The FASEB Journal • Research Communication
Rim formation is not a prerequisite for distribution of cone photoreceptor outer segment proteins Shannon M. Conley, Muayyad R. Al-Ubaidi, Zongchao Han, and Muna I. Naash1 Department of Cell Biology, University of Oklahoma Health Sciences Center, Oklahoma City, Oklahoma, USA Retinal degeneration slow (RDS/PRPH2) is critical for the formation of the disc/lamella rim in photoreceptor outer segments (OSs), but plays a different role in rods vs. cones. Without RDS, rods fail to form OSs, however, cones lacking RDS (in the rdsⴚ/ⴚ/ Nrlⴚ/ⴚ) exhibit balloon-like OSs devoid of lamellae. We show that distribution of most proteins in the lamella and PM domains is preserved even in the absence of RDS, rim, and lamella structures. However, the rim protein prominin-1 exhibits altered trafficking and OS localization, suggesting that proper targeting and distribution of rim proteins may require RDS. Our ultrastructural studies show that in cones, OS formation is initiated by the growth of opsin-containing membrane with RDS-mediated rim formation as a secondary step. This is directly opposite to rods and significantly advances our understanding of the role of the rim in cone OS morphogenesis. Furthermore, our results suggest that the unique folded lamella architecture of the cone OS may maximize density or proximity of phototransduction proteins, but is not required for OS function or for protein distribution and retention in different membrane domains.—Conley, S. M., Al-Ubaidi, M. R., Han, Z., Naash, M. I. Rim formation is not a prerequisite for distribution of cone photoreceptor outer segment proteins. FASEB J. 28, 3468 –3479 (2014). www.fasebj.org ABSTRACT
Key Words: RDS 䡠 degeneration 䡠 morphogenesis 䡠 retina The photoreceptor outer segment (OS) comprises stacks of membranous discs/lamellae circumscribed by a hairpin-like rim. Rod discs are sealed and enclosed by the plasma membrane (PM), while cone lamellae are contiguous with the PM. Lamellae have one portion of their circumference/rim surrounded by PM (i.e., closed),
Abbreviations: CC, connecting cilium; CMS, cone matrix sheath; CNG, cyclic nucleotide gated; COS, cone outer segment; ECM, extracellular matrix; IF, immunofluorescence; IRBP, interphotoreceptor retinoid-binding protein; IS, inner segment; NP, nanoparticle; ONL, outer nuclear layer; OS, outer segment; P, postnatal day; PM, plasma membrane; PNA, peanut agglutinin; PRPH2, peripherin-2; RDS, retinal degeneration slow; ROM-1, rod outer segment membrane protein 1; RP1, retinitis pigmentosa 1; RPE, retinal pigment epithelium; TEM, transmission electron microscopy; WT, wild type 3468
while the remainder is exposed or “open” to the extracellular space (1, 2). Proteins localize to different domains in these OSs: the PM, the discs/lamellae, and the rim. For example, opsins are found in the discs/ lamellae, the cyclic nucleotide gated (CNG) channel is found in the PM, and the tetraspanin protein retinal degeneration slow [RDS; also known as peripherin-2 (PRPH2), Fig. 1A] is found in the rim. RDS is required for rim formation (3, 4), and is thought to promote membrane curvature, flattening, and fusion (5–7). The closed portion of the cone rim is similar in orientation to the rod rim (Fig. 1B) and contains RDS. However, the open portion of the rim curves in the opposite direction (with regard to the orientation of proteins in the membrane, Fig. 1B), and RDS is unlikely to be there, since the curvature is backwards from that promoted by RDS. Instead, recent studies in Xenopus laevis have shown that the cholesterol-binding protein prominin-1 (also known as CD133) is found along these open lamellar edges in cones, while RDS resides on opposing closed rims (8). Tetraspanins are known to organize membrane microdomains known as the tetraspanin web (9), and it has been hypothesized that RDS performs a similar function in photoreceptors. In mice lacking RDS (rds⫺/⫺), no rod OSs are formed. To understand what happens to cones in the absence of RDS, we used the Nrl⫺/⫺ mouse model in which developing rods convert to S-cone-like cells (10, 11). In contrast to rods lacking RDS, cones in the rds⫺/⫺/Nrl⫺/⫺ model develop open OSs that lack lamellae and rim structures (12). These data suggest that RDS and rim formation are differently required for OS biogenesis and function in rods vs. cones. Although RDS is required for rim formation in rods and cones, little is known about the mechanisms of rim formation or the precise role of rim formation during OS morphogenesis. Two different models of rod morphogenesis have been proposed (13, 14), but the differences between them remain unresolved, and they 1 Correspondence: OUHSC, Department of Cell Biology, 940 Stanton L. Young Blvd. BMSB 781, Oklahoma City, OK 73104. E-mail: [email protected] doi: 10.1096/fj.14-251397 This article includes supplemental data. Please visit http:// www.fasebj.org to obtain this information.
0892-6638/14/0028-3468 © FASEB
Figure 1. OS ultrastructure and membrane organization. A) RDS is a tetraspanin protein with 4 transmembrane domains, a large extracellular (cones)/intradiscal (rods) loop (EC2/D2), and cytoplasmic N/C termini. The EC2/D2 loop has a highly conserved (green) region and a hypervariable region (orange) and is glycosylated (Y-shaped hexagons). B) In rods, RDS is located along the circumference of the disc, with the glycosylated EC/D2 loop pointing toward the intradiscal space, and is thought to play a part in mediating rim curvature. In cones, RDS is found in a similar localization along the closed edge of the lamellae (i.e., portion surrounded by PM), but the curvature of the open edge of the discs is the opposite direction (with regard to the orientation of the proteins in the membrane) as closed discs and is not likely to contain RDS. Dark gray, PM; light gray, lamella membrane. Opposing membrane leaflets are solid/dashed lines. C, D) Immunogold labeling of cone OSs (COSs) was undertaken on retinal sections collected at P30 using antibodies against S-opsin. Shown are representative structures from WT and rds⫺/⫺ (C), and Nrl⫺/⫺ and rds⫺/⫺/Nrl⫺/⫺ (D). Scale bar ⫽ 500 nm.
do not directly address rim formation. Even less is known about cone morphogenesis. Here we take advantage of the simplified OSs of cones lacking RDS to study the role of rim formation in OS morphogenesis, and particularly in the processes that govern the formation and maintenance of discrete OS membrane domains. We show that despite the lack of lamellae, these cone OSs (COSs) retain distribution of OS proteins in distinct membrane domains, and that reintroduction of RDS leads to OS structures with intermediate attempts at rim/lamella formation.
MATERIALS AND METHODS Animal husbandry Double-knockout rds⫺/⫺/Nrl⫺/⫺ mice were generated from rds⫺/⫺ mice (generously provided by Dr. Neeraj Agarwal, National Eye Institute, Bethesda, MD, USA), and Nrl⫺/⫺ mice (generously provided by Dr. Anand Swaroop, National Eye Institute). Mice were housed in a 12-h dark-light cycle with light intensity ⬃30 lux. Animal breeding and experimental CONE OUTER SEGMENT MORPHOGENESIS
protocols were approved by the University of Oklahoma Health Sciences Center Institutional Animal Care and Use Committee and are consistent with those recommended by the Association for Research in Vision and Ophthalmology. Subretinal injection and processing Mice (rds⫺/⫺/Nrl⫺/⫺) at postnatal d 30 (P30) were subretinally injected as described previously (15) with compacted DNA nanoparticles (NPs) carrying wild-type (WT) mouse RDS gene (NMP) under the control of the human interphotoreceptor retinoid-binding protein (hIRBP) promoter, which drives expression in rods and cones (16, 17). These NPs were generated as described previously (18), and have been shown to effectively drive RDS gene expression in photoreceptors as soon as 2 d postinjection (15, 16). At d 7 postinjection, eyes were enucleated, fixed, and processed for immunofluorescence (IF) or transmission electron microscopy (TEM) as described below. IF Mice at P30 were euthanized, and eyes were enucleated, fixed, embedded, and cryosectioned (10 or 25 m sections), and slides were then processed for IF as described previously 3469
(19). Briefly, sections were either incubated in ice-cold methanol at ⫺20°C for 20 min or immediately rehydrated, then treated with 1% sodium borohydride (Sigma Aldrich, St. Louis, MO, USA) for 2–5 min, washed with water and phosphate buffered saline (PBS), blocked (5% BSA, 2% donkey serum, 1% fish gelatin, and 1% TX-100 in PBS), then incubated with the primary antibodies overnight at 4°C. Slides were then washed with PBS and incubated with the secondary antibodies for 1 h at room temperature, followed by additional washing and mounting using ProlongGold with DAPI (Life Technologies, Grand Island, NY, USA). Primary antibodies are as follows: retinitis pigmentosa 1 (RP1) at 1:1000 dilution (generously provided by Dr. Eric Pierce, Massachusetts Eye and Ear Infirmary, Boston, MA, USA; ref. 20); acetylated ␣-tubulin at 1:100 dilution (T7451; SigmaAldrich); cone S-opsin at 1:500 dilution (H-17; Santa Cruz Biotechnology, Santa Cruz, CA, USA); CNGA3 at 1:250 dilution (generously provided by Dr. Xi-Qin Ding, University of Oklahoma Health Sciences Center; ref. 21); Na⫹K⫹ATPase at 1:500 dilution (a5-c, Developmental Studies Hybridoma Bank, Iowa City, IA, USA); GNAT2 at 1:500 dilution (sc-390; Santa Cruz Biotechnology); RDS mAB 2B7, RDS-CT, ROM-1 at 1:1000 dilution (generated in house and previously characterized; refs. 19, 22); cone Na⫹K⫹Ca2⫹ exchanger at 1:200 dilution (NCKX2; generously provided by Dr. Jonathan Lytton, University of Calgary Health Sciences Center, Calgary, Canada; ref. 23); prominin-1 at 1:200 dilution (13A4, eBioscience, San Diego, CA, USA); M-opsin at 1:35,000 dilution (generously provided by Dr. Cheryl Craft, University of Southern California, Los Angeles, CA, USA; ref. 24). Secondary antibodies were AlexaFluor 488-, 555-, or 647-conjugated donkey IgGs (Life Technologies), used at 1:1000 dilution. AlexaFluor 488- or 555-conjugated peanut agglutinin (PNA; Life Technologies) was used to visualize the cone matrix sheath (CMS). For each antibody combination, the IF experiment was repeated ⱖ3 times using eyes from 3–5 independent animals. Microscope image acquisition All images were captured using a Olympus BX-62 coupled to a spinning disk confocal unit (Olympus DSU; Olympus, Tokyo, Japan). All images were captured at room temperature using either a ⫻60 oil objective with a numerical aperture of 1.42 or a ⫻100 oil objective with a numerical aperture of 1.4. Images were captured at room temperature using a Hamamatsu DCAM-API camera (Hamamatsu Photonics, Hamamatsu, Japan). Image acquisition was done in Slidebook 4.2.0.3 (Intelligent Imaging Innovations, Inc., Denver, CO, USA), using the confocal filter set. Z stacks were deconvolved using the nearest neighbors algorithm, and 3D reconstructions were done using the 3D volume view command in Slidebook. All images shown (except 3D reconstructions of stacks) are single planes from the captured stack and are not collapsed or projection images. Figures were assembled in Adobe Photoshop CS5 (Adobe Systems, San Jose, CA, USA). TEM and immunogold labeling Eyes for electron microscopy were fixed and embedded as described previously (25). Thin sections of ⬃800 Å were collected on copper 75/300 mesh grids, then further processed and stained. Embedding and sectioning for immunogold labeling was done as described previously (19, 26). Thin sections were collected on nickel 75/300 mesh grids. Primary antibodies were anti-S-opsin rabbit polyclonal, used at 1:10 (generously shared by Dr. Cheryl Craft, University of Southern California, Los Angeles, CA, USA), and secondary anti3470
Vol. 28
August 2014
bodies (1:50) were AuroProbe 10 nm gold-conjugated goat anti-rabbit IgG; (GE/Amersham, Piscataway, NJ, USA). All were imaged using a Jeol 100CX electron microscope (Jeol Ltd., Akishima, Japan) at 60 keV accelerating voltage (25).
RESULTS Open COSs are formed in the absence of RDS In WT mouse retinas, only ⬃3–5% of photoreceptors are cones, making them quite difficult to study. We have observed that RDS appears to have different roles in rods vs. cones and can cause both rod- and conedominant disease, piquing interest in the role of RDS and the rim in COS biogenesis and function. We therefore generated the rds⫺/⫺/Nrl⫺/⫺ model and observed the aforementioned lamella-free OSs (12). Strikingly, these OSs retained ⬃50% of photopic ERG response, indicating that the highly modified folded lamella ultrastructure was not required for phototransduction. This divergence between the anatomical and physiological behavior of cones lacking RDS prompted us to undertake more detailed studies of the characteristics of these rim free OSs. Our first step here was to determine whether cones lacking RDS in the WT background also adopt this open OS structure. We performed immunogold labeling with antibodies against S-opsin (to identify COSs) coupled with TEM. As expected, in the WT and Nrl⫺/⫺, S-opsin is highly concentrated in the folded lamellae. In contrast, in both the rds⫺/⫺ and rds⫺/⫺/Nrl⫺/⫺, S-opsin immunoreactivity is present on unfolded, membranous OS structures that lack the normal flattened lamellae seen in the WT (Fig. 1C, D). These results confirmed that lamella-free COSs form as a genuine outcome of RDS deficiency, in both the WT and Nrl⫺/⫺ models. Targeting of COS proteins to the PM/lamellae is unaffected by RDS deficiency The photoreceptor OS is connected to the inner segment (IS) by the connecting cilium (CC). Photoreceptor proteins are synthesized in the IS, and those destined for the OS are then transported to their final destination through the CC. Because of the open membranous structures adopted by the COSs in the absence of RDS, we wanted to confirm the integrity of the CC. We therefore colabeled retinal sections from adult WT, rds⫺/⫺, Nrl⫺/⫺, and rds⫺/⫺/Nrl⫺/⫺ mice with the ciliary marker, RP1 (green; ref. 20 and Fig. 2A, B), and the axoneme marker, acetylated ␣-tubulin (red; Fig. 2A, B). All IF images are single planes from a confocal stack unless indicated otherwise. The Nrl⫺/⫺ and rds⫺/⫺/Nrl⫺/⫺ retinas are characterized by rosettes, in which the outer nuclear layer (ONL) forms the outer layer, with ISs and OSs protruding toward the center of the rosette (Fig. 2A, right panels, R; refs. 10, 12). To identify cones in the WT and rds⫺/⫺, sections were counterstained with the cone matrix marker, PNA (blue). We observed that in both the presence and
The FASEB Journal 䡠 www.fasebj.org
CONLEY ET AL.
