Vision 2011; 17:3224-3233
Received 21 March 2011 | Accepted 10 December 2011 | Published 14 December 2011
Amitavo Mitra, Yashodhan Chinchore, Ronald Kinser, Patrick J. Dolph
Department of Biological Sciences, Dartmouth College, Hanover, NH
Correspondence to: Patrick J. Dolph, Department of Biological Sciences, Dartmouth College, 6044 LSC, Hanover, NH, 03755; Phone: (603) 646-1092; FAX: (603) 646-1347; email: firstname.lastname@example.org
Dr. Yashodhan Chinchore is now at the Department of Genetics, Harvard Medical School, 77 Avenue Louis Pasteur, Boston, MA 02115
Purpose: In this study we investigated the biochemical and cell biologic characteristics of flies expressing two novel dominant alleles of the major rhodopsin encoding gene neither inactivation nor afterpotential E (ninaE) in a heterozygous background.
Methods: Presence of the deep pseudopupil in flies was assayed 5 days post eclosion. For structural analysis, 1-μm-retinal cross sections were obtained from fixed and resin-embedded Drosophila heads. Confocal microscopy was performed on dissected retinas stained with antibodies specific for rhodopsin, NinaA, and F-actin. Rhodopsin levels were determined by western and slot blot analysis.
Results: Dominant rhodopsin mutants showed progressive age-dependent and light-independent loss of the deep pseudopupil, without any apparent retinal degeneration at the morphological level. Expression of mutant rhodopsin caused rhodopsin to mislocalize to the cell body and the endoplasmic reticulum compartment. Mutant rhodopsin also caused loss of solubility of wild-type rhodopsin and its accumulation presumably as a high molecular mass complex in the photoreceptor cell body.
Conclusions: In heterozygous mutant flies, there is loss of wild-type rhodopsin immunoreactivity on a western assay but less reduction using slot blot analysis. This suggests that mutant rhodopsin is likely inducing the misfolding and insolubility of wild-type rhodopsin. Localization of rhodopsin revealed that in mutant flies, wild-type rhodopsin is mislocalized to the cell body and the endoplasmic reticulum.
Retinitis pigmentosa (RP) is a diverse collection of genetically inherited diseases that is characterized by loss of visual acuity and retinal degeneration in humans [1-3]. The heterogeneity of the disease can be appreciated by the fact that RP can be inherited as an autosomal dominant (ADRP), autosomal recessive, or X-linked disease . Despite the multimodal inheritance pattern, ADRP accounts for almost a quarter of all cases of RP [5,6]. Mutations in the rhodopsin gene account for the majority or most the underlying genetic determinants of ADRP cases in worldwide surveys [7-10], thus making the study of rhodopsin physiology in the context of RP an important avenue in elucidating the molecular mechanisms of RP. Despite the fact that mutations in a single gene (rhodopsin) are responsible for most cases of RP, mechanistic details might be complicated since in some cases not only does the same mutation in different people exhibit variability with respect to severity of their disease but also different amino acid substitutions at the same position can lead to distinct phenotypes [11,12].
The Drosophila phototransduction pathway has been studied in detail and has been established as a model system to elucidate mechanisms of retinal degeneration [13-15]. Even though the vertebrate and Drosophila phototransduction cascades have a different organization, they share anatomic and molecular features, making Drosophila an appropriate model. The Drosophila eye is a compound eye that consists of about 800 individual repeating units known as ommatidia. Individual ommatidia have about 20 cells out of which eight are photoreceptor cells. The phototransduction machinery in photoreceptor cells is localized to actin-rich microvillar structures known as rhabdomeres that are functionally equivalent to vertebrate outer segments. Loss of individual rhabdomeres within photoreceptors and/or the loss of the ommatidial array are indicative of retinal degeneration. The vertebrate and invertebrate light-stimulated signal transduction pathways are thematically similar, as evidenced by several common proteins [13,16].
