The R838S Mutation in Retinal Guanylyl Cyclase 1 (RetGC1) Alters Calcium Sensitivity of cGMP Synthesis in the Retina and Causes Blindness in Transgenic Mice*

Substitutions of Arg838 in the dimerization domain of a human retinal membrane guanylyl cyclase 1 (RetGC1) linked to autosomal dominant cone-rod degeneration type 6 (CORD6) change RetGC1 regulation in vitro by Ca2+. In addition, we find that R838S substitution makes RetGC1 less sensitive to inhibition by retinal degeneration-3 protein (RD3). We selectively expressed human R838S RetGC1 in mouse rods and documented the decline in rod vision and rod survival. To verify that changes in rods were specifically caused by the CORD6 mutation, we used for comparison cones, which in the same mice did not express R838S RetGC1 from the transgenic construct. The R838S RetGC1 expression in rod outer segments reduced inhibition of cGMP production in the transgenic mouse retinas at the free calcium concentrations typical for dark-adapted rods. The transgenic mice demonstrated early-onset and rapidly progressed with age decline in visual responses from the targeted rods, in contrast to the longer lasting preservation of function in the non-targeted cones. The decline in rod function in the retina resulted from a progressive degeneration of rods between 1 and 6 months of age, with the severity and pace of the degeneration consistent with the extent to which the Ca2+ sensitivity of the retinal cGMP production was affected. Our study presents a new experimental model for exploring cellular mechanisms of the CORD6-related photoreceptor death. This mouse model provides the first direct biochemical and physiological in vivo evidence for the Arg838 substitutions in RetGC1 being the culprit behind the pathogenesis of the CORD6 congenital blindness.

Retinal membrane guanylyl cyclase (RetGC) 2 is one of the key enzymes in vertebrate visual signaling. Rods and cones respond to light by closing cGMP-gated plasma membrane channels in the outer segment as cGMP becomes hydrolyzed by phosphodiesterase PDE6 and then re-open the channels, after PDE6 activity becomes quenched and cGMP becomes re-synthesized by RetGC (1)(2)(3)(4)(5). Two RetGC isozymes expressed in photoreceptors, RetGC1 and RetGC2 (6 -8), bind calcium-sensor proteins, GCAPs (7)(8)(9)(10)(11)(12), which decelerate the cyclase activity at high intracellular calcium concentrations in the dark and accelerate it in the light, when the influx of calcium through the cGMP-gated channels is shut off (13)(14)(15)(16)(17)(18)(19). RetGC also binds retinal degeneration 3 (RD3) protein (20 -21), which prevents cyclase activation by GCAPs (22)(23) and promotes accumulation of RetGC1 in the outer segments (21,24,25). RetGC1 (human gene GUCY2D) accounts for most of the cGMP synthesis in mammalian photoreceptors (26,27), therefore mutations in RetGC1 that disable the cyclase activity (28 -33) cause recessive blindness at birth, Leber congenital amaurosis type 1 (LCA1) (33), mostly a non-degenerative or partially degenerative (31,32) loss-of-function hereditary retinal disease. Different from the LCA1 type of severe congenital blindness linked to the RetGC1 gene, cone-rod dystrophy type 6 (CORD6), is a progressive early-onset loss of photoreceptors (34 -35). The most frequent mutations in GUCY2D found in the CORD6 patients of various ethnic groups are substitutions of the Arg 838 (34 -45). Studies of the recombinant RetGC1 heterologously expressed in cultured non-photoreceptor cells (44 -46) indicate that the CORD6-related substitutions in Arg 838 do not inactivate the cyclase, but rather alter its regulation by calciumsensor proteins, in particular guanylyl cyclase activating protein 1 (GCAP1). However, no experimental evidence has been available to date to demonstrate that the CORD6-linked mutations actually alter cGMP synthesis regulation in living photoreceptors of the retina and cause loss of visual function. In the present study, we describe a new transgenic model that expresses a CORD6-linked variant R838S RetGC1 in mouse photoreceptors and demonstrate that this mutation when expressed in vivo in the living retina alters the biochemistry of the cyclase regulation and disables the physiological function of the retina by causing photoreceptor death.

R838S Desensitizes Negative Regulation of RetGC1 by GCAP1
and RD3-The RetGC1 encoded by the cDNA fragment intended for the transgenic expression in the retina was first tested for Ca 2ϩ sensitivity, apparent affinity for GCAP1, and sensitivity to inhibition by RD3, by being expressed in HEK293 cells under control of a CMV promoter (46) (Fig. 1). Consistently with the original in vitro observations by Ramamurthy et al. (45), and later by Peshenko et al. (46), the half-maximal deceleration of the GCAP1-activated cyclase by calcium was shifted toward nearly 2.5-fold higher concentrations (Fig. 1A). The apparent affinity of the R838S RetGC1 for the Mg 2ϩ -liganded activator form of GCAP1 (19) became nearly 5-fold higher than normal (K MgGCAP1 327 Ϯ 10 nM (S.D.) versus 1585 Ϯ 80 nM in wild type) (Fig. 1B). In addition to its being the main reason for less effective deceleration by Ca 2ϩ (44 -46), observed in Fig.  1A, the higher affinity for GCAP1 proportionally correlated with a reduced sensitivity of the GCAP1-activated RetGC1 to the inhibition by RD3: the K1 ⁄ 2 RD3 rose from 2.6 Ϯ 1.2 versus 11.4 Ϯ 0.4 nM, respectively (p Ͻ 0.01), in the R838S RetGC1 (Fig. 1C).
