Modulation of Opsin Apoprotein Activity by Retinal

The bovine opsin apoprotein activates transducin, although at a much reduced level than light-activated rhodopsin (Surya, A., Foster, K., and Knox, B. (1995) J. Biol. Chem. 270, 5024–5031). The ability of retinal to modulate opsin apoprotein activity was investigated using a guanyl nucleotide exchange assay on transducin. 11-cis-Retinal reacted with opsin at 22 °C to (a) reform pigment having maximal absorbance at 500 nm and (b) reduce opsin activity by >80%. Pigment formation also occurred at 0 °C with a t½ of 260 min. However, unlike rhodopsin formed at 22 °C (R22), the rhodopsin formed at 0 °C (R0) activated transducin with the same half-saturating concentration as opsin in an exhaustive binding assay. Thus, the formation of a protonated Schiff base associated with 500 nm absorbance does not by itself lead to the inactivation of opsin. The R0 conformation was partially inactivated by incubation at 22 °C (t½ = 61 ± 9 min), suggesting that it may be an intermediate conformation in the regeneration of rhodopsin.

The vertebrate rod cell photoreceptor, rhodopsin, is a member of a larger class of G-protein-coupled transmembrane receptors. The rhodopsin holoprotein consists of the apoprotein, opsin, covalently bound to a 11-cis-retinal chromophore by a protonated Schiff base (1). The sole action of light is to isomerize retinal from a cis to trans configuration, thereby initiating a series of conformational changes in rhodopsin that modulate its affinity for the rod cell G-protein, transducin (for reviews, see Refs. 2 and 3). While the dark-adapted rhodopsin molecule has negligible affinity for transducin, metarhodopsin(II), a conformation formed within milliseconds of light absorption, is a highly active form of the receptor that rapidly binds and catalyzes nucleotide exchange on hundreds of transducin molecules (4). Metarhodopsin(II) decays to the opsin apoprotein and free all-trans-retinal.
Opsin is able to activate transducin and is a form of the receptor that has low activity (5). In contrast to the high turnover number and rapid kinetics, of transducin activation by metarhodopsin(II), opsin activity is characterized by a low turnover number and slower kinetics, which results in a second-order rate constant for transducin activation by opsin that is approximately 30-fold lower than that for metarhodopsin(II) (5). Furthermore, while metarhodopsin(II)-transducin com-plexes are readily isolated in the absence of GTP, similar conditions fail to yield opsin-transducin complexes suggesting that the interaction is of low-affinity (5)(6)(7). There are also significant differences in the pH profile of transducin activation by metarhodopsin(II) and opsin. Unlike metarhodopsin(II), opsin is unable to activate transducin at pH Ͼ 9 (5), and it is possible that the high affinity that metarhodopsin(II) has for transducin arises from the contributions of basic side chains that are unavailable in opsin.
The apparent low affinity and slow kinetics of transducin activation by opsin rule out a significant role in the rod cell's rapid ON response to light, which is almost entirely mediated by metarhodopsin(II). However, the ability of the apoprotein to activate transducin is a potential source of phototransduction noise and may provide a means by which the cell adapts to varying light intensities, as has been observed in bleaching adaptation (8 -14). Conversely, the need for a mechanism to reduce apoprotein noise and increase the sensitivity of the rod cell during low levels of ambient light is clear (15).
One such mechanism is the regeneration of dark-adapted rhodopsin from opsin and 11-cis-retinal, a process which results in the reduction in opsin's ability to activate transducin (5) and a dramatic reduction in cell sensitivity (11,14). However, the mechanism by which retinal alters the apparent affinity of opsin for transducin is not known. The retinal binding site is located within the hydrophobic core of the protein while the transducin activation site lies on the cytoplasmic face (e.g. Refs. 16 and 17). While the binding of retinal leads to the formation of a protonated Schiff base between opsin and 11cis-retinal and a large shift in the absorption maximum of the chromophore, it is not known whether Schiff base formation plays a direct role in altering the affinity of opsin for transducin. In this report, we have further characterized the effect of 11-cis-retinal on opsin activity to understand the mechanism by which it reduces the affinity of opsin for transducin. We have determined that the inhibitory effect of 11-cis-retinal is not solely due to the formation of a protonated Schiff base between retinal and opsin. Rather, it appears that a conformational change, presumably at the transducin activation site, is essential for the deactivation of opsin.

