Stability of dark state rhodopsin is mediated by a conserved ion pair in intradiscal loop E-2.

The rhodopsin crystal structure reveals that intradiscal loop E-2 covers the 11-cis-retinal, creating a "retinal plug." Recently, we noticed the ends of loop E-2 are linked by an ion pair between residues Arg-177 and Asp-190, near the highly conserved disulfide bond. This ion pair appears biologically significant; it is conserved in almost all vertebrate opsins and may occur in other G-protein-coupled receptors. We report here that the Arg-177/Asp-190 ion pair is critical for the folding and stability of dark state rhodopsin. We find ion pair mutants that regenerate with retinal are functionally and spectrally wild-type-like yet thermally unstable in their dark state because of rapid hydrolysis of the retinal Schiff base linkage. Surprisingly, Arrhenius analysis indicates that the activation energies for the hydrolysis process are similar between the ion pair mutants and wild-type rhodopsin. Furthermore, the ion pair mutants do not show increased reactivity toward hydroxylamine, suggesting that their instability is not caused by an increased exposure to bulk solvent. Our results indicate that the loop E-2 ion pair is important for rhodopsin stability and thus suggest that retinitis pigmentosa observed in patients with Asp-190 mutations may in part be the result of thermally unstable rhodopsin proteins.

Rhodopsin, the dim light photoreceptor of rod cells, is the best characterized member of the superfamily of G-proteincoupled receptors (GPCRs), 1 (1)(2)(3)(4)(5)(6)(7)(8). Rhodopsin consists of a chain of 348 amino acids, approximately half of which form a cluster of seven membrane-spanning helices located within the membrane (Fig. 1). The rhodopsin chromophore, 11-cisretinal, resides in the middle of these helices attached to lysine 296 through a protonated Schiff base linkage (9,10). Interactions of amino acid side chains, as well as water molecules within the chromophore-binding pocket with the retinal, result in the 500 nm absorbance maxima for dark state rhodopsin (11,12). Dim light vision begins when the 11-cis-retinal chromophore in rhodopsin absorbs a photon and is converted to all-trans-retinal. This change in retinal configuration initiates a series of photo-intermediates and conformational changes in the protein, culminating in a 380 nm absorbing species called metarhodopsin II (MII), the "active conformation" which is able to bind and activate the G-protein transducin (3,6,13).
Recently, high resolution crystal structures of rhodopsin have been obtained (2,14,15). These structures confirm some of the previous hypotheses about the rhodopsin structure, such as the general arrangement of the transmembrane helices, the locations of the disulfide bond, and glycosylation sites (16 -21). However, they also revealed several surprises. One of the most intriguing aspects was the high degree of order in the intradiscal loops (the equivalent to the extracellular loops in other GPCRs and hereby denoted as such). Especially intriguing is loop E-2, which connects helices IV and V (residues 173-198) and forms a twisted ␤-hairpin that lies alongside the retinal chromophore, potentially forming a "lid" or "plug" across the retinal-binding pocket (2,14,15) (Fig. 1). This unexpected finding has led to a number of new questions. What role does the structure of loop E-2 play in the stability and function of rhodopsin? Does it help provide a place for retinal to bind, or does retinal binding induce structure in loop E-2? If the loop E-2 structure is present in the apoprotein (opsin), how does retinal get into and out of the binding pocket?
We have recently begun to address some of these questions, and in the process noticed an ion pair, Arg-177/Asp-190, is present on the ends of loop E-2 (Fig. 1B). This ion pair, Arg-177/Asp-190, is conserved in almost all vertebrate rhodopsins 2 and may also be present in other GPCRs (Fig. 2). Furthermore, the potentially important functional role of this ion pair is suggested by the fact that mutations at residue Asp-190 in rhodopsin are found in patients with autosomal dominant retinitis pigmentosa (ADRP) (22)(23)(24)(25)(26).
In this work we report our investigations into the structural and functional role of the Arg-177/Asp-190 ion pair. Our primary finding is that the ion pair helps stabilize the dark state rhodopsin structure. We find that mutations to the ion pair either result in opsin proteins that do not regenerate with 11-cis-retinal or, if they do regenerate, undergo rapid retinal Schiff base hydrolysis in the dark state. Surprisingly, the active MII signaling state and MII decay processes are not affected by mutations at the ion pair. We also find that the ion pair mutations do not increase the susceptibility of Schiff base attack by the bulk solvent (as judged by hydroxylamine reactivity assays) nor affect the activation energy barrier for Schiff base hydrolysis. These results illustrate the importance of loop E-2 in retinal binding, rhodopsin stability, and retinal release processes.

