cGMP/Protein Kinase G Signaling Suppresses Inositol 1,4,5-Trisphosphate Receptor Phosphorylation and Promotes Endoplasmic Reticulum Stress in Photoreceptors of Cyclic Nucleotide-gated Channel-deficient Mice*

Background: Cone photoreceptors undergo endoplasmic reticulum stress-associated apoptosis in CNG channel deficiency. Results: Suppressing cGMP/PKG signaling enhances inositol 1,4,5-trisphosphate receptor 1 (IP3R1) phosphorylation and inhibits endoplasmic reticulum stress and cone death. Conclusion: cGMP/PKG signaling regulates IP3R1 activity and promotes endoplasmic reticulum stress in CNG channel deficiency. Significance: Understanding of the mechanism(s) of photoreceptor degeneration is essential for therapeutic strategy development. Photoreceptor cyclic nucleotide-gated (CNG) channels play a pivotal role in phototransduction. Mutations in the cone CNG channel subunits CNGA3 and CNGB3 are associated with achromatopsia and cone dystrophies. We have shown endoplasmic reticulum (ER) stress-associated apoptotic cone death and increased phosphorylation of the ER Ca2+ channel inositol 1,4,5-trisphosphate receptor 1 (IP3R1) in CNG channel-deficient mice. We also presented a remarkable elevation of cGMP and an increased activity of the cGMP-dependent protein kinase (protein kinase G, PKG) in CNG channel deficiency. This work investigated whether cGMP/PKG signaling regulates ER stress and IP3R1 phosphorylation in CNG channel-deficient cones. Treatment with PKG inhibitor and deletion of guanylate cyclase-1 (GC1), the enzyme producing cGMP in cones, were used to suppress cGMP/PKG signaling in cone-dominant Cnga3−/−/Nrl−/− mice. We found that treatment with PKG inhibitor or deletion of GC1 effectively reduced apoptotic cone death, increased expression levels of cone proteins, and decreased activation of Müller glial cells. Furthermore, we observed significantly increased phosphorylation of IP3R1 and reduced ER stress. Our findings demonstrate a role of cGMP/PKG signaling in ER stress and ER Ca2+ channel regulation and provide insights into the mechanism of cone degeneration in CNG channel deficiency.

specific neural retina leucine zipper transcriptional factor, promotes differentiation of rods, and NRL deficiency produces a cone-only retina (14). Because cones include only 2-3% of the total photoreceptor population in the wild-type mouse retina, the use of the Cnga3 Ϫ/Ϫ /Nrl Ϫ/Ϫ and Cngb3 Ϫ/Ϫ / Nrl Ϫ/Ϫ mouse lines enabled us to explore the cellular alterations and biochemical events in CNG channel-deficient cones to understand the mechanism(s) of cone degeneration.
In addition to ER stress, CNG channel deficiency is associated with elevation of cellular cGMP levels. Retinal cGMP levels in Cnga3 Ϫ/Ϫ /Nrl Ϫ/Ϫ mice sharply increased at postnatal day 8 (P8), peaked around P10 -15, remained high through P30 -60, and returned to near control levels at P90 (15). The cGMP elevation pattern correlated with apoptotic cone death (8,12). Using Cnga3 Ϫ/Ϫ /Gucy2e Ϫ/Ϫ mice lacking retinal guanylyl cyclase 1 (retGC1), an enzyme responsible for biosynthesis of cGMP in photoreceptors, we showed that cGMP accumulation contributed to cone death in the absence of CNG channels. Cone density and expression levels of cone proteins were significantly increased in Cnga3 Ϫ/Ϫ /Gucy2e Ϫ/Ϫ mice, compared with Cnga3 Ϫ/Ϫ mice (15). We also showed that the activity and expression levels of cGMP-dependent protein kinase (protein kinase G, PKG) were significantly increased, suggesting a potential role of cGMP/PKG signaling in cone death.
In this study, we investigated whether cGMP/PKG signaling contributes to ER stress and regulates IP 3 R1 phosphorylation in CNG channel-deficient cones. Treatment with PKG inhibitors and deletion of retGC1 effectively reduced apoptotic cone death and Müller glial cell activation and increased expression levels of cone proteins in CNG channel-deficient mice. Furthermore, inhibition of cGMP/PKG signaling significantly increased phosphorylation of IP 3 R1 and reduced ER stress. Findings from this work support a cGMP/PKG-regulated, IP 3 R1-associated ER stress/apoptosis in CNG channel-deficient cones.

