Coordination between Production and Turnover of Interphotoreceptor Retinoid-Binding Protein in Zebrafish

Abstract

purpose. Interphotoreceptor retinoid-binding protein (IRBP), which is secreted by the photoreceptors of most vertebrates, is the major soluble protein component of the interphotoreceptor matrix (IPM). Recent studies suggest that IRBP is short lived in the IPM (half-life, ∼11 hours). The mechanisms coordinating the production and removal of IRBP are not known. Zebrafish provide a useful system to study the regulation of these two processes, because its IRBP mRNA levels are under circadian regulation. In the present study, the relationship between the quantity of IRBP, the rate of its turnover, and the expression of its mRNA in the zebrafish retina were examined.

methods. Full-length zebrafish IRBP was expressed in Escherichia coli and an antiserum generated against purified recombinant IRBP. Western and protein dot blot analyses and indirect immunofluorescence were used to define the temporal and spatial patterns of IRBP expression in the adult zebrafish. In vivo and in vitro metabolic labeling experiments were used to examine the regulation of IRBP turnover by both environmental light and the light–dark cycle.

results. Despite the known rhythmicity in IRBP mRNA expression, neither the amount of IRBP nor its localization changes significantly during the light–dark cycle. IRBP is rapidly removed from the zebrafish eye (half life, ∼7 hours). This rapid turnover is independent of environmental lighting conditions during subjective day and is more rapid during the day than at night.

conclusions. Because the amount of IRBP remains constant throughout the day, the enhanced daytime IRBP mRNA expression may function to compensate for an increased turnover of the protein during the day. These findings suggest that the processes of IRBP production and removal are coordinately regulated.

The interphotoreceptor matrix (IPM) mediates critical interactions between the neural retina and the retinal pigment epithelium (RPE). ^{1 2} This complex matrix is composed of interphotoreceptor retinoid-binding protein (IRBP), S-laminin, ³ growth factors, ^{4 5} specific domains enriched in lectin-binding glycoconjugates, ^{4 5 6} metalloproteases, ⁷ and hyaluronan. ⁸ In the normal retina the concentrations of individual IPM components appear to be carefully regulated, and abnormal expression of matrix components has been associated with disease. ^{9 10 11 12} Little is known about the mechanisms regulating the amount of any IPM constituent. The concentration of any molecule in the subretinal space depends on the rates of both its introduction into and its removal from the IPM. The rates of both production and removal may be linked to the photoreceptor circadian oscillator, ¹³ which regulates not only the expression of many retinal genes but also photoreceptor outer segment disc shedding, ¹⁴ a potential mechanism for IRBP removal. 

As a first step toward understanding the processes that regulate the concentrations of the molecular components of the IPM, we examined the expression and turnover of IRBP, the major soluble protein component of the matrix. IRBP is a glycolipoprotein that has long been thought to mediate the transfer of retinoids between the photoreceptors and the RPE during the visual cycle. ^{14 15 16} Targeted disruption of the IRBP gene in mice shows that although IRBP is necessary for photoreceptor survival, it is not essential for a normal rate of visual pigment regeneration. ¹⁷ Low levels of IRBP have been reported for some disease states. ^{12 18 19 20 21} However, the mechanism responsible for abnormal IRBP expression has not been established for any of these conditions. 

In most vertebrates IRBP is secreted by both rod and cone photoreceptors. ^{22 23} In the subretinal space IRBP is short lived and is rapidly cleared from the IPM. ²⁴ The mechanism of this turnover is not known. Our long-term goal is to understand how the process of IRBP production and the mechanism of its removal from the matrix are coordinated. 

The zebrafish offers a potentially useful system for studying the relationship between the production of IRBP and the processes responsible for its removal from the IPM. In this animal, IRBP is produced primarily by the cones ²⁵ and to a lesser extent by the RPE. ²⁶ We have shown that zebrafish IRBP mRNA expression by non-UV cones is circadian, with higher levels of expression occurring during the day than at night. ²⁵ In contrast, UV cones express IRBP mRNA at similar levels throughout the day. ²⁵ In this study, we asked whether the amount of IRBP simply mirrors the expression of its mRNA or whether IRBP and its mRNA are differentially expressed. The former suggests that a higher concentration of IRBP in the IPM is required during the day than at night. The latter suggests a coordination between IRBP production and degradation. 

Methods

Animals

All experiments were approved by the University of Virginia Animal Care and Use Committee and were conducted in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. Adult zebrafish (Danio rerio) were maintained at 28.5°C on a 12-hr–12-hr light–dark cycle for at least 2 weeks before use. Overhead fluorescent lamps provided a maximum irradiance in the aquaria of approximately 10.7 μW/cm². Zebrafish were maintained according to Westerfield ²⁷ and killed by decapitation with brain pithing. 

