Abstract
Retinoids act as chromophore co-factors for light-detecting rhodopsin proteins. In vertebrates, retinoids also actively regulate gene expression. Whether retinoids regulate gene expression in Drosophila for a specific biological function remains unclear. Here, we report that Drosophila fatty acid binding protein (fabp) is a retinoid-inducible gene required for Rhodopsin-1 (Rh1) protein homeostasis and photoreceptor survival. Specifically, we performed a photoreceptor-specific gene expression profiling study in flies bearing a misfolding-prone Rhodopsin-1 (Rh1) mutant, ninaEG69D, which serves as a Drosophila model for Retinitis Pigmentosa. ninaEG69D photoreceptors showed increased expression of genes that control Rh1 protein levels, along with a poorly characterized gene, fabp. We found that in vivo fabp expression was reduced when the retinoids were deprived through independent methods. Conversely, fabp mRNA was induced when we challenged cultured Drosophila cells with retinoic acid. In flies reared under light, loss of fabp caused an accumulation of Rh1 proteins in cytoplasmic vesicles. fabp mutants exhibited light-dependent retinal degeneration, a phenotype also found in other mutants that block light-activated Rh1 degradation. These observations indicate that a retinoid-inducible gene expression program regulates fabp that is required for Rh1 proteostasis and photoreceptor survival.
Author Summary Rhodopsins are light-detecting proteins that use retinoids as chromophore co-factors. In vertebrates, retinoids also actively regulate gene expression. Whether retinoids regulate Rhodopsin function aside from its role as a chromophore remains unclear. Here, we report that Drosophila fatty acid binding protein (fabp) is a retinoid-inducible gene required for Rhodopsin-1 (Rh1) protein homeostasis and photoreceptor survival. Specifically, we found that fabp is among the genes induced by a misfolding-prone Rhodopsin-1 (Rh1) mutant, ninaEG69D, which serves as a Drosophila model for Retinitis Pigmentosa. We further found that fabp induction in ninaEG69D photoreceptors required retinoids. fabp was required in photoreceptors to help degrade light-activated Rh1. In the absence of fabp, Rh1 accumulated in cytoplasmic vesicles in a light-dependent manner, and exhibited light-dependent retinal degeneration. These observations indicate that a retinoid-inducible gene expression program regulates fabp that is required for Rh1 proteostasis and photoreceptor survival.
Introduction
Rhodopsins are G-protein coupled proteins associated with retinal chromophores to detect light and initiate signal transduction (1). As in mammals, Drosophila has multiple Rhodopsins, including ninaE (neither inactivation nor afterpotential) that encodes the Rhodopsin-1 (Rh1) protein expressed in R1 to R6 photoreceptors (2-4). Functional Rh1 is covalently attached to the 11-cis-3-hydroxyretinal chromophore, which is derived from dietary vitamin A (5-7). ninaE loss of function results in an impairment of light detection (4, 8).
Abnormal Rh1 protein homeostasis is a frequent cause of retinal degeneration. One class is caused by a group of ninaE missense mutations that dominantly cause progressive age-related retinal degeneration (9, 10). These alleles are analogous to human rhodopsin mutations that underlie age-related retinal degeneration in Autosomal Dominant Retinitis Pigmentosa (ADRP) patients (11, 12). Using the Drosophila ninaEG69D allele as a model, we previously established that these mutations impose stress in the endoplasmic reticulum (ER), which contributes to retinal degeneration (13, 14). The human rhodopsin allele that is most frequently found associated with ADRP, the P23H mutant, similarly causes ER stress in mammalian cells (15).