Experimental Procedures
Mice, Antibodies, and Other Materials-The Cnga3 Ϫ/Ϫ (9), Nrl Ϫ/Ϫ (14), Gucy2e Ϫ/Ϫ (16), and Cnga3 Ϫ/Ϫ /Nrl Ϫ/Ϫ (12,13) mouse lines were generated as described previously. Cnga3 Ϫ/Ϫ / Nrl Ϫ/Ϫ /Gucy2e Ϫ/Ϫ line was generated by cross-breeding. All mice were maintained under cyclic light (12-h light-dark) conditions. During the light cycle, cage illumination was ϳ7 foot-candles. All animal maintenance and experiments were approved by the local Institutional Animal Care and Use Committee (University of Oklahoma Health Sciences Center, Oklahoma City, OK) and conformed to the guidelines on the care and use of animals adopted by the Society for Neuroscience and the Association for Research in Vision and Ophthalmology (Rockville, MD).
Primary antibodies used in this study are listed in Table 1. Horseradish peroxidase (HRP)-conjugated anti-rabbit and anti-mouse secondary antibodies were purchased from Kirkegaard & Perry Laboratories Inc. (Gaithersburg, MD). Fluorescent goat anti-rabbit and goat anti-mouse antibodies were obtained from Invitrogen. All other reagents were purchased from Sigma, Bio-Rad, and Invitrogen.
PKG Activity Assay-PKG activity in the retinal lysate was assayed using the CycLex cGMP-dependent protein kinase assay kit (purchased from MBL International, Woburn, MA) as we described previously (15). This immune colorimetric assay analyzes PKG activity by measuring the levels of phosphorylated PKG substrate, which is phosphorylated by PKG family members PKGI and PKGII. The phosphorylated substrate is detected by the phospho-specific monoclonal antibody 10H11 that recognizes the phosphothreonine 68/119 residues on the substrate. Briefly, retinal proteins (100 g) in kinase buffer (200 l) were added to the assay plates coated with recombinant PKG substrate containing threonine 68/119 residues. The plates were incubated in the presence of Mg 2ϩ and ATP for 30 min at 30°C to phosphorylate the substrate. After washing, 100 l of the HRP-conjugated, phosphospecific antibody was added to each well, and samples were incubated for 1 h at room temperature. The wells were washed and substrate reagent was added, leading to a color change from colorless to blue catalyzed by HRP. The reaction was stopped with stop solution, which changes the color from blue to yellow, and the absorbance at 450 nm was measured with a SpectraMax 190 microplate spectrophotometer (Molecular Devices). Each reaction was performed in duplicate. Results are an average of three to four independent experiments using retinas prepared from five to six mice.
Eye Preparation, Immunofluorescence Labeling, and Confocal Microscopy-We prepared mouse eye cross-sections for immunohistochemical analysis as described previously (11). Briefly, euthanasia of mice was performed by CO 2 asphyxiation, and mouse eyes were enucleated and fixed with Prefer (Anatech Ltd., Battle Creek, MI) for 25-30 min at room temperature. The superior portion of the cornea was marked with a green dye for orientation before enucleation. Fixed eyes were then trans-ferred in 70% ethanol and stored at 4°C until the eyes were processed and embedded in paraffin. We prepared 5-m-thick paraffin sections passing vertically through the retina along the vertical meridian passing through the optic nerve head using a Leica microtome (Leica Biosystems, Buffalo Grove, IL). Immunofluorescence labeling was performed as described previously (11). Briefly, eye sections were blocked with PBS containing 5% BSA and 0.5% Triton X-100 for 1 h at room temperature. Antigen retrieval was performed by incubating tissues in 10 mM sodium citrate buffer, pH 6.0, for 30 min in a 70°C water bath. Primary antibody incubation was performed at room temperature for 2 h (see Table 1 for antibody dilutions). Following fluorescence-conjugated secondary antibody incubation and rinses, slides were mounted and coverslipped. Fluorescent signals were imaged using an Olympus FV1000 confocal laser scanning microscope (Olympus, Melville, NY) and FluoView imaging software (Olympus, Melville, NY). Fluorescence intensity of GFAP in the retinal sections was analyzed as described previously (20,21). Data were analyzed and graphed using GraphPad Prism software (GraphPad software, San Diego, CA).
