It is commonly assumed that photoreceptor (PR) outer segment (OS) morphogenesis is reliant upon the presence of peripherin/rds, hereafter termed Rds. In this study, we demonstrate a differential requirement of Rds during rod and cone OS morphogenesis. In the absence of this PR-specific protein, rods do not form OSs and enter apoptosis, whereas cone PRs develop atypical OSs and are viable. Such OSs consist of dysmorphic membranous structures devoid of lamellae. These tubular OSs lack any stacked lamellae and have reduced phototransduction efficiency. The loss of Rds only appears to affect the shape of the OS, as the inner segment and connecting cilium remain intact. Furthermore, these structures fail to associate with the specialized extracellular matrix that surrounds cones, suggesting that Rds itself or normal OS formation is required for this interaction. This study provides novel insight into the distinct role of Rds in the OS development of rods and cones.
The mammalian retina is comprised of both rod and cone photoreceptors (PRs), which initiate the phototransduction cascade upon excitation of their visual pigment by a photon of light. In both PR types, the outer segment (OS) is comprised of stacks of membranous discs in rods and lamellae in cones, which house and compartmentalize the proteins used in the phototransduction cascade. It is commonly thought that the proper development of these organelles is directly linked to normal PR cell function and viability; indeed, mutations in proteins specific to the OS (e.g., the rod visual pigment, rhodopsin) cause a multitude of blinding diseases (Molday, 1998). In both PRs, the plasma membrane undergoes further ultrastructural reorganization to form the discs of rod OSs and lamellae of cone OSs (Steinberg et al., 1980; Arikawa et al., 1992). In cones, the membrane lamellae are open and physically contiguous with the plasma membrane, whereas in rods, they become sealed, forming distinct membranous structures (discs) that are separated from the plasma membrane by cytosol. Rod and cone PRs also use redundant and analogous proteins for structural development and phototransduction, and many proteins have a conserved function in both PR cell types (Molday, 1998). The precise mechanism of OS morphogenesis is still a matter of active investigation even though the basic features of the process have been known for nearly 40 yr. However, a role for the PR-specific protein Rds (product of the retinal degeneration slow gene) in this process has been suggested based upon its localization to the disc rim, and in vitro data also suggest a fusogenic role for Rds in OS membrane assembly (Steinberg et al., 1980; Molday et al., 1987; Arikawa et al., 1992; Ritter et al., 2004; Damek-Poprawa et al., 2005).
Rds (also known as peripherin/rds or peripherin-2) is a tetraspanning transmembrane protein that is preferentially expressed in the OSs of rod and cone PRs (Molday et al., 1987; Connell and Molday, 1990; Wrigley et al., 2002). In the rod-dominated wild-type (WT) mouse retina, the loss of Rds causes a failure of OS generation, a greatly diminished response to light, and a slow degeneration of the PR cell layer (Sanyal et al., 1980; Sanyal and Jansen, 1981; Reuter and Sanyal, 1984; Jansen et al., 1987; Travis et al., 1989). However, these observations are limited by the fact that in the WT mouse retina, the PR population is comprised mostly of rods (>95%), making the study of cones difficult in this animal model. Although Rds is clearly requisite for normal rod OS morphogenesis and function, a similar requirement for Rds in cone PRs has, as of yet, not been established. Furthermore, human mutations in Rds manifest as rod or cone dystrophies with varying severity (Kohl et al., 1998; van Soest et al., 1999; Musarella, 2001), suggesting this protein has distinct functions in rod and cone PRs.
Recently, a knockout of neural retina leucine zipper (Nrl−/−) has been described in which rod PRs fail to develop and the retina consists entirely of cone PRs (Mears et al., 2001). Several studies have demonstrated the legitimacy and utility of the Nrl−/− mouse model as an excellent resource for studying cone PRs (Mears et al., 2001; Yoshida et al., 2004; Yu et al., 2004; Daniele et al., 2005; Nikonov et al., 2005). In this study, we took advantage of this model to assess the role of Rds in cone PRs, generating a double knockout mouse that lacked both Nrl and Rds (Nrl−/−/Rds−/−). We report that in the absence of Rds, cones form atypical OSs consisting of distended membranous structures that do not resemble morphologically normal lamellae. This is in striking contrast to rods, where no OSs form at all in the absence of Rds. Furthermore, these noncanonical cone OSs are capable of phototransduction with minimally reduced sensitivity, which is in marked contrast to the phenotype observed in the Rds−/− retina, where rod function is barely detectable. Finally, our results also suggest that Rds has a role in maintaining interactions between the OS and the specialized extracellular matrix that surrounds cone PRs (the so-called cone matrix sheath [CMS]; Hollyfield et al., 1989; Johnson et al., 1989; Hageman et al., 1995).
