In order to identify compounds that may target Nr2e3, we searched PubChem (https://pubchem.ncbi.nlm.nih.gov/) for top-scoring hit compounds previously identified to interact with Nr2e3 in transfected CHO-S cells in a luciferase-based assay (PubChem Assay IDs: 602229, 624378, 624394, and 651849). As a secondary screen for the initial hits we used primary retinal cell cultures and assayed Rhodopsin expression because it is a well-defined target of Nr2e3 signaling and is expressed at high levels exclusively in rod photoreceptors (Cheng et al., 2004; Peng et al., 2005; Haider et al., 2009). We dissociated retina from postnatal day 5 (P5) mice and cultured them in media containing the small molecules. After treatment for 2 days, we assessed Rhodopsin expression with an immunofluorescence-based assay (Nakamura et al., 2016). One compound, PR3 (Figure 1A), showed robust reduction in Rhodopsin compared to DMSO and PR1 treatment (Figure 1B). We confirmed this finding with qPCR analysis using intact retinal explant cultures from P4 mice treated with a 0.3 μM dose of PR1 or PR3. Similar to the immunofluorescence assay, treatment with PR3 resulted in reduced Rhodopsin but not Otx2, a rod transcription factor upstream of Nr2e3, compared to DMSO and PR1 (Figure 1C).

Figure 1 with 1 supplement see all

Download asset Open asset

Figure 1—source data 1

Source data for Figure 1.

https://doi.org/10.7554/eLife.30577.006

Download elife-30577-fig1-data1-v1.xlsx

Mutations in Nr2e3 result in an increased number of S Opsin+ photoreceptors as well as a reduction in rod gene expression (Haider et al., 2006; Haider et al., 2000; Cheng et al., 2011; Peng et al., 2005; Chen et al., 2005; Cheng et al., 2006). To determine if PR3 treatment also affects cone gene expression, we explanted intact retinas from P11 wild type mice in media containing DMSO or PR3 for 3 days. We used intact retinas for this experiment to assess changes in dorsal and ventral retina independently. After fixation and whole mount immunostaining, we counted S Opsin+ cells in the dorsal and ventral retina. Similar to Nr2e3 mutations, treatment with PR3 resulted in an increase in the number of S Opsin+ cells in the ventral, but not dorsal retina (Figure 1D–E).

PR3 was initially identified as a chemical modulator of Nr2e3 in a luciferase-based assay that identified ligands by disruption of the Nr2e3-NCoR dimer complex and had a calculated IC₅₀ of 0.07 μM in this assay (PubChem Assay IDs: 602229, 624378, 624394, and 651849). To confirm a direct Nr2e3-PR3 interaction, we used isothermal titration calorimetry (ITC). Consistent with our other assays, ITC qualitatively showed a direct interaction between PR3 and Nr2e3 (Figure 1F; estimated K_d of 67 μM using a one site model). The initial screen that identified PR3 demonstrated its effects on Nr2e3 in a co-repression assay with NCoR; however, the effects we observed on rod gene expression suggested that PR3 also inhibits the co-activator function of Nr2e3. To explore this possibility, we assessed the effects of PR3 on the ability of Nr2e3 to cooperate with Nrl and Crx, two other transcription factors known to form a co-activation complex for rod gene expression. We first transfected HEK cells with Nr2e3, Nrl, and Crx and measured activation of the Rhodopsin promoter in the presence of DMSO or PR3. PR3 strongly reduced Rhodopsin promoter activation (Figure 1—figure supplement 1A). We next used co-immunoprecipitation to assess whether PR3 affects the interaction between Nr2e3 and Nrl in HEK 293 cells. We found that PR3 caused an increase in the binding of Nr2e3 to Nrl (Figure 1—figure supplement 1B–C). Several reports have shown that ligand binding to nuclear receptors stabilizes interactions with co-activators by stabilizing the structure of the ligand-binding domain and co-activator binding sites (Onishi et al., 2010; Jia et al., 2009; Forrest and Swaroop, 2012; Fu et al., 2014). It is not clear why stabilizing the complex would reduce its activity; however, it may be that stabilizing interactions with Nr2e3 prevents Nrl and Crx from interacting with consensus sites on the DNA, or alternatively prevents the recruitment of other components of the transcriptional machinery. Nevertheless, our results indicate that PR3 affects the formation or stabilization of the complex among these critical rod gene regulators.