EXPERIMENTAL PROCEDURES
Protein Preparations and Assays-Rhodopsin-and opsin-containing rod outer segment membranes were prepared from frozen dark-adapted bovine retinae as described earlier (5). Bovine transducin was purified according to Ref. 18, and transducin activation was measured by a guanyl nucleotide exchange assay using a non-hydrolyzable GTP analog, [ 35 S]GTP␥S 1 (5). The reactions were carried out at 22°C in dim red light for 2 h in 10 mM Tris acetate (pH 7.0), containing 100 mM NaCl, 5 mM MgCl 2 , and 5 mM 2-mercaptoethanol.
Pigment Formation with 11-cis-Retinal-All procedures were carried out under dim red light. 11-cis-Retinal (19) used in the reactions was stored as an ethanolic stock solution, and the volume was adjusted such that the concentration of ethanol during regeneration did not exceed 2%. Regenerations were carried out in 10 mM sodium phosphate (pH 6.5), at room temperature or on ice. Argon was blown into the reaction tubes, which were then incubated at the desired temperature on a nutator. Aliquots from the regenerated material were solubilized in 1% dodecyl maltoside at 0°C for 45 min, and then centrifuged at 50,000 ϫ g at 0°C for 30 min to remove any insoluble material. In kinetic experiments regeneration was stopped by the addition of hydroxylamine prior to solubilization. Hydroxylamine treatment did not change the measured extent of regeneration. Absorption spectra were recorded at 0 -4°C using a Beckmann DU-640 single beam spectrophotometer fitted with a water-jacketed cuvette holder connected to a circulating water bath. The samples were bleached for 2 min at 0°C with light from a 300-watt projector ( Ͼ 535 nm), and dark-light difference spectra were obtained. The percent protonated Schiff base formed at each time point was calculated as follows. The 280 nm absorbance of unregenerated opsin was divided by 500 nm absorbance from the difference spectra to yield a A 280 /A 500 ratio. This ratio was divided by the A 280 /A 500 ratio of the urea-stripped rhodopsin from which the opsin was prepared and expressed as a percentage.
In experiments in which complete Schiff base formation was desired, opsin was incubated with 10-fold excess 11-cis-retinal for a minimum of 12 h. Pigment formed in the presence of excess 11-cis-retinal at 0°C is henceforth called R 0 , while that formed at 22°C will be called R 22 . In an individual regeneration experiment the percent regeneration could be determined to within 5%. The extent of regeneration varied from 80 to 90% at 0°C and from 85 to 95% at 22°C.
In experiments in which the regeneration temperature was varied, 11-cis-retinal was added to opsin on ice, and the mixture rapidly transferred to the desired temperature and incubated overnight. In each experiment, transducin assays were performed at 22°C on four samples: opsin, R 0 , R 22 , and the sample regenerated at the test temperature. Duplicate measurements of transducin activation were normalized to the opsin value, and the mean was plotted as a percentage.

RESULTS
Regeneration at Low Temperature-To examine the inhibition of opsin activity by 11-cis-retinal, the reaction was slowed by decreasing the regeneration temperature. This permitted us to determine if the inhibition occurred in one or more steps. We chose 0°C for the lowest temperature, to avoid the addition of glycerol or other solvents, and 22°C as our standard temperature. Although regeneration of bovine rhodopsin from opsin and 11-cis-retinal has been extensively studied, there are only a few examples of regeneration at low temperatures (20,21). Furthermore, it was necessary to characterize our opsin preparation which was essentially free of other rod cell proteins and retinoids by extensive washing with urea, hydroxylamine, and bovine serum albumin (5).