Cellular mechanisms that regulate rhodopsin protein levels affect retinal degeneration. Flies bearing one copy of the ninaEG69D allele have total Rh1 protein levels reduced by more than half, indicating that both the mutant and the wild type Rhodopsin-1 proteins undergo degradation in these flies (9, 10). Three ubiquitin ligases that specialize in the degradation of misfolded endoplasmic reticulum (ER) proteins mediate the degradation of Rh1 in ninaEG69D flies (16). Overexpression of these ubiquitin ligases can delay the onset of retinal degeneration in Drosophila ninaEG69D flies, indicating that excessive misfolded Rh1 is a contributing factor to retinal degeneration (14, 16).
Functional wild type Rh1 proteins also undergo degradation after being activated by light. Specifically, this occurs after light-activated Rh1, also referred to as metarhodopsin (M), engages with Arrestin that mediates feedback inhibition (17). Rh1 forms a stable complex with Arrestin and together undergo endocytosis for degradation (18-21). Such degradation of activated Rh1 is essential, as too much Rh1 accumulation in the endosome/lysosome defective photoreceptors results in light-dependent retinal degeneration (18, 22-26). In Rh1 endocytosis/degradation defective mutants, retinal degeneration could be delayed by conditions that reduce overall Rh1 levels (24, 25, 27), indicating that too much active Rh1 is a cause of retinal degeneration. These aspects appear to be conserved across phyla, as the human rhodopsin mutants that exhibit high affinities for Arrestin display endosomal abnormalities and are associated with severe forms of ADRP (28, 29).
Retinoids are among the molecules implicated in regulating Rh1 protein levels. Deprivation of vitamin A, which serves as a precursor for the retinal chromophore, causes a reduction in overall Rh1 levels (30-34). Such an effect is largely attributed to the importance of chromophores in Rh1 protein maturation. Aside from its role as a rhodopsin cofactor, retinoids regulate gene expression in vertebrates. The best characterized transcription factors that mediate this response are nuclear hormone receptors, including RAR and RXR, which become transcriptional activators upon binding to all-trans or 9-cis retinoic acid (35). Cellular Retinoic Acid Binding Protein -I and -II (CRABP-1, -II) bind to the lipophilic retinoic acids and deliver them to RAR and RXR in the nucleus (36, 37). In addition to mediating the RA signaling response, CRABP-II itself is induced by RA signaling (38). Whether these mechanisms are conserved in Drosophila has not been examined in detail in part because the Drosophila genome does not encode a RAR homolog (7, 39). Intriguingly, several Drosophila genes require the retinoid precursor vitamin A for their gene expression (40, 41). One such gene is highroad, whose expression is induced by retinoic acids in cultured cells and mediates the degradation Rh1 in ninaEG69D/+ (42). These results suggested that retinoid-mediated gene expression programs may be involved in Rh1 homeostasis.
Here, we report that Rh1 protein levels are regulated by the Drosophila CRABP homolog, fatty acid binding protein (fabp). Similar to CRABP-II, fabp expression is induced by retinoids. Loss of fabp enhances total Rh1 levels in ninaE wild type and G69D mutant backgrounds. Moreover, loss of fabp causes light-dependent retinal degeneration. Our results indicate that this retinoic acid inducible gene controls Rh1 protein levels, and this regulatory axis is essential for photoreceptor survival.
Results
Photoreceptor-specific gene expression profiling shows fabp induction in ninaEG69D eyes
To better understand how photoreceptors respond to stress imposed by the ninaEG69D allele, we performed a photoreceptor-specific gene expression profiling analysis. We specifically employed a previously described approach in which the expression of the nuclear envelope-localized EGFP::Msp300KASH is driven in specific cell-types through the Gal4/UAS system to isolate the EGFP-labeled nuclei for RNA-seq analysis (43, 44). We used the Rh1-Gal4 driver to isolate ninaE expressing R1 to R6 photoreceptor nuclei from the adult fly ommatidia (Figure 1A). Microscopy imaging confirmed that anti-EGFP beads enriched the EGFP::Msp300KASH-tagged nuclei (Figures 1B, C). RNA-seq was performed with nuclei isolated from ninaE wild type and ninaEG69D/+ photoreceptors.