Although PR3 was identified in an assay for Nr2e3, it is also possible that it interacts with other nuclear receptors. Two related nuclear receptors that are expressed in photoreceptors are Errb and Rorb. Deletion of these genes in mice has effects on rod gene expression (Onishi et al., 2010; Jia et al., 2009; Forrest and Swaroop, 2012; Fu et al., 2014). However, small molecule modulators of ERRb induce rapid rod death (Onishi et al., 2010), and so it is unlikely that PR3 acts through this pathway. Loss of function Rorb mutations reduce rod gene expression and lead to a phenotype more like what we observe with PR3 treatment. To determine whether PR3 might also act through antagonism of RORb we transfected HEK 293 cells with RORb and Crx and measured activation of the S Opsin promoter, a known target of RORb. In transfected HEK 293 cells, RORb and Crx synergistically activate the S Opsin promoter (Srinivas et al., 2006; Liu et al., 2017). Treatment with PR3 did have a small effect on RORb-Crx driven activation of the S Opsin promoter (Figure 1—figure supplement 1D); however, this effect was much smaller than what we found for PR3 on activation of the Rhodopsin promoter by Nr2e3, Nrl, and Crx (Figure 1—figure supplement 1A). To further evaluate a role for PR3 as an RORb antagonist, we analyzed downstream RORb target gene expression in developing retina. One target that is specific to RORb, but not regulated by Nr2e3 is Prdm1 (Liu et al., 2017; Wang et al., 2014; Mills et al., 2017). Prdm1 expression was unchanged in P0 retinal explants treated with PR3 (Figure 1—figure supplement 1E), supporting our hypothesis that the effects of PR3 are primarily mediated through modulation of Nr2e3.

To further evaluate whether PR3 has other targets in the retina, we tested PR3 in Rd7 retinas, which harbor a spontaneous mutation in Nr2e3 (Akhmedov et al., 2000; Haider et al., 2001; Chen et al., 2006). We explanted adult (P23-35) retinas from wild type and Rd7 mice in media containing DMSO or PR3 and measured Rhodopsin expression by qPCR. Consistent with the hypothesis that the effects of PR3 on rod gene expression are mediated through Nr2e3, PR3 caused a significant reduction in Rhodopsin expression in wild type retinas but not in retinas from Rd7 mice (Figure 1G). While this experiment confirms Nr2e3 specificity of PR3, it remains unclear why Rd7 retinas exhibit only moderate reductions in rod gene expression (Corbo and Cepko, 2005; Peng et al., 2005; Chen et al., 2005; Haider et al., 2001), while PR3 treatment or CRISPR-Cas9 deletion of Nr2e3 (Zhu et al., 2017) result in substantial decreases. This suggests that there is some developmental compensation that is not present when the gene is deleted or inhibited in postmitotic rods; alternatively, PR3 may be acting in a dominant-negative manner, and inhibiting the ability of the Nr2e3-Nrl-Crx complex to function properly.

Nr2e3 signaling is important for rod photoreceptor cell fate, development and maturation, and maintenance of expression. To determine the effect of PR3 treatment on gene expression in postmitotic retinal cells, we systemically treated (intraperitoneal injection; IP) wild type mice with PR3 or vehicle at P12. At P13, 24 hr after the injection, we collected the retinas for global transcriptome analysis by RNA sequencing. As hypothesized, we observed a decrease in most rod photoreceptor-specific transcripts (Figure 2A–B). Similar to conditional knockout of Nrl in adult mice and knockdown of Nrl by CRISPR/Cas9 in postmitotic photoreceptors, we did not observe global increases in cone gene expression (Montana et al., 2013; Yu et al., 2017). Nevertheless, we did observe an increase in the number of S Opsin+ cells in the ventral retina with PR3 treatment in vivo (Figure 2C–D). Genes expressed in both rod and cone photoreceptors, eg. Crx and Otx2, showed no difference in expression between control and PR3 treatment.

