regeneration of rhodopsin to determine whether the overall function of the visual cycle was affected by the unusual distribution of retinoids noted during recovery from a flash. Dark-adapted wt and CRBPI ^−/− mice eyes contained approximately 650 picomoles rhodopsin/eye, similar to values we have reported for other mice. ^{5 6} Recovery from a flash occurred at approximately the same rate in both types of mice (Fig. 5) , indicating that the absence of CRBPI did not affect the overall function of the visual cycle to a marked extent. 

Rates of Dark Adaptation

To determine whether the absence of CRBPI affects the rate of recovery of visual sensitization after a flash we measured full-field ERG responses in dark-adapted mice and in mice recovering in the dark from exposure to intense illumination that bleached approximately 99% of their visual pigment. ERG responses to a test flash were recorded at intervals during the subsequent dark adaptation period. The a-wave amplitudes at 7 ms after the test flash are plotted in Figure 6 versus the time of dark adaptation after the conditioning illumination. This analysis showed that the recovery of the test flash responses after conditioning illumination, a measure of dark adaptation, was delayed by approximately twofold compared with the wt control animals. 

Discussion

Targeted disruption of the gene encoding CRBPI resulted in anatomically normal mice with reduced hepatic stores of retinyl esters and a sixfold faster whole-body turnover of vitamin A. ²² These findings strongly support the proposed role of CRBPI in promoting esterification of retinol through hepatic LRAT and clearly indicate that the binding protein is unlikely to be required for the oxidation of all-trans-retinol to all-trans-retinal. We were led to examine the visual cycle of these animals in more detail because of the presence of CRBPI in the RPE and Müller cells of the retina and because of the reported importance of LRAT in supplying substrate for the isomerase reaction. ³² Our studies revealed perturbations in the distributions of visual-cycle retinoids in dark-adapted mice and in mice recovering from a flash that bleached approximately 35% of their visual pigment. The overall effect on the visual cycle of these changes in retinoid composition was relatively mild. However, the perturbations provide insight into the mechanism driving diffusion of retinol from photoreceptor cells into RPE, further define the flux of the visual cycle during recovery from a flash and demonstrate an important role of CRBPI in delivering all-trans-retinol for esterification. 

Diminished Amounts of Retinyl Esters

Studies in vitro demonstrated that the physiologic substrate of LRAT is likely to be a complex of all-trans-retinol and CRBPI. ^{16 17} In addition, other studies with cultured cells have provided evidence that CRBPI delivers all-trans-retinol to LRAT for esterification. ^{20 21} In contrast, the physiologic substrate for acylCoA-retinol acyltransferase (ARAT) is likely to be all-trans-retinol dissolved in membranes. ¹⁶ The reduced amounts of retinyl esters in liver ²² from CRBPI ^−/− mice are consistent with these findings. In this study, we found that the amounts of retinyl esters were also reduced in dark-adapted mouse eyes to approximately 33% of the wt amounts (Fig. 3) . In addition, the transient increase in retinyl ester during recovery from a flash, observed in the wt mouse, was reduced in CRBP1 ^−/− mice. The retinyl ester that we observed in CRBPI ^−/− mice could result from ARAT activity, although previous studies have suggested that ARAT in bovine RPE is low. ³³ It is also possible that LRAT has low activity in the absence of CRBPI, because the enzyme is known to process free retinol. ^{15 23 24}  

Impaired Retinol Diffusion

Although the mechanisms of vitamin A uptake are not known, several studies have suggested that CRBPI and LRAT are involved. For example, overexpression of CRBPI results in increased uptake and esterification of all-trans-retinol in cells that express LRAT ²¹ but not in those that express ARAT. Similar findings have been reported for Caco-2 cells (a cell line derived from small intestinal epithelium), which take up and esterify more retinol after transfection with CRBPI. ²⁰  

The mechanism of retinol translocation from photoreceptor to RPE cells is not understood. IRBP was initially postulated to mediate the diffusion of retinoids, either through a passive mechanism ³⁴ or an active mechanism possibly involving a receptor for IRBP. ^{35 36} However, studies of the visual cycle in IRBP ^−/− mice demonstrated that the rate of visual pigment regeneration and the distribution of visual-cycle retinoids were normal during recovery from a flash. ⁶ Because the rod photoreceptor cells degenerate in the absence of IRBP, ³⁷ perhaps a major role of the protein is in protecting the cells from high concentrations of free retinoids. 

During recovery from a flash, all-trans-retinol transiently accumulated in the neural retina of CRBPI ^−/− but not wt mice (Fig. 4) . Thus, a concentration gradient by itself cannot establish the normal rate of diffusion of all-trans-retinol, and it appears that metabolic processing of the vitamin to retinyl ester in RPE plays an important role. Perhaps the driving force for the translocation is provided by the combination of high-affinity binding of all-trans-retinol with CRBPI (K _d, 0.1 nM) ³⁸ and delivery of the vitamin to LRAT in a protein-bound form. However, it is important to note that the absence of CRBPI results in slower diffusion of retinol out of the neural retina, not an absolute block in retinol diffusion (Fig. 3) , indicating that this process occurs in the absence of CRBPI, albeit more slowly. 