R838S RetGC1 Expressed in Mouse Rods Alters Ca 2ϩ Sensitivity of the cGMP Synthesis in the Retina-To express the human CORD6 cyclase mutation in mouse rods, cDNA coding for the R838S RetGC1 was transferred from the expression vector used in HEK293 cells ( Fig. 1) and inserted under the control of rod opsin promoter (47) in the transgene construct ( Fig. 2A) injected in mouse eggs (see "Experimental Procedures" for details). Although CORD6 affects both rods and cones, we chose the promoter that is efficient only in rods to keep the non-expressing cones as a negative control group of photoreceptors in the same transgenic retinas and thus verify that any physiological changes we may observe specifically originate from the transgene-targeted photoreceptors. Founders screened for the presence of the transgenic construct ( Fig. 2A) and passing the transgene to the progeny were then bred to wild type C67B6J mice for at least 4 consecutive generations to establish the R838S RetGC1 expressing lines (R838S ϩ ) 379 and 362 for subsequent detailed analysis. As will be demonstrated further in this article, the line 379 developed more gradually deterioration (arbitrarily dubbed "slower" line) and the line 362, a more severe phenotype ("faster" line). The properties of those two mouse lines will be described in that order in the following sections.
At the age of 3 weeks, RetGC activity tested in the retinas of the R838S ϩ mice (slower line 379) was markedly shifted outside the physiological levels of the intracellular Ca 2ϩ when compared with the retinas from their wild type non-transgenic littermates (K1 ⁄ 2Ca 110 versus 59 nM, respectively) (Fig. 2B). However, only a very slight, if any, increase in the total RetGC1 content was detectable by Western immunoblotting in the young transgenic mice, indicating that the expression of mutant cyclase hardly exceeded 30% of the total RetGC1 levels (Fig. 2C). Proteins related to RetGC1 regulation, GCAP1, GCAP2, and RD3, were also present in the R838S ϩ retinas. Immunofluorescence of RetGC1 in the slower line 379 retinas, FIGURE 1. Altered biochemical characteristics of a recombinant CORD6 R838S RetGC1. RetGC1 was expressed in HEK293 cells and tested in vitro prior to making it into the transgene-expression construct. A, the calcium sensitivity of wild type (closed symbols) and R838S (R838S ϩ , open symbols) RetGC1. The cyclase activity was assayed at varied concentrations of free Ca 2ϩ set by Ca 2ϩ /Mg 2ϩ /EGTA buffers at a constant 0.9 mM free Mg 2ϩ . The data (mean Ϯ S.D.) from two independent sets of experiments were normalized as the percentage of maximal activity in each cyclase preparation and fitted using the Hill function, already containing differentiated rods and cones but not yet affected at that age by the subsequent massive retinal degeneration (described in the following sections), also resembled that of the wild type (Fig. 2D). The RetGC1 immunofluorescence signal was present both in rods and cones of the R838S ϩ retinas and the wild type, and may have appeared slightly brighter in some R838S ϩ rods (Fig. 2D), the difference from the wild type was by far not obvious. Although the overall expression pattern argued that there would be no obvious abnormalities in the localization of the transgenic CORD6 variant, the presence of the endogenous mouse RetGC1 proved to be a major obstacle in attempts to estimate how much of the human transgene product was expressed and where it was located.