Localization of Rds in Nrl−/− PRs
To examine the subcellular localization of Rds in cones, we used immunogold cytochemistry and ultrastructural analysis. The majority of Rds immunoreactivity was localized to the rim region of both rod and cone OSs in the WT retina, as previously observed (Fig. 1; Arikawa et al., 1992). A similar pattern of labeling was observed in the cone OSs of the Nrl−/− retina (Fig. 1), establishing the legitimacy of this model for studying Rds function in cones.
Differentiation status of Nrl−/−/Rds−/− PRs
Successful crossbreeding of Nrl−/− and Rds−/− mice was verified with PCR genotyping (not depicted) and quantitative RT-PCR (qRT-PCR; Fig. 2 a). To determine the impact of the loss of Rds on PR differentiation, we used qRT-PCR to examine the mRNA levels of several genes involved in phototransduction and maintenance of OS structure in the adult retina (Fig. 2, b–d). For all of the genes examined, similar expression levels were observed in the retinas of Nrl−/− and Nrl−/−/Rds−/− mice. We detected no expression of rod opsin (Rho), rod transducin (Gnat1), or the rod cyclic nucleotide-gated channel (Cnga1) in either genetic background, and the level of rod phosphodiesterase (Pde6a) was substantially reduced as compared with WT and Rds−/− retinas (Fig. 2 b). In contrast, expression levels of cone phototransduction genes (S-Opsin, Gnat2, Pde6c, and Cngb3) were markedly increased in the Nrl−/−/Rds−/− retinas relative to the WT retinas and were at comparable levels with those observed in the Nrl−/− retinas (Fig. 2 c; Mears et al., 2001; Yoshida et al., 2004; Yu et al., 2004). The expression levels of two retinal genes that produce proteins localized to the disc region of PR OSs were also examined. Rod OS membrane protein (Rom-1) and prominin-1 (Prom-1) were expressed at nearly identical levels in the Nrl−/− and Nrl−/−/Rds−/− retinas (Fig. 2 d), indicating that these genes were not up-regulated to compensate for the loss of Rds.
Nrl−/−/Rds−/− mice are capable of phototransduction
To evaluate retinal function in vivo, we used electroretinography (ERG) to examine the electrical response of the retina to light stimulation, distinguishing between rod and cone responses by varying the illuminance conditions before delivering a test flash. Multiple light intensities were used to stimulate light-adapted (photopic) WT, Nrl−/−, Rds−/−, and Nrl−/−/Rds−/− mice at postnatal day 30 (P30; Fig. 3 a). Serial photopic ERGs demonstrated a complete absence of any electrical signal from the Rds−/− retina regardless of the light intensity. In the WT and Nrl−/− retinas, a b-wave signal was observed at −0.04 log cd sm−2 and enlarged with increasing flash intensities. A similar pattern of waveforms was detected in the Nrl−/−/Rds−/− retina, starting at 0.37 log cd sm−2, which is suggestive of reduced phototransduction sensitivity of nearly 1 log unit. Quantification at 1.89 log cd sm−2 revealed that photopic b-wave amplitudes were undetectable in Rds−/− mice but significantly higher in Nrl−/− and Nrl−/−/Rds−/− mice as compared with the WT control (Fig. 3 b). It is likely that the lack of photopic ERG signal in Rds−/− mice results from either the toxic effect of the rapidly apoptosing rod PRs that is detrimental to the surviving cones or loss of the structural support of the rods. At this intensity of light, Nrl−/−/Rds−/− mice show nearly identical a- and b-waves regardless of whether scotopic (dark adapted) or photopic (light adapted) ERGs were recorded, signifying that the ERG responses of this mouse model originate in the cone PRs (Fig. S1). By 2 mo of age, the Nrl−/− photopic b-wave had decreased significantly, as previously reported (Daniele et al., 2005), and a declining trend in the photopic ERG amplitude of Nrl−/−/Rds−/− was observed throughout the time course examined (Fig. 3 c). At 12 mo of age, the b-wave amplitudes of both strains were not significantly different from those recorded from age-matched WT mice. These data demonstrate that the retina of Nrl−/−/Rds−/− is functionally cone dominated, similar to the Nrl−/− retina, and capable of significant levels of phototransduction, unlike the Rds−/− retina. We then hypothesized that the decreased photopic b-wave amplitude in Nrl−/−/Rds−/− mice compared with that of Nrl−/− mice may represent altered phototransduction efficiency and sensitivity because of a structural abnormality resulting from the loss of Rds.
