Learn more: PMC Disclaimer | PMC Copyright Notice
NudC regulates photoreceptor disk morphogenesis and rhodopsin localization
Abstract
The outer segment (OS) of rod photoreceptors consist of a highly modified primary cilium containing phototransduction machinery necessary for light detection. The delivery and organization of the phototransduction components within and along the cilium into the series of stacked, highly organized disks is critical for cell function and viability. How disks are formed within the cilium remains an area of active investigation. We have found nuclear distribution protein C (nudC), a key component of mitosis and cytokinesis during development, to be present in the inner segment region of these postmitotic cells in several species, including mouse, tree shrew, monkey, and frog. Further, we found nudC interacts with rhodopsin and the small GTPase rab11a. Here, we show through transgenic tadpole studies that nudC is integral to rod cell disk formation and photoreceptor protein localization. Finally, we demonstrate that short hairpin RNA knockdown of nudC in tadpole rod photoreceptors, which leads to the inability of rod cells to maintain their OS, is rescued through coexpression of murine nudC.—Boitet, E. R., Reish, N. J., Hubbard, M. G., Gross, A. K. NudC regulates photoreceptor disk morphogenesis and rhodopsin localization.
In vertebrate rod photoreceptor cells, the rod outer segment (ROS) is a heavily modified primary cilium that contains the phototransduction components necessary to propagate a physiologic response to light. The phototransduction components are highly concentrated within stacks of membranous disks, which are initially synthesized in the rod inner segment (IS). Thus, an efficient vectorial transport system is required to move these components packaged in vesicles from the IS region to the connecting cilium (CC) at the base of the OS (1–3). Investigation into the formation of these disks has demonstrated that the plasma membrane evaginates at the actin-rich base of the OS to form flattened lamellae exposed to the extracellular space. These flattened lamellae grow to fill the entire diameter of the OS and then fuse to the OS plasma membrane to form an enclosed disk (4–7). As newly synthesized disks are added from the base of the OS, disks at the apical end of the OS are phagocytosed by retinal pigmented epithelial (RPE) cells. Maintenance of the disk-turnover cycle is critical for maintaining proper OS length and effective cellular functioning (8, 9). Several proteins have been identified as facilitators in the packaging, delivery, or release of cargo as it moves along this transport system; however, numerous other proteins may be involved that have yet to be identified or characterized. During the course of our investigation into the role of the small GTPase rab11a in trafficking and packaging of the GPCR rhodopsin, we discovered an additional protein component of the rhodopsin, rab11a complex, and here identify it as the mouse homolog of nuclear distribution protein C (nudC) (10).
NudC was originally identified in the filamentous fungus Aspergillus nidulans via a temperature-sensitive mutant nudCL146P, where it was found to be required for proper nuclear migration during asexual reproduction (11, 12). The larger family of Nud genes in A. nidulans is recognized as coding for components of dynein, including the light, intermediate, and heavy chains, as well as dynactin complexes (12–14). NudC is highly conserved among eukaryotes both structurally and functionally, though the N-terminal region is expanded in metazoans (13, 15). Equivalent mutations in human nudC, nudCL279P, and rodent nudC, nudCL280P, corresponding to nudCL146P in A. nidulans show a variety of phenotypes in cell culture, including perinuclear accumulation of microtubules, increased distance between the centrosome and nucleus, and reduced stability of the interaction partner lissencephaly protein 1 (Lis1) (14, 16–18). These results imply that nudC has regulatory roles in mitosis, cytokinesis, cell migration, and F-actin dynamics. Furthermore, Zhang and colleagues have confirmed that nudC regulates actin dynamics by stabilizing the F-actin modulator cofilin 1 in zebrafish (19). Overall, the common feature in these various functions appears to be the stabilization of microtubule, dynein, and dynactin complexes (20, 21). Investigation of the human nudC gene has expanded our understanding of the gene product and emphasized the importance of nudC not only to individual cells but to critical responses carried out by many cells in order to maintain cellular homeostasis across many cell types (13). However, there has been no previous evidence of a role for nudC in postmitotic senescent cells. NudC was identified in an exhaustive list of proteins within the proteome of the mouse photoreceptor cilium complex (22, 23), but the molecular mechanism of its action in photoreceptors remains unknown.
After identifying nudC as a protein component of the rhodopsin rab11a complex, we wanted to further investigate its involvement in the formation and maintenance of the Xenopus laevis ROSs. Transgenic animals were created expressing an mCherry (mC)-tagged murine full-length nudC (mC-nudC) or mutant (mC-nudCL280P) nudC as well as an mVenus (mV)-tagged short hairpin RNA (shRNA), which targets both the L homeolog and S homeolog of found nudC X. laevis photoreceptors. We found that the dominant negative mutation nudCL280P caused irregular stacking of the OS disk membranes and elongated nascent disks, suggesting the participation of nudC in the formation of disks and maintenance of the sensory cilium. This observed finding was also present and more severe in transgenic tadpoles expressing an shRNA targeting X. laevis nudC. Finally, we found that rhodopsin was mislocalized in both nudCL280P-expressing and nudC knockdown tadpoles, further supporting a role for nudC in the transport and delivery of rhodopsin to the OS. Our data demonstrated that nudC has a critical role in photoreceptor health and maintenance through interactions with proteins involved in the transport, delivery, and packaging of rhodopsin into disk membranes.