To better estimate the level of human CORD6 RetGC1 expression and localization, we transferred the transgene expression into a genetic background lacking the endogenous mouse RetGC1, by consecutively crossing the R838S ϩ mice from the slower line 379 with the Gucy2e (mouse ortholog of RetGC1, Ref. 48) gene knock-out mice (49) (Fig. 3A). Semiquantitative estimates by luminescence of the R838S RetGC1 band in Gucy2e Ϫ/Ϫ R838S ϩ retinas (Fig. 3A, right panel) indicated that under half the normal levels of R838S RetGC1 were sufficient to shift Ca 2ϩ sensitivity of cGMP synthesis in the presence of the endogenous mouse RetGC1 (Fig. 2B). No less important, having interference from the endogenous RetGC1 eliminated (Fig. 3B) also made it possible to more clearly establish the R838S RetGC1 localization in Gucy2e Ϫ/Ϫ R838S ϩ mice by immunofluorescence staining (Fig. 3C). In contrast to the endogenous RetGC1, the transgenically expressed R838S RetGC1 was present in rods but not in cones (Fig. 3C), consistently with the expected activity of the rod opsin promoter used for its expression (47). Except for being absent in cones, the R838S RetGC1 signal in rods was similar to the wild type albeit not uniform between different rods. Most importantly, the R838S RetGC1 was found in the rod outer segments, with no indication of being retained in the inner segment (  The 3.6-kbp R838S RetGC1 cDNA was inserted between the 4.2-kbp mouse rod opsin gene fragment harboring the promoter region and the 0.3-kb BGH polyadenylation signal-containing fragment as described under "Experimental Procedures." The presence of the transgenic construct in the tail DNA samples from F 0 mice was tested by PCR amplification of the transgene-specific 0.33-kb product, between the positions of the For1 and Rev1 primers as indicated in the upper panel and described under "Experimental Procedures." Lines developing progressive blindness (slower in line 379 and faster in line 362) were derived from the two positive founders and analyzed as described further under "Results." B, Ca 2ϩ sensitivity of the cGMP synthesis rate in the dark-adapted retinas extracted from 3-week-old transgene-positive (R838S ϩ , red symbols) and transgene-negative (wild type, black symbols) littermate retinas of the slower R838S RetGC1 line 379, assayed at the indicated free Ca 2ϩ concentrations as described under "Experimental Procedures." The data (mean Ϯ S.D.) were normalized as the percent of maximal catalytic activity of the cyclase in each preparation and fitted with the Hill function, where A max and A min are the respective maximal and minimal activity, [Ca], free Ca 2ϩ concentration; K1 ⁄2Ca , free Ca 2ϩ concentration causing 2-fold inhibition of the cyclase activity; and H, the Hill coefficient. The average K1 ⁄2Ca in five independently tested transgenic mice (R838S ϩ ) increased to 110 Ϯ 26 from 59 Ϯ 14 nM in four wild type siblings (p Ͻ 0.03). The shaded area approximates the physiological range of the free Ca 2ϩ concentrations in photoreceptors between light-adapted and darkadapted states (73). C, Western immunoblotting of the retinas from 3-week-old wild type (WT) and transgene-positive (R838S ϩ ) siblings probed for RetGC1, GCAP1, GCAP2, RD3, Gt␣, and ␤-actin. D, transgenic expression of the CORD6 RetGC1 mutant does not change the overall pattern of RetGC1 immunofluorescence in the R838S ϩ retina. Paraformaldehyde-fixed retina sections from 3-week-old littermates, either wild type (upper panels) or transgene-positive (bottom panels), were probed with anti-RetGC1 antibody (red) and peanut agglutinin (PA, green); the nuclei were counterstained with TOPRO3 iodide (pseudo-blue). The arrowheads indicate anti-RetGC1 fluorescence in cone outer segments (COS), below the rod outer segment (ROS) layer, identified by the surrounding PA-stained cone sheaths superimposed on the anti-RetGC1 fluorescence (right panels). The samples were processed and stained in parallel and the images were recorded in the same experiment, using identical settings for the anti-RetGC1 fluorescence imaging. responses as early as 1 month of age (Fig. 4, A and B). The average maximal scotopic a-wave (negative voltage deflection by hyperpolarization of the rod photoreceptors in response to a bright flash) was reduced in the R838S ϩ mice to half the normal amplitude (305 Ϯ 81 versus 573 Ϯ 37 V in wild type, p Ͻ 0.0001), but the flash strength producing a half-maximal response was not changed (3.5 ϫ 10 3 Ϯ 1.5 ϫ 10 3 versus 3.8 ϫ 10 3 Ϯ 1.3 ϫ 10 3 R*/rod, p ϭ 0.63) (Fig. 4B). In a similar fashion, rod-driven scotopic b-wave (positive deflection by depolarization of bipolar cells in the inner retina) was reduced in amplitude by one-third (710 Ϯ 179 versus 1055 Ϯ 115 V, p Ͻ 0.0001), without a change in half-maximum flash strength (266 versus 261 R*/rod) (Fig. 4, A and B). This suggested that rodbased vision in the R838S mice was affected by reduction of the rod numbers, rather than loss of light sensitivity by individual rods.
The rod vision in line 379 further declined between 1 and 3 months of age (Fig. 5). For the combined age group of 3 to 6 months, the maximal scotopic a-wave amplitude was already suppressed 4-fold (121 Ϯ 29 V, n ϭ 38) compared with the wild type littermates (479 Ϯ 94 V, n ϭ 23; p Ͻ 0.0001), and the b-wave more than 2-fold (453 Ϯ 109 versus 931 Ϯ 206 V, respectively, p Ͻ 0.0001). Reduction of the response amplitudes in the whole R838S ϩ group combining all tested ages ( Fig. 5B) was also significant (p Ͻ 0.0001) for both aand b-waves.