TUNEL Assay-The terminal deoxynucleotidyltransferase dUTP nick end-labeling (TUNEL) was performed to evaluate photoreceptor apoptotic death as described previously (12). We used paraffin-embedded retinal sections and the In Situ Cell Death Fluorescein Detection kit (Roche Diagnostics) in this analysis. Immunohistochemical labeling was imaged using an Olympus FV1000 confocal laser scanning microscope. The total TUNEL-positive cells in the outer nuclear layer that passed through the optical nerve were counted and averaged from four sections from one eye. The averages from at least four eyes were obtained. Data were analyzed and graphed using GraphPad Prism software (GraphPad Software, San Diego).
cGMP ELISA-cGMP level in the retinal lysate was measured by ELISA using the cyclic GMP complete kit (Assay Designs, Farmingdale, NY) as we described previously (15). Briefly, dissected retinas were homogenized in 0.1 M HCl. The acidic supernatants were used, and the assays were performed per the manufacturer's instructions. We used a SpectraMax 190 microplate spectrophotometer (Molecular Devices, CA) to measure the absorbance at 405 nm. Each reaction was performed in duplicate. Results are an average of three to four independent experiments using retinas prepared from five to eight mice.

Retinal Protein Preparation, SDS-PAGE, and Western Blot
Analysis-Protein SDS-PAGE and Western blotting were performed as described previously (12). Briefly, retinas were homogenized in homogenization buffer A (20 mM HEPES-NaOH, pH 7.4, 5 mM EDTA, 320 mM sucrose, containing protease and phosphatase inhibitor mixture (catalog no. 04906387001, Roche Applied Science)), and the homogenates were centrifuged at 1000 ϫ g for 10 min at 4°C. The resulting supernatant and pellet were subjected to extraction of cytosolic/membrane proteins and nuclear protein, respectively. To separate membrane protein from cytosolic protein, the supernatant was centrifuged at 16,000 ϫ g for 30 min at 4°C, and the resulting pellet was used as membrane fraction. The nuclei protein was extracted by resuspending the pellet in buffer B (20 mM HEPES-NaOH, pH 7.4, 5 mM EDTA, 320 mM sucrose, containing protease and phosphatase inhibitor mixture as described above) and sonicating for 10 s at medium speed using an ultrasonic cell disruptor (Masonix, model XL2000) twice, allowing a 30-s recovery between disruptions, followed by incubation on ice for 1 h with gentle agitation. After incubation, the solubilized homogenate was centrifuged at 16,000 ϫ g for 35 min at 4°C, and the resulting supernatant was used as the nuclear fraction. Protein concentration of the membrane, cytosol, and nuclei preparations was determined using the protein assay kit from Bio-Rad. For protein separation and detection, retinal protein preparations were subjected to SDS-PAGE and transferred onto polyvinylidene difluoride membranes. Following blocking in 5% nonfat milk at room temperature for 1 h, blots were incubated with primary antibody overnight at 4°C (see Table 1 for antibody dilutions). After rinsing in Tris-buffered saline with 0.1% Tween 20, blots were incubated with HRPconjugated secondary antibodies (1:20,000) for 1 h at room temperature. SuperSignal West Dura Extended Duration chemiluminescent substrate (Pierce) was used to detect binding of the primary antibodies to their cognate antigens. HyBlot CL autoradiography films (Denville Scientific, Inc., Metuchen, NJ) were used to develop the target proteins, and Adobe Photoshop CS5 was used to analyze the signal density.

cGMP/PKG Signaling and Photoreceptor Degeneration
onset apoptosis (8,12). Retinas of these mice also showed a remarkable elevation of [cGMP] and increased PKG activity (12,15,22). Furthermore, we provided evidence that the cGMP accumulation contributed to cone degeneration (15). To determine whether the beneficial effects of cGMP reduction are associated with PKG signaling, we examined the effects of PKG inhibition on apoptotic cone death. We used two commonly used PKG inhibitors, KT5823 and (R p )-8-Br-cGMPS (17)(18)(19) in our study. We found that treatment with PKG inhibitor significantly reduced photoreceptor apoptosis in Cnga3 Ϫ/Ϫ /Nrl Ϫ/Ϫ mice. As shown in Fig. 1A, PKG activity in Cnga3 Ϫ/Ϫ /Nrl Ϫ/Ϫ retinas was strongly reduced by PKG inhibitors. Subsequently, the increased TUNEL labeling was significantly reduced in KT5823-treated mice and was nearly completely abolished in (R p )-8-Br-cGMPS-treated mice (Fig. 1, B and C). In addition, caspase-7 cleavage was abolished (Fig. 1D), and CHOP expression was reduced (Fig. 1E) in Cnga3 Ϫ/Ϫ /Nrl Ϫ/Ϫ mice treated with PKG inhibitor, compared with vehicle-treated controls.