Differential gene expression analysis showed 182 genes whose expression changed with adjusted p values below 0.01 (Supplementary Table 1). Among the most highly induced genes was gstD1 (Figure 1D), which was also identified as an ER stress-inducible gene in a separate study performed with larval imaginal discs (Brown et al., under review). The ER chaperone encoding cnx99A was also induced (Figure 1D), consistent with the previous report that ninaEG69D imposes ER stress in photoreceptors (13).
Also, notable from the differential gene expression analysis was the induction of genes that could affect Rh1 levels. ninaE was itself induced in ninaEG69D samples (Figure 1D). Since ninaEG69D/+ flies have very low Rh1 levels (9, 10), we speculate that increases in ninaE transcription may be part of a feedback homeostatic response. Also induced were genes Arrestin 1 (Arr1), Arrestin 2 (Arr2) and culd (Figure 1D), which promote the degradation of light-activated Rh1 in photoreceptors (17, 18, 23, 45).
Also, among the ninaEG69D-induced genes was fabp (Figure 1D), which encodes a protein homologous to human CRABP-1, -II and FABP5 (Figure 1E). We validated the induction of fabp mRNA in ninaEG69D/+ through q-RT PCR (Figure 1F). The human homologs of fabp are known to bind all trans RA with high affinity (36, 37, 46, 47). Notably, CRABP-II is one of the well-characterized RA inducible genes in mammalian cells (38). fabp drew our interest because a retinoic acid-inducible Drosophila gene, hiro, regulates Rh1 levels in ninaEG69D/+ flies (42).
fabp expression is regulated by Vitamin A and retinoids
To test if Drosophila fabp is also regulated by retinoic acids (RA), we examined fabp mRNA levels through RT-qPCR in Drosophila S2 culture cells treated with or without 10mM RA. We found that RA treated cells showed an increase in fabp transcripts after 60 minutes of RA exposure (Figure 2A, B).
To examine if fabp expression in fly tissues is affected by the availability of Vitamin A and its metabolites, we examined fabp levels in the mutants of ninaD and santa maria that have impaired transport of carotenoids, the precursors of retinoids. These mutants are devoid of retinoids in the retina, as evidenced by defective rhodopsin maturation and light detection (31, 34). We found that these mutants had reduced FABP protein as assessed through western blot of fly head extracts (Figure 2C). Consistently, the mutants also had reduced fabp mRNA levels as assessed through q-RT-PCR (Figure 2D).
These observations prompted us to examine if other ninaEG69D-inducible genes require santa maria and ninaD for proper expression. We focused on candidates known to be involved in Rh1 protein regulation. Among those tested, the mRNAs of Arr1 and Arr2 were found to be reduced in ninaD or santa maria mutant backgrounds (Figure 2E, F). Not all genes involved in Rh1 homeostasis were affected in these mutants. For example, fatty acid transport protein (fatp) is a gene whose loss-of-function increases Rh1 protein levels (27). fatp mRNA levels were affected neither in the mutant backgrounds of ninaD nor santa maria (Figure 2G). These results indicate that the expression of fabp, Arr1, and Arr2 specifically require retinoid and carotenoid transporters, ninaD and santa maria.
An fabp protein trap line shows carotenoid-dependent expression in the larval intestine
To independently validate the carotenoid-dependent expression of fabp in vivo, we utilized the fabp protein trap line CA06960. This P-element insertion line has a GFP with splice donor and acceptor sites, designed to make fusion proteins with the endogenous fabp coding sequence (Figure 3A). Anti-GFP western blot of fly extracts confirmed the expression of a GFP-fused protein in adult fly head extracts with the predicted size (Figure 3B).
