Differential Light-induced Responses in Sectorial Inherited Retinal Degeneration*

Background: Two new rhodopsin mutations associated with the rare form sector retinitis pigmentosa (RP) have been found. Results: Characterization of both rhodopsin mutant proteins shows different progression correlating with a different behavior of rhodopsin upon light exposure. Conclusion: Light plays an important role in triggering sector RP. Significance: Other mechanisms, in addition to protein misfolding, underlie GPCR dysfunction in pathological processes. Retinitis pigmentosa (RP) is a group of genetically and clinically heterogeneous inherited degenerative retinopathies caused by abnormalities of photoreceptors or retinal pigment epithelium in the retina leading to progressive sight loss. Rhodopsin is the prototypical G-protein-coupled receptor located in the vertebrate retina and is responsible for dim light vision. Here, novel M39R and N55K variants were identified as causing an intriguing sector phenotype of RP in affected patients, with selective degeneration in the inferior retina. To gain insights into the molecular aspects associated with this sector RP phenotype, whose molecular mechanism remains elusive, the mutations were constructed by site-directed mutagenesis, expressed in heterologous systems, and studied by biochemical, spectroscopic, and functional assays. M39R and N55K opsins had variable degrees of chromophore regeneration when compared with WT opsin but showed no gross structural misfolding or altered trafficking. M39R showed a faster rate for transducin activation than WT rhodopsin with a faster metarhodopsinII decay, whereas N55K presented a reduced activation rate and an altered photobleaching pattern. N55K also showed an altered retinal release from the opsin binding pocket upon light exposure, affecting its optimal functional response. Our data suggest that these sector RP mutations cause different protein phenotypes that may be related to their different clinical progression. Overall, these findings illuminate the molecular mechanisms of sector RP associated with rhodopsin mutations.


Retinitis pigmentosa (RP) is a group of genetically and clinically heterogeneous inherited degenerative retinopathies caused by abnormalities of photoreceptors or retinal pigment epithelium in the retina leading to progressive sight loss. Rhodopsin is the prototypical G-protein-coupled receptor located in the vertebrate retina and is responsible for dim light vision.
Here, novel M39R and N55K variants were identified as causing an intriguing sector phenotype of RP in affected patients, with selective degeneration in the inferior retina. To gain insights into the molecular aspects associated with this sector RP phenotype, whose molecular mechanism remains elusive, the mutations were constructed by site-directed mutagenesis, expressed in heterologous systems, and studied by biochemical, spectroscopic, and functional assays. M39R and N55K opsins had variable degrees of chromophore regeneration when compared with WT opsin but showed no gross structural misfolding or altered trafficking. M39R showed a faster rate for transducin activation than WT rhodopsin with a faster metarhodopsinII decay, whereas N55K presented a reduced activation rate and an altered photobleaching pattern. N55K also showed an altered retinal release from the opsin binding pocket upon light exposure, affecting its optimal functional response. Our data suggest that these sector RP mutations cause different protein pheno-types that may be related to their different clinical progression. Overall, these findings illuminate the molecular mechanisms of sector RP associated with rhodopsin mutations.
RP 7 refers to a group of inherited visual diseases that cause retinal degeneration and its incidence is around 1 in 4000 people (1). In the initial phase of the condition, RP patients have impaired dark adaptation and suffer from night blindness. Progression of the disease involves the appearance of so-called tunnel vision, lack of outer lateral vision, and the half-peripheral visual field is reduced to 30 -50°, and it can be accompanied by a decrease of the visual acuity. At later stages, signs include attenuated retinal vessels and intraretinal bone pigment distributed circumferentially around the mid-periphery in the fundus; myopia and astigmatism are also common (2).
Rhodopsin (Rh) is the prototypical member of the G-protein coupled receptor (GPCR) superfamily of membrane proteins, which includes over 900 members (3). The structures of several members of this superfamily have been resolved in the past decade using x-ray crystallography (4,5). Rh has been crystallized in various states, such as the dark-state form, the ligandfree form opsin structure, and the G protein-interacting conformation of opsin (6 -8).