In stark contrast to the rod vision, response from both midwave (505 nm) and shortwave (365 nm) sensitive cones in light-adapted mice (photopic ERG) remained strong at the age when their rod a-wave was already reduced to rudimentary (Fig. 5, A and C). The averaged S-cone ( Fig. 5C) and M-cone (data not shown) a-wave amplitudes in the R838S ϩ mice between 4 and 6 months of age showed neither substantial nor statistically significant reduction compared with the non-transgenic littermates. Likewise, the cone b-wave amplitude showed no evidence for reduction compared with the wild type (Fig. 5C).
The R838S RetGC1 Causes Progressive Retinal Degeneration in Transgenic Mice-The decline in rod photoreceptor function in R838S ϩ mice coincided with the progressive degeneration of the photoreceptor cell layer (Fig. 6). Erosion of the rod outer segments (ROS) and reduction of the total count of the photoreceptor nuclei (164 Ϯ 27 versus 233 Ϯ 6 per 100-m retina section length in wild type; p ϭ 0.043) in the outer nuclear layer (ONL) of the retina was evident in mice aged 6 weeks and rapidly progressed with age, so that by the age of 3 to 4 months less than one-half and by 6 months, only one-third of the photoreceptor nuclei remained in the retina (76 Ϯ 17 versus 247 Ϯ 22 per 100 m in wild type littermates; p Ͻ 0.0001) (Fig. 6B). Because only 3% of all photoreceptors in a mouse retina are cones (50), the major reduction of the ONL nuclei count mostly reflected a rapid loss of the transgenic rods.
Stronger Degenerative Phenotype Accompanies a Larger Shift in Ca 2ϩ Sensitivity of the Retinal Guanylyl Cyclase-In the transgenic retinas from the faster line 362, the right-shift in Ca 2ϩ sensitivity of the cGMP synthesis measurable at 3 weeks of age was larger than in the slower line 379 (to K1 ⁄ 2Ca 148 nM) (Fig. 7A). Although the expression level in that line was difficult to quantify by immunoblotting due to an early onset of severe retinal degeneration in both Gucy2e ϩ/ϩ and Gucy2e Ϫ/Ϫ genetic backgrounds (data not shown), based on the change in Ca 2ϩ sensitivity the level of R838S RetGC1 expression in the line 362 was at least twice the amount expressed in the slower line 379.
The R838S ϩ mice from the faster line 362 demonstrated a more severe loss of rod function: the amplitude of scotopic ERG a-wave became negligible and the b-wave rudimentary in comparison with the wild type at 3.5 months of age (Fig. 7B, left  panel). The residual activity evoked by the bright flash in darkadapted animals most likely originated from cones, because both M and S cones remained functional and produced pho- topic ERG responses (Fig. 7B, right panel). Retinal degeneration was also more pronounced (Fig. 7C): less than 20% of photoreceptor nuclei remained in the outer nuclear layer after 3 months of age and less than 6% could be identified in the single remaining incomplete row of the nuclei after 5 months (Fig.  7D).

The Arg 838 Mutation in RetGC1 Alters Physiology of Photoreceptors and Triggers Degeneration in a Living Retina
CORD6 is typically diagnosed as an early-onset blinding disorder, frequently by 7 years of age, first as impaired central and color vision (cone function), later involving rods, by progressing to the peripheral visual field with age (34,35). Among the  genes linked to CORD6 in multiple genetic studies (34,45,51,52), GUCY2D coding for a human RetGC1 has been found most frequently affected by substitutions in codon 838 replacing Arg 838 with Cys, Gly, His, Pro, or Ser. Three of these substitutions, Cys 838 , His 838 and Ser 838 , were shown to cause a prominent shift in Ca 2ϩ sensitivity of the cyclase regulation by GCAP1 when heterologously expressed in cultured HEK293 (44 -46). However, the physiological effects of the Arg 838 substitutions in RetGC1 have not been demonstrated in mammalian photoreceptors. One attempt to heterologously express the CORD6 RetGC1 variant in vivo using zebrafish larvae documented changes in cone morphology detectable in the central retina, but did not impair visual responses in the fish (53).
Our present study provides the first direct biochemical and physiological evidence that a CORD6-related mutation alters photoreceptor physiology in a manner consistent with the in vitro studies of heterologously expressed Arg 838 mutants of RetGC1 (44 -46). We demonstrate that R838S RetGC1 shifts Ca 2ϩ sensitivity of the cyclase regulation in living rods of the mouse retina and also, by triggering degeneration of photoreceptors, causes early-onset progressive blindness, even at the levels of expression lower than the endogenous RetGC1 (Fig. 3). The photoreceptor death caused by the R838S RetGC1 (Figs. 5-7) in our studies is a striking contrast to the restoration of vision by transgenic overexpression of wild type RetGC1, rescuing rods and cones lacking functional guanylyl cyclase (54 -57).