Production of Recombinant Zebrafish IRBP and Antiserum

The small size of the zebrafish eye precludes extraction and purification of sufficient quantities of native IRBP for biochemical characterization and antiserum production. Therefore, we expressed in Escherichia coli full-length zebrafish IRBP as a thioredoxin fusion protein. Thioredoxin promotes soluble expression of many recombinant proteins in E. coli, ^{28 29} including IRBP. ³⁰ To generate the thioredoxin–zebrafish IRBP fusion protein, we used the expression vector pThioHis (Invitrogen, San Diego, CA). This system incorporates a histidine patch on the surface of the thioredoxin (“His-patch” thioredoxin), ³¹ allowing the recombinant protein to be purified by metal-chelate affinity chromatography. pThioHis uses the trc promoter, and protein expression is induced by isopropyl-β-d-thiogalactopyranoside. We had previously isolated and sequenced an intron-containing full-length cDNA for zebrafish IRBP. ²⁵ An intronless cDNA was amplified from total ocular RNA using sense primer acaaggtacccggggatcctttctctcccacacttattgc and the antisense primer cttaaggtcgactatagagaggctatatcatttggcttccgt (the segment of each primer corresponding to the pThioHis multiple cloning region is in italics; the nucleotides corresponding to the N-terminal residue and the stop codon are underlined). The cDNA was excised from pCRII with EcoRI and ligated into the BamHI-XbaI site of pTrxFus (Invitrogen). The reading frame was confirmed by DNA sequencing and the construct used to transform E. coli (ToplO; Invitrogen). 

The temperature and duration of protein expression were optimized in pilot cultures to maximize yield of soluble recombinant protein. Fermentations were performed in a 7-liter reactor (Applikon, Foster City, CA). After the cells reached optical density (OD)₅₅₀ = 0.5, the temperature was lowered to 30°C before protein expression was induced. Cells were grown for 5 more hours, harvested by centrifugation, resuspended in 50 mM Tris (pH 7.4) and 100 mM NaCl, and ruptured with a French pressure cell. The zebrafish IRBP-thioredoxin fusion protein was purified from the soluble fraction by arsenic-based affinity chromatography. ^{32 33} Approximately 9 mg (114 nanomoles) of purified recombinant zebrafish IRBP was obtained per liter of E. coli. Rabbits received one intradermal and one subcutaneous injection of 125 μg purified recombinant zebrafish IRBP-thioredoxin fusion protein suspended in 0.5 ml Freund’s complete adjuvant. Resultant antiserum was characterized by enzyme-linked immunosorbent assay (ELISA) and found to have no reactivity to thioredoxin (data not illustrated). 

Western and Dot Blot Assays

Whole zebrafish eyes were frozen in liquid N₂ and homogenized in phosphate-buffered saline (PBS) containing 1 mM phenylmethylsulfonyl fluoride (PMSF), 1 mM EDTA, 1.5 μg/ml leupeptin, 1.5 μg/ml pepstatin A, and 3.0 μg/ml aprotinin. Insoluble material was removed by centrifugation at 10,000g for 30 minutes at 4°C. Proteins were precipitated from the supernatant in 6% trichloracetic acid (TCA) and 0.1% deoxycholic acid, washed with acetone followed by ethanol, and resuspended in Laemmli buffer containing dithiothreitol (DTT). ³⁴ Protein equivalent to three eyes was fractionated by sodium dodecyl sulfate (SDS)-10% polyacrylamide gel electrophoresis (PAGE) and transferred to nitrocellulose membrane (NitroBind; Micron Separations, Westboro, MA). The blot was washed in rinsing solution (150 mM NaCl, 27 mM KCl, 25 mM Tris [pH 8.0], 0.005% nonylphenyl-polyethylene glycol, and 0.001% Tween 20). Blocking was performed for 2 hours in rinsing solution containing 0.5% bovine serum albumin (BSA; Sigma, St. Louis, MO), 0.03% nonfat dry milk (Bio-Rad, Hercules, CA). This was followed by overnight incubation in anti-zebrafish IRBP serum (or preimmune serum) diluted 1:1000 in blocking solution. The blot was washed in rinsing solution and incubated for 2 hours with [¹²⁵I]-goat anti-rabbit IgG (ICN, Costa Mesa, CA) in blocking solution. 

For dot blots, two eyes per tube were homogenized as for Western blot analysis. Nitrocellulose membrane was soaked in Tris-buffered saline (TBS: 150 mM NaCl, 27 mM KCl, 25 mM Tris [pH 8.0]) for 10 minutes and placed in a dot blot manifold (Bio-Rad). Proteins corresponding to one third of one eye were applied to each well and allowed to filter by gravity. Wells were washed with TBS containing 0.05% Tween 20. The nitrocellulose membrane was then dried and processed as described earlier. Blots were analyzed using a PhosphorImager (Molecular Dynamics, Sunnyvale, CA). 