Rh is the light-sensitive protein of the rod cell in the retina and is concentrated in the discs of the rod outer segment (9,10). In its basal resting state, it consists of a polypeptide chain, opsin, and its cognate chromophore, the vitamin A aldehyde derivative 11-cis-retinal, which acts as an inverse agonist preventing G-protein activation in the dark (3,11). The retinal chro-mophore is responsible for light absorption in the visual process (12), and, upon photon capture, it isomerizes to its all-trans configuration yielding the active Rh photointermediate metarhodopsin II (MetaII) capable of binding and activating the G-protein transducin (Gt) (13)(14)(15).
Mutations in the opsin gene are the most common cause of dominant RP. The large number of mutations and their different location along the polypeptide chain indicate a highly sophisticated arrangement of physicochemical interactions in the three-dimensional structure of the native molecule. Although protein misfolding or altered trafficking is considered to be the major biochemical features of many RP mutations in Rh (16 -19), some mutations do not induce major structural defects (20,21) but cause changes in G-protein binding and activation (22,23) or the formation of altered photointermediates (24,25), among other phenotypes (17,(25)(26)(27)(28). Sector RP is an atypical form of RP, ranging from stationary to slowly progressive evolution of the retinal degenerative pattern, characterized by regional areas of bone spicule pigmentation, subnormal electroretinogram, and visual-field defects, usually in the inferior quadrant of the retina (29,30).
It is important to obtain detailed knowledge of the characteristics of mutant proteins and to correlate the molecular defect observed with the clinical phenotype of the affected RP patients. For this purpose, we expressed, purified, and characterized recombinant receptors containing mutations at amino acid Asn-55 and Met-39, initially associated with sector RP, to determine the effect of these mutations on Rh stability and functionality. Residue Asn-55 1.50 is part of the conserved GX 3 N motif (31,32) in the first transmembrane helix (TM) of the receptor and has been proposed to be involved in a net of conserved allosteric interactions important for signal transduction (33). Met-39 1.34 is also located at the first TM, in one of the sites of the proposed retinal channel, where retinal can enter the protein and leave upon Schiff base (SB) breakage after photoactivation (34). We found that these two mutants show similar structural instability in the dark and significantly different photobleaching and G-protein activation responses. In particular, N55K shows a clearly altered response to illumination in contrast to M39R behavior. These differences at the molecular level may be correlated with the observed different clinical progressions for the two mutations.

EXPERIMENTAL PROCEDURES
Patients and Clinical Data-Patients and families were recruited from the inherited retinal clinics at Moorfields Eye Hospital as part of a larger study. Informed consent was obtained from all participants, and all investigations were conducted in accordance with the principles of the Declaration of Helsinki. Institutional Review Board/Ethics Committee approval was obtained from the Moorfields Eye Hospital Local Ethics Committee.
Materials-All chemicals were purchased from either Fisher or Sigma except where stated. 11-cis-Retinal was kindly provided by Dr. R. Crouch and the NEI, National Institutes of Health, Bethesda, MD. Purified Rh-1D4 was obtained from Cell Essentials (Boston, MA) and was coupled to CNBr-activated Sepharose 4B Fast Flow (Amersham Biosciences); secondary HRP-conjugated goat antibody was provided by Santa Cruz Biotechnology (Heidelberg, Germany). n-Dodecyl-D-maltoside (DM) was purchased from Affymetrix (Maumee, OH). The nanomer peptide corresponding to the last nine amino acids of Rh was synthesized by Serveis Cientificotècnics (Universitat de Barcelona, Barcelona, Spain). Site-directed mutagenesis was performed by using Agilent Pfu polymerase and DpnI restriction enzyme and following QuikChange mutagenesis instructions (Agilent, Santa Clara, CA). EndoH (catalog no. P0702L) and PNGaseF (P0704L) were from New England Biolabs (Hitchen, UK).
Cell Culture Materials-COS-1 cells were purchased from the American Type Culture Collection (Manassas, VA), and SK-N-SH human Caucasian neuroblastoma cells were from the European Collection of Cell Cultures (Salisbury, UK). Tissue culture media were provided from PAA (Austria) or Sigma; antibiotics and serum were from Sigma; Opti-MEM reduced serum media were from Invitrogen, and the polymer used for transfection, polyethyleneimine 25 kDa (PEI), was from Polysciences (Warrington, PA). For cellular localization, the primary antibodies Rh-1D4 and Rh-4D2 against rod opsin were gifts from Robert Molday (Dept. of Biochemistry and Molecular Biology, University of British Columbia, Vancouver, Canada). Rabbit polyclonal anti-BiP anti-GRP78/BiP (GL-19) (catalog no. G8918) was from Sigma. Goat anti-mouse Alexa Fluor 488 and goat anti-rabbit Alexa Fluor 594 secondary antibodies conjugated IgGs and Lipofectamine and Plus reagent were purchased from Invitrogen. Goat anti-mouse secondary antibody conjugated with horseradish peroxidase was from Pierce.