Materials
Except where noted below, all buffers and chemicals were purchased from either Fisher or Sigma. Protease inhibitor tablets and GTP␥S were purchased from Roche Molecular Biochemicals. Dodecyl maltoside (DM) was purchased from Anatrace (Maumee, OH), and GBX red light filters were from Eastman Kodak Co. Polystyrene columns (2-ml bed volume) were purchased from Pierce. Frozen bovine retinas were from J. A. Lawson Co. (Lincoln, NE). Transducin was purified from rod outer segments as described previously (27). DNA oligonucleotides were purchased from Qiagen/Operon (Alameda, CA). Restriction endonucleases were from New England Biolabs (Beverly, MA). 11-cis-Retinal was a generous gift from Dr. R. Crouch (Medical University of South Carolina and NEI, National Institutes of Health). The rho1D4 antibody was purchased from the National Cell Culture Center (Minneapolis, MN). The nonapeptide corresponding to the C terminus of rhodopsin was acquired from the Emory University Microchemical Facility (Atlanta, GA). Cuvettes were purchased from Uvonics (Plainview, NY). Bandpass filters and long-pass filters were purchased from Oriel (Stratford, CT). The 30% acrylamide/bisacrylamide solution (37.5:1) was purchased from Bio-Rad. Goat anti-mouse (H ϩ L) conjugated with peroxidase and SuperSignal West Pico Luminol/Enhancer Solution were obtained from Pierce.

Construction and Expression of Rhodopsin Mutants
Site-directed mutagenesis was performed using a cassette-based strategy as described previously in the pMT4 plasmid (28,29), as well as overlap extension PCR (30) to generate EcoRI and NotI fragments containing either the R177C, R177K, R177Q or D190C, D190E, D190N mutations in the synthetic bovine rhodopsin gene (28). The sequences for the primers are as follows: R177C, 5ЈCGTCGGCTGGTCTTGCTAC-ATCCCGGAG3Ј; R177K, 5ЈGCTCGTCGGCTGGTCTAAGTACATCCC-GGAGGGCATGCAGTGC3Ј: R177Q, 5ЈGCTCGTCGGCTGGTCTCAGT-ACATCCCGGAGGGCATGCAGTGC3Ј; D190C, 5ЈCTCGTGCGGGATC-TGCTACTACACGCCG3Ј; D190E, 5ЈGGAGGGCATGCAGTGCTCGTG-CGGGATCGAGTACTACACGCCGC3Ј; and D190N, 5ЈGGAGGGCATG-CAGTGCTCGTGCGGGATCAACTACTACACGCCGC3Ј. Subsequent to PCR mutagenesis, the PCR fragments were subcloned into the pMT4 vector containing the synthetic gene of rhodopsin using XhoI and PstI restriction sites, and double mutants were constructed using the BsrI restriction site. Cysteine mutants were subcloned into the pMT4 plasmid theta, a synthetic gene of rhodopsin in which the potentially reactive background cysteine residues 140, 316, 322 and 323 were replaced with serines (31,32). All mutations were confirmed by the dideoxynucleotide sequencing method. The mutant rhodopsin proteins were transiently expressed in COS-1 cells using the DEAE-dextran method, and cells were harvested 56 -72 h after transfection as described previously (33,34).