Figure 2

Download asset Open asset

Figure 2—source data 1

Source data for Figure 2.

https://doi.org/10.7554/eLife.30577.008

Download elife-30577-fig2-data1-v1.xlsx

We next used electron microscopy (EM) to perform an ultrastructural analysis of photoreceptor morphology after PR3 treatment. We carried out IP injections of vehicle or PR3 in wild type mice for three consecutive days starting at P11. At P14, 24 hr after the third injection, we euthanized the mice and processed their retinas for EM. Photoreceptor outer segments begin to form during the second and third postnatal week; mutations in Nr2e3 lead to an impairment in rod outer segment formation, and we predicted that PR3 treatment would affect their development in a similar way (Haider et al., 2006). As predicted, PR3 treatment prevented outer segment development; outer segments of PR3-treated photoreceptors were strikingly truncated compared to controls (Figure 2E). We did not observe any indication of photoreceptor apoptosis induced by PR3 treatment upon examination of outer nuclear layer (ONL) nuclei (Figure 2E), further indicating that the effects on rod development were not due to an increase in cell death.

We have recently shown that reductions in rod gene expression caused by treatment with PR1 were sufficient to slow the degeneration of Rho^P23H photoreceptors in vitro (Nakamura et al., 2016). The Rho^P23H mutation causes misfolding of Rhodopsin in rod photoreceptors, which leads to activation of the unfolded protein response and eventually results in rod and cone death (Nguyen et al., 2014). In Rho^P23H mice, most rod photoreceptors undergo apoptosis by the end of the third postnatal week (Sakami et al., 2011).

To determine whether we could prevent photoreceptor degeneration in this model, we treated Rho^P23H mice with PR3 or vehicle from P12-P14 until P21, during the period of rod photoreceptor death (Figure 3A). At P21, we assessed visual function with ERGs and euthanized the mice for histological and qPCR analyses. At P21, control Rho^P23H mice had only 2–3 rows of photoreceptors remaining in their ONL (Figure 3B–C). Rods were sparse and there were few remaining cones (S Opsin+ and Cone Arrestin+ photoreceptors). Control Rho^P23H mice had minimal scotopic and photopic b-wave amplitudes by ERG analysis (Figure 4A, B, D and E). By contrast, retinas from PR3-treated Rho^P23H mice had several rows of rod and cone photoreceptors in the ONL (Figure 3B–C). The surviving cones in the PR3-treated retinas were more elongated and healthier than in the DMSO control retinas. We confirmed our histological results with qPCR on retinas from control and PR3 Rho^P23H mice and found that treated mice had more expression of Recoverin and Rhodopsin, indicating greater photoreceptor cell survival (Figure 3D). ERG analysis of PR3-treated Rho^P23H mice showed significantly elevated scotopic and photopic b-wave amplitudes at most stimulation intensities compared to littermate controls (Figure 4A, C, D and F). Together, these data support the conclusion that PR3 treatment prevented structural and functional degeneration of photoreceptors in this model of RP.

Figure 3

Download asset Open asset

Figure 3—source data 1

Source data for Figure 3.

https://doi.org/10.7554/eLife.30577.010

Download elife-30577-fig3-data1-v1.xlsx

Figure 4

Download asset Open asset

Figure 4—source data 1

Source data for Figure 4.

https://doi.org/10.7554/eLife.30577.012

Download elife-30577-fig4-data1-v1.xlsx

In this study we successfully prevented photoreceptor degeneration in the Rho^P23H mouse, the first report of successful treatment of this RP model with a small molecule in vivo. Our strategy was to reduce the expression of photoreceptor genes by targeting the rod-specific nuclear receptor Nr2e3 with a small molecule modulator. Treatment with PR3 decreased rod gene expression, and was sufficient to functionally and structurally preserve photoreceptors in the Rho^P23H mouse. Previous studies have shown that genetic manipulation of the rod photoreceptor differentiation pathway may be useful for the treatment of multiple RP models. Conditional deletion of Nrl in adult mouse rods prevents degeneration in the Rho^−/− model of recessive RP (Montana et al., 2013). More recently, knockdown of Nrl or Nr2e3 by AAV-CRISPR/Cas9 gave long-term histological and functional preservation of photoreceptors in numerous RP models (Zhu et al., 2017; Yu et al., 2017). Our report now shows that a small molecule targeting this same regulatory pathway is also effective at slowing rod degeneration in a particularly aggressive RP model and provides a novel target for medical therapy of retinal degeneration.