Construction, Expression, and Purification of Mutant Rh-Opsin mutants were constructed with a site-directed mutagenesis kit (QuikChange, Stratagene) using a synthetic bovine opsin gene as a template (36). The expression and purification of the visual receptors were performed as described previously (37). Briefly, plasmids encoding the WT or mutant Rh genes were transfected into five COS-1 cells plates at 85% confluence by using PEI with 30 g of plasmid DNA per 145-cm plate. Cells were harvested 48 -60 h after transfection, and pigments were reconstituted with 10 M 11-cis-retinal in intact cells and then solubilized with 1% (w/v) DM in Buffer A. After centrifugation, the supernatant was incubated with a Sepharose 4B coupled to the Rh-1D4 antibody. The resin was washed with buffer C, and pigments were subsequently eluted with the same buffer but containing 100 M nonamer peptide.
Subcellular Localization-SK-N-SH cells were maintained and transfected as described (35). 24 h after transfection, cells were fixed with 4% paraformaldehyde for 15 min and permeabilized in 0.5% Triton X-100 for 10 min. Nonspecific binding was blocked using blocking buffer (3% BSA and 10% serum of the secondary antibodies species in PBS) for 1 h. Rh-1D4 primary antibody (1:5000), anti-BiP (1:100), or Rh-4D2 primary antibody (1:100) in blocking buffer was added for 1 h. Secondary Alexa Fluor 488 or 594 was used at 1:1000 in blocking buffer for 1 h. Images were taken using a Carl Zeiss LSM 710 laserscanning confocal microscope. The images were exported from LSM Browser and prepared using Adobe Photoshop and Illustrator CS4. Cell morphology studies scored the predominant localization of rod opsin on the plasma membrane (PM), or strong overlap with an endoplasmic reticulum (ER) marker, as a percentage of total transfected cells. Four fields of ϳ100 cells were counted for each condition.
Western Blotting and Glycosidase Digestion-SK-N-SHtransfected cells were lysed for 15 min at 4°C in PBS buffer containing 1% DM and 2% protease inhibitor mixture in PBS. Cell lysates were centrifuged for 15 min at 12,000 ϫ g and at 4°C. For deglycosylation reactions, 10 g of total protein in DM-soluble cell lysate was digested with EndoH or PNGase F for 2 h at 37°C. 10 g of total protein was added to 2ϫ Laemmli loading buffer before being subjected to SDS-PAGE and Western blotting. For opsin immunodetection, Rh-1D4 was used at 1:1500 and goat anti-mouse HRP (Pierce) was used at 1:50,000 in 5% (w/v) Marvel Milk in PBS with 0.1% (v/v) Tween buffer (PBS-T).
Absorbance and Fluorescence Spectroscopy Instrumentation-All measurements were made on a Cary 100Bio spectrophotometer (Varian, Australia), equipped with water-jacketed cuvette holders connected to a circulating water bath. Temperature was controlled by a Peltier accessory connected to the spectrophotometer. All spectra were recorded in the 250 -650-nm range with a bandwidth of 2 mm, a response time of 0.1 s, and a scan speed of 300 nm/min. The spectral ratio, which was used as a measure of pigment yield and stability, is defined as absorbance at 280 nm divided by absorbance at the visible max value. All fluorescence assays were performed by using a Photon Technologies QM-1 steady-state fluorescence spectrophotometer (PTI Technologies, Birmingham, NJ). Sample temperature was controlled with a cuvette holder Peltier accessory TLC 50 (Quantum Northwest, Liberty Lake, WA) connected to a hybrid liquid coolant system Reserator XT (Zalman, Garden Grove, CA). All fluorescence spectra were carried out by exciting the samples for 2 s at 295 nm and a bandwidth slit of 0.5 nm and blocking the excitation beam for 28 s with a beam shutter to avoid photobleaching of the sample (38). Tryptophan emission was monitored at 330 nm and a bandwidth slit of 10 nm.
