Light-dependent Activation of Rod Transducin by Pineal Opsin*

The pineal gland expresses a unique member of the opsin family (P-opsin; Max, M., McKinnon, P. J., Seidenman, K. J., Barrett, R. K., Applebury, M. L., Takahashi, J. S., and Margolskee, R. F. (1995) Science 267, 1502–1506) that may play a role in circadian entrainment and photo-regulation of melatonin synthesis. To study the function of this protein, an epitope-tagged P-opsin was stably expressed in an embryonic chicken pineal cell line. When incubated with 11-cis-retinal, a light-sensitive pigment was formed with a λmax at 462 ± 2 nm. P-opsin bleached slowly in the dark (t 1/2 = 2 h) in the presence of 50 mm hydroxylamine. Purified P-opsin in dodecyl maltoside activated rod transducin in a light-dependent manner, catalyzing the exchange of more than 300 mol of GTPγS (guanosine 5′-O-(3-thiotriphosphate))/mol of P-opsin. The initial rate for activation (75 mol of GTPγS bound/mol of P-opsin/min at 7 μm) increased with increasing concentrations of transducin. The addition of egg phosphatidylcholine to P-opsin had little effect on the activation kinetics; however, the intrinsic rate of decay in the absence of transducin was accelerated. These results demonstrate that P-opsin is an efficient catalyst for activation of rod transducin and suggest that the pineal gland may contain a rodlike phototransduction cascade.

. The phototransduction pathway that leads to these two effects has not yet been elucidated, and the relationship between the acute and phase-shift responses is uncertain.
Acute photo-suppression of chick pineal melatonin synthesis and photo-entrainment of the circadian clock respond to a spectrally broad range of wavelengths with equivalent sensitivity between 450 and 500 nm (12,13). A rhodopsin-like photopigment was postulated to mediate pineal phototransduction based on several observations: sensitivity to light in the visible region; an action spectrum consistent with vitamin A-based pigments (12); the existence of 11-cis-and all-trans retinal in the pineal gland (14 -16); opsin immunoreactivity (17); and immunochemical identification in the pineal gland of retinal proteins such as transducin (18), arrestin (19), and others (20).
Using molecular cloning, a chicken pinealocyte-specific cDNA was recently identified that encodes a protein with strong similarity to the visual opsins (1,21). This protein, pineal opsin (P-opsin), is ϳ40% identical to the visual pigments. P-opsin, but none of the known chicken visual opsins, was found to be expressed in pinealocytes by RNase protection and in situ hybridization, 2 suggesting that P-opsin might mediate pineal phototransduction. A consideration of P-opsin's amino acid sequence suggests that it binds 11-cis-retinal and activates transducin. To characterize the spectral and biochemical properties of P-opsin, an epitope-tagged P-opsin was expressed in embryonic chick pineal cells (CP3 cells; Ref. 22). P-opsin was regenerated with 11-cis-retinal and purified by immunoaffinity chromatography to obtain absorption spectra. In addition, we characterized the ability of expressed P-opsin to catalyze light-dependent guanyl nucleotide exchange on bovine rod transducin.

EXPERIMENTAL PROCEDURES
Construction of the P-opsin Expression Vector-To create the epitopetagged P-opsin cDNA (23), an EcoRI-ApaI fragment of the P-opsin gene was cloned into pMT4 (24) together with the synthetic linker (sense, CCATGCAGATGTCACCGCAGCGGGGCTGAGGAACAAGGTGATGC-CAGCACACCCCGTGGAGACTAGTCAGGTGGCTCCTGCTTGAGC; antisense, GGCCGCTCAAGCAGGAGCCACCTGACTAGTCTCCACG-GGGTGTGCTGGCATCACCTTGTTCCTCAGCCCCGCTGCGGTGAC-ATCTGCATGGGGCC), that recreates the carboxyl terminus of P-opsin sequence  and adds the 1D4 epitope and termination codon of bovine rhodopsin (341-348). The resulting cDNA was subcloned as an EcoRI-NotI fragment into pBK-RSV (with the prokaryotic promoter removed; Stratagene) and was confirmed by DNA sequencing.