OS structure of cone PRs lacking Rds
To examine the localization of phototransduction and OS structural proteins, we used immunohistochemistry on retinal sections from P30 WT, Nrl−/−, Rds−/−, and Nrl−/−/Rds−/− mice (Fig. 4 a). Rds immunoreactivity was abundant in the OS of both WT and Nrl−/− retinas but absent from Rds−/− and Nrl−/−/Rds−/− retinas. The Rds-associated protein Rom-1 displayed an identical pattern of labeling as Rds, which is in support of previous observations that Rds is essential for Rom-1 targeting to the OS (Tam et al., 2004). Double labeling with monospecific primary antibodies revealed the presence of rod visual pigment (rhodopsin) and short wavelength cone visual pigment (S-opsin) in WT OSs and in the OS remnants of the Rds−/− retinas; however, retinas from Nrl−/− and Nrl−/−/Rds−/− mice showed only S-opsin immunoreactivity localized in the subretinal space and rosettelike structures and retinal folds. To better define the morphology of the structures labeled by anti–S-opsin, thick sections (25 μM) were examined with confocal microscopy, and rotations of the image stacks were performed (Fig. 4 b). The OS of Nrl−/− retinas showed punctate labeling of tightly packed cone OS lamellae that were mostly aligned with the retinal pigment epithelium (RPE), which is typical of OS structure. However, in the Nrl−/−/Rds−/− retina, a markedly different structure was observed. Instead of normal cone OSs, numerous tubular-shaped structures were located in the subretinal space, representing dysmorphic cone OSs. These structures were also longer than the punctate OS of Nrl−/− retinas and were not aligned with the RPE. Upon rotation of the image stack, a latticelike network of these structures was observed, possibly reflecting interactions between these OSs and the RPE microvilli.
Rds is required for lamellae formation in cone OSs
Retinal histology and ultrastructure were evaluated by light and electron microscopy, respectively, using plastic-embedded tissue sections from P30 WT, Nrl−/−, Rds−/−, and Nrl−/−/Rds−/− mice (Fig. 5). The characteristic heterochromatin clumps distinct to cone PR nuclei was noticeable in both Nrl−/− and Nrl−/−/Rds−/− retinas (Fig. 5). The outer nuclear layer was well organized in the WT and Rds−/− retinas but was disrupted by whorls and rosettelike structures in the Nrl−/− retina and, to a lesser extent, in the Nrl−/−/Rds−/− retina. Nrl−/−/Rds−/− retinas exhibited an undulating morphology, with numerous retinal folds from the outer nuclear layer toward the inner nuclear layer, which failed to form complete rosettelike structures that were observed typically in retinas of Nrl−/− mice. Furthermore, retinas from Nrl−/−/Rds−/− mice were prone to detachment from the RPE along many regions regardless of the fixation method used (not depicted). The lack of distinct rod OSs was apparent in all panels of Fig. 5 except those corresponding to retinas from WT eyes. Ultrastructural analysis demonstrated typical cone OS membrane structures in the Nrl−/− retina; however, dysmorphic membranous structures lacking the characteristic stacked cone OS lamellae were observed in Nrl−/−/Rds−/− retinas. At a higher magnification, these lamellae-less OSs were distended and seemed to conform in shape to the space in which they were located (Fig. 5 b). Large balloon-shaped contiguous membrane structures also were observed, representing cone OSs that have failed to undergo normal morphogenesis. S-opsin immunoreactivity was detected solely in these membranous structures of the Nrl−/−/Rds−/− retina (Fig. 5 c), further confirming their identity as cone OSs.