Spectral Characterization of Purified Rh-Samples were bleached with a 150-watt power source equipped with an optic fiber guide and using a 495-nm cutoff filter. Dark-adapted pigments were illuminated for 30 s to ensure complete photocon-version to 380-nm absorbing species. Acidification was carried out, immediately after photobleaching, by the addition of 2 N H 2 SO 4 .
Thermal Bleaching Assay-Rh thermal stability was monitored as described previously (39). Briefly, Rh bleaching rates, in the dark, were obtained by monitoring the decrease of absorbance at max of the visible spectral band as a function of time at 48°C. Spectra were recorded every 2.5 min, and half-life times were determined by fitting the experimental data to single-exponential curves using Sigma Plot version 11.0 (Systat Software, Chicago).
Hydroxylamine Assay-A solution of 1 M hydroxylamine hydrochloride, adjusted to pH 7, was added to dark-adapted samples in Buffer C (final concentration of 50 mM), and successive spectra were recorded every 2.5 min to monitor the loss of pigment (A max ) and formation of retinal oxime (A 365 nm ). Reactions were carried out in the dark at 20°C (37). Initial velocity was determined by fitting the experimental data to either single-exponential curves or a linear plot using Sigma Plot version 11.0 (Systat Software).
MetaII Decay Measurements-The MetaII active conformation decay process was followed in real time by fluorescence spectroscopy (38). Briefly, 0.5 M Rh in Buffer C was kept at 20°C for 10 min and illuminated for 30 s. The t1 ⁄ 2 values of Trp fluorescence increase, due to retinal release, were determined by fitting the experimental data to a single-exponential curve using Sigma Plot version 11.0 (Systat Software, Inc.).
Gt Activation Assay-G-protein activation was determined in COS-1 cell membranes, prepared as described previously (37), and the ability of opsin and Rh to activate Gt was monitored with a radionucleotide filter binding assay by measuring the uptake of [S 35 ]GTP␥S by Gt purified from bovine retinas. The assays were performed by mixing 10 nM Rh or opsin in membranes with 500 nM Gt in Gt buffer at room temperature. Reactions were initiated by the addition of Rh or opsin in the dark, and samples were filtrated after different incubation times, either in the dark or after illumination, to determine the amount of bound [S 35 ]GTP␥S.

RESULTS
Clinical Findings-Of the 73 unrelated families presenting at the inherited retinal disease clinics at Moorfields with retinal degeneration, in which RHO was the established causative gene, three families were found to harbor the variant p.N55K. Clinical data were available on five affected individuals from these three families. In family GC18466, the female proband first noticed nyctalopia in her 30s and was diagnosed with RP at age 35 years. Visual acuities were normal when measured at age 45 years. A de novo mutation was assumed because neither parent had symptoms nor did they harbor the mutation. The fundus and OCT findings for this patient are shown in Fig. 1, A-J. Two of three children were found presymptomatically to have inferior retinal pigmentation suggestive of RP at ages 13 and 14 years. In family GC16563, the 26-year-old female proband was examined by an optometrist following the incidental finding of retinal signs in her mother. Despite inferior retinal pigmentation, no symptoms had developed up until the last review at age 39 years (Fig. 1, E and F). The International Society for Clinical Electrophysiology of Vision standard at this age was normal. In the third family, GC17079, a female patient at the age of 37 years had night-blindness and field loss, the latter enough to fail the driving standard in the United Kingdom. Visual acuities were 6/9 right and 6/5 left. There was a family history compatible with autosomal dominant segregation but no other relatives were examined.
Of the 73 unrelated RHO families, six families were identified with p.M39R heterozygous mutations (GC97, -109, -414, -1379, -4740, and -19,172). All of them consistently showed an inferior hemisphere that was more affected than the superior. The eye fundus autofluorescence images for patients from some of these families are shown in Fig. 2. In contrast to p.N55K, however, the retinal degeneration was more severe and progressed to the superior retina with age (Fig. 2).
