Loss of the Effector Function in a Transducin-α Mutant Associated with Nougaret Night Blindness*

A missense mutation, G38D, was found in the rod transducin α subunit (Gαt) in individuals with the Nougaret form of dominant stationary night blindness. To elucidate the mechanism of Nougaret night blindness, we have examined the key functional properties of the mutant transducin. Our data show that the G38D mutation does not alter the interaction between Gαtand Gβγt or activation of transducin by photoexcited rhodopsin (R*). The mutant Gαt has only a modestly (∼2.5-fold) reduced k cat value for GTP hydrolysis. The GTPase activity of GαtG38D can be accelerated by photoreceptor regulator of Gprotein signaling, RGS9. Analysis of the GαtG38D interaction with cGMP phosphodiesterase revealed marked impairment of the mutant effector function. GαtG38D completely fails to bind the inhibitory PDE γ subunit and activate the enzyme. Altogether, our results demonstrate a novel molecular mechanism in dominant stationary night blindness. In contrast to known forms of the disease caused by constitutive activation of the visual cascade, the Nougaret form has its origin in attenuated visual signaling due to loss of effector function by transducin G38D mutant.

A missense mutation, G38D, was found in the rod transducin ␣ subunit (G␣ t ) in individuals with the Nougaret form of dominant stationary night blindness. To elucidate the mechanism of Nougaret night blindness, we have examined the key functional properties of the mutant transducin. Our data show that the G38D mutation does not alter the interaction between G␣ t and G␤␥ t or activation of transducin by photoexcited rhodopsin (R*). The mutant G␣ t has only a modestly (ϳ2.5-fold) reduced k cat value for GTP hydrolysis. The GTPase activity of G␣ t G38D can be accelerated by photoreceptor regulator of G protein signaling, RGS9. Analysis of the G␣ t G38D interaction with cGMP phosphodiesterase revealed marked impairment of the mutant effector function. G␣ t G38D completely fails to bind the inhibitory PDE ␥ subunit and activate the enzyme. Altogether, our results demonstrate a novel molecular mechanism in dominant stationary night blindness. In contrast to known forms of the disease caused by constitutive activation of the visual cascade, the Nougaret form has its origin in attenuated visual signaling due to loss of effector function by transducin G38D mutant.
The visual transduction cascade in vertebrate rod photoreceptor cells is among the best understood G protein signaling systems. In the outer segments of rod photoreceptor cells (ROS), 1 the visual G protein, transducin (G t ), couples the light receptor, rhodopsin (R), to the effector enzyme, cGMP phosphodiesterase (PDE). Photolyzed rhodopsin (R*) binds G t and induces GDP/GTP exchange on the G␣ t subunit, which upon dissociation from R* and G␤␥ t activates PDE. To activate the enzyme, G␣ t ⅐GTP interacts with the inhibitory PDE ␥ subunits (P␥) and prevents them from blocking catalytic sites on the PDE ␣␤ subunits. The following decrease in intracellular cGMP concentration leads to a closure of cGMP-gated channels in the ROS plasma membrane (1)(2)(3). An analogous visual cascade operates in cone photoreceptor cells. Unlike the cones that mediate high intensity daytime color vision, rod photoreceptors are designed for dim light vision during nighttime.
Defects in the genes encoding key components of the visual transduction cascade have been linked to a number of retinal diseases. Certain forms of retinitis pigmentosa, a progressive photoreceptor degeneration, are caused by mutations in rhodopsin (4), ␣and ␤-subunits of PDE (5-7), or the ␣-subunit of the cGMP-gated channel (8). Mutations in rhodopsin (9,10), PDE␤ (11), or G␣ t (12) have been identified in different forms of congenital stationary night blindness. Stationary night blindness is not associated with retinal degeneration and manifests itself in the inability to see in the dark, whereas daytime vision is largely unaffected. A missense G␣ t mutation, G38D, was found in the Nougaret form of congenital stationary night blindness (12). Initially, it was thought that Nougaret night blindness resulted from a loss of rod function. However, a recent study using electroretinographic analysis of Nougaret patients has indicated the presence of a detectable albeit subnormal rod function coupled with slight impairment of cone function (13). A clear outcome of the disease is a significant loss of rod light sensitivity (13).
