Constitutive Activation of Photoreceptor Guanylate Cyclase by Y99C Mutant of GCAP-1

Photoreceptor membrane guanylate cyclases (RetGC) are regulated by calcium-binding proteins, GCAP-1 and GCAP-2. At Ca2+ concentrations below 100 nm, characteristic of light-adapted photoreceptors, guanylate cyclase-activating protein (GCAPs) activate RetGC, and at free Ca2+ concentrations above 500 nm, characteristic of dark-adapted photoreceptors, GCAPs inhibit RetGC. A mutation, Y99C, in human GCAP-1 was recently found to be linked to autosomal dominant cone dystrophy in a British family (Payne, A. M., Downes, S. M., Bessant, D. A. R., Taylor, R., Holder, G. E., Warren, M. J., Bird, A. C., and Bhattachraya, S. S. (1998) Hum. Mol. Genet. 7, 273–277). We produced recombinant Y99C GCAP-1 mutant and tested its ability to activate RetGC in vitro at various free Ca2+ concentrations. The Y99C mutation does not decrease the ability of GCAP-1 to activate RetGC. However, RetGC stimulated by the Y99C GCAP-1 remains active even at Ca2+ concentration above 1 μm. Hence, the cyclase becomes constitutively active within the whole physiologically relevant range of free Ca2+ concentrations. We have also found that the Y99C GCAP-1 can activate RetGC even in the presence of Ca2+-loaded nonmutant GCAPs. This is consistent with the fact that cone degeneration was dominant in human patients who carried such mutation (Payne, A. M., Downes, S. M., Bessant, D. A. R., Taylor, R., Holder, G. E., Warren, M. J., Bird, A. C., and Bhattachraya, S. S. (1998) Hum. Mol. Genet. 7, 273–277). A similar mutation, Y104C, in GCAP-2 results in a different phenotype. This mutation apparently does not affect Ca2+ sensitivity of GCAP-2. Instead, the Y104C GCAP-2 stimulates RetGC less efficiently than the wild-type GCAP-2. Our data indicate that cone degeneration associated with the Y99C mutation in GCAP-1 can be a result of constitutive activation of cGMP synthesis.

Ca 2ϩ enters outer segments (OS) of vertebrate photoreceptors through cGMP-gated Na ϩ /Ca 2ϩ channels in the plasma membranes. These channels are open in the dark, but they become closed in the light, because illumination stimulates cGMP hydrolysis by phosphodiesterase. Ca 2ϩ is constantly extruded from the OS by a light-independent Na ϩ /K ϩ , Ca 2ϩ exchanger, therefore interruption of Ca 2ϩ influx through the channels decreases the intracellular free Ca 2ϩ concentration (10,14,15), and that stimulates cGMP resynthesis in photoreceptors (10,16). This Ca 2ϩ feedback mechanism is essential for the recovery and light adaptation of photoreceptors (10).
RetGC itself is not sensitive to Ca 2ϩ , but it can interact with Ca 2ϩ sensor proteins, GCAP-1 and GCAP-2 (3,4,(11)(12)(13). A unique property of GCAPs is that they can be either activators or inhibitors of RetGC (17): at Ca 2ϩ concentrations below 100 nM, characteristic of light-adapted photoreceptors, GCAPs activate the cyclase, and at free Ca 2ϩ concentrations above 500 nM, characteristic of dark-adapted photoreceptors, GCAPs inhibit RetGC. GCAP-1 and GCAP-2 have four EF-hand Ca 2ϩbinding domains, and GCAPs can be turned into constitutive activators of RetGC by mutations that inactivate the ability of their EF-hands to bind Ca 2ϩ (17)(18).
The intracellular level of cGMP may be important not only for the phototransduction, but also for the viability of photoreceptors. Several types of rod or cone degeneration have been linked to the mutations in those photoreceptor proteins that regulate either synthesis or hydrolysis of cGMP (19 -23). Recently Payne et al. (1) described a new case of human autosomal dominant cone dystrophy associated with a point mutation in GCAP-1 gene. In this paper we present the evidence that this mutation, Y99C, causes a dramatic change in Ca 2ϩ sensitivity of GCAP-1. As a result, RetGC stimulated by the Y99C GCAP-1 remains active even at high free Ca 2ϩ concentrations. We also demonstrate that the corresponding mutation in GCAP-2 produces a different effect. Our data indicate that dominant cone degeneration associated with the Y99C substitution in GCAP-1 can be caused by permanent activation of cGMP synthesis.