Molecular Modeling of the RP Mutants-The repertoire of Rh crystal structures (7, 8, 40 -43) suggests the existence of a channel between TM1 and TM7 that retinal uses for entry/exit of its binding cavity (34). Met-39 lies at the mouth of this channel, where the crystal structures revealed the presence of fatty acids, lipids, and detergent molecules (Fig. 3A), supporting the hypothesis that this region might accommodate hydrophobic retinal in an intermediate step. Fig. 3B shows that Phe-293, adjacent to Met-39, exhibits a conformational change upon receptor activation. Specifically, in the crystal structures representing the inactive states (Fig. 3B, orange schematic), Phe-293 is buried within the binding pocket and closes the channel by interacting with Leu-40, whereas in those representing active structures (Fig. 3B, red schematic), Phe-293 has moved toward the membrane-exposed face of TM1 and interacts with Tyr-43 (Fig. 3B). This conformational change is accompanied by a parallel shift of TM6 and TM7 toward TM1. Consequently, the presence of a positively charged residue, such as Arg-39, nearby this network, would potentially alter the exit/entry of retinal through the ligand channel, and it may also affect the conformational changes involving helical motions ensuing receptor photoactivation.
In addition, our molecular models suggest that the M39R substitution may alter the TM1-TM7 interface and the network of interactions between TM1, -6, and -7, suggesting problems on receptor stability, and the rate of both Gt activation and retinal release, due to the location of M39R in the proposed entry/exit retinal channel (34).

Subcellular Localization of RP Mutants in SK-N-SH Cells-
The heterologous expression of rod opsin in cell culture can be used to monitor protein biogenesis and traffic and to facilitate protein purification for functional studies. We used SK-N-SH neuroblastoma cells to study the biogenesis, degradation, aggregation, and traffic of WT and mutant forms, as this cell line has been used extensively in the past (37). WT rod opsin was mainly localized on the PM (94% of transfected cells) and did not overlap with the ER marker BiP (Fig. 4, A and C). Occasionally WT rod opsin was predominantly retained in the ER (5% of cells) or formed intracellular inclusions (3% of cells) (Fig.  4, A and C). The M39R and N55K amino acid substitutions in rod opsin did not have a major effect on traffic in SK-N-SH cells  DECEMBER 26, 2014 • VOLUME 289 • NUMBER 52 (Fig. 4A). Both mutants were localized to the PM in the majority of cells (88 and 80%, respectively), and only a few cells showed predominant retention in the ER and overlap with BiP staining or inclusion formation (Fig. 4, A and C). The traffic of WT, N55K, and M39R rod opsin to the PM was confirmed by staining nonpermeabilized transfected SK-N-SH cells with the Rh antibody 4D2, which recognizes an extracellular epitope (Fig.  4B). The electrophoretic mobility of the mutant proteins on Western blotting was similar to WT rod opsin (Fig. 4D), but the expression level was generally reduced. Digestion of cell lysates with the glycosidase EndoH showed that, like WT, most of the N55K and M39R rod opsin protein was EndoH-resistant and therefore had trafficked beyond the ER (Fig. 4D).

UV-visible Spectral Characterization of Purified Mutants-
The RP mutant opsins were purified, and their UV-visible spectra were investigated (Fig. 5). N55K and M39R were purified at a lower yield when compared with WT, and they also had a ratio A 280 /A max of 5.4 and 6, ϳ2.2and 2.4-fold higher when compared with the WT ratio, respectively (Table 1). This could indicate folding problems in these mutants that may not allow complete regeneration with 11-cis-retinal. The dark spectra of N55K showed a blue shift of 3 nm in the visible absorbance band when compared with that of the WT. Upon illumination, the N55K spectrum also showed a specific behavior, without a complete conversion of the visible band to the 380-nm absorbing species. This remaining band, observed upon illumination, had a similar visible wavelength maximum as the dark pigment blue-shifted only 3 nm, suggesting conversion to a photointermediate with a retinal binding pocket similar to the dark pigment, including the presence of a protonated SB linkage. This behavior is different from other RP mutants, such as G51V and G89D, that also showed a blue-shifted photointermediate, possibly reflecting changes in the retinal binding pocket (24, 45).
The M39R mutant had photobleaching behavior similar to that of the WT protein (compare insets in Fig. 5).