Rds is required for interactions between the cone OS and the extracellular matrix
To further elucidate the cause of the abnormal morphology observed in the Nrl−/−/Rds−/− retina, immunohistochemistry was performed to visualize PR-connecting cilia and inner segments (ISs) in the retinas of P30 mice (Fig. 6 a). The connecting cilium/axoneme and IS appeared intact in both models, as observed by labeling with antibodies to acetylated α-tubulin and to Na/K-ATPase, respectively. Hence, the loss of Rds resulted in gross structural changes in the OS but appeared to have no morphological effect on other PR structures or cellular compartments. We also examined localization of the CMS in relation to cone OSs (Fig. 6 b). This sheath surrounds the IS and OS of cone PRs to mediate adhesive interactions between the retina and RPE (Hollyfield et al., 1989; Johnson et al., 1989; Hageman et al., 1995). In retinal sections from WT and Nrl−/− mice, the CMS was labeled by fluorescently conjugated peanut agglutinin (PNA) and appeared within the IS and OS layers. Furthermore, S-opsin immunoreactivity was observed within the OS layer where PNA staining was detected. Interestingly, PNA labeling was not detected in the layer containing the dysmorphic cone-derived OS membranes of the Nrl−/−/Rds−/− retina but was detected solely within the IS layer.
In this study, we have demonstrated that the loss of Rds in cone PRs does not affect the differentiation of PRs but causes the formation of morphologically novel distended, membranous OS structures that, nonetheless, are capable of phototransduction. These structures contain S-opsin yet lack the compartmentalization of lamellae, directly implicating Rds in this process during cone OS morphogenesis. Rod PRs lacking Rds fail to form OS structures and possess rhodopsin solely in the tip of the cilium and IS membrane, causing minimal phototransduction activity and subsequent PR degeneration (Reuter and Sanyal, 1984; Nir and Papermaster, 1986). In contrast, cone PRs lacking Rds form altered OS structures lacking normal lamellar organization but, nonetheless, are capable of phototransduction, albeit with reduced sensitivity. Furthermore, the loss of Rds in cones with concomitant abortive lamellar formation prevents the extracellular matrix from forming around the cone OS, further reinforcing the notion that the OS lamellae are required for proper interactions between cone PRs and their extracellular environment.
In these studies, we have used the naturally occurring rds mutant mouse on a C57BL/6 background, and no photopic ERG signal is detectable using our methods. Previous investigations using rds mutant mice on a 020/A genetic background revealed a nominal scotopic ERG that would also include the response of surviving rods (Reuter and Sanyal, 1984). In that study, the ERGs may have been more sensitive, as they were performed by placing a needle electrode into the anterior chamber, whereas our method utilizes a looped platinum electrode placed on the cornea. These differences in genetic background and ERG methodology could explain the variation in results obtained between previous work (Reuter and Sanyal, 1984) and this study.
The data presented here support a model of cone OS membrane morphogenesis that predicts OS lamellae rim formation to be a second stage of morphogenesis after evagination of the plasma membrane from the connecting cilium (Steinberg et al., 1980). A previous study demonstrated ultrastructural localization of Rds in the rim regions in cones opposite to the connecting cilium where the membrane invaginates; however, in the rod PR, the OS plasma membrane is separated from the discs, and Rds localizes to the rim on both sides of the disc (Arikawa et al., 1992). Several studies have also shown the fusogenic properties of Rds (Boesze-Battaglia and Goldberg, 2002; Ritter et al., 2004; Damek-Poprawa et al., 2005), which further implicate its role in disc membrane morphogenesis. Based upon these models, we propose that the loss of Rds in cone PRs causes a morphogenic event in which plasma membrane evagination occurs but invagination fails, resulting in the formation of a dysmorphic OS organelle devoid of lamellae. Interestingly, these tubular-like OS membrane structures observed in the Nrl−/−/Rds−/− retina are consistent with a model of cone OS morphogenesis by which growth of the plasma membrane occurs bidirectionally from the connecting cilium (Eckmiller, 1987). Further studies to reveal the subcellular compartmentalization of phototransduction proteins in this novel structure may reveal the exact purpose for the utilization of the disc membrane shape in the normal (WT) OS. This phenotype also demonstrates an inherently different role for Rds in rod versus cone PRs. It appears that cones only require Rds for membrane pinching to form the OS lamellae; however, in rods, Rds may have an additional role in OS development because its absence results in a more severe morphogenic outcome whereby the OS does not form and the rod PRs undergo apoptosis.