Purification of Rhodopsin Mutants
Mutant rhodopsin proteins were expressed and harvested essentially as described previously (34). Briefly, five 15-cm plates of transfected COS-1 cells were washed twice with 7 ml of cold PBSSC buffer per plate, pelleted, and subsequently resuspended in 10 ml of cold PBSSC (pH 6.5) containing 0.5 mM PMSF. The opsin mutants were then regenerated with 10 M 11-cis-retinal at 4°C for 1 h, and an additional 5 M of 11-cis retinal was then added and regeneration allowed to proceed for an additional 1 h (35).
The purification of the rhodopsin mutants proceeded essentially as the original procedure (33), except small polystyrene columns were used for washes and elution (36). Cells were solubilized in 5 ml of buffer A containing 0.5 mM PMSF at 4°C for 1 h and then centrifuged to pellet the unsolubilized fraction. The supernatant was mixed with 200 l of rho1D4 antibody-Sepharose beads (binding capacity ϳ1 g of rhodopsin/g of resin) in buffer B containing 0.5 mM PMSF and nutated at 4°C for 4 -5 h. The slurry was subsequently transferred to polystyrene columns and washed once with 50 ml of buffer C followed by a 40-ml wash with buffer D by gravity filtration. Samples were eluted in 350-l (hydrogen-bound to Arg-177), as well as the 11-cis-retinal chromophore, and the disulfide bond between cysteine residues Cys-110 and Cys-187 in loop E-2. These same features are present in all three rhodopsin crystal structures (2,14,15). The model shown was prepared using coordinates from 1L9H (15) using the program WebLab.
FIG. 2. Sequence alignment of loop E-2 for rhodopsin, ␤ 2 -adrenergic, and CB1 cannabinoid receptors. Notice that the charged residues found in rhodopsin at Arg-177 and Asp-190 are mirrored in ␤ 2 -adrenergic receptor (␤AR) and cannabinoid receptor (CB1). For comparison purposes conserved proline, tryptophan, and cysteine residues are also highlighted.
fractions of buffer D containing 200 M nonapeptide corresponding to the rho1D4 antibody epitope (the last nine amino acids of the C terminus of rhodopsin). 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.

Immunoblot Analysis of Rhodopsin Mutant Cell Membranes
COS cells expressing rhodopsin mutants were pelleted and resuspended in 1 ml/plate of buffer E and homogenized on ice. The homogenates were then centrifuged at 40,000 ϫ g for 45 min at 4°C, and the pellets were washed with 5 ml of buffer F and subsequently resuspended in buffer F. Protein concentrations of the resuspended membrane pellets were determined by a modified Dc protein Assay from Bio-Rad. The manufacturer's instructions were followed except for the addition 1.45% SDS to each well. Aliquots of the membrane preparation were snap-frozen and stored at Ϫ80°C until use. SDS-PAGE was performed according to Laemmli (37), using a 5% stacking gel and a 10% resolving gel. The protein bands were electrotransferred onto Immobilon-P transfer membranes (Millipore) and detected using the rho1D4 monoclonal antibody as described previously (19). Protein expression levels were determined using a Bio-Rad PhosphorImager, and pixel densities were determined using a GS-525 molecular imaging system using supplied software.

UV-visible Absorption Spectroscopy
All 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. For concentration calculations, a molar extinction coefficient value (⑀ 500 ) for WT rhodopsin was taken to be 40 600 M Ϫ1 cm Ϫ1 (38). 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 wavelength Ͼ490-nm long-pass filter. This light treatment was found to be adequate for full conversion of all samples. Extinction coefficients were determined for each dark state mutant species as described previously in buffer D at 15°C (34,39). The presence of a protonated Schiff base (PSB) in the MII state for each mutant was verified by adding H 2 SO 4 to a pH of 1.9 immediately after photobleaching and then measuring the absorbance spectrum to assay the presence of a spectral species at 440 nm (which indicates a PSB) (40).

Thermal Bleaching of Rhodopsin Samples
Absorbance Measurements-Thermal decay rates were followed by UV-visible spectroscopy in buffer D. Specific temperatures were maintained using water-jacketed cuvette holders connected to a circulating water bath. Temperature was monitored through emersion of a digital thermometer into the sample chamber, with an accuracy of approximately Ϯ0.2°C. Thermal stability of the mutants was determined by first measuring the samples from 650 to 250 nm at 1-min intervals at a given temperature. Thermal decay rates were then measured by monitoring the decrease of the 500 nm absorbing dark state species from these measurements over time (41)(42)(43). Base-line drift was corrected for by normalizing all spectra to an absorbance of 0 at 650 nm.
Fluorescence Measurements-Thermal decay rates were also measured by monitoring the increase in tryptophan fluorescence at 330 nm, caused by the release of retinal from the chromophore-binding pocket (44). The experimental set up was similar to that of the retinal release assay (described below) except that the samples were not photobleached. All thermal decay data was analyzed using mono-exponential decay (absorbance experiments) or mono-exponential rise to maxima (fluorescence experiments) fitting algorithms in Sigma Plot (Jandel Scientific software).

