Binding and activation of rod outer segment phosphodiesterase and guanosine triphosphate binding protein by disc membranes: influence of reassociation method and divalent cations

Attempts to optimize the recovery of light-stimulated phosphodiesterase activity following reassociation of the hypotonically extractable proteins derived from retinal rod segments with hypotonically stripped disc membranes lead to the following observations: the best reassociations were obtained by mixing proteins and stripped disc membranes under hypotonic conditions and slowly increasing the salt concentration; the binding of G-protein and phosphodiesterase to stripped disc membrane occurs in less than 5 minutes and the recovery of light-stimulated phosphodiesterase activation in response to subsaturating stimulus levels requires 2-3 h to plateau. Stripped disc membranes and proteins were reassociated in 'isotonic' buffers containing KCl/NaCl, KCl/NaCl plus Mg2+, or KCl/NaCl plus Ca2+. Large fractional rhodopsin bleaches produced nearly identical light-stimulated phosphodiesterase activities in each of these samples and in the control rod outer segment membranes. Rod outer segment membranes and reassociated stripped disc membrane samples containing divalent cations showed similar phosphodiesterase activities in response to low fractional rhodopsin bleaches (e.g. less than or equal to 0.1%), however, samples devoid of divalent cations during reassociation required rhodopsin bleaches up to 10-fold larger to elicit comparable phosphodiesterase activities. These results suggest that not all phosphodiesterase and/or G-protein molecules bound to the disc membrane surface are equivalent with regard to their efficiency of activation by bleached rhodopsin and that divalent cations can modulate the distribution of G-protein and/or phosphodiesterase between these populations.     

Visual transduction in rod outer segments

The visual transduction cascade of the retinal rod outer segment responds to light by decreasing membrane current. This ion channel is controlled by cyclic GMP which is, in turn, controlled by its synthesis and degradation by guanylate cyclase and phosphodiesterase, respectively. When light bleaches rhodopsin there is an induced exchange of GTP for GDP bound to the alpha subunit of the retinal G-protein, transducin (T). The T alpha.GTP then removes the inhibitory constraint of a small inhibitory subunit (PDE gamma) on the retinal cGMP phosphodiesterase (PDE). This results in activation of the PDE and in hydrolysis of cGMP. Recently both low and high affinity binding sites have been identified for PDE gamma on the PDE alpha/beta catalytic subunits. The discovery of two PDE gamma subunits, each with different binding affinities, suggests that a tightly regulated shut-off mechanism may be present.     

A monoclonal antibody to guanine nucleotide binding protein inhibits the light-activated cyclic GMP pathway in frog rod outer segments

A monoclonal antibody that blocks the light-activated cyclic GMP (cGMP) pathway in frog photoreceptor outer segments (ROS) has been obtained. The antibody (4A) inhibits guanine nucleotide binding to G-protein, the intermediate that links rhodopsin excitation to cGMP phosphodiesterase (PDE), inhibiting light-induced PDE activity as a consequence. Antibody inhibition of the light-activated cGMP pathway is complete at a stoichiometry of approximately one antibody per G-protein in the mixture, which indicates high specificity of the inhibition. Inhibition is more pronounced than that caused by PDE inhibitors such as isobutylmethylxanthine (IBMX) or Ro 20-1724. Antibody 4A has the further effect of inhibiting the phosphorylation of two low molecular weight proteins, components I and II, whose phosphorylation normally can be stimulated by raising cGMP levels. The inhibition is not overridden by adding cGMP, which suggests that the G-protein influences these phosphorylations by a pathway distinct from its action on cGMP concentration. Antibody 4A may prove useful as a probe of the relevance of the cGMP pathway to visual transduction in living photoreceptors. Six other monoclonal antibodies to G-protein, as well as six monoclonal antibodies to rhodopsin and one to PDE, do not block light-activated guanine nucleotide binding, PDE activity, or ROS protein phosphorylations.     

