Contribution of protein kinase C to the phosphorylation of rhodopsin in intact retinas.

Similar to other G protein-coupled receptors, the visual receptor, rhodopsin, is phosphorylated by both a substrate-regulated kinase, rhodopsin kinase, and a second messenger-regulated kinase, protein kinase C. In the present study, the extent of involvement of protein kinase C in the light-dependent phosphorylation of rhodopsin in intact retinas was assessed using a specific activator (phorbol ester) and specific inhibitor (calphostin C) of protein kinase C. Kinetic analysis of rhodopsin phosphorylation following different illumination conditions revealed that hyperactivation of protein kinase C with phorbol ester resulted in a relative increase in rhodopsin phosphorylation that peaked 10-15 min after the onset of illumination. Following this period, the rate of rhodopsin dephosphorylation was increased in the phorbol ester-treated retinas, so that by about 30 min the amount of phosphorylation was similar to that in control retinas. Treatment of retinas with calphostin C, a potent regulatory domain-directed inhibitor of protein kinase C, resulted in an approximately 50% reduction in the light-dependent phosphorylation of rhodopsin. This inhibitor had no effect on the activity of rhodopsin kinase in vitro Last, we show that frog rhodopsin is phosphorylated in vitro by protein kinase C from frog rod outer segments, indicating that this kinase could directly modulate rhodopsin in vivo In conclusion, the present results reveal that the kinetics of rhodopsin phosphorylation/dephosphorylation differ markedly, depending on whether protein kinase C or rhodopsin kinase activity dominates, and that, under the conditions studied, protein kinase C contributes to approximately half of the phosphorylation of rhodopsin in intact frog retinas.

Similar to other G protein-coupled receptors, the visual receptor, rhodopsin, is phosphorylated by both a substrate-regulated kinase, rhodopsin kinase, and a second messenger-regulated kinase, protein kinase C. In the present study, the extent of involvement of protein kinase C in the light-dependent phosphorylation of rhodopsin in intact retinas was assessed using a specific activator (phorbol ester) and specific inhibitor (calphostin C) of protein kinase C. Kinetic analysis of rhodopsin phosphorylation following different illumination conditions revealed that hyperactivation of protein kinase C with phorbol ester resulted in a relative increase in rhodopsin phosphorylation that peaked 10 -15 min after the onset of illumination. Following this period, the rate of rhodopsin dephosphorylation was increased in the phorbol ester-treated retinas, so that by about 30 min the amount of phosphorylation was similar to that in control retinas. Treatment of retinas with calphostin C, a potent regulatory domain-directed inhibitor of protein kinase C, resulted in an approximately 50% reduction in the light-dependent phosphorylation of rhodopsin. This inhibitor had no effect on the activity of rhodopsin kinase in vitro. Last, we show that frog rhodopsin is phosphorylated in vitro by protein kinase C from frog rod outer segments, indicating that this kinase could directly modulate rhodopsin in vivo. In conclusion, the present results reveal that the kinetics of rhodopsin phosphorylation/dephosphorylation differ markedly, depending on whether protein kinase C or rhodopsin kinase activity dominates, and that, under the conditions studied, protein kinase C contributes to approximately half of the phosphorylation of rhodopsin in intact frog retinas.
The involvement of multiple kinases in the regulation of specific signaling pathways is a common cellular theme. For G protein-coupled pathways, two classes of kinases phosphorylate receptors: substrate-regulated G protein-coupled receptor kinases and second messenger-regulated kinases such as protein kinase A and protein kinase C (1). In the examples studied to date, phosphorylation by G protein-coupled receptor kinases is typically involved in homologous desensitization (the phosphorylation follows direct stimulation of the receptor), and phosphorylation by second messenger-regulated kinases is in-volved in heterologous desensitization (the phosphorylation is independent of the liganded state of the receptor and may be initiated by activation of the kinase by a separate receptor) (1,2).
Phototransduction serves as the archetypal transduction pathway, and extensive studies with photoreceptor proteins have provided invaluable insight into the regulation of not only this pathway but numerous other G protein-coupled receptor signaling pathways as well (3)(4)(5). Activation of the heptahelical receptor, rhodopsin, by a photon of light results in activation of the G protein transducin, which in turns activates a cGMP phosphodiesterase. The ensuing drop in cGMP levels in photoreceptor outer segments causes hyperpolarization of the plasma membrane by closing cGMP-gated channels, thus converting a biochemical signal into an electrophysiological response (6). Phosphorylation of the light-activated receptor by rhodopsin kinase, followed by binding of arrestin, effectively quenches the signal. Ca 2ϩ levels, which are lowered when the cGMP-gated channels close, regulate deactivation and recovery stages of phototransduction. Two targets of Ca 2ϩ have been identified in photoreceptors: guanylate cyclases, which are responsible for restoring cGMP levels, are activated at low Ca 2ϩ concentrations by calmodulin-like proteins (7), and the activity of rhodopsin kinase is inhibited by another Ca 2ϩ -binding protein, recoverin, at high concentrations of Ca 2ϩ (8,9).