Thermodynamic Calculations of Thermal Decay Rates
Activation energies (E a ) were determined by applying rate data to the Arrhenius equation: k ϭ Ae ϪEa/(RT) . Thermodynamic parameters ⌬H ‡ , ⌬G ‡ , and ⌬⌬G ‡ were calculated from the rate data as described previously (41,45). Briefly, the following thermodynamic Equations 1-3 were used, where R is the universal gas constant; T is the temperature; k 1 is the thermal decay rate, h is Planck's constant, and k B is the Boltzmann constant.

Measurement of the Rate of Retinal Release and/or MII Decay by Fluorescence Spectroscopy
The MII stability was assessed by measuring the time course of retinal release occurring after MII formation using a Photon Technologies QM-1 steady state fluorescence spectrophotometer (44). Each measurement was carried out using 100 l of a 0.25 M mutant sample in buffer D, and sample temperature was maintained as described above. After the samples were photobleached to the MII state (see above), the retinal release measurements were carried out at the appropriate temperature by exciting the sample for 3 s (excitation wavelength ϭ 295 nm, 1 ⁄4-nm bandwidth slit setting) and then blocking the excitation beam for 42 s to avoid further photobleaching the samples. Tryptophan fluorescence emission was monitored at 330 nm (12-nm bandwidth slit setting), and this cycle was repeated for up to 100 min during each measurement. To determine the t1 ⁄2 values for retinal release, experimental data were analyzed using a mono-exponential rise to maxima fit in Sigma Plot (Jandel Scientific software). In this manner a series of MII decay rates was obtained at 5, 10, 15, 20, 25, 30, and 35°C, and their rates were applied to the Arrhenius equation, k ϭ Ae ϪEa/(RT) , to determine the activation energy (E a ) of the retinal release process for each mutant rhodopsin.

Determination of Transducin (G T ) Activation Rates
Activation of G T by rhodopsin was monitored using fluorescence spectroscopy at 10°C as described previously (34, 46 -48). The excitation wavelength was 295 nm (2-nm bandwidth), and fluorescence emission was monitored 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 allowed to stir for 300 s. The reaction was then initiated with the addition of GTP␥S to a final concentration of 5 M, and the increase in fluorescence was followed for an additional 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.

Effect of R117Q Thermal Decay on Ability to Activate Transducin
Activation of G T by rhodopsin was monitored as described above. Activation assays were first performed on freshly thawed WT and R177Q stocks. The stocks were next incubated in the dark at 37°C to facilitate thermal decay of the 500 nm absorbing species, and aliquots were withdrawn at the indicated time points and assayed for G T activation.

Hydroxylamine Reactivity
Hydroxylamine reactivity of the dark state was determined for purified rhodopsins by monitoring the rate of 500 nm absorbance decrease after the addition of hydroxylamine (pH 6.0) to the samples in buffer D to a final concentration of 50 mM at the indicated temperatures (49). Base-line drift was corrected as described above (see "Thermal Bleaching of Rhodopsin Samples").