MATERIALS AND METHODS
Research animals
All animal studies were conducted in compliance with the Guide for the Care and Use of Laboratory Animals [National Institutes of Health (NIH), Bethesda, MD, USA] and were approved by the Institutional Animal Care and Use Committee of the University of Alabama at Birmingham (UAB; Birmingham, AL, USA).
Isolation of retinal proteins and immunoprecipitation experiments
Retinal extracts for immunoprecipitation (IP) were taken from bovine retinas and prepared using published methods (24, 25). Proteins were separated by SDS-PAGE gel and were identified using in-gel trypsin digestion and liquid chromatography-tandem mass spectrometry as previously described in Reish et al. (10). The presence of rab11a in these experiments was confirmed via Western blot with a polyclonal rabbit antibody from Thermo Fisher Scientific (Waltham, MA, USA). Reverse pull-down experiments were performed using an mAb against nudC produced in rabbits from Abcam (Cambridge, MA, USA), and subsequent confirmation of the presence of rab11a and rhodopsin precipitation were carried out in a similar manner to that previously described in Reish et al. (10). For confirmation of nudC presence in multiple species, retinal samples of bovine, mouse, and frog were prepared for Western blotting as described in Reish et al. (10). Dot blots were performed as Zhu et al. (20) described, using 5% X. laevis tadpole eye homogenate probed with the monoclonal nudC antibody and the standard with glyceraldehyde-3-phosphate dehydrogenase (GAPDH) antibody from Abcam. The monoclonal rhodopsin antibody B6-30N was a kind gift of W. Clay Smith (University of Florida, Gainesville, Florida).
X. laevis transgenesis
Transgenic X. laevis tadpoles were generated as described in Sparrow et al (26). The Xenopus rod opsin expression vector, or pXOP0.8, a gift from Orson Mortiz (University of British Columbia, Vancouver, BC, Canada) (27), was modified to insert I-SceI recognition sites near the PciI and AflII sites of the vector to generate pXOPO.8/I2. The vectors pCDNA3.1 mC-nudC and pCDNA3.1 mC-nudCL280P were created by amplifying nudC from mouse retinal RNA and introducing the L280P mutation by site-directed mutagenesis. The mC-nudC and mC-nudCL280P fusions were then digested with NheI/NotI and ligated into the NheI/NotI sites of pN1/I2 or the unmethylated XbaI/NotI sites of pXOP0.8/I2. For transgenesis, DNA was mixed with X. laevis sperm and injected together into eggs.
mV-miR-30–based shRNA constructs
The miR-30 knockdown system was designed after a proven strategy first used in zebrafish, in which an RNA Pol II promoter drives expression of a yellow fluorescent protein, mV, and an miR-30–based hairpin (10, 28). Previous experiments in our lab and similar experiments performed in X. tropicalis have demonstrated the ability of transgenesis to introduce stable RNA interference (29). The miR-30 recognition sequence and cloning sites were generated as previously described (10, 30). Hairpin sequences were produced using the pSM2 algorithm (http://cancan.cshl.edu/RNAi_central/RNAi.cgi?type=shRNA), and the following sequence targeting both the long and short homeologs of X. laevis nudC was selected: AACACATTCTTCAGTTTTCTG. The hairpins were created by PCR amplifying the designed oligo with the primer TGCTGTTGACAGTGAGCGCAACACATTCTTCAGTTTTCTGTAGTGAAGCCACAGATGTACAGAAAACTGAAGAATGTGTTTTGCCTACTGCCTCGGA, digesting the product with XhoI and EcoRI, and then ligation into the corresponding sites in the Xenopus rod opsin expression vectors. The hairpin targeting firefly luciferase was based on luc1309 and was similarly cloned into the Xenopus rod opsin expression vector (31).
Immunohistochemistry
Enucleated eyes from 3-wk-old mice of either sex were placed in 4% paraformaldehyde in 0.1 M PBS (pH 7.4) overnight at 4°C. Following an additional overnight incubation in 30% sucrose in PBS at 4°C, eyes were embedded in Tissue-Tek O.C.T. (Sakura Finetek, Torrance, CA, USA), flash frozen in isopentane cooled to near freezing by liquid nitrogen, and sectioned at 10-μm intervals using a cryomicrotome. For tadpole sections, whole tadpoles of either sex at 2 or 4 wk of age were fixed overnight in 4% paraformaldehyde at 4°C. Following fixation, the eyes were processed for immunohistochemistry (IHC) as previously described for the mouse eyes. Previously prepared retinal sections from Tupaia belangeri (northern tree shrew) were provided from the UAB Tree Shrew Core. Prepared Macaca mulatta (rhesus monkey) retinal sections were a gift from Dr. Paul Gamlin at UAB. For staining of retinal sections from all species, sections were treated with heat-induced epitope retrieval by boiling in 10 mM sodium citrate and 0.05% Tween (pH 6.0) for 20 min and cooling to room temperature for 30 min. Primary antibodies used were the monoclonal anti-nudC and monoclonal B6-30N, for rhodopsin, as previously described. Wheat germ agglutinin (WGA) from Thermo Fisher Scientific conjugated to Alexa Fluor 555 or Alexa Fluor 647 was used in some experiments to counterstain the photoreceptor OSs. Images of stained retinal sections were acquired on an Olympus IX81 spinning disk confocal microscope with ×60 oil-immersion objective, a PerkinElmer UltraView 6FE-US spinning disk confocal microscope (PerkinElmer, Waltham, MA, USA) attached to a Nikon TE2000-U (Nikon, Tokyo, Japan) with a ×60 oil-immersion objective, or a Zeiss LSM 800 Airyscan confocal microscope (Carl Zeiss, Oberkochen, Germany) with an ×63 oil-immersion objective attached to an Axiocam-506 CCD camera (Carl Zeiss). ImageJ software (NIH) was used to compile maximum projection z-stacks and process acquired images.