Several reports over the past decade have implicated a role for protein kinase C in photoreceptor function; its presence in rod outer segments (10 -12) and its phosphorylation of rhodopsin (13,14), transducin (15), the ␥ subunit of the cGMP-phosphodiesterase (16), and arrestin (17) have been described. Protein kinase C is activated by the lipid second messenger diacylglycerol through binding of this ligand to a membranetargeting domain (C1) that also binds phorbol esters (18,19). Some isozymes of protein kinase C are sensitive to Ca 2ϩ , which binds to a separate membrane-targeting domain (C2) and thus also increases the enzyme's affinity for membranes (18). Generation of diacylglycerol, the key activator of protein kinase C, has been reported to be stimulated by light in photoreceptors (20 -23), consistent with a potential role for this kinase in phototransduction.
Evidence that the phosphorylation of rhodopsin by protein kinase C is physiologically relevant comes from the finding that hyperactivation of protein kinase C in intact retinas, by treatment with phorbol esters, alters the light-dependent phosphorylation of rhodopsin (24,25). In vitro studies reveal remarkable similarities between the enzymology of phosphorylation of rhodopsin by protein kinase C and the phosphorylation of other G protein-coupled receptors by second messenger kinases (26). Most notably, phosphorylation is independent of the activation state of the receptor (13,14) and results in decreased coupling of the receptor to G protein (13). Taken together, the in situ and in vitro data are consistent with direct modulation of rhodopsin by protein kinase C. Nonetheless, the extent of involvement of protein kinase C in rhodopsin phosphorylation in vivo remains to be established.
In the present study, specific inhibition of protein kinase C by calphostin C was used to determine the contribution of protein kinase C in the light-dependent phosphorylation of rhodopsin in intact frog retinas. We found that this kinase contributes to approximately 50% of the phosphorylation of rhodopsin. Furthermore, using phorbol esters to hyperactivate the kinase, we show that the kinetics of rhodopsin phosphorylation and dephosphorylation differ markedly, depending on whether protein kinase C or rhodopsin kinase activities dominate.
Rhodopsin Phosphorylation in Situ-Freshly dissected retinas from dark-adapted (12-14 h) frogs were incubated individually in 1 ml of amphibian culture medium (ACM; 65 mM NaCl, 35 mM NaHCO 3 , 3 mM KCl, 2 mM CaCl 2 , 1 mM MgSO 4 , 10 mM Na-HEPES (pH 7.3), 10 g⅐ml Ϫ1 phenol red, 1 mg⅐ml Ϫ1 casamino acids, 10 mM ␤-D-glucose, 0.1% bovine serum albumin) containing 20 -100 Ci⅐ml Ϫ1 carrier-free [ 32 P]orthophosphate and gassed with 95% O 2 , 5% CO 2 at room temperature (21-23°C). After 30 -120 min, 100 l of ACM containing an 11-fold concentrate of PMA or calphostin C was added to gently agitated tubes to yield final concentrations of 1-1000 nM PMA and 1 M calphostin C. An appropriate concentration (0.1%) of Me 2 SO (solvent for PMA and calphostin C) was added to control tubes. One retina from each frog served as the experimental and the other as the control. After a 15-120-min incubation with Me 2 SO or pharmacophore, retinas were exposed to one of three light regimes by flash or continuous illumination. A Vivitar flash unit (model 285) was used to provide a single flash of one of two intensities. For the lower intensity, the flash unit was set at 1 ⁄16 maximal intensity and filtered by 20 layers of lens tissue, which act as a neutral density filter, with measured attenuation of 1.2 log units. The higher intensity was effected by an unfiltered flash from the unit at 1 ⁄2 maximal intensity. For a third condition of light exposure, retinas were illuminated by 500-lux cool fluorescent lighting for 1-30 min. Some samples were not exposed to light. After the flashes, retinas were incubated under dim red light for 1-60 min. Retinas were gently agitated (approximately 40 rpm) during incubations. All procedures were carried out under dim red light (15-W bulb, No. 1A Safelight Filter, Eastman Kodak Co.) unless otherwise stated.