Opsin in rod outer segment membranes was reconstituted with 10-fold excess 11-cis-retinal at 0 or 22°C, and Schiff base formation was determined by dark-light difference spectra, after quenching the reaction at various times with hydroxylamine. Schiff base formation occurred at 0°C with a half-time of approximately 5 h and reached its final extent of Ͼ85% by 14 h (Fig. 1A). In contrast, regeneration at 22°C occurred in minutes (Ref. 22, also see Fig. 5). The extent of regeneration varied between 90 and 95% at 22°C, and 85 and 90% at 0°C as determined by difference spectra (Fig. 1B). The small difference in the extent was reproducible and could be due to small differences in the extinction coefficient of the two species of rhodopsin or could represent ϳ5% of unregenerated opsin at 0°C. The dark-adapted rhodopsin conformations formed at 0°C and 22°C are termed R 0 and R 22 , respectively.
Schiff Base Formation Is Not Sufficient to Inhibit Opsin Activity-The ability of rhodopsin regenerated at 0 and 22°C to activate transducin in the dark was assayed by measuring the stimulation of [ 35 S]GTP␥S exchange on purified transducin in exhaustive binding assays (4,5,23). When opsin was incubated extensively with varying amounts of 11-cis-retinal at 22°C, the increase in the amount of bound 11-cis-retinal (Fig. 2, left panel, open circles) was accompanied by a corresponding decrease in the original activity (Fig. 2, left panel, closed circles). The half-maximal retinal concentrations for these two processes were 2.7 and 3.9 M, respectively, approximately stoichiometric to the opsin concentration. In contrast, a dramatic difference was observed when opsin was reconstituted at 0°C. While a protonated Schiff base was formed between opsin and 11-cis-retinal (Fig. 2, right panel, open circles), only a minor inhibition (Ͻ20%) of opsin activity was observed even at the highest 11-cis-retinal concentration tested (Fig. 2, right panel, closed circles). Thus, the transducin activation site on opsin was not perturbed by the covalent binding of retinal in a protonated Schiff base at 0°C.
A complication arose from extended incubations at 22°C. Typically 30 -40% of the opsin activity was lost during overnight incubations of opsin at this temperature (Fig. 3, open and the extent of protonated Schiff base formation (A 500 ) was estimated from dark minus light difference spectra (). The time required for 50% spectral regeneration was 3.7 h, as determined from a single parameter exponential fit (solid line). The extent of regeneration at 0°C was comparable to that at 22°C (triangle). B, difference spectra of R 0 and R 22 . Opsin and 11-cis-retinal were mixed together at 22 or 0°C as described in A and the reaction allowed to proceed for 20 h. Samples were solubilized with 1% dodecyl maltoside, and spectra were recorded both in the dark and following a 5-min bleach with Ͼ 535 nm light and dark minus light difference spectra were obtained. The positive peak with a 500-nm maximum indicates the formation of a protonated Schiff base, while the negative peak with a 360-nm minimum indicates the formation of retinaloxime from bleached rhodopsin. bars). However, the inhibition of opsin activity due to incubation with 11-cis-retinal was significantly greater (Fig. 3, third solid bar). When R 22 activity was normalized to the activity of opsin samples incubated at 22°C, the ratio ranged from 0 to 40% (Fig. 3, hatched bars). Thus, the extent of inhibition caused by 11-cis-retinal was always much greater than that due to temperature effects on the apoprotein. Two further experiments show that the inhibition of activity caused by 11-cis-retinal occurs by a different mechanism than the temperature-dependent decay of opsin activity (e.g. as a result of denaturation of the apoprotein, Ref. 24). First, the rate of inhibition of opsin activity at 22°C by 11-cis-retinal (Fig. 4A, closed circles) occurred significantly faster than the thermal decay of opsin activity (Fig. 4A, open circles). Furthermore, the rate of inhibition of opsin activity by 11-cis-retinal (Fig. 4A, closed circles) was indistinguishable from the rate of Schiff base formation at 22°C (Fig. 4B, open triangles), and did not require extended incubation. The loss of activity with regeneration is in contrast to the known stabilizing effect that 11-cisretinal has on the apoprotein (25). In fact, incubation at 22°C for several days did not alter rhodopsin's ability to activate transducin in the light (data not shown). In summary, the mechanism by which 11-cis-retinal reduces opsin's affinity for transducin is distinct from the temperature-dependent decay of opsin activity.