Transmission electron microscopy
Xenopus laevis eyes from 2- or 4-wk-old tadpoles were isolated and placed in 0.5× Karnovsky’s buffer, 2.5% glutaraldehyde, 2.0% paraformaldehyde, 0.2 M sodium cacodylate (pH 7.2) for at least 15 min. Fixed eyes were then microwaved at 120 W for 80 s, followed by 350 W for 40 s. Eyes were then placed on a rotator at room temperature for 15 min followed by 3 washes in 0.1 M sodium cacodylate buffer (pH 7.2). Samples were then postfixed in 1% osmium and 1.5% ferrocyanide in 0.2 M sodium cacodylate buffer (pH 7.2) and then microwaved at 120 W for 80 s, followed by 120 W for 40 s, then incubated on a rotator at room temperature in the dark for 30 min. After this incubation, samples were washed 3 times with millipore water. The X. laevis eyes were then placed in 1% aqueous uranyl acetate and microwaved at 120 W for 80 s and then at 120 W for 40 s. The samples for transmission electron microscopy (TEM) were then washed 3 times with millipore water and dehydrated with successive increasing concentrations of acetone, microwaving samples at 120 W for 40 s after each acetone wash. Finally, the isolated eyes were infiltrated with resin and the tissue was embedded with 100% Epon 812 in a 60–70°C oven overnight. Ultrathin sections were then acquired and imaged in the UAB high-resolution imaging facility using an FEI Tecnai T12 Spirit 20–120-kv Transmission Electron Microscope (Thermo Fisher Scientific) with a digital camera (Advanced Microscopy Techniques, Woburn, MA, USA).
Statistical analyses
All statistical analyses were performed using SPSS software (IBM SPSS, Chicago, IL, USA). All biochemistry experiments were carried out at least 3 times yielding similar results. For all immunohistochemical or transgenic studies where n is reported, the n represents the number of separate animals. For transgenic tadpole studies, animals with morphologic defects were excluded from analyses and no other method to exclude outliers from data sets was utilized. Dot blots were quantified with ImageJ utilizing GAPDH as a standard, analyzed with a 1-way ANOVA, and then a Tukey’s honest significance test for post hoc comparison. The number of overgrown disks for each group was tested against wild type using the Fisher’s exact test, with a significance defined as a value of P < 0.05.
RESULTS
Coimmunoprecipitation studies on rab11a using bovine retinas identify a novel interaction with nudC
We previously identified the small GTPase rab11a as a binding partner of rhodopsin whose interaction depends on the integrity of rhodopsin’s carboxy terminus (10). To identify additional members of the rab11a rhodopsin complex, we performed IP pull-down experiments with rab11a antibodies and bovine retinal extract to capture candidate proteins. In addition to rab11a pulling down rhodopsin, another major band (∼40 kDa) was observed in the Coomassie-stained SDS-PAGE gel (Fig. 1A). Following proteolytic digestion of the excised gel slice, mass spectrometric analysis was performed and the protein was identified as nudC. The presence of rab11a in pull-down fractions for this experiment was confirmed via Western blot (Fig. 1B). NudC as an additional protein involved in the rab11a rhodopsin trafficking complex was further confirmed through reverse pull-down experiments using bovine retinal extract and a monoclonal nudC antibody as bait. Western blot analysis of the reverse pull-down fractions demonstrated that both rab11a and rhodopsin were pulled down in these experiments (Fig. 1C). Next, we used Western blotting to identify nudC expression in bovines, mice, and frogs. Retinal extracts from all species tested showed a single band (∼40 kDa in mice and bovines and ∼38 kDa in frogs) when probing the blot with monoclonal nudC antibodies (Fig. 1D).
NudC binds to rab11a and rhodopsin in bovine retinas and is present in mouse and frog retinas. A) IP using an anti-rab11a pAb antibody as bait followed by SDS-PAGE revealed 2 proteins binding to rab11a, which were identified by mass spectroscopy as nudC and rhodopsin. B) Western blot probed for rab11a following IP of bovine retinal lysates using rab11a antibody as bait. C) Western blots from reverse pull-down experiments using bovine retinal extract and a nudC antibody as bait support an interaction between nudC, rab11a, and rhodopsin. D) Western blotting confirms nudC is found in retinal extracts prepared from bovines, mice, and frogs. Ab, rab11a polyclonal antibody eluate; C, control IgG; IB, immunoblotting; MW, molecular weight.