Quantitative RT-PCR

IRBP mRNA levels were compared at midlight and middark using real-time reverse transcription–polymerase chain reaction (RT-PCR). This technique uses a sequence-specific, fluorescently tagged oligonucleotide probe to detect PCR products as they accumulate. The probe contains both a reporter fluorescent dye and a quencher dye. When the target sequence is present, the probe anneals to the PCR product, and the quencher is cleaved by the 5′ nuclease activity of Taq DNA polymerase. This cleavage greatly increases the fluorescent signal obtained from the probe. The probe is then cleaved from the target sequence, allowing continuation of primer extension. The thermal cycler (ABI Prism 7700; Perkin-Elmer Applied Biosystems, Foster City, CA) is equipped with a laser and CCD camera to measure fluorescence after each amplification cycle. Reactions are characterized by the number of PCR cycles required for the fluorescent signal to cross a threshold level above baseline. 

Total ocular RNA equivalent to 5% of one zebrafish eye was extracted from aliquots of the homogenized whole zebrafish eyes collected at midlight and middark used in the dot blot assays. RNA was reverse transcribed using random hexamer primers. PCR primers were 5′-CAGCAACATCCCTGCACTTC-3′ (forward primer) and 5′-ACTTTTGATGAGCGCGATGA-3′ (reverse primer). The sequence-specific IRBP fluorescent probe sequence was 5′-6 FAM-CCAATGAACCCCACACCCGAGAATGT-TAMRA-3′, where FAM represents the fluorescent dye and TAMRA the fluorescent quencher. Cycling conditions were as follows: (RT step) 25°C for 10 minutes, 48^oC for 30 minutes, 95^oC for 5 minutes; (PCR step) 50^oC for 2 minutes, 95^oC for 10 minutes, and 40 cycles of 95^oC for 15 seconds, 60^oC for 1 minute. Fluorescence intensity was measured after each cycle using the sequence detection system (ABI Prism 7700), and data were analyzed with the accompanying software (ABI Prism; Perkin–Elmer Applied Biosystems). 

Metabolic Studies

Both in vivo and in vitro approaches were used for IRBP labeling studies. For in vivo labeling, adult zebrafish were anesthetized by brief (∼15 seconds) immersion in ice-cold water. Zebrafish received a single systemic injection of[ ³⁵S]methionine (Amersham Life Sciences, Arlington Heights, IL). The injection was performed using a 10 μl syringe (Hamilton, Reno, NV) fused to a 5-cm beveled 30-gauge needle. The needle was inserted into the ventral midline just caudal to the pectoral fins; it was then passed under the skin caudally for a few millimeters, and 5 μl (250 μCi) of[ ³⁵S]methionine was injected. Fish were out of water for 15 to 25 seconds during the injection procedure. The postinjection survival rate was more than 95%. After injection, fish were maintained in cyclic light for 4 hours to 2 weeks. Whole eyes (10 per tube) were frozen in liquid N₂. Eyes were homogenized in 1 ml of NET buffer (50 mM Tris-HCl [pH 7.5], 150 mM NaCl, 1 mM EDTA, 0.1% nonylphenyl-polyethylene glycol, 0.25% gelatin, 0.02% sodium azide) containing protease inhibitors (above), and insoluble material was removed by centrifugation. Before immunoprecipitation, samples were incubated with preimmune serum followed by a 50% suspension of protein A-Sepharose CL-4B beads in 20 mM sodium phosphate [pH 8.0], 150 mM NaCl for 30 minutes at 4°C. Beads were collected by centrifugation and discarded. The supernatant was incubated with 25 μl anti-IRBP serum at 4°C for 1 hour followed by protein A-Sepharose for 1 hour. Beads were washed three times in NET buffer and resuspended in Laemmli buffer containing DTT. 

Two sets of in vitro experiments were performed. In the first, the rate of IRBP turnover was compared in light versus dark during the day (light–dark in vitro experiments). In the second study, IRBP turnover was compared during the day versus at night (circadian in vitro experiments). For in vitro light–dark labeling studies, eyes were enucleated 3 hours after light onset. For circadian studies, eyes were enucleated just after light onset and just before light offset. To facilitate diffusion of medium into the posterior chamber, the lens was removed through a corneal incision. The retina–RPE eyecups were incubated in Dulbecco’s modified Eagle’s medium (Gibco, Gaithersburg, MD) without methionine supplemented with 0.015% vol/vol HEPES and equilibrated with 95% O2/5%CO2.[ ³⁵S]methionine (0.5 mCi/ml) was added, and eyes incubated for 1 hour at 22°C with gentle agitation. The chase consisted of four 15-minute washes with medium containing a fivefold excess of unlabeled methionine. After 1 hour in chase medium, tubes (six eyes per tube) were frozen in liquid N₂. For light–dark studies, the remaining eyes were incubated for another 10 hours in either light or dark before they were collected in groups of six and frozen. For circadian studies, remaining eyes were incubated for another 8 hours in light or dark without interruption of their regular light–dark cycle. For both studies IRBP was immunoprecipitated from the undetached retina–RPE eyecups, as described. 