Numerous rhodopsin mutations were isolated in Drosophila screens in the late 1960s [17-20], many of which cause retinal degeneration in fly photoreceptors. In a more recent screen, dominant neither inactivation nor afterpotential E (ninaE) alleles that undergo retinal degeneration have been isolated, some of which correspond to mutations in the same residues of human rhodopsin associated with ADRP . Relevance for a Drosophila model of RP was further established when it was found that the most frequently occurring mutation in ADRP, a proline substitution at position 23 by histidine, faithfully recapitulated the dominant degenerative phenotype when engineered into the Drosophila rhodopsin gene .
Quantification of the rhodopsin present in such mutant flies, especially for rhodopsin mutants, is a widely used assay in all studies, but the lack of any detailed insight into the fate of rhodopsin has led to questioning how low levels of rhodopsin lead to rhodopsin-mediated retinal degeneration. The endoplasmic reticulum (ER) has been implicated to play a role in ADRP [10,23], and only recently has the importance of the accumulation of misfolded rhodopsin and its clearance mechanism been elucidated [24,25]. Accumulation of rhodopsin in photoreceptors, which potentially can be prone to aggregation and/or resistant to proper maturation/degradation, may contribute to the underlying mechanism(s) of retinal degeneration in phototransduction mutants, which otherwise show variability in functional and morphological phenotypes.
In this study we report two new alleles of the Drosophila major rhodopsin gene (ninaE), ninaER11 and ninaER12. These two alleles are dominant and exert their effect in a light-independent manner. Using a combination of cell biologic and molecular approaches, we describe characteristics of these two novel alleles along with three previously characterized alleles as a comparison. We show that in heterozygous mutant flies, there is a significant accumulation of rhodopsin, part of which may be mislocalized to the cell body and the ER along with the disruption of the ER pattern within the cell body of photoreceptors. Mutant rhodopsin expression exerts its dominant effect by inducing misfolding and insolubility of wild-type rhodopsin.
Drosophila melanogaster stocks ninaE5, ninaED1, and ninaED2 were obtained from the Bloomington stock center (Bloomington, IN). The ninaER11 and ninaER12 alleles were identified in a previous arrestin 2 (arr2) degeneration enhancer screen . All ninaE alleles were crossed into a white (w1118) background, and progeny were collected that were haploid for mutant and wild-type rhodopsin. Flies were reared in constant darkness.
At least 15 flies of each genotype were reared in complete darkness for 5 days and tested for the presence of deep pseudopupil (DPP) at 24-h intervals. Flies with a distinct DPP were given a score of 5; flies with faint to diffuse DPP were given a score of 3; and flies with no DPP were given a score of 0. The average of the observations was reported as the DPP score of the fly population within each genotype for each day.
Heads of flies were bisected and immersed in ice-cold 2% glutaraldehyde in 0.1 M phosphate buffer. An equal volume of 2% osmium tetroxide was added and incubated for 30 min on ice. The glutaraldehyde/osmium tetroxide solution was removed, and the eyes were washed with 0.1 M phosphate buffer, followed by treatment with 2% osmium tetroxide for 60 min. The eyes were then serially dehydrated with increasing ice-cold ethanol (15%, 30%, 50%, 70%, and 90%) for 10 min each on ice. This was followed by two 10-min incubation periods in 100% ethanol at room temperature and two 10-min incubations in propylene oxide (Electron Microscopy Sciences, Hatfield, PA). The heads were incubated in a 50% propylene oxide/50% Durcupan ACM Fluka (Sigma, St. Louis, MO) mixture overnight on a rotator at room temperature, followed by a 4-h incubation in 100% Durcupan. Eyes were then embedded in 100% Durcupan and cured at 70 °C overnight. Cross sections (1 μm) were cut using a Sorvall ultra microtome MT-1 (Sorvall, Newtown, CT). The sections were stained with toluidine blue and observed on a Zeiss Axioplan 2 microscope using a 63×/2 numerical aperture (NA) oil immersion objective. Digital images were captured using an Optronics DEI-750 camera Optronics DEI-750 camera and associated software (Carl Zeiss Optronics, Oberkochen, Germany). At least three flies were analyzed for each genotype and treatment paradigm.