C57Bl/6 (Jackson Stock No: 000664), Rho^P23H (Sakami et al., 2011)(Jackson Stock No: 017628), and Nr2e3^Rd7 (Jackson Stock No: 004643) mice were used at the indicated ages. All mice were housed by the Department of Comparative Medicine at the University of Washington and protocols were approved by the University of Washington Institutional Animal Care and Use Committee. The research was carried out in accordance with the ARVO statement for the Use of Animals in Ophthalmic and Vision Research. For all experiments, a sample size of at least four mice per condition was chosen to ensure adequate power to detect a pre-specified effect size. From each litter, half of the animals were randomly assigned to the control group and the other half to the experimental group, and no animals were excluded.

Photoregulin3 was identified by searching previous small molecule screens with PubChem for Nr2e3 interacting molecules. It was initially obtained from ChemDiv and then synthesized and purified in large quantities in the lab after initial screening. For in vivo experiments, mice were injected intraperitoneally with PR3 dissolved in DMSO at 10 mg/kg.

Retinas were dissected from postnatal day 5 (P5) mice and dissociated by treatment with 0.5% Trypsin diluted in calcium- and magnesium-free HBSS for 10 min at 37°C. Trypsin was inactivated by adding an equal volume of FBS and cells were pelleted by centrifugation at 4°C and resuspended in media (Neurobasal-A containing 1% FBS, 1% N2, 1% B27, 1% Pen/Strep, and 0.5% L-Glutamine). For the immunofluorescence assay, cells were plated into 96-well black walled, clear bottom tissue culture plates at a density of 1 retina/5 wells. Small molecules were diluted in media and were added the day following dissociation. After two days of treatment, cells were fixed with 4% PFA for 20 min at room temperature, blocked with blocking solution (10% Normal Horse Serum and 0.5% Triton X-100 diluted in 1X PBS) for 1 hr at room temperature, and incubated overnight at 4°C with primary antibodies generated against Rhodopsin (1:250; Rho4D2, Gift from Dr. Robert Molday, UBC) diluted in blocking solution. The following day, wells were washed with 1X PBS and then incubated with species appropriate, fluorescently labeled secondary antibodies diluted in blocking solution for 1 hr at room temperature. Wells were washed three times, counterstained with ToPro3, and the entire plate was imaged using a GE Typhoon FLA 9400 imager. Optical density measurements were obtained from the plate scans using ImageJ software and Rhodopsin expression was normalized to ToPro3 nuclear stain.

RNA from retinas was isolated using TRIzol (Invitrogen) and cDNA was synthesized using the iScript cDNA synthesis kit (Bio-Rad). SSO Fast (Bio-Rad) was used for quantitative real-time PCR. For analysis, values were normalized to Gapdh (ΔC_T) and ΔΔC_T between DMSO and compound-treated samples was expressed as percent of DMSO treated controls (100*2^ΔΔCt). t-tests were performed on ΔC_T values. The following primer sequences were used: Gapdh (F: GGCATTGCTCTCAATGACAA, R: CTTGCTCAGTGTCCTTGCTG), Rhodopsin (F: CCCTTCTCCAACGTCACAGG, R: TGAGGAAGTTGATGGGGAAGC), Opn1sw (F: CAGCATCCGCTTCAACTCCAA, R: GCAGATGAGGGAAAGAGGAATGA), Recoverin (F: ACGACGTAGACGGCAATGG, R: CCGCTTTTCTGGGGTGTTTT), Otx2 (F: CCGCCTTACGCAGTCAATG, R: GAGGGATGCAGCAAGTCCATA), and Prdm1 (F: TTCTCTTGGAAAAACGTGTGGG, R: GGAGCCGGAGCTAGACTTG).

Intact retinas without RPE from mice at the indicated ages and genotypes were explanted on 0.4 μm pore tissue culture inserts in media (Neurobasal-A containing 1% FBS, 1% N2, 1% B27, 1% Pen/Strep, and 0.5% L-Glutamine) containing DMSO or 0.3–1.0 μM PR3. Full media changes were performed every other day.

Explants or eye cups were fixed with 4% PFA for 20 min at room temperature, blocked with blocking solution (10% normal horse serum and 0.5% Triton X-100 diluted in 1X PBS) for 1 hr at room temperature, and incubated overnight at 4°C with primary antibodies generated against S Opsin (1:400, SCBT, sc-14363). The following day, the retinas were washed with 1X PBS, and then incubated with a species appropriate, fluorescently-labeled secondary antibody diluted in blocking solution overnight, followed by washing with 1X PBS and DAPI staining. The retinas were transferred to slides and coverslipped with Fluoromount-G (SouthernBiotech). An Olympus FluoView FV1000 was used for confocal microscopy. Cells were counted from single plane confocal images taken at fixed settings.