The phenotype of LRAT-null mice has not been reported. However, based on the information presented herein, it can be predicted that all-trans-retinol would also accumulate in the neural retinas of LRAT-null mice, perhaps to a greater extent than observed in the current study. In addition, the phenotype of this animal may reveal the identity of the substrate for the isomerization reaction. 

Visual Pigment Regeneration

It was initially surprising that a reduced rate of retinol diffusion did not result in a major delay in regeneration of 11-cis-retinal/rhodopsin (Figs. 3 5) . However, the additional slow step occurs upstream of a slow step in the normal visual-cycle isomerization of all-trans-retinyl ester to 11-cis-retinol in the RPE (Fig. 1) . ⁵ Thus, the rate of generation of the final product (11-cis-retinal/rhodopsin) would be only minimally affected, unless the new slow step were slower than the existent downstream slow step. 

Dark Adaptation

Dark adaptation requires regeneration of the visual pigment that was lost during bleaching (regain of quantum catch) and the quenching of photointermediates that actively desensitize the visual system. ^{39 40} These active intermediates are likely to include opsin and complexes of opsin with all-trans-retinal. ³⁹ Dark adaptation was delayed in CRBPI ^−/− mice (Fig. 6) and the rate of reduction of all-trans-retinal was modestly slower (Fig. 3) , perhaps because of a mass action effect of the accumulation of all-trans-retinol. The delayed dark adaptation could result from slower decay of all-trans-retinal, because a small amount of photointermediate profoundly affects the sensitivity of the visual system. ^{41 42} However, the biochemical and ERG experiments used different bleaching and anesthetization conditions, and so they are not directly comparable. 

In summary, the results demonstrate that CRBPI plays an important role in generating the driving force for diffusion of all-trans-retinol from the neural retina to the RPE. This is most likely accomplished through its ability to deliver all-trans-retinol to LRAT, the main retinol esterification enzyme of RPE. 

 Supported in part by National Eye Institute Grants EY01730, EY02317, and EY06641, by the Howard Hughes Medical Research Foundation, and by an unrestricted award from Research to Prevent Blindness to the University of Washington. JCS is a Senior Scientific Investigator of Research to Prevent Blindness.

 Submitted for publication November 8, 2001; revised January 11, 2002; accepted January 17, 2002.

 Commercial relationships policy: N.

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Figure 1.

 

A simplified version of the vertebrate rod visual cycle. Reactions known to occur in RPE are shown above the pair of horizontal lines, and those known to occur in the rod photoreceptor cell are shown below. The extracellular compartment separating these two cells must be traversed by retinoids during regeneration of visual pigments. The postulated roles for CRBPI in binding all-trans-retinol and facilitating its esterification by LRAT are derived from published results and those in this study. ISOM, isomerase; 11-RDH, 11-cis-retinol dehydrogenase; prRDH, photoreceptor all-trans-retinol dehydrogenase.

Figure 2.

 

HPLC traces obtained during recovery from a flash. Dark-adapted mice were subjected to a flash and kept in the dark for 15 minutes. Animals were then killed, the eyes dissected, and retinoids extracted and analyzed by HPLC. Note the reduced amount of retinyl ester and the larger amount of all-trans-retinol in the extract from CRBPI ^−/− mice, relative to the wt control. Component 1, retinyl esters; 2, syn 11-cis-retinal oxime; 3, syn all-trans-retinal oxime; 4, anti 11-cis-retinal oxime; 5, all-trans-retinol; 6, anti-all-trans-retinal oxime; IS, internal standard.

Figure 3.

 

Composition of visual-cycle retinoids during recovery from a flash. Retinoids were extracted from the eyes of mice before a flash (dark) and at the times indicated during recovery from the flash in the dark. The bars depict the mean ± SD (n > 5). The time course shown above all-trans-retinal in the top panel applies to all retinoids in both panels.

Figure 4.

 

Accumulation of all-trans-retinol in neural retina. Eyes were obtained from mice before (dark adapted) and 30 minutes after a flash (30 minutes dark) and dissected into neural retina (top) and RPE-choroid fractions (bottom). Retinoids were extracted and analyzed by HPLC. The all-trans-retinol accumulated in neural retina and not in the RPE-choroid of CRBPI ^−/− mice. The mean ± SD (n = 4) is shown.

Figure 5.

 

Rhodopsin regeneration. Eyes of dark-adapted mice were dissected before a flash (values to the left of zero hour) and at the times indicated on the abscissa after the flash. The mean ± SD (n ≥ 3) is shown.

Figure 6.

 

Time course of dark adaptation. The figure illustrates the time course for recovery of the ERG a-wave amplitude at in wt and CRBPI ^−/− mice after a flash that bleached approximately 99% of their visual pigment. a-Wave amplitudes 7 ms after the stimulus flash are shown. The mean ± SD is shown (n ≥ 3). The smooth curves are fits to the data with an equation that describes the decay of an equivalent background light. ⁴³ Dark adaptation was delayed in CRBPI ^−/− mice relative to wt mice. Because of differences in experimental protocol, the time courses cannot be directly compared with those in other experiments.