Purification of P-opsin Photopigment-P-opsin-positive CP3 cells were grown in flasks in Dulbecco's modified Eagle's medium/F-12 medium and 10% fetal bovine serum containing G418 (60 g/ml), 0.4 mM retinal acetate and penicillin-streptomycin. Half of the growth media and cells were harvested every other day and replaced by fresh media. Cells were recovered by centrifugation, washed once with 1ϫ buffer Y1 (1ϫ ϭ 50 mM HEPES, pH 6.6, 140 mM NaCl, 3 mM MgCl 2 ), resuspended in a small volume of 1ϫ buffer Y1. All subsequent procedures were carried out at 0 -4°C. Cells were pooled to obtain ϳ6 ϫ 10 8 cells, centrifuged at 40,000 ϫ g for 1 h and resuspended in 20 ml of 0.1ϫ buffer Y1 containing 50 g/ml each of aprotinin, leupeptin, pepstatin, and phenylmethylsulfonyl fluoride. The cell slurry was disrupted by passage through progressively narrower syringe needles and centrifuged at 1500 ϫ g for 30 min to pellet nuclei. Membranes were recovered from the supernatant by centrifugation at 40,000 ϫ g for 1 h. The membrane pellet was resuspended in 9 ml of 1ϫ buffer Y2 (ϭ buffer Y1 containing 20% (v/v) glycerol and supplemented with 50 g/ml each of aprotinin, leupeptin, pepstatin, and phenylmethylsulfonyl fluoride). All subsequent steps were performed under dim red light (Kodak no. 2) at 4°C. Membranes and 11-cis-retinal (final concentration, 5 M) were incubated overnight, solubilized by the addition of 10% (w/v) dodecyl maltoside (DM) to a final concentration of 1% DM, and incubated for 2 h. Insoluble material was removed by centrifugation at 40,000 ϫ g for 1 h. The P-opsin protein was purified on 1D4-Sepharose by immunoaffinity chromatography (26,27). Rhodopsin from bovine retina was purified in parallel with P-opsin. Lipid was added to the samples from a sonicated stock of egg phosphatidylcholine (PC, Sigma) to a final concentration of 0.2% egg PC (w/v) and 0.2% DM (w/v) in 1ϫ buffer Y2. Following incubation in dark on ice for at least 1 h, the P-opsin lipiddetergent mixture was diluted for subsequent experiments.
Absorption Spectroscopy-UV-visible absorption spectra of pigments were performed at 20°C in buffer Y2 containing 0.1% DM. Spectra were recorded with a Beckman DU 640 single beam spectrophotometer equipped with a water-jacketed cuvette holder. The pigment concentrations were determined spectroscopically from the absorbance maxima, using the extinction coefficient of rhodopsin. Pigment was treated in the dark with 50 mM hydroxylamine (pH 6.7), and absorbance was monitored at various times following addition. The half time for the reaction was determined by fitting the absorbance at 462 nm versus time to a single exponential decay function (SigmaPlot software).
Transducin Activation-Bovine transducin purification and [ 35 S]GTP␥S exchange assays have been described elsewhere (28,29). Purified transducin bound 0.9 -1.0 mol of [ 35 S]GTP␥S/mol of protein in the presence of excess, light-activated rhodopsin. The concentration of active transducin in each assay was determined independently by the amount of [ 35 S]GTP␥S bound in response to an excess of light-activated rhodopsin. Reactions were carried out at 22°C in 0.01% DM, 10 mM Tris acetate (pH 7.0), 100 mM NaCl, 5 mM MgCl 2 , 5 mM 2-mercaptoethanol, 2 M [ 35 S]GTP␥S (1400 -4000 cpm/pmol) with transducin and initiated by exposing the samples to light from a 300-watt projector (Eastman Kodak) at a distance of 50 -60 cm using a high-pass colored glass cut-off filter (Ͼ515 nm, Edmund Scientific Inc., Barrington, NJ). All data were corrected for [ 35 S]GTP␥S binding to transducin in the absence of pigment (typically 5-8% of the maximum). The lifetime of the active species was determined by illuminating ( max Ͼ 515) a solution of P-opsin or rhodopsin (in buffers as indicated in the figure legends) at 22°C. At the times indicated, an aliquot of the pigment solution was added to reaction buffer containing transducin and [ 35 S]GTP␥S and the reaction proceeded for 1 h before filtering through nitrocellulose. Data were fit using SigmaPlot (Jandel).