Eyecups were fixed in 4% PFA in 1X PBS for 20 min at room temperature and then cryoprotected in 30% sucrose in 1X PBS overnight at 4°C. Samples were embedded in OCT (Sakura Finetek), frozen on dry ice, and then sectioned at 16–18 μm on a cryostat (Leica). Slides were blocked with a solution containing 10% normal horse serum and 0.5% Triton X-100 in 1X PBS for 1 hr at room temperature and then stained overnight at 4°C with primary antibodies (Rho4D2 at 1:250 from Dr. Robert Molday, S Opsin at 1:400 from SCBT: sc-14363, Cone Arrestin at 1:1000 from Millipore: AB15282, Otx2 at 1:200 from R and D Systems: BAF1979) diluted in blocking solution. Slides were washed three times with 1X PBS the following day and then incubated in fluorescently labeled secondary antibodies diluted in blocking solution for 2 hr at room temperature, stained with DAPI, washed, and coverslipped using Fluoromount-G (SouthernBiotech). An Olympus FluoView FV1000 was used for confocal microscopy. Cells were counted from single plane confocal images taken at fixed settings. Counts in the central retina were taken adjacent to the optic nerve head (50 μm from the nerve head on the ventral side) and counts in the peripheral retina were taken 50 μm from the peripheral edge on the ventral side.

NR2E3 protein (aa 90–410) was expressed as an His8-MBP-TEV fusion protein from the expression vector pVP16 (DNASU Plasmid ID: HsCD00084154). E. coli BL21 (DE3) cells were grown to an OD600 of 1, and then induced with 0.2 mM IPTG at 16°C overnight. Cells were harvested, resuspended in extract buffer (20 mM Tris pH 8, 200 mM NaCl, 10% glycerol, 5 mM 2-mercaptoethanol, and saturated PMSF diluted 1:1,000), and then lysed by sonication on ice. Lysates were centrifuged at 4°C and the supernatant was loaded onto an equilibrated column containing 5 mL of Ni-NTA agarose (Qiagen). The column was washed with 20 mM Tris pH 8, 1M NaCl, 5 mM 2-mercaptoethanol, and 40 mM imidazole, and then the protein was eluted with 20 mM Tris pH 8, 200 mM NaCl, 5 mM 2-mercaptoethanol, and 100 mM imidazole. The fusion protein was incubated with TEV overnight at 4°C and then the His8-MBP tags were separated from NR2E3 by ion exchange chromatography. For isothermal titration calorimetry, 100 μM PR3 was injected into 20 μM NR2E3 in 10 mM Sodium Phosphate buffer pH 8 with 50 mM NaCl and 0.5% DMSO in a MicroCal ITC-200 (Malvern) and the data was analyzed with Origin 7.0 software.

RNA from retinas was isolated using TRIzol (Invitrogen) and total RNA integrity was checked using an Agilent 4200 TapeStation and quantified with a Trinean DropSense96 spectrophotometer. RNA-seq libraries were prepared from total RNA using the TruSeq RNA Sample Prep kit (Illumina) and a Sciclone NGSx Workstation (PerkinElmer). Library size distributions were validated using an Agilent 4200 TapeStation. Additional Library quality control, blending of pooled indexed libraries, and cluster optimization were performed using Life Technologies’ Invitrogen Qubit Fluorometer. RNA-seq libraries were pooled (4-plex) and clustered onto a flow cell lane. Sequencing was performed using an Illumina HiSeq 2500 in rapid mode employing a paired-end, 50 base read length (PE50) sequencing strategy.

Mice were euthanized by CO₂, and then perfused with 0.9% saline followed by 4% glutaraldehyde in 0.1 M sodium cacodylate buffer. Eye cups were fixed in 4% glutaraldehyde in 0.1 M sodium cacodylate buffer, washed with 0.1 M sodium cacodylate buffer, and then post-fixed in 2% osmium tetroxide. After fixation, eye cups were washed with water, dehydrated through a graded series of ethanol, incubated in propylene oxide and then epon araidite, polymerized overnight at 60°C, and then sectioned at a thickness of 70 nm. Images were obtained using a JEOL JEM-1230 electron microscope.