EXPERIMENTAL PROCEDURES
Recombinant GCAP-1 and GCAP-2-Recombinant GCAP-1 and GCAP-2 were expressed in Escherichia coli according to the procedure described previously in detail (17,24), except that we used BLR(DE3)pLysS E. coli strain (Novagen) instead of BL21(DE3)pLysS. Myristoylated GCAP-2 was expressed as described previously (24). The N terminus of GCAP-1 is a poor substrate for yeast N-myristoyltransferase (NMT; Ref. 25). Substitution D6S makes it a better substrate (25), that allows us to produce GCAP-1, which is Ͼ90% myristoylated and is fully capable of regulating RetGC (Fig. 1). To make the GCAP-1 expression system, a cDNA encoding GCAP-1 was isolated from a bovine retinal cDNA library (a gift from Dr. D. Oprian, Brandeis University), amplified by polymerase chain reaction using forward primer AAAAAACCCATGGGGAACATTATGAGCGGTAAGTCGGTG and reverse primer ATATATGGATCCTTAAAGAGTAGGCAGTGAGCTCA. The resulting 0.65-kilobase pair fragment was inserted into the NcoI/BamHI restriction endonucleases sites of pET11d vector (Novagen) and expressed under the lac-controlled T7 promoter in the BLR(DE3)pLysS E. coli strain (Novagen) that harbored a plasmid encoding yeast NMT (a gift from Dr. J. Gordon, Washington University) as described previously (24). To produce Y 3 C substitutions fragments of GCAPs cDNAs were amplified by polymerase chain reaction using Pfu polymerase (Stratagene) and spliced by "splicing by overlap extension" (26). Pairs of primers encoding the base substitutions were: GGTACT-TCAAGCTCTGCGACGTGGACGGCAA and TTGCCGTCCACGTCGC-AGAGCTTGAAGTACC for making Y99C GCAP-1 and AGTGGACCT-TCAAGATCTGCGACAAGGACCGCAA and TTGCGGTCCTTGTCGC-AGATCTTGAAGGTCCACT for making Y104C GCAP-2. Mutant GCA-P-1 and GCAP-2 were expressed using the same method (24). Expressed proteins were purified as described previously (24) using chromatography on Sephacryl S-100 column. Positions of the mutations were verified by automated DNA sequencing (ABI Prizm, Perkin-Elmer). Calculated average isotopic mass for the myristoylated Y99C used in this study is 23,500.00. The actual average isotopic mass of purified Y99C GCAP-1 found by electrospray mass-spectrometry was 23,500.0 Ϯ 1.3. The nonmyristoylated form was undetectable.
RetGC Activity Assay-Washed bovine OS membranes (containing both RetGC-1 and RetGC-2) were prepared, depleted of endogenous GCAPs, reconstituted with recombinant GCAPs, and assayed as described previously (12,24). The assay mixtures ( . Reaction mixtures were incubated under infrared illumination for 12 min at 30°C. The reaction was stopped by heating for 2 min at 95°C. Samples were chilled on ice, centrifuged, and analyzed by TLC using fluorescent plastic-backed polyethylenimine cellulose plates (Merck). After development in 0.2 M LiCl, cGMP spots were visualized under UV illumination, cut, eluted with 1 ml of 2 M LiCl, mixed with 10 ml of an Ecolume scintillation mixture, and both 3 H and 32 P radioactivity were counted. [ 3 H]cGMP was used as an internal standard to ensure the absence of cGMP hydrolysis by light-sensitive phosphodiesterase. Ca/EGTA buffers were prepared according to (27).

RESULTS AND DISCUSSION
Mutation Y99C Affects Ca 2ϩ Sensitivity of GCAP-1-GCAPs are highly conserved proteins (28). Human, mouse, and bovine GCAP-1 are virtually identical within their EF-hands regions, and this is also true for GCAP-2 (Ref. 19; also, see Fig. 1, top  panel). When expressed as recombinant proteins, both GCAP-1 and GCAP-2 stimulate RetGC in a Ca 2ϩ -sensitive manner as it is shown in Fig. 1. It is also important to notice that GCAPs regulate RetGC within the submicromolar range of free Ca 2ϩ concentrations. The exact free Ca 2ϩ concentrations in rods and cones of mammals and humans have yet to be determined, but in dark-adapted resting rods of lower vertebrate, the free Ca 2ϩ concentration is near 550 nM, and it decreases to near 50 nM (10) after strong illumination. Therefore we consider the submicromolar range of free Ca 2ϩ as "physiologically relevant." To evaluate the potential functional significance of Y99C substitution reported by Payne et al. (1), we replicated this mutation in recombinant GCAP-1. The mutant protein was expressed in E. coli and purified as described under "Experimental procedures." We have found that the Y99C substitution does not hamper the ability of GCAP-1 to stimulate RetGC (Fig. 1A). Instead, the Y99C GCAP-1 fails to inhibit the cyclase at high Ca 2ϩ , so that RetGC remains equally active within the whole range of free Ca 2ϩ concentrations between 6 nM an 1 M. The Ca 2ϩ sensitivity of RetGC in the presence of the mutant GCAP-1 is decreased to such extent that the cyclase remains at near 50% of its maximal activity at Ca 2ϩ concentrations higher than 10 M.