P-opsin Expressed in CP3 Cells Targets to the Cell Mem-
brane-The small size of the chicken pineal gland prohibits ready purification of P-opsin or other phototransduction components for functional assays. Therefore, P-opsin was expressed in transfected cells. To enable purification (26) of the expressed protein, an 8-amino acid epitope derived from bovine rhodopsin (amino acids 341-348) was added to the carboxyl terminus of P-opsin. This minor carboxyl addition is not expected to significantly change the properties of P-opsin based on results from proteolytic cleavage of bovine rhodopsin or spectral characterization and transducin activation by similarly tagged pigments (27, 30 -33). All experiments reported below were performed with the epitope-tagged P-opsin.
Previous workers have used a monkey kidney cell line (COS1; Refs. 31 and 34) or a human embryonic kidney cell line (HEK293; Ref. 35) to express functional visual proteins. Western blot analysis showed levels of P-opsin protein expressed in COS1 cells comparable to protein levels obtained for bovine rhodopsin (ϳ4 g/10 7 COS1 cells; data not shown). However, there was barely detectable pigment formed upon incubation of cells with 11-cis-retinal (6 ng of pigment/10 7 COS1 cells or ϳ0.15% pigment formation). This extremely low level of Popsin pigment formation was similar to that observed previously using HEK293 cells (21).
A chicken embryonic pineal cell line (CP3; Ref. 22), which does not express P-opsin ( Fig. 1), was chosen as an alternative expression system. CP3 cells were transfected with a plasmid encoding P-opsin under the control of the Rous sarcome virus promoter. Individual G418-resistant clones varied by more than 10-fold in the level of mRNA for P-opsin (Fig. 1A), reaching levels comparable to that in pineal gland in lines 4 and 9. Line 9 had the strongest immunofluorescence signal (Fig. 1B) and thus was chosen to produce the P-opsin for in vitro experiments.
Immunostaining of epitope-tagged P-opsin was most prominent around the cell periphery, indicating that P-opsin was targeted to the cell membrane (Fig. 1B). The parental cell line did not show any reactivity to the 1D4 antibody. CP3 cellexpressed P-opsin was only found in the membrane fraction (data not shown), and Western blots of the membranes showed two broad bands with apparent molecular masses of ϳ40 and ϳ43 kDa (Fig. 1C). The presence of two labeled bands suggests FIG. 1. P-opsin expression in lines of stably transfected CP3 cells. A, RNase protection assays, using P-opsin (top) and actin (bottom) probes, were performed on RNA (ϳ20 g each) from eight individual lines of CP3 cells stably transfected with P-opsin, the parental (CP3) cell line, and chicken pineal gland (Pin). The sizes of the protected products for P-opsin and actin are 213 and 100 base pairs, respectively. B, CP3 and P-opsin-transfected CP3 line 9 cells were labeled using the 1D4 antibody and a fluorescein isothiocyanate-conjugated secondary antibody. C, immunochemical detection using the 1D4 antibody on a Western blot of total membrane protein (10 g) from P-opsin line 9 and CP3 membranes labeled two broad bands having molecular masses of ϳ 43 and 40 kDa. that P-opsin is heterogeneously glycosylated. This was confirmed by treatment with peptide N-glycosidase F, which transformed both original bands to the same higher mobility band (data not shown). From the deduced amino acid sequence, there are two potential N-linked glycosylation sites (1). The sizes of the bands in P-opsin expressing CP3 membranes suggest that both sites are utilized. P-opsin Binds 11-cis-Retinal to Form a Photopigment-To determine if P-opsin can form a photopigment, membranes were incubated with 11-cis-retinal in the dark. The purified P-opsin exhibited a peak absorption at 462 Ϯ 2 nm ( Fig. 2A) with a profile similar in shape to that of rhodopsin ( Fig. 2A,  inset), with some distortion in the blue region, due to protein absorption (280 nm).