Rationale for Choice of Loop E-2 Ion Pair Mutations-Amino
acid mutations were constructed based on their ability to disrupt or potentially restore the ion pair charge interaction, while introducing minimal steric perturbation. Thus, residue Arg-177 was mutated to R177K (conserved charge) and R177Q (neutral substitution). Residue Asp-190 was mutated to D190E (conserved charge) and D190N (ADRP mutation, neutral charge). Mutants R177C and D190C were constructed to enable chemical modification of the single cysteine residues. We did not analyze other ADRP-associated point mutations at site Asp-190 (D190A, D190G, and D190Y), because previous reports suggest these mutations are defective in folding, trafficking, and/or chromophore binding (23,24,26).
Characterization of Rhodopsin Mutants-Expressed ion pair mutant rhodopsins were analyzed for expression levels, proper post-translational modifications, ability to bind 11-cis-retinal, and photobleaching properties. Immunoblot analysis of mu-tants expressed from transfected COS cells indicates all mutants expressed to similar levels comparable with that of wildtype rhodopsin (Fig. 3A). Mutants R177C, D190C, and D190E did not regenerate in our hands and were therefore not further characterized. These mutants also had abnormal glycosylation patterns in comparison to WT rhodopsin in that they did not exhibit the characteristic glycosylation smear pattern when expressed in COS cells (Fig. 3A). Furthermore, mutants defective in chromophore binding tended to form large molecular weight aggregates relative to both WT and other mutants (Fig. 3A).
Immunoblot analysis of recombinant rhodopsins purified using the rho1D4 monoclonal antibody reveal a band pattern similar to that of wild-type rhodopsin, with an apparent molecular mass of ϳ40 kDa and the characteristic heterogeneous glycosylation smear due to overexpression in a COS cell system (Fig. 3B), (50). Mutants capable of regenerating with 11-cisretinal formed characteristic rhodopsin-like pigments and could be purified to obtain spectral ratios (A 280 /A 500 ) between 1.6 and 1.8. The Arg-177 and Asp-190 single point mutants exhibited normal photobleaching behavior with respect to formation of a blue-shifted max Ϸ380 nm species (characteristic of the MII intermediate) (9). Acidification of these photobleached samples generated a max ϭ 440 nm species, indicating the presence of a protonated retinal Schiff base (PSB) (40). These results are compiled in Table I, and a representative example of the photobleaching behavior is shown for mutants R177Q and D190N (Fig. 4A). Similar to the single mutants R177Q and D190N, the R177Q/D190N double mutant shows wild-type behavior in terms of expression levels, post-translational modifications, and chromophore binding. However, it did exhibit perturbed photobleaching properties. Although capable of forming both a spectral MII species and a PSB, following illumination a residual species with a max of ϳ480 nm persisted up to 10 h after illumination (data not shown). The cause of this is not known, although similar effects have been reported for other rhodopsin point mutations such as G90S and L226C (29,34,51).
Retinal Release Rates and Activation Energies for Metarhodopsin II Decay Measured by Fluorescence Spectroscopy-To determine potential effects the ion pair mutations may have on the stability of the MII active signaling species of rhodopsin, the activation energies for retinal release were determined. The rate of retinal release occurring during the decay of the MII species was measured using a fluorescence-based assay at 20°C (44). Under the conditions used for this assay, the t1 ⁄2 of retinal release for WT rhodopsin at 20°C in buffer D was 13 Ϯ 0.5 min (n ϭ 3), comparable with the 13-15-min values reported previously (34,43,48,52) for both ROS-purified and COS-expressed rhodopsin. Somewhat unexpectedly, the corresponding t1 ⁄2 values for the ion pair mutant rhodopsins were similar to that of WT rhodopsin. The values for each of the mutants are compiled in Table I. The activation energy for the metarhodopsin II decay process was obtained by monitoring the rate of fluorescence increase in buffer D at seven different temperatures (5,10,15,20,25,30, and 35°C). The rate of fluorescence increase in all cases was temperaturedependent, and Arrhenius plots of these measurements indicated a temperature-dependent linear relationship for all mutants (Fig. 4B). From these plots an activation energy (E a ) of 20.2 kcal/mol was obtained for purified WT rhodopsin in DM, in good agreement with values reported previously (34,44,53). Arrhenius plots of the retinal release rates for the ion pair mutants show nearly equal E a values ( Fig. 4B and Table I).
Transducin Activation by Ion Pair Mutants-To assess the potential functional effects, the ion pair mutants were tested for their ability to activate transducin using a fluorescencebased assay that measures the increase in tryptophan fluorescence of the G T␣ -GTP␥S species (46, 48, 54). All ion pair mutations that regenerated with retinal are functionally active, and representative examples are presented in Fig. 4C. The results for transducin activation are compiled in Table I as initial rates of fluorescence increase relative to WT rhodopsin.