Figure 2. The connecting cilia are maintained in the rds⫺/⫺/Nrl⫺/⫺. A) IF was used to evaluate the localization of two different cilia markers, RP1 (green) and acetylated ␣-tubulin (red), in the indicated genotypes. PNA (blue) labeling of cone extracellular matrix was used to identify cones in the WT and rds⫺/⫺ backgrounds. B) Higher-magnification views of the CC. Arrows indicate CC and extending axonemes. DAPI (gray) marks nuclei. C) TEM images of the indicated genotypes were used to examine CC ultrastructure in the presence and absence of RDS. The CC is intact and similar in structure in all examined genotypes. CC, connecting cilium; IS, inner segment; ONL, outer nuclear layer; R, rosette. Scale bars ⫽ 10 m (A, B); 500 nm (C).
absence of RDS, cilia are formed and RP1 colocalized with acetylated ␣-tubulin in the region of the CC (Fig. 2A, B, arrows). Ultrastructural examination (Fig. 2C) confirmed the presence of connecting cilia, microtubules, and basal bodies in all models. Having established that the structural connection between the IS and OS is maintained in the open, balloon-like COSs (Fig. 1D, right panel), we turned our attention to evaluating whether OS proteins were properly trafficked from the IS to the OS. We evaluated S-opsin (red; Fig. 3A–D), a marker for OS lamellae and the cone CNG channel (red; Fig. 3E–H), a marker of OS PM. We colabeled retinal sections with antibodies against RP1 to label the CC (green; Fig. 3A, C, E, G), Na⫹K⫹ATPase to label the ISs (blue, Fig. 3A–C, E–G; and green, Fig. 3D, H). We observed that both S-opsin and CNGA3 properly reached the OS in the rds⫺/⫺/ Nrl⫺/⫺; the IS marker did not colocalize with these OS proteins, and while some colocalization was observed between RP1 and the OS proteins, substantial S-opsin and CNGA3 labeling was detected beyond the CC (i.e., toward the center of the rosette). To confirm that the OS proteins were not stuck in the IS, we assembled 3D reconstructions from 25 m confocal stacks and CONE OUTER SEGMENT MORPHOGENESIS
observed that the majority of CNGA3 and S-opsin (red; Fig. 3D, H) labeling did not colocalize with Na⫹K⫹ATPase (green; Fig. 3D, H) in either the Nrl⫺/⫺ or the rds⫺/⫺/Nrl⫺/⫺; only a small amount of colocalization was observed at the most apical region of the IS, likely corresponding to newly synthesized protein. Similar to the case in the Nrl⫺/⫺ we also observed proper OS targeting in WT and rds⫺/⫺ retinas (Fig. 3C, G). Distribution of COS proteins to different membrane subdomains is preserved in the absence of RDS Having established that gross OS protein targeting is not impaired in cones lacking RDS, we next evaluated distribution of proteins to specific subdomains in the open OSs of RDS-free cones. We used S-opsin as a marker for the lamella subdomain (green; Fig. 4), while CNGA3 was used as a marker for the PM (red; Fig. 4). Little is known about how these discrete subdomains are formed and maintained in cones or how their distinct protein content is retained. Although these two markers sort to different membrane domains, they exhibited overlapping labeling in the Nrl⫺/⫺ (and WT; not shown) when visualized using light microscopy 3471
Figure 3. Proteins are properly targeted to the OS in cones lacking RDS. A–C, E–G) IF with the ciliary marker, RP1 (green), the IS marker, Na⫹K⫹ATPase (blue), and either S-opsin (red, A–C) or CNGA3 (red, E–G). Nuclei are counterstained with DAPI (gray). B, F) Higher-magnification images from A (B) and E (F). C, G) Retinal sections from the WT and rds⫺/⫺. Arrows indicate cones. D, H) 3D reconstructions of image stacks captured from 25 m sections in which IF was used to label the IS marker Na⫹K⫹ATPase (green) and either the lamella marker S-opsin (red, D) or the PM marker CNGA3 (red, H). Nuclei are counterstained in DAPI (blue). Grid dimensions and parameters are indicated in each individual image. R, rosette. Scale bars ⫽ 10 m.
(Fig. 4A, top) because in normal cones, the lamellae are partially enclosed by the PM, and the two domains are quite close to each other (as seen in Fig. 1D). Remarkably, in the rds⫺/⫺/Nrl⫺/⫺, S-opsin and CNGA3 did not colocalize in the expanded, lamella-free COS structures (Fig. 4A, bottom). To confirm that this separation was not an artifact of the rosettes, we assembled 3D reconstructions of confocal stacks collected in both rosettes and the region neighboring the retinal pigment epithelium (RPE; Fig. 4B). The two membrane markers were present in completely separate layers within the open OS structures of the rds⫺/⫺/Nrl⫺/⫺ retina; with S-opsin present in a layer distal (i.e., toward the RPE or toward the inside of the rosette) to that where CNGA3 was detected. To determine whether this preserved localization to distinct membrane domains in the absence of lamellae/RDS was a general phenomenon or merely limited to S-opsin and CNG, we undertook labeling for a series of other proteins that target to those two membrane domains. We first examined another integral mem3472
Vol. 28
August 2014
brane lamella protein, M-opsin, which is coexpressed with S-opsin in most murine cones (27). We observed extensive colocalization of M- and S-opsins in all models (Supplemental Fig. S1A, B) indicating that M-opsin was sorted to the same membrane domain as S-opsin. We next assessed whether nonintegral membrane proteins, which are nevertheless associated with a specific membrane domain, retained their distribution. Cone transducin (GNAT2) is peripherally anchored to the lamella membrane by acyl and farnesyl groups so that it can be in proximity to opsin (28). In WT and Nrl⫺/⫺ cones, GNAT2 (green; Supplemental Fig. S1C, D) was found in the lamella subdomain with S-opsin (red; Supplemental Fig. S1C, D), both proteins colocalized when visualized under light microscopy. Similarly, in the rds⫺/⫺ and rds⫺/⫺ Nrl⫺/⫺ we observed that GNAT2 properly localized to the layer containing S-opsin (Supplemental Fig. S1C, D, bottom). Unfortunately, M-opsin, GNAT2, and CNGA3 antibodies are all rabbit polyclonals, so colabeling with them was not possible. These results confirm that lamella-associated proteins
The FASEB Journal 䡠 www.fasebj.org
CONLEY ET AL.
domain as CNG, a layer distinct from that occupied by lamella proteins. We next examined the localization of a final PM protein, the cone Na⫹K⫹Ca2⫹ exchanger (23, 29) which is involved in the regulation of OS cation concentrations and calcium levels (23, 30, 31). Surprisingly, in the rds⫺/⫺/Nrl⫺/⫺, the Na⫹K⫹Ca2⫹ exchanger (red; Fig. 5C, D) localized to the S-opsin layer, unlike CNGA3 and PNA, indicating that only specific PM proteins retain their specific distribution in the absence of lamellae. Prominin-1 partially codistributes with S-opsin and CNG In addition to the PM and lamella membrane domains, properly assembled OSs have a third distinct membrane domain, the rim, which forms the perimeter of
Figure 4. OS membrane distribution is preserved in the rds⫺/⫺/Nrl⫺/⫺. A) IF was used to evaluate lamella (S-opsin, green) and PM (CNGA3, red) protein localization at P30 in the indicated genotypes. Nuclei are counterstained with DAPI (gray). Images show single planes from confocal stacks. Scale bar ⫽ 10 m. B) 3D reconstruction of a 25 m section from indicated genotypes labeled with S-opsin (green) and CNGA3 (red). Nuclei are counterstained with DAPI (blue). Grid sizes are shown in the images. R, rosette.
distributed together to a domain distinct from PM proteins even in the absence of RDS and structural lamellae/rims. We next assessed the distribution of other PM-associated proteins (i.e., in addition to CNG). One of the features of the COS PM is that it interacts with the cone extracellular matrix (ECM), known as the CMS. In normal cones, both in the WT (e.g., in Fig. 2A) and Nrl⫺/⫺, the matrix surrounds the OS/IS, shown by imaging PNA (a green CMS marker; Fig. 5A, top) and S-opsin or CNGA3 (blue and red, respectively; Fig. 5A, arrows and inset). In the rds⫺/⫺/Nrl⫺/⫺, S-opsin labeling is found in a distinct layer from PNA (Fig. 5B; observe blue and green layers in the left panel); however, CNGA3 colocalizes with PNA (Fig. 5A, B, middle panels; observe yellow). These results suggest that PM proteins in these open COS structures maintain interactions with the CMS and are found in the same CONE OUTER SEGMENT MORPHOGENESIS
Figure 5. Localization of additional PM markers in the rds⫺/⫺/Nrl⫺/⫺. A, B) IF on P30 retinal sections from Nrl⫺/⫺ (A) and rds⫺/⫺/Nrl⫺/⫺ (B) was used to evaluate localization of the CMS (PNA, green) relative to the two COS membrane microdomains, lamella (S-opsin, blue) and PM (CNGA3, red). Nuclei are counterstained with DAPI (gray). Arrows indicate regions of colocalization. Insets: Higher-magnification views. C, D) IF was performed on P30 retinal sections from Nrl⫺/⫺ (C) and rds⫺/⫺/Nrl⫺/⫺ (D). Na⫹K⫹Ca2⫹ exchanger is shown in red and S-opsin in green in COSs of Nrl⫺/⫺ and rds⫺/⫺/Nrl⫺/⫺ retinas. Nuclei are counterstained with DAPI (blue). Arrows indicate areas of colocalization. R, rosette. Scale bars ⫽ 10 m. 3473
the disc, and we were interested in the distribution of other rim proteins, such as RDS’ nonglycosylated homologue, rod outer segment membrane protein-1 (ROM-1), and the cholesterol binding protein prominin-1, in the absence of RDS. In rods, the entire circumference of the rim is surrounded by PM, while in cones, a portion of the rim is open. It has been shown that in amphibian cones, RDS is found in the portion of the rim that is surrounded by PM, while prominin-1 is found in the open portion of the rim (8). ROM-1 is virtually undetectable in the absence of RDS (26); however, we assessed the localization of prominin-1 in the Nrl⫺/⫺ and rds⫺/⫺/Nrl⫺/⫺ retinas in relation to other OS markers. We colabeled with prominin-1 (red; Fig. 6) and S-opsin (lamellae, green; Fig. 6A) or CNGA3 (PM, green; Fig. 6B). Interestingly, we observed partial colocalization between prominin-1 and both S-opsin and CNGA3 (Fig. 6A, B, yellow arrows), indicating that prominin-1 does not localize specifically to the lamellae or PM domain. However, many areas expressed S-opsin or CNGA3 but not prominin-1 (in both the Nrl⫺/⫺ and rds⫺/⫺/Nrl⫺/⫺), an observation which suggests that prominin-1 does not distribute uniformly throughout the OS (i.e., there are some parts of the OS that lack prominin-1).
Because we did not see complete colocalization of prominin-1 with the OS markers CNGA3 and S-opsin, we wanted to confirm that it was actually being targeted to the OS in the absence of RDS. We therefore colabeled with prominin-1 (red) and the IS marker Na⫹K⫹ATPase (green; Fig. 6C, D). Interestingly, we observed that a substantial amount of prominin-1 accumulated in the cone IS in the absence of RDS (Fig. 6C, arrows; magnified in Fig. 6E), although some was also detected in the OS (Fig. 6C, arrowheads; magnified in Fig. 6E). Prominin-1 was also mislocalized in rod and cone photoreceptors lacking RDS in the WT background (Fig. 6D, magnified in Fig. 6E). To determine whether this mislocalization was due to the lack of RDS or merely the presence of malformed OS, we examined prominin-1 localization in the rhodopsin knockout retina (rho⫺/⫺). While cones in this model are fairly normal at early time points, rods have only tiny sacs of PM in the OS with no normal discs (although RDS and ROM-1 are present in the OS; ref. 32). We did not observe accumulation of prominin-1 in the IS of the rho⫺/⫺ retina (Fig. 6D, bottom panels; magnified in Fig. 6E) suggesting that prominin-1 mislocalization in the rds⫺/⫺ and rds⫺/⫺/Nrl⫺/⫺ may arise due to the absence of RDS and/or the rim structure.