An important property of 11-cis-retinal (A 1 )-based pigments is an increase in bandwidth with decreasing max (36). Using the long wavelength side of the P-opsin peak to estimate bandwidth, we found a half-bandwidth of 4590 cm Ϫ1 for P-opsin and 4230 cm Ϫ1 for rhodopsin ( max ϭ 500 nm). The observed value for the P-opsin pigment agrees with the value (4600 cm Ϫ1 ) predicted from the visual pigment nomogram (36). A dark-light difference spectrum in the presence of hydroxylamine (Fig. 2B) shows that the 462 nm absorbance disappears with a corresponding increase in absorbance at 363 nm, consistent with the formation of retinaloxime. Purified P-opsin, regenerated with 11-cis-retinal, was denatured in the dark at pH 1.8. This treatment stably traps protonated Schiff base linkages (440 nm) and otherwise eliminates interactions of retinal with the apoprotein (37). Acid-trapped P-opsin exhibited a shift in peak absorbance from 462 nm to 440 nm, with a slight increase in bandwidth (Fig. 2C), consistent with results from rhodopsin (37). The ratio of the acidic to the control peak heights indicates that the extinction coefficient for P-opsin is similar to that of rhodopsin (ϳ40,000 M Ϫ1 cm Ϫ1 ), although a more precise estimate of the extinction coefficient was not possible due to the significant increase in absorbance at short wavelengths following acid denaturation.
The absorption spectrum of P-opsin had a higher than expected ratio of 280 nm to 462 nm peaks (ϳ5), when compared with rhodopsin (280/500 nm ratio of ϳ1.6; Ref. 38) due to proteins in the preparation that do not absorb at 462 nm: either misfolded P-opsin or other co-purifying proteins. A silverstained gel of purified P-opsin showed two prominent P-opsin bands (Fig. 3B, bands 2 and 3) that were reactive with the 1D4 antibody in a corresponding Western blot (Fig. 3A). One distinct band and several minor ones appeared in the silverstained gel that did not label with 1D4 (Fig. 3B). The unlabeled protein(s) did not appear in 1D4-Sepharose immunopurified material from parental CP3 cell membranes (data not shown); thus, the presence of non-P-opsin proteins in the immunopurified preparation results from an interaction with P-opsin. The high ratio of A 280 /A 462 is primarily due to the presence of non-opsin protein. However, there may also be a small amount of misfolded or non-functional P-opsin in the preparation.
In darkness cone pigments bleach rapidly in the presence of hydroxylamine, but typical rhodopsins are stable (39 -43). Bleaching of the pigment in the presence of hydroxylamine in darkness indicates that the chromophore is accessible to the aqueous environment. We added hydroxylamine to purified P-opsin pigment in the dark and observed the effect on the absorption spectrum over time (Fig. 4). P-opsin bleached in the presence of hydroxylamine. However, the reaction occurred slowly, with a half time of decay of approximately 2 h; absorbance at 462 nm persisted even after 5 h.
P-opsin Activates Rod Transducin-Visual pigments interact in a light-dependent fashion with transducins and stimulate the exchange of bound guanyl nucleotide. Sequence homology with visual opsins suggests that P-opsin may interact with a transducin. In fact, transducin immunoreactivity has been observed in chicken pineal (18) and rod, but not cone, transducin ␣ subunit mRNA has been detected in chicken pineal gland. 2   FIG. 2. UV-visible spectra of purified P-opsin pigment. A, the spectrum of P-opsin pigment in 0.1% dodecyl maltoside following regeneration with 11-cis retinal, and purification on 1D4-Sepharose. Inset, expanded view of the visible region of P-opsin and bovine rhodopsin for comparison. B, difference spectrum obtained by subtracting the spectrum of P-opsin exposed to light in the presence of 50 mM hydroxylamine from a dark spectrum taken before light exposure and the addition of hydroxylamine. C, acid denaturation of P-opsin. Absorbance spectra of purified P-opsin before (control) and after addition of acid to pH 1.8 (ϩacid). The peak absorbance of P-opsin shifted from 462 nm to 440 nm upon addition of acid.

FIG. 3. Purification of P-opsin.