Rhodopsin phosphorylation from retinas was analyzed by one of two methods. In one, 300 l of ice-cold buffer containing 80 mM sodium phosphate (pH 7.3), 20 mM EDTA, 20 mM EGTA was added to each sample, retinas were vortexed for 10 s to break off rod outer segments, and tubes were placed on ice for 2 min to allow retinas to settle. Aliquots (1 ml) of the rod outer segment-enriched supernatant were centrifuged at 16,000 ϫ g for 30 s. The sedimented rod outer segments were quickly resuspended in 1 ml of ice-cold buffer containing 110 mM NaCl, 5 mM Na-HEPES (pH 7.3), 1 mM EDTA, 1 mM EGTA and centrifuged again, and the pellet was resuspended in 50 l of 1% SDS. To reduce viscosity, samples were frozen at Ϫ80°C, thawed, and centrifuged at 16,000 ϫ g for 15 min. Aliquots of the supernatant (10 -20 l) were mixed with a 5-fold concentrate of SDS-PAGE sample buffer (250 mM Tris-HCl, pH 6.8, 5% SDS, 50 mM ␤-mercaptoethanol, 0.01% bromphenol blue, 50% glycerol). Samples were analyzed by SDS-PAGE (10% acrylamide) and stained with Coomassie Blue R-250; protein in each lane was quantified by densitometric analysis on a Molecular Dynamics computing densitometer, and 32 P incorporation into rhodopsin was quantified by Phos-phorImager analysis (Molecular Dynamics) or scintillation counting of excised rhodopsin bands. This method yielded good quality rod outer segments with clean separation of rhodopsin from other proteins, but there was concern that phosphorylation reactions were not quenched instantly. In the other method, 100% trichloroacetic acid (125 l) was added to retinas in ACM. The tubes were vortexed for 3 s and then placed on ice for 5 min. Samples were then centrifuged for 15 s at 20,000 ϫ g. The retinal pellets were resuspended by sonication (for 10 s with a Branson Probe Sonifer-250, power setting on 2) in 500 l of 50 mM Tris-HCl buffer (pH 7.5) containing 10 mM EDTA, 10 mM EGTA, and 20 mM sodium phosphate and then centrifuged again for 20 min at 23,000 ϫ g, 2°C. Pellets were resuspended in 1% SDS (200 l) and sonicated again, and after a second centrifugation (20 min, 23,000 ϫ g, 22°C), two aliquots (40 l) of the supernatant were mixed with 10 l of a 5-fold concentrate of SDS-PAGE sample buffer. One sample was kept at room temperature while the other was heated for 15 min at 100°C to oligomerize rhodopsin molecules (this treatment resulted in quantitative disappearance of monomeric rhodopsin on SDS-PAGE and served as the background for calculation of 32 P incorporation into rhodopsin). Both methods yielded qualitatively similar results.
Relationship between Light Stimulus and Rhodopsin Photoactivation-Frog retinas in ACM (without [ 32 P]orthophosphate) were subjected to various flash intensities or continuous light under the same conditions used in the in situ experiments above. After illumination, rod outer segment-enriched supernatants were prepared as in the first method above, except that 20 mM hydroxylamine was included in the vortexing mixture. Rod outer segments were isolated by sucrose density gradient centrifugation (29) and washed twice by centrifugation at 1 The abbreviations used are: PMA, phorbol myristate acetate; ACM, amphibian culture medium; DTT, dithiothreitol; PAGE, polyacrylamide gel electrophoresis. 2 N. M. Greene, unpublished data.

FIG. 1. Level of photoexcited rhodopsin resulting from exposure of dark-adapted intact frog retinas to different flash intensities.