Three fly heads of each genotype were homogenized in 25 μl Laemmli 1× sodium dodecyl sulfate (SDS) loading buffer. Each lysate (20 μl) was subjected to SDS PAGE (SDS–PAGE) followed by western analysis . The primary antibodies used were anti-Rhodopsin1 (Rh1; 1:1,000; Developmental Studies Hybridoma Bank, Iowa City, IA), anti-Rh1–1D4 (1:100,000; gift from Dr. John Hwa, Weill Cornell Medical College, New York, NY), anti-Rab7 (1:10,000) , and anti-Arrestin2 (Arr2; 1:1,000). Secondary antibodies used were horseradish peroxidase (HRP) conjugated antimouse or antigoat (1:5,000). Arr2 and Rab7 were used as loading controls.
Densitometric analysis of Rh1, Arr2, and Rab7 was performed using the BioRad ImageQuant software (Amhersham Bioscience, Piscataway, NJ). Each experiment was repeated independently three times.
For slot blot analysis, two wild-type heads were ground in 20 μl 1× SDS loading buffer, and the total volume was adjusted to 200 μl with 1× SDS loading buffer/1 M urea in 0.5× PBS (0.15 M NaCl, 0.002 M KCl, 0.01 M Na2HPO4, 0.001 M KH2PO4, pH 7.4). Serial dilutions were prepared, and samples equivalent to 1, 0.5, 0.25, and 0.125 heads were applied to a BioDot apparatus (BioRad, Hercules, CA). OptiTran BA-S 85-supported nitrocellulose membrane (Whatman, Piscataway, NJ) was used for the protein transfer. For heterozygous mutants, lysate concentrations were doubled. The membrane was denatured by incubating in 0.2 M sodium hydroxide for 30 min at room temperature followed by blocking in 5% milk in 1× PBS+0.1% Tween-20 for another 30 min. Antibody incubations were performed as in western analysis.
Densitometric analysis of total Rh1 levels in wild-type and mutant flies was performed as described for western blot analysis. Total rhodopsin level in heterozygous mutant flies relative to wild-type flies was calculated and plotted by comparing rhodopsin levels of mutants at various concentrations to those of wild-type samples. Rhodopsin levels for each sample were arrived at by averaging the values obtained from three independent repetitions.
Retinas from adult flies were prepared for whole-mount immunostaining as described previously . The primary antibodies used were anti-Rh1 (1:50; Developmental Studies Hybridoma Bank, Iowa City, IA), anti-Rh1–1D4 (1–800; a gift from Dr. John Hwa, Yale University), anti-Rh1 (1:50; polyclonal antibody; gift from Dr. C. Zuker, Columbia University, New York, NY), and anti-NinaA (1:150; gift from Dr. Craig Montell, Johns Hopkins University, Baltimore, MD). Rhabdomeres were visualized by staining for F-actin using rhodamine or Alexa-568-conjugated phalloidin (Molecular Probes, Carlsbad, CA). Secondary antibodies were antimouse- or antirabbit-conjugated Alexa-488 or Alexa-647 (1:300; Molecular Probes).
Images were captured on a Leica TCS SP confocal laser-scanning microscope (Leica Microsystems, Heidelburg, Germany) and Nikon A1RSi Confocal Workstation (Nikon, Melville, NY). Image processing was done with Adobe Photoshop (San Jose, CA). Retinas from at least three flies in each genotype were processed.
The Student t test was used for statistical comparison between wild-type controls and the various mutant genotypes. Differences were considered statistically significant at p<0.05.
Previously, we isolated a collection of genetic enhancers of arr2 degeneration . From that collection, two alleles were genetically mapped to the ninaE locus, which encodes for the major Drosophila rhodopsin, Rh1. Sequencing of the ninaE gene from these alleles indicated that they both had point mutations within the coding region. The ninaER11 allele has a proline to leucine mutation in the first cytoplasmic loop at position 84, and the ninaER12 allele has a serine to isoleucine mutation in the fourth transmembrane domain at position 177 (Figure 1).