IHC and transgenic expression of nudC demonstrates similar localization in photoreceptors of all species tested
Immunofluorescence with monoclonal antibodies against nudC in wild-type mouse retinal sections displayed pan-retinal distribution with signal detected in the RPE cells, the photoreceptor IS, outer nuclear layers (ONLs), inner nuclear layers and at synaptic terminals (Fig. 2A). Retinal sections from wild-type tree shrew (Fig. 2B) and rhesus monkey (Fig. 2C) demonstrated intense IS labeling of nudC. In wild-type tadpole retinal sections, nudC is primarily localized within the IS region and RPE cell layer, with limited signal detected within the ONL and inner nuclear layer (Fig. 2D). A concentrated IS localization of nudC was the most prominent signal observed in the photoreceptors of all species, and most species displayed no signal within the OS. The exception was the cone dominant tree shrew retina, which showed prominent signal along the ciliary axoneme of the OS (Fig. 2B, >).
NudC and nudCL280P localizes to the IS and ONL of photoreceptors and can be knocked down upon expression of shRNA against nudC. A–D) IHC of retinal cryosections probing with antibody indicates endogenous nudC (red) is expressed throughout the retinas of mice (A) and is primarily localized to the IS of tree shrew (B), monkey (C) and frog (D) retinas. Photoreceptors are counterstained with WGA (green). NudC stains the axoneme of cone cells (B). A′–D′) View of endogenous nudC IHC shown without counterstains. E, F) Expression and localization of mC-nudC (E, yellow) or mC-nudCL280P (F, yellow) in transgenic X. laevis, counterstained with WGA (green). E′, F′) View of mC-nudC (E′) or mC-nudCL280P (F′) localization shown without counterstains. G) IHC from transgenic tadpole retinas expressing nudC shRNA (green) probing with anti-nudC (red). ^, successful depletion of nudC protein in cells expressing nudC shRNA (*). H) NudC (red) protein levels are unaffected by expression of shRNA targeting firefly luciferase (green). G′, H′) View of nudC IHC with nuclear counterstain alone. DAPI (blue), nuclear marker. OPL, outer plexiform layer. Scale bar, 20 μm.
Similar to the IHC results, transgenic expression of the fusion protein mC-nudC in tadpoles showed distribution throughout the IS (Fig. 2E) and presence at the synapse of transgenic rod photoreceptors (unpublished data) with undetectable levels of expression in the OS. The subcellular distribution of mC-nudCL280P in transgenic tadpole rod photoreceptors was primarily observed in the IS, typically as bright puncta, and in the perinuclear region. The mC-nudCL280P transgene was not observed within the OS (Fig. 2F) nor at the synapse (unpublished data).
shRNA expressed under the X. laevis opsin promotor knocks down nudC in transgenic tadpoles
To further characterize the role of nudC in rod cells, we created an shRNA directed against both X. laevis nudC homeologs expressed under the Xenopus opsin promotor. The shRNA was produced with an mV fluorescent marker to demarcate the cells expressing the transgene. By immunohistochemically counterstaining for nudC in the retinas from 2-wk-old tadpoles, we showed a lack of nudC within transgenic rod cells (Fig. 2G, ^) expressing the nudC shRNA (Fig. 2G, *). Expression of nudC protein was not affected by the control shRNA targeting firefly luciferase (Fig. 2H). Dot blot quantification and analysis showed lower nudC expression in animals with shRNA against nudC compared with nontransgenic (NTG) animals (Fig. 3A). Additionally, there appears to be an increase of nudC in transgenic tadpoles expressing mC-nudC or mC-nudCL280P when compared with wild type (Fig. 3B). Statistical analysis of all groups with a 1-way ANOVA demonstrated an F(4,27) = 4.64, P = 0.007; n = 4–7 per group. Tukey’s honest significant difference test (HSD) test for mC-nudCL280P vs. hairpin found a value of P < 0.01 and all other comparisons were P > 0.05. Thus, whereas expression of nudC shRNA trended to decrease the overall levels of nudC relative to wild type, statistical analyses only showed a significant difference between mC-nudCL280P and the shRNA knockdown (*). The control shRNA, mV-firefly luciferase shRNA, did not alter nudC levels compared with wild-type tadpoles. Rescue of nudC knockdown via shRNA was seen when mC-nudC was coexpressed in transgenic tadpoles returning overall nudC protein concentration to approximate wild-type levels. Relative nudC amounts in lysates from whole tadpole eyes were normalized to GAPDH.
Dot blots of whole tadpole eyes demonstrate a decrease of nudC in shRNA-transgenic X. laevis photoreceptors. Expression of mC-nudCL280P or nudC-shRNA leads to rhodopsin mislocalization, which is resolved when coexpressing nudC-shRNA and mC-nudC. A) Dot blot showing nudC-expression changes in NTG, mC-nudC, mC-nudCL280P, nudC-shRNA, nudC-shRNA plus mC-nudC rescue, and luciferase shRNA control transgenic animals. B) Quantification of relative nudC amounts in lysate from whole tadpole eyes with expression normalized to GAPDH (n = 4–7 per group; 1-way ANOVA, F(4,27) = 4.64, P = 0.007; Tukey’s HSD test for mC-nudCL280P vs. hairpin, P < 0.01 (*); all other comparisons, P > 0.05). C–F) IHC for rhodopsin (green) in retina cryosections of transgenic X. laevis expressing mC-nudC (C, red), mC-nudCL280P (D, red), nudC-shRNA (E, red), and coexpression nudC-shRNA (cyan) and mC-nudC (red) as a rescue experiment (F). (D, E) Intense band of rhodopsin staining at the base of the OS (^). C′-F′) View of rhodopsin IHC without counterstains highlighting protein mislocalization. E) Aberrant rhodopsin in the IS (*). Nuclei marked by DAPI (blue). Scale bars, 20μm.
NudCL280P and shRNA knockdown of nudC in X. laevis photoreceptors results in rhodopsin mislocalization
To determine if the expression of mC-nudC or mC-nudCL280P altered the subcellular localization of rhodopsin, we stained retinal sections of transgenic tadpoles with a monoclonal antibody against rhodopsin (B6-30N) and visualized the sections using confocal microscopy. Rhodopsin localized normally to the OS with little to no staining in the IS of animals expressing the mC-nudC fusion protein (Fig. 3C). However, in animals expressing the mC-nudCL280P fusion protein, an intense banding pattern of rhodopsin staining was noticed at the base of the OS, near the site of new disk formation (Fig. 3D, ^). Animals expressing mC-nudCL280P also showed increased IS staining for rhodopsin. Transgenic expression of the nudC shRNA resulted in rhodopsin mislocalization in the IS (Fig. 3E, *), as well as the distinctive banding staining of rhodopsin at the base of the OS (Fig. 3E, ^). Rescue experiments that coexpressed nudC shRNA and mC-nudC resolved the rhodopsin mislocalization and the intense banding at the base of the OS (Fig. 3F). There were no differences detected in OS length or overall retinal health between mC-nudC and mC-nudCL280P animals at 4 wk postfertilization (not shown).
Altered expression of nudC and continued trafficking of rhodopsin contributes to membrane disruption and the overgrowth of disk evaginations at the base of the OS
To uncover the etiology of the intense rhodopsin banding seen at the proximal OS in Figs. 3D (^) and E (<), we determined the ultrastructure of the IS and OS of the transgenic tadpole retinas by TEM. At 2 and 4 wk post fertilization NTG tadpole retinas appear healthy with consistently sized disk membranes stacked in the sensory cilium (Fig. 4A, D). TEM images from animals expressing mC-nudC showed photoreceptor OS that were comparable to NTG animals with only 8% (2/25, P = 0.606, Fisher’s exact test) of photoreceptors imaged showing a disruption of disk stacking in the sensory cilium at 2 and 4 wk postfertilization (Fig. 4B, E). However, 58% (30/52, P < 0.00001, Fisher’s exact test) of the OS of mC-nudCL280P transgenic tadpoles showed dysregulated disk membranes with ectopic disk spiraling out of the OS at 2 wk postfertilization (Fig. 4C, ←) and at 4 wk postfertilization (Fig. 4F, →).
Transgenic expression of mC-nudCL280P but not mC-nudC causes dysmorphic OS disk membranes. A, B) Ultrathin sagittal sections taken from 2-wk-postfertilization wild-type retinas (A) and mC-nudC transgenic retinas (B) show (^) properly formed rod and cone photoreceptors using TEM. C) X. laevis mC-nudCL280P photoreceptors contain overgrown disc membranes and ectopic disc formation (arrow) at 2 wk postfertilization. D–F), Ultrathin sagittal retinal sections from 4-wk-postfertilization wild-type retinas (D) and mC-nudC (E) contain photoreceptors with normal disks formed (^) in the OS, whereas mC-nudCL280P 4-wk animals (F) show ectopic disk overgrowth (arrow). Scale bars, 500 nm. Asterisk (*) indicates oil droplets. WT, wild type.
Knockdown of nudC resulted in the formation of disrupted, tubulovesicular membranes at the base of the OS at 2 wk postfertilization (Fig. 5A, #). At 4 wk postfertilization, the disrupted membranes at the base of the OS were absent, however evagination overgrowths of the OS were observed (Fig. 5A, ←). Taken together, 92% (68/74, P < 0.00001, Fisher’s exact test) of all photoreceptors imaged from tadpoles expressing nudC shRNA displayed dysregulated OS disks.
Knockdown of nudC in rod photoreceptors results in overgrowth of disk membranes and ectopic disk formation. A–D) 2-wk-postfertilization tadpole ultrathin retina sections imaged by TEM. A) Transgenic tadpoles expressing nudC shRNA show dysregulation of new disk formation at the base of ROSs. An accumulation of tubulovesicular membranes near the base of ROSs was observed (#). A′) Tubulovesicular membrane accumulation at the base of the OS in nudC shRNA retinas has been resolved; however, overgrowth of discs at the base of the OS is observed (arrows). B) Phenotype associated with expression of nudC shRNA is rescued by coexpression of mC-nudC with no dysregulation of disk membranes observed. B′) Rescue of nudC shRNA via coexpression of mC-nudC results in normal formation of ROSs based on ultrastructure. C) NudC shRNA is not rescued by coexpression of mC-nudCL280P, and large overgrown disks are present at base of OS (arrow). C′) Animals coexpressing mC-nudCL280P and nudC shRNA do not show normal OS disk membranes and have ectopic disks formed at the base of the OS (arrows). D) Control shRNA directed against firefly luciferase showed no apparent phenotype in X. laevis photoreceptors. A′–D′) Ultrathin sections from 4-wk-postfertilization tadpole retinas imaged by TEM. D′) Knockdown of firefly luciferase via shRNA shows no adverse effects of OS disk formation. Carets (^) indicate normal OS disk ultrastructure. Asterisk (*) shows oil droplets contained in X. laevis photoreceptors. Scale bars, 500 nm.