Intact frog retinas were exposed to increasing light levels, and the amount of photoexcited rhodopsin was measured as described under "Materials and Methods." The arrow at a indicates the lower flash intensity used in the present study ( Fig. 2 and Fig. 3, B and E); it results in the photoexcitation of 6% rhodopsin. The arrow at b indicates the higher flash intensity used in the present study ( Fig. 2 and Fig. 3, C and F); it delivers 2.8 log units more light and results in the photoexcitation of 80% rhodopsin. Data represent mean Ϯ S.E. 300,000 ϫ g, 10 min, 2°C, in buffer containing 10 mM sodium phosphate, pH 7.3, 110 mM NaCl. Final pellets were solubilized in 80 l of 50 mM sodium phosphate buffer, pH 7.0, containing 1% octylglucoside and incubated for 15 min on ice, and insoluble material was removed by centrifugation at 100,000 ϫ g for 5 min, 2°C. The A 278 /A 498 absorbance ratios of supernatants were measured before and after complete bleaching (5,000 lux, 2 min) of rhodopsin in the presence of 20 mM hydroxylamine. Fig. 1 shows the amount of bleached rhodopsin in intact frog retinas calculated by this method as a function of flash intensity. Arrows indicate the two intensities used in the in situ phosphorylation experiments; setting the flash at 1 ⁄16 maximal power and using a 1.2-log unit neutral density filter bleached 6 Ϯ 3% of the rhodopsin (arrow a), and an unfiltered flash set at 1 ⁄2 maximal intensity bleached 80 Ϯ 6% of the rhodopsin (arrow b). Ten minutes of 500-lux continuous illumination bleached 97 Ϯ 1% of the rhodopsin (not shown). Octylglucosidesolubilized rhodopsin from purified dark-adapted frog rod outer segments was illuminated under the same light regimes as intact retinas, and the amount of bleaching, as a function of light intensity, was determined as described above. Similar levels of bleaching resulted from illumination of intact retinas or solubilized rhodopsin from purified rod outer segments.
Phosphorylation of rhodopsin by protein kinase C was determined as described above except that substrate was replaced with 0.5 M rhodopsin in urea-stripped membranes and Triton X-100/lipid mixed micelles were omitted. In some experiments, 20 -100 nM PMA was present. For calphostin inhibition studies, 0 -1 M of this inhibitor was included in phosphorylation mixtures that also contained 20 nM PMA to ensure full activation of protein kinase C. Phosphorylation by rhodopsin kinase was performed under similar conditions, except that protein kinase C was replaced with rhodopsin kinase and assays did not include PMA (PMA does not affect rhodopsin phosphorylation by rhodopsin kinase in vitro (14)). Reactions were quenched by the addition of 5 l of a 5-fold concentrate of SDS-PAGE sample buffer, and samples were analyzed by SDS-PAGE (10% acrylamide). 32 P incorporation into rhodopsin was assessed as described above for in situ experiments.
Isolation of Protein Kinase C from Frog Rod Outer Segments-Protein kinase C from rod outer segments was isolated following the procedure of Udovichenko et al. (32). Briefly, rod outer segments from 20 frogs were collected under dim red light, purified by two sequential sucrose gradients (29), and lysed in 10 mM Tris, pH 7.5, at 4°C, 1 mM MgCl 2 , 2 mM DTT, 1 g⅐ml Ϫ1 leupeptin, 1 g⅐ml Ϫ1 aprotinin, 1 g⅐ml Ϫ1 pepstatin, 0.5 mM phenylmethylsulfonyl fluoride, and 0.1 mg⅐ml Ϫ1 sodium L-ascorbate. Membranes were washed three times by centrifugation at 300,000 ϫ g for 10 min in this buffer. Membranes were resuspended in buffer containing 10 mM Tris-HCl, pH 7.5, at 4°C, 2 mM EGTA, 2 mM EDTA, 5 mM DTT, 1 g⅐ml Ϫ1 leupeptin, 1 g⅐ml Ϫ1 aprotinin, 1 g⅐ml Ϫ1 pepstatin, 0.5 mM phenylmethylsulfonyl fluoride (buffer A) to extract protein kinase C, and sample was centrifuged at 100,000 ϫ g for 1 h at 4°C. The supernatant, containing protein kinase C, was applied to a 50-l DEAE-Sephacel column equilibrated with buffer A, and the column was washed with 2 ml of buffer A. Protein kinase C was eluted with a step gradient (50-l steps in a 1-ml total volume) of 0 -500 mM NaCl in buffer A, in 25 mM NaCl increments.  1 and 2) and phosphor image showing radioactivity (lanes 3 and 4) of a heated sample (lanes 1 and 3) and a sample kept at room temperature (lanes 2 and 4) are shown. The area of the gel analyzed is indicated by the boxes; the arrow indicates the position of rhodopsin. The heated sample contains little rhodopsin in this area and serves as the background level of protein and radioactivity.
Fractions (50 l) were collected and assayed for protein kinase C activity.
Calcium Concentration-Concentrations of free Ca 2ϩ were calculated using a computer program kindly provided by Dr. Claude Klee that takes into account pH, Ca 2ϩ , Mg 2ϩ , K ϩ , Na ϩ , EGTA, EDTA, and ATP concentrations (33).