Activity at pH Ͼ 9 -One of the key features of opsin activity is its dependence on pH. Opsin shows activity over a broad pH range (5.8 -7.4) but is inactive at pH Ͼ 9 unlike metarhodopsin-(II) (5). To determine whether the activity of the R 0 conformation was opsin-like at high pH, R 0 was assayed with transducin at pH 9.5 (Table I). Opsin exhibited no activity at pH 9.5. In the dark, R 0 also failed to activate transducin, suggesting similarities in the interaction of transducin with both opsin and R 0 ( Table I, A). As expected, R 22 was also inactive in the dark at pH 9.5. In contrast, both R 0 and R 22 activated transducin to the same extent at pH 9.5 when exposed to light, suggesting that these two different conformations form metarhodopsin(II) ( Table  I, B). In fact, the similarity in their activity suggests that there is no temperature-driven decay of activity in the R 22 sample. R 0 Is Inactivated at 22°C-To examine the stability of R 0 , incubations at 22 or 0°C before assay with transducin were performed (Fig. 3, solid bars). Open bars, overnight incubation of opsin at 22°C decreased opsin activity in transducin assays by approximately 40%. In contrast, regeneration with 11-cis-retinal at 22°C decreased opsin activity by greater than 90% (see third solid bar). Hatched bars, R 22 and R 0 activity normalized to activity from control opsin samples incubated overnight at 22 and 0°C, respectively, showing that 11-cis-retinal causes a major decrease in apoprotein activity at 22°C. Solid bars, R 0 and R 22 were preincubated at 22 and 0°C, respectively, for 2 h before assay with transducin. While preincubation at 22°C for 2 h decreased R 0 activity by greater than 50%, cooling R 22 to 0°C did not alter its activity.
22°C resulted in a 56 Ϯ 7% drop in its activity. In contrast, preincubation of R 22 at 0°C for 2 h did not result in any appreciable change in transducin activation. Since the amount of transducin activated is proportional to the concentration of the active species (R 0 ), the change in the amount of transducin activated as a function of preincubation time was used to determine the rate of inactivation of R 0 at 22°C. This rate of inactivation of R 0 was slow, with a t1 ⁄2 ϭ 61 Ϯ 9 min (Fig. 5). Thus R 0 may represent a reaction intermediate in the formation of dark-adapted rhodopsin from opsin and 11-cis-retinal, stabilized at 0°C.
Inhibition as a Function of Regeneration Temperature-To determine the effect of regeneration temperature on the extent of inhibition of opsin activity, the activity of dark rhodopsin was measured after regeneration to completion with excess 11-cis-retinal, at a range of temperatures between 0 and 22°C (Fig. 6). The percent regeneration, as estimated from difference spectra, was greater than 85% at each of the temperatures tested (data not shown). Regeneration at 0°C preserved the most apoprotein activity, while higher regeneration temperatures showed more inactivation. The extended times for regeneration and assay precluded higher temperature resolution. Another technique will be required to further resolve this transition. Taken together with the earlier experiments, these results suggest that the active and inactive states of the transducin activation domain on dark-adapted rhodopsin are separated by a thermal barrier. DISCUSSION To further our understanding of the structural requirements for transducin activation, we have used a guanyl nucleotide exchange assay with a non-hydrolyzable GTP analog (4,5,23,26) to measure the activation of transducin. This assay is able to detect activity that arises from low affinity interactions between the receptor and transducin since it measures the cumulative number of activated transducin molecules. This is advantageous when compared with assays such as light-scattering, which require tight association between the receptor and the G-protein, and initial rate assays, which require higher sensitivity. Using this approach, the relatively weak activation of transducin by the bovine opsin apoprotein was observed, by measuring both the rate of activation and a titration of the  a Activity was measured in exhaustive binding assays using 50 pmol of transducin and 500 pmol of [ 35 S]GTP␥S in a 250-l volume. The numbers represent the picomoles of transducin activated.
b Assays were carried out in dim red light using 20 pmol of unregenerated opsin (control), R 0 or R 22 .
c Assays were carried out under illumination with Ͼ 535 nm light from a 300-watt projector, using 4 pmol of unregenerated opsin (control), R 0 or R 22 . number of transducin molecules activated during the lifetime of the active receptor (5). Both measures reflect signal transduction properties of the apoprotein. To study the relative effect of various chemical and physical constraints on this activity, experiments were carried out in the presence of excess transducin and GTP␥S, using concentrations of opsin in the linear range as determined from the titration curve.