Phosphorylation of rhodopsin and quenching of cyclic GMP phosphodiesterase activation by ATP at weak bleaches

Light and GTP-dependent cyclic GMP phosphodiesterase activation of rod disk membranes is rapidly quenched by ATP. Maximum speed of this effect occurs only with the weakest bleaches. Though it has been proposed that ATP mediates its effect through rapid phosphorylation of bleached rhodopsin, previous workers have found phosphorylation kinetics too slow by more than an order of magnitude to be causal in quenching of cyclic GMP phosphodiesterase activation. In this report, we use preparations retaining more endogenous rhodopsin kinase, higher specific activity ATP, and cyclic GMP phosphodiesterase quenching conditions to show that ATP-dependent multiple phosphorylation of rhodopsin at very weak bleaches (10(-5)) is complete in less than 2 s, easily compatible with cyclic GMP phosphodiesterase quench times of 4 s measured under identical conditions. Thus, it seems likely that previous efforts to achieve high 32P counts by using large bleaches have produced conditions of substrate saturation where much longer times to completion are caused by a very large ratio of substrate to enzyme velocity. Such conditions are not appropriately compared to those that support rapid quenching. We conclude that the speed of rhodopsin phosphorylation is, in fact, adequate to explain ATP quenching of cyclic GMP phosphodiesterase activation.     

Protein complement of rod outer segments of frog retina

Rod outer segments (ROS) from frog retina have been purified by Percoll density gradient centrifugation, a procedure that preserves their form and intactness. One- and two-dimensional electrophoretic analysis reveals a smaller number of proteins than is observed in many cell organelles and permits quantitation of the 20 most abundant polypeptides. Rhodopsin accounts for 70% of the total protein (3 X 10(9) copies/outer segment), and approximately 70 other polypeptides are present at more than 6 X 10(4) copies/outer segment. Another 17% of the total protein is accounted for by the G-protein (3 X 10(8) copies/outer segment) that links rhodopsin bleaching and the activation of cyclic GMP phosphodiesterase (PDE). The phosphodiesterase accounts for 1.5% of the protein (1.5 X 10(7) copies/outer segment), and a 48,000-dalton component that binds to the membrane in the light accounts for a further 2.6%. The function of approximately 90% of the total protein in the outer segment is known, and two-thirds of the non-rhodopsin protein is accounted for by enzyme activities associated with cyclic GMP metabolism. The relative molar abundance of rhodopsin, G-protein, and PDE is 100:10:1. Apart from these major membrane-associated proteins, most of the other proteins are cytosolic. Thirteen other polypeptides are found at an abundance of one or more copies per 1000 rhodopsins, nine soluble and four membrane-bound, and their abundance relative to rhodopsin has been quantitated. ROS have been separated into subcellular fractions which resolve three classes of soluble, extrinsic membrane, and integral membrane proteins. A listing of the proteins that are phosphorylated and their subcellular localization is given. Approximately 25 phosphopeptides are detected, and most are in the soluble fraction. Fewer phosphorylated proteins are associated with the purified outer segments than with crude ROS. Distinct patterns of phosphorylation are associated with intact rods incubated with [32P]Pi and broken rods incubated with [gamma-32P]ATP.     

Sites of arrestin action during the quench phenomenon in retinal rods

The target proteins for arrestin (48 kDa protein) action during the quench of cGMP phosphodiesterase (PDE) activation in retinal rod disk membranes were identified by the use of a cross-linking reagent. A heterobifunctional, cleavable, photo-activatable cross-linker (sulfo-SADP) was coupled to purified arrestin. Under precise weak visible light bleach and nucleotide conditions of quench, the cross-linker was UV flash-activated at a time when quench was well established. The target proteins covalently linked to arrestin by cross-linker activation were identified by immunoblotting. In the presence of ATP arrestin cross-linked to both PDE and rhodopsin during the quench phenomenon. Removal of ATP from the reaction mixture essentially abolished the cross-link with PDE, just as ATP omission abolishes quench, but significantly increased the cross-link to rhodopsin. The absence of a cross-link to the plentiful beta-subunit of transductin, as well as the results of competition studies employing arrestin without attached cross-linker, suggest that the observed cross-links are specific and reflect true binding interactions of arrestin during quench. The data are consistent with a model of quench in which photolyzed rhodopsin (R*) catalyzes the formation of an activated form of arrestin, which dissociates from R* in the presence of ATP, and binds to PDEs, thereby deactivating them.     