RESULTS
Previously we showed that PMA alters the phosphorylation of rhodopsin in intact rat retinas (24,25). To understand better the effect of protein kinase C activation on rhodopsin phosphorylation, we undertook a systematic study of the time-dependent phosphorylation of rhodopsin after different illumination conditions. For these studies, we chose frog retinas because they have been well characterized in short term culture. They are more stable in culture medium than mammalian retinas, so that they are more suited to the present experiments, which require longer incubation periods than those of our previous studies (24,25).
As a first step, we characterized the light-dependent phosphorylation of rhodopsin initiated by three light treatments (Fig. 1): exposure to a filtered flash of light resulting in photoexcitation of approximately 6% of rhodopsin, exposure to a flash of light resulting in photoexcitation of approximately 80% of rhodopsin, and exposure to continuous room lighting (approximately 500 lux) resulting in photoexcitation of approximately 97% of rhodopsin. Fig. 2 shows the time-dependent increase in rhodopsin phosphorylation after treatment of intact frog retinas with these light regimes; these data are compiled from 35 separate experiments. All lighting conditions resulted in a rapid phosphorylation of rhodopsin, with a half-time of approximately 2 min. Relative to the phosphorylation resulting from photoexcitation of 97% of the rhodopsin, approximately 12 and 50% of the rhodopsin was phosphorylated following the two different flash intensities, which photoexcited 6 and 80% of the rhodopsin, respectively. Thus, conditions were established where the kinetics and extent of phosphorylation of control retinas, in the absence of any stimulation other than light, were characterized. We next studied how hyperactivation of protein kinase C, resulting from PMA treatment of retinas, altered the kinetics of rhodopsin phosphorylation. The data in Fig. 3 show the effect of PMA treatment on rhodopsin phosphorylation relative to that of untreated retinas (for these experiments, one retina from each frog was treated with PMA and the other with Me 2 SO, the solvent for PMA). For all illumination conditions, the effects of phorbol esters were biphasic. In general, PMA first caused an increase in rhodopsin phosphorylation relative to controls; this increase was greatest at about 10 -15 min after the onset of illumination. After this time, the relative difference in rhodopsin phosphorylation level between PMA-treated retinas and control retinas decreased, so that by approximately 30 min after the onset of illumination there was little difference. What was unique to each illumination condition was the magnitude of the phorbol ester effect. In response to the flash photoexciting 6% of the rhodopsin, PMA caused an approximately 20% increase in rhodopsin phosphorylation that was sustained for 20 min (Fig. 3, B and E). This increase was Fig. 2. A-D, relative effect of PMA. One retina from each frog was treated with 1 M PMA, and the other retina served as a control. Retinas were incubated for 30 min and then either exposed to one of the light stimuli described in the legend to Fig. 2 or kept in the dark. 32 P incorporation into rhodopsin was measured for retinas not exposed to light (A), retinas exposed to a flash resulting in photoexcitation of 6% of the rhodopsin (B), retinas exposed to a flash resulting in photoexcitation of 80% of the rhodopsin (C), and retinas exposed to 500-lux continuous illumination, resulting in photoexcitation of 97% of the rhodopsin (D). Data are expressed as the level of rhodopsin phosphorylation (normalized to the amount of rhodopsin in each sample) of the PMA-treated retina relative to that of the control retina. They represent mean Ϯ S.E. and were compiled from 6 (n ϭ 12), 12 (n ϭ 100), 13 (n ϭ 104), and 5 (n ϭ 26) separate experiments for A, B, C, and D, respectively. E-G, time courses of rhodopsin phosphorylation in control (solid lines) and PMA-treated (dotted lines) retinas. Curves a, b, and c from Fig. 2A are shown in E, F, and G, respectively, together with curves taken from the data in B, C, and D, respectively. followed by a marked decrease in the relative level of rhodopsin phosphorylation in PMA-treated retinas, such that by 60 min, it was 50% less relative to controls. In retinas exposed to the flash of light resulting in 80% photoexcitation of rhodopsin, PMA treatment resulted in 80% greater rhodopsin phosphorylation by 15 min, although the initial phosphorylation level in PMA-treated retinas was actually less (Fig. 3, C and F). As with the lower flash intensity, rhodopsin phosphorylation in PMA-treated retinas became progressively less relative to control with incubation times exceeding 15 min, except that at the higher flash intensity, the amount of phosphorylated rhodopsin in PMA-treated retinas did not drop below that in control retinas. In retinas exposed to continuous illumination, the same biphasic trend was evident, with the greatest increase in rhodopsin phosphorylation due to PMA occurring after 15 min (Fig. 3, D and G). No effect of PMA was detected in the absence of illumination (Fig. 3A).