Indirect Immunofluorescence

Adult zebrafish eyes were collected at midlight and middark. To avoid disturbing the animals during collection, groups of animals were separated into large beakers approximately 10 hours before collection. Dark-adapted specimens were collected under dim red light. Corneas were slit to facilitate the diffusion of fixative into the globe. Eyes were fixed in 4% paraformaldehyde in 0.10 M phosphate buffer (pH 7.4) for 12 hours at 4°C. The tissues were dehydrated through graded ethanols and embedded in paraffin. Three-micrometer sections on polylysine-coated slides were cleared, rehydrated, and incubated in blocking solution (0.5% BSA; 2% calf serum; 0.08% Triton X-100 in PBS) before overnight incubation at 4°C in anti-zebrafish IRBP serum diluted 1:1000 in blocking solution. Sections were then incubated for 2 hours in Oregon Green-conjugated goat anti-rabbit IgG (Molecular Probes, Eugene, OR) diluted 1:200 in blocking solution. Sections were protected from light during this and all preceding steps. 4′,6-diamidino-2-phenylindole (DAPI, Sigma) was used as a nuclear counterstain. Sections were mounted in fluorescent mounting medium (Fluoromount-G; Southern Biotechnology, Birmingham, AL). Fluorescence microscopy was performed on an Axioplan 2 microscope (Zeiss, Thornwood, NY) equipped with a CCD camera (Microview; Princeton Instruments, Trenton, NJ) and imaging software (MetaMorph; Universal Imaging, West Chester, PA). 

Results

Immunofluorescent Localization of IRBP in the Zebrafish Retina

The distribution of IRBP in the retina at midlight and middark was studied by indirect immunofluorescence (Fig. 1) . Representative hematoxylin and eosin–stained sections, which are shown alongside the immunofluorescence, correlate the anatomic changes due to retinomotor movements to the different fluorescence patterns at midlight and middark. ^{35 36 37} During light adaptation cone photoreceptors contracted inward and rod photoreceptors elongated, whereas RPE pigment granules migrated into apical processes. In darkness the positions were reversed: pigment granules migrated sclerad into the RPE cell body, whereas rods contracted and cones elongated. Thus, in a light-adapted retina, cones were positioned for light capture, whereas RPE pigment granules shielded rods from illumination. In darkness, rods were positioned for maximum light capture with cones located behind them. 

At both midlight and middark IRBP was restricted to the region between the neural retina and the RPE. At midlight (Fig. 1C) the central portion of this region was less fluorescent than that at midnight. Intense fluorescence surrounded cones (Fig. 1C , inset). At middark (Fig. 1F) the fluorescence intensity was more uniform throughout the region between the RPE and the neural retina. IRBP appeared to outline the outer segments of both photoreceptor types, although rod outer segments were obscured by RPE pigment granules at midlight. At middark (Fig. 1F) , IRBP immunofluorescence outlined both rod and cone outer segments. 

Western and Dot Blot Analyses

The quantity of IRBP throughout the light–dark cycle was monitored by Western and dot blot analyses. As in our previous study, ²⁵ to minimize errors associated with tissue dissections, we extracted total soluble proteins from intact globes rather than from IPM obtained from detached retinas. The Western blot analysis is shown in Figure 2 . Eyes were collected in pairs at 3-hour intervals throughout the light–dark cycle. At each time point a single immunoreactive band of relative molecular mass (M_r) of 76.6 was identified, consistent with the size of zebrafish and other teleost IRBPs. ²⁵ Phosphorimaging did not show a significant difference in the amount of IRBP at any time point in the light–dark cycle. 

Having established the specificity of the antiserum, we further analyzed IRBP levels by dot blot analysis, which allowed the sample size to be efficiently increased. Figure 3 shows the relationship between loading level and signal intensity. Binding specificity was further established by comparison to BSA incubated with immune serum and to total soluble ocular proteins incubated with either preimmune serum or with immune serum preadsorbed with recombinant IRBP (Fig. 3B) . For the experimental samples shown in Figure 4a loading level of 0.3 eyes/dot was selected. This loading level fell within the dynamic range of the assay while providing adequate signal intensity. 

Figure 4A shows the mean IRBP dot densities at 3-hour intervals throughout the light–dark cycle. IRBP levels did not change significantly during the 24-hour period. The constant level of IRBP in the whole eye was in contrast to the known circadian expression of IRBP mRNA (also in the whole eye; overlay in Fig. 4A ). ²⁵ To show that IRBP mRNA levels were rhythmic in the same animals used for the dot blot assay, we used quantitative real-time RT-PCR to compare the level of IRBP mRNA expression at midlight with that at middark. The real-time RT-PCR approach allows for measurement of the amplified product after each PCR cycle using a fluorescent probe. The reaction is characterized by the number of cycles required for the fluorescent signal to rise above the noise (i.e., reach a threshold value). The data in Figure 4B show that the fluorescence intensity for IRBP mRNA increased more rapidly for eyes collected at midlight (cycles to reach threshold: 18.06 ± 0.36) than at middark (cycles to reach threshold: 19.98 ± 0.32). The shift of 1.92 cycles between midlight and middark tubes was compared with a standard curve (Fig. 4B , inset) generated by amplifying IRBP mRNA from varying amounts of total ocular RNA. This comparison yielded an estimate of the increase in IRBP mRNA of 4.94-fold at midlight versus middark. This estimate is consistent with the previously reported rhythmicity of IRBP mRNA expression. ²⁵  