Similar to our previous experiments, coexpression of mC-nudC rescued the phenotype after knockdown of nudC by shRNA and abnormal membrane formations or evagination overgrowths were seen only 13% (11/82, P = 0.1315, Fisher’s exact test) of the time in 2- and 4-wk animals (Fig. 5B). The overgrown disk phenotype was not rescued by coexpressing nudC shRNA with mC-nudCL280P, however we did observe a decrease in abnormal membrane formation with 35% (24/67, P = 0.00001, Fisher’s exact test) of photoreceptors imaged displaying dysregulated OSs. Although abnormal OS disk were commonly observed in 2-wk animals (Fig. 5C, ←), a partial rescue seems evident at 4-wk post fertilization with rod cells appearing to have smaller evagination overgrowths and a reduction in disk stacking defects (Fig. 5C, ←). No observable phenotypes were detected in animals expressing the control shRNA hairpin (Fig. 5D).
DISCUSSION
A major goal of protein biology is to understand the roles proteins play within an individual cell and how they contribute to overall cell health. The role a given protein has in 1 cell type may vastly differ from the role it plays in other types of cells. NudC has been shown to be actively involved in many processes needed for cellular development and replication; however a function in the mature, postmitotic rod photoreceptor was unknown prior to this study. Here we report the presence and localization of nudC in vertebrate photoreceptors and propose a novel role of nudC in disk membrane formation and rhodopsin trafficking.
NudC was first identified in our pull-down experiments using rab11a antibodies as bait, which eluted both nudC and rhodopsin from bovine retinal extract (Fig. 1A), as well as rab11a (Fig. 1B). We also performed the reverse pull-down experiment using nudC antibodies as bait which eluted rhodopsin and nudC from mouse retinal extract (Fig. 1C). Next, we used a Western blot to confirm nudC was expressed in several species of vertebrate retinas, suggesting an evolutionarily conserved role for nudC in retinal cells (Fig. 1D).
The interactions we detected between nudC, rab11a, and rhodopsin argue that these proteins are part of the same trafficking pathway. However, there are conflicting data in the literature regarding species-specific roles of rab11a in photoreceptors which are difficult to reconcile with this work and our previous efforts (32). A closely related protein to rab11a, rab11b, may have differential expression across species and may account for the difference in the mouse data compared with X. laevis. Ying et al. attempted to compensate for this by expressing the S25N mutation of rab11b in rab11a knockdown mice; however, our own data demonstrate that the S25N mutation of rab11a is only mildly if at all toxic in Xenopus rods (10, 32). We consider that the interaction of rab11a and nudC is worthy of investigation from the perspective of cell biology generally and in photoreceptors specifically as this interaction has not previously been reported in the literature.
IHC revealed the presence of nudC throughout mouse, tree shrew, monkey, and frog retinas including within the photoreceptors. NudC primarily localized to the IS near the microtubule-organizing center at the base of the cilium (Fig. 2A–D). Because the cellular machinery necessary to synthesizes new protein is located within the IS and proteins destined for the OS are trafficked by dynein and dynactin to the base of the CC and the nud family of genes are associated with dynein motor proteins and dynactin, these IHC results were as expected. We suspect that the nudC, rab11a and rhodopsin interaction is important for newly synthesized rhodopsin to be trafficked apically from the IS to the CC and OS.
It is also possible that nudC’s role in the transport of rhodopsin is in post-Golgi trafficking toward the OS via participation in other previously described mechanisms such as the contact between the COOH-terminal cytoplasmic tail of rhodopsin and the dynein light chain, Tctex-1; this potential interaction should be investigated in the future (33). Because nudC showed relatively little localization to the ROS we suspect that its interaction is completed after rhodopsin is packaged into newly formed disk membranes. In contrast to the other species tested, the cone dominant tree shrew retina exhibited extended axonemal labeling of nudC (Fig. 2B, ^). This could indicate nudC has a different role in the formation and maintenance of the OS in cone cells and would be an interesting focus of future investigations.
Similar to the results of IHC, transgenic animal expressing mC-nudC confirmed nudC localization to the IS (Fig. 2E). However, the expression of mC-nudCL280P (Fig. 2F) resulted in more punctate signal within the IS compared with the more dispersed signal detected in animals expressing mC-nudC (Fig. 2E). This punctate signal may be the result of the known defects nudCL280P has on dynein mediated movement, which could restrict the migration of mC-nudCL280P and its cargo within the IS of photoreceptors.
Dot blot analyses revealed a non-significant reduction of nudC in animals expressing nudC shRNA compared with NTG animals (Fig. 3A, B) and there are 2 possibilities that could explain this result. First, due to the mosaic expression pattern of transgenes in the allotetraploid genome of X. laevis, some rod cells do not express the shRNA. This is a common limitation of this model and can dilute the effect of the shRNA knockdown when examining protein extraction (34). A second possibility is that because nudC is expressed in other cells of the retina and the shRNA under the rod opsin promotor is only expressed in rod cells, the effects of the knockdown may not be as apparent when examining preparations of whole retina lysates. Regardless, knockdown of nudC via shRNA was confirmed by the lack of nudC antibody labeling (Fig. 2G, ^) in rod cells expressing nudC shRNA (Fig. 2G, *). Cells not expressing the shRNA targeting nudC were observed in the same retinas and the IS was labeled by the nudC antibody (Fig. 2G). Additionally, nudC antibody staining in the IS of rod cells expressing the control shRNA targeting firefly luciferase shows that nudC levels were unaffected in these cells (Fig. 2H).