Functional consequences of the G38D mutation for transducin signaling have not been examined, although a Gly residue at the corresponding position in other GTP-binding proteins has been actively investigated. This Gly is located within the conserved P-loop with the consensus sequence GXXXXK(S/T) that binds the ␣and ␤-phosphate of GDP and GTP in the superfamily of G proteins (14). One of the most common transforming mutations in p21 ras is a substitution of the corresponding residue Gly-12 by Val. This mutation essentially blocks the GTPase activity of p21 ras and prevents its stimulation by GAPs leading to constitutive activation of p21 ras -mediated pathways (15)(16)(17). Recently, a similar observation has been made for the analogous mutant of G␣ i1 , G42V, which hydrolyzes GTP with a 30-fold lower rate than that of the parent protein (18). The ability of G␣ i1 G42V to inhibit the effector enzyme, adenylyl cyclase, was not evaluated. Interestingly, G␣ z , in which the Gly residue is replaced by Ser, has a characteristically low k cat for GTP hydrolysis (19). The biochemical properties of p21 ras G12V and G␣ i1 G42V seem to point to the constitutive activity of the G␣ t G38D mutant as a cause of Nougaret night blindness. However, the alternative possibility that the G␣ t G38D mutation leads to an inactive visual cascade remained. Such a possibility is supported by analysis of an analogous G␣ s mutant, G␣ s G49V. The rate of GTP hydrolysis for G␣ s G49V is ϳ4-fold lower than that for G␣ s , and the GTP-bound mutant is capable of activation of adenylyl cyclase in vitro (20). However, G␣ s G49V was a very poor activator of adenylyl cyclase when expressed in a cyc Ϫ S49 cell line (21). The loss of light sensitivity of the Nougaret rod photoreceptors could potentially be caused either by the desensitization due to constitutive activity of G␣ t or by reduced ability to transduce a visual signal due to impaired G␣ t function.
To identify the molecular mechanism of Nougaret night blindness, we have introduced the G38D substitution into G␣ t -like, effector-competent, G␣ t /G␣ i chimeric ␣-subunit. An extensive examination of the functional properties of this mutant has revealed a complete loss of the effector function.

EXPERIMENTAL PROCEDURES
Materials-[ 35 S]GTP␥S (1160 Ci/mmol) and [ 32 P]NAD (1000 Ci/ mmol) were purchased from Amersham Pharmacia Biotech. Restriction enzymes were from New England Biolabs. T4 DNA ligase was from Roche Molecular Biochemicals. Cloned Pfu DNA polymerase was from Stratagene. TPCK-treated trypsin was from Worthington. All other chemicals were from Sigma or Fisher. Bovine ROS membranes were prepared as described previously (22). Urea-washed ROS membranes (uROS) were prepared according to protocol in Yamanaka et al. (23). G␤␥ t was purified according to Kleuss et al. (24).
Site-directed Mutagenesis-The G␣ t *G38D mutant was obtained using the pHis6-G␣ t * vector for expression of the Lys-248/Asp-251 3 His-244/Asn-247 (HN) mutant of G␣ t /G␣ i1 chimera 8 (25,26) as a template in the first round of PCR amplification. A forward primer carrying a NcoI site and a start codon was paired with a reverse primer coding the G38D mutation. This PCR product was used as a forward primer in the second round of PCR with pHis6-G␣ t * as a template and a reverse primer containing a BamHI site overlapping codons for G␣ t -207-209. The PCR product was digested with NcoI and BamHI and subcloned into pHis6-G␣ t *. The mutant sequence was confirmed by automated DNA sequencing at the University of Iowa DNA Core Facility. For expression of G␣ t * and G␣ t *G38D, 500 ml of 2ϫ TY medium was inoculated with 10 ml of overnight culture of BL21(DE3) cells transformed with pHis6-G␣ t * or pHis6-G␣ t *G38D. At A 600 ϭ 0.8, expression was induced by 30 M isopropyl-1-thio-␤-D-galactopyranoside for 3 h at 24°C. Cells were pelleted and kept frozen at Ϫ70°C. Purifications of recombinant proteins were carried out as described previously (25,26). Yields for 80 -90% pure G␣ proteins were normally 3-5 mg.
cDNA coding for RGS9 -284-461 (27,28) was PCR-amplified from the human retinal cDNA library (provided by J. Nathans, Johns Hopkins University, Baltimore, MD) using primers carrying the NdeI and BamHI restriction sites. The PCR product was subcloned into the pET15b vector for expression of RGS9 -284-461 as a His-tagged protein in Escherichia coli. RGS9 -284-461 was expressed and purified essentially as described previously (28).