Since ADRP accounts for the majority of cases of diagnosed RP, we are primarily interested in the dominant phenotype exerted by the mutant protein on its wild-type counterpart. To test whether these rhodopsin alleles had dominant phenotypes, we reared heterozygous ninaE mutant flies in complete darkness and assayed them for loss of the DPP . The DPP is a virtual image of the trapezoidal array of multiple rhabdomeres near the center of curvature of the Drosophila compound eye . The loss of the DPP indicates a reduction in the amount of rhodopsin or a disruption in the ommatidial structure. Both ninaER11/+ and ninaER12/+ flies reared in the dark underwent progressive loss of DPP, with only about 10% of flies showing no loss of DPP by day 5 post eclosion (Figure 2). The loss of DPP in mutant heterozygous flies compared to control flies for every time point is significant with a p value less than 0.05. Although a small number of flies retained their DPP at the end of our assay time frame, the data clearly showed the trend of loss of the DPP in a light-independent manner. The continued presence of detectable DPP in a minor percentage of flies is likely due to the qualitative basis of the DPP assay and/or variation in the phenotype presentation within the fly population at the time of our assay. These data suggest that ninaER11 and ninaER12 are dominant alleles of ninaE and that the loss of DPP is age dependent and light independent, suggesting underlying retinal degeneration.
Since heterozygous mutant flies show loss of DPP, we examined aged flies to observe the loss of rhabdomeres and/or loss of structural integrity of the ommatidial array. Flies aged for 3 weeks in complete darkness did not show any readily identifiable hallmarks of degeneration and had a full complement of seven rhabdomeres per ommatidium and a regular ommatidial array (Figure 3A-C). Moreover, analysis of heterozygous animals reared in a 12 h:12 h light–dark cycle for up to 30 days showed minimal changes in photoreceptor cell structure and rhabdomere morphology (Figure 3D-F). These results are also consistent with previous reports on published ninaE alleles. It should be noted that the previously identified ninaED1 dominant allele showed degeneration only when exposed to continuous light and showed minimal degeneration in dark-reared flies only after 40 days . We also analyzed ninaER11 and ninaER12 homozygous flies under two different aging paradigms (7-day and 3-weeks old, dark reared) by means of cross sections and did not observe any noticeable retinal degeneration (data not shown).
The lack of major structural defects is not necessarily in conflict with the observed early loss of DPP data. It has been previously shown that the loss of DPP can be correlated to reduced rhodopsin content even when the photoreceptors have intact rhabdomeres [29,31]. Moreover, although rhodopsin is required for photoreceptor morphogenesis and viability [32,33], it has been previously shown that flies with a 70% reduction in photopigment level  or a reduction in rhodopsin levels to <3% in photoreceptors  have an overall normal photoreceptor morphology.
Three mutations in the human rhodopsin gene have been identified that cause congenital stationary night blindness [36-38]. In this form of retinopathy, similar to RP, loss of vision in low-light conditions is an early clinical observation, but unlike RP where there is a prominent loss of rods and cones, in congenital stationary night blindness there is no loss of rods and cones. These histological results are similar to those observed in the Drosophila mutants in this study as no loss of rhabdomeres in photoreceptor cells was observed.
Our histological data indicate that in mutant photoreceptors low levels of rhodopsin may be properly trafficked to the rhabdomeres to maintain rhabdomeric integrity. To test whether rhabdomeric rhodopsin is present in these flies and whether there is any mislocalization of rhodopsin, we stained whole retinas with anti-Rh1 and the ER-specific marker anti-NinaA. The ninaA gene in the Drosophila genome encodes for an eye-specific homolog of cyclosporine A-binding protein , which functions as an Rh1-specific chaperone and is involved in its maturation and transport from the ER to the rhabdomeric membranes [40-42]. We wanted to ascertain if, in our mutants, rhodopsin is localized to the ER, indicating maturation/trafficking defect/delay.
In dark-reared wild-type flies, Rh1 is localized to the base of rhabdomeres in a crescent-like shape and no cytoplasmic rhodopsin can be observed (Figure 4C). NinaA shows a diffused staining pattern throughout the cell body (Figure 4B). All mutant photoreceptors show some rhabdomeric localization of rhodopsin, suggesting that some proportion of rhodopsin may be correctly trafficked to its site of activity. This rhabdomeric rhodopsin may explain the long term viability of photoreceptors and lack of morphological defects. However, both mutants analyzed exhibit some defect in rhodopsin localization where a significant amount is mislocalized to the cell body.