We also examined the effects of PKG inhibition on the activity of cAMP-response element-binding protein (CREB), which is a known substrate of PKG. We found that the level of phospho-CREB was significantly reduced in Cnga3 Ϫ/Ϫ /Nrl Ϫ/Ϫ mice treated with (R p )-8-Br-cGMPS, compared with vehicle-treated controls (Fig. 1F), supporting the effectiveness of PKG inhibitor treatment.
IP 3 R plays a critical role in the cellular Ca 2ϩ regulation, including ER Ca 2ϩ homeostasis and responses to cellular Ca 2ϩ perturbation and stress. We have shown an increased IP 3 R1 phosphorylation in Cnga3 Ϫ/Ϫ /Nrl Ϫ/Ϫ retinas (12). In this study, we examined the effects of PKG inhibitor on IP 3 R1 phosphorylation. The level of phospho-IP 3 R1 was about 30% higher in Cnga3 Ϫ/Ϫ /Nrl Ϫ/Ϫ mice than in Nrl Ϫ/Ϫ mice (Fig. 5,  A and B). Following treatment with the PKG inhibitor, the level of phospho-IP 3 R1 increased further. The level of phospho-IP 3 R1 in Cnga3 Ϫ/Ϫ /Nrl Ϫ/Ϫ mice increased by about 33% following treatment with (R p )-8-Br-cGMPS (Fig. 5A) and increased by about 12% following treatment with KT5823 (Fig. 5B), compared with vehicle-treated Cnga3 Ϫ/Ϫ / Nrl Ϫ/Ϫ controls. Thus, inhibition of PKG activity reduced ER stress and increased IP 3 R1 phosphorylation in CNG channel-deficient mice.
In this work, we also examined the activity of the other arms/ pathways of ER stress in CNG channel-deficient mice and the effects of PKG inhibition. These arms include the activating transcription factor 6 (ATF6) and the serine/threonine-protein kinase/endoribonuclease (IRE1) pathways. Activation of the ATF6 arm is characterized by cleavage of the pro-protein and production of the cleaved form. Using the anti-ATF6 antibody that recognizes both the pro-form and the cleaved form of the protein, our analysis showed the presence of the cleaved ATF6 in Cnga3 Ϫ/Ϫ /Nrl Ϫ/Ϫ retinas, which was barely detected in Nrl Ϫ/Ϫ retinas (Fig. 6A). The cleaved form of ATF6 in Cnga3 Ϫ/Ϫ /Nrl Ϫ/Ϫ retinas was increased by about 4.5fold, compared with Nrl Ϫ/Ϫ retinas (Fig. 6A). Moreover, treatment with PKG inhibitor significantly suppressed cleavage of ATF6 (Fig. 6A). Activity of the IRE1 arm was evaluated by examining the expression levels of phospho-IRE1␣ using antiphospho-IRE1␣ antibody. As shown in Fig. 6B, the level of phospho-IRE1␣ in Cnga3 Ϫ/Ϫ /Nrl Ϫ/Ϫ retinas was increased by about 1.8-fold, compared with Nrl Ϫ/Ϫ retinas. Treatment with the PKG inhibitor reduced its phosphorylation, but the differences did not reach statistical significance (Fig. 6B).

Contribution of cGMP/PKG Signaling to Apoptotic Cone
Death-This study using a PKG inhibitor and deletion of retGC1, which effectively abolished PKG activation, establishes a role for PKG signaling in cone death. Suppressing   AUGUST 21, 2015 • VOLUME 290 • NUMBER 34 cGMP production and inhibiting PKG activity significantly decreased apoptotic cone death, caspase-7 cleavage, and Müller glial cell activation in Cnga3 Ϫ/Ϫ /Nrl Ϫ/Ϫ mice. The effectiveness of PKG inhibitor treatment was evident by the strong suppression of the enzyme's activity and by effective reduction of the activation of CREB (Fig. 1). Thus, our exper-

cGMP/PKG Signaling and Photoreceptor Degeneration
imental data support the view that cGMP/PKG signaling activated in CNG channel deficiency contributes to cone death and that the toxicity of elevated cGMP is primarily mediated via PKG signaling.