In Vitro Transcription-Translation of G␣ t and G␣ t G38D-In vitro transcription-translations were carried out in the TNT T7 quick-coupled reticulocyte lysate system (Promega) according to the manufacture's recommendations. A plasmid pHis6-G␣ t G38D for in vitro translation of G␣ t G38D was obtained by excising the BamHI/HindIII fragment from pHis6-G␣ t *G38D and replacing it with the corresponding fragment from pHis6-G␣ t (25) containing wild-type bovine G␣ t cDNA. Plasmids pHis6-G␣ t and pHis6-G␣ t G38D were added to the translation mixture at a concentration of 4 g/ml. Transcription-translations were carried out for 2 h at 30°C with addition of 20 M methionine or, where indicated, 1 M [ 35 S]methionine. Translation mixtures using cold methionine (450 l) were combined with 50-l aliquots of translation mixtures using [ 35 S]methionine and applied onto a MonoQ HR 5/5 column (Amersham Pharmacia Biotech) equilibrated with 30 mM Tris-HCl buffer (pH 8.0) containing 2 mM MgSO 4 . The 35 S-labeled G␣ t or G␣ t G38D were eluted with a 0 -0.5 M NaCl/30-min gradient and concentrated using Microcon ® centrifugal filter devices with cutoff of 10,000 Da (Millipore). To determine levels of in vitro translation of G␣ t and G␣ t G38D, transcription-translations were carried out in the presence of 1 M [ 35 S]methionine and varying concentrations (1-10 M) of unlabeled methionine. The 35 S-labeled G␣ t or G␣ t G38D was then precipitated, and the percentage of incorporation of radioactive label was calculated by liquid scintillation counting according to the manufacture's (Promega) recommendations. Taking into account a ratio of 12 Met residues/G␣ t molecule, the calculated in vitro translation levels of G␣ t and G␣ t G38D were 25-30 nM.
Trypsin Protection Assay-G␣ t * or G␣ t *G38D (20 M) was incubated for 5 min at 25°C in 20 mM HEPES buffer (pH 8.0) containing 130 mM NaCl, 50 M GDP, and 5 mM MgSO 4 (buffer A). Where indicated, 10 mM NaF and 30 M AlCl 3 were included in the buffer. The GTP␥S-bound G␣ t * or G␣ t *G38D were prepared by preincubation of G␣ proteins (25 M) with bleached uROS (1 M rhodopsin) and G␤␥ t (3 M) in the presence of 50 M GTP␥S for 30 min at 25°C, followed by centrifugation for 60 min at 100,000 ϫ g to remove the membranes. For the trypsinprotection test of in vitro translated G␣ t or G␣ t G38D, 0.5-l aliquots of translation mixtures with additions of 1 M [ 35 S]methionine were diluted into 20 l of buffer A. The GTP␥S-bound G␣ t or G␣ t G38D were prepared by preincubation of translation mixtures with bleached uROS (0.5 M rhodopsin) and G␤␥ t (0.5 M) in the presence of 10 M GTP␥S for 10 min at 25°C. Trypsin digestions were performed with 25 g trypsin/ml for 15 min at 25°C and stopped with simultaneous addition of SDS sample buffer and heat treatment (100°C, 5 min).
GTP␥S Binding Assay-G␣ proteins (1 M) alone, mixed with 2 M G␤␥ t or 2 M G␤␥ t and uROS membranes (100 nM rhodopsin) were incubated for 3 min at 25°C. Binding reactions were started by addition of 5 M [ 35 S]GTP␥S (0.2 Ci). Aliquots of 50 l were withdrawn at the indicated times, mixed with 1 ml of ice-cold 20 mM Tris-HCl (pH 8.0) buffer containing 130 mM NaCl and 10 mM MgSO 4 , and passed through Whatman cellulose nitrate filters (0.45 m). The filters were then washed two times with the same buffer (2 ml, ice-cold) and counted in a liquid scintillation counter after dissolution in a 3a70B mixture. The k app values for the binding reactions were calculated by fitting the data to the equation, GTP␥S bound (% bound) ϭ 100% (1 Ϫ e -kt ).
Pertussis Toxin-catalyzed ADP-ribosylation-G␣ t * or G␣ t *G38D (0.5 M each) were mixed with G␤␥ t (0 -3 M) in 50 l of 20 mM Tris-HCl (pH 8.0) buffer containing 2 mM MgSO 4 , 2 mM dithiothreitol, 1 mM EDTA, 10 M GDP, and 3 g/ml pertussis toxin (preactivated with 100 mM dithiothreitol and 0.25% SDS for 10 min at 30°C). The reaction was started by addition of 5 M [ 32 P]NAD and allowed to proceed for 1 h at 25°C. Afterward, reaction mixtures were diluted with 1 ml of ice-cold 20 mM Tris-HCl (pH 8.0) buffer containing 100 mM NaCl and filtered through Whatman cellulose-nitrate filters. The filters were washed four times with the same buffer and counted in a liquid scintillation counter. In some experiments, 10-l aliquots were withdrawn from reaction mixtures, mixed with sample buffer for SDS-PAGE, and analyzed by electrophoresis in 12% gels.
Single-turnover GTPase Assay-Single-turnover GTPase activity measurements were carried out in suspensions of uROS membranes (5 M rhodopsin) reconstituted with G␣ t * or G␣ t *G38D (2 M) and G␤␥ t (2 M) essentially as described in Refs. 29 and 30. The GTPase reaction was initiated by addition of 50 nM [␥-32 P]GTP (0.2 Ci). The GTPase rate constants were calculated by fitting the experimental data to an exponential function: % GTP hydrolyzed ϭ 100(1 Ϫ e -kt ), where k is a rate constant for GTP hydrolysis.
Fluorescence Assays-Fluorescence assays of interaction between G␣ t * or G␣ t *G38D and P␥ labeled with 3-(bromoacetyl)-7-diethyl aminocoumarin at Cys68 (P␥BC) were performed on a F-2000 fluorescence spectrophotometer (Hitachi) in 1 ml of 20 mM HEPES (pH 7.5) buffer containing 100 mM NaCl, 5 mM dithiothreitol, and 4 mM MgCl 2 essentially as described in Ref. 31. The GTP␥S-bound G␣ t * or G␣ t *G38D were prepared as described for the trypsin-protection assay. When the AlF 4 Ϫ -activated G␣ t * or G␣ t *G38D were tested, the buffer contained 30 M AlCl 3 , 10 mM NaF, and 1 M GDP. Fluorescence of P␥BC (10 nM) was monitored with excitation at 445 nm and emission at 495 nm. Concentration of P␥BC was determined using ⑀ 445 ϭ 53,000.
PDE Activation Assay-PDE was extracted from ROS membranes and purified as described previously (32). HoloPDE (0.2 nM) was reconstituted with G␣ t * or G␣ t *G38D (1-2 M) and 2 M G␤␥ t in suspensions of uROS membranes containing 2 M rhodopsin. GTP␥S (10 M) was added to the reaction mixture, and PDE activity was measured using 100 M [ 3 H]cGMP similarly as described previously (33).
Other Methods-Protein concentrations were determined by the method of Bradford (34) using IgG as a standard or using calculated extinction coefficients at 280 nm. SDS-PAGE was performed by the method of Laemmli (35) in 12% acrylamide gels. Fitting of the experimental data was performed with nonlinear least squares criteria using GraphPad Prizm (version 2) software.

RESULTS
Expression and Trypsin Sensitivity of the G␣ t *G38D Mutant-A mutation, G38D, was introduced into the G␣ t -like G␣ t / G␣ i1 chimeric protein (25), which contained ϳ94% of G␣ t residues including the two key G␣ t effector residues His-244 and Asn-247 (26). This chimeric G␣ t , referred to as G␣ t *, is fully functional and able to bind the P␥ subunit and stimulate PDE activity comparably to native G␣ t (26). G␣ t * was chosen as a template for mutagenesis because it has efficient expression, which allows complete examination and comparison of functional properties of G␣ t * and G␣ t *G38D. Expression levels in E. coli of soluble mutant G␣ t *G38D were analogous to those of G␣ t * (ϳ4 -5 mg/liter of culture). Activation of G␣ subunits induces a conformational change that protects the G␣ switch II region from proteolysis with trypsin. Fig. 1 demonstrates the results of the trypsin-protection assay for G␣ t * and G␣ t *G38D. Both proteins were capable of undergoing an activational conformational change upon R*-induced binding of GTP␥S (Fig. 1). However, the GDP⅐AlF 4 Ϫ -induced conformation of G␣ t *G38D is different from that of G␣ t *, as indicated by the lack of its resistance to trypsin in the presence of AlF 4 Ϫ . Such a tryptic resistance pattern of G␣ t *G38D is consistent with similar results reported for the G42V mutant of G␣ i1 (18).
Kinetics of R*-dependent GTP␥S binding to G␣ t * and G␣ t *G38D-The ability of R* to activate G␣ t * and G␣ t *G38D was examined by measuring their GTP␥S-binding kinetics in the presence of G␤␥ t . The rate of GTP␥S-binding is controlled by a rate-limiting GDP release of G␣ subunits. Similarly to native G␣ t (25), G␣ t * and G␣ t *G38D in the absence (data not shown) or in the presence of G␤␥ t (Fig. 2) displayed very slow intrinsic rates of GDP release as measured by GTP␥S binding. This observation, that the G38D mutation does not significantly affect a high affinity of G␣ t for GDP, is in agreement with the GDP binding properties of G␣ i1 G42V (18). In the presence of uROS membranes (100 nM R*) and G␤␥ t (2 M), the rates of GTP␥S binding were markedly enhanced (Fig. 2). The comparable R*-induced GTP␥S binding rates by G␣ t * (k app ϭ 0.17 min Ϫ1 ) and G␣ t *G38D (k app ϭ 0.11 min Ϫ1 ) indicate that rhodopsin recognition in the mutant G␣ t * is generally not impaired.
Pertussis Toxin-catalyzed ADP-ribosylation of G␣ t * and G␣ t *G38D-Pertussus toxin effectively ADP-ribosylates G␣ t ⅐GDP at Cys-347 (36,37). The heterotrimeric complex, G␣␤␥ t is a notably better substrate than G␣ t alone (38), making ADP-ribosylation a useful tool to assess G␣ t interaction with G␤␥ t . The GTP␥S binding properties of G␣ t *G38D in the presence of R* and G␤␥ t indicated that the mutation did not grossly alter the G␣ t * binding to G␤␥ t . However, an excess of G␤␥ t was used in the binding assay to avoid the influence of potential defects in the G␣ t *G38D/G␤␥ t coupling on activation of G␣ t *G38D by R*. To further examine the interaction of G␣ t *G38D with G␤␥ t , we carried out a pertussis toxin-catalyzed ADP-ribosylation of G␣ t * and G␣ t *G38D in the presence of increasing concentrations of G␤␥ t . The dose dependences of G␤␥ t -supported ADP-ribosylation of G␣ t * and G␣ t *G38D were similar, suggesting that G␣ t *G38D retains intact interaction with G␤␥ t (Fig. 3).
GTPase Activity of G␣ t *G38D and Effects of RGS9 -Unaltered interaction of G␣ t *G38D with G␤␥ t and activation by R* allowed examination of mutant GTPase activity under single turnover conditions ([GTP] Ͻ [G␣␤␥ t *]). The GTPase activities of G␣ t * and G␣ t *G38D were measured in the reconstituted system with G␤␥ t and uROS membranes. uROS membranes lack the activity of a photoreceptor GAP, RGS9 (27). The calculated k cat for GTP hydrolysis by G␣ t * was 0.020 Ϯ 0.003 s Ϫ1 (Fig. 4A). G␣ t *G38D hydrolyzed GTP with a notably lower rate (k cat of 0.008 Ϯ 0.0004 s Ϫ1 ) (Fig. 4B). The reduction in the k cat value for GTP hydrolysis caused by the G␣ t *G38D mutation (ϳ2.5-fold) is proportional to that observed in the G49V mutant of G␣ s (20) but considerably smaller than a 30-fold decrease in the k cat of the G␣ i1 G42V mutant (18). The GTPase activity of the p21 ras G12V mutant is insensitive to the p21 ras GAP (17). We tested the effects of a truncated RGS9 protein (amino acids 284 -461) containing the RGS domain on rates of GTP hydrolysis for G␣ t * and G␣ t *G38D. Addition of 5 M RGS9 -284-461 resulted in acceleration of the GTPase activity of G␣ t by 6-fold (k cat 0.12 Ϯ 0.01 s Ϫ1 ) (Fig. 4A) and of G␣ t *G38D by ϳ3-fold (k cat 0.023 Ϯ 0.002 s Ϫ1 ).
Effector Properties of G␣ t *G38D-A fluorescence read-out assay was used to study the interaction between G␣ t *G38D and the P␥ subunit. It utilizes the P␥ subunit labeled at Cys-68 with the fluorescent probe, 3-(bromoacetyl)-7-diethyl aminocoumarin (P␥BC) (31). Binding of G␣ t to P␥BC causes a large FIG. 3. The G␤␥ t -dependent ADP-ribosylation of G␣ t * and G␣ t *G38D. Pertussis toxin-catalyzed ADP-ribosylation of G␣ t * or G␣ t *G38D G␣ t * (0.5 M) was carried out in the presence of increasing concentrations of G␤␥ t as described under "Experimental Procedures." A, [ 32 P]ADP-ribosylation of G␣ t * and G␣ t *G38D is analyzed by liquid scintillation counting. B, aliquots from the ADP-ribosylation reaction mixtures were analyzed by SDS-PAGE in 12% gels followed by autoradiography.
increase in the probe fluorescence. Using this assay, affinities of G␣ t *⅐GTP␥S (K d 1.7 Ϯ 0.3 nM) or G␣ t *⅐AlF 4 Ϫ (K d 3.2 Ϯ 0.3) nM) for P␥BC were similar (Fig. 5). Remarkably, G␣ t *G38D in both the GTP␥S-and AlF 4 Ϫ -activated conformations showed no detectable interaction with P␥BC (Fig. 5). To test the possibility that G␣ t *G38D binds P␥BC without causing a fluorescence increase, we investigated the binding of G␣ t * to P␥BC in the presence of high concentrations of G␣ t *G38D. No competition between G␣ t * and G␣ t *G38D for binding to P␥BC was detected as the K d value for the G␣ t */P␥BC interaction was essentially unchanged in the presence of 100 nM G␣ t *G38D (Fig. 5A). This result suggests the G␣ t mutation leads to a loss of the effector function. To confirm this conclusion, we evaluated the ability of G␣ t *G38D to stimulate activity of holoPDE reconstituted with uROS membranes and G␤␥ t in the presence of GTP␥S. For comparison, G␣ t * (2 M) was capable under these conditions of stimulating basal PDE activity by ϳ18-fold (Fig. 6). G␣ t *G38D failed to activate cGMP hydrolysis (Fig. 6). Moreover, excess of the GTP␥S-bound G␣ t *G38D did not interfere with activation of holoPDE by G␣ t *, further supporting the lack of competition between G␣ t * and G␣ t *G38D for the effector molecule (Fig. 6).
Binding of P␥BC to in Vitro Translated G␣ t and G␣ t G38D-We utilized in vitro translations of G␣ t and G␣ t G38D to confirm that the G38D mutation causes a loss of effector function not only in the background of the chimeric G␣ t * protein but in the wild-type G␣ t as well. In vitro translation is known to produce a functional G␣ t (39). Although yields of recombinant proteins using in vitro translation are typically very low, a high sensitivity fluorescence binding assay allows examination of the interaction between in vitro translated G␣ t and P␥BC. The calculated levels of in vitro translations for G␣ t and G␣ t G38D were 25-30 nM. The trypsin-protection assay for in vitro translated G␣ t and G␣ t G38D demonstrated patterns very similar to those of G␣ t * and G␣ t *G38D expressed in E. coli. Both G␣ t and G␣ t G38D displayed a characteristic trypsin-protection upon R*-induced binding of GTP␥S, but the GDP⅐AlF 4 Ϫ -bound conformation of G␣ t G38D was not resistant to trypsin (Fig. 7A). Addition of G␣ t , purified from an in vitro translation mixture, to P␥BC in the presence of AlF 4 Ϫ led to a large fluorescence change, whereas an analogously prepared G␣ t G38D had no effect (Fig. 7B). DISCUSSION Heterotrimeric G proteins transduce a variety of extracellular signals (neurotransmitters, hormones, light, olfactory and taste signals) from specific cell surface receptors to intracellular effectors. Mutations in G␣ subunits that lead to abnormal signaling are known to cause a large number of diseases (40). Abnormal G protein signaling might result from either attenuated or elevated signal transduction. Mutations of G␣ subunits leading to lower protein stability, reduced ability for interaction with and activation by cognate receptors, inability to dissociate G␤␥, or diminished capacity to activate effectors would attenuate G protein transduction. Mutations causing an increase in spontaneous G␣ GDP/GTP exchange, a lower GTPase activity, or impairment of interaction with GAP proteins are often a source of excessive G protein signaling.
Recently, a mutation in the visual G protein, G␣ t G38D, was found in patients with the Nougaret form of dominant stationary night blindness (12). Although psychophysical and electrophysiological consequences of the Nougaret night blindness have been thoroughly investigated (13), the molecular mechanism of the disease remained unclear. Despite a lack of biochemical characterization of G␣ t G38D, a wealth of information has been accumulated on mutations of an analogous Gly residue in small and heterotrimeric G proteins. The G12V mutation in p21 ras produces constitutively active signaling by inhibiting the p21 ras GTPase activity and abolishing its stimulation by GAP proteins (15)(16)(17). Likewise, the G␣ i1 G42V mutant has a 30-fold lower k cat for GTP hydrolysis in comparison to the wild-type G␣ i1 and it is insensitive to RGS proteins (18). By analogy with the p21 ras and G␣ i1 mutants, constitutive activity of G␣ t G38D becomes the most appealing model for abnormal function of rod photoreceptors in the Nougaret pedigree. Persistent activation of PDE by the GTP-bound G␣ t G38D would also provide a simple explanation to the dominant inheritance of the disease (12). Moreover, such a hypothesis would be consistent with existing biochemical evidence on dominant night blindness caused by two mutations in rhodopsin, G90D and A292E (9,10). Both rhodopsin mutants are constitutively active and capable of activating transducin even in the absence of the retinal chromophore (9,10). The PDE ␤-subunit gene, a third gene implicated in dominant stationary night blindness, carries a missense mutation H258N (11). Although properties of the PDE mutant are not known, it has been suggested that it might be constitutively active due to impaired interaction with the P␥ subunit (11). Finally, an Oguchi disease, a recessive form of stationary night blindness, was found to be associated with null mutations in the rhodopsin kinase gene (41). Rhodopsin kinase is intimately involved in inactivation of R*. Transgenic mice lacking rhodopsin kinase displayed larger and longer than normal single-photon responses (42). The excessive signaling is thought to saturate rod responses at abnormally low light-intensities in Oguchi disease (42). Similarly to light adaptation, constitutive activation of the visual cascade would cause desensitization of rod photoreceptors by lowering cGMP levels, keeping the cGMP-gated channels closed, the plasma membrane hyperpolarized, and lowering intracellular Ca 2ϩ concentration. Supporting the hypothesis for constitutive activity of G␣ t G38D, some of the abnormalities seen in the rod and cone functions of Nougaret patients have been simulated by light adaptation of the normal retina (13). However attractive, constitutive activity toward effectors of G␣ mutants with substitution of the Gly residue has not been firmly established. Examination of G␣ s G49V revealed that it can normally stimulate adenylyl cyclase in the reconstituted system (20), but only poorly in cyc Ϫ S49 cells expressing the mutant (21).
To elucidate the biochemical mechanism of Nougaret night blindness, we investigated the key functional properties of G␣ t *G38D, such as interaction with R* and G␤␥ t , GTPase activity, interaction with RGS9, binding the P␥ subunit, and the ability to stimulate PDE. G␣ t *G38D interaction with G␤␥ t and activation by R* was found largely intact. In comparison with G␣ t *, the mutant had only a very modest ϳ2.5-fold reduction in the k cat value for GTP hydrolysis. The decrease in G␣ t *G38D GTPase activity was significantly smaller than that seen in the G␣ i1 G42V or G42S mutants (18). In addition, unlike G␣ i1 G42V (18), G␣ t *G38D retained reduced ability to interact with RGS proteins, particularly, with a photoreceptorspecific RGS9. In contrast, the effector function of G␣ t *G38D is markedly impaired. G␣ t *G38D fails to bind P␥ and activate PDE. The inability of the G38D mutant, in a background of wild-type G␣ t , to interact with the effector molecule was demonstrated using in vitro translated G␣ t and G␣ t G38D. Lack of trypsin protection of G␣ t G38D (or G␣ t *G38D) in the presence of AlF 4 Ϫ indicates that the active conformation of the switch II region of the mutant G␣ t may differ from that in wild-type transducin. The switch II region of G␣ t is an essential effector binding domain (26,43), and alterations in this region provide a plausible rationale for the loss of effector function. Moreover, similar patterns of sensitivity to trypsin between active conformations of G␣ i1 G42V (18) and G␣ t G38D allow us to speculate that these mutations may have caused analogous conformational changes. A crystal structure of the GTP␥S bound G␣ i1 G42V shows that the Val-42 side chain forces the peptide planes of the switch II residues 203-206 to rotate leading to disruption of ionic contacts between Arg-205 and the switch III/␣3-helix residues Asp-237 and Glu-245 (18). If a similar conformational change is caused by the G␣ t G38D mutation, it would break the linkage between the G␣ t switch II Arg-201 and the switch III/␣3-helix residues Glu-232 and Glu-241 (44). This linkage is central to the ability of G␣ t to assume the effectorcompetent conformation (26,45). All mutations of the G␣ t switch II region residues that are involved in the linkage with switch III/␣3-helix have resulted in a severe impairment of G␣ t effector function (26).
An absolute inability of G␣ t G38D to activate PDE may only in part account for desensitization of rod photoreceptors in the Nougaret form of dominant night blindness. Heterozygous Nougaret patients, who presumably have about 50% functional transducin, display more than 99% reduction in rod sensitivity (13). The observed rod sensitivity loss implies that G␣ t G38D somehow interferes with expression of the wild-type allele, or the mutant has dominant negative properties. We demonstrated that G␣ t *G38D is unable to prevent PDE activation by G␣ t *. Despite the lack of evidence to support the dominant negative nature of the G38D mutant, such a possibility cannot be ruled out. G␣ t *G38D is fully capable of interaction with R*, and under dim light conditions and very low concentrations of R*, the mutant G␣ t may potentially compete with wild-type G␣ t for R*. Such a competition would slow the rate of formation of G␣ t ⅐GTP and, consequently, attenuate PDE activation.
Overall, our results demonstrate a novel molecular mechanism in dominant stationary night blindness. In contrast to stationary night blindness caused by constitutive activation of the visual cascade by rhodopsin mutants (9,10), the Nougaret form has its basis in decreased visual signaling due to loss of transducin effector function. FIG. 7. Properties of in vitro translated G␣ t and G␣ t G38D. A, the trypsin-protection test. Translation mixtures containing G␣ t or G␣ t G38D (0.5 l each) were diluted into 20 l of 20 mM HEPES buffer (pH 8.0) containing 130 mM NaCl and 5 mM MgSO 4, and treated with trypsin (25 g/ml) for 15 min at 25°C in the presence of 10 M GDP or 10 M GDP with 30 M AlCl 3 and 10 mM NaF. The GTP␥S-bound G␣ t and G␣ t G38D were obtained as described under "Experimental Procedures." The proteolytic patterns were analyzed using fluorography of 12% polyacrylamide gels. B, interaction of in vitro translated G␣ t and G␣ t G38D with P␥BC. The relative fluorescence change (F/F o ) of P␥BC (10 nM) (excitation at 445 nm, emission at 495 nm) was determined after addition of G␣ t ⅐GDP or G␣ t G38D⅐GDP (8 nM each) in the presence of AlF 4 Ϫ .