To investigate the role of ligand binding on opsin activity, we have characterized the regeneration of dark-adapted rhodopsin from opsin and 11-cis-retinal, and have identified a temperature-sensitive reaction intermediate in the regeneration process. R 0 , which can be trapped by carrying out the regeneration reaction at 0°C, has activity similar to that of the apoprotein and activates transducin despite having a protonated Schiff base. R 0 is thermally unstable and decays at room temperature to inactive dark-adapted rhodopsin, suggesting that in addition to Schiff base formation, a temperature-dependent conformational change, presumably at the transducin activation site, is required to reset the protein to its inactive state. Our findings are in qualitative agreement with those of Okada et al. (21), who showed that the post-metarhodopsin(II) activity in their preparations was only inhibited by the addition of 11-cis-retinal at temperatures above 13°C. We note that while the transducin binding assay is a powerful probe of conformational change, limitations in thermal resolution of conformational change are inherent due to the lengthy regeneration and assay periods.
The regeneration of bovine opsin has been studied extensively, and detailed information is available on the steric requirements for retinal binding to the hydrophobic domain of the protein (27,28). Competition studies have shown that the binding of 11-cis retinal to opsin is a two-step process with hydrophobic recognition preceding protonated Schiff base formation (29 -31). Although the two stages of retinal binding to opsin are distinct, even the slower Schiff base formation is rapid at room temperature, in contrast to the findings at lower temperatures, where reported rates of Schiff base formation in digitonin are significantly lower (20).
While the 11-cis-retinal-binding site is buried within the hydrophobic core of the protein (32), the transducin activation site lies on the cytoplasmic face of rhodopsin (2,16,17,26). Thus to alter the affinity of opsin for transducin, the binding of retinal should induce a conformational change at the transducin activation site. These changes have been detected by proteolysis experiments (33), cysteine labeling (34), and EPR studies (35,36).
A key consequence of retinal binding is the formation of a Schiff base, whose protonation state correlates strongly with the active conformation of the protein. In inactive dark-adapted rhodopsin ( max ϭ 500 nm), the Schiff base is protonated and early photointermediates of light-activated rhodopsin have altered spectral properties, but the Schiff base remains protonated until the metarhodopsin(I)-metarhodopsin(II) transition has occurred. The early intermediates, including metarhodopsin(I) ( max ϭ 470 nm), are not able to activate transducin. Moreover, it has been shown that deprotonation of the Schiff base is obligatory for the formation of the high affinity transducin activation site in metarhodopsin(II) ( max ϭ 380 nm) (37). An additional step required for the formation of the active conformation has been suggested to involve proton uptake by Glu-134 at the cytoplasmic face of the protein (38).
Just as the activation of rhodopsin appears to involve Schiff base deprotonation and conformational change at the transducin activation site, complete inactivation of the protein requires the formation of a protonated Schiff base and conformational change at the transducin activation site. This conformational change requires thermal energy and can be decoupled from protonated Schiff base formation by simply lowering the regeneration temperature. Recently, Arnis and Hofmann (39) have shown that the flash-induced conversion of metarhodopsin(II) to rhodopsin takes place in discrete steps that involve the protonation of the Schiff base and a spectrally silent proton transfer from the protein to the surrounding solvent. They also demonstrated that the different conformations of dark rhodopsin interacted differently with transducin. The experiments presented in the present study demonstrate thermally distinct states of dark-adapted rhodopsin that have different affinities for transducin. We speculate that conformations identified in Ref. 39 are similar to those identified here and note that the existence of thermal intermediates representing different conformations of the protein is reminiscent of the thermal photointermediates that are observed on the bleaching of rhodopsin (40,41).