The contribution of PKG signaling to photoreceptor death was previously shown in the rd1 mouse, a model of recessive retinitis pigmentosa caused by a nonsense mutation in the rod Pde6b gene. These mice also display ER stress-associated photoreceptor apoptosis (28,29). cGMP accumulation has been shown to cause cell death in rd1 mice by overactivating the CNG channel, Ca 2ϩ overload (30,31), and by activating PKG signaling. Treatment of rd1 retina with PKG inhibitor reduced apoptotic photoreceptor death and improved rod survival (32). Thus, excessive activation of PKG signaling is harmful to photoreceptors, and our work provides insights into how cGMP/ PKG signaling contributes to photoreceptor death.

Contribution of cGMP/PKG Signaling to ER Stress-CNG
channel-deficient mice display early-onset ER stress, shown by studies at the protein and mRNA expression levels. The expression levels of phospho-eIF2␣ and CHOP were increased in CNG channel-deficient mice (12). Gene expression profiling studies demonstrated up-regulation of EIF2/ER stress pathways (13). In addition to examining the levels of phospho-eIF2␣, we examined the activity of the other two arms/pathways of ER stress and detected increased cleavage of ATF6 and increased expression levels of phospho-IRE1␣ in Cnga3 Ϫ/Ϫ / Nrl Ϫ/Ϫ retinas, compared with Nrl Ϫ/Ϫ retinas. The relevance of ER stress in cone degeneration is highlighted by a recent report that links mutations in ATF6 gene with achromatopsia (33). This is the first nonphotoreceptor-specific gene thus far identified to associate with achromatopsia. ER stress was also indirectly supported by increased levels of calpains, Bcl-2/Bcl-x, FIGURE 5. Reduced levels of phospho-eIF2␣ and increased levels of phospho-IP 3 R1 in retinas of Cnga3 ؊/؊ /Nrl ؊/؊ mice treated with PKG inhibitor. Shown are representative images of the Western blot detection of phospho-eIF2␣ and phospho-IP 3 R1 in retinas of Cnga3 Ϫ/Ϫ /Nrl Ϫ/Ϫ mice treated with (R p )-8-Br-cGMPS (A) or KT5823 (B) and corresponding densitometric analysis. The relative expression levels in Cnga3 Ϫ/Ϫ /Nrl Ϫ/Ϫ mice were normalized to the values in Nrl Ϫ/Ϫ mice. Data are represented as means Ϯ S.E. of measurements from four assays using retinas prepared from 4 to 5 mice. Unpaired Student's t test was used for determination of the significance of differences between the drug-treated and vehicle-treated mice (*, p Ͻ 0.05; **, p Ͻ 0.01; ***, p Ͻ 0.001). AUGUST 21, 2015 • VOLUME 290 • NUMBER 34

JOURNAL OF BIOLOGICAL CHEMISTRY 20887
and caspase-12/caspase-7 cleavage (12). Thus, the activation of all three ER stress pathways and processing of the ER stressassociated caspases suggest that cone death in CNG channel deficiency is associated with ER stress. Experiments with the ER chemical chaperone TUDCA and the molecular chaperone 11-cis-retinal further support this view. Treatment with TUDCA effectively suppressed caspase-7 cleavage and phosphorylation of eIF2␣ and increased cone-specific protein expression levels (13). Treatment with 11-cis-retinal significantly improved cone survival, manifested as increased levels of cone proteins (data not shown). To understand the mechanism of ER stress and determine whether the elevated cGMP/PKG signaling contributes to ER stress, we examined the effects of PKG inhibitor and Gucy2e deletion. We found that inhibition of cGMP production and PKG activity effectively reduced ER stress in CNG channel-deficient mice, supporting a contribution of cGMP/PKG signaling to ER stress. Involvement of cGMP/PKG signaling in ER stress and apoptosis was previously shown in pancreatic ␤ cells and myocardial cells (34,35). This study for the first time demonstrated the regulatory role of cGMP/PKG signaling in ER stress in neural retinal cells.