Figure 6. Prominin-1 partially colocalizes with lamella and PM proteins in the absence of RDS but also accumulates in the IS. A–C) IF was used to evaluate the distribution of prominin-1 in the Nrl⫺/⫺ and rds⫺/⫺/Nrl⫺/⫺ (red) in relation to the lamellae (A, S-opsin, green), PM (B, CNGA3, green), and IS (C, Na⫹K⫹ATPase, green). D) WT, rds⫺/⫺, and rho⫺/⫺ retinal sections were labeled for prominin-1 (red) and the IS marker Na⫹K⫹ATPase. E) Higher-magnification views from C, D. Nuclei are counterstained with DAPI (blue). Arrows indicate areas of colocalization; arrowheads indicate areas of prominin-1 that do not colocalize with the IS marker. R, rosette. Scale bars ⫽ 10 m. 3474
Vol. 28
August 2014
The FASEB Journal 䡠 www.fasebj.org
CONLEY ET AL.
Reintroduction of RDS leads to lamella and rim formation Combined, the above results suggested that RDS is necessary for the formation of the rim region and morphogenesis of proper cone ultrastructure, but not for the formation or maintenance of distinct protein domains within that structure. We next asked what the outcome of reintroduction of RDS would be on cone morphogenesis and membrane organization. We took two approaches to this question; first, we examined the cones of the rds⫹/⫺/Nrl⫺/⫺; and second, we subretinally injected rds⫺/⫺/Nrl⫺/⫺ mice with compacted DNA NPs
carrying a photoreceptor-specific RDS expression vector (IRBP-NMP, previously characterized in ref. 16). Prior to assessing the structure of these retinas, we confirmed that both RDS and ROM-1 were detected in rds⫹/⫺/Nrl⫺/⫺ and IRBP-NMP injected rds⫺/⫺/Nrl⫺/⫺ (blue and red; Fig. 7A). On the ultrastructural level, the COSs of the Nrl⫺/⫺ exhibited nicely stacked, flattened discs (Fig. 7B, C, pink arrows and lines) aligned along the axoneme (Fig. 7B, C, orange arrows and lines), while the rds⫺/⫺/ Nrl⫺/⫺ exhibited balloon-like OSs lacking rims or lamellae. In the Nrl⫺/⫺, we observe a mix of closed (i.e., surrounded by PM; Fig. 7B, C, blue arrows and lines)
Figure 7. RDS supplementation initiates lamella formation in the rds⫺/⫺/Nrl⫺/⫺. RDS supplementation was performed genetically (rds⫹/⫺/Nrl⫺/⫺) or by subretinal injection of IRBP-NMP NPs carrying WT RDS (IRBP-NMP in rds⫺/⫺/Nrl⫺/⫺ mice). A) IF was performed using antibodies against RDS (blue) and ROM-1 (red) with DAPI labeling of nuclei (gray). Arrows indicate areas of colocalization (magenta). B) TEM images of COSs in the indicated genotypes or treatment groups. C) Diagrams representing the hypothetical morphology observed in these groups. Colors correspond to arrows in the TEM images. Scale bars ⫽ 10 m (A); 2 m (B). R, rosette. CONE OUTER SEGMENT MORPHOGENESIS
3475
and open rims (i.e., not surrounded by PM; Fig. 7B, C, purple arrows and lines). We hypothesize that these open rims could be the sites of prominin-1 localization, however, a conclusion on this point must await the availability of a mammalian prominin-1 antibody that works on immunogold cytochemistry. The rds⫹/⫺/ Nrl⫺/⫺ OSs showed sworls of membrane that nonetheless had pinched and flattened rims (Fig. 7B, black arrows). Similarly, in the rds⫺/⫺/Nrl⫺/⫺ retinas that were injected with IRBP-NMP, membrane sworls and flattened rims (Fig. 7B, black arrows) were also detected, but the lamellae were much larger than in the Nrl⫺/⫺ or rds⫹/⫺/Nrl⫺/⫺. While we would predict that the flattened rims in the rds⫹/⫺/Nrl⫺/⫺ and IRBP-NMPinjected rds⫺/⫺/Nrl⫺/⫺ would come in both open and closed varieties as in the normal cone, the highly aberrant sworl structure and lack of clearly identifiable PM make it difficult to conclusively determine the status of these rims. Consistent with the presence of rims and lamellae (albeit improperly sized and stacked) in the COSs of the rds⫹/⫺/Nrl⫺/⫺ and IRBP-NMP injected rds⫺/⫺/Nrl⫺/⫺, CNGA3, RDS, and S-opsin exhibited partial colocalization (similar to the Nrl⫺/⫺, not shown).
DISCUSSION Here we present evidence demonstrating that the segregation of PM and lamella membrane domains is preserved in the open OS of cones lacking RDS (Fig. 8A). We show that cone opsins and other lamella markers occupy a distinct area of COS than the PM protein CNGA3 in these unfolded COSs of the rds⫺/⫺/Nrl⫺/⫺. We further show that a marker of the CMS, PNA, colocalizes with PM proteins (such as CNG) rather than with markers of the lamellae, consistent with predicted interactions between the PM proteins and the components of the ECM. We also show that all studied proteins except the rim protein prominin-1 are properly targeted from the IS to the COS. PM, rim, and lamella proteins are thought to traffic independently in rods (33–35). Studies on trafficking in cones are also being done, however, less is known about the processes which govern cone OS (COS) targeting (36). We show that prominin-1 accumulates in the ISs of photoreceptors lacking RDS (although some is found in the OS as well). These data suggest that RDS or rim structures may be required for proper OS targeting of prominin-1. Historically, the question of whether rim proteins traffic together has been difficult to answer, since the other primary rim protein, ROM-1, is virtually undetectable in the absence of RDS. However our data here suggest that rim proteins, even if they localize to different regions of the rim (open vs. closed) may traffic together, although definitive results on this must await additional study. A primary role of tetraspanins as a protein family is the organization of functional membrane microdomains (9, 37), and RDS may play a similar role in the 3476
Vol. 28
August 2014
Figure 8. Model of OS protein distribution and morphogenesis. A) Diagram summarizing the distribution of OS membrane markers in WT and RDS-free cones. B) Hypothetical model of COS morphogenesis showing 6 steps in the formation of normal COSs, and the final equivalent in the absence of RDS. Step 1: initial PM covering of nascent connecting cilium/axoneme. Step 2: accumulation of opsin-containing membrane (pink). Step 3: accumulation of rim proteins (blue) and formation of the rim. Step 4: lamella growth and alignment with the axoneme. Steps 5 and 6: formation and growth of additional lamellae to result in traditional stacked morphology.
photoreceptor. However, RDS also plays a structural role as it is required for the formation of the physical rim (4, 12). Here we show that introducing RDS into the lamella-free OS of the rds⫺/⫺/Nrl⫺/⫺, either genetically or by subretinal injection of DNA NPs, leads to the formation of a whorl of flattened membrane structures reminiscent of lamellae, although they are larger than normal lamellae. This size anomaly is likely due to an insufficient quantity of RDS; as the amount of RDS
The FASEB Journal 䡠 www.fasebj.org
CONLEY ET AL.
is critical for proper lamella formation and OS morphogenesis in cones and rods (38). The role of RDS in promoting flattening of discs/lamellae is further highlighted by in vitro studies showing that RDS expression induces a flattened morphology in microsomal vesicles (6). Further, we have shown that transgenic mice expressing a mutant form of RDS (C150S) which cannot form normal higher-order oligomeric complexes exhibit open, nonflattened lamellae in their COSs (19). However, empirically it is quite difficult to distinguish between functions of RDS tied to formation of the physical rim structure and those tied to gathering rim proteins to these structures. Tracking COS membrane organization and the ultrastructure of COS in models that lack RDS or have varying amounts of RDS has led us to hypothesize a model of COS biogenesis in which rim formation is not the essential initiating step (in contrast to rods, which form no OSs without rims/RDS). We hypothesize that lamella growth and formation is initiated as cone-opsin carrying membrane (Fig. 8B, pink, step 2) accumulates in a region of the PM (Fig. 8B, black) primed for membrane growth (Fig. 8B, green). Subsequently, RDScontaining membrane (Fig. 8B, blue, steps 3–5) initiates rim curvature/attachment to the axoneme and flattened, stacked lamellae are formed (Fig. 8B, step 6). Final sizing of the lamellae is likely governed by a combination of several factors, including ROM-1 and the actin cytoskeleton (39, 40). In the rds⫺/⫺/Nrl⫺/⫺, the lack of RDS would therefore lead simply to continued growth of the portion of membrane containing lamella proteins without formation of rim or lamella structures (Fig. 8B, bottom, pink). Notably, this region of membrane, which we have termed “primed for growth,” may contain the lipid and/or protein components that are responsible for maintaining the separate distribution of PM and lamella proteins. In the rds⫺/⫺/ Nrl⫺/⫺ this results in a separate layer of PM/lamella proteins, while in the normal cone (Nrl⫺/⫺), this membrane would be found in the open portion of the cone rims, while the RDS-containing rims would be found along the closed portion of the lamellae (Fig. 8B, green and blue, respectively). The fraction of open rim varies; in amphibians, the majority of the cone rim is open (2), while in mammals, a much smaller portion is open, and many discs are completely enclosed (as in rods; refs. 1, 41). The conversion from open to closed discs in mammalian cones appears to be dynamic, with lamellae undergoing episodic fusion with the PM over time (1), but it is not clear what the functional significance of these differences are. The observation that lamella-associated proteins (Sopsin and GNAT2) localize to a distinct membrane domain from the PM-associated protein CNGA3 suggests that the distribution of some proteins in discrete COS membrane domains does not require lamellae, rims, or RDS. However, we also show that, unlike CNG, the Na⫹K⫹Ca2⫹ exchanger, a PM protein, does not segregate specifically to the PM domain. Instead, it distributes throughout the COS in the rds⫺/⫺/ CONE OUTER SEGMENT MORPHOGENESIS
Nrl⫺/⫺. These data indicate that the distribution of all OS PM proteins is not governed in the same way, and suggest the presence of multiple sorting processes, some of which are preserved in the absence of lamellae and some of which are not. Interestingly, the rod Na⫹K⫹Ca2⫹ exchanger has been shown to directly interact with the rod CNG ␣ subunit (30, 42, 43), however, this interaction is absent in cones (44), suggesting that in addition to multiple sorting pathways in cones, there may also be differences in the way proteins are arranged in the two cell types. The next logical question is how some proteins are retained in separate domains (PM vs. lamella). We have shown that lamellae/RDS are not universally required for this process, however several potential mechanisms exist. First, the lipid content of the membrane may play a role in maintaining protein segregation. Second, other membrane proteins or scaffold/cytoskeletal proteins may help keep the domains segregated. Finally, the ECM may play a critical role. We show that the CMS marker PNA colocalizes with the PM protein CNGA3, and interactions between the PM and the ECM could be critical for maintaining the segregated protein distribution we observe. The extent to which each of these (or other unknown mechanisms) may contribute to our observation is not clear and may be the focus of further study. While we have no direct evidence for the identity of the components which facilitate the separation of discrete domains, some evidence suggests that prominin-1 may be one such component. First, we show that in the rds⫺/⫺/Nrl⫺/⫺ S-opsin and CNGA3 exhibit only partial colocalization with prominin-1; that is, there are extensive lamella (S-opsin) and PM (CNGA3) membrane regions that do not coincide with prominin-1 labeling, suggesting that prominin-1 may be present at the barrier between the two membrane domains. Second, while results from amphibians do not universally translate to mammals, X. laevis cones clearly exhibit prominin-1 along lamella edges open to the extracellular space and RDS and the axoneme marker acetylated ␣-tubulin along the opposite edge (8). Third, evidence from the prominin-1 knockout mouse shows that prominin-1 is not required for membrane folding or rim formation, but is required for proper disc/lamella growth and alignment (45). Finally, prominin-1 has some molecular features which make it likely to be involved in membrane organization. It binds cholesterol and localizes to specific lipid microdomains (46). More importantly, it always localizes to curved membrane regions, for example microvilli in epithelial cells or protrusions of the PM (such as lamellipodia) in nonepithelial cells (47). Future studies will focus on clearly testing these hypotheses regarding morphogenesis and the potential role of prominin-1 in mammalian cones. This study demonstrates that the folded lamella ultrastructure of the COS and the presence of RDS are not required for COS membrane organization and basic morphogenesis. Combined with our previous work showing that lamellae/RDS are not required for 3477
cone phototransduction in the rds⫺/⫺/Nrl⫺/⫺, these data suggest that cones and rods are significantly different in their dependence on the highly modified OS ultrastructure and in the involvement of the rim in the process of morphogenesis, since rods without RDS fail to form even nascent OS discs and exhibit no signs of phototransduction. Understanding the biological processes which underlie cone photoreceptor OS and protein distribution and morphogenesis is essential from both a cell biological standpoint and for our understanding of the pathology of RDS-associated disease. This work was supported by the U.S. National Eye Institute (EY10609, M.I.N.; EY018512, S.M.C.); the Foundation Fighting Blindness (M.I.N. and M.R.A.); the Knights Templar Eye Foundation (Z.H.), and the Oklahoma Center for the Advancement of Science and Technology (S.M.C., Z.H., and M.I.N.). The authors thank Mr. Justin Burnett, Ms. Barb Nagel, and Ms. Reem Alomer for their technical assistance. The compacted DNA NPs were generated in collaboration with Dr. Mark Cooper at Copernicus Therapeutics, Inc. (Cleveland, OH, USA). Antibodies (as detailed in Materials and Methods) were generously shared by Dr. Cheryl Craft (University of Southern California, Los Angeles, CA, USA), Dr. Jonathan Lytton (University of Calgary Health Sciences Center, Calgary, AB, Canada), Dr. Xi-Qin Ding (University of Oklahoma Health Sciences Center), Dr. Eric Pierce (Massachusetts Eye and Ear Infirmary, Boston, MA, USA), and Dr. David Garbers (University of Texas Southwestern Medical Center, Dallas, TX, USA). Mice were generously shared by Dr. Anand Swaroop, (Nrl⫺/⫺, National Eye Institute, Bethesda, MD, USA), and Dr. Neeraj Agarwal (rds⫺/⫺, U.S. National Institutes of Health, Bethesda, MD, USA). The authors declare no conflicts of interest.