Thermal and Chemical Stability in the Dark-The chemical reagent hydroxylamine, added in the dark, decreased the visible maximal absorbance, especially in the case of N55K. Here, the initial rate of the hydroxylamine-induced bleaching process was about 50-fold faster than that of WT Rh (Table 1 and Fig.  6A), suggesting that the SB in this mutant is more exposed to the environment so the reagent can enter, break the SB, and sequester the retinal from the binding pocket. Hydroxylamine also affected the chemical stability of the M39R mutant but to a lesser extent than N55K. Thermal bleaching was studied at 48°C by monitoring the decrease of absorbance at the visible max over time. Both RP mutants were unstable in the darkstate (Table 1 and Fig. 6B). To improve protein stability, 1,2didocosahexaenoyl-sn-glycero-3-phosphocholine lipid was used, previously shown to increase Rh stability (46), but it did not help to stabilize N55K, when compared with Rh from rod outer segments (data not shown).
MetaII Stability-The stability of the active state of purified Rh was measured by monitoring the fluorescence increase upon illumination (Fig. 7). MetaII stability t1 ⁄ 2 of the N55K mutant was similar to that of WT Rh, showing that this mutation does not affect initial retinal release upon illumination, whereas for the M39R mutant the release occurred faster (t1 ⁄ 2 ϭ 10.8 min). When a fluorescence plateau was reached, hydroxylamine was added to test whether there was still retinal remaining in the binding pocket. Strikingly, N55K reached a plateau at a lower fluorescence intensity and showed an additional increase in Trp fluorescence emission, upon hydroxylamine addition to the sample, reflecting the presence of residual retinal in the binding pocket for a long time after bleaching (Fig.  7C). To determine whether this could be a differential feature of sector RP mutants, retinal release was compared with that of the G51V RP mutant; position Gly-51 is located one helical turn below Asn-55, and G51V also showed a different photobleaching behavior when compared with WT similar to N55K (45). In this case, however, the G51V MetaII assay showed a similar behavior to WT (data not shown).
Gt Activation for WT and Mutants-Gt activation by Rh and opsin forms was monitored by using a GTP␥S 35 binding assay where the protein was incubated in a mixture containing the G-protein. The amount of GTP␥S 35 bound was determined by a radioactive binding assay (37). The opsin form (i.e. with no retinal) was used to determine whether the mutations caused constitutive activity. We observed that these mutations do not activate Gt in the absence of the ligand, and thus they are not constitutively active (Fig. 8). Photoactivated mutant Rh activated Gt but with altered kinetics. Compared with the WT, M39R activated Gt at a faster rate, whereas N55K activated the G protein at a slower rate, 3.4-and 0.3-fold respectively ( Table  1). The N55K result is compatible with the proposed impor-  A and B, representative images of rod opsin WT, M39R, and N55K localization. SK-N-SH cells were transfected with WT and mutant rod opsins, and 24 h later, cells were fixed. A, Rh-1D4 primary antibody (green) was used to compare localization with BiP (red) following cell permeabilization with Triton X-100. B, Rh antibody to an extracellular epitope 4D2 was used on nonpermeabilized cells to show traffic to the cell surface. Images were taken using a Carl Zeiss LSM 710 laser-scanning confocal microscope. Scale bar, 10 m. C, quantification of opsin localization. Cells were scored for the predominant opsin localization at the PM (open bars) or overlap with an ER marker (gray bars) in ϳ400 cells. Error bars represent Ϯ 2SE. D, Western blot of WT, N55K, and M39R transfected SK-N-SH cell lysates for rod opsin expression detected with Rh-1D4. 10 g of soluble cell lysate protein (S) was compared with 10 g of lysate digested with EndoH (E) or PNGase F (P). The asterisk highlights the position of deglycosylated rod opsin, and two asterisks indicate the position of opsin dimer. ER-retained opsin is sensitive to EndoH, whereas all opsin is sensitive to PNGase F and will tend to dimerize. The position of molecular mass markers is indicated on the left in kDa. DECEMBER 26, 2014 • VOLUME 289 • NUMBER 52

DISCUSSION