A, Western blot of P-opsin (P-opsin) and bovine rhodopsin (Rho) purified by 1D4 immunoaffinity chromatography. The proteins were visualized using 1D4 antibody and antimouse HRP followed by ECL detection. Samples were maintained in the dark prior to the addition of SDS loading buffer to prevent dimerization due to pigment bleaching. B, silver-stained SDS-PAGE of purified rhodopsin (Rho) and P-opsin (P-opsin). Numbered arrows (1-3) indicate bands that label with 1D4 antibody in Western blots in A.
These results and others (44) suggest that P-opsin and rod transducin proteins are present in the same cells (pinealocytes) and may provide a phototransduction pathway.
We investigated the interaction of P-opsin with exogenously added bovine transducin, since bovine and chick ␣ transducins are nearly identical (45) and bovine transducin is readily available. Nucleotide exchange assays were performed using purified components under conditions that produce high activity of similarly prepared bovine rhodopsin (46) and violet cone opsin (27). Transducin activation by P-opsin was strictly light-dependent (i.e. in the dark, nucleotide exchange was the same as transducin alone), and addition of phospholipid did not alter the rate of transducin activation (Fig. 5A). Under these conditions, the reaction was complete by 10 min and was not limited by the available transducin ( Fig. 5A; see below). Varying concentrations of purified P-opsin were exposed to light and nucleotide exchange reactions were allowed to go to completion. The total number of activated transducins increased with increasing P-opsin concentration, eventually reaching saturation (Fig. 5B, closed symbols). Approximately 50% of the transducin that could be activated by excess rhodopsin was activated by P-opsin, indicating that the cessation of activity was not caused by depletion of unactivated transducin. Moreover, the concentration of P-opsin required to produce 50% of the maximum turnovers was similar for both transducin concentrations (6 nM P-opsin for 3 M transducin compared with 3 nM P-opsin for 1 M transducin), and substantially below the concentration of transducin throughout the reaction. At concentrations below 1 M transducin, the fraction of transducin activated dropped significantly (Fig. 5B). Thus, it appears that P-opsin exhibits an interaction-dependent inactivation; in contrast, bovine rhodopsin is able to activate Ͼ95% of the available transducin. In the linear portion of the P-opsin concentration range, P-opsin catalyzed the exchange of GTP␥S on a large number of transducin molecules (162 GTP␥S bound/P-opsin and 80 GTP␥S bound/P-opsin at 3 M and 1 M transducin respectively). Exogenously added lipid did not have a significant effect on the total number of transducins activated (Fig. 5B, open symbols).
The initial rates of GTP␥S exchange at different transducin concentrations were measured using light-activated P-opsin and bovine rhodopsin. Comparing the slopes of the linear portions of the curves (Fig. 6), the initial rate of nucleotide exchange catalyzed by rhodopsin was 2-3 times faster than that achieved with P-opsin. These results indicate that the efficiency of P-opsin interaction with rod transducin (the binding affinity, catalytic turnover rate, or both) is lower than that of rhodopsin. The addition of phospholipid decreased the initial FIG. 4. Hydroxylamine sensitivity of P-opsin pigment. A, spectra were obtained from P-opsin that had been maintained in darkness for various times following addition of 50 mM hydroxylamine. The initial spectrum was taken prior to addition of hydroxylamine. The remaining spectra were taken at times indicated in B. At the end of hydroxylamine exposure in the dark, light caused no further change. B, disappearance of the 462 nm absorbance from A as a function of time following addition of hydroxylamine. The line represents an exponential fit with a half-time of the reaction of 2 h. rate of activation by 1.5-2-fold. In contrast, rhodopsin exhibited a significant, 4-fold increase in activity in the presence of lipid. Compared with P-opsin in lipid micelles, rhodopsin had a 15-20-fold faster initial rate.