Notably, the effect of R838S RetGC1 is specific for the targeted photoreceptors, because the non-targeted cones remain functional even at times when severe degeneration of the transgenic rods has already occurred and most of the rod visual function has already become impaired (Figs. 5 and 7). Although half of the R838S ϩ rods in the slower line 379 died by the age of 3 months, the degeneration did not progress to eliminate all remaining rods within the following three months. It is likely that transgene expression under the rhodopsin promoter in that line was not uniform among all rods (47) and those of them who expressed the CORD6 cyclase died out more quickly, whereas those with lower levels of expression sustained lesser damage and lingered for much longer. In the faster line 362, where the average transgene expression was higher, more strongly affecting Ca 2ϩ sensitivity of cGMP production, all but a few photoreceptors degenerated within 6 months (Fig. 7). The massive death of rods in the line 362 eventually provoked the secondary loss of cones as well, because after 5 months virtually no ERG responses were detected (data not shown).

Relevance and Limitations for the Mouse Model in Studying GUCY2D CORD6 Mechanisms
CORD6 is a dominant congenital disease diagnosed at a young age as a progressive loss of cone and then rod function (34 -38). Consistently with the dominant character of the disease, heterozygous CORD6 RetGC1 transgenes in our model alter physiology and trigger degeneration of rods, whereas their non-transgenic siblings retain visual function and normal retinal morphology at all tested ages. The loss of the rod function in transgenic mice becomes clearly detectable after 1 month (Fig.  4) and progresses with age, which is also generally consistent with the clinical manifestation of CORD6 in humans. At the same time, the design of our model does not mimic all clinical symptoms of the human CORD6. Unlike in human patients, cones do not express the mutated cyclase in our model, by its design, and therefore remain functional for much longer than rods. However, we reason that this mouse model can be adequately used for studying photoreceptor death related to CORD6. Rods account for nearly 97% of all photoreceptors in a mouse retina (50), therefore, the vast majority of photoreceptors undergoing physiological changes and apoptotic death caused by the CORD6 RetGC1 constitute a fair model in further studies of cellular mechanisms of CORD6 blindness and testing potential therapy. Having easily been distinguishable by their electrophysiology cones as a small non-diseased photoreceptor population should be beneficial, because it provides an additional means to assess nonspecific side effects of the potential therapy applications.
GUCY2D CORD6 presents an even bigger challenge for gene therapy than recessive GUCY2D LCA1. In animal models for the LCA1, vision can be sustainably restored by expression of a normal RetGC1 allele in dysfunctional rods and cones, when delivered via adeno-associated virus (54 -57). Consequently, such gene therapy is expected to be perspective in the treatment of a human GUCY2D LCA1 (31). In contrast, to suppress expression of a dominant disease-causing GUCY2D allele in CORD6, more complex approaches would likely be required, such as CRISPR-mediated destruction of the CORD6 allele or suppression of the CORD6-specific transcripts by short interference RNA. Development of these approaches in treating CORD6 would require rigorous evaluation of their efficiency and specificity. For these types of studies, the availability of a GUCY2D CORD6 mouse model would be highly beneficial, provided that such a model is comprehensively characterized in-depth and the mechanisms of photoreceptor degeneration are fully understood.

What Processes Downstream of the Mutated CORD6 RetGC1 Cause Photoreceptor Degeneration?
Although our study decisively demonstrates that the mutation of Arg 838 in RetGC1 causes blindness in vivo, the pathway(s) leading to photoreceptor death remains to be delineated. At present, several non-mutually exclusive scenarios that could, separately or by acting in concert, cause the photoreceptor death need to be considered as follows.
Deregulation of RetGC Activity in the Dark-Negative Ca 2ϩ feedback regulates RetGC1 via Ca 2ϩ -sensor proteins (1,5,10), predominantly by GCAP1 (58,59). The biochemical studies of CORD6 mutations in ReGC1 pioneered by J. Hurley's group (44,45) demonstrated that Arg 838 substitutions alter Ca 2ϩ sensitivity of RetGC regulation by affecting coiled-coil interactions in the cyclase dimerization domain. Although RetGC1 is not a calcium-sensitive protein, the dimerization domain is part of the GCAP1 and GCAP2 binding interface (59 -61), therefore, change in its structure unbalances RetGC1 affinities for two competing with each forms of GCAP1, Mg 2ϩ -liganded (activator) and Ca 2ϩ -liganded (inhibitor) (46). Because the mutant cyclase binds the activator form of GCAP1 with higher affinity (44 -46), it now takes higher Ca 2ϩ concentrations, converting a larger fraction of GCAP1 into the inhibitor state, to outcompete the activator form of GCAP1 and decelerate the cyclase (46). Deregulation of RetGC1 is the main reason for photoreceptor death in mouse models expressing Y99C or E155G GCAP1 mutants (58,62,63). When RetGC1 deceleration by Ca 2ϩ in the dark becomes less sensitive, a larger fraction of cGMP-gated channels in the plasma membrane maintains Ca 2ϩ influx in the dark, eventually provoking the onset of apoptosis (62,63). Stimulation of cGMP-dependent protein kinase activity can also lead to mouse photoreceptor death independently of the Ca 2ϩ pathway (64,65). In our present study, faster rod degeneration accompanied a larger shift in Ca 2ϩ sensitivity of the cGMP synthesis in the retina (Fig. 7), and this supports RetGC1 deregulation as the most probable cause of GUCY2D CORD6. However, the extent to which the cyclase reduces its sensitivity in the R838S ϩ retinas (Figs. 2B and 7A) appears less drastic as in the mouse models expressing Y99C or E155G GCAP1 (58,63). Therefore, it would be premature at this point to conclude that the change in Ca 2ϩ sensitivity of the CORD6 cyclase alone is fully sufficient to cause photoreceptor death in our experimental model.