Understanding the molecular mechanism(s) underlying retinal degeneration is the first step toward finding an efficient treatment for RP. Here, we analyzed the consequences on the stability and functionality of Rh caused by the two newly identified RP mutations, M39R and N55K, located in the first TM of the receptor. Spectral characterization showed that M39R and N55K behave differently from the WT showing decreased chro-mophoric visible bands than the WT protein after purification in detergent solution. Despite the fact that the pattern observed in our spectra could be compatible with some sort of protein misfolding, as reported previously for other RP mutations, such as T58R (50), the results obtained from the cellular localization clearly indicate that these mutants trafficked to the PM at a similar level to WT opsin, suggesting that these mutations do not significantly alter the folding of Rh in the ER of cell culture systems. Moreover, Western blot analysis of these mutant proteins showed a slightly different mobility and lower level of expression when compared with the WT, which may reflect differences in the degradation rate and/or glycosylation pattern (Fig. 4C). However, digestion with the glycosidase EndoH, which digests ER resident glycoproteins, confirmed that most of the N55K and M39R mutant protein trafficked normally. Therefore, if the lower expression levels are due to enhanced degradation, this must be efficient, as the mutant proteins did not accumulate in the ER. Taking all this evidence into account, we suggest that the decrease in the level of chromophore present in the purified samples is potentially due to protein instability during the purification process, rather than an ER folding problem. Because chromophore stability of these two mutants was 90% lower than that of WT when treated at 48°C (Table 1), but they still reached the PM in over 80% of cells (in contrast with the ER retention levels seen as one of the key features for class II mutants (28)), both mutants are more consistent with a class IV phenotype of near normal native folding but with inherent structural instability (20). The crystal structures suggest that the effects of M39R on Rh could be related to the involvement of the residue in a proposed entry/exit channel for retinal between TM1 and TM7 (34).
In the case of the N55K substitution, the mutation at the conserved Asn 1.50 residue affected retinal release during the MetaII decay. Addition of hydroxylamine to the sample, after the fluorescence signal had reached a plateau, caused further release of retinal, indicating that some retinal was still present inside the protein a long time after photobleaching. This retinal trapping would undoubtedly alter the visual cycle, affecting the normal retinal turnover rate (51). This behavior was not detected for the G51V RP mutant (45) located only one helix turn away from Asn-55, which also shows a different photobleaching behavior when compared with WT Rh. In this case, the similar retinal release pattern to WT suggests that the retinal trapping found in the N55K mutant may be a distinctive feature of this form of sector RP. Our molecular modeling results suggest that the Lys residue interferes in the network between TM1, -2, and -7, due to the proximity of Asn-55, Asp-83, and Asn-302, including water molecules (44). Additionally, Fourier-transform infrared difference spectra experiments indicated that Asp-83 2.50 should be protonated (52), which would be affected by the introduction of Lys in this environment.
The molecular differences observed for the N55K and M39R mutated proteins can be associated with the physicochemical nature of the introduced amino acid residue as a result of the mutation. Both mutations are located at the same helix (TM1) of Rh, and both introduce positively charged amino acids of a similar volume. However, the N55K substitution appears to have a specific behavior. Interestingly, another mutation to Lys in the TM domain of Rh, M207K, was shown to have an unusual autosomal dominant retinal phenotype characterized by a heterogeneous distribution (53). It is possible that mutations introducing a Lys, at the transmembrane core of Rh, cause unusual phenotypes as a result of affecting the entry/exit paths of the retinal chromophore after isomerization and SB hydrolysis.
The intrinsic structural features of the receptor-bound chromophore, in some Rh RP mutants, could result in differences in keeping the retinal in its binding pocket and altering its release mechanism. This could be the starting cause of the retinal degeneration process. Thus, the presence of unstable free opsin proteins, either in the dark or at different stages of the photoactivation process, could result in protein aggregation and ultimately in cell death. The presence of fully functional Rh is important for maintaining healthy photoreceptor cells. Rh knock-out mice have been reported to be unable to generate photoreceptor rod outer segments (54). In the case of N55K, a critical factor, in addition to its inherent instability, is the altered behavior upon illumination (and associated altered retinal release) that could affect the Rh photocycle disrupting normal retinal accessibility and protein turnover.