Rhodopsin localization in transgenic mC-nudC tadpoles did not show any aberration compared with wild type (Fig. 3C). However, expression of mC-nudCL280P resulted in mislocalization of rhodopsin in the IS of transgenic rods (Fig. 3D). Interestingly, we observed an increase in intensity of rhodopsin staining at the base of the OS in the mC-nudCL280P animals (Fig. 3D, ^). The distinctive band of rhodopsin labeling suggests a dysregulation of disk formation or an unusual accessibility of rhodopsin antibody in this region due to presence of nudCL280P. This may be the result of impaired vesicle transport to or from the Golgi, as nudC is known to be involved in dynein movement of organelles and vesicles within cells, or due to dysregulation of the F-actin network at the base of the OS. Upon knockdown of nudC, we also detected high levels of rhodopsin mislocalization in the IS (*), as well as, the intense band of rhodopsin (^) staining at the proximal base of the OS similar to the mC-nudCL280P tadpole suggesting a similar mechanism of impaired vesicular transport (Fig. 3E). By coexpressing both the nudC shRNA and mC-nudC, we demonstrate a rescued phenotype with normal localization of rhodopsin and resolution of the significant band at the base of the OS (Fig. 3F). These data suggest a critical role for nudC in rhodopsin vesicle trafficking and the formation disks in the OS.
Our next step was to investigate the ultrastructure of the disks at the base of the ROS in transgenic tadpoles with the knockdown of nudC expression using TEM. TEM further confirmed the disruption of the IS/OS divide and suggests that the intense banding of rhodopsin at the base of these transgenic animals may be caused by an overgrowth of newly formed disk. The animals expressing mC-nudCL280P displayed a phenotype change with dysregulated disk membranes and some ectopic disks at 2 wk (Fig. 4C, ←) and 4 wk postfertilization (Fig. 4F, ←). This phenotype may be caused by the presence of endogenous nudC still present at normal levels with the variable mosaic expression of the mC-nudCL280P transgene in the photoreceptors.
Animals expressing the nudC shRNA, in which endogenous nudC levels were decreased compared with wild type, revealed a dramatic increase in membranous material forming tubulovesicular membranes at the base of the OS in 2 wk–postfertilization animals (Fig. 5A, #). At 4 wk postfertilization, in animals with nudC shRNA, these extraneous membranes were resolved, but large evagination outgrowths were present, which spiraled out of the OS (Fig. 5A, ←). Overall, retinas from animals expressing the nudC shRNA consistently displayed tubulovesicular membranes and disk overgrowth. As mentioned earlier, because of the allotetraploid nature of X. laevis and incomplete transgene expression, phenotypically normal rods were also seen in retinas expressing the nudC shRNA. However, in all transgenic retinas from animals expressing mC-nudCL280P (n = 5), the shRNA targeting nudC (n = 4), or both the nudC shRNA and mC-nudCL280P (n = 5) displayed defective membranes when observed by TEM. Comparatively, NTG (n = 5), shRNA luciferase control (n = 4), and mC-nudC (n = 4) transgenic retinas as well as rescue animals expressing both nudC shRNA and mC-nudC (n = 5) had significantly fewer defects in OS membrane structure. The limited defects observed in these transgenic retinas were most likely caused by artifacts of fixation.
A rescued phenotype was demonstrated when nudC shRNA and mC-nudC were coexpressed in transgenic animals (Fig. 5B, B). The numbers of observed defective rod cells were similar to those of NTG animals. However, in 2 wk–postfertilization transgenic tadpoles expressing both nudC shRNA and mC-nudCL280P, there were considerably larger outgrowths of disk membranes seen at the base of the OS compared with retinas expressing only transgenic mC-nudCL280P. We suspect that, in rod photoreceptors, nudCL280P is a dominant negative mutant that leads to the semidominant phenotype of slightly overgrown disk membranes from the OS because of the presence of endogenous nudC. The semidominant phenotype is exacerbated when endogenous nudC levels are reduced through the expression of nudC shRNA. Comparable evagination overgrowths were seen at 2 wk postfertilization but seemed to be partially resolved by 4 wk postfertilization (Fig. 5C). The reduced severity of disk overgrowth observed at 4 wk may be the result of decreased expression of the transgene within the cell or possibly the activation of compensatory pathways sufficient to decrease disk malformation.
Our data suggest that nudC interacts with rhodopsin and is involved with its trafficking; however, nudC’s role may be more complex, such as regulating the growth and incorporation of rhodopsin into newly forming disk membranes at the base of the OS (Fig. 6). For instance, our data show that dysregulation of newly formed disks is strikingly similar to a study by Williams et al., which showed evagination outgrowths from the base of the OS following inhibition of F-actin dynamics by cytochalasin D treatment (35, 36). Further studies are needed to connect the mechanism by which nudC regulates F-actin dynamics and how dysregulation of F-actin dynamics results in overgrown disk in the OS.