FIG. 3. Effect of phorbol myristate acetate on rhodopsin phosphorylation in intact frog retinas. Frog retinas were incubated with [ 32 P]orthophosphate under dim red light as in
To test whether the PMA-induced effect resulted from direct stimulation of protein kinase C, we determined the effect of the biologically inactive enantiomer 4-␣-PMA on rhodopsin phosphorylation in situ. Although 4-␣-PMA caused a slight increase in rhodopsin phosphorylation, PMA caused at least a 2-fold increase above this level, with a half-maximal effect at 2 nM (data not shown). Thus, the major effect of PMA is stereospecific and does not arise from perturbation of rod outer segment membranes or nonspecific interactions with rod outer segment components.
As an alternative approach to studying the effect of protein kinase C on rhodopsin phosphorylation, we investigated the effect of a specific protein kinase C inhibitor, calphostin C, on rhodopsin phosphorylation. Calphostin C binds the phorbol ester-binding (C1) domain of protein kinase C (34) and at concentrations up to 1 M has no significant effect on other kinases (35). Fig. 4 shows that calphostin C specifically inhibits the protein kinase C-catalyzed phosphorylation of rhodopsin in vitro (filled circles), with no detectable inhibition of the rhodopsin kinase-catalyzed phosphorylation of rhodopsin (open circles). Protein kinase C was half-maximally inhibited by 400 nM calphostin C, with 1 M causing approximately 70% inhibition. Note that this inhibition occurred in the presence of 20 nM PMA, which, without calphostin, results in full activation of the kinase; greater inhibition by calphostin would be expected at subsaturating concentrations of activator. In marked contrast, 1 M calphostin C had no effect on rhodopsin kinase activity. Thus, calphostin C, like PMA, serves as a specific agent for examining protein kinase C effects. Fig. 5 shows that treatment of intact retinas with 1 M calphostin C resulted in a marked reduction in the light-dependent phosphorylation of rhodopsin. Compilation of data from four separate experiments revealed a 2-fold reduction in 32 P incorporation into rhodopsin from calphostin-treated retinas relative to controls (as in the preceding experiments, one retina from each frog served as the experimental and the other as the control). These data reveal that at least 50% of the 32 P incorporated into rhodopsin in response to continuous room illumination (97% bleach) depends on protein kinase C.
The foregoing data establish that protein kinase C is a major contributor to the phosphorylation of rhodopsin in vivo. To test whether frog rhodopsin serves as a direct substrate for protein kinase C, urea-stripped membranes from purified frog rod outer segments were incubated with recombinant protein kinase C␣ and Mg 2ϩ -ATP. Fig. 6A shows that rhodopsin in these membranes was phosphorylated effectively by protein kinase C in a Ca 2ϩ -dependent manner. As reported for bovine rhodopsin (14), protein kinase C phosphorylated dark-adapted (stippled  ). After 30 min, retinas were exposed to continuous room lighting (500 lux) for the indicated lengths of time, and 32 P incorporation into rhodopsin was determined as described under "Materials and Methods." Data are expressed as the level of 32 P incorporation into rhodopsin normalized to the amount of rhodopsin in each sample. Data from four separate experiments (n ϭ 34) were compiled. columns) and photoexcited (open columns) rhodopsin equally well. Under the conditions of the assay, we found that approximately 1 mol of phosphate could be incorporated per mol of rhodopsin, although, as the curve in Fig. 6B indicates (it has not reached a plateau), this value underestimates the maximal stoichiometry.
Last, we investigated whether protein kinase C isolated from frog rod outer segments was also an effective kinase for frog rhodopsin. Protein kinase C was partially purified from frog rod outer segments by ion exchange chromatography (Fig. 7A), with Ca 2ϩ /lipid-stimulated activity eluting at approximately 200 mM NaCl. Due to the limited amount of material, we used a microcolumn (50-l volume) and obtained about 45 milliunits of partially pure rod outer segment protein kinase C from 30 retinas. This protein kinase C phosphorylated rhodopsin in a Ca 2ϩ -dependent manner (Fig. 7B), revealing that, like the protein kinase C in bovine outer segments (12,13,36), it is Ca 2ϩ -regulated. DISCUSSION The present results establish that 1) protein kinase C is a major contributor to the phosphorylation of rhodopsin in intact frog retinas; 2) the kinetics of rhodopsin phosphorylation/dephosphorylation differ markedly, depending on whether or not protein kinase C is hyperactivated; and 3) frog rhodopsin is phosphorylated in vitro by protein kinase C.