Reduced Affinity for RD3-One should also consider for future studies a new intriguing observation, that the R838S mutation results in a much weaker affinity of the GCAP1-activated RetGC1 for RD3 (Fig. 1C). RD3 evidently plays a dual functional role in photoreceptors (66): it enhances the efficiency of RetGC expression and accumulation in rod and cone outer segments and blocks RetGC1 activity, most probably up until the cyclase reaches the outer segment. The lack of RD3 causes retinal degeneration in the affected humans and in the rd3/rd3 mouse strain (20). Evidently, both proposed functions of RD3 contribute to photoreceptor survival, because photoreceptors in mice lacking RD3 degenerate even faster than in mice completely devoid of both RetGC1 and RetGC2 (23). We find that R838S RetGC1 reaches the outer segment in a seemingly normal fashion (Fig. 3), however, the reduced affinity of the activated RetGC1 for RD3 (Fig. 1C) would undoubtedly affect the proposed function of RD3 to transiently block premature cyclase activation in the inner segment (22,66).
Unfolded Protein Response-Mutations in some proteins have shown propensity to provoke photoreceptor death by causing endoplasmic reticulum (ER) stress triggering apoptotic unfolded protein response (65,67). We find that the CORD6 mutant of RetGC1 is functional in vivo (Figs. 2 and 6) and properly accumulates in the rod outer segments (Fig. 3, C and D) rather than the inner segments where the ER is located. Therefore, the unfolded protein response scenario seems less likely for the GUCY2D CORD6. Yet it cannot be completely ruled out, because RetGC1 is a membrane protein and therefore could produce ER stress if some of it is improperly incorporates in the ER membranes.
Further studies will determine which of these scenarios take place in the diseased photoreceptors. Detailed characterization of the model will need to address the potential impact of the R838S RetGC1 on single-cell electrophysiology and/or other cGMP-based signaling pathways potentially altered by the CORD6-linked mutation.

Experimental Procedures
Animals-All animal work was conducted in accordance with the Public Health Service guidelines and approved by the Institutional Animal Care and Use Committee. The wild type C57B6 strain was purchased from JAX Research/Jacksons Laboratories. The mouse Gucy2e knock-out line (49), from Dr. David Garbers laboratory (University of Texas), was maintained heterozygous, by crossing to the C57B6 for 10 generations, before breeding into a homozygous Gucy2e Ϫ/Ϫ state. Mice of a mixed SJL/C57B6 background were used to produce founders for the R838S RetGC1 transgenic mice. Both male and female mice were used in experiments.
Transgenic Mice-The expression construct was assembled in a Stratagene pBlueScript plasmid as follows. Bovine growth hormone (BGH) gene 0.3-kb fragment containing the polyadenylation signal was PCR-amplified from Invitrogen pRC-CMV-based plasmid harboring wild type RetGC1 cDNA from previously published studies (46) using a forward 5Ј-GTGGG-TGAAATAAAGGCATACTGTCTTCC and a reverse 5Ј-AAA-AAGCGGCCGCCAGCTGGTTCTTTCCGCCTCAGA primer and inserted into the XbaI/NotI sites of pBlueScript, yielding the intermediate construct "pBlue-BGH." A 4.2-kb KpnI/XhoI rod opsin gene promoter-containing gene fragment (47) was transferred from the pRho-mp3 construct (14,47) into the KpnI/XhoI sites of the pBlue-BGH construct, thus yielding intermediate construct "pRho-BGH" containing both rod opsin promoter and the BGH polyadenylation signal fragment. A HindIII/XbaI RetGC1 cDNA fragment including the ribosomerecognition motif and the translation initiation codon, were subcloned from the pRCCMV construct (46,68) into the Hin-dIII/XbaI restriction sites of pBlueScript, excised using the ClaI/XbaI digest, and inserted into the ClaI/XbaI sites of the pRho-BGH vector. To generate the final construct, pRho-R838S-BGH, the BssHII/XbaI fragment of the RetGC1 cDNA was replaced by that containing the 2512 C 3 A transversion coding for the R838S substitution, excised from the construct (46) previously used for the R838S RetGC1 expression in HEK293 cells (Fig. 1). The pRho-R838S-BGH construct ( Fig.  2A) was verified by restriction nuclease digest and then by DNA sequencing on both strands. A 8.3-kb PvuI/NotI fragment excised from the pRho-R838S-BGH construct and purified from the agarose gel was injected in fertilized mouse egg pronuclei (service of the University of Michigan Transgenic Animal Model Core). The tail DNA samples from the F 0 founders and later of their progeny from a series of consecutive breeding to C57B6J mice were tested for the presence of the transgene by PCR using a forward primer (For1), 5Ј-CGATCCAAACAAC-ATCTGCGGT, and a reverse primer (Rev1), 5Ј-CAGCCAAA-CCCTGTCTCCCTCAT, amplifying a 0.33-kb transgene-specific region ( Fig. 2A). Three lines were initially established from the founders designated by numbers 362, 377, and 379, but the phenotype of the line 377 in the initial testing (not shown) was similar to that of line 379, and therefore only lines 379 and 362 were selected for the subsequent detailed analysis described under "Results." After consecutive breeding of the transgenepositive mice to C57B6J, the comparative study was conducted on the R838S ϩ mice from the lines 379 and 362 using their transgene-negative littermates as the wild type control.