Thermal Stability in the Dark State-The most dramatic perturbation induced by the ion pair mutations was on the stability of the dark state structure. Thermal stabilities of dark state WT and ion pair mutant rhodopsins were determined by measuring the loss of the 500 nm absorbing species over time as described under "Experimental Procedures." An example of this assay is depicted for mutant R177Q at 37°C in Fig. 5A. The loss of the 500 nm species directly correlates with a loss of ability to activate transducin (Fig. 5, B and C). Additionally, the decrease in absorbance at 500 nm reflects a loss of the chromophore Schiff base linkage as judged by decay of the acid-denatured 440 nm species over the duration of the thermal decay assay (Fig. 5, D and E). Furthermore, we conclude the retinal is leaving the chromophore binding pocket after the hydrolysis because the rate of the loss of the 500 nm absorbing  (19). B, immunoblot of purified ion pair rhodopsin mutants. Purified recombinant rhodopsin mutants were prepared as described under "Experimental Procedures" and probed using the rho1D4 antibody. ROS purified rhodopsin was included as a control. Notice that purification removes the lower weight species. species correlates with the rate of tryptophan fluorescence increase, and irradiation of the sample with light following a plateau in signal does not cause a further fluorescence increase (Fig. 5F) (44).
All of the ion pair mutants showed significantly expedited rates of thermal decay in comparison to WT rhodopsin as judged by their loss in absorbance at 500 nm and increase in fluorescence at 330 nm (Table II). A comparison of the thermal decay rates at 55°C monitored by absorbance is shown in Fig. 6. Note that the thermal decay rate of ROS-purified rhodopsin was similar to that of WT recombinant rhodopsin purified from COS cells (38.5 Ϯ 3.0 and 37 min at 55°C,  Table I. B, Arrhenius plots of retinal release rates from the MII state of ion pair mutant rhodopsins. The rate constants were obtained from the retinal release assays (see "Experimental Procedures") performed in buffer D at pH 6.0, with temperatures ranging from 5 to 35°C. For comparison, values obtained for wild-type rhodopsin are overlaid in black open circles fit with black lines. The activation energy (E a ) values determined from these assays are reported in Table I. C, example of transducin activation by ion pair mutants. Transducin activation was measured by monitoring the increase in G T␣ tryptophan fluorescence that occurs upon MII stimulation of the G T␣ -GTP␥S complex formation. The arrow indicates the time of GTP␥S addition. The relative rates of transducin activation of the mutants compared with WT rhodopsin are reported in Table I. respectively). The activation energies for the thermal absorbance decay processes were determined by monitoring the loss of the 500 nm absorbing species over time at 7 different temperatures (37, 41, 45, 47.5, 50, 52.5, and 55°C). In all cases, the rate of loss in 500 nm absorbance was temperature-dependent, and Arrhenius plots indicate a similar temperature-dependent relationship for all mutants (Fig. 7). The Arrhenius plots are clearly concave, suggesting at least two different rate-limiting processes may occur during the temperature-dependent absorbance decay. With this in mind,  b Activation energies (E a ) and thermodynamic parameters (⌬G ‡ , ⌬⌬G ‡ , ⌬H ‡ ) were obtained from linear regression of Arrhenius plots (Fig. 7), for further details see "Discussion." c ND, not determined. d Increase in tryptophan fluorescence was monitored once at 330 nm to monitor dark state thermal decay rates.
two linear regressions were used to approximate the activation energies for the two apparent processes (55-47.5 and 47.5-37°C, respectively). From this analysis, the E a for WT rhodopsin was determined to be ϳ16 kcal/mol at 37°C and 103 kcal/mol at 55°C. The thermodynamic parameters E a , ⌬G ‡ , ⌬H ‡ , and ⌬⌬G ‡ were estimated from the rate data for WT and mutant rhodopsins using equations described previously (see Table II) (41,45). Hydroxylamine Reactivity-Hydroxylamine reactivity experiments showed that the ion pair mutants were not more susceptible to hydroxylamine in the dark state. These assays were carried out for purified ROS rhodopsin and each mutant sample, and the decay of the dark state 500 nm absorbing species was monitored in buffer D at either 20, 37, or 55°C over time following the addition of hydroxylamine (pH 6.0) to a final concentration of 50 mM. WT rhodopsin purified from retinal sources and from expressed COS cells was found to be inert to hydroxylamine in the dark state, as described previously (49).
Intriguingly, none of the ion pair mutants exhibited any increased reactivity toward hydroxylamine treatment in the dark state at 20, 37, and 50°C (Fig. 8).  Table II. FIG. 7. Arrhenius plots of dark state thermal decay rates show ion pair mutants have similar activation energies for retinal hydrolysis yet faster rates than wild-type rhodopsin. The rate constants were obtained from the dark state thermal absorbance decay experiments (see Fig. 6), performed in buffer D, with the temperatures ranging from 37 to 55°C (WT, filled circles; R177K, filled diamonds; R177Q, open squares; D190N, open triangles). The concave plot suggests two different processes may lead to thermal decay. The E a and thermodynamic values approximated for these processes are presented in Table II.