2. 3. 4.
5. 6. 7. 8.
9. 10.
3478
Anderson, D. H., Fisher, S. K., and Steinberg, R. H. (1978) Mammalian cones: disc shedding, phagocytosis, and renewal. Invest. Ophthalmol. Vis. Sci. 17, 117–133 Eckmiller, M. S. (1987) Cone outer segment morphogenesis: taper change and distal invaginations. J. Cell Biol. 105, 2267– 2277 Molday, R. S., Hicks, D., and Molday, L. (1987) Peripherin. A rim-specific membrane protein of rod outer segment discs. Invest. Ophthalmol. Vis. Sci. 28, 50 –61 Arikawa, K., Molday, L. L., Molday, R. S., and Williams, D. S. (1992) Localization of peripherin/rds in the disk membranes of cone and rod photoreceptors: relationship to disk membrane morphogenesis and retinal degeneration. J. Cell Biol. 116, 659 – 667 Conley, S. M., Cai, X., and Naash, M. I. (2008) Nonviral ocular gene therapy: assessment and future directions. Curr. Opin. Mol. Ther. 10, 456 –463 Wrigley, J. D., Ahmed, T., Nevett, C. L., and Findlay, J. B. (2000) Peripherin/rds influences membrane vesicle morphology. Implications for retinopathies. J. Biol. Chem. 275, 13191–13194 Boesze-Battaglia, K. (2000) Fusion between retinal rod outer segment membranes and model membranes: functional assays and role for peripherin/rds. Methods Enzymol. 316, 65–86 Han, Z., Anderson, D. W., and Papermaster, D. S. (2012) Prominin-1 localizes to the open rims of outer segment lamellae in Xenopus laevis rod and cone photoreceptors. Invest. Ophthalmol. Vis. Sci. 53, 361–373 Hemler, M. E. (2001) Specific tetraspanin functions. J. Cell Biol. 155, 1103–1107 Mears, A. J., Kondo, M., Swain, P. K., Takada, Y., Bush, R. A., Saunders, T. L., Sieving, P. A., and Swaroop, A. (2001) Nrl is
Vol. 28
August 2014
12.
13. 14. 15.
16.
17. 18. 19.
20. 21.
22.
REFERENCES 1.
11.
23.
24.
25.
26.
27.
28.
29.
required for rod photoreceptor development. Nat. Genet. 29, 447–452 Daniele, L. L., Lillo, C., Lyubarsky, A. L., Nikonov, S. S., Philp, N., Mears, A. J., Swaroop, A., Williams, D. S., and Pugh, E. N., Jr. (2005) Cone-like morphological, molecular, and electrophysiological features of the photoreceptors of the Nrl knockout mouse. Invest. Ophthalmol. Vis. Sci. 46, 2156 –2167 Farjo, R., Skaggs, J. S., Nagel, B. A., Quiambao, A. B., Nash, Z. A., Fliesler, S. J., and Naash, M. I. (2006) Retention of function without normal disc morphogenesis occurs in cone but not rod photoreceptors. J. Cell Biol. 173, 59 –68 Chuang, J. Z., Zhao, Y., and Sung, C. H. (2007) SARA-regulated vesicular targeting underlies formation of the light-sensing organelle in mammalian rods. Cell 130, 535–547 Steinberg, R. H., Fisher, S. K., and Anderson, D. H. (1980) Disc morphogenesis in vertebrate photoreceptors. J. Comp. Neurol. 190, 501–508 Cai, X., Conley, S. M., Nash, Z., Fliesler, S. J., Cooper, M. J., and Naash, M. I. (2010) Gene delivery to mitotic and postmitotic photoreceptors via compacted DNA nanoparticles results in improved phenotype in a mouse model of retinitis pigmentosa. FASEB J. 24, 1178 –1191 Cai, X., Nash, Z., Conley, S. M., Fliesler, S. J., Cooper, M. J., and Naash, M. I. (2009) A partial structural and functional rescue of a retinitis pigmentosa model with compacted DNA nanoparticles. PLoS One 4, e5290 Han, Z., Conley, S. M., Makkia, R. S., Cooper, M. J., and Naash, M. I. (2012) DNA nanoparticle-mediated ABCA4 delivery rescues Stargardt dystrophy in mice. J. Clin. Invest. 122, 3221–3226 Farjo, R., Skaggs, J., Quiambao, A. B., Cooper, M. J., and Naash, M. I. (2006) Efficient non-viral ocular gene transfer with compacted DNA nanoparticles. PLoS One 1, e38 Chakraborty, D., Conley, S. M., Stuck, M. W., and Naash, M. I. (2010) Differences in RDS trafficking, assembly and function in cones versus rods: insights from studies of C150S-RDS. Hum. Mol. Genet. 19, 4799 –4812 Liu, Q., Zuo, J., and Pierce, E. A. (2004) The retinitis pigmentosa 1 protein is a photoreceptor microtubule-associated protein. J. Neurosci. 24, 6427–6436 Fitzgerald, J. B., Malykhina, A. P., Al-Ubaidi, M. R., and Ding, X. Q. (2008) Functional expression of cone cyclic nucleotidegated channel in cone photoreceptor-derived 661W cells. Adv. Exp. Med. Biol. 613, 327–334 Chakraborty, D., Ding, X. Q., Fliesler, S. J., and Naash, M. I. (2008) Outer segment oligomerization of Rds: evidence from mouse models and subcellular fractionation. Biochemistry (Mosc.) 47, 1144 –1156 Li, X. F., Kiedrowski, L., Tremblay, F., Fernandez, F. R., Perizzolo, M., Winkfein, R. J., Turner, R. W., Bains, J. S., Rancourt, D. E., and Lytton, J. (2006) Importance of K⫹-dependent Na⫹/Ca2⫹-exchanger 2, NCKX2, in motor learning and memory. J. Biol. Chem. 281, 6273–6282 Zhu, X., Brown, B., Li, A., Mears, A. J., Swaroop, A., and Craft, C. M. (2003) GRK1-dependent phosphorylation of S and M opsins and their binding to cone arrestin during cone phototransduction in the mouse retina. J. Neurosci. 23, 6152–6160 Stricker, H. M., Ding, X. Q., Quiambao, A., Fliesler, S. J., and Naash, M. I. (2005) The Cys214 –⬎Ser mutation in peripherin/ rds causes a loss-of-function phenotype in transgenic mice. Biochem. J. 388, 605–613 Chakraborty, D., Ding, X. Q., Conley, S. M., Fliesler, S. J., and Naash, M. I. (2009) Differential requirements for retinal degeneration slow intermolecular disulfide-linked oligomerization in rods versus cones. Hum. Mol. Genet. 18, 797–808 Craig, S. E., Calinescu, A. A., and Hitchcock, P. F. (2008) Identification of the molecular signatures integral to regenerating photoreceptors in the retina of the zebra fish. J. Ocul. Biol. Dis. Infor. 1, 73–84 Bigay, J., Faurobert, E., Franco, M., and Chabre, M. (1994) Roles of lipid modifications of transducin subunits in their GDPdependent association and membrane binding. Biochemistry (Mosc.) 33, 14081–14090 Schwarzer, A., Kim, T. S., Hagen, V., Molday, R. S., and Bauer, P. J. (1997) The Na/Ca-K exchanger of rod photoreceptor exists as dimer in the plasma membrane. Biochemistry (Mosc.) 36, 13667–13676
The FASEB Journal 䡠 www.fasebj.org
CONLEY ET AL.
30.
31. 32.
33.
34. 35.
36.
37. 38.
39.
Bauer, P. J. (2002) Binding of the retinal rod Na⫹/Ca2⫹-K⫹ exchanger to the cGMP-gated channel indicates local Ca(2⫹)signaling in vertebrate photoreceptors. Ann. N. Y. Acad. Sci. 976, 325–334 Kaupp, U. B., and Seifert, R. (2002) Cyclic nucleotide-gated ion channels. Physiol. Rev. 82, 769 –824 Lem, J., Krasnoperova, N. V., Calvert, P. D., Kosaras, B., Cameron, D. A., Nicolo, M., Makino, C. L., and Sidman, R. L. (1999) Morphological, physiological, and biochemical changes in rhodopsin knockout mice. Proc. Natl. Acad. Sci. U. S. A. 96, 736 –741 Fariss, R. N., Molday, R. S., Fisher, S. K., and Matsumoto, B. (1997) Evidence from normal and degenerating photoreceptors that two outer segment integral membrane proteins have separate transport pathways. J. Comp. Neurol. 387, 148 –156 Kizhatil, K., Baker, S. A., Arshavsky, V. Y., and Bennett, V. (2009) Ankyrin-G promotes cyclic nucleotide-gated channel transport to rod photoreceptor sensory cilia. Science 323, 1614 –1617 Nishimura, D. Y., Fath, M., Mullins, R. F., Searby, C., Andrews, M., Davis, R., Andorf, J. L., Mykytyn, K., Swiderski, R. E., Yang, B., Carmi, R., Stone, E. M., and Sheffield, V. C. (2004) Bbs2-null mice have neurosensory deficits, a defect in social dominance, and retinopathy associated with mislocalization of rhodopsin. Proc. Natl. Acad. Sci. U. S. A. 101, 16588 –16593 Craig, S. E., Thummel, R., Ahmed, H., Vasta, G. R., Hyde, D. R., and Hitchcock, P. F. (2010) The zebrafish galectin Drgal1-l2 is expressed by proliferating Muller glia and photoreceptor progenitors and regulates the regeneration of rod photoreceptors. Invest. Ophthalmol. Vis. Sci. 51, 3244 –3252 Hemler, M. E. (2005) Tetraspanin functions and associated microdomains. Nat. Rev. Mol. Cell Biol. 6, 801–811 Nour, M., Ding, X. Q., Stricker, H., Fliesler, S. J., and Naash, M. I. (2004) Modulating expression of peripherin/rds in transgenic mice: critical levels and the effect of overexpression. Invest. Ophthalmol. Vis. Sci. 45, 2514 –2521 Clarke, G., Goldberg, A. F., Vidgen, D., Collins, L., Ploder, L., Schwarz, L., Molday, L. L., Rossant, J., Szel, A., Molday, R. S.,
CONE OUTER SEGMENT MORPHOGENESIS
40. 41. 42. 43. 44.
45.
46.
47.
Birch, D. G., and McInnes, R. R. (2000) Rom-1 is required for rod photoreceptor viability and the regulation of disk morphogenesis. Nat. Genet. 25, 67–73 Vaughan, D. K., and Fisher, S. K. (1989) Cytochalasin D disrupts outer segment disc morphogenesis in situ in rabbit retina. Invest. Ophthalmol. Vis. Sci. 30, 339 –342 Young, R. W. (1971) The renewal of rod and cone outer segments in the rhesus monkey. J. Cell Biol. 49, 303–318 Kim, T. S., Reid, D. M., and Molday, R. S. (1998) Structurefunction relationships and localization of the Na/Ca-K exchanger in rod photoreceptors. J. Biol. Chem. 273, 16561–16567 Schwarzer, A., Schauf, H., and Bauer, P. J. (2000) Binding of the cGMP-gated channel to the Na/Ca-K exchanger in rod photoreceptors. J. Biol. Chem. 275, 13448 –13454 Matveev, A. V., Quiambao, A. B., Browning Fitzgerald, J., and Ding, X. Q. (2008) Native cone photoreceptor cyclic nucleotidegated channel is a heterotetrameric complex comprising both CNGA3 and CNGB3: a study using the cone-dominant retina of Nrl⫺/⫺ mice. J. Neurochem. 106, 2042–2055 Zacchigna, S., Oh, H., Wilsch-Brauninger, M., Missol-Kolka, E., Jaszai, J., Jansen, S., Tanimoto, N., Tonagel, F., Seeliger, M., Huttner, W. B., Corbeil, D., Dewerchin, M., Vinckier, S., Moons, L., and Carmeliet, P. (2009) Loss of the cholesterol-binding protein prominin-1/CD133 causes disk dysmorphogenesis and photoreceptor degeneration. J. Neurosci. 29, 2297–2308 Roper, K., Corbeil, D., and Huttner, W. B. (2000) Retention of prominin in microvilli reveals distinct cholesterol-based lipid micro-domains in the apical plasma membrane. Nat. Cell Biol. 2, 582–592 Weigmann, A., Corbeil, D., Hellwig, A., and Huttner, W. B. (1997) Prominin, a novel microvilli-specific polytopic membrane protein of the apical surface of epithelial cells, is targeted to plasmalemmal protrusions of non-epithelial cells. Proc. Natl. Acad. Sci. U. S. A. 94, 12425–12430 Received for publication February 14, 2014. Accepted for publication April 7, 2014.