Guanylyl Cyclase Assays-RetGC1 activity was assayed as previously described in detail (26,69). Briefly, the typical assay mixture contained in 25 l of 30 mM MOPS-KOH (pH 7.2), 60 mM KCl, 4 mM NaCl, 1 mM DTT, 2 mM Ca 2ϩ /EGTA buffer, MgCl 2 added to maintain 0.9 mM free Mg 2ϩ (unless indicated otherwise), purified GCAP1, and RD3 at concentrations indicated in the pertinent figure legends, 0.3 mM ATP, 4 mM cGMP, 1 mM GTP, 10 mM creatine phosphate, 0.5 unit of creatine phosphokinase (Sigma), 1 Ci of [␣-32 P]GTP, 0.1 Ci of [8-3 H]cGMP (PerkinElmer Life Sciences), phosphodiesterase inhibitors zaprinast and dipyridamole, and variable concentrations of GCAPs, RD3, and Ca 2ϩ . Membrane fractions from transfected HEK293 cell cultures expressing RetGC1 under control of the CMV promoter as previously described in detail (70) were assayed under ambient illumination. The reaction was incubated for 40 min at 30°C and quenched by heating at 95°C for 2 min. Mouse retinas for RetGC activity measurements were excised from dark-adapted 3-week-old mice using infrared illumination and a dissecting microscope fitted with an Excalibur infrared goggles as described (26), wrapped in aluminum foil, frozen in liquid N 2 , and stored at Ϫ70°C prior to their use in cyclase activity assays. The assays containing retinal samples were conducted under infrared illumination using Kodak number 11 infrared filters and Excalibur infrared binocular goggles, and the incubation time for the reaction was 12 min. The [ 32 P]cGMP product of the reaction was separated from the substrate using thin-layer chromatography on fluorescentback polyethyleneimine cellulose plates (Merk), eluted in 2 M LiCl solution, and quantified by liquid scintillation counting using ScintiSafe scintillation mixture (Fisher Scientific) containing 20% ethanol. Ca 2ϩ /EGTA buffers maintaining variable free Ca 2ϩ concentrations at physiological for the photoreceptors (71) 0.9 mM free Mg 2ϩ were prepared using the Tsien and Pozzan method (72) and verified by fluorescent indicator dyes as previously described (73).
ERG-Mice were dark-adapted overnight, their pupils were dilated by 1% tropicamide and 2.5% phenylephrine ophthalmic eye drops under dim red safelight illumination, and the mice were dark-adapted for another 15 min. Full-field electroretinography was performed in the dark using a Maxwellian view projection-based Phoenix Research Laboratories Ganzfeld ERG2 setup. Mice were anesthetized by inhalation of 1.7-1.9% isoflurane (VEDCO), delivered by a Kent Scientific SomnoSuite Small Animal Anesthesia System at the flow rate of 50 ml/min, and remained on a heated pad during the entire procedure to maintain body temperature. To evoke scotopic ERG responses, 0.1-1-ms 505-nm light pulses ranging between 1.9 and 1.1 ϫ 10 6 photoisomerizations (R*) per rod at the mouse retinal surface were delivered through a infrared camera-guided corneal electrode/LED light source in 1-s to 3-min intervals for different flash strengths, and typically 3-10 traces for each flash strength series were averaged. Constant 505-nm illumination, equivalent of 33 ϫ 10 3 R*/rod/s, was used as a background light to record photopic (cone) ERG responses (74) to 1-ms 505-and 365-nm pulses, equivalent of 5.8 ϫ 10 5 R*/rod, delivered at 1-s intervals; traces were averaged from 15 to 20 trials for each animal.
GCAP1 and RD3-Myristoylated bovine GCAP1 D6S was expressed from a Novagen pET11d vector in BLR(DE3) Escherichia coli strain harboring N-myristoyl transferase and purified to ϳ95% electrophoretic homogeneity from inclusion bodies using urea extraction, hydrophobic, and size exclusion chromatography as described previously in detail (75,76). Human RD3 was expressed from pET11d vector in BL21(DE3) CodonPlus E. coli strain (Stratagene/Agilent Technologies), extracted from the inclusion bodies and purified to ϳ95% electrophoretic homogeneity as previously described in detail (22,23).