Rhodopsin kinase prepared from bovine rod disk membranes quenches light activation of cGMP phosphodiesterase in a reconstituted system

Rhodopsin kinase was extracted into a buffer containing 200 mM KCl and no MgCl2. The activity of the enzyme was stabilized with the use of a mixture of protease inhibitors, aprotinin, benzamidine, leupeptin, and pepstatin. The extract consisted of three major proteins of molecular weight (Mr) 65,000, 56,000, and 37,000, of which the Mr 65,000 protein was identified with the kinase activity since preparations containing the other proteins had no kinase activity and the Mr 65,000 protein was phosphorylated when the extract was incubated with ATP. A reconstituted cGMP phosphodiesterase (PDE) system consisting of peripheral protein-depleted rod disk membranes (RDM), GTP binding protein (G-protein), and PDE was used to test the effectiveness of the rhodopsin kinase preparation in mediating the ATP-dependent quench of light activation of PDE. In the absence of kinase, light-activated PDE activity lasted several minutes. In its presence, ATP and to a lesser extent GTP quenched the activation about as rapidly as in rod disk membranes. The influence of kinase was unaffected by increasing G-protein or PDE content of the reconstituted system but was slowed down by brighter flashes, showing that quench was caused by the inactivation of bleached rhodopsin and not of PDE or G-protein.     

The dynamics of phosphodiesterase activation in rods and cones

Phototransduction starts with the activation of a rhodopsin (respectively, coneopsin) molecule, located in the outer segment of rod (respectively, cone) photoreceptors. The subsequent amplification pathway proceeds via the G-protein transducin to the activation of phosphodiesterase (PDE), a G-protein coupled effector enzyme. In this article, we study the dynamics of PDE activation by constructing a Markov model that is based on the underlying chemical reactions including multiple rhodopsin phosphorylations. We derive explicit equations for the mean and the variance of activated PDE. Our analysis reveals that a low rhodopsin lifetime variance is neither necessary nor sufficient to achieve reliable PDE activation. The numerical simulations show that during the rising phase the variability of PDE activation is much lower compared to the recovery phase, and this property depends crucially on the transducin activation rates. Furthermore, we find that the dynamics of the activation process greatly differs depending on whether rhodopsin or PDE deactivation limits the recovery of the photoresponse. Finally, our simulations for cones show that only very few PDEs are activated by an excited photopigment, which might explain why in S-cones no single photon response can be observed.     

Oxidative stress employs phosphatidyl inositol 3-kinase and ERK signalling pathways to activate cAMP phosphodiesterase-4D3 (PDE4D3) through multi-site phosphorylation at Ser239 and Ser579

RAW macrophages, which express the PDE4D3 and PDE4D5 cAMP phosphodiesterase isoforms, exhibited increased PDE4 activity when challenged with H2O2 in a fashion that was negated by treatment with the cell permeant antioxidant, N-acetyl cysteine and by diphenyleneiodonium chloride, an inhibitor of NADPH oxidase. In Cos1 cells transfected to express PDE4D3, challenge with H2O2 caused a rapid increase in both the activity and phosphorylation of PDE4D3. Lysates from H2O2-treated COS cells caused the phosphorylation of purified, recombinant PDE4D3 at two sites. One was the established ERK phosphorylation site at Ser579, located at the extreme C-terminus of the catalytic unit, and the other was a novel site at Ser239, located at the extreme N-terminus of the catalytic unit. Double Ser239Ala:Ser579Ala mutation of PDE4D3 prevented its H2O2-dependent phosphorylation both in vitro and in intact COS cells. Phosphorylation of PDE4D3 at Ser579 was ablated by treating COS cells with the MEK inhibitor, PD98059, which also negated activation. The activity of the Ser239Ala:Ser579Ala double mutant, and the Ser579Ala single PDE4D3 mutant was unaffected by H2O2 challenge of COS cells, whilst the Ser239Ala mutant was inhibited. Wortmannin inhibited the H2O2-dependent phosphorylation of PDE4D3 in COS cells by around 50%, whilst it fully ablated phosphorylation at Ser239 as well as ablating activation of PDE4D3. Neither immunodepletion of p70S6 kinase nor siRNA-mediated knockdown of mTor inhibited the H2O2-dependent phosphorylation of PDE4D3 at Ser239. Activation of PDE4D3 by challenge with H2O2 was not additive with activation through protein kinase A (PKA)-mediated phosphorylation of PDE4D3. Challenge with H2O2 did not alter PKA-mediated phosphorylation of PDE4D3 at Ser54. H2O2 dependent phosphorylation of PDE4D3, at Ser239 and Ser579, did not alter the sensitivity of PDE4D3 to inhibition by the selective PDE4 inhibitor, rolipram. An unknown protein kinase acting downstream of phosphatidyl inositol 3-kinase phosphorylates PDE4D3 at Ser239. This switches the effect of phosphorylation by ERK at Ser579 from inhibition to activation. We propose that phosphorylation at Ser239 attenuates interaction between either UCR2 or the UCR1/UCR2 module and the PDE4 catalytic unit so as to re-programme the functional outcome effect of phosphorylation by ERK. We identify a novel process through which reactive oxygen species activate long PDE4 isoforms so as to reduce cAMP levels and thereby promote inflammatory responses.     