3479
Rim formation is not a prerequisite for distribution of cone photoreceptor outer segment proteins Shannon M. Conley, Muayyad R. Al-Ubaidi, Zongchao Han, and Muna I. Naash1 Department of Cell Biology, University of Oklahoma Health Sciences Center, Oklahoma City, Oklahoma, USA Retinal degeneration slow (RDS/PRPH2) is critical for the formation of the disc/lamella rim in photoreceptor outer segments (OSs), but plays a different role in rods vs. cones. Without RDS, rods fail to form OSs, however, cones lacking RDS (in the rdsⴚ/ⴚ/ Nrlⴚ/ⴚ) exhibit balloon-like OSs devoid of lamellae. We show that distribution of most proteins in the lamella and PM domains is preserved even in the absence of RDS, rim, and lamella structures. However, the rim protein prominin-1 exhibits altered trafficking and OS localization, suggesting that proper targeting and distribution of rim proteins may require RDS. Our ultrastructural studies show that in cones, OS formation is initiated by the growth of opsin-containing membrane with RDS-mediated rim formation as a secondary step. This is directly opposite to rods and significantly advances our understanding of the role of the rim in cone OS morphogenesis. Furthermore, our results suggest that the unique folded lamella architecture of the cone OS may maximize density or proximity of phototransduction proteins, but is not required for OS function or for protein distribution and retention in different membrane domains.—Conley, S. M., Al-Ubaidi, M. R., Han, Z., Naash, M. I. Rim formation is not a prerequisite for distribution of cone photoreceptor outer segment proteins. FASEB J. 28, 3468 –3479 (2014). www.fasebj.org ABSTRACT
Key Words: RDS 䡠 degeneration 䡠 morphogenesis 䡠 retina The photoreceptor outer segment (OS) comprises stacks of membranous discs/lamellae circumscribed by a hairpin-like rim. Rod discs are sealed and enclosed by the plasma membrane (PM), while cone lamellae are contiguous with the PM. Lamellae have one portion of their circumference/rim surrounded by PM (i.e., closed),
Abbreviations: CC, connecting cilium; CMS, cone matrix sheath; CNG, cyclic nucleotide gated; COS, cone outer segment; ECM, extracellular matrix; IF, immunofluorescence; IRBP, interphotoreceptor retinoid-binding protein; IS, inner segment; NP, nanoparticle; ONL, outer nuclear layer; OS, outer segment; P, postnatal day; PM, plasma membrane; PNA, peanut agglutinin; PRPH2, peripherin-2; RDS, retinal degeneration slow; ROM-1, rod outer segment membrane protein 1; RP1, retinitis pigmentosa 1; RPE, retinal pigment epithelium; TEM, transmission electron microscopy; WT, wild type 3468
while the remainder is exposed or “open” to the extracellular space (1, 2). Proteins localize to different domains in these OSs: the PM, the discs/lamellae, and the rim. For example, opsins are found in the discs/ lamellae, the cyclic nucleotide gated (CNG) channel is found in the PM, and the tetraspanin protein retinal degeneration slow [RDS; also known as peripherin-2 (PRPH2), Fig. 1A] is found in the rim. RDS is required for rim formation (3, 4), and is thought to promote membrane curvature, flattening, and fusion (5–7). The closed portion of the cone rim is similar in orientation to the rod rim (Fig. 1B) and contains RDS. However, the open portion of the rim curves in the opposite direction (with regard to the orientation of proteins in the membrane, Fig. 1B), and RDS is unlikely to be there, since the curvature is backwards from that promoted by RDS. Instead, recent studies in Xenopus laevis have shown that the cholesterol-binding protein prominin-1 (also known as CD133) is found along these open lamellar edges in cones, while RDS resides on opposing closed rims (8). Tetraspanins are known to organize membrane microdomains known as the tetraspanin web (9), and it has been hypothesized that RDS performs a similar function in photoreceptors. In mice lacking RDS (rds⫺/⫺), no rod OSs are formed. To understand what happens to cones in the absence of RDS, we used the Nrl⫺/⫺ mouse model in which developing rods convert to S-cone-like cells (10, 11). In contrast to rods lacking RDS, cones in the rds⫺/⫺/Nrl⫺/⫺ model develop open OSs that lack lamellae and rim structures (12). These data suggest that RDS and rim formation are differently required for OS biogenesis and function in rods vs. cones. Although RDS is required for rim formation in rods and cones, little is known about the mechanisms of rim formation or the precise role of rim formation during OS morphogenesis. Two different models of rod morphogenesis have been proposed (13, 14), but the differences between them remain unresolved, and they 1 Correspondence: OUHSC, Department of Cell Biology, 940 Stanton L. Young Blvd. BMSB 781, Oklahoma City, OK 73104. E-mail: [email protected] doi: 10.1096/fj.14-251397 This article includes supplemental data. Please visit http:// www.fasebj.org to obtain this information.
0892-6638/14/0028-3468 © FASEB
Figure 1. OS ultrastructure and membrane organization. A) RDS is a tetraspanin protein with 4 transmembrane domains, a large extracellular (cones)/intradiscal (rods) loop (EC2/D2), and cytoplasmic N/C termini. The EC2/D2 loop has a highly conserved (green) region and a hypervariable region (orange) and is glycosylated (Y-shaped hexagons). B) In rods, RDS is located along the circumference of the disc, with the glycosylated EC/D2 loop pointing toward the intradiscal space, and is thought to play a part in mediating rim curvature. In cones, RDS is found in a similar localization along the closed edge of the lamellae (i.e., portion surrounded by PM), but the curvature of the open edge of the discs is the opposite direction (with regard to the orientation of the proteins in the membrane) as closed discs and is not likely to contain RDS. Dark gray, PM; light gray, lamella membrane. Opposing membrane leaflets are solid/dashed lines. C, D) Immunogold labeling of cone OSs (COSs) was undertaken on retinal sections collected at P30 using antibodies against S-opsin. Shown are representative structures from WT and rds⫺/⫺ (C), and Nrl⫺/⫺ and rds⫺/⫺/Nrl⫺/⫺ (D). Scale bar ⫽ 500 nm.
do not directly address rim formation. Even less is known about cone morphogenesis. Here we take advantage of the simplified OSs of cones lacking RDS to study the role of rim formation in OS morphogenesis, and particularly in the processes that govern the formation and maintenance of discrete OS membrane domains. We show that despite the lack of lamellae, these cone OSs (COSs) retain distribution of OS proteins in distinct membrane domains, and that reintroduction of RDS leads to OS structures with intermediate attempts at rim/lamella formation.
MATERIALS AND METHODS Animal husbandry Double-knockout rds⫺/⫺/Nrl⫺/⫺ mice were generated from rds⫺/⫺ mice (generously provided by Dr. Neeraj Agarwal, National Eye Institute, Bethesda, MD, USA), and Nrl⫺/⫺ mice (generously provided by Dr. Anand Swaroop, National Eye Institute). Mice were housed in a 12-h dark-light cycle with light intensity ⬃30 lux. Animal breeding and experimental CONE OUTER SEGMENT MORPHOGENESIS
protocols were approved by the University of Oklahoma Health Sciences Center Institutional Animal Care and Use Committee and are consistent with those recommended by the Association for Research in Vision and Ophthalmology. Subretinal injection and processing Mice (rds⫺/⫺/Nrl⫺/⫺) at postnatal d 30 (P30) were subretinally injected as described previously (15) with compacted DNA nanoparticles (NPs) carrying wild-type (WT) mouse RDS gene (NMP) under the control of the human interphotoreceptor retinoid-binding protein (hIRBP) promoter, which drives expression in rods and cones (16, 17). These NPs were generated as described previously (18), and have been shown to effectively drive RDS gene expression in photoreceptors as soon as 2 d postinjection (15, 16). At d 7 postinjection, eyes were enucleated, fixed, and processed for immunofluorescence (IF) or transmission electron microscopy (TEM) as described below. IF Mice at P30 were euthanized, and eyes were enucleated, fixed, embedded, and cryosectioned (10 or 25 m sections), and slides were then processed for IF as described previously 3469
(19). Briefly, sections were either incubated in ice-cold methanol at ⫺20°C for 20 min or immediately rehydrated, then treated with 1% sodium borohydride (Sigma Aldrich, St. Louis, MO, USA) for 2–5 min, washed with water and phosphate buffered saline (PBS), blocked (5% BSA, 2% donkey serum, 1% fish gelatin, and 1% TX-100 in PBS), then incubated with the primary antibodies overnight at 4°C. Slides were then washed with PBS and incubated with the secondary antibodies for 1 h at room temperature, followed by additional washing and mounting using ProlongGold with DAPI (Life Technologies, Grand Island, NY, USA). Primary antibodies are as follows: retinitis pigmentosa 1 (RP1) at 1:1000 dilution (generously provided by Dr. Eric Pierce, Massachusetts Eye and Ear Infirmary, Boston, MA, USA; ref. 20); acetylated ␣-tubulin at 1:100 dilution (T7451; SigmaAldrich); cone S-opsin at 1:500 dilution (H-17; Santa Cruz Biotechnology, Santa Cruz, CA, USA); CNGA3 at 1:250 dilution (generously provided by Dr. Xi-Qin Ding, University of Oklahoma Health Sciences Center; ref. 21); Na⫹K⫹ATPase at 1:500 dilution (a5-c, Developmental Studies Hybridoma Bank, Iowa City, IA, USA); GNAT2 at 1:500 dilution (sc-390; Santa Cruz Biotechnology); RDS mAB 2B7, RDS-CT, ROM-1 at 1:1000 dilution (generated in house and previously characterized; refs. 19, 22); cone Na⫹K⫹Ca2⫹ exchanger at 1:200 dilution (NCKX2; generously provided by Dr. Jonathan Lytton, University of Calgary Health Sciences Center, Calgary, Canada; ref. 23); prominin-1 at 1:200 dilution (13A4, eBioscience, San Diego, CA, USA); M-opsin at 1:35,000 dilution (generously provided by Dr. Cheryl Craft, University of Southern California, Los Angeles, CA, USA; ref. 24). Secondary antibodies were AlexaFluor 488-, 555-, or 647-conjugated donkey IgGs (Life Technologies), used at 1:1000 dilution. AlexaFluor 488- or 555-conjugated peanut agglutinin (PNA; Life Technologies) was used to visualize the cone matrix sheath (CMS). For each antibody combination, the IF experiment was repeated ⱖ3 times using eyes from 3–5 independent animals. Microscope image acquisition All images were captured using a Olympus BX-62 coupled to a spinning disk confocal unit (Olympus DSU; Olympus, Tokyo, Japan). All images were captured at room temperature using either a ⫻60 oil objective with a numerical aperture of 1.42 or a ⫻100 oil objective with a numerical aperture of 1.4. Images were captured at room temperature using a Hamamatsu DCAM-API camera (Hamamatsu Photonics, Hamamatsu, Japan). Image acquisition was done in Slidebook 4.2.0.3 (Intelligent Imaging Innovations, Inc., Denver, CO, USA), using the confocal filter set. Z stacks were deconvolved using the nearest neighbors algorithm, and 3D reconstructions were done using the 3D volume view command in Slidebook. All images shown (except 3D reconstructions of stacks) are single planes from the captured stack and are not collapsed or projection images. Figures were assembled in Adobe Photoshop CS5 (Adobe Systems, San Jose, CA, USA). TEM and immunogold labeling Eyes for electron microscopy were fixed and embedded as described previously (25). Thin sections of ⬃800 Å were collected on copper 75/300 mesh grids, then further processed and stained. Embedding and sectioning for immunogold labeling was done as described previously (19, 26). Thin sections were collected on nickel 75/300 mesh grids. Primary antibodies were anti-S-opsin rabbit polyclonal, used at 1:10 (generously shared by Dr. Cheryl Craft, University of Southern California, Los Angeles, CA, USA), and secondary anti3470
Vol. 28
August 2014
bodies (1:50) were AuroProbe 10 nm gold-conjugated goat anti-rabbit IgG; (GE/Amersham, Piscataway, NJ, USA). All were imaged using a Jeol 100CX electron microscope (Jeol Ltd., Akishima, Japan) at 60 keV accelerating voltage (25).
RESULTS Open COSs are formed in the absence of RDS In WT mouse retinas, only ⬃3–5% of photoreceptors are cones, making them quite difficult to study. We have observed that RDS appears to have different roles in rods vs. cones and can cause both rod- and conedominant disease, piquing interest in the role of RDS and the rim in COS biogenesis and function. We therefore generated the rds⫺/⫺/Nrl⫺/⫺ model and observed the aforementioned lamella-free OSs (12). Strikingly, these OSs retained ⬃50% of photopic ERG response, indicating that the highly modified folded lamella ultrastructure was not required for phototransduction. This divergence between the anatomical and physiological behavior of cones lacking RDS prompted us to undertake more detailed studies of the characteristics of these rim free OSs. Our first step here was to determine whether cones lacking RDS in the WT background also adopt this open OS structure. We performed immunogold labeling with antibodies against S-opsin (to identify COSs) coupled with TEM. As expected, in the WT and Nrl⫺/⫺, S-opsin is highly concentrated in the folded lamellae. In contrast, in both the rds⫺/⫺ and rds⫺/⫺/Nrl⫺/⫺, S-opsin immunoreactivity is present on unfolded, membranous OS structures that lack the normal flattened lamellae seen in the WT (Fig. 1C, D). These results confirmed that lamella-free COSs form as a genuine outcome of RDS deficiency, in both the WT and Nrl⫺/⫺ models. Targeting of COS proteins to the PM/lamellae is unaffected by RDS deficiency The photoreceptor OS is connected to the inner segment (IS) by the connecting cilium (CC). Photoreceptor proteins are synthesized in the IS, and those destined for the OS are then transported to their final destination through the CC. Because of the open membranous structures adopted by the COSs in the absence of RDS, we wanted to confirm the integrity of the CC. We therefore colabeled retinal sections from adult WT, rds⫺/⫺, Nrl⫺/⫺, and rds⫺/⫺/Nrl⫺/⫺ mice with the ciliary marker, RP1 (green; ref. 20 and Fig. 2A, B), and the axoneme marker, acetylated ␣-tubulin (red; Fig. 2A, B). All IF images are single planes from a confocal stack unless indicated otherwise. The Nrl⫺/⫺ and rds⫺/⫺/Nrl⫺/⫺ retinas are characterized by rosettes, in which the outer nuclear layer (ONL) forms the outer layer, with ISs and OSs protruding toward the center of the rosette (Fig. 2A, right panels, R; refs. 10, 12). To identify cones in the WT and rds⫺/⫺, sections were counterstained with the cone matrix marker, PNA (blue). We observed that in both the presence and
The FASEB Journal 䡠 www.fasebj.org
CONLEY ET AL.