Protein Kinase C Contributes Significantly to the Phosphorylation of Rhodopsin in Situ-Specific inhibition of protein kinase C by calphostin C provides a unique tool to evaluate the contribution of protein kinase C in the phosphorylation of rhodopsin in intact frog retinas. This inhibitor binds the phorbol ester-binding (C1) domain of protein kinase C and thus inhibits the regulation, rather the catalytic activity, of protein kinase C (34). This inhibitor was shown to have no effect on the activity of rhodopsin kinase in vitro, at concentrations causing about 70% inhibition of PMA-activated protein kinase C. It should be noted that inhibition of protein kinase C by this molecule is light-dependent and appears to be an oxidation step (37,38). This property served an advantage in our experiments in that inhibition of protein kinase C was initiated by light exposure.
Inhibition of protein kinase C in retinas by calphostin C revealed that this kinase contributes to about half of the phosphorylation of rhodopsin under the conditions studied. Specifically, the extent of rhodopsin phosphorylation in intact retinas was decreased by approximately 50% in the presence of 1 M calphostin C. Similar use of kinase-specific inhibitors to study the ␤-adrenergic signaling pathway suggested that protein kinase A and ␤-adrenergic receptor kinases each contribute 40 -50% toward the desensitization of the ␤2-adrenergic receptor in intact A431 cells at high (Ͼ70%) receptor occupancy (39). These conditions would be analogous to the high levels of rhodopsin photoexcitation in the our calphostin C experiments. In the ␤-adrenergic system, protein kinase A phosphorylation dominated over that of the G protein-coupled receptor kinase at ligand concentrations resulting in Ͻ10% receptor occupancy (39). More recently, the use of inhibitors and dominant negative kinase mutants revealed that protein kinase C and G protein-coupled receptor kinase 2 each contribute equally (about 40 -50%) toward the phosphorylation of a different G protein-coupled receptor, the type 1 angiotensin II receptor, in 293 cells (40). Thus, the phototransduction pathway shares in common with at least two other G protein receptor transduction pathways equal regulation of the receptor by a second messenger-regulated kinase and a G protein-coupled receptor kinase.
Effect of Phorbol Esters on Kinetics of Rhodopsin Phosphorylation in Intact Retinas-Previously we reported that hyperactivation of protein kinase C with phorbol esters had varying effects on the phosphorylation of rhodopsin in intact retinas. To resolve why under some conditions phorbol ester treatment caused an increase in rhodopsin phosphorylation and in other cases a decrease, we undertook here a systematic study of the kinetics of the phorbol ester effect at different levels of rhodopsin photoexcitation. Examination of the kinetics of rhodopsin phosphorylation revealed that PMA had a similar effect, irrespective of the amount of photoexcited rhodopsin. First, PMA caused a progressive increase in rhodopsin phosphorylation, which peaked around 10 -15 min. Then, at later times, PMA caused a progressive decrease in rhodopsin phosphorylation relative to controls, such that by about 30 min the level of rhodopsin phosphorylation in PMA-treated retinas and control retinas was similar. Thus, the overall effect of hyperactivation of protein kinase C is a relatively short-term increase in rhodopsin phosphorylation (Fig. 3, E, F, and G).
The progressive nature of the increase in rhodopsin phosphorylation, at earlier times, in PMA-treated retinas could arise because protein kinase C phosphorylation is slower than that of rhodopsin kinase. In the absence of PMA, the half-time for the light-dependent phosphorylation of rhodopsin was approximately 2 min for the conditions studied, irrespective of stimulus intensity. But in PMA-treated retinas exposed to the higher intensity flash or continuous light, it was evident that the half-time was significantly increased (by at least 2-fold) (Fig. 3,  F and G). Thus, the phosphorylation by protein kinase C appeared to display slower kinetics than that by rhodopsin kinase. By comparison, the phosphorylation of the ␤-adrenergic receptor by protein kinase A is significantly slower relative to that catalyzed by the ␤-adrenergic receptor kinase in permeabilized cells (41).