Retinal Morphology-Mice were anesthetized with a lethal dose of ketamine/xylazine injection, perfused through the heart with phosphate-buffered saline (PBS), and then with 2.5% glutaraldehyde in PBS. The eyes were surgically removed and fixed overnight in 2.5% glutaraldehyde, 2.5% formaldehyde, PBS solution (Electron Microscopy Sciences) at 4°C. The fixed eyes were washed three times for 15 min each in PBS, soaked in PBS overnight, processed for paraffin embedding, sectioned, and stained with hematoxylin/eosin (AML Laboratories, Saint Augustine, FL). The retinas were photographed using an Olympus Magnafire camera mounted on an Olympus BX21 microscope. The photoreceptor nuclei in the outer nuclear layer of the retina were counted from the 425-m fragment of the 5-m thick retina section, between the optic nerve and the periphery, and the density of the nuclei per 100 m were averaged from several frames for each retina. The average nuclei counts were then typically compared using 3-5 retinas of different genotypes.
Immunofluorescence-Mice were perfused with the fixative solution as described above except that 10% freshly prepared formaldehyde solution in PBS was used instead of glutaraldehyde. The eyes were then dissected and the eyecups were placed into 10% formaldehyde/PBS for 5 min and then 5% formaldehyde/PBS for 1 h at room temperature. The fixed eyecups were impregnated with 30% sucrose solution in PBS for cryoprotection, placed in OCT medium (Electron Microscopy Sciences), and frozen at Ϫ70°C. The cryosections were taken using a Hacker-Bright cryogenic microtome, dried for 1 h at room temperature, and stored in Ϫ70°C. The sections were washed three times in PBS containing 0.1 M glycine (pH 7.4), blocked for 1 h at 30°C with the same solution containing 5% bovine serum albumin and 0.1% Triton X-100, incubated overnight at 4°C, and then 1 h at room temperature with anti-RetGC1 rabbit polyclonal antibody against catalytic domain of RetGC1 (68), washed with PBS solution three times for 15 min each, incubated with the 1:400 diluted donkey anti-rabbit Alexa Fluor 568-or Alexa Fluor 488-conjugated antibody (Molecular Probes), and washed four times with PBS for 15 min each at room temperature. Where indicated, fluorescein-labeled peanut agglutinin (Vector Laboratories) was added with the secondary antibody to label cone sheaths, and the sections were covered with TOPRO3 iodide-containing Vectashield mounting medium (Vector Laboratories) to counterstain the nuclei. Confocal images were acquired using an Olympus FV1000 Spectral instrument controlled by FluoView FV10-ASW soft-ware, collecting in a sequential mode the images of emission excited by 488-, 543-, and 635-nm lasers, assigned as, respectively, green, red, and pseudo-blue color channels. Where indicated, fluorescence was superimposed on a differential interference contrast image. No changes were made to the original images except for minor gamma correction applied to the whole image for more clear presentation in the print.
Immunoblotting-RD3 polyclonal antibody 10929 was produced in rabbit (22); GCAP1, GCAP2, and RetGC1 rabbit polyclonal antibodies were reported previously (17,68), G t ␣ antibody was purchased from Santa Cruz Biotechnology and ␤-actin antibody from GeneTex. Retinas from 3-week-old mice were excised in 50 l of 10 mM Tris-HCl (pH 7.4), containing 100 M phenylmethylsulfonyl fluoride and 10 g/ml of leupeptin, homogenized in 250 l of Laemmli SDS-containing sample buffer (Sigma), and subjected to electrophoresis in a 7% polyacrylamide gel containing 0.1% SDS. The volume of the retinal material in R838S RetGC1 versus wild type was corrected for a small decrease in photoreceptor nuclei count at that age. Following the electrophoresis, the proteins were transferred to iBlot PVDF mini stack cassette (Novex) for 22 min using a Invitrogen/Life Sciences iBlot instrument. The membranes were transiently stained by Ponceau S (Sigma) dye solution in 1% acetic acid to mark the positions of molecular mass markers, de-stained by a series of washes in water and Tris-buffered saline (Fisher Scientific) containing 0.5% Tween 20 (TTBS), blocked by Super Block (ThermoFisher) solution in TTBS, probed by rabbit polyclonal anti-RetGC1 antibody, and the luminescence was developed using peroxidase-conjugated goat anti-rabbit polyclonal IgG (Cappel/MP Biomedical) and a Pierce SuperSignal Femto substrate protocol (Thermo Scientific). The images were acquired using a Fotodyne Luminous FX imager and subjected to densitometry using National Institutes of Health ImageJ software.
Statistics-The data are mean average Ϯ S.D. Statistical significance of the differences was evaluated by unpaired/unequal variance Student's t test using Synergy Kaleidagraph 4 software.