DISCUSSION
Early studies by Khorana and others (50,[55][56][57][58][59] led to the hypothesis that the rhodopsin intradiscal domain plays a crucial role in maintaining proper protein folding, correct posttranslational modifications, trafficking, and 11-cis-retinal binding. Consistent with this theory, a number of point mutations that naturally occur in this region result in ADRP, an inherited human disease causing retina degradation (23, 24, 26, 60 -62). As noted in the Introduction, the rhodopsin crystal structures reveal a high degree of order and structure in the intradiscal region. This region of the protein is proposed to be structurally critical for maintaining the electrostatic and hydrogen-bonded network surrounding the retinal chromophore (15). Most notably, loop E-2 forms a twisted ␤-sheet which lies across the retinal chromophore (2,14,15). Through analysis of the loop E-2 region we noticed an ion pair Arg-177/Asp-190 is present on either end of this loop structure. Additionally, we noticed that residue Arg-177 is hydrogen-bonded to the backbone carbonyl of residue P7, a residue found at the turn of loop E-1 in the N terminus of the proteins. The fact that residues Arg-177 and Asp-190 interact with many residues within the intradiscal region of rhodopsin suggests that the ion pair may play a significant role in maintaining the structural integrity of the "retinal plug" domain. The published rhodopsin crystal structures show very little difference in the region around the Arg-177/Asp-190 ion pair (analysis of all three structures using the program Swiss-PDB Viewer shows that the 87 amino acid side chains within 16 Å of Arg-177 show a root mean square deviation of 1.0 Å or less (2,14,15)). However, we do notice a difference in the placement of residue Asn-200, which exhibits a different rotameric flip between structures 1HZX, 1L9H, and 1F88, and thus exhibits alternate hydrogen bonding to residue Asp-190 in the different structures (2,14,15). The present report details our studies on the structural and functional effects caused by disrupting the Arg-177/ Asp-190 intradiscal ion pair located on either end of loop E-2 (Fig. 1B).
General Characteristics of Mutants-The majority of the single ion pair mutants we created, expressed to similar levels comparable with WT rhodopsin, underwent proper glycosylation (as judged by whole cell lysate immunoblotting, Fig. 3) and regenerated with 11-cis-retinal. Additionally, with the exceptions of mutants R177C, D190C, and D190E, the mutants were properly folded, as judged by their ability to bind the 11-cisretinal chromophore, and produced a wild type like A 280 /A 500 ratio ( Table I). The fact that most of the Asp-190 mutations were unable to bind retinal is in agreement with previous reports of other Asp-190 mutations, which also were found to be defective in retinal binding (23,26,58). One possible reason for the sensitivity of this site to mutations may be that residue Asp-190 is partially buried and makes contacts with several residues (Ile-189 and Tyr-Y191), which form part of the retinal binding pocket (Fig. 1B) (2,15,52). Abrogation of these contacts by mutations to residue Asp-190 may thus distort the retinal-binding pocket thereby making it inaccessible or sterically unfavorable for proper binding of 11-cis-retinal. Additionally, it is also possible that the R177C and D190C mutants are not able to regenerate with 11-cis-retinal because improper disulfide bonds are formed in the final folded structure, as these residues are in close proximity to the conserved Cys-110/ Cys-187 disulfide pair, as well as residue Cys-185 (Fig. 1B). This explanation is supported by recent findings, which show that improper disulfide bonds form as a result of mutations to this region (63)(64)(65). Furthermore, the fact that the glycosylation patterns of whole cell lysate immunoblots of mutants R177C, D190C, and D190E are different suggests the misfold-ing may have already occurred by the time the protein reached the endoplasmic reticulum, and thus the lack of regeneration is not simply due to an ultra-fast rate of retinal hydrolysis after regeneration (Fig. 3A). These three mutants also tend to form high molecular weight aggregates when analyzed on SDS gels, and such aggregation has been suggested to result in proteins that eventually become degraded in the endoplasmic reticulum rather than proceeding through the Golgi to the plasma membrane (24,66).