Schematic representation of the regulation of newly formed disk membranes by nudC in X. laevis rod photoreceptors. NudC localizes primarily to the IS and synapse in normal rods with uniform disk membranes in the OS. NudCL280P is present in the IS but does not localize to the synapse. Newly formed disks grow ectopically, spiraling out of the OS, either contained within or external to the plasma membrane. Knockdown of nudC results in drastically overgrown disk membranes and the mislocalization of rhodopsin. Arf, ADP ribosylation factor; ASAP, arf-GAP with SH3 domain, ankryn repeat and pH domain containing protein; NudC, nuclear distribution protein C; rab, ras-related protein.
Additionally, formation of vesicular tubular structures in the OS of nudC shRNA knockdown animals may be caused by impaired autophagy, the process by which rod cells regulate the synthesis, degradation, and recycling of their products. This process is completed through the lysosomal machinery and is dependent on the intracellular movement of vesicles and organelles (37). It is well documented that rod cells of X. laevis undergo many changes during development, such as the required removal of their oil droplets and the incorporation of lipid membrane into their OS (38, 39). Rod cells expressing either nudCL280P or the shRNA knockdown under the rod opsin promoter were observed, with evagination overgrowth and highly dysregulated membranes that contained oil droplets (Fig. 4C) near healthy cones containing oil droplets (Fig. 5A). Thus, the loss of nudC or the expression of nudCL280P may result in the inability of rod cells to lose their oil droplet and transition to the mature rod state. Although little is known about the molecular mechanisms regulating oil droplet removal from rod cells, our data suggest that nudC has a role in degrading these lipids either directly or by regulating the movement of the organelles responsible for oil droplet removal. The Xenopus opsin promoter used in all plasmids was sequenced, verified in accordance with other previous work (40, 41), and should not express in cones, suggesting the observed cells containing overgrown disk and an oil droplet were, in fact, rod cells that do not fully mature when nudC expression is altered. There is also a possibility that the shRNA is being transferred or delivered to cone photoreceptors through Muller cells or by some other method. Further investigation is necessary to fully understand the role nudC plays in rod cell development and the transition from an early photoreceptor to a mature rod cell.
Other studies have shown that nudC is expressed in motile cilia (42) and, more recently, in nonmotile cilia (19). In 1 study, Zhang and colleagues showed that knockdown of nudC or cofilin 1, an actin-modulating protein, causes panciliary defects in zebrafish. This includes increased ciliary length in RPE cells but no described retinal dysfunction (19). Our data do not show a statistically significant increase in OS ciliary length in transgenic rods expressing shRNA against nudC, presumably because of the inherent regulation of disk length caused by RPE phagocytosis (2, 43) (unpublished data). Despite these reports, there have been limited studies on the functions of nudC in ciliary trafficking, and no studies have investigated the role of nudC in GPCR trafficking or its regulation of the actin cytoskeleton in the formation of OS disk membranes. Thus, we may have identified a novel component of the rhodopsin GPCR-trafficking pathway and photoreceptor disk formation.
Further investigation into the interaction of nudC with rab11a and rhodopsin could be helpful in understanding how rab11a mediates its regulatory roles in recycling-endosome trafficking and Golgi-to-cilium trafficking. This research could also provide insight as to why nudC, a protein known for its role in cytokinesis, is expressed throughout the postmitotic cells of the retina. Overall, further investigation of nudC’s regulation of the actin cytoskeleton in rod photoreceptors may shed light on the enigmatic molecular mechanisms underlying OS disk formation.
ACKNOWLEDGMENTS
The authors thank Melissa F. Chimento and the University of Alabama at Birmingham (UAB) High Resolution Imaging Facility for their help with transmission electron microscopy sample preparations and image acquisition, Russell Veale and UAB Tree Shrew Core for their kind gift of T. belangeri retinal sections and Paul Gamlin (UAB) for the gift of retinal cryosections from M. mulatta, David Redden and the UAB Center for Clinical and Translational Science for statistical support, TJ Hollingsworth (University of Tenessee Health Science Center, Memphis, TN, USA) for artistic assistance, Xiaogang Cheng (UAB) for technical assistance, and Elizabeth Sztul and Skyler Boehm (both of UAB) for helpful discussions. This work was supported by the U.S. National Institutes of Health (NIH), National Eye Institute Grants R01EY019311 (to A.K.G) and P30EY003039 (to the UAB Vision Science Research Center), and a grant from the E. Matilda Ziegler Foundation for the Blind (to A.K.G.). The authors declare no conflicts of interest.
Glossary
| CC | connecting cilium |
| GAPDH | glyceraldehyde-3-phosphate dehydrogenase |
| IHC | immunohistochemistry |
| IP | immunoprecipitation |
| IS | inner segment |
| mC | mCherry |
| mV | mVenus |
| NTG | nontransgenic |
| nudC | nuclear distribution protein C |
| ONL | outer nuclear layer |
| OS | outer segment |
| ROS | rod OS |
| RPE | retinal pigmented epithelial |
| shRNA | short hairpin RNA |
| TEM | transmission electron microscopy |
| UAB | University of Alabama at Birmingham |
| WGA | wheat germ agglutinin |
AUTHOR CONTRIBUTIONS
All authors helped design experiments, analyze data, and write the manuscript; E. R. Boitet and N. J. Reish performed research; A. K. Gross contributed reagents and other resources; and M. G. Hubbard was essential in statistical analyses and editing the manuscript.