In the third instar larva, the fabpCA06960 line had GFP expression detectable in several regions of the intestine (Figure 3C, E). Such expression was abolished when the flies were reared in Vitamin A deficient food (Figure 3D). Consistently, the expression of GFP was suppressed in the mutant backgrounds of ninaD and santa maria (Figure 3F, G). In adult flies, the GFP signal was most prominent in the female abdomen, which was reduced in the ninaD mutant background (Figure 3H). These results independently support the idea that fabp expression depends on carotenoids.
Loss of fabp increases Rh1 protein levels
To test possible Rh1 regulation by retinoids, we examined several Drosophila homologs of mammalian RA signaling mediators. The candidate genes we analyzed included enzymes that convert Vitamin A to retinoids (e.g. ninaB (48)), nuclear hormone receptors (e.g. knl and eg) and fabp. For our assay, we used ninaEG69D/+ flies, which have drastically reduced Rh1 protein levels as compared to ninaE wild type flies (Figure 4A). Specifically, we drove the expression of RNAi lines that target the genes of interest in the photoreceptors of these ninaEG69D/+ flies using the Rh1-Gal4/UAS system (Figure 4A). An RNAi line that targeted fabp showed a reproducible effect of partially enhancing Rh1 levels as assessed through western blots of fly head extracts (Figure 4A, B).
To validate fabp RNAi results, we employed an fabp loss of function allele, EY02678, which has a P-element inserted near an exon-intron boundary (Figure 4C). This allele has strongly reduced FABP expression as assessed through western blot (Figure 4D). We found that ninaEG69D/+ Rh1 levels increased in the fabpEY02678 -/- background (Figure 4E, F), validating the results with fabp RNAi. We further found that the loss of fabp increased Rh1 levels even in the ninaE wild type flies (Figure 4E, F). When we re-introduced fabp expression in fabp mutant flies using the eye specific GMR-Gal4 driver, Rh1 protein levels were restored to those levels of wild type controls (Figures 4E, F). These results indicate that fabp affects general Rh1 protein levels.
Regulation of Rh1 by fabp is light dependent
Since wild type Rh1 proteins are most notably degraded through light-dependent endocytosis (18-21), we examined whether fabp regulation of Rh1 was light-dependent. We found that fabp mutants showed higher Rh1 levels when the flies were reared under light. Such effect was not seen in flies that were reared in dark (Figure 5A, B).
To examine the pattern of Rh1 distribution in photoreceptors, we performed anti-Rh1 immuno-labeling in the adult Drosophila retina. In control wild type flies, Rh1 is predominantly detected in the rhabdomeres of R1 to R6 photoreceptor cells organized in a trapezoidal pattern (Figure 5C-E). In fabp -/- flies reared under light, however, there were additional anti-Rh1 signals in intracellular vesicles (Figure 5F).
Vesicular Rh1 signals reportedly appear in flies exposed to light, becoming even more prominent in mutants that have defects in Rh1 trafficking to the lysosome (24, 25). We found that vesicular Rh1 patterns in fabp -/- eyes were also light-dependent, as extra-rhabdomeric anti-Rh1 signals mostly disappeared in flies raised under constant darkness (Figure 5G). Together, these results suggest that light-activated Rh1 localize to intracellular vesicles, and these proteins are stabilized in fabp mutants.
fabp mutants show light-dependent retinal degeneration that is suppressed in the ninaEG69D/+ background
To examine whether fabp mutants affect retinal degeneration, we used Rh1-GFP flies with their photoreceptors labeled with green fluorescence. An intact ommatidia has R1-R6 photoreceptors arranged in a trapezoidal pattern that is readily visible as pseudopupils in live flies under low power microscopes (Figure 6A-D). Under standard conditions in which the flies were exposed to moderate light (see Methods), most wild type flies maintained this trapezoidal pattern of Rh1-GFP pseudopupils for the first thirty days after eclosion (Figure 6E, black line). fabpEY02678 -/- flies, on the other hand, showed signs of severe age-related retinal degeneration under otherwise identical conditions: Specifically, a few flies of this genotype began showing the loss of Rh1-GFP pseudopupils at day 15, with almost all examined flies having signs of retinal degeneration by day 28 (Figure 6E, red line; 6F). The difference between wild type controls and fabpEY02678 -/- was statistically significant (Log-rank test, p < 0.0001). Retinal degeneration in fabp mutants was light-dependent, as those reared in the dark did not exhibit signs of photoreceptor degeneration (Figure 6G). The light-dependent nature of photoreceptor degeneration correlated with fabp’s effect on Rh1 levels.