Potential ER Ca 2ϩ Dysregulation and Impaired ER Function in CNG Channel Deficiency-As a Ca 2ϩ store, the ER is responsible for sequestering excess intracellular Ca 2ϩ and releasing ER luminal Ca 2ϩ stores when intracellular Ca 2ϩ levels are low. ER Ca 2ϩ homeostasis is crucial for cellular Ca 2ϩ signaling and the ER's function as a site of cellular protein processing. ER requires high levels of luminal Ca 2ϩ to maintain proper protein folding and exportation. Low luminal Ca 2ϩ is known to cause an accumulation of incorrectly folded proteins and the subsequent unfolded protein response and ER stress (36 -39). ER Ca 2ϩ homeostasis is regulated by three ER Ca 2ϩ channels/ pumps as follows: the IP 3 R and ryanodine receptors for Ca 2ϩ efflux out of the ER into the cytosol, and the sarco/endoplasmic reticulum Ca 2ϩ -ATPase for Ca 2ϩ influx into the ER (Fig. 8). The activity of these channels/pumps is primarily regulated by cellular Ca 2ϩ levels, their respective ligands, and numerous signaling molecules/pathways. As nonselective cation channels, CNG channels are the main source of the Ca 2ϩ inward currents in the outer segments of photoreceptors and play a pivotal role in the light response/adaptation and cellular Ca 2ϩ homeostasis. Cones lacking functional CNG channels likely suffer from cellular Ca 2ϩ perturbation/cytosolic Ca 2ϩ reduction. In mouse rods, shutting down the influx through the CNG channels in the light lowers free cytoplasmic Ca 2ϩ nearly 10-fold (40). The lowered cytosolic Ca 2ϩ level in CNG channel deficiency is also supported by the remarkable elevation in cellular cGMP levels, because cGMP production in photoreceptors is tightly nega-

cGMP/PKG Signaling and Photoreceptor Degeneration
tively regulated by the cytosolic Ca 2ϩ level via the guanylate cyclase-activating proteins/retinal guanylate cyclase regulatory axis (41). Cytosolic Ca 2ϩ reduction would likely lead to a compensating increase in ER Ca 2ϩ release via ER Ca 2ϩ -releasing channels (and probably a reduced ER Ca 2ϩ influx via the sarco/ endoplasmic reticulum Ca 2ϩ -ATPase pumps), which may subsequently lead to excess ER Ca 2ϩ release/ER Ca 2ϩ reduction and interfere with ER function. Nevertheless, the typical phenotypes of CNG channel deficiency are impaired opsin trafficking and opsin mislocalization (8,20), reflecting impaired ER function/protein folding, and ER stress. It is worth mentioning that there are other Ca 2ϩ channels in the inner segment membrane, including transient receptor potential channels, and these channels could also play a role in the cytosolic Ca 2ϩ regulation (42,43).
cGMP/PKG Signaling Contributes to ER Stress in CNG Channel Deficiency, Potential Involvement of the ER Ca 2ϩ Channel IP 3 R1-In addition to impaired cytosolic Ca 2ϩ homeostasis directly caused by lack of the functional CNG channels, the activated cGMP/PKG signaling may play a role in ER stress, via its regulation on ER Ca 2ϩ channels. Our results showing increased IP 3 R1 phosphorylation in CNG channel-deficient mice on GC1 knock-out background or after treatment with the PKG inhibitor support a role of cGMP/PKG signaling in ER Ca 2ϩ regulation. Expressed in all cell types, IP 3 R channels play a crucial role in cytosolic and ER Ca 2ϩ homeostasis. Reduced cytosolic Ca 2ϩ or increased inositol 1,4,5-trisphosphate levels are known to increase IP 3 R activity/release of Ca 2ϩ from the ER (44 -46). Receptor phosphorylation is also a known factor regulating the channel's activity. Although the reported data were controversial, IP 3 R phosphorylation has been shown to decrease the channel's activity and reduce Ca 2ϩ release from the ER in a variety of experimental conditions (47)(48)(49)(50)(51)(52). This work indicates increased IP 3 R1 phosphorylation by knocking out GC1 or inhibiting PKG, accompanied by reduced ER stress and apoptosis. The results suggest that cGMP/PKG signaling suppresses IP 3 R1 phosphorylation, promotes Ca 2ϩ release from the ER, and subsequently causes ER Ca 2ϩ reduction/ER malfunction. How cGMP/PKG signaling suppresses phosphorylation of IP 3 R1 in CNG channel-deficient cones remains to be determined and may involve complex mechanisms. Numerous molecules/pathways, including PKA (53,54), PKC (49), PKG (47), AKT (48,55), and ERK (52), have been reported to regulate IP 3 R phosphorylation. Furthermore, PKG has multiple targets, and cross-talks among pathways exist.