In ninaER11/+ flies, a proportion of rhodopsin is mislocalized in a punctate-staining pattern in the cell body. Many of the large rhodopsin-positive puncta also stain positive for NinaA. In contrast to NinaA distribution in wild-type flies, NinaA in ninaER11/+ is limited to the cell body periphery and is no longer localized throughout the cell body (Figure 4E-H). Flies heterozygous for the ninaER12 mutation show extensive diffused rhodopsin mislocalization to the cell body that does not seem to localize with NinaA. In contrast, NinaA is localized in a wild-type distribution (Figure 4I-L). As we use NinaA as a marker for the ER, we speculate, based on the restrictive ER staining in ninaER11 heterozygous flies, that the normal ER distribution within photoreceptor cells in a mutant background may be affected. Another mutually nonexclusive possibility is that the majority of the NinaA is relocated within the ER to areas where rhodopsin is present and likely concentrated. We see similar results with the previously published dominant ninaE allele ninaED2 (data not shown). We also see normal NinaA distribution in another previously published dominant ninaE allele ninaED1 (data not shown); this distribution is similar to our observations in ninaER12/+ flies. This suggests that there exists heterogeneity within the dominant ninaE alleles in presentation of defects at the cellular level. This collection of data suggests that rhodopsin is mislocalized to the cell body and may also be accumulating in the ER, resulting in changes in the ER distribution pattern and/or localization of NinaA within the ER.
Previously it has been shown that mutations in rhodopsin cause a reduction in protein levels. When protein levels are determined by western blot analysis, ninaER11/+ and ninaER12/+ show significantly (10% [p<0.0001] and about 6% [p<0.001], respectively) rhodopsin levels relative to the wild type (Figure 5A,C). Since all of the animals analyzed were heterozygous and contained a haploid content of wild-type rhodopsin, the loss of rhodopsin immunoreactivity below 50% indicates that the mutant rhodopsin is triggering the loss of rhodopsin in this assay.
In human cases of ADRP, mutant rhodopsin has been shown to be prone to aggregation and to form high molecular weight complexes . Abnormal accumulation of protein aggregates is also associated with diverse human pathologies, such as Huntington’s disease, Alzheimer disease, and spinocerebellar ataxia . Therefore, we explored whether rhodopsin becomes insoluble in the dominant mutants. We reasoned that insoluble aggregated rhodopsin might not be detectable on immunoblots since the protein complexes may be unable to enter the polyacrylamide gel matrix. However, insoluble proteins may be detectable by slot blot analysis since this method does not require the proteins to be solublized. Interestingly, slot blot data indicate that there is significantly more rhodopsin in the dominant mutants than detected by western blot. Flies heterozygous for the ninaER11 and ninaER12 mutations show approximately twofold to fourfold more total rhodopsin than on western blots. These differences are statistically significant (p<0.001; Figure 5B,C). This is also evident in the immunolocalization data in Figure 4. Even though there is no detectible protein on a western blot, both ninaER11 and ninaER12 heterozygotes exhibit Rh1 immunoreactivity. In the case of ninaER12, levels appear even higher than the wild type.
We were interested in whether other previously characterized rhodopsin alleles also show a difference between soluble (western blot) and total (slot blot) rhodopsin levels. We chose to analyze ninaED1, ninaED2, and ninaE5 alleles (Figure 1). Similar to the ninaE11 and ninaE12 alleles in this study, these rhodopsin mutations have previously been shown to induce low levels of rhodopsin on western blots and minimal retinal degeneration . ninaED1/+ and ninaED2/+ flies show a significant severe loss of rhodopsin immunoreactivity on a western blot, less than 10% of wild-type levels (p<0.0001; Figure 5A, C). This is comparable to previously described rhodopsin levels in these two mutants . We have found that ninaE5/+ flies exhibit approximately 15% rhodopsin levels (p<0.001) compared to about 1.5% in ninaE5/+ flies relative to the wild type, as reported elsewhere  (Figure 5A,C).