Metabolic Labeling Studies

The turnover of IRBP was characterized by in vivo and in vitro metabolic labeling experiments. Both approaches used immunoprecipitation to isolate the radiolabeled IRBP from total ocular-soluble proteins. The specificity of the immunoprecipitation assay is shown in Figure 5 . Anti-zebrafish IRBP serum was incubated with the total soluble ocular proteins in the absence (lanes 2 and 4) and presence (lanes 3 and 5) of an excess of purified recombinant zebrafish IRBP-thioredoxin fusion protein. In the Coomassie blue–stained gel (left) the major band at M_r of approximately 55 represents the immunoglobulins. The band at M_r of approximately 82 in lane 3 is the added recombinant IRBP-thioredoxin fusion protein. The fluorogram (right) shows the radiolabeled immunoprecipitated proteins. The radiolabeled protein band of M_r 76.6, which is the most prominent band in lane 4, has an electrophoretic mobility identical with that of zebrafish IRBP identified by Western blot in Figure 2 . When the antibody was preadsorbed with the recombinant IRBP the M_r of 76.6 band (and fainter higher and lower bands) disappeared. This triplet (M_rs = 82.4, 76.6, and 73.5) is identified by the bracket to the right of the fluorogram. Faint bands at M_rs of approximately 200 and 43 in lane 4 are prominent in the preadsorbed control (lane 5). These bands and another at M_r of approximately 65 may represent proteins that bind to the recombinant zebrafish IRBP. Whether these proteins normally interact with IRBP in the IPM is under study in our laboratory. 

In Figure 6 zebrafish were killed 4 hours to 14 days after a single systemic injection of [³⁵S]methionine. The fluorograms in Figure 6A represent the emergence and disappearance of [³⁵S]IRBP and[ ³⁵S]opsin. Note that the immunoprecipitated[ ³⁵S]IRBP bands represent a triplet similar to that seen in Figure 5 . [³⁵S]IRBP accumulated in the eye for 12 to 24 hours after injection before declining. By 14 days[ ³⁵S]IRBP levels had decreased to approximately that seen at 4 hours. [³⁵S]opsin, which migrates on SDS-PAGE as a group of bands with M_r of 36 to 39, ^{38 39} accumulated in the zebrafish eye at approximately the same rate as [³⁵S]IRBP. Unlike [³⁵S]IRBP, the amount of[ ³⁵S]opsin remained constant for approximately 12 days before declining by 14 days after injection. 

To better achieve a pulse label of photoreceptor proteins, we exposed isolated retina-RPE eyecups to a pulse of[ ³⁵S]methionine followed by a chase containing an excess of unlabeled methionine. The incorporation of[ ³⁵S]methionine into IRBP and transducin were measured immediately after the chase and after 10 hours in either constant light or constant dark. The specific activity of transducin did not significantly change during the 10-hour chase incubation in either light or dark (Fig. 7) . In contrast, IRBP-specific activity decreased fourfold in both light and dark. 

To compare the rate of IRBP turnover during the day with that during the night, we repeated the [³⁵S]methionine pulse–chase paradigm using retina-RPE eyecups collected just after light onset with those collected just before light offset. The incubations were terminated 1 hour and 8 hours after the pulse–chase. IRBP was immunoprecipitated from the total ocular soluble proteins. Results are shown in Figure 8 . The specific activity of IRBP was lower in the 1-hour nighttime group than in the 1 hour daytime group. This is consistent with the fact that the level of IRBP mRNA expression in the morning is higher than that at night. At 8 hours, the level of [³⁵S]IRBP decreased approximately 60% for the daytime group. For the nighttime group, there was not a significant change in the amount of[ ³⁵S]IRBP between the 1- and 8-hour time points. 

Discussion

IRBP appeared to distribute itself throughout the IPM. Although IRBP mRNA expression is largely cone specific, ²⁵ indirect immunofluorescence showed that the protein was not restricted to the cone matrix sheath but was associated with the IPM of rods as well as cones. Although the immunofluorescence showed IRBP associated with the matrix immediately surrounding rod and cone outer segments, the finding that IRBP turnover occurs at a faster rate than that of either opsin or transducin suggests that IRBP may not be tightly bound to outer segments. It also suggests that the mechanism of IRBP turnover likely differs from that of opsin and transducin, each of which are turned over through outer segment disc shedding. The possibility that IRBP can move within the matrix was also suggested by others who observed changes in the spatial distribution of IRBP in the light compared with the dark. In the present study, the pattern of IRBP immunofluorescence was different at midlight compared with middark but probably for a different reason. We found that the outer segment region showed less immunofluorescence at midlight than at middark. Although this pattern is reminiscent of that for albino rats, ⁴⁰ in the present study, the reduced immunofluorescence in the central region of the subretinal compartment at midlight was probably due to quenching of the fluorescent signal by RPE pigment granules. 