The disruption of CMS interactions with the dysmorphic OS as observed in the Nrl−/−/Rds−/− retina suggests either a direct involvement of Rds in tethering the CMS to the OS or a more generalized dependence on normal OS structure for PR–matrix adhesional competence. The CMS is required for a variety of cell–matrix and cell–cell interactions, including trophic and metabolic interactions with the adjacent RPE (Hollyfield et al., 1989; Johnson et al., 1989; Hageman et al., 1995). The failed establishment of the CMS around these OSs suggests that disc morphogenesis (or normal OS formation) is somehow integrally linked to CMS association. Future biochemical studies to determine the precise role of Rds in the development and maintenance of this association may provide further insights into the compositional and functional differences between rod- and cone-associated extracellular matrix.
These observations also have implications regarding therapeutic strategies for treating human diseases involving Rds mutations that cause diseases specific to cone PRs. It is possible that a complete absence of Rds may be more advantageous than having mutant isoforms of Rds in cone PRs because human mutations in Rds that initially display cone-specific dysfunction could be caused by detrimental protein associations (e.g., aggregation of mutant Rds protein). In this regard, the use of RNA interference methodologies to silence Rds specifically in cone PRs may be a beneficial approach.
Materials And Methods
All mice were bred into and assessed on a C57BL/6 background. All experiments and animal maintenance were approved by the local Institutional Animal Care and Use Committee (Oklahoma City, OK) and conformed to the guidelines on the care and use of animals adopted by the Society for Neuroscience and the Association for Research in Vision and Ophthalmology (Rockville, MD). The Nrl−/− mice were provided by A. Swaroop (University of Michigan Kellogg Eye Center, Ann Arbor, MI).
Total RNA was extracted from the retinas of a single mouse using TRIzol reagent (Invitrogen) and DNase treated with RNase-free DNase I (Promega). Reverse transcription was performed using an oligo-dT primer and Superscript III reverse transcriptase (Invitrogen). Primers for all genes were designed to span introns as to avoid amplification from genomic DNA. Primers for Rds spanned exon 2, where a 9-kb genomic insertion of a viral element causes the loss of Rds in the rds mouse, so amplification only occurred in those samples harboring a WT allele. All primer sequences are available in Table S1 . qRT-PCR was performed in triplicate on each cDNA sample using a real-time PCR detection system (iCycler; Bio-Rad Laboratories), and ΔcT values were calculated against the neuronal housekeeping gene hypoxanthine phosphoribosyltransferase (Hprt). Hprt was assigned an arbitrary expression level of 10,000, and relative gene expression values were calculated by the following calculation: relative expression = 10,000/2ΔcT, where ΔcT = (gene cT – Hprt cT). This was repeated with three independent samples for each genotype, and the mean expression value is presented with the SD. Agarose gel electrophoresis and disassociation curve analysis were performed on all PCR products to confirm proper amplification.
ERG analyses were performed as previously described (Cheng et al., 1997). In brief, after a minimum of 4-h dark adaptation, animals were anesthetized by intramuscular injection of 85 mg/kg ketamine and 14 mg/kg xylazine. For the assessment of scotopic response, a stimulus intensity of 1.89 log cd s m−2 was presented to the dark-adapted dilated eyes in a Ganzfeld (GS-2000; Nicolet). The amplitude of the scotopic a-wave was measured from the prestimulus baseline to the a-wave trough. The amplitude of the b-wave was measured from the trough of the a-wave to the crest of the b-wave. To evaluate photopic response, animals were light adapted for 5 min under a light source of 1.46 log cd m−2 intensity. Afterward, a strobe flash was presented to the dilated eyes in the Ganzfeld with various intensities (−0.99–2.86 log cd s m−2). The amplitude of the photopic b-wave was measured from the trough of the a-wave to the crest of the b-wave. Significance was determined using one-way analysis of variance and post-hoc tests using Bonferroni's pairwise comparisons (Prism, version 3.02; GraphPad).