Figure 2. The connecting cilia are maintained in the rds⫺/⫺/Nrl⫺/⫺. A) IF was used to evaluate the localization of two different cilia markers, RP1 (green) and acetylated ␣-tubulin (red), in the indicated genotypes. PNA (blue) labeling of cone extracellular matrix was used to identify cones in the WT and rds⫺/⫺ backgrounds. B) Higher-magnification views of the CC. Arrows indicate CC and extending axonemes. DAPI (gray) marks nuclei. C) TEM images of the indicated genotypes were used to examine CC ultrastructure in the presence and absence of RDS. The CC is intact and similar in structure in all examined genotypes. CC, connecting cilium; IS, inner segment; ONL, outer nuclear layer; R, rosette. Scale bars ⫽ 10 m (A, B); 500 nm (C).
absence of RDS, cilia are formed and RP1 colocalized with acetylated ␣-tubulin in the region of the CC (Fig. 2A, B, arrows). Ultrastructural examination (Fig. 2C) confirmed the presence of connecting cilia, microtubules, and basal bodies in all models. Having established that the structural connection between the IS and OS is maintained in the open, balloon-like COSs (Fig. 1D, right panel), we turned our attention to evaluating whether OS proteins were properly trafficked from the IS to the OS. We evaluated S-opsin (red; Fig. 3A–D), a marker for OS lamellae and the cone CNG channel (red; Fig. 3E–H), a marker of OS PM. We colabeled retinal sections with antibodies against RP1 to label the CC (green; Fig. 3A, C, E, G), Na⫹K⫹ATPase to label the ISs (blue, Fig. 3A–C, E–G; and green, Fig. 3D, H). We observed that both S-opsin and CNGA3 properly reached the OS in the rds⫺/⫺/ Nrl⫺/⫺; the IS marker did not colocalize with these OS proteins, and while some colocalization was observed between RP1 and the OS proteins, substantial S-opsin and CNGA3 labeling was detected beyond the CC (i.e., toward the center of the rosette). To confirm that the OS proteins were not stuck in the IS, we assembled 3D reconstructions from 25 m confocal stacks and CONE OUTER SEGMENT MORPHOGENESIS
observed that the majority of CNGA3 and S-opsin (red; Fig. 3D, H) labeling did not colocalize with Na⫹K⫹ATPase (green; Fig. 3D, H) in either the Nrl⫺/⫺ or the rds⫺/⫺/Nrl⫺/⫺; only a small amount of colocalization was observed at the most apical region of the IS, likely corresponding to newly synthesized protein. Similar to the case in the Nrl⫺/⫺ we also observed proper OS targeting in WT and rds⫺/⫺ retinas (Fig. 3C, G). Distribution of COS proteins to different membrane subdomains is preserved in the absence of RDS Having established that gross OS protein targeting is not impaired in cones lacking RDS, we next evaluated distribution of proteins to specific subdomains in the open OSs of RDS-free cones. We used S-opsin as a marker for the lamella subdomain (green; Fig. 4), while CNGA3 was used as a marker for the PM (red; Fig. 4). Little is known about how these discrete subdomains are formed and maintained in cones or how their distinct protein content is retained. Although these two markers sort to different membrane domains, they exhibited overlapping labeling in the Nrl⫺/⫺ (and WT; not shown) when visualized using light microscopy 3471
Figure 3. Proteins are properly targeted to the OS in cones lacking RDS. A–C, E–G) IF with the ciliary marker, RP1 (green), the IS marker, Na⫹K⫹ATPase (blue), and either S-opsin (red, A–C) or CNGA3 (red, E–G). Nuclei are counterstained with DAPI (gray). B, F) Higher-magnification images from A (B) and E (F). C, G) Retinal sections from the WT and rds⫺/⫺. Arrows indicate cones. D, H) 3D reconstructions of image stacks captured from 25 m sections in which IF was used to label the IS marker Na⫹K⫹ATPase (green) and either the lamella marker S-opsin (red, D) or the PM marker CNGA3 (red, H). Nuclei are counterstained in DAPI (blue). Grid dimensions and parameters are indicated in each individual image. R, rosette. Scale bars ⫽ 10 m.
(Fig. 4A, top) because in normal cones, the lamellae are partially enclosed by the PM, and the two domains are quite close to each other (as seen in Fig. 1D). Remarkably, in the rds⫺/⫺/Nrl⫺/⫺, S-opsin and CNGA3 did not colocalize in the expanded, lamella-free COS structures (Fig. 4A, bottom). To confirm that this separation was not an artifact of the rosettes, we assembled 3D reconstructions of confocal stacks collected in both rosettes and the region neighboring the retinal pigment epithelium (RPE; Fig. 4B). The two membrane markers were present in completely separate layers within the open OS structures of the rds⫺/⫺/Nrl⫺/⫺ retina; with S-opsin present in a layer distal (i.e., toward the RPE or toward the inside of the rosette) to that where CNGA3 was detected. To determine whether this preserved localization to distinct membrane domains in the absence of lamellae/RDS was a general phenomenon or merely limited to S-opsin and CNG, we undertook labeling for a series of other proteins that target to those two membrane domains. We first examined another integral mem3472
Vol. 28
August 2014
brane lamella protein, M-opsin, which is coexpressed with S-opsin in most murine cones (27). We observed extensive colocalization of M- and S-opsins in all models (Supplemental Fig. S1A, B) indicating that M-opsin was sorted to the same membrane domain as S-opsin. We next assessed whether nonintegral membrane proteins, which are nevertheless associated with a specific membrane domain, retained their distribution. Cone transducin (GNAT2) is peripherally anchored to the lamella membrane by acyl and farnesyl groups so that it can be in proximity to opsin (28). In WT and Nrl⫺/⫺ cones, GNAT2 (green; Supplemental Fig. S1C, D) was found in the lamella subdomain with S-opsin (red; Supplemental Fig. S1C, D), both proteins colocalized when visualized under light microscopy. Similarly, in the rds⫺/⫺ and rds⫺/⫺ Nrl⫺/⫺ we observed that GNAT2 properly localized to the layer containing S-opsin (Supplemental Fig. S1C, D, bottom). Unfortunately, M-opsin, GNAT2, and CNGA3 antibodies are all rabbit polyclonals, so colabeling with them was not possible. These results confirm that lamella-associated proteins
The FASEB Journal 䡠 www.fasebj.org
CONLEY ET AL.
domain as CNG, a layer distinct from that occupied by lamella proteins. We next examined the localization of a final PM protein, the cone Na⫹K⫹Ca2⫹ exchanger (23, 29) which is involved in the regulation of OS cation concentrations and calcium levels (23, 30, 31). Surprisingly, in the rds⫺/⫺/Nrl⫺/⫺, the Na⫹K⫹Ca2⫹ exchanger (red; Fig. 5C, D) localized to the S-opsin layer, unlike CNGA3 and PNA, indicating that only specific PM proteins retain their specific distribution in the absence of lamellae. Prominin-1 partially codistributes with S-opsin and CNG In addition to the PM and lamella membrane domains, properly assembled OSs have a third distinct membrane domain, the rim, which forms the perimeter of
Figure 4. OS membrane distribution is preserved in the rds⫺/⫺/Nrl⫺/⫺. A) IF was used to evaluate lamella (S-opsin, green) and PM (CNGA3, red) protein localization at P30 in the indicated genotypes. Nuclei are counterstained with DAPI (gray). Images show single planes from confocal stacks. Scale bar ⫽ 10 m. B) 3D reconstruction of a 25 m section from indicated genotypes labeled with S-opsin (green) and CNGA3 (red). Nuclei are counterstained with DAPI (blue). Grid sizes are shown in the images. R, rosette.
distributed together to a domain distinct from PM proteins even in the absence of RDS and structural lamellae/rims. We next assessed the distribution of other PM-associated proteins (i.e., in addition to CNG). One of the features of the COS PM is that it interacts with the cone extracellular matrix (ECM), known as the CMS. In normal cones, both in the WT (e.g., in Fig. 2A) and Nrl⫺/⫺, the matrix surrounds the OS/IS, shown by imaging PNA (a green CMS marker; Fig. 5A, top) and S-opsin or CNGA3 (blue and red, respectively; Fig. 5A, arrows and inset). In the rds⫺/⫺/Nrl⫺/⫺, S-opsin labeling is found in a distinct layer from PNA (Fig. 5B; observe blue and green layers in the left panel); however, CNGA3 colocalizes with PNA (Fig. 5A, B, middle panels; observe yellow). These results suggest that PM proteins in these open COS structures maintain interactions with the CMS and are found in the same CONE OUTER SEGMENT MORPHOGENESIS
Figure 5. Localization of additional PM markers in the rds⫺/⫺/Nrl⫺/⫺. A, B) IF on P30 retinal sections from Nrl⫺/⫺ (A) and rds⫺/⫺/Nrl⫺/⫺ (B) was used to evaluate localization of the CMS (PNA, green) relative to the two COS membrane microdomains, lamella (S-opsin, blue) and PM (CNGA3, red). Nuclei are counterstained with DAPI (gray). Arrows indicate regions of colocalization. Insets: Higher-magnification views. C, D) IF was performed on P30 retinal sections from Nrl⫺/⫺ (C) and rds⫺/⫺/Nrl⫺/⫺ (D). Na⫹K⫹Ca2⫹ exchanger is shown in red and S-opsin in green in COSs of Nrl⫺/⫺ and rds⫺/⫺/Nrl⫺/⫺ retinas. Nuclei are counterstained with DAPI (blue). Arrows indicate areas of colocalization. R, rosette. Scale bars ⫽ 10 m. 3473
the disc, and we were interested in the distribution of other rim proteins, such as RDS’ nonglycosylated homologue, rod outer segment membrane protein-1 (ROM-1), and the cholesterol binding protein prominin-1, in the absence of RDS. In rods, the entire circumference of the rim is surrounded by PM, while in cones, a portion of the rim is open. It has been shown that in amphibian cones, RDS is found in the portion of the rim that is surrounded by PM, while prominin-1 is found in the open portion of the rim (8). ROM-1 is virtually undetectable in the absence of RDS (26); however, we assessed the localization of prominin-1 in the Nrl⫺/⫺ and rds⫺/⫺/Nrl⫺/⫺ retinas in relation to other OS markers. We colabeled with prominin-1 (red; Fig. 6) and S-opsin (lamellae, green; Fig. 6A) or CNGA3 (PM, green; Fig. 6B). Interestingly, we observed partial colocalization between prominin-1 and both S-opsin and CNGA3 (Fig. 6A, B, yellow arrows), indicating that prominin-1 does not localize specifically to the lamellae or PM domain. However, many areas expressed S-opsin or CNGA3 but not prominin-1 (in both the Nrl⫺/⫺ and rds⫺/⫺/Nrl⫺/⫺), an observation which suggests that prominin-1 does not distribute uniformly throughout the OS (i.e., there are some parts of the OS that lack prominin-1).
Because we did not see complete colocalization of prominin-1 with the OS markers CNGA3 and S-opsin, we wanted to confirm that it was actually being targeted to the OS in the absence of RDS. We therefore colabeled with prominin-1 (red) and the IS marker Na⫹K⫹ATPase (green; Fig. 6C, D). Interestingly, we observed that a substantial amount of prominin-1 accumulated in the cone IS in the absence of RDS (Fig. 6C, arrows; magnified in Fig. 6E), although some was also detected in the OS (Fig. 6C, arrowheads; magnified in Fig. 6E). Prominin-1 was also mislocalized in rod and cone photoreceptors lacking RDS in the WT background (Fig. 6D, magnified in Fig. 6E). To determine whether this mislocalization was due to the lack of RDS or merely the presence of malformed OS, we examined prominin-1 localization in the rhodopsin knockout retina (rho⫺/⫺). While cones in this model are fairly normal at early time points, rods have only tiny sacs of PM in the OS with no normal discs (although RDS and ROM-1 are present in the OS; ref. 32). We did not observe accumulation of prominin-1 in the IS of the rho⫺/⫺ retina (Fig. 6D, bottom panels; magnified in Fig. 6E) suggesting that prominin-1 mislocalization in the rds⫺/⫺ and rds⫺/⫺/Nrl⫺/⫺ may arise due to the absence of RDS and/or the rim structure.