It is important to emphasize that the level of phospho-IP 3 R1 in Cnga3 Ϫ/Ϫ /Nrl Ϫ/Ϫ mice always increased relative to Nrl Ϫ/Ϫ mice (Figs. 5 and 7) (12). A possible explanation for this observation is compensatory regulation. The Ca 2ϩ release from the ER in CNG channel deficiency is likely increased (and Ca 2ϩ entry is probably decreased) to compensate for the reduced cytosolic Ca 2ϩ level, which could potentially lead to a reduction of ER Ca 2ϩ level. Indeed, alteration of ryanodine receptors, including increased phosphorylation and expression level, can be detected in CNG channel-deficient retinas (data not shown). As a consequence, the function of IP 3 R could be suppressed to protect the ER from excess Ca 2ϩ release/Ca 2ϩ depletion. The suppressive regulation of IP 3 R has been shown to protect hippocampal neurons from apoptosis both in vitro and in an animal model of ER stress (56). The protective effects of cGMP/ PKG signaling inhibition on ER stress and cone death, accompanied with increased IP 3 R phosphorylation, suggest that the regulatory suppression of IP 3 R1 in CNG channel deficiency is insufficient to compensate, likely due to the elevated cGMP/PKG signaling. Thus, the elevated cGMP/PKG signaling in CNG channel deficiency may interfere with the compensatory inhibition of IP 3 R1, leading to excess Ca 2ϩ release from the ER and ER stress. Based on our findings, we postulate that the reduced cytosolic Ca 2ϩ and increased cGMP/PKG signaling in CNG channel deficiency promotes Ca 2ϩ release from the ER, leading to ER Ca 2ϩ dysregulation/ER stress and apoptotic death (Fig. 8). FIGURE 7. Reduced levels of phospho-eIF2␣ and increased levels of phospho-IP 3 R1 in Cnga3 ؊/؊ /Nrl ؊/؊ /Gucy2e ؊/؊ retinas. Shown are representative images of the Western blot detection of phospho-eIF2␣ and phospho-IP 3 R1 in retinas of Cnga3 Ϫ/Ϫ /Nrl Ϫ/Ϫ /Gucy2e Ϫ/Ϫ , Cnga3 Ϫ/Ϫ /Nrl Ϫ/Ϫ , and Nrl Ϫ/Ϫ mice at P30 (A) and the corresponding densitometric analysis (B). The relative expression levels in Cnga3 Ϫ/Ϫ /Nrl Ϫ/Ϫ and Cnga3 Ϫ/Ϫ /Nrl Ϫ/Ϫ /Gucy2e Ϫ/Ϫ mice were normalized to the values in Nrl Ϫ/Ϫ mice. Data are represented as means Ϯ S.E. of measurements from 3 to 4 assays using retinas prepared from 4 to 5 mice. Unpaired Student's t test was used for determination of the significance of differences between the drug-treated and vehicle-treated mice (**, p Ͻ 0.01).

Presence of a Nonapoptotic cGMP/PKG-independent Death
Mechanism in CNG Channel Deficiency-We found that TUNEL-positive staining was almost completely abolished in Gucy2e-deleted mice and in PKG inhibitor-treated Cnga3 Ϫ/Ϫ / Nrl Ϫ/Ϫ mice (see Figs. 1 and 2). However, the cone rescue was only partial (Fig. 3) (15). These findings suggest that apoptosis may not be the only mechanism in cone death; a nonapoptotic, cGMP/PKG-independent mechanism may exist. The early-onset apoptotic cone death appears to be mediated mainly by cGMP/PKG signaling. The correlation between the time course of cGMP elevation and the TUNEL labeling (15) also supports this view.
In summary, this study shows that treatment with the PKG inhibitor or deletion of Gucy2e suppressed cone apoptosis and improved cone survival in CNG channel-deficient mice, demonstrating a role of cGMP/PKG signaling in cone death. ER stress was decreased by treatment with the PKG inhibitor and deletion of Gucy2e, indicating cGMP/PKG signaling dependence. Moreover, suppressing cGMP/PKG signaling increased IP 3 R1 phosphorylation, suggesting a role of ER Ca 2ϩ channel regulation in ER stress and cone death. This work provides insights into ER Ca 2ϩ channel regulation/ER stress in cone death and improves our understanding of the mechanism of cone degeneration in CNG channel deficiency.