The progressive decrease in rhodopsin phosphorylation observed after 30 min postillumination in PMA-treated retinas could be accounted for by several mechanisms. First, the protein kinase C phosphorylation site could be more sensitive to dephosphorylation than the rhodopsin kinase site. At least in vitro, the primary sites of phosphorylation of bovine rhodopsin by protein kinase C and rhodopsin kinase differ (14), 3 so that the phosphatase sensitivity of these sites could also differ (see further discussion below). Second, protein kinase C activation could stimulate phosphatase activity. Although Ser/Thr phosphatases directly regulate protein kinase C structure and function (42), the reverse has not been established; nonetheless, phorbol esters have been shown to indirectly modulate the activity of protein phosphatase 1 in intact cells (43). Third, selective binding of secondary proteins such as arrestin to receptor phosphorylated by rhodopsin kinase (44), but not protein kinase C, could decrease the phosphatase accessibility of the rhodopsin kinase-phosphorylated receptor. In this regard, arrestin has been shown to regulate ␤-adrenergic receptor phosphorylated by G protein-coupled receptor kinase but not protein kinase A (45). Whether or not protein kinase C regulates phosphatase activity in photoreceptors remains to be established.
The kinetics of the light-dependent phosphorylation and dephosphorylation of rhodopsin reported here for intact frog retinas are similar to in vivo data on frogs reported by Kü hn (46). Kü hn found that 20-min illumination of dark-adapted frogs resulted in a level of phosphorylation that was constant between 20 and 35 min after the onset of illumination and then decayed with a half time of approximately 30 min. Although these conditions are not identical to those in this report, the kinetics of phosphorylation/dephosphorylation observed under our conditions are remarkably similar.
The kinetics of light-dependent phosphorylation of rhodopsin have been examined more recently in mouse retinas as a function of phosphorylated residue (47). Two residues were found to be phosphorylated in vivo, Ser-334 and Ser-338. In striking contrast, only one of these sites, Ser-338, is a primary phos-FIG. 7. The protein kinase C from frog photoreceptors is Ca 2؉ -dependent. A, protein kinase C was extracted from double sucrose density gradient-purified frog rod outer segments as described under "Materials and Methods." This extract was applied to a 50-l DEAE-Sephacel column, and protein kinase C was eluted by a step gradient of 0 -0.5 M NaCl. Protein kinase C activity in each fraction toward phosphorylation of the protein kinase C-selective peptide (27) was measured with phosphatidylserine/diacylglycerol/Triton X-100 mixed micelles and in the presence of 500 M Ca 2ϩ (filled circles) or 1 mM EGTA (open circles), as described under "Materials and Methods." B, partially purified protein kinase C from frog rod outer segments (pooled fractions 7-9 from A; 0.3 units⅐ml Ϫ1 ) was used to phosphorylate frog photoexcited rhodopsin (0.5 M) from urea-stripped membranes in the presence or absence of 100 M Ca 2ϩ . Data represent mean Ϯ S.E. phorylation site by rhodopsin kinase in vitro using bovine rhodopsin as the substrate; secondary phosphorylations occur in vitro at Ser-343 and Thr-336 (48 -50). The novel site, Ser-334, can be phosphorylated to only 10% under appropriate pH and temperature conditions in vitro (51), suggesting that rhodopsin kinase is not the physiological regulator of this site. However, this site lies within the primary phosphorylation domain of protein kinase C, using bovine rhodopsin as the substrate (14). Interestingly, the kinetics of phosphorylation at both in vivo sites differ markedly. In response to three continuous flashes of light, both residues are phosphorylated equally; however, Ser-338 is phosphorylated and dephosphorylated significantly more rapidly compared with Ser-334. In contrast, continuous illumination results in significantly greater phosphorylation of Ser-334 than Ser-338, and both sites are dephosphorylated at comparable rates (47). Although the conditions in these in vivo experiments are clearly different from ours, it is noteworthy that in our experiments PMA caused a progressive increase in rhodopsin phosphorylation relative to the control, suggesting that phosphorylation of the protein kinase C site(s) lagged behind that of the rhodopsin kinase. The slower kinetics of the phorbol ester-simulated phosphorylation would be consistent with regulation of the more slowly phosphorylated Ser-334 site by protein kinase C. Curiously, the dephosphorylation at Ser-334 is slow relative to that of Ser-338, yet PMA appeared to accelerate the dephosphorylation of rhodopsin. If Ser-334 is the protein kinase C site, the possibility that PMA activates a rhodopsin-directed phosphatase could account for the accelerated dephosphorylation in our experiments.
Conclusion-The use of a specific activator and a specific inhibitor of protein kinase C reveals that this kinase is a major contributor to the light-dependent phosphorylation of rhodopsin. Because the phosphorylation by protein kinase C is regulated by second messengers, rather than by substrate conformation, the protein kinase C-catalyzed phosphorylation of rhodopsin may be analogous to the heterologous regulation of other G protein-coupled receptors. Whether protein kinase C in photoreceptors is activated by signals separate from rhodopsin activation, perhaps arising from feedback pathways from elsewhere in the retina, remains to be explored.