Tissue fixation and sectioning were performed as previously described (Stricker et al., 2005). In brief, eyes from P30 mice were enucleated and fixed in 4% PFA/PBS for 16 h before paraffin embedding. Tissue sections (10-μm thickness) were obtained with a microtome, deparaffinized, rehydrated as described previously (Nour et al., 2004), and blocked in 5% BSA/PBS for 30 min at RT. Slides were briefly washed with PBS and incubated with the primary antibody in 1× BSA/PBS for 2 h at RT followed by a brief wash in PBS and incubation with the secondary antibody in 1× BSA/PBS for 30 min at RT. After a brief washing in PBS, Vectashield with DAPI (Vector Laboratories) was applied, and the slide was coverslipped. Primary antibodies (with dilutions) and sources were as follows: anti–Rds-CT (1:200); anti-Rom1 (1:200); monoclonal anti-rod opsin (Rho 1D4; 1:1,000) provided by R. Molday (University of British Columbia, Vancouver, Canada; Hicks and Molday, 1986); rabbit anti–mouse S-opsin (1:500), a gift from C. Craft and X. Zhu (Doheny Eye Institute, University of Southern California, Los Angeles, CA); acetylated α-tubulin (1:200) from Sigma-Aldrich; and anti–Na+/K+-ATPase (1:100) from the Developmental Studies Hybridoma Bank (University of Iowa, Iowa City, IA; Lebovitz et al., 1989). All secondary antibodies (FITC or Cy3 conjugates; Jackson ImmunoResearch Laboratories) were applied at a dilution of 1:1,000 from the original stock. Before incubation with antibodies against acetylated α-tubulin and Na+/K+-ATPase, antigen retrieval was performed by incubating slides in 10 mM citrate buffer, pH 3.0, for 30 min at 37°C followed by a brief rinsing in PBS. For PNA staining, AlexaFluor488-conjugated PNA (Invitrogen) was applied at a 1:200 dilution during the incubation with secondary antibody. Sections were viewed at RT with a microscope (Axioskop 50; Carl Zeiss MicroImaging, Inc.) in the autoexpose mode using a 40, 63, or 100× objective. Images were captured with a digital camera (Axiocam HR; Carl Zeiss MicroImaging, Inc.) using Axiovision 3.1 software (Carl Zeiss MicroImaging, Inc.).
Transmission electron microscopy, light microscopy (histology), and plastic-embedment immunogold cytochemistry
Methods used for tissue collection and processing for plastic-embedment light and electron microscopy and immunohistochemistry were as previously described (Stricker et al., 2005). For conventional light and electron microscopy, mice were perfused with 0.1 M sodium phosphate buffer, pH 7.4, containing 2% (vol/vol) PFA and 2% (vol/vol) glutaraldehyde; for plastic-embedment immunohistochemistry, the buffered fixative contained 2% PFA and 0.1% glutaraldehyde. For light microscopy, tissue sections (0.75–1-μm thickness) were viewed and photographed with a photomicroscope (BH-2; Olympus) in the autoexpose mode using a 20 or 60× DplanApo objective, and images were collected with a digital camera system (DXM-1200; Nikon). For electron microscopy, Spur's resin-embedded or (for immunogold) LR White–embedded tissue sections (silver–gold) were viewed with an electron microscope (100EX; JEOL). For immunohistochemistry, primary antibodies (see previous section) were used at a 1:10 dilution; secondary antibodies (AuroProbe 10-nm gold-conjugated goat anti–rabbit IgG; GE Healthcare) were used at a 1:50 dilution.
Online supplemental material
Table S1 presents primers that were designed to generate amplicons of 180–300 bp using Primer3 software (http://frodo.wi.mit.edu/cgi-bin/primer3/primer3_www.cgi). Fig. S1 shows mice that were dark adapted for a minimum of 4 h or light adapted for at least 5 min, and scotopic ERG analyses were performed as described above.
We thank Drs. Muayyad Al-Ubaidi, Alan J. Mears, and Sepideh Zareparsi for critical evaluation of the manuscript and Dr. Anand Swaroop for providing us with breeding pairs of Nrl−/− mice used in this study. We also thank Drs. Robert S. Molday, Cheryl Craft, and Xuemei Zhu for supplying antibodies used in this study.
This study was supported by grants from the National Institutes of Health (EY10609 to M.I. Naash, EY07361 to S.J. Fliesler, and Core Grant for Vision Research EY12190 to M.I. Naash), the Foundation Fighting Blindness (to M.I. Naash), the Norman J. Stupp Foundation Charitable Trust (to S.J. Fliesler), and by an unrestricted departmental grant from Research to Prevent Blindness (to S.J. Fliesler). M.I. Naash is a recipient of the Research to Prevent Blindness James S. Adams Scholar Award.
Abbreviations used in this paper: CMS, cone matrix sheath; ERG, electroretinography; IS, inner segment; OS, outer segment; PNA, peanut agglutinin; PR, photoreceptor; qRT-PCR, quantitative RT-PCR; RPE, retinal pigment epithelium; WT, wild type.