The spectral behaviors for the mutants that did regenerate with 11-cis-retinal were WT-like in their ability to form an MII absorbing species and a PSB upon acid denaturation (Fig. 4A). One exception was the double mutant R177Q/D190N, although WT-like in its MII and PSB characteristics, which demonstrated a residual absorbing species with a max of ϳ480 nm following illumination (data not shown). This ϳ480 nm absorbing species may be an MII-like intermediate containing a PSB, as was suggested for mutations at site Gly-90 (34,67,68), or it may be an MIII intermediate, due to an altered equilibrium between the MII and MIII states (69,70). However, why the double mutation shows this characteristic and the single mutants do not is unknown.
The Arg-177/Asp-190 Ion Pair Is Not Required for MII Formation, Signaling, or Stability-We were surprised to find that the Arg-177/Asp-190 ion pair does not appear critical for the formation or maintaining the stability of the MII state. All ion pair mutants show retinal release rates similar to that of WT rhodopsin in buffer D at 20°C (13 Ϯ 0.5 min, Table I), and activation energies (E a ) for retinal release from MII are almost identical to the 20.2 kcal/mol obtained for WT rhodopsin ( Fig.  4B and Table I) (34,44,53). These results are in contrast to our previous findings on mutations within the chromophore-binding pocket, which increased the retinal release rate from MII (although the activation energy for retinal release were unaffected (34)). Notably, disruption of the Cys-110/Cys-187 disulfide bond in the intradiscal region of rhodopsin dramatically disrupts MII decay and stability (41). In contrast, the regenerated ion pair mutants all show normal MII function and decay, suggesting they must still contain a normal Cys-110/Cys-187 disulfide bond. Furthermore, the ion pair mutants are functionally active, as they could activate the G-protein transducin, although not quite to the full extent of WT rhodopsin ( Fig. 4C and Table I). Taken together, these results suggest that the intradiscal ion pair is not critical for the formation or stability of the active signaling MII state.
Arg-177/Asp-190 Ion Pair Helps Stabilize the Dark State Rhodopsin Structure-The most dramatic effect exhibited by the ion pair mutants was a sharp decrease in the stability of their dark state structures in comparison to WT rhodopsin. Recent studies (43,71,72) have begun to address carefully the thermodynamics of rhodopsin protein stability. In the present work, our conclusions about rhodopsin thermal stability are primarily based on monitoring the stability of the retinal Schiff base linkage, which we measure as the loss of absorbance at 500 nm. We feel confident that these measurements report on Schiff base hydrolysis and release of retinal from the binding pocket, for the following reasons. The loss in absorbance at 500 nm correlates with both the loss of the PSB (as observed by a decrease in the acid-denatured 440 nm species over time) as well as the increase in tryptophan fluorescence at 330 nm ( Fig.  5 and Table II). Furthermore, the loss of absorbance at 500 nm also correlates with the loss in ability of the protein to activate transducin (Fig. 5C). Taken together these results indicate that during the thermal decay of the dark state structure the Schiff base is hydrolyzed and the retinal leaves the chromophorebinding pocket. of interactions that occur within the protein to provide receptor stability and allow receptor activation and attenuation.