Visual pigments respond to light by going through a series of photo-intermediates leading to a metastable state (metarhodopsin II) that stimulates transducin (47). The metarhodopsin II state decays to the apoprotein plus all-trans-retinal, and the time course of the decay plays an important role in phototransduction. To determine the stability of its active state, P-opsin was exposed to light for various times prior to addition to the nucleotide exchange reaction. At low detergent concentrations (0.01%), P-opsin exhibited a slow decay in active conformation, with a half-time of about 60 min (Fig. 7). Under the same conditions, rhodopsin had a slightly faster decay, with a halftime of about 30 min. Surprisingly, P-opsin did not decay at higher detergent concentrations. When phospholipid was added, P-opsin's decay was dramatically accelerated, with a half-time of about 4 min. The increase in decay of metarhodopsin II in phospholipid compared with detergent has been previously described (reviewed in Ref. 48). Thus, P-opsin produces a meta-stable active state when exposed to light, with properties that parallel metarhodopsin II. In comparison to other opsins, P-opsin remains active 5 times longer than violet cone opsin (27) and its active state decays more slowly than that of green cone opsin (as estimated by spectroscopic measurements; Refs. 49 and 50).

DISCUSSION
Is P-opsin a Rhodopsin-like Photopigment?-We have investigated several properties of P-opsin and made comparisons to the visual pigments. First, CP3-expressed P-opsin bound 11cis-retinal with a maximum absorbance at 462 nm, in good agreement with the dark-light difference spectrum ( max ϳ470 nm) obtained from a membrane preparation of P-opsin-expressing HEK 293 cells (21). The spectral profile of P-opsin (adjusted for protein contaminants) is very similar to that predicted for vitamin A 1 -based pigments. Acid denaturation of P-opsin in the dark indicated that P-opsin binds 11-cis-retinal with a Schiffbase linkage. Although P-opsin reacted in the dark with hydroxylamine, unlike rhodopsin, the kinetics were much slower than cone pigments, which are bleached within minutes of exposure (51,52). This suggests that the accessibility of the Schiff base to water-soluble reagents in P-opsin is intermediate between rod and cone opsins.
In contrast to the visual pigments (26,30,33), COS1-expressed P-opsin did not fold efficiently into a conformation that bound 11-cis-retinal (similar results were found in human embryonic kidney cells; Ref. 21). The reason for this is unclear, but may reflect a need for a pineal-specific protein (also found in CP3 cells) that is required for proper folding and chromophore binding. In the past few years, a number of extra-retinal photoreceptor opsins have been cloned. These include VA-opsin (53), RGR (54 -56), parapinopsin (57), melanopsin (58), and several opsin homologs to P-opsin (41, 59 -61). Attempts to express these proteins in a functional state have generally failed. This suggests that non-retinal opsins may require specialized conditions or proteins for proper folding. Recently, fish VA-opsin, expressed in Cos I cells, was shown to bind 11-cisretinal and undergo bleaching when exposed to light (62). The max of VA-opsin (465 nm) is very similar to that shown here for P-opsin. The results reported here are the first biochemical studies of a extra-retinal opsin and show that many of the transduction properties exhibited by rhodopsin are conserved for P-opsin.
What Determines the Number of Transducin Molecules Activated by P-opsin?-We have shown that P-opsin activates rod transducin in a light-dependent manner that is similar to rhodopsin but distinct from violet cone opsin (27). Moreover, Popsin has a meta-stable active conformation that is very similar to metarhodopsin II. Taken together, these observations show that P-opsin is a rhodopsin-like photopigment. Although Popsin activity is comparable to rhodopsin, the lower efficiency suggests that P-opsin may have a lower affinity or slower turnover than rhodopsin. These differences may be intrinsic to P-opsin activity, due to subtle sequence differences between bovine and chicken rod transducins, or in vitro conditions may not be optimized for P-opsin. Therefore, quantitative comparisons between the two proteins require caution. Nonetheless, an FIG. 7. Lifespan of light-activated P-opsin. P-opsin was exposed to light (Ͼ515 nm) at 22°C in three buffers: 0.1% DM (closed squares), 0.01% DM (closed circles), and 0.02% egg PC/DM (open circles). At the indicated times the pigment was diluted into reaction buffer to a final detergent concentration of 0.01% DM containing transducin and [ 35 S]GTP␥S. Nucleotide exchange assays were allowed to proceed for an additional 1 h. The remaining ability of each pigment to activate transducin was normalized to the pigment's activity that was immediately assayed following exposure to light. The lines represent fits to an exponential decay, and the decay for rhodopsin in 0.01% DM is shown (inverted triangles) for comparison. unusual feature of the in vitro activation of transducin by P-opsin is the inability to activate all of the transducin that can be activated by rhodopsin in comparable assays. Several explanations can be envisaged. First, an inhibitor (i.e. product inhibition) may be produced during activation e.g. G␤␥ or G␣-GTP␥S. Second, the transducin preparation may not be homogeneous, but may contain distinct "pools" or complexes, which interact differentially with P-opsin and rhodopsin. Third, there may be a limiting impurity in the G protein necessary for P-opsin activity. Further work is needed to resolve these important possibilities.