As reported previously, ninaEG69D/+ flies showed age-related retinal degeneration that started occurring around day 17, with most flies exhibiting retinal degeneration at day 30 (Figure 6E, black dotted line). Surprisingly, flies containing ninaEG69D/+ in the fabp -/- background had a significantly delayed course of retinal degeneration, with most flies still showing intact Rh1-GFP pseudopupils 30 days after eclosion. While surprising, such genetic interaction with ninaEG69D is not unprecedented. Previous studies found that mutants that increase wild type Rh1 levels, such as fatty acid transport protein (fatp), cause severe retinal degeneration. Such retinal degeneration is suppressed in the ninaEG69D/+ background (27).
Discussion
The biological role of retinoic acid signaling is now well-delineated in vertebrates. However, the biological role and the mechanism of retinoid-mediated gene expression in Drosophila have remained unclear. Here, we showed that the expression of Drosophila fabp is regulated by vitamin A and retinoids. We found that fabp regulates Rh1 protein levels and loss of fabp results in retinal degeneration.
In Drosophila, the major phenotype associated with vitamin A deficiency is the loss of visual function. Significantly lower levels of Rh1 protein are detected under these conditions, as vitamin A deficiency would deplete the chromophore 11-cis 3-hydroxy retinal, which is normally required for proper Rh1 maturation. Our data presented here indicates that there is an additional layer of Rh1 regulation through retinoid-inducible fabp.
Several pieces of evidence presented here support the idea that fabp is involved in the endosomal/lysosomal degradation of light-activated Rh1. Specifically, we found that Rh1 levels increase in fabp mutants when flies were reared under light, but not when the flies were reared in constant darkness. Furthermore, immunohistochemical analysis shows that Rh1 accumulates in intracellular vesicles of fabp mutants only when the flies were reared under light. Since it is now well-documented that light-activated Rh1 undergoes endocytosis and lysosomal degradation (18-21), we interpret that fabp is specifically involved in this process.
We further note that the accelerated retinal degeneration phenotype of fabp mutants are reminiscent of other genetic conditions that result in endosomal accumulation of Rh1 in response to light. Examples of this type include mutations in norpA, culd, retromer complex proteins, and fatty acid transport protein (18, 25, 27, 45). As these genes normally regulate light-dependent internalization of Rh1, it is likely that excessive levels of light-activated Rh1 is contributing to the retinal degeneration phenotype under these conditions.
If it is indeed the excessive Rh1 levels in fabp mutants that are the cause of retinal degeneration, reduction of Rh1 in these flies would delay retinal degeneration. This is what we observed when we examined the fabp phenotype in the ninaEG69D/+ background, which drastically reduces overall Rh1 levels. The way fabp and ninaEG69D genetically interacted with each other was interesting in that ninaE G69D/+ flies normally show age-related retinal degeneration, which was also delayed in the fabp mutant background. This hints at the possibility that too low Rh1 levels is a contributing factor of retinal degeneration in ninaEG69D/+ photoreceptors. Partial restoration functional Rh1 in the fabp -/- background appears to help delay retinal degeneration in ninaEG69D/+ eyes.
In conclusion, we showed that fabp expression is regulated by retinoids and carotenoids in Drosophila. Our results indicate that retinoids not only serve as chromophores for rhodopsins, but have additional roles in regulating Rh1 protein levels. It remains to be examined whether mammalian CRAPBs similarly regulate rhodopsin levels and affect retinal degeneration in response to retinoic acids.