Besides the structural consequences of these mutations, the big question is why the affected region is restricted to a specific part of the retina when the Rh mutation is present all over the retina. There is a clear variability in the clinical phenotypes of the novel sector RP mutations studied. M39R, for instance, initially shows a sector phenotype, but over time the degenerative process spreads to the whole retina, and the patient develops a classical RP phenotype. However, P23H, the most common autosomal dominant RP mutation, has been described to be heterogeneous, and some people showed a sector phenotype, and other people showed the typical RP form, suggesting the influence of external factors (55). To date, most of the studies have focused on the idea that light is an important external factor underlying sector RP (56,57), and a different light distribution has been proposed in the different sections of the retina, based on a theoretical model (58). Previous reports on mice carrying a transgene with the P23H mutation involved housing the animals under overhead light or in the dark, and higher photoreceptor degeneration was observed in the inferior retina a The ratio of A 280 /A max provides a measurement of sample purity and ability of opsin to form pigment with the chromophore. b Thermal bleaching t1 ⁄ 2 values of purified samples in Buffer C at 48°C were derived from the time course decay of the visible absorbance band at max . c Hydroxylamine reactivity of the dark state pigments was at 20°C. Values were determined by monitoring the rate of max absorbance decrease after the addition of 50 mM hydroxylamine, pH 7.0, to the samples in Buffer C and adjusted to a single exponential decay function for the N55K RP mutant or a linear function for WT and M39R. For a better comparison of initial velocity values, relative data are shown in parentheses. d MetaII stability t1 ⁄ 2 values were determined by fitting the experimental data to a single-exponential curve of Trp fluorescence increase, due to the retinal release at 20°C (38). e Gt activation initial rate was calculated as previously described (45). In all cases, mean values and standard error were obtained from three independent experiments. FIGURE 6. Chemical and thermal stability for WT and RP mutants. WT, M39R, and N55K in buffer C were treated with 50 mM hydroxylamine, pH 7.0, at 20°C for the chemical stability assay. Thermal stability assay was carried out at 48°C. A, chemical reagent stability was determined by monitoring the decrease of max over time (at 498 nm for WT and M39R and at 495 nm for N55K). B, thermal stability assay plot showing the decrease of absorbance at the max in the visible region over time. FIGURE 7. MetaII decay for WT and RP mutants. WT, M39R, and N55K in Buffer C were illuminated, and the Trp fluorescence increase at 330 nm was monitored to follow ligand release (38). After fluorescence stabilization, 50 mM hydroxylamine, pH 7.0, was added to confirm complete retinal release. Fluorescence intensity spectra of MetaII decay for WT, M39R and N55K are shown. DECEMBER 26, 2014 • VOLUME 289 • NUMBER 52 of mice reared under an overhead light (59). Moreover, a recent study in Xenopus laevis has provided further evidence of the effects of light on sector RP-derived degeneration (60). Despite these studies, no definitive answer has been provided about the link between light and sector RP, and it is not clear whether the direct action of light on Rh, or other unknown light-sensitive factors located in the retina, may cause this particular phenotype. The detailed molecular basis of light-dependent translocation of proteins from the inner segment to the outer segment in photoreceptor cells, such as arrestin (61), or from the outer segment to the inner segment, like Gt (62), is still largely undeciphered, but appears to be a finely regulated machine, and a small change in one of the partners could significantly affect the function of the others. Here, we show that N55K presents a different behavior upon illumination by forming a photointermediate unable to completely finish the retinal release process from its binding pocket. We suggest that this supports the direct action of light on Rh as the trigger for the observed sector phenotype. Thus, different times and/or dose exposures to light, in combination with malfunction of the receptor due to the mutation, may be the cause of sector RP. The mechanism could also indirectly involve other light-sensitive proteins or factors so far unidentified, but confirming this aspect would require further investigation.

Molecular Mechanism of Sector Retinitis Pigmentosa
In view of the clinical phenotypes observed for the two mutations reported herein, we can ask the following question. Is sector RP really that uncommon? Our results emphasize that sector RP etiology is not crystal clear. It could be that in many instances sector RP is an initial step, before the development of a typical RP phenotype. Some mutations might cause unnoticeable effects early in disease, and RP would be diagnosed as a typical form later on after severe visual impairment arises (52). For example, in one M39R family, the fundus of a son shows a sector phenotype ( Fig. 2A), whereas his father (Fig. 2E) has a more typical RP form. Time will tell if the son will develop the typical form like his father or will remain as sector RP. According to our findings, we believe that the son will develop typical RP when older. In contrast, N55K appears relatively stable and restricted to the inferior retina in the families we have studied, and the protein has unique properties upon illumination, suggesting this is genuine sector RP.