There is an apparent discrepancy between the measured time course of activation of transducin by P-opsin (Fig. 5A) and measurements of the decay of the light-activated conformation (Fig. 7). From the decay results, a persistent (Ͼ60 min) stimulation of GTP␥S exchange is expected, yet the reaction is complete by 10 min. This discrepancy can not be explained by substrate limitation, since transducin and GTP␥S are in excess. Rather, it suggests that there is a turnover-dependent loss of activated P-opsin. One plausible mechanism is that interaction of P-opsin with transducin causes a change in Popsin conformation, leading to an accelerated release of alltrans-retinal. There have been a number of reports showing that transducin effects the lifetime of metarhodopsin II (63,64). From our data, this effect of transducin on P-opsin is less dramatic in the presence of phospholipid where the rate of decay is already accelerated. Thus, biochemical measurements of P-opsin interaction with transducin are complicated both by an unstable active state and potential allosteric effects resulting from P-opsin/transducin interaction.
Does P-opsin Mediate Circadian Phototransduction in the Pineal Gland?-That P-opsin forms a light-sensitive pigment which activates rod transducin suggests that P-opsin may be involved in one or both of the pineal gland's response to light. Studies of the acute inhibition of N-acetyltransferase activity (12,13) and the circadian phase shift (13) show a broad sensitivity profile with the largest response occurring between 450 and 500 nm. No difference in wavelength sensitivity is found between acute inhibition of melatonin and phase shifts of the circadian clock, suggesting that both processes are controlled by the same or similar photopigment(s). The available physiological data do not distinguish between single and multiple photopigments; however, they are consistent with the P-opsin absorbance spectrum and suggest that P-opsin participates in both responses.
The acute suppression of N-acetyltransferase activity is inhibited by pertussis toxin (65) and involves the reduction of cAMP (66 -68). It is possible that a retinal-like phototransduction pathway, employing a cyclic nucleotide phosphodiesterase, mediates the acute response to light. Additional support for this hypothesized phototransduction cascade comes from the expression in pinealocytes of rod transducin and rodlike ␤and ␥-subunits of the retinal cGMP phosphodiesterase, as well as a cone ␣ cGMP-phosphodiesterase subunit. 3 Activation of a phosphodiesterase may alter the level of cAMP and activity of third messengers and result in lowered activity of AA-NAT (68), perhaps via its degradation (69), leading to diminished melatonin biosynthesis.
The second response to light by chick pinealocytes involves photoentrainment of the circadian clock that controls the rhythmic synthesis of melatonin. This pathway is insensitive to pertussis toxin (65) and is separate from the rod transducin pertussis toxin-sensitive acute pathway. If P-opsin pigment mediates photoentrainment, it must do so via a secondary process unrelated to its activation of rod transducin. Opsins, like other G-protein coupled receptors, may regulate effectors via multiple G-protein pathways. For example, when expressed in Xenopus oocytes, rhodopsin activates membrane currents in the absence of transducin (70,71). The pineal gland expresses multiple G proteins in addition to rod transducin. Okano et al. (72) report the presence of G i2 , G i3 , and two G o splice variants in pineal gland cDNA libraries, and we find those messages as well as G i1 and G z in pinealocytes by RPA and in situ hybridization. 2 Of these, G z is the only one known to be pertussis toxin-insensitive (73). It will be of interest to test P-opsin's ability to activate G z and the other G-proteins that are expressed in the pineal gland.