Comparative analysis of cone and rod transducins using chimeric Gα-subunits

Abstract

The molecular nature of transducin-α subunits (Gα_t) may contribute to the distinct physiology of cone and rod photoreceptors. Biochemical properties of mammalian cone Gα_t2-subunits and their differences with rod Gα_t1 are largely unknown. Here, we examined properties of chimeric Gα_t2 in comparison with its rod counterpart. The key biochemical difference between the rod- and cone-like Gα_t was ~10-fold higher intrinsic nucleotide exchange on the chimeric Gα_t2. Presented mutational analysis suggests that weaker interdomain interactions between the GTPase (Ras-like) domain and the helical domain in Gα_t2 are in part responsible for its increased spontaneous nucleotide exchange. However, the rates of R*-dependent nucleotide exchange of chimeric Gα_t2 and Gα_t1 were equivalent. Furthermore, chimeric Gα_t2 and Gα_t1 exhibited similar rates of intrinsic GTPase activity as well as similar acceleration of GTP hydrolysis by the RGS domain of RGS9. Our results suggest that the activation and inactivation properties of cone and rod Gα_t–subunits in an in vitro reconstituted system are comparable.

The two types of photoreceptors in the vertebrate retina, rods and cones, mediate vision at dim and bright light, respectively. Cones produce small rapidly-decaying responses and are much less sensitive to light than rods. The molecular basis for the differences in physiology of rods and cones are poorly understood. The phototransduction cascades in rods and cones utilize homologous components including visual pigments, heterotrimeric G proteins (transducins), and cGMP-phosphodiesterases (PDE6) (1–3). In rods, a robust amplification in the signaling cascade is achieved due to a high rate of transducin activation by photolyzed rhodopsin (Meta II or R*) and rapid hydrolysis of cGMP by transducin-activated PDE6. Consistent with the low sensitivity of cones, the signal amplification in mouse cone phototransduction was found to be lower (4). The molecular differences between major cone and rod signaling proteins have been examined as a probable origin of the physiological differences (5–9). Current evidence does not implicate the signaling properties of rod and cone visual pigments. Expression of rhodopsin and red cone pigment in Xenopus cones and rods, respectively, yielded responses identical to those of native Xenopus photoreceptors (5). Recent physiological analysis points to an important role of transducin α-subunits in shaping specific photoresponses (7). Rods of GNAT2C transgenic mice with substitution of rod Gα_t1 for cone Gα_t2 displayed decreased sensitivity, reduced rate of activation, and accelerated recovery characteristic of cone photoreceptors (7). Reduced rate of cascade activation may result from a slower rate of R*-dependent activation of the Gα_t2 complex with rod-specific Gβ₁γ₁ or less efficient activation of PDE6 by Gα_t2 (10). Accelerated recovery of GNAT2C rods suggests faster GTP hydrolysis in the transition complex of Gα_t2 with PDE6 and the RGS9-1 GTPase accelerating protein (GAP) complex (11). Thus, Gα_t2 may have a higher intrinsic rate of GTP hydrolysis and/or a greater GTPase potentiation by the GAP complex than Gα_t1.

In contrast to well-characterized rod Gα_t1, biochemical properties of Gα_t2 have not been investigated. Native Gα_t1 is readily available for biochemical analyses, whereas the sparsity of cones in mammalian retina impedes isolation of native Gα_t2. In addition, a wealth of information about Gα_t1 properties, including its interaction with R*, PDE6 and RGS9, was developed using robust bacterial expression of transducin-like Gα_t1/Gα_i1 chimeras (12–14). Here, we applied a chimera approach to analyze key signaling properties of Gα_t2. A Gα_t2/Gα_i1 chimera (Gα_t2′) was produced and investigated in comparison to analogous Gα_t1/Gα_i1 chimera Chi8 (13), termed hereafter Gα_t1′. Gα_t2′ and Gα_t1′ are ~94% identical to Gα_t2 and Gα_t1, respectively, which is significantly greater than the degree of homology between Gα_t2 and Gα_t1 (~82%). Gα_t2′ showed a large ~10-fold increase in spontaneous nucleotide exchange rate in comparison to Gα_t1′. A series of chimeric and mutant Gα_t subunits delineated Gα_t2′ residues responsible for high intrinsic nucleotide exchange rate. In contrast, the rates of R*-dependent nucleotide exchange were comparable for Gα_t2′ and Gα_t1′ reconstituted with Gβ₁γ₁. Furthermore, Gα_t2′ and Gα_t1′ exhibited similar rates of intrinsic GTPase activity as well as similar acceleration of GTP hydrolysis by RGS9.

Experimental procedures

Materials

Guanosine 5′-[γ-³⁵S]thiotriphosphate triethylammonium salt (GTPγS; 1100 Ci/mmol), guanosine 5′-[γ-³²P]triphosphate triethylammonium salt (~5000Ci/mmol), and nicotinamide adenine dinucleotide ([³²P]NAD; 800 Ci/mmol) were from PerkinElmer. Pertussis toxin was from Sigma. Bovine outer rod segment (ROS) membranes were prepared as described (15). Urea-washed ROS membranes (uROS) were prepared according to a published protocol (16). Recombinant Gβ₁γ₁ complex was expressed using the baculovirus/sf9 cell system and isolated as described (17). The RGS-domain of RGS9 (RGS9d, aa 284–461) was expressed and purified as previously described (14).

Cloning, experession and purification of chimeric Gα_t subunits

To obtain chimera Gα_t2′-2 (Fig. 1), the full-length Gα_t1′ (Chi8) sequence was PCR-amplified from the plasmid (13) with a pair of sequence-specific primers (reverse primer contained BamHI site) and cut with BamHI to isolate a 430-bp fragment coding the C-terminal portion of Gα_t1′. This fragment was inserted into the large fragment of the pET15b-Gα_t2 vector (9) digested with BamHI, thereby replacing the C-terminal portion of Gα_t2 with the corresponding fragment of Gα_t1′. Correct orientation of the insert was selected by sequencing the construct. Gα_t2′ was generated from Gα_t2′-2 by replacing three Gα_t1-specific residues, Val301, Glu305 and Arg310, with the corresponding Gα_t2-specific residues Ser305, Asp309, and Lys314. The triple mutant was produced using the QuikChange mutagenesis protocol (Stratagene). The same mutagenesis procedure was used to generate the A144S mutant of Gα_t2′and the following single, double, and triple mutants of Gα_t1′: S140A, K117P/S120V, S153N/D154Q, and V159T/T160D/G162E.

Figure 1.

(A). Schematic representation of Gα_t chimeras. The switch III region (S3, Gα_t1-228-236) (black) is identical in Gα_t1 and Gα_t2. (B). Coomassie blue-stained SDS-gel showing purified Gα_t chimeras. Samples represent equal fractions of the preparations from 0.5 liter bacterial cultures.

Chimeras Gα_t2′-3 and Gα_t2′-4 were produced by swapping the residues 1–116 and 117–215 of Gα_t1′ with residues 1–120 and 121–219 of Gα_t2, respectively. The Gα_t2′-1-120 sequence was amplified from the Gα_t2′-template using a 5′-primer containing an NcoI site and a 3′-primer with a flanking Gα_t1′ sequence. The resulting PC product was paired with a 3′ primer containing a HindIII site in PC amplification from the Gα_t1′ template to produce the Gα_t2′-3 sequence, which was then cloned into the NcoI/HindIII sites of the pHis6 vector (13). To produce Gα_t2′-4, the Gα_t1′-1-116 sequence was amplified from the Gα_t1′ template using a 5′-primer containing an NcoI site and a 3′-primer with a flanking Gα_t2′ sequence. The resulting PC product was paired with a 3′ primer containing a XhoI site in PC amplification from the Gα_t2′-2 template to produce the Gα_t2′-4 sequence, which was then cloned into the NcoI/XhoI sites of the pET15b vector. Upon sequence verification, protein expression of Gα_t1′, Gα_t2′, their derivatives and mutants, was induced in BL21(DE3) codonPlus competent cells with 30 μM IPTG at 16°C overnight. His₆-tagged proteins were purified on His-bind Ni-NTA resin (Novagen) followed by an ion-exchange chromatography on Uno-Q1 column (Bio-Rad). Fractions containing functional Gα-subunits were dialyzed against buffer containing 40% glycerol overnight and stored at −20°C.

GTPγS binding assay

Chimeric Gα_t subunits (1 μM) alone, or mixed with 1 μM Gβ₁γ₁ and uROS membranes (0.25, 1 or 2 μM rhodopsin), were incubated for 2 min at 25ºC in the presence of light. Binding reactions were started with the addition of 5 μM or 100 nM [³⁵S]GTPγS. Aliquots of 15 μl were withdrawn at the indicated times, mixed with 1 ml ice-cold 20 mM Tris-HCl (pH 8.0) buffer containing 130 mM NaCl, 2 mM MgSO₄, and 1 mM GTP, passed through Whatman cellulose nitrate filters (0.45 μm), and washed three times with 3 ml of the same buffer without GTP. The filters were dissolved in 5 ml of a xylene-based 3a70B counting cocktail (RPI Corp.) and [³⁵S]GTPγS was measured in a liquid scintillation counter (18). The k_app values for the binding reactions were calculated by fitting data with equation %GTPγS bound=100(1-e^−kt). Groups of measurements were compared with two-tailed unpaired t test.

GTPase activity assays

Single-turnover GTPase activity measurements were carried out in suspensions of uROS membranes (10 μM rhodopsin) reconstituted with chimeric Gα_t subunits (1 μM) and Gβ₁γ₁ (1 μM) in 10 mM Tris-HCl (pH 7.5) buffer containing 100 mM NaCl and 4 mM MgSO₄ (18, 19). Where indicated, RGS9d (3 or 6 μM) was added. After incubation for 5 min at 25ºC, 100 nM [γ-³²P]GTP was added to the mixtures to initiate the reaction. GTPase hydrolysis was quenched at the indicated times by mixing 10 μl aliquots with 100 μl of 6 % (v/v) perchloric acid. Nucleotides were precipitated with 700 μl of 10% (w/v) charcoal suspension in phosphate-buffered saline, and free [³²P_i] was measured by liquid scintillation counting. GTPase rate constants were calculated by fitting the data equation %GTP hydrolyzed = 100(1-e^−kt), where k_cat is the rate constant for GTP hydrolysis.

Pertussis toxin-catalyzed ADP-ribosylation

Chimeric Gα_t subunits (0.5 μM each) were mixed with Gβ₁γ₁ (0.5–2 μM) in 50 μl of 20 mM Tris-HCl (pH 8.0) buffer containing 2 mM MgSO₄, 2 mM dithiothreitol, 1 mM EDTA, 10 μM GDP and 5 μg/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 μM [³²P]NAD and allowed to proceed for 1 hr at 25°C. 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. Aliquots (10 μl) were withdrawn from reaction mixtures, mixed with sample buffer for SDS-PAGE and analyzed by SDS-PAGE and autoradiography.

Results

Increased spontaneous nucleotide exchange in chimeric Gα_t2 subunits

To examine the biochemical properties of cone transducin-α, we generated a chimeric Gα_t2′-subunit, which is a counterpart of previously characterized Gα_t1/Gα_i1 chimera Chi8 or Gα_t1′ (13) (Fig. 1A). This choice of a chimeric template is based on the efficient bacterial expression of Gα_t1′ and the fact that it is 94% identical to Gα_t1 and appears to recapitulate its essential signaling characteristics. Indeed, expression of functional Gα_t2′ in E. coli was similarly robust, and the protein is readily purified (Fig. 1B).

In comparison to Gα_i1, native Gα_t1 and Gα_t1/Gα_i1 chimeras containing 215 N-terminal residues of Gα_t1 have been shown to have very slow rates of spontaneous guanine nucleotide exchange rate (13). In agreement, the basal rate of GTPγS-binding to Gα_t1′ under our experimental conditions was 0.0050±0.0006 min⁻¹. In contrast to Gα_t1′, Gα_t2′ showed markedly higher intrinsic GTPγS-binding rate (0.048±0.003 min⁻¹) (Fig. 2). Next, we tested chimeric Gα_t2′-2, containing 219 N-terminal residues of Gα_t2 (Fig. 1A). The basal GTPγS-binding rate for Gα_t2′-2 was similar to that of Gα_t2′ (Fig. 2), suggesting that the N-terminal part of Gα_t2′ is responsible for the increased nucleotide exchange. Chimeras Gα_t2′-3 and Gα_t2′-4 were produced by swapping the residues 1–116 and 117–215 of Gα_t1′ with the corresponding residues of Gα_t2 (Fig. 1A). The rate of GTPγS binding to Gα_t2′-4 was nearly as high as that for Gα_t2′, whereas Gα_t2′-3 demonstrated relatively small ~1.7-fold increase in the binding rate over Gα_t1′ (p=0.016) (Fig. 2). The Gα_t determinants of the spontaneous nucleotide exchange within the Gα_t1(117–215) region were further probed by mutational analysis of Gα_t1′. This region contains only a few residues that are different between rod and cone Gα_t, but strongly conserved within each transducin family (Suppl. Fig 1). The following single, double, and triple mutations replacing rod-specific residues with their cone-specific counterparts were introduced into Gα_t1′: S140A, K117P/S120V, S153N/D154Q, and V159T/T160D/G162E (Fig. 3A). The double and triple mutations did not significantly alter the GTPγS-binding rate of Gα_t1′, while the S140A substitution led to a moderate 1.8-fold increase in the k_app value for GTPγS binding (p=0.001) (Fig. 3B). We then examined if the reverse mutation A144S in Gα_t2′ would alter its nucleotide exchange kinetics (Fig 3C,D). Indeed, the GTPγS-binding rate of the A144S mutant was 1.7-fold lower than that of Gα_t2′(p=0.004) (Fig. 3D).

Figure 2.

(A). Spontaneous GTPγS binding to Gα_t1′ and Gα_t2′-chimeras. The binding of GTPγS to Gα_t1′ and Gα_t2′-chimeras (1 μM each) was initiated with the addition of 5 μM [³⁵S]GTPγS. Gα-bound GTPγS was counted by withdrawing aliquots at the indicated times and passing them through Whatman cellulose nitrate filters. Results from one of four similar experiments are shown. (B) k_app values (min⁻¹) (mean±SE) of intrinsic GTPγS binding to chimeric Gα_t-proteins were calculated from four experiments such as shown in A.

Figure 3.

(A, C)Coomassie blue-stained SDS-gel showing purified mutant Gα_t1′- and Gα_t2′-subunits. Samples represent equal fractions of the preparations from 0.5 liter bacterial cultures. (B, D) k_app values (min⁻¹) (mean±SE) of intrinsic GTPγS binding to mutant Gα_t1′ and Gα_t2′ calculated from four binding experiments for each mutant. (E). k_app values (min⁻¹) (mean±SE) of intrinsic GTPγS binding to mutant Gα_t2′ (1 μM) in the presence of 1 and 2 μM Gβ₁γ₁.

Binding of GTPγS to Gα subunits with high spontaneous nucleotide exchange such as Gα_i′ or Gα_s′ can often be inhibited by Gβγ-subunits (20, 21). Gβ₁γ₁ had no effect on the basal rate of GTPγS binding to Gα_t2′ (Fig. 3E).

Pertussis toxin-catalyzed ADP-ribosylation

Binding of Gβ₁γ₁-subunits to Gα_t1 facilitates pertussis toxin-catalyzed ADP-ribosylation at Cys347 of Gα_t1 (22–24). We utilized this reaction to assess the interactions of Gα_t1′ and Gα_t2′with Gβ₁γ₁. Pertussis toxin-catalyzed ADP-ribosylation of Gα_t1′ and Gα_t2′ were carried out in the absence or presence of increasing concentrations of Gβ₁γ₁. The basal level of ADP-ribosylation in the absence of Gβ₁γ₁ was somewhat higher for Gα_t2′ compared to Gα_t1′ (Fig. 4). The dose-dependencies of Gβ₁γ₁-supported ADP-ribosylation were comparable for Gα_t1′ and Gα_t2′ suggesting similar submicromolar K_d values for Gβ₁γ₁ binding (Fig. 4).

Figure 4. The Gβ₁γ₁-dependent ADP-ribosylation of Gα_t1′ and Gα_t2′.

Pertussis toxin-catalyzed ADP-ribosylation of Gα_t1′ and Gα_t2′ (0.5 μM each) was carried out in the presence of increasing concentrations of Gβ₁γ₁. (A) [32P]ADP-ribosylation of Gα_t1′ and Gα_t2′ is analyzed by liquid scintillation counting (mean±SE, n=3). (B) Aliquots from the ADP-ribosylation reaction mixtures were analyzed by SDS-PAGE followed by autoradiography.

Rhodopsin-catalyzed activation of chimeric Gα_t2 and Gα_t1 reconstituted with Gβ₁γ₁

The efficiencies of activation of Gα_t2′ and Gα_t1′ by R* were measured using a GTPγS-binding assay and reconstitution of the Gα-subunits with Gβ₁γ₁ and uROS. Suspensions of uROS containing 0.25 μM and 1 μM R* were used to ensure that the GTPγS-binding rates are submaximal. In the presence of uROS containing 0.25 μM R*, the rates of R*-dependent GTPγS binding to Gα_t2′ (k_app 0.0045±0.0005 s⁻¹ or 0.27 min⁻¹) and Gα_t1′(k_app 0.0048±0.0008 s⁻¹ or 0.29 min⁻¹) were similar and much greater than the unstimulated rates (Fig. 5). In the presence of uROS containing 1 μM R*, the GTPγS-binding rates to Gα_t2′ and Gα_t1′ were higher still (~0.011 s⁻¹ or 0.7 min⁻¹) (Fig. 5). No significant differences were seen in the activation kinetics of the two Gα_t proteins.

Figure 5.

(A) Rhodopsin-catalyzed GTPγS binding to Gα_t1′ and Gα_t2′. The binding of GTPγS to Gα_t1′ and Gα_t2′ (1 μM each) in the presence of 1 μM Gβ₁γ₁ and uROS membranes (0.25 or 1 μM rhodopsin) was initiated with the addition of 5 μM [³⁵S]GTPγS. Gα-bound GTPγS was counted by withdrawing aliquots at the indicated times and passing them through Whatman cellulose nitrate filters. Results from one of four similar experiments are shown. (B) k_app values (s⁻¹) (mean±SE) of intrinsic GTPγS binding to chimeric Gα_t-proteins were calculated from four experiments such as shown in A.

GTPase activity of chimeric Gα_t2 and Gα_t1 and the effects of RGS9d

The catalytic rates of GTP hydrolysis for Gα_t2′ and Gα_t2′-2 were determined in comparison to Gα_t1′ using a single turnover assay (GTP=100 nM ≪ Gα_tβ₁γ₁=1 μM) and a high concentration of uROS (10 μM R*). Using uROS containing 2 μM R* and 100 nM GTPγS, the rates of GTPγS-binding to Gα_t1′ and Gα_t2′ were significantly faster than the measured rates of GTP hydrolysis (Suppl. Fig. 2). Thus, the guanine nucleotide binding rates did not limit GTPase reactions under our experimental conditions. The intrinsic GTPase activity of Gα_t2′ (k=0.024±0.001 s⁻¹) was not significantly different from that of Gα_t1′ (k=0.023±0.001 s⁻¹) (Fig. 6). The rate of GTP hydrolysis by Gα_t2′-2 was also similar to that of Gα_t1′ (Suppl. Fig. 3).

Figure 6. Single turnover GTPase assays of Gα_t1′ and Gα_t2′. Effects of RGS9d.

(A) Single-turnover GTPase activity measurements were carried out in suspensions of uROS membranes (10 μM rhodopsin) reconstituted with chimeric Gα_t subunits (1 μM) and Gβ₁γ₁ (1 μM). Where indicated, RGS9d was added. Reactions were started with the addition of 100 nM [γ-³²P]GTP and free ³²P_i was measured by liquid scintillation. Results from one of three similar experiments are shown. (B) The k_cat values (s⁻¹) (mean±SE) for GTP hydrolysis by chimeric Gα_t-proteins were calculated from three experiments such as shown in A.

The ability of the RGS9-284-461 domain to stimulate GTPase activity of Gα_t2′ and Gα_t1′ was measured at two RGS-protein concentrations, 3 and 6 μM. RGS9-284-461 comparably stimulated GTPase activities of the cone and rod chimeric Gα-subunits. The levels of GTPase activity of Gα_t2′ and Gα_t1′ were ~ 2.2–2.5 and ~3-fold higher in the presence of 3 μM and 6 μM RGS9-284–461, respectively (Fig. 6B).

Discussion

The role of transducin-α subunit in setting the sensitivity and kinetics of light responses of rods and cones has been actively debated (6, 7, 25). Electrophysiological recordings from mouse cones indicate that in mouse S- and M-cones the amplification of phototransduction is 2–3 fold lower and the inactivation is considerably faster than in rods (4). In view of the similar catalytic efficiencies of rod and cone PDE6 (9) and the smaller size of cone outer segment, the lower amplification implies at least 5-fold lower rate of PDE6 activation per activated pigment molecule in mouse cones compared to rods (4). Replacing the rod Gα_t1 with the cone Gα_t2 reduced the amplification in mouse rods by ~2-fold, and speeded up the recovery phase by ~2 – fold, suggesting that the differences in the rod and cone responses might be attributed to a significant extent to the nature of Gα_t (7). We sought to determine the biochemical basis underlying the physiological changes induced by Gα_t2 in GNAT2C rods. Although, the functional GTPγS-bound form of Gα_t2 can be isolated following transducin expression in E. coli (9), this preparation cannot be used to study the activation of Gt by R* or Gα_t2 GTPase activity. Therefore, we produced and examined chimeric Gα_t2–like proteins.

The key biochemical difference between rod-like Gα_t1′ and its cone counterpart Gα_t2′ was found to be markedly higher intrinsic nucleotide exchange of Gα_t2′. The region primarily responsible for the increased nucleotide exchange was mapped to residues Gα_t2 (121–219). Gα-subunits of heterotrimeric G proteins are composed of two domains, a GTPase (Ras-like) domain and a helical domain. GDP (or GTP) is buried between the two domains (26, 27). A network of interactions between the two domains is involved in control of basal nucleotide exchange. Mutations disrupting the interdomain interactions increased the basal nucleotide exchange in Gα_i (28). Within Gα_t2 (121–219), a conserved cone-specific residue Ala144 corresponds to the conserved rod-specific Ser140. Ser140 from the helical domain interacts with Asp227 and Lys273 from the Ras-like domain (26, 27). The model of the Ser→Ala substitution using the structure of Gα_tGDP (27) demonstrates that the interdomain interactions involving Ser140 are disrupted (Fig. 7). Previous study found no evidence for the role of Ser140 in controlling Gα_t1 nucleotide exchange (29). However, a trypsin-protection assay as readout of Gα_t1 activation is severely limited in terms of the kinetic resolution. Our results suggest that the Ser→Ala substitution in Gα_t2 contributes to its high intrinsic nucleotide rate. The Gα_t1′Ser140Ala mutation increased, whereas the Gα_t2′Ala144Ser mutation decreased the nucleotide exchange rates of the respective Gα subunits. The effects of these mutations on the basal GTPγS binding rates were moderate, suggesting the involvement of other residues. Gβ₁γ₁ did not inhibit the high nucleotide exchange rate of Gα_t2. This agrees with the impairment of the interdomain contacts in Gα_t2 as the cause of its high nucleotide exchange rate. Gβ binds only to the N-terminus and the Ras-like domain of Gα, and, thus may not influence the interdomain interface (12). The finding that the GTPγS-bound Gα_t2 can be isolated through the spontaneous nucleotide exchange supports the notion that native Gα_t2 also has high intrinsic nucleotide exchange rate (9). The functional significance of this phenomenon, if any, remains to be determined. Spontaneous activation of G_t2 in cones may elevate basal activity of PDE6, potentially altering the sensitivity and kinetics of cone light responses (30).

Figure 7. Interdomain interactions of Ser140 in rod Gα_t1 are lost in the S140A mutant.

In the structure of Gα_t1GDP (24), Gα_t1Ser140 from the helical domain interacts with Asp227 and Lys273 from the Ras-like domain (left). The S140A mutation was introduced into the structure of Gα_t1GDP using the Swiss-PdbViewer (v.4) (44). The mutation disrupts the interdomain interactions in Gα_t1GDP (right). The images were produced using PyMOL 1.4.1 (45).

Our experiments revealed no significant differences in the binding of Gβ₁γ₁ to Gα_t1′ and Gα_t2′ (Fig. 4), nor in the efficiency of Gα_t1′ and Gα_t2′ activation by R* in the presence of Gβ₁γ₁ (Fig. 5). However, the rates of activation of non-N-acylated recombinant Gα in the R*, Gβ₁γ₁-reconstituted system observed previously and in this study are well below the rates of activation of native Gt under similar conditions (13, 31, 32). Thus, our activation paradigm may not be capable of detecting small differences in the R*-Gt coupling for native Gα_t2 and Gα_t1. Moreover, the potential role of cone-specific Gβ₃γ₈ to transducin/R* coupling has not been investigated in this study. The interdomain interactions in Gα-subunits appear to be essential to receptor-mediated activation of G proteins. Mutant Gα_s-subunits with disrupted interdomain interactions showed decreased ability for activation by the β-adrenergic receptor (33). The relatively weak interdomain interface in Gα_t2 may potentially reduce the Gα_t2 coupling to R*. Our results seem to favor the idea that the lower amplification in GNAT2C rods is linked to the lower efficiency of Gα_t2 coupling to rod PDE6 rather than to R*. Yet, the potency of rod PDE6 activation by Gα_t2GTPγS in solution is not significantly lower than the activation by Gα_t1GTPγS (9). Nonetheless, rod PDE6 activation by Gα_t1 on the membrane is markedly more potent than in solution (34). Alternatively, small reductions in both couplings, R*/Gα_t2 and Gα_t2/PDE6, may result in the combined 2-fold lower amplification in GNAT2C rods. In cones, however, a short lifetime of photoactivated cone pigments due to rapid spontaneous decay and pigment phosphorylation, and a lower rate of Gt₂ activation, may dictate the activation phase of photoresponses (8, 35, 36).

The intrinsic GTPase activities of Gα_t2′ and Gα_t1′ were found to be similar. The k_cat for the unstimulated GTP hydrolysis by Gα_t1′ is equivalent to the previously reported k_cat values of native Gα_t1 (14, 37). Furthermore, the intrinsic GTPase activities of Gα_t1 and Gα_i1 are comparable (14). Thus, the chimeric Gα_t-subunits appear to faithfully reflect intrinsic GTPase activities of Gα_t2 and Gα_t1, which, in all probability, are similar. The GTPase activity of transducins in vivo is far greater than its intrinsic activity owing to the GAP activity of the membrane-bound RGS9 protein complex with the cooperative input from the effector PDE6 (38, 39). The faster rate of transducin inactivation in cones compared to rods may result from the greater potency or the higher expression levels of the RGS9 complex (14, 40). The former possibility is suggested by the finding that RGS9 is a much more potent GAP for Gα_t1 compared to Gα_i1, and the selectivity determinants reside in the Gα_t1 helical domain (14). Nonetheless, the lack of the difference in the effects of the RGS9d on the GTPase activities of Gα_t2 and Gα_t1 indicates that the efficacies of RGS9 towards cone and rod transducins are not grossly different. This result is consistent with the strong conservation of RGS9-contact residues of transducin (41) and supportive of RGS9 protein levels as a determining factor in transducin inactivation. The rate of transducin inactivation is regulated by the membrane attachment of the RGS9 GAP complex (42, 43). Therefore, further studies with the use of the N-acylated Gα_t subunits are needed to probe the selectivity of the RGS9 GAP complex.

Abbreviations

Gα_t1

rod transducin α-subunit

Gα_t2

cone transducin α-subunit

Gα_t2′

Gα_t2/Gα_i1 chimera

PDE6

photoreceptor phosphodiesterase-6

R*

photoexcited rhodopsin

RGS9

regulator of G-protein signaling 9

GTPγS

guanosine 5′-O-(3-thiotriphosphate)