GCAP-1 and GCAP-2 are nearly 40% identical to each other (12,28). However, despite the overall functional and structural similarity between GCAP-2 and GCAP-1, a similar substitution, Y104C, in GCAP-2 results in different biochemical phenotype than the Y99C mutation in GCAP-1 (Fig. 1B). The Y104C GCAP-2 inhibits activation of RetGC at low Ca 2ϩ , but its Ca 2ϩ sensitivity remains practically unaffected. Hence, the region adjacent to the EF-3 in GCAP-1 apparently plays a different role in RetGC regulation than the same region in GCAP-2. The difference between the Y99C GCAP-1 and the Y104C GCAP-2 in our experiments is consistent with other recent observations. First, unlike GCAP-1, inactivation of EF-3 in GCAP-2 has relatively minor effect on the regulatory properties of GCAP-2 (17,18). Second, EF-2 is very important for the Ca 2ϩ sensitivity of GCAP-2 (17), although it was postulated not to be essential for the activity of GCAP-1 (18). Third, a calcium-myristoyl switch has been postulated to be critical for the GCAP-1 activity (29); however myristoylation has only a minor significance for the general regulatory properties of GCAP-2 (24).
The Y99C GCAP-1 Competes with the Wild Type GCAP-1 and GCAP-2-Payne et al. (1) reported that the Y99C mutation in GCAP-1 gene had a dominant phenotype. The question is why does the presence of the normal allele(s) of GCAP(s) not protect cone cells from degeneration?
Even though the exact level of GCAP-1 and GCAP-2 expression in cones and rods has not been unambiguously defined, it has been well established that both GCAP-1 and GCAP-2 are expressed in photoreceptors (11-13, 29 -32). Several antibodies were raised in different laboratories that could detect both GCAP-1 and GCAP-2 in rods (12,30) and in cones (30,31) (some conclusions about the distribution of GCAP-1 and GCAP-2 in rods versus cones (29) were at variance apparently because of the different masking of GCAP-2 epitopes in animal species (30)). Both immunocytochemical (13, 30 -32) and in situ hybridization analyses (11) indicate that GCAP-1 is strongly expressed in cones. At the same time, Y99C mutation in GCAP-1 results only in cone dystrophy, and rods appear to be unaffected (1). This fact suggests that GCAP-1 is either not functioning in rods, or its concentration in rods is insignificant for RetGC regulation. It also strongly argues that GCAP-1 plays an important role in RetGC regulation in cones. On the other hand, GCAP-2 was initially found in rod outer segments (12). This localization of GCAP-2 in rods has been confirmed by other groups (30,32). However, a lower level of GCAP-2 expression in cones has also been detected (30,32). It is therefore possible that the normal alleles of GCAP-1 and GCAP-2 can both be present in the affected human cones along with the Y99C GCAP-1. Based on that assumption, we tested whether Y99C GCAP-1 could activate RetGC in the presence of both Ca 2ϩ -loaded GCAP-1 and GCAP-2 in vitro.
We have found that the Y99C GCAP-1 efficiently competes with Ca 2ϩ -loaded GCAP-1 and GCAP-2 and prevents their inhibitory effect at free Ca 2ϩ as high as 1 M (Fig. 2). The addition of nonmutant GCAP-1 and GCAP-2 increases the EC 50 for the RetGC activation by the Y99C GCAP-1, but it does not prevent RetGC from being activated by the mutant protein ( Fig. 2A). The Y99C GCAP-1 stimulates RetGC in the presence of equimolar concentrations of either wild type GCAP-1 or GCAP-2 (Fig. 2B). The normal GCAP-1 and GCAP-2 are able to only partially decrease RetGC activity stimulated by the Y99C GCAP-1, at free Ca 2ϩ above 1 M. Therefore, given that the intracellular free Ca 2ϩ in human photoreceptors in the dark is within the micromolar range (10), the Y99C mutation should be able to cause an excessive synthesis of cGMP in resting photoreceptors, even in the presence of normal GCAPs. That could explain the dominant phenotype of Y99C mutation in GCAP-1 found in vivo.
The steady-state dark/resting level of free cGMP in photoreceptors is maintained at the level of 3-4 M, and that keeps a few percent of cGMP gated ion channels in the open state (10). It is not immediately apparent why and how the Y99C GCAP-1 effect on RetGC activity would cause photoreceptors to degenerate. So, we can only suggest various scenarios that could potentially lead to the cell death. It is likely that constitutive synthesis of cGMP, especially when it is not balanced by phosphodiesterase activity in the resting photoreceptors, may increase the steady-state level of the free cGMP in the cytoplasm. In such case, high cGMP level, for example, would be able to alter the activity of the cyclic nucleotide-regulated protein kinase(s). Also, higher than normal cGMP concentrations can keep too many cGMP gated channels open in the dark and create excessive influx of both Na ϩ and Ca 2ϩ . Because the activity of RetGC in the presence of the Y99C GCAP-1 is not completely inhibited even by [Ca 2ϩ ] free above 10 M (Figs. 1A and 2B), cGMP synthesis may continue until the free concentration of Ca 2ϩ (and perhaps Na ϩ ) in the cell dramatically exceeds the normal resting level. That may affect cellular metabolism in general. For example, more ATP will be constantly utilized to extrude both Na ϩ and Ca 2ϩ from the resting cell. The elevated intracellular Ca 2ϩ concentrations could be also toxic for other vital cell functions.