Quantitative RT-PCR experiments indicated that IRBP mRNA levels were approximately five times higher at midlight than at middark. This is consistent with our previous mRNA dot blot study that showed a four- to sevenfold increase in the amount of IRBP mRNA at midlight compared with middark. ²⁵ In situ hybridization studies have shown that nighttime IRBP mRNA expression is restricted to the UV-sensitive cone photoreceptors. In contrast, all cone subtypes express IRBP mRNA during the day. ²⁵  

Taken together, our Western blot and immunofluorescence studies showed that the quantity of IRBP in the zebrafish eye remained constant throughout the light–dark cycle. This is in contrast to the circadian expression of its mRNA. One explanation for the differential expression of IRBP and its mRNA is that the rate of removal of IRBP is an important regulator of its concentration in the matrix. To explore this issue we used metabolic labeling to examine the turnover of IRBP. Because the small size of the zebrafish eye precludes the use of intraocular injections for protein labeling, we used in vivo systemic injections and in vitro pulse–chase paradigms. 

The in vivo systemic injections significantly underestimated the IRBP turnover rate compared with the in vitro pulse–chase studies. The main issue is that [³⁵S] released from the breakdown of labeled proteins was partly reincorporated into newly synthesized IRBP. Because of this tracer recycling, the estimate of an IRBP half-life of 6 days (based on the data from the[ ³⁵S]methionine systemic injections) is an overestimate of the true half-life. Nevertheless, these in vivo experiments show that the turnover of IRBP was significantly faster than that of opsin. Opsin turnover occurs by RPE internalization and degradation during the circadian process of photoreceptor outer segment disc shedding. ¹⁴ Removal of[ ³⁵S]opsin occurred when the radiolabeled discs reached the distal tip of the outer segment and were internalized by the RPE. Because of the time involved in the displacement of the labeled band to the RPE we did not see a reduction in radiolabeled opsin until after 12 days. In contrast, the amount of radiolabeled IRBP began to decrease only 2 days after the injection. 

The retina-RPE eyecup pulse–chase experiments showed that the rate of IRBP turnover was faster than estimated by the in vivo study. In the eyecup preparation the biological half-life of IRBP during the day (in either light or dark) was approximately 7 hours. This estimate is consistent with recent studies in Xenopus laevis in which intraocular injections of carboxyl-terminal labeled leucine ([1-¹⁴C]leucine) were used to estimate the IRBP turnover rate. ²⁴ The early decarboxylation of leucine during its degradation liberates the radiolabel as[ ¹⁴C]CO₂, which is not significantly recycled into new protein. In the present in vitro study, recycling of the radiolabel was not an issue, because in the eyecup preparation, the pulse could be followed by a chase. Thus, the expense of [1-¹⁴C]leucine is avoided. Our data suggest that during the day, IRBP is rapidly removed from the subretinal space by a process that is not dependent on environmental lighting conditions. 

IRBP turnover could not be detected at night, offering a possible explanation for the differential expression of IRBP and its mRNA. It is plausible that the rate of IRBP production is matched to the rate of its removal from the IPM. According to this working model, gene expression and protein turnover in the zebrafish retina are coordinated to achieve a constant concentration of IRBP in the IPM throughout the day. Higher daytime IRBP mRNA levels may compensate for an increased daytime degradation of IRBP. At night, IRBP mRNA levels may decrease to match the reduced rate of IRBP degradation. It should be pointed out that this model relies on the assumption that the level of IRBP mRNA is directly related to IRBP synthesis. Although our data do not rule out a contribution of regulatory mechanisms controlling ribosome access to the translation start codon, ⁴¹ the finding that IRBP production is lower at night (when the mRNA level is lowest) than during the day (Fig. 8) suggests that IRBP mRNA expression directly correlates with the rate of IRBP synthesis. 

The notion that the mechanism(s) responsible for the removal of IRBP from the subretinal space plays an important role in regulating the extracellular IRBP concentration is consistent with the work of others. Light-deprived mice show a marked reduction in IRBP mRNA without a decrease in the amount of IRBP. ⁹ The maintenance of IRBP levels despite reduced IRBP mRNA expression could be explained by a decreased rate of IRBP removal during light deprivation. Others have noted that retinol deprivation or replenishment depresses or enhances IRBP expression, respectively, without altering the level of its mRNA or the rate of its gene transcription. ⁴² The accumulation of IRBP in the vitiligo mouse despite normal IRBP mRNA levels also suggests that if we are to understand what controls the concentrations of matrix components, we must consider not only the mechanisms regulating IPM production but also those responsible for removal of specific matrix components from the subretinal compartment. 