Mice were dark adapted overnight (12–18 hr). All subsequent steps were carried out under dim red light. Mice were placed in an anesthesia chamber and anesthetized with 1.5–3% isoflurane gas. Mice were transferred from the anesthesia chamber to a heated platform maintained at 37 ˚C and positioned in a nose cone to maintain a constant flow of Isoflurane. Drops of 1% tropicamide and 2.5% phenylephrine Hydrochloride were applied to each eye. A reference needle electrode was placed subdermally on the top of the head and a ground needle electrode was placed subdermally in the tail. Drops of 1.5% methyl cellulose were applied to each eye and contact lens electrodes were placed over each eye.

Dim red light was turned off and the platform was positioned inside of an LKC Technologies UTAS BigShot ganzfeld and a series of flashes of increasing intensity were delivered scotopically. A series of photopic flashes were performed immediately following the series of scotopic flashes.

HEK293 cells were transfected with 1 μg of the luciferase reporter BR-225Luc (Dr. Shiming Chen) or S Opsin 600 pGL3 (Dr. Douglas Forrest), 1 ng of the control pRL-CMV (Promega) and 100 ng of hNRL-pCMVSport6 (Open Biosystems), hCRX-pCMVSport6 (Open Biosystems), hNR2E3-pcDNA3.1/HisC (Dr. Shiming Chen), or hRORB-pLenti6.2/V5-DEST (DNASU) in 24-well plates using Lipofectamine 3000 reagent (Thermo Fisher Scientific). Transfection reagents were removed the following day and replaced with media containing DMSO or 1 μM PR3 for 2 days. Cells were lysed and firefly and renilla luciferase activity was measured with the Dual-Luciferase Reporter Assay System (Promega) using a 1420 Multilabel Victor3V plate reader.

HEK293 cells were transfected in 6-well tissue culture plates with lipofectamine 3000 (Thermo Fisher Scientific) and 800 ng of each hNRL-pCMVSport6 (Open Biosystems) and hNR2E3-pcDNA3.1/HisC (Dr. Shiming Chen) in Opti-MEM media. Transfection reagents were removed after 24 hr and replaced with media containing DMSO or 1 μM PR3 for 2 days. Cells were lysed with Co-IP lysis buffer (25 mM Tris-HCl pH 7.5, 150 mM NaCl, 1 mM EDTA, 1% Triton X-100, 5% glycerol and 1X protease inhibitor cocktail). Sheep anti-mouse IgG magnetic Dynabeads were incubated with anti-Nr2e3 antibody (5 μg/precipitation) diluted in Co-IP buffer for 2 hr at 4°C. Equal volume of lysate were then added to the antibody-coated beads and incubated overnight at 4°C. The following day, beads were washed four times with Co-IP buffer and then incubated at 85°C for 15 min in 1X sample buffer diluted in Co-IP buffer. Samples were loaded and run in a 4–20% SDS gel (Bio-Rad). Protein was transferred to a PVDF membrane (Thermo Fisher Scientific), blocked (5% BSA and 0.1% Tween 20 in 1X PBS) for at least 1 hr at room temperature and stained with primary antibodies anti-Nrl (Chemicon) or Anti-Nr2e3 (R and D systems) diluted in blocking solution overnight at 4°C. Membranes were washed with 0.1% Tween 20 in 1X PBS and then incubated with Clean-Blot IP Detection Reagent (Thermo Fisher Scientific) diluted in blocking solution for 1 hr at room temperature. Signals were visualized on X-ray film with SuperSignal West Dura Extended Duration Substrate (Thermo Fisher Scientific) and quantified using ImageJ software.

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.

The authors acknowledge the support of NIH R01 EY021374 (to T Reh, S Ding and K Zhang), 1PO1 GM081619-01 (T Reh) and the Allen Distinguished Investigator Award (T Reh). The authors thank members of the Reh and Berminghan-McDonogh labs for helpful critique, and Dr. John Sumida and members of Drs. Jason Smith and Wenqing Xu labs for technical assistance and advice, and Dr. Timothy Cherry for comments on the manuscript. The authors also thank Jonathan Linton from the Vision Core for assistance with ERGs.

Animal experimentation: All experiments were performed in strict accordance with the protocols approved by the University of Washington Institutional animal care and use committee (IACUC).

© 2017, Nakamura et al.

This article is distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use and redistribution provided that the original author and source are credited.