Author contributions
H.H. performed all experiments. H.H. and H.D.R. together designed experiments and analyzed data. H.D.R. wrote the manuscript draft with H.H.’s edits and inputs.
Competing interests
The authors declare no competing interests.
Materials and Methods
Fly Genetics
All fly crosses were maintained in 25 °C. Unless otherwise stated, flies were reared with a standard cornmeal-agar diet supplemented with molasses. Vitamin A deficient food was made by mixing 12 g yeast, 1.5 g agar, 7.5 g sucrose, 30 mg cholesterol, 3.75 ml of 1.15M Nippagin, 720 μl propionic acid in distilled water volume of 150 ml.
Uas-fabp had EGFP fused in frame with the fabp’s N-terminal coding sequence. EGFP-fabp was subcloned into the pUAST plasmid, and the resulting construct was injected by Best Gene, Inc., to generate the uas-fabp transgenic line.
We used the following flies that had been reported previously: Rh1-Gal4 (49), Rh1-GFP (50), ninaEG69D (9), santa maria1 (34), ninaD1 (31), uas-dicer2 (51), UAS-EGFP::Msp-300KASH (44). fabpCA06960 (52) and fabpEY02678 were obtained from the Bloomington Drosophila Stock Center (stock numbers #50808 and #15579, respectively).
The RNAi lines used are as follows: uas-lacZ RNAi (53), uas-fabp RNAi (Bloomington Stock Center # 34685), uas-fatp RNAi (Bloomington Stock Center # 55273), uas-ninaB RNAi (Bloomington Stock Center #34994), uas-knrl RNAi (Bloomington Stock Center # 36664), uas-eg RNAi (Bloomington # 35234). These lines were crossed to the female virgins of the genotype: Rh1-Gal4; ninaEG69D/TM6B. We collected non-TM6B progeny of these crosses to examine Rh1 protein and RNA.
Photoreceptor-specific nuclear RNA extraction
We followed a published protocol to isolate Rh1-Gal4>UAS-EGFP::Msp-300KASH-positive nuclei (44). In brief, approximately 500 adult fly heads (from flies within 5 days of eclosion) per genotype were lysed in ice-cold nuclear isolation buffer (10 mM HEPES-KOH, pH 7.5; 2.5 mM MgCl2; 10 mM KCl) with a dounce homogenizer. The homogenate was filtered through a 40µm Flowmi cell strainer (WVR, cat. #BAH136800040), and the filtrate was incubated with anti-EGFP-coupled protein G Dynabeads (Invitrogen, cat. #10003D) for 1 hour at 4°C. The beads were collected using a magnetic microcentrifuge tube holder (Sigma, cat. #Z740155). Following washes with wash buffer (PBS, pH 7.4; 2.5mM MgCl2), the beads were resuspended in a final volume of 150µL of wash buffer. Then the post-isolation nuclei were suspended in 1mL of Trizol reagent (Life Technologies, cat. #15596018) for RNA extraction following standard procedures. Prior to RNA precipitation with isopropanol, 0.3M sodium acetate and glycogen were added to facilitate visualization of the RNA pellet. We then suspended the pellet in RNAse-free water and purified it using a Qiagen RNeasy MinElute cleanup kit (Qiagen, cat. #74204) following standard protocols.
Preparation of cDNA libraries, RNA-seq and data processing
The NYU Genome Technology Center performed library preparation and RNA sequencing. We quantified RNA on an Agilent 2100 BioAnalyzer (Agilent, cat. #G2939BA). For cDNA library preparation and ribodepletion, we utilized a custom Drosophila Nugen Ovation Trio low-input library preparation kit (Tecan Genomics), using approximately 20 ng total RNA per sample. For sequencing, we performed paired-end 50bp sequencing of samples on an Illumina NovaSeq 6000 platform (Illumina, cat. #20012850) using half of a 100 cycle SP flow cell (Illumina, cat. #20027464). We used the bcl2fastq2 Conversion software (v2.20) to convert per-cycle BCL base call files outputted by the sequencing instrument (RTA v3.4.4) into the fastq format in order to generate per-read per-sample fastq files. For subsequent data processing steps, we used the Seq-N-Slide automated workflow developed by Igor Dolgalev (https://github.com/igordot/sns). For read mapping, we used the alignment program STAR (v2.6.1d) to map reads of each sample to the Drosophila melanogaster reference genome dm6, and for quality control we used the application Fastq Screen (v0.13.0) to check for contaminating sequences. We employed featureCounts (Subread package v1.6.3) to generate matrices of read counts for annotated genomic features. For differential gene statistical comparisons between groups of samples contrasted by genotype, we used the DESeq2 package (R v3.6.1) in the R statistical programming environment. We excluded genes with baseMean counts less than 300 so as to avoid artifacts due to varying extent of nuclei purification.
Immunofluorescence and Western Blots
We followed standard protocols for western blots and whole mount immuno-labeling experiments using the following primary antibodies: Mouse monoclonal 4C5 anti-Rh1 (Developmental Studies Hybridoma Bank, used at 1:5000 for western blots), anti-β tubulin antibody (Covance #MMS-410P), Rabbit anti-GFP (Invitrogen #A-6455), anti-FABP antibody (54).
RT-PCR
We performed qRT-PCR using Power SYBR green master mix kit (Thermo Fisher). The primer sequences are as follows:
Rpl15F: AGGATGCACTTATGGCAAGC
Rpl15R: GCGCAATCCAATACGAGTTC
FatpF: CTCCCGGTGAGTGCAATAGCTT
FatpR: GCGGTGTGGTACAAAGGCAA
Arr1F: CATGAACAGGCGTGATTTTGTAG
Arr1R: TTCTGGCGCACGTACTCATC
Arr2F: TCGATGGAGTGATTGTGGTGG
Arr2R: GCGACCATAGCGATAGGTGG
Fabp-1F: CCGAGGTCTCAGTGTGCTC
Fabp-1R: CCGAGGTCTCAGTGTGCTC
Fabp-2F: CACAGTGGAGGTGACCTTGG
Fabp-2R: GATGCTCTTGACGTTGCGAC
TubF: CTCAGTGCTCGATGTTGTCC
TubR: GCCAAGGGAGTGTGTGAGTT
Retinal degeneration assay
We performed all retinal degeneration assays in the cn, bw -/- background to eliminate eye pigments that otherwise affect the course of retinal degeneration. The flies were incubated in the 25 °C incubator with 1000 lux of light. For retinal degeneration assays under constant darkness, the flies were reared in an enclosed cardboard box in the 25 °C incubator. Retinal degeneration was assessed based on green fluorescent pseudopupils originating from the Rh1-GFP transgene. We interpreted clear trapezoidal pseudopupils as evidence in intact photoreceptors, while its disappearance was construed as a sign of retinal degeneration. The number of flies analyzed for each genotype in Figure 6E is as follows:
wild type, 48 flies; fabpEY02678, 52 flies; ninaEG69D/+, 50 flies; ninaEG69D, fabpEY02678/ fabpEY02678, 32 flies.
For Figures 6F and G, 50 flies were analyzed for each genotype.
Quantification and statistics
To quantify proteins in gels, we measured average pixel intensities of western blot bands using Image J, and normalized them to anti-β tubulin bands. Graphs were generated after at least three independent measurements and p values were calculated using a paired t-test. For retinal degeneration assays, we used the Log-rank (Mantel-Cox) text. Graphs were made using the Graphpad Prism program. All error bars represent SEM (Standard error of the mean).
Acknowledgements
We thank Drs. Vikke Weake, Jason Gerstner for fly lines and antibodies. This work was supported by the NIH grant R01 EY020866 to H.D.R.