Protein kinase Calpha-induced p115RhoGEF phosphorylation signals endothelial cytoskeletal rearrangement

Heterotrimeric G-proteins of the Galpha12/13 family activate Rho GTPase through the guanine nucleotide exchange factor p115RhoGEF. Because Rho activation is also dependent on protein kinase Calpha (PKCalpha), we addressed the possibility that PKCalpha can also induce Rho activation secondary to the phosphorylation of p115RhoGEF. Studies were made using human umbilical vein endothelial cells in which we addressed the mechanisms of PKCalpha-induced Rho activation and its consequences on actin cytoskeletal changes. We observed that PKCalpha associated with p115RhoGEF within 1 min of thrombin stimulation and p115RhoGEF phosphorylation was dependent on PKCalpha. Inhibition of PKCalpha-dependent p115RhoGEF phosphorylation prevented the thrombin-induced Rho activation, indicating that the response occurred downstream of PKCalpha phosphorylation of p115RhoGEF. The regulator of G-protein signaling domain of p115RhoGEF, a GTPase activating protein for G12/13, also prevented thrombin-induced Rho activation, indicating the parallel requirement of G12/13 in signaling Rho activation via p115RhoGEF. These data demonstrate a pathway of Rho activation involving PKCalpha-dependent phosphorylation of p115RhoGEF. Thus, Rho activation in endothelial cells and the subsequent actin cytoskeletal re-arrangement require the cooperative interaction of both G12/13 and PKCalpha pathways that converge at p115RhoGEF.     

A 48 kDa protein arrests cGMP phosphodiesterase activation in retinal rod disk membranes

Photolyzed rhodopsin (R) catalyzes GTP-binding to alpha-transducins (T alpha); T alpha X GTPs then activate cGMP phosphodiesterase (PDE). PDE activation is arrested by ATP in two ways: (i) initial velocity is suppressed, and (ii) PDE velocity rapidly returns to preactivation levels (turnoff). Arrestin (a 48 kDa protein) markedly enhances turnoff while not affecting initial velocity. Arrestin in the presence of ATP achieves rapid turnoff by directly inhibiting activated PDE, as indicated by its ability to inhibit the direct activation of PDE by T alpha X GMP--PNP (guanylyl-imidodiphosphate). Double reciprocal plots reveal a competition between arrestins and activated transducins for sites on PDE. Blocking R phosphorylation blocks initial velocity suppression but does not disturb rapid turnoff. Our data suggest a 2-fold mechanism for PDE deactivation: (i) formation of T alpha X GTPs is suppressed by R phosphorylation, while (ii) activation of PDE by T alpha X GTPs is competitively inhibited by arrestins when ATP is present.     

Inactivation of photoexcited rhodopsin in retinal rods: the roles of rhodopsin kinase and 48-kDa protein (arrestin)

The inactivation of excited rhodopsin in the presence of ATP, rhodopsin kinase, and/or arrestin has been studied from its effect on the two subsequent steps in the light-induced enzymatic cascade: metarhodopsin II catