Rhodopsin activation exposes a key hydrophobic binding site for the transducin alpha-subunit C terminus.

Conformational changes enable the photoreceptor rhodopsin to couple with and activate the G-protein transducin. Here we demonstrate a key interaction between these proteins occurs between the C terminus of the transducin alpha-subunit (G(Talpha)) and a hydrophobic cleft in the rhodopsin cytoplasmic face exposed during receptor activation. We mapped this interaction by labeling rhodopsin mutants with the fluorescent probe bimane and then assessed how binding of a peptide analogue of the G(Talpha) C terminus (containing a tryptophan quenching group) affected their fluorescence. From these and other assays, we conclude that the G(Talpha) C-terminal tail binds to the inner face of helix 6 in a retinal-linked manner. Further, we find that a "hydrophobic patch" comprising key residues in the exposed cleft is required for transducin binding/activation because it enhances the binding affinity for the G(Talpha) C-terminal tail, contributing up to 3 kcal/mol for this interaction. We speculate the hydrophobic interactions identified here may be important in other GPCR signaling systems, and our Trp/bimane fluorescence methodology may be generally useful for mapping sites of protein-protein interaction.

How might this movement enable coupling with transducin? We and others have hypothesized that helix 6 movement opens a cleft that provides a binding site for the transducin ␣ subunit C-terminal tail (G T␣ C terminus) (7,16,17) (Fig. 1). Several lines of evidence show that the rhodopsin cytoplasmic face directly interacts with the G T␣ C terminus. For example, peptides and peptide analogues corresponding to the G T␣ C terminus can bind and stabilize MII rhodopsin (18 -20), and binding induces structure in these peptides (21)(22)(23)(24). Rhodopsin mutagenesis studies show that residues in intracellular loops i2 and i3 are important in this interaction (25)(26)(27)(28), and a photocross-linking probe on loop i3 at residue 240 on rhodopsin cross-links to this region of G T␣ (29). However, direct evidence for where and how the G T␣ C terminus and rhodopsin interact is lacking.
In the present manuscript we set out to test whether the G T␣ C terminus binds to the cleft formed during rhodopsin activation through the use of a combination of site-directed labeling and fluorescence methods. Using these approaches, we were able to map where these regions interact by monitoring the tryptophan quenching of bimane fluorescence from modified cysteine residues on rhodopsin. Our results indicate that the G T␣ C-terminal tail does indeed bind to the inner face of helix 6, and we obtained evidence that critical hydrophobic contacts in this region control the affinity of the rhodopsin-transducin interaction, with key residues contributing up to 3 kcal/mol binding energy. We propose that the lack of these contacts is the underlying defect in several mutant rhodopsin proteins and speculate that the activation mechanism described and the hydrophobic interactions identified here may be universally important in other GPCR signaling systems.
The peptides used in this study are variants of the high affinity analogues of G T ␣ C-terminal 11 residues 340 -350 originally developed by Hamm and co-workers (19,20). All of the peptides contain free N and C termini and were synthesized by the Emory Microchemical Facility (Atlanta, GA) and analyzed by Micro high pressure liquid chromatography analysis and matrix-assisted laser desorption ionization mass spectrometric analysis. The name and amino acid sequence for each peptide are as follows: 23S, NH2 VLEDLKSCGLF COOH , and W23SV, NH2 WVLEDLKSVGLF COOH . All of the peptides were prepared in Buffer D at pH 6.0. The sulfhydryl-reactive fluorescent probes 1-(2maleimidylethyl)-4-(5-(4-methoxyphenyl)oxazol-2-yl)pyridinium methanesulfonate (PyMPO maleimide), and (2-pyridyl)dithiobimane (PDTbimane) were purchased from Molecular Probes (Eugene, OR) and Toronto Research Chemicals (North York, Canada), respectively. All of the remaining materials used in this study are similar to those described previously (30).
PDT-bimane Labeling of Rhodopsin Mutants and Determination of Labeling Efficiency-Labeling of samples with PDT-bimane and subsequent purification was carried out when the protein was bound to the 1D4 matrix as described previously for monobromobimane (7). After elution from the immunoaffinity column, a spectrum of each elution fraction was recorded (described below), and the purified samples were either used immediately or snap frozen in liquid N 2 and stored at Ϫ80°C. The amount of PDT-bimane label incorporation and the presence of unreacted free label were determined as follows (59). Labeling efficiency was determined by incubating the samples in the presence of 50 M TCEP reducing agent at 4°C for 10 min in buffer D followed by the addition of 10% trichloroacetic acid to each sample to precipitate the protein. The samples were incubated at 4°C for 15 min in this mixture and then centrifuged at 100,000 ϫ g for 30 min, and fluorescence emission scans were carried out on both sets of supernatants using 380-nm excitation and monitoring emission from 410 to 605 nm. These data were then used to determine the concentration of PDT-bimane in each sample by comparison with a standard curve generated from serial dilutions of stock PDT-bimane. To determine the amount of free, unreacted label, the above procedure was carried out except the protein was first precipitated by the addition of trichloroacetic acid, and any remaining signal was attributed to label not attached to the protein.
UV-visible Absorption Spectroscopy-All of the UV-visible absorption spectra were recorded with a Shimadzu UV-1601 spectrophotometer at 20°C using a bandwidth of 2 nm, a response time of 1 s, and a scan speed of 500 nm/min unless otherwise noted. The concentration of protein was calculated using a molar extinction coefficient value (⑀ 500 ) for WT rhodopsin of 40,600 M Ϫ1 cm Ϫ1 (40), and the molar extinction coefficient (⑀ 380 ) of PDT-bimane was 5000 M Ϫ1 cm Ϫ1 . The samples were photobleached in buffer D by illumination for 30 s (at a 6-Hz flash rate) with a Machine Vision Strobe light source (EG&G) equipped with a Ͼ495-nm-wavelength long pass filter. This light treatment was found to be adequate for full conversion of all samples to the MII state.
Measurement of the Rate of Retinal Release and/or MII Decay by Fluorescence Spectroscopy-The MII stability of the PDT-bimane-labeled mutants was assessed by monitoring the fluorescence increase of opsin during retinal release using a Photon Technologies QM-1 steadystate fluorescence spectrophotometer and standard instrument settings (39,41).
Peptide Inhibition of MIII Formation-The ability to inhibit MIII formation was used to monitor peptide binding to each mutant rhodopsin. In these assays, the samples were incubated in the presence or absence of 100 M peptide W23SV (in buffer D, pH 6.0) on ice for 10 min, and UV-visible spectra were recorded every 5 min from 650 to 250 nm at 10°C. Following an initial dark state scan, the samples were photobleached to the MII state, and the effect of the peptide on MIII formation was monitored at 460 nm over time (see Fig. 2A). Instrument base-line drift was corrected for by setting the absorbance of all spectra to 0 at 650 nm. For comparison purposes duplicate experiments were performed on each sample in the absence of peptide (see Figs. 2B and 3A).
Peptide Blocking of Rhodopsin Mutant V250C-The reactivity of rhodopsin mutant V250C toward the sulfhydryl-specific probe PyMPO maleimide was carried out as described previously (7). Briefly, 3 g of mutant rhodopsin was reacted with 25-fold excess PyMPO for 1 min in Buffer F, pH 7.2, at 4°C in either the dark state or following photobleaching (30 s) to the MII state. To test the ability of peptide W23SV to block PyMPO labeling of mutant V250C the sample was preincubated with 100-fold excess of peptide prior to photobleaching and to the addition of PyMPO. All of the reactions were quenched by the addition of L-cysteine to a 1 mM final concentration. The samples were subsequently resolved by SDS-polyacrylamide gel electrophoresis, and label incorporation was visualized using a Bio-Rad Gel Documentation Instrument. The background mutant (which contains no reactive cysteine residues) was used as a control to show PyMPO labeling specificity. Following label visualization, the gels were subjected to Coomassie staining to assure equal distribution of protein in each of the sample wells.
KI Fluorescence Quenching Measurements-Steady-state fluorescence quenching measurements of V250B were carried out as described previously (7). Briefly, 125 l of a 250 nM sample was excited at 380 nm (0.25-nm bandpass setting), and emission was monitored from 405 to 605 nm (12-nm-bandpass setting). Five separate samples with KI concentrations ranging from 0 to 25 mM were measured in both the dark state and in the MII state following illumination with Ͼ490 nm light for 30 s. Separate samples were also measured in a similar manner following preincubation (10 min on ice) with 100 M peptide W23SV. The salt concentration of each sample was kept constant at 25 mM by the addition of a corresponding amount of KCl, and Na 2 S 2 O 3 was added to 0.10 mM to inhibit the formation of I 3 Ϫ . The resulting data were plotted as fluorescence intensity at 466 nm versus concentration of quenching agent (KI) to calculate the Stern-Volmer quenching constant, K SV , according to Equation 1 below.
FIG. 1. Light activation of rhodopsin opens a cleft to enable binding of the G T ␣ C terminus. A, model of the rhodopsin cytoplasmic face based on the rhodopsin crystal structure (51). The black spheres indicate the ␣-carbon sites for the cysteine residues used in the present study. The proposed outward helix 6 movement during rhodopsin activation is shown as a gray helix and is based on EPR derived distance constraints (5). The possible outward movement of helix 3 is also denoted (3). B, proposed model of the hydrophobic cleft formed by rhodopsin activation. A model of the peptide used in the present studies, peptide W23SV, is based on NMR structural coordinates of a undecapeptide corresponding to residues (340 -350) of the G T␣ C terminus in its MII bound state (22). The Trp residue in this peptide is shown for visualization purposes. The residues shown to be important for transducin activation (Leu 226 , Thr 229 , and Val 230 ) are shown in gray. In both models the rhodopsin extracellular domain and portions of cytoplasmic loop i3 and the C-terminal tail have been removed for clarity.
Fluorescence Quenching of PDT-bimane-labeled Rhodopsin Samples by Peptide W23SV-Unless noted otherwise, the steady-state fluorescence measurements of PDT-bimane-labeled rhodopsin mutants were carried out at 10°C using 500 nM sample in buffer D, pH 6.0. All of the spectra were buffer-subtracted and corrected for sample dilution when necessary. The fluorescence emission spectra were recorded from 405 to 610 nm (15-nm bandpass) unless otherwise noted, while being excited at 380 nm (0.25-nm bandpass). Dark state spectra were recorded first, followed by photobleaching the sample to the MII state. Note that conversion to the MII state results in a large increase in fluorescence intensity from the labeled samples because of the shift in rhodopsin absorbance from 500 to 380 nm, which removes spectral overlap with PDT-bimane fluorescence, thereby relieving the quenching caused by energy transfer from the bimane to retinal. Peptide quenching experiments were performed by the addition of peptide in buffer D, pH 6.0, to MII rhodopsin to the desired concentration. In control experiments we find that the fluorescence intensity of the labeled samples in the MII state remains relatively constant throughout the short time course of the peptide quenching assays (data not shown). Additionally, less than 8% of the MII photointermediate decays over the time course of the assay. The data were acquired using the program Felix (Photon Technologies Inc.) and plotted using Sigma Plot (Jandel Scientific Software).
Determination of Transducin (G T ) Activation Rates-Activation of G T by rhodopsin was monitored using fluorescence spectroscopy at 10°C, using described previously conditions (5,39,42,43). Briefly, this involved exciting the samples at 295 nm (2-nm bandwidth) and monitoring fluorescence emission at 340 nm (12-nm bandwidth). Photobleached mutant rhodopsin (see above) was added to a concentration of 5 nM to the reaction mixture consisting of 250 nM G T in 10 mM Tris, pH 7.2, 2 mM MgCl 2 , 100 mM NaCl, 1 mM dithiothreitol, and 0.01% DM, and the mixture was allowed to stir for 300 s. The reaction was then initiated by adding GTP␥S to a final concentration of 5 M, and the fluorescence increase was monitored for 2000 s. To calculate the initial activation rates, the slopes of the initial fluorescence increase following GTP␥S addition were determined through the data points covering the first 60 s.
Peptide Titration Assays and Calculation of EC 50 and ⌬⌬G ‡ Values-The EC 50 values for quenching of PDT-bimane-labeled mutant rhodopsin samples by W23SV were determined by titrating the samples with increasing amounts of peptide. All of the experiments were performed at 10°C to minimize the decay of the MII species. It was necessary to account for the decay of the MII species (t1 ⁄2 Ϸ 38 min at 10°C; see Table  II) during the time course of the assay. Thus, the titration experiments were performed by splitting the samples in two and titrating each sample separately with different concentrations of peptide. In this manner dose-response curves were generated in less than 10 min, in which time less than 8% of the MII species had decayed. These peptide quenching data were combined and analyzed using nonlinear regression by fitting the data to a sigmoidal dose-response (variable slope) equation (four-parameter logistic equation; see Equation 2 below) (20).
In Equation 2, "minimum" is the Y value for fluorescence at the bottom plateau (fully quenched) and "maximum" is the Y value for fluorescence at the top plateau (not quenched). logEC 50 is the logarithm of the EC 50 , the concentration that gives a response halfway between maximum and minimum. The ⌬⌬G ‡ values for mutants K141B/226A, K141B/229A, and K141B/230A were calculated using the EC 50 values to reflect binding K eq , and thus the ⌬⌬G ‡ values were calculated according to Equation 3 below.
In Equation 3, R is the universal gas constant, T is the temperature, and K eq is the EC 50 for peptide quenching.

RESULTS
Peptide W23SV Binding Protects V250C on the Inner Face of Helix 6 -Previously, we established that helix 6 movement in rhodopsin can be detected by two simple methods: through monitoring an increased reactivity of a cysteine residue introduced into the inner face of this helix at site Val 250 and by monitoring the increased solvent accessibility of a fluorescent probe attached at this site (7). Here we used these methods to probe the proposed G T␣ C terminus binding site in conjunction with a high affinity peptide analogue of the G T␣ C terminus. This peptide, W23SV (sequence WVLEDLKSVGLF), is a derivative of a peptide originally constructed by Hamm and coworkers (called 23S (19,20)), in which we added a tryptophan residue to the N terminus to act as a quenching group. We also exchanged the reactive cysteine residue with a valine (20) to ensure that it would not react with or complicate the cysteine studies.
At the start of these experiments, we established that peptide W23SV binds rhodopsin mutant V250C by monitoring its ability to inhibit formation of a 460-nm photoproduct, called MIII, which normally occurs as the MII species decays to its inactive forms (44,45) (Fig. 2A). In addition, MII stabilization studies carried out by monitoring "extra MII" in digitonin (21) indicated that both peptide 23S and W23SV were able to bind to light-activated rhodopsin (data not shown).
As seen in Fig. 2B, binding of peptide W23SV completely blocks the reactivity of V250C in the MII state toward the cysteine-labeling agent PyMPO maleimide (Fig. 2B). Similar results were obtained using peptide 23S (data not shown). However, because of possible cross-reactivity of the cysteine in peptide 23S with the fluorescent probes, the remaining work was performed using peptide W23SV. Binding of peptide W23SV reduces the solvent accessibility of a fluorescent probe (bimane) attached at site 250. As indicated from the Stern-Volmer plot (Fig. 2C), the W23SV peptide shields the fluorescent probe from quenching by KI only in the MII state. Taken together, these results provide compelling evidence that the G T␣ C terminus binds to the cleft formed during rhodopsin activation. However, they do not formally rule out the possibility that the results described above are due to an indirect effect on the environment surrounding site V250B caused by the peptide binding. To alleviate this potential concern, we turned to a novel approach, described below.
Mapping Rhodopsin/G T␣ C-terminal Interactions-To more directly map where the G T␣ C terminus binds on the face of rhodopsin, we turned to a novel method we have recently developed for mapping protein-protein interactions. This approach exploits the fact that a Trp residue substantially quenches the fluorescence of a bimane label only when the two are essentially within contact distance (46). Because bimane is a small probe (similar in size to a Trp residue), this new method clearly defines localized sites of interaction, because substantial quenching only occurs if the Trp and bimane are able to contact one another during the excited state of the bimane fluorophore.
The goal in these experiments was to see whether the Trp residue in peptide W23SV could quench the fluorescence of bimane labels attached at different sites on the cytoplasmic face of rhodopsin (Fig. 1). We carried out this experiment using rhodopsin mutants containing single reactive cysteine residues introduced at sites 140, 141, 150, 228, 248, 250, and 316 in a nonreactive background mutant (called Theta (31, 37)). As can be seen in Fig. 1, these sites are distributed throughout the rhodopsin cytoplasmic face. These mutants were expressed, purified, and labeled with bimane at a 1:1 label to protein ratio (Scheme 1).
Photobleaching and MII decay assays confirmed that the bimane label did not perturb the structure or stability of these samples. The labeling and functional characterization of these bimane-labeled rhodopsin mutants are presented in Table I.
Before carrying out the mapping experiments, we first established that 100 M peptide W23SV could bind to all of the bimane-labeled mutants, as demonstrated by its ability to block MIII formation (Fig. 3A). We next assessed the effect that the Trp quenching group on peptide W232SV has on the fluorescent properties of the individually bimane-labeled rhodopsin mutants. We found that most of the mutants showed no fluo-rescence change upon peptide W23SV binding (Fig. 3B). Only two mutants, C140B and K141B (where B indicates a PDTbimane label), showed a substantial and reproducible fluorescence decrease, with one mutant, F228B, showing a small decrease as well. As seen in Fig. 3C, the quenching observed for C140B, K141B, and F228B occurred in a dose-dependent fashion, with peptide quenching exhibiting a titration of effect. The data from these experiments are summarized in graphical form in Fig. 4.
Importantly, we find that the W23SV-induced quenching is reversible, because the addition of an excess of peptide without the tryptophan residue (peptide 23S) can fully displace W23SV quenching (Fig. 5A). Furthermore, the quenching is specific for the MII form of rhodopsin, because it does not occur if peptide W23SV is added following the complete decay of the MII species (Fig. 5B) or following cleavage of the retinal Schiff base with hydroxylamine (NH 2 OH) as shown in Fig. 5C. Finally, we find that peptide W23SV does not bind to dark state rhodopsin (data not shown). Taken together, these results suggest peptide W23SV binds to the cleft formed in MII rhodopsin in an orientation that places its N-terminal tryptophan within near contact distance of the bimane probe located at positions Cys 140 and Lys 141 , and to a much lesser extent, Phe 228 (Fig. 1A).
Hydrophobic Interactions Govern Rhodopsin/G T␣ C Terminus Affinity-Interestingly, the rhodopsin crystal structure shows that sites that exhibit Trp/bimane quenching (C140B, K141B, and F228B) are near several residues (Leu 226 , Thr 229 , and Val 230 ; Fig. 1B) previously identified by cysteine scanning studies to be important for transducin activation (31). We hypothesized that these residues may form an unappreciated "hydrophobic patch" that is required for rhodopsin/transducin interactions. To test this theory, we measured the contribution of these residues to G T␣ C-terminal binding, using the above mentioned peptide quenching assay. First, we substituted ala-  4 and 8). Bottom panel, Coomassie staining of the same gel illustrating equal amounts of protein is present in each lane. C, peptide W23SV binding shields a fluorescent label (bimane) at site V250C from the fluorescence-quenching agent KI. Stern-Volmer analysis shows the label at V250C in the dark state (, K SV ϭ 2.7 M Ϫ1 ) becomes more accessible to KI quenching in the MII state (E, K SV ϭ 70.5 M Ϫ1 ). However, binding of peptide W23SV (100 M) protects the label from collisions with KI in the MII state (Ⅺ, K SV ϭ 19.4 M Ϫ1 ), although it has no effect on the dark state (OE, K SV ϭ 2.5 M Ϫ1 ). SCHEME 1. Reaction of PDT-bimane with cysteine residues incorporated into rhodopsin. The label is attached through a reversible disulfide bond. a All of the samples are rhodopsin cysteine mutants generated in the Theta background and labeled specifically at the introduced cysteine residue with the PDT-bimane probe.
b Pure rhodopsin typically exhibits a protein (A 280 ) to chromophore (A 500 ) absorbance ratio of 1.6 to 1.8.
c The amount of PDT-bimane label incorporated into the rhodopsin single cysteine mutants are presented as the number of labels/rhodopsin mutant (label/sample). The error values are S.E. (n ϭ 2). d The amount of free label remaining in each sample following labeling and purification is presented as a percentage of the total sample (for further details see ''Experimental Procedures'').
e The MII decay rate of the labeled samples was determined in buffer D as described under ''Experimental Procedures'' at 10°C. Under these identical conditions the MII decay t1 ⁄2 for unlabeled WT rhodopsin is 37.5 min.
nines at positions Leu 226 , Thr 229 , and Val 230 into the K141C mutant and then reacted the cysteine at site 141 with the PDT-bimane label. Characterization of these labeled mutants, K141B/L226A, K141B/T229A, and K141B/V230A indicated that they were all similar to WT in terms of expression, chromophore binding, and rates of retinal release (Table II).
When these mutants were assayed for their ability to stimulate GTP␥S exchange in the whole transducin protein, all were found to be severely impaired in comparison with K141B alone or WT rhodopsin ( Fig. 6A and Table III). However, the peptide quenching assays indicate that these mutants were all able to bind peptide W23SV, but with greatly reduced affinities. As seen in Fig. 6B and Table III, substantially more peptide was required to see the quenching effect, resulting in EC 50 values for "hydrophobic patch" mutants more than 160 times greater than that of K141B alone (200 nM versus 32 M) (Table III). DISCUSSION A growing body of evidence suggests that an outward movement of transmembrane helix 6 on the cytoplasmic face of the rhodopsin is a key step in receptor activation. This movement may be the only large scale conformational change required for receptor activation (47) and may occur in all GPCRs (15). We and others have proposed that the role of this movement may be to open up a cleft that provides a binding site for the transducin ␣ subunit C-terminal tail as well as to expose other important binding sites in rhodopsin (7, 10 -14, 17, 26, 48 -50). The major aim of the present work has been to directly test this hypothesis.
Using variants of a high affinity peptide analogue of the G T␣ C terminus, we obtained evidence that the G T ␣ C terminus FIG. 3. Peptide W23SV binding and quenching properties of PDT-bimane-labeled rhodopsin mutants. A, peptide W23SV can bind to all PDT-bimane-labeled samples as indicated by its ability to inhibit MIII photoproduct. In the absence of peptide, the PDT-bimanelabeled mutants accumulates increasing amounts of the MIII photoproduct (460 nm) during decay of the active MII species (q, black), whereas preincubation of the samples with 100 M W23SV peptide inhibits this process (OE, red). B, steadystate fluorescence emission spectra of MII PDT-bimane-labeled rhodopsin mutants (500 nM) in the absence (black) and presence (red) of 100 M peptide W23SV. Notice the decrease in fluorescence intensity for C140B and K141B and to a lesser extent F228B upon addition of peptide W23SV. C, titration of MII PDT-bimanelabeled rhodopsin mutants (500 nM) with increasing amounts of peptide W23SV. Mutants C140B, K141B, and F228B exhibit dose-dependent saturation of PDTbimane quenching, whereas no quenching is observed for the other mutants.  Table I. binds to MII rhodopsin in a manner that occludes residue V250C, consistent with the hypothesis that the G T␣ C-terminal tail is binding into the cleft formed by helix 6 movement during rhodopsin activation. Our mapping studies suggest that pep-tide W23SV binds in a position that places the N-terminal Trp residue near the bimane label on residues C140B and K141B and to a lesser extent F228B, although our results do not indicate how deeply into rhodopsin the G T ␣ C terminus penetrates upon binding. Interestingly, site V248B shows no quenching interaction, yet the dark state structures of rhodopsin show it to be in close proximity to residues Cys 140 , Lys 141 , and Phe 228 (51)(52)(53). These findings are consistent with the  a All of the samples contain a single alanine substitution generated in the K141C background and labeled specifically at the introduced cysteine residue with the PDT-bimane probe.
b Pure rhodopsin typically exhibits a protein (A 280 ) to chromophore (A 500 ) absorbance ratio of 1.6 to 1.8.
c The amount of PDT-bimane label incorporated into the rhodopsin mutants is presented as the number of labels/rhodopsin mutant (label/ sample). d The amount of free label remaining in each sample following labeling and purification is presented as a percentage of the total sample (for further details see ''Experimental Procedures'').
e The MII decay rate of the labeled samples was determined in buffer D as described under ''Experimental Procedures'' at 10°C. Under these identical conditions the MII decay t1 ⁄2 for unlabeled WT rhodopsin is 37.5 min.
FIG. 6. Mutations in a "hydrophobic patch" on rhodopsin reduce but do not abolish peptide W23SV binding affinity. A, transducin activation of PDT-bimane-labeled mutants (5 nM) in comparison with WT. Labeled mutant K141B shows near WT levels of transducin activation. In contrast, equivalent amounts of hydrophobic patch mutants K141B-L226A, K141B-T229A, and K141B-V230A all show greatly reduced abilities to activate transducin. B, monitoring the bimane quenching of these labeled mutants (500 nM) upon titration with W23SV indicates that the mutants can bind the C-terminal tail of G T␣ but do so with greatly reduced affinity. The plots show two separate titrations per mutant (OE, q), and the values are given in Table III. a All of the samples contain a single alanine substitution generated in the K141C background and labeled specifically at the introduced cysteine residue with the PDT-bimane probe.
b The relative initial rate of G T activation is represented by the rate of fluorescence increase obtained from the slope of the fluorescence measurements in the first 60 s after the addition of GTP␥S relative to that of WT rhodopsin (100%).
c Peptide quenching EC 50 values were calculated from dose-response curves generated for the combined peptide quenching data. d ⌬⌬G ‡ values were calculated by comparing the quenching EC 50 data from the alanine containing K141B mutants with that of K141B.
proposed movement of helix 6 during receptor activation, which would alter the location of site Val 248 further away from the sites that show quenching near helix 3.
In addition, our results illustrate that all of the sites labeled with bimane were able to bind the G T␣ C terminus peptide analogue (Fig. 3A). This suggests that none of the labeled residues participate in, or at least are required for, providing the direct binding interface between the receptor and G T␣ Cterminal peptide. We have chosen not to model the exact nature of this binding, because we feel our data do not yet warrant doing so. However, our results do provide an additional constraint to the rhodopsin G T␣ C terminus interaction that will facilitate future more detailed modeling studies.
Importantly, our studies show that the G T␣ C terminus binding is specific for the MII state and that the presence of the retinal Schiff-base linkage is required to maintain the exposure of the cleft required for interaction with the C-terminal tail of G T␣ . Interactions between the i3 loop of rhodopsin and the C terminus of transducin have previously been implicated in conferring efficient and specific coupling and activation (28,54). Our results show residues Leu 226 , Thr 229 , and Val 230 , which lie in this region, play a key role by imparting high affinity binding for the transducin G T␣ C-terminal tail with each of these "hydrophobic patch" residues contributing upwards of 2.6 kcal/mol to the binding energy (Table III). Note that formally our data do not clarify whether this interaction is required to maintain a structural motif in this region that enables G T␣ binding or is required for some other function, such as inducing GDP/GTP exchange. Regardless, the binding energy contributed by these residues appears critical for rhodopsin/transducin interactions, because residues Leu 226 and Val 230 are completely conserved in all type-1 vertebrate rhodopsins (www.gpcr.org/7tm/, version 8.0). Interactions between these receptor side chains with the C-terminal region of Gprotein ␣-subunits may also aid in conferring G-protein subtype specificity.
In summary, our results suggest that the role of helix 6 movement during MII formation is to provide a binding site on the cytoplasmic face of rhodopsin for the G T␣ C terminus. This movement appears to open a cleft and expose a hydrophobic patch, which directly interacts with the G T␣ C terminus and increases the affinity of transducin binding. Our results also suggest that hydrophobic interactions between rhodopsin and transducin may play a heretofore unappreciated but key role in the coupling of these two proteins, as has been suggested for other protein-protein complexes (56). Finally, we speculate that the approach we have described here may prove useful for assessing the role of other movements that occur during activation of rhodopsin and its interaction with other proteins (8,55,57,58).