However, using slot blot analysis, there is a significant sixfold (p<0.001) and 15-fold (p<0.05) more detectable rhodopsin in ninaED1/+ and ninaED2/+ flies, respectively, while ninaE5/+ shows about threefold (p<0.05) more rhodopsin compared to wild-type levels (Figure 5B,C). These data suggest that rhodopsin in dominant ninaE alleles may not be efficiently degraded, and instead much of the rhodopsin may still be present in a form that is resistant to solubility. The data also suggest that western blot analysis may not be sufficient to detect proteins that are prone to aggregate, and other methods that do not require protein solubilization should be used.
Our biochemical data indicates that there is a loss of both wild-type and mutant rhodopsin in the heterozygous flies. To establish that mutant rhodopsin expression specifically induces loss of wild-type rhodopsin immunoreactivity, we expressed epitope-tagged rhodopsin in the presence of the rhodopsin variant. We took advantage of a variant of Rh1 where the C-terminus of Drosophila Rh1 is replaced by the bovine rhodopsin C-terminus . A monoclonal antibody, ID4, specifically recognizes the C-terminus of bovine rhodopsin  and therefore will recognize tagged rhodopsin (1D4) but not endogenous rhodopsin (+/+; Figure 6A, lane 1; Figure 6F-H). In this way we can monitor the presence of wild-type rhodopsin independently from mutant rhodopsin.
When ID4-tagged wild-type rhodopsin (Rh1–1D4) is expressed along with mutant rhodopsin, there is a significant decrease in immunoreactivity of tagged wild-type rhodopsin. The decrease in immunoreactivity on immunoblots closely parallels the data observed when all Rh1 is analyzed (Figure 6A). Compared to flies expressing Rh1–1D4 alone, we detected 16.6% rhodopsin in flies of ninaER11/Rh1–1D4 (p<0.01), 5% in ninaER12/Rh1–1D4 (p<0.0001), 13.5% in ninaED1/Rh1–1D4 (p<0.01), and 5.4% in ninaED2/Rh1–1D4 (p<0.0001; Figure 6B). This loss of wild-type rhodopsin indicates that mutant rhodopsin is specifically triggering the loss of immunoreactivity of wild-type rhodopsin in this assay.
Since we have demonstrated that mutant rhodopsin is forcing the loss of solubility, which is likely due to misfolding of wild-type rhodopsin, we predicted that wild-type rhodopsin is also mislocalized in the mutant photoreceptors. We chose to look at ninaER11/+ flies specifically as we had observed multiple Rh1-positive puncta in the cell body of dark-reared animals. Immunolocalization data demonstrate that there are significantly more wild type Rh1–1D4-positive vesicles mislocalized to the cell body compared to haploid Rh1–1D4 flies alone (Figure 6C-E, I-K). We were unable to perform slot blots using the 1D4 antibody, as is shown in Figure 5. However, even though there are severely reduced levels of Rh1 on western blots for ninaER11/Rh1–1D4 animals, there are detectable levels using immunofluorescence, indicating that Rh1 is still present but is in a conformation that cannot be revealed using western analysis.
Overall, our data indicate that mutant rhodopsin expression exerts its dominant effect by inducing mislocalization and insolubility of wild-type rhodopsin. This results in the likely impaired clearance and accumulation of rhodopsin that may underlie the cause of cell death.
We would like to thank Ann Lavanway for her help in confocal microscopy and Surachai Supattapone and Eric Schaller for advice on the biochemical analysis of rhodopsin. The confocal microscope used for Figure 4 was supported in part by National Science Foundation Grant DBI-9970048 to Roger D. Sloboda. Funds to purchase the Nikon A1RSi Confocal Workstation used in Figure 6 were provided by NSF award #DBI-1039423 and Dartmouth. This work was supported in part by a grant from the National Institute of Neurologic Disorders and Stroke. The authors have no competing financial interest related to this manuscript.