Because its Stokes’ radius is greater than the photoreceptor Müller cell zonulae adherens exclusion limit, IRBP cannot simply leave the IPM by diffusion but must be cleared from the subretinal compartment by endocytosis and/or proteolysis. ⁴³ Although certain IPM components may be removed enzymatically, ⁸ IRBP proteolytic fragments are not observed in IPM extracts, and RPE-conditioned media do not degrade native IRBP. ⁴⁴ RPE phagocytosis of shed photoreceptor outer segment discs is another possible uptake pathway, and immunoelectron microscopy has detected IRBP within RPE phagosomes. ^{45 46} However, several lines of evidence suggest that disc shedding is not the primary mechanism of IRBP turnover. First, in our in vivo and in vitro metabolic labeling experiments, IRBP turned over faster than either opsin or transducin. Because disc shedding is the primary mechanism of both opsin and transducin turnover, the finding that IRBP was removed faster than either opsin or transducin suggests that disc shedding is not the primary mechanism of IRBP turnover. Second, in our in vitro experiments the pulse labeling was performed after most disc shedding would have occurred. Phagocytosis of rod discs occurs in all vertebrates shortly after lights on. ¹⁴ While the timing of zebrafish rod disc shedding has not been studied, zebrafish retinomotor activity shows rhymthmicity similar to that of other teleosts (Gregory M. Cahill, personal communication, August, 2000). If zebrafish rod disc shedding is also rhythmic, occurring as in other teleosts shortly after light onset, ^{47 48} then the observed clearance of IRBP could not occur through phagocytosis of rod discs since the animals were not used until 3 hours after light onset. Third, although our experiments measured IRBP removal from the whole eye, a similar conclusion has been reached in studies of IRBP turnover in Xenopus. ⁴⁶ In those studies in which the larger size of the Xenopus eye permitted separate analysis of IRBP in the retina, IPM, and RPE, the amount of IRBP in the RPE did not increase after disc shedding. ⁴⁶  

A potential mechanism for IRBP removal from the IPM is nonphagocytic endocytosis. Endosomes containing IRBP have been observed in both the RPE and photoreceptors. ^{45 46} Furthermore, there is evidence that photoreceptor IRBP uptake is receptor mediated and targeted to the lysosomal system. ⁴⁹ Taken together, the above data implicate nonphagocytic endocytosis as a likely mechanism for IRBP removal. Ongoing studies in our laboratory will define the mechanism(s) responsible for the coordination of IRBP production and its removal from the IPM. 

⁴  Present address: Virginia Merrill Bloedel Hearing Research Center, Department of Otolaryngology-HNS, University of Washington, Box 357923, Seattle, WA 98195-7923.

 Supported by National Institutes of Health Grant EY09412 (FG-F); the Thomas F. Jeffress and Kate Miller Memorial Trust (FG-F); a Wyeth–Ayerst Laboratories Scholarship through the Business and Professional Women’s Foundation (LLC); and an unrestricted grant from Research to Prevent Blindness to the Department of Ophthalmology at the University of Virginia Health Sciences Center.

 Submitted for publication February 22, 2000; revised May 15, 2000; accepted June 9, 2000.

 Commercial relationships policy: N.

 

Figure 1.

 

Immunolocalization of IRBP in the adult zebrafish retina at midlight (A, B, and C) and middark (D, E, and F). Hematoxylin and eosin–stained sections (A, D) and schematic diagrams (B, E) are shown to correlate immunofluorescence patterns (C, F) with retinomotor activity. (A, B) At midlight, cones were contracted, rods were elongated, and RPE processes were filled with pigment granules to the level of the cone inner segments. (C) Immunolocalization of IRBP at midlight. IRBP was restricted to the region between the outer limiting membrane and the apical RPE border. There was relatively less fluorescence in the middle of this region. Inset (from region indicated by ∗): Higher magnification shows IRBP immunofluorescence outlining cone outer segments (arrow). (D, E) At middark cones were elongated, rods were contracted, and RPE pigment granules collected in the RPE soma. (F) Immunolocalization of IRBP at middark. The distribution of IRBP was similar to that at midlight except that the fluorescence was more uniform throughout the IPM, and there was a zone of decreased fluorescence near the RPE. Inset (from region indicated by ∗): IRBP outlined both rod (arrow) and cone (arrowhead) outer segments at middark. Nuclei were labeled with DAPI. ROS, rod outer segments; PG, pigment granules; CN, cone nuclei; OLM, outer limiting membrane; RN, rod nuclei; OPL, outer plexiform layer; INL, inner nuclear layer; IPL, inner plexiform layer; GC, ganglion cell layer; SS, short single cone; LS, long single cone; DC, double cone; R, rod photoreceptor. Scale bar, (A, C, D, F) 50 μm; (C, F; insets) 20 μm.

Figure 2.

 

Western blot analysis of IRBP throughout the light–dark cycle. (A) Representative Western blot analyses are shown for each time point. The time (in hours relative to light onset) is shown at the top of each panel. The arrow indicates the IRBP band (M_r = 77). (B) The average IRBP band density is shown for each time point. Error bars represent ± SEM for four pairs of eyes per time point.

Figure 3.

 

Dot blot assay for quantification of IRBP. Soluble ocular proteins equivalent to 0.3 eyes were immobilized on nitrocellulose and treated with anti-zebrafish IRBP serum followed by ^I25I-labeled goat anti-rabbit IgG. Signal intensity was quantified by phosphorimaging. (A) Phosphorimager dot signal intensity versus fraction of zebrafish eye. Each point represents the average of two measurements. Representative dots corresponding to the first five plotted data points are shown below the curve. (B) Control dot blots. Top: Dot loaded with 200 μg BSA; blot was incubated with immune serum (dot density, 40). Middle: Dot loaded with soluble ocular proteins equivalent to 0.3 eye; blot was incubated with preimmune serum (dot density, 51). Bottom: Dot loaded with soluble ocular proteins equivalent to 0.3 eye; blot was incubated with rabbit immune serum preadsorbed with recombinant zebrafish IRBP (dot density, 12).

Figure 4.

 

IRBP dot blot and real-time RT-PCR analysis. (A) IRBP levels throughout the light–dark cycle. Zebrafish eyes were homogenized in pairs, and total soluble ocular proteins equivalent to 0.3 eye were loaded per dot (error bars indicate ± SEM, n = 8 dots per time point). Representative dot signals are shown above each bar. Equivalent dots incubated with preimmune serum are shown at far right (Pre). IRBP mRNA levels in the whole eye (curve) are shown superimposed on the protein level for each time point (replotted from Rajendran et al. ²⁵ ). (B) Real-time RT-PCR analysis of IRBP mRNA levels at midlight and middark. The analysis was performed on aliquots of the same samples used in Figure 3A . Fluorescence values (mean ± SEM) for RNA collected at midlight (n = 5) and middark (n= 4) are shown for each PCR cycle after cycle 10. Fluorescence intensity increased more rapidly in tubes containing RNA collected at midlight than at middark. Inset: Standard curve obtained from loading increasing amounts of total RNA. Amount of RNA loaded is plotted against the number of PCR cycles required for the fluorescent signal to reach threshold value.

Figure 5.

 

Immunoprecipitation of IRBP from eyes of zebrafish injected systemically with [³⁵S]methionine.[ ³⁵S]IRBP was immunoprecipitated from the soluble fraction of homogenized whole eyes. Left: Coomassie blue–stained gel; right: its fluorogram. Lane 1: Molecular weight standards. Lanes 2 and 4: Total soluble ocular proteins incubated with anti-zebrafish IRBP serum. The darkest band in the IRBP triplet has electrophoretic mobility (M_r = 76.6) equal to IRBP identified by Western blot (Fig. 2) . Lanes 3 and 5: The immune serum was preadsorbed with recombinant zebrafish IRBP. The recombinant IRBP is the M_r = 82 band in lane 3 (the recombinant IRBP is larger than native IRBP because of the thioredoxin fusion tag). Preadsorption of immune serum with recombinant IRBP (lane 5) resulted in elimination of three bands with M_rs of 82.4, 76.6, and 73.5.

Figure 6.

 

In vivo metabolic labeling studies of IRBP and opsin turnover. Adult zebrafish were injected with [³⁵S]methionine and maintained in cyclic light for 4 hours to 14 days after injection. IRBP was immunoprecipitated using rabbit anti-zebrafish IRBP. Opsin was resolved from the insoluble fraction by SDS-10% PAGE. (A) Representative phosphorimaging bands of IRBP and opsin are shown for various time points after injection of[ ³⁵S]methionine. (B) Time course of turnover of [³⁵S]IRBP (•) and[ ³⁵S]opsin (○). Each point represents one to two experiments in which 10 eyes were pooled; opsin data are normalized to IRBP peak expression value. The maximum biological half-life of IRBP is 6 days.

Figure 7.

 

IRBP turnover in light versus dark during the day. Retina-RPE eyecups were subjected to a [³⁵S]methionine pulse–chase. A first group of eyecups was collected 1 hour after the chase. Two remaining groups of eyecups were incubated for an additional 10 hours in either light or dark. (A) Representative phosphorimaging bands of immunoprecipitated ³⁵S-IRBP and ³⁵S-transducin (as well as Coomassie blue–stained bands used to calculate transducin specific activity) are shown for each collection time point. (B) Quantification of IRBP and transducin in light and dark. IRBP band densities decreased significantly during 10 hours of either light or darkness, whereas transducin specific activity remained high (n = 5–7 tubes of six eyes each for each time point; mean ± SEM).

Figure 8.

 

IRBP turnover during the day versus the night. Eyecups from animals killed just after light onset (day) or just before light offset (night) were subjected to a [³⁵S]methionine pulse–chase and then incubated for 1 or 8 hours without interruption of their established light–dark cycle (i.e., “day” eyecups were incubated in the light, and “night” eyecups were incubated in the dark). IRBP was immunoprecipitated and the amount of radiolabel associated with the IRBP band measured by phosphorimaging (representative bands are shown in the inset). [³⁵S]IRBP signal intensity decreased approximately 60% during the day but changed little overnight. Data represent mean ± SEM for four to five tubes (six eyes per tube).

The authors thank Oswald Steward and Thomas Briese for assistance with RT-PCR experiments and Ellen Van Niel and Yonde Bao for preparation of plasmid constructs.