Figure 6. Prominin-1 partially colocalizes with lamella and PM proteins in the absence of RDS but also accumulates in the IS. A–C) IF was used to evaluate the distribution of prominin-1 in the Nrl⫺/⫺ and rds⫺/⫺/Nrl⫺/⫺ (red) in relation to the lamellae (A, S-opsin, green), PM (B, CNGA3, green), and IS (C, Na⫹K⫹ATPase, green). D) WT, rds⫺/⫺, and rho⫺/⫺ retinal sections were labeled for prominin-1 (red) and the IS marker Na⫹K⫹ATPase. E) Higher-magnification views from C, D. Nuclei are counterstained with DAPI (blue). Arrows indicate areas of colocalization; arrowheads indicate areas of prominin-1 that do not colocalize with the IS marker. R, rosette. Scale bars ⫽ 10 m. 3474
Vol. 28
August 2014
The FASEB Journal 䡠 www.fasebj.org
CONLEY ET AL.
Reintroduction of RDS leads to lamella and rim formation Combined, the above results suggested that RDS is necessary for the formation of the rim region and morphogenesis of proper cone ultrastructure, but not for the formation or maintenance of distinct protein domains within that structure. We next asked what the outcome of reintroduction of RDS would be on cone morphogenesis and membrane organization. We took two approaches to this question; first, we examined the cones of the rds⫹/⫺/Nrl⫺/⫺; and second, we subretinally injected rds⫺/⫺/Nrl⫺/⫺ mice with compacted DNA NPs
carrying a photoreceptor-specific RDS expression vector (IRBP-NMP, previously characterized in ref. 16). Prior to assessing the structure of these retinas, we confirmed that both RDS and ROM-1 were detected in rds⫹/⫺/Nrl⫺/⫺ and IRBP-NMP injected rds⫺/⫺/Nrl⫺/⫺ (blue and red; Fig. 7A). On the ultrastructural level, the COSs of the Nrl⫺/⫺ exhibited nicely stacked, flattened discs (Fig. 7B, C, pink arrows and lines) aligned along the axoneme (Fig. 7B, C, orange arrows and lines), while the rds⫺/⫺/ Nrl⫺/⫺ exhibited balloon-like OSs lacking rims or lamellae. In the Nrl⫺/⫺, we observe a mix of closed (i.e., surrounded by PM; Fig. 7B, C, blue arrows and lines)
Figure 7. RDS supplementation initiates lamella formation in the rds⫺/⫺/Nrl⫺/⫺. RDS supplementation was performed genetically (rds⫹/⫺/Nrl⫺/⫺) or by subretinal injection of IRBP-NMP NPs carrying WT RDS (IRBP-NMP in rds⫺/⫺/Nrl⫺/⫺ mice). A) IF was performed using antibodies against RDS (blue) and ROM-1 (red) with DAPI labeling of nuclei (gray). Arrows indicate areas of colocalization (magenta). B) TEM images of COSs in the indicated genotypes or treatment groups. C) Diagrams representing the hypothetical morphology observed in these groups. Colors correspond to arrows in the TEM images. Scale bars ⫽ 10 m (A); 2 m (B). R, rosette. CONE OUTER SEGMENT MORPHOGENESIS
3475
and open rims (i.e., not surrounded by PM; Fig. 7B, C, purple arrows and lines). We hypothesize that these open rims could be the sites of prominin-1 localization, however, a conclusion on this point must await the availability of a mammalian prominin-1 antibody that works on immunogold cytochemistry. The rds⫹/⫺/ Nrl⫺/⫺ OSs showed sworls of membrane that nonetheless had pinched and flattened rims (Fig. 7B, black arrows). Similarly, in the rds⫺/⫺/Nrl⫺/⫺ retinas that were injected with IRBP-NMP, membrane sworls and flattened rims (Fig. 7B, black arrows) were also detected, but the lamellae were much larger than in the Nrl⫺/⫺ or rds⫹/⫺/Nrl⫺/⫺. While we would predict that the flattened rims in the rds⫹/⫺/Nrl⫺/⫺ and IRBP-NMPinjected rds⫺/⫺/Nrl⫺/⫺ would come in both open and closed varieties as in the normal cone, the highly aberrant sworl structure and lack of clearly identifiable PM make it difficult to conclusively determine the status of these rims. Consistent with the presence of rims and lamellae (albeit improperly sized and stacked) in the COSs of the rds⫹/⫺/Nrl⫺/⫺ and IRBP-NMP injected rds⫺/⫺/Nrl⫺/⫺, CNGA3, RDS, and S-opsin exhibited partial colocalization (similar to the Nrl⫺/⫺, not shown).
DISCUSSION Here we present evidence demonstrating that the segregation of PM and lamella membrane domains is preserved in the open OS of cones lacking RDS (Fig. 8A). We show that cone opsins and other lamella markers occupy a distinct area of COS than the PM protein CNGA3 in these unfolded COSs of the rds⫺/⫺/Nrl⫺/⫺. We further show that a marker of the CMS, PNA, colocalizes with PM proteins (such as CNG) rather than with markers of the lamellae, consistent with predicted interactions between the PM proteins and the components of the ECM. We also show that all studied proteins except the rim protein prominin-1 are properly targeted from the IS to the COS. PM, rim, and lamella proteins are thought to traffic independently in rods (33–35). Studies on trafficking in cones are also being done, however, less is known about the processes which govern cone OS (COS) targeting (36). We show that prominin-1 accumulates in the ISs of photoreceptors lacking RDS (although some is found in the OS as well). These data suggest that RDS or rim structures may be required for proper OS targeting of prominin-1. Historically, the question of whether rim proteins traffic together has been difficult to answer, since the other primary rim protein, ROM-1, is virtually undetectable in the absence of RDS. However our data here suggest that rim proteins, even if they localize to different regions of the rim (open vs. closed) may traffic together, although definitive results on this must await additional study. A primary role of tetraspanins as a protein family is the organization of functional membrane microdomains (9, 37), and RDS may play a similar role in the 3476
Vol. 28
August 2014
Figure 8. Model of OS protein distribution and morphogenesis. A) Diagram summarizing the distribution of OS membrane markers in WT and RDS-free cones. B) Hypothetical model of COS morphogenesis showing 6 steps in the formation of normal COSs, and the final equivalent in the absence of RDS. Step 1: initial PM covering of nascent connecting cilium/axoneme. Step 2: accumulation of opsin-containing membrane (pink). Step 3: accumulation of rim proteins (blue) and formation of the rim. Step 4: lamella growth and alignment with the axoneme. Steps 5 and 6: formation and growth of additional lamellae to result in traditional stacked morphology.
photoreceptor. However, RDS also plays a structural role as it is required for the formation of the physical rim (4, 12). Here we show that introducing RDS into the lamella-free OS of the rds⫺/⫺/Nrl⫺/⫺, either genetically or by subretinal injection of DNA NPs, leads to the formation of a whorl of flattened membrane structures reminiscent of lamellae, although they are larger than normal lamellae. This size anomaly is likely due to an insufficient quantity of RDS; as the amount of RDS
The FASEB Journal 䡠 www.fasebj.org
CONLEY ET AL.
is critical for proper lamella formation and OS morphogenesis in cones and rods (38). The role of RDS in promoting flattening of discs/lamellae is further highlighted by in vitro studies showing that RDS expression induces a flattened morphology in microsomal vesicles (6). Further, we have shown that transgenic mice expressing a mutant form of RDS (C150S) which cannot form normal higher-order oligomeric complexes exhibit open, nonflattened lamellae in their COSs (19). However, empirically it is quite difficult to distinguish between functions of RDS tied to formation of the physical rim structure and those tied to gathering rim proteins to these structures. Tracking COS membrane organization and the ultrastructure of COS in models that lack RDS or have varying amounts of RDS has led us to hypothesize a model of COS biogenesis in which rim formation is not the essential initiating step (in contrast to rods, which form no OSs without rims/RDS). We hypothesize that lamella growth and formation is initiated as cone-opsin carrying membrane (Fig. 8B, pink, step 2) accumulates in a region of the PM (Fig. 8B, black) primed for membrane growth (Fig. 8B, green). Subsequently, RDScontaining membrane (Fig. 8B, blue, steps 3–5) initiates rim curvature/attachment to the axoneme and flattened, stacked lamellae are formed (Fig. 8B, step 6). Final sizing of the lamellae is likely governed by a combination of several factors, including ROM-1 and the actin cytoskeleton (39, 40). In the rds⫺/⫺/Nrl⫺/⫺, the lack of RDS would therefore lead simply to continued growth of the portion of membrane containing lamella proteins without formation of rim or lamella structures (Fig. 8B, bottom, pink). Notably, this region of membrane, which we have termed “primed for growth,” may contain the lipid and/or protein components that are responsible for maintaining the separate distribution of PM and lamella proteins. In the rds⫺/⫺/ Nrl⫺/⫺ this results in a separate layer of PM/lamella proteins, while in the normal cone (Nrl⫺/⫺), this membrane would be found in the open portion of the cone rims, while the RDS-containing rims would be found along the closed portion of the lamellae (Fig. 8B, green and blue, respectively). The fraction of open rim varies; in amphibians, the majority of the cone rim is open (2), while in mammals, a much smaller portion is open, and many discs are completely enclosed (as in rods; refs. 1, 41). The conversion from open to closed discs in mammalian cones appears to be dynamic, with lamellae undergoing episodic fusion with the PM over time (1), but it is not clear what the functional significance of these differences are. The observation that lamella-associated proteins (Sopsin and GNAT2) localize to a distinct membrane domain from the PM-associated protein CNGA3 suggests that the distribution of some proteins in discrete COS membrane domains does not require lamellae, rims, or RDS. However, we also show that, unlike CNG, the Na⫹K⫹Ca2⫹ exchanger, a PM protein, does not segregate specifically to the PM domain. Instead, it distributes throughout the COS in the rds⫺/⫺/ CONE OUTER SEGMENT MORPHOGENESIS
Nrl⫺/⫺. These data indicate that the distribution of all OS PM proteins is not governed in the same way, and suggest the presence of multiple sorting processes, some of which are preserved in the absence of lamellae and some of which are not. Interestingly, the rod Na⫹K⫹Ca2⫹ exchanger has been shown to directly interact with the rod CNG ␣ subunit (30, 42, 43), however, this interaction is absent in cones (44), suggesting that in addition to multiple sorting pathways in cones, there may also be differences in the way proteins are arranged in the two cell types. The next logical question is how some proteins are retained in separate domains (PM vs. lamella). We have shown that lamellae/RDS are not universally required for this process, however several potential mechanisms exist. First, the lipid content of the membrane may play a role in maintaining protein segregation. Second, other membrane proteins or scaffold/cytoskeletal proteins may help keep the domains segregated. Finally, the ECM may play a critical role. We show that the CMS marker PNA colocalizes with the PM protein CNGA3, and interactions between the PM and the ECM could be critical for maintaining the segregated protein distribution we observe. The extent to which each of these (or other unknown mechanisms) may contribute to our observation is not clear and may be the focus of further study. While we have no direct evidence for the identity of the components which facilitate the separation of discrete domains, some evidence suggests that prominin-1 may be one such component. First, we show that in the rds⫺/⫺/Nrl⫺/⫺ S-opsin and CNGA3 exhibit only partial colocalization with prominin-1; that is, there are extensive lamella (S-opsin) and PM (CNGA3) membrane regions that do not coincide with prominin-1 labeling, suggesting that prominin-1 may be present at the barrier between the two membrane domains. Second, while results from amphibians do not universally translate to mammals, X. laevis cones clearly exhibit prominin-1 along lamella edges open to the extracellular space and RDS and the axoneme marker acetylated ␣-tubulin along the opposite edge (8). Third, evidence from the prominin-1 knockout mouse shows that prominin-1 is not required for membrane folding or rim formation, but is required for proper disc/lamella growth and alignment (45). Finally, prominin-1 has some molecular features which make it likely to be involved in membrane organization. It binds cholesterol and localizes to specific lipid microdomains (46). More importantly, it always localizes to curved membrane regions, for example microvilli in epithelial cells or protrusions of the PM (such as lamellipodia) in nonepithelial cells (47). Future studies will focus on clearly testing these hypotheses regarding morphogenesis and the potential role of prominin-1 in mammalian cones. This study demonstrates that the folded lamella ultrastructure of the COS and the presence of RDS are not required for COS membrane organization and basic morphogenesis. Combined with our previous work showing that lamellae/RDS are not required for 3477
cone phototransduction in the rds⫺/⫺/Nrl⫺/⫺, these data suggest that cones and rods are significantly different in their dependence on the highly modified OS ultrastructure and in the involvement of the rim in the process of morphogenesis, since rods without RDS fail to form even nascent OS discs and exhibit no signs of phototransduction. Understanding the biological processes which underlie cone photoreceptor OS and protein distribution and morphogenesis is essential from both a cell biological standpoint and for our understanding of the pathology of RDS-associated disease. This work was supported by the U.S. National Eye Institute (EY10609, M.I.N.; EY018512, S.M.C.); the Foundation Fighting Blindness (M.I.N. and M.R.A.); the Knights Templar Eye Foundation (Z.H.), and the Oklahoma Center for the Advancement of Science and Technology (S.M.C., Z.H., and M.I.N.). The authors thank Mr. Justin Burnett, Ms. Barb Nagel, and Ms. Reem Alomer for their technical assistance. The compacted DNA NPs were generated in collaboration with Dr. Mark Cooper at Copernicus Therapeutics, Inc. (Cleveland, OH, USA). Antibodies (as detailed in Materials and Methods) were generously shared by Dr. Cheryl Craft (University of Southern California, Los Angeles, CA, USA), Dr. Jonathan Lytton (University of Calgary Health Sciences Center, Calgary, AB, Canada), Dr. Xi-Qin Ding (University of Oklahoma Health Sciences Center), Dr. Eric Pierce (Massachusetts Eye and Ear Infirmary, Boston, MA, USA), and Dr. David Garbers (University of Texas Southwestern Medical Center, Dallas, TX, USA). Mice were generously shared by Dr. Anand Swaroop, (Nrl⫺/⫺, National Eye Institute, Bethesda, MD, USA), and Dr. Neeraj Agarwal (rds⫺/⫺, U.S. National Institutes of Health, Bethesda, MD, USA). The authors declare no conflicts of interest.
2. 3. 4.
5. 6. 7. 8.
9. 10.
3478
Anderson, D. H., Fisher, S. K., and Steinberg, R. H. (1978) Mammalian cones: disc shedding, phagocytosis, and renewal. Invest. Ophthalmol. Vis. Sci. 17, 117–133 Eckmiller, M. S. (1987) Cone outer segment morphogenesis: taper change and distal invaginations. J. Cell Biol. 105, 2267– 2277 Molday, R. S., Hicks, D., and Molday, L. (1987) Peripherin. A rim-specific membrane protein of rod outer segment discs. Invest. Ophthalmol. Vis. Sci. 28, 50 –61 Arikawa, K., Molday, L. L., Molday, R. S., and Williams, D. S. (1992) Localization of peripherin/rds in the disk membranes of cone and rod photoreceptors: relationship to disk membrane morphogenesis and retinal degeneration. J. Cell Biol. 116, 659 – 667 Conley, S. M., Cai, X., and Naash, M. I. (2008) Nonviral ocular gene therapy: assessment and future directions. Curr. Opin. Mol. Ther. 10, 456 –463 Wrigley, J. D., Ahmed, T., Nevett, C. L., and Findlay, J. B. (2000) Peripherin/rds influences membrane vesicle morphology. Implications for retinopathies. J. Biol. Chem. 275, 13191–13194 Boesze-Battaglia, K. (2000) Fusion between retinal rod outer segment membranes and model membranes: functional assays and role for peripherin/rds. Methods Enzymol. 316, 65–86 Han, Z., Anderson, D. W., and Papermaster, D. S. (2012) Prominin-1 localizes to the open rims of outer segment lamellae in Xenopus laevis rod and cone photoreceptors. Invest. Ophthalmol. Vis. Sci. 53, 361–373 Hemler, M. E. (2001) Specific tetraspanin functions. J. Cell Biol. 155, 1103–1107 Mears, A. J., Kondo, M., Swain, P. K., Takada, Y., Bush, R. A., Saunders, T. L., Sieving, P. A., and Swaroop, A. (2001) Nrl is
Vol. 28
August 2014
12.
13. 14. 15.
16.
17. 18. 19.
20. 21.
22.
REFERENCES 1.
11.
23.
24.
25.
26.
27.
28.
29.
required for rod photoreceptor development. Nat. Genet. 29, 447–452 Daniele, L. L., Lillo, C., Lyubarsky, A. L., Nikonov, S. S., Philp, N., Mears, A. J., Swaroop, A., Williams, D. S., and Pugh, E. N., Jr. (2005) Cone-like morphological, molecular, and electrophysiological features of the photoreceptors of the Nrl knockout mouse. Invest. Ophthalmol. Vis. Sci. 46, 2156 –2167 Farjo, R., Skaggs, J. S., Nagel, B. A., Quiambao, A. B., Nash, Z. A., Fliesler, S. J., and Naash, M. I. (2006) Retention of function without normal disc morphogenesis occurs in cone but not rod photoreceptors. J. Cell Biol. 173, 59 –68 Chuang, J. Z., Zhao, Y., and Sung, C. H. (2007) SARA-regulated vesicular targeting underlies formation of the light-sensing organelle in mammalian rods. Cell 130, 535–547 Steinberg, R. H., Fisher, S. K., and Anderson, D. H. (1980) Disc morphogenesis in vertebrate photoreceptors. J. Comp. Neurol. 190, 501–508 Cai, X., Conley, S. M., Nash, Z., Fliesler, S. J., Cooper, M. J., and Naash, M. I. (2010) Gene delivery to mitotic and postmitotic photoreceptors via compacted DNA nanoparticles results in improved phenotype in a mouse model of retinitis pigmentosa. FASEB J. 24, 1178 –1191 Cai, X., Nash, Z., Conley, S. M., Fliesler, S. J., Cooper, M. J., and Naash, M. I. (2009) A partial structural and functional rescue of a retinitis pigmentosa model with compacted DNA nanoparticles. PLoS One 4, e5290 Han, Z., Conley, S. M., Makkia, R. S., Cooper, M. J., and Naash, M. I. (2012) DNA nanoparticle-mediated ABCA4 delivery rescues Stargardt dystrophy in mice. J. Clin. Invest. 122, 3221–3226 Farjo, R., Skaggs, J., Quiambao, A. B., Cooper, M. J., and Naash, M. I. (2006) Efficient non-viral ocular gene transfer with compacted DNA nanoparticles. PLoS One 1, e38 Chakraborty, D., Conley, S. M., Stuck, M. W., and Naash, M. I. (2010) Differences in RDS trafficking, assembly and function in cones versus rods: insights from studies of C150S-RDS. Hum. Mol. Genet. 19, 4799 –4812 Liu, Q., Zuo, J., and Pierce, E. A. (2004) The retinitis pigmentosa 1 protein is a photoreceptor microtubule-associated protein. J. Neurosci. 24, 6427–6436 Fitzgerald, J. B., Malykhina, A. P., Al-Ubaidi, M. R., and Ding, X. Q. (2008) Functional expression of cone cyclic nucleotidegated channel in cone photoreceptor-derived 661W cells. Adv. Exp. Med. Biol. 613, 327–334 Chakraborty, D., Ding, X. Q., Fliesler, S. J., and Naash, M. I. (2008) Outer segment oligomerization of Rds: evidence from mouse models and subcellular fractionation. Biochemistry (Mosc.) 47, 1144 –1156 Li, X. F., Kiedrowski, L., Tremblay, F., Fernandez, F. R., Perizzolo, M., Winkfein, R. J., Turner, R. W., Bains, J. S., Rancourt, D. E., and Lytton, J. (2006) Importance of K⫹-dependent Na⫹/Ca2⫹-exchanger 2, NCKX2, in motor learning and memory. J. Biol. Chem. 281, 6273–6282 Zhu, X., Brown, B., Li, A., Mears, A. J., Swaroop, A., and Craft, C. M. (2003) GRK1-dependent phosphorylation of S and M opsins and their binding to cone arrestin during cone phototransduction in the mouse retina. J. Neurosci. 23, 6152–6160 Stricker, H. M., Ding, X. Q., Quiambao, A., Fliesler, S. J., and Naash, M. I. (2005) The Cys214 –⬎Ser mutation in peripherin/ rds causes a loss-of-function phenotype in transgenic mice. Biochem. J. 388, 605–613 Chakraborty, D., Ding, X. Q., Conley, S. M., Fliesler, S. J., and Naash, M. I. (2009) Differential requirements for retinal degeneration slow intermolecular disulfide-linked oligomerization in rods versus cones. Hum. Mol. Genet. 18, 797–808 Craig, S. E., Calinescu, A. A., and Hitchcock, P. F. (2008) Identification of the molecular signatures integral to regenerating photoreceptors in the retina of the zebra fish. J. Ocul. Biol. Dis. Infor. 1, 73–84 Bigay, J., Faurobert, E., Franco, M., and Chabre, M. (1994) Roles of lipid modifications of transducin subunits in their GDPdependent association and membrane binding. Biochemistry (Mosc.) 33, 14081–14090 Schwarzer, A., Kim, T. S., Hagen, V., Molday, R. S., and Bauer, P. J. (1997) The Na/Ca-K exchanger of rod photoreceptor exists as dimer in the plasma membrane. Biochemistry (Mosc.) 36, 13667–13676
The FASEB Journal 䡠 www.fasebj.org
CONLEY ET AL.
30.
31. 32.
33.
34. 35.
36.
37. 38.
39.
Bauer, P. J. (2002) Binding of the retinal rod Na⫹/Ca2⫹-K⫹ exchanger to the cGMP-gated channel indicates local Ca(2⫹)signaling in vertebrate photoreceptors. Ann. N. Y. Acad. Sci. 976, 325–334 Kaupp, U. B., and Seifert, R. (2002) Cyclic nucleotide-gated ion channels. Physiol. Rev. 82, 769 –824 Lem, J., Krasnoperova, N. V., Calvert, P. D., Kosaras, B., Cameron, D. A., Nicolo, M., Makino, C. L., and Sidman, R. L. (1999) Morphological, physiological, and biochemical changes in rhodopsin knockout mice. Proc. Natl. Acad. Sci. U. S. A. 96, 736 –741 Fariss, R. N., Molday, R. S., Fisher, S. K., and Matsumoto, B. (1997) Evidence from normal and degenerating photoreceptors that two outer segment integral membrane proteins have separate transport pathways. J. Comp. Neurol. 387, 148 –156 Kizhatil, K., Baker, S. A., Arshavsky, V. Y., and Bennett, V. (2009) Ankyrin-G promotes cyclic nucleotide-gated channel transport to rod photoreceptor sensory cilia. Science 323, 1614 –1617 Nishimura, D. Y., Fath, M., Mullins, R. F., Searby, C., Andrews, M., Davis, R., Andorf, J. L., Mykytyn, K., Swiderski, R. E., Yang, B., Carmi, R., Stone, E. M., and Sheffield, V. C. (2004) Bbs2-null mice have neurosensory deficits, a defect in social dominance, and retinopathy associated with mislocalization of rhodopsin. Proc. Natl. Acad. Sci. U. S. A. 101, 16588 –16593 Craig, S. E., Thummel, R., Ahmed, H., Vasta, G. R., Hyde, D. R., and Hitchcock, P. F. (2010) The zebrafish galectin Drgal1-l2 is expressed by proliferating Muller glia and photoreceptor progenitors and regulates the regeneration of rod photoreceptors. Invest. Ophthalmol. Vis. Sci. 51, 3244 –3252 Hemler, M. E. (2005) Tetraspanin functions and associated microdomains. Nat. Rev. Mol. Cell Biol. 6, 801–811 Nour, M., Ding, X. Q., Stricker, H., Fliesler, S. J., and Naash, M. I. (2004) Modulating expression of peripherin/rds in transgenic mice: critical levels and the effect of overexpression. Invest. Ophthalmol. Vis. Sci. 45, 2514 –2521 Clarke, G., Goldberg, A. F., Vidgen, D., Collins, L., Ploder, L., Schwarz, L., Molday, L. L., Rossant, J., Szel, A., Molday, R. S.,
CONE OUTER SEGMENT MORPHOGENESIS
40. 41. 42. 43. 44.
45.
46.
47.
Birch, D. G., and McInnes, R. R. (2000) Rom-1 is required for rod photoreceptor viability and the regulation of disk morphogenesis. Nat. Genet. 25, 67–73 Vaughan, D. K., and Fisher, S. K. (1989) Cytochalasin D disrupts outer segment disc morphogenesis in situ in rabbit retina. Invest. Ophthalmol. Vis. Sci. 30, 339 –342 Young, R. W. (1971) The renewal of rod and cone outer segments in the rhesus monkey. J. Cell Biol. 49, 303–318 Kim, T. S., Reid, D. M., and Molday, R. S. (1998) Structurefunction relationships and localization of the Na/Ca-K exchanger in rod photoreceptors. J. Biol. Chem. 273, 16561–16567 Schwarzer, A., Schauf, H., and Bauer, P. J. (2000) Binding of the cGMP-gated channel to the Na/Ca-K exchanger in rod photoreceptors. J. Biol. Chem. 275, 13448 –13454 Matveev, A. V., Quiambao, A. B., Browning Fitzgerald, J., and Ding, X. Q. (2008) Native cone photoreceptor cyclic nucleotidegated channel is a heterotetrameric complex comprising both CNGA3 and CNGB3: a study using the cone-dominant retina of Nrl⫺/⫺ mice. J. Neurochem. 106, 2042–2055 Zacchigna, S., Oh, H., Wilsch-Brauninger, M., Missol-Kolka, E., Jaszai, J., Jansen, S., Tanimoto, N., Tonagel, F., Seeliger, M., Huttner, W. B., Corbeil, D., Dewerchin, M., Vinckier, S., Moons, L., and Carmeliet, P. (2009) Loss of the cholesterol-binding protein prominin-1/CD133 causes disk dysmorphogenesis and photoreceptor degeneration. J. Neurosci. 29, 2297–2308 Roper, K., Corbeil, D., and Huttner, W. B. (2000) Retention of prominin in microvilli reveals distinct cholesterol-based lipid micro-domains in the apical plasma membrane. Nat. Cell Biol. 2, 582–592 Weigmann, A., Corbeil, D., Hellwig, A., and Huttner, W. B. (1997) Prominin, a novel microvilli-specific polytopic membrane protein of the apical surface of epithelial cells, is targeted to plasmalemmal protrusions of non-epithelial cells. Proc. Natl. Acad. Sci. U. S. A. 94, 12425–12430 Received for publication February 14, 2014. Accepted for publication April 7, 2014.
3479