The recombination reaction between 11-cis-retinal and the opsin apoprotein provides an opportunity to independently monitor Schiff base formation and transducin activation. It was shown earlier that room temperature recombination resulted in both protonated Schiff base formation and a significant decrease in opsin activity (5). It has now been shown that the rate of Schiff base formation and the rate of conformational change at the cytoplasmic face are similar when the recombination reaction is carried out at room temperature. However, when recombination is carried out at 0°C, protonated Schiff base formation occurs without significant change in transducin activation. Furthermore, raising the temperature of opsin reconstituted at 0°C (R 0 ) results in a lowering of the activity, suggesting that R 0 is a thermal intermediate in the recombination reaction between opsin and 11-cis-retinal. Finally, the slow rate of decay of R 0 to R 22 explains why R 0 activity is observed in room temperature assays.
The ability of the apoprotein to activate transducin in a light-independent manner is a potential source of transduction noise in the rod cell. There is physiological evidence that the generation of noise could be a means of desensitizing the rod cell to adapt to higher intensities of light (42-45). Lamb and co-workers (45) showed that two processes occur in the rod cell immediately after a bleach. There is an increase in the basal current which is superimposed with photon-like quantum noise events. They also proposed that both these events originated at the level of the opsin molecule. We propose that the post-flash increase in basal current observed is caused by an accumulation of the active opsin apoprotein and that the quantum noise events are caused by the increased turnover of transducin when opsin recombines with the released all-trans-retinal. 2 However, the rod cell reverts to its low noise state during periods of extended darkness (15,44,45). This decrease in noise coincides with the increased transport of 11-cis-retinal from the retinal pigment epithelium into the photoreceptor layer of the retina in the dark (46). In fact, 11-cis-retinal has been shown to increase the sensitivity of isolated salamander rod (11,12) and cone cells (14) that have been bleached. Thus, physiologically, the reformation of rhodopsin from opsin and 11-cis-retinal leads to a decrease in photoreceptor noise and consequent increase in sensitivity which is in agreement with in vitro biochemical data from this study that show that 11-cisretinal causes a dramatic reduction in transducin activation by opsin.
The magnitude of opsin activity required to initiate the biochemical events that underlie adaptation have been made from extrapolations of physiological data (13,47). These estimates are significantly lower than the activity of opsin measured in our in vitro assays. However, physiological measurements do not exclude the effect of quenching mechanisms, such as phosphorylation and arrestin binding, that could effect opsin and metarhodopsin(II) in different ways. In fact, the relatively high opsin activity that is measured in vitro may provide an additional reason for the shut off of the visual cascade by specialized proteins such as rhodopsin kinase and arrestin in vivo. Biochemical estimates of opsin activity vary greatly, most likely due to the use of opsin in non-native membranes (e.g. from transfected COS1 cells (48)) or at non-optimal, non-physiological conditions (e.g. pH 8 (49)). Experiments reported here indicate a much more active apoprotein, requiring several inactivation mechanisms.
In contrast to the inhibitory effect of 11-cis-retinal, all-transretinal enhances opsin's ability to activate transducin (5, 47, 49 -53) 2 and its ability to act as a substrate for rhodopsin kinase as previously reported (54). The enhanced activation of opsin by all-trans-retinal is another mechanism that could cause the observed post-flash noise in the rod cell (44,45). Thus, in the rod cell, the rate of cessation of noise probably depends on both the rate of reformation of rhodopsin from opsin and 11-cis-retinal and the rate of reduction of all-trans-retinal to all-trans-retinol. The ability of 11-cis-and all-trans-retinal to combine with opsin and alter the sensitivity of the photoreceptors may provide the rationale for the rapid reduction of alltrans-retinal to all-trans-retinol by retinol dehydrogenase and the subsequent transport of retinoids out of the retina into the pigment epithelium. It follows then that the nighttime transport of 11-cis-retinal from the pigment epithelium back to the retina (46) coincides with increased photoreceptor sensitivity probably due to the reduction of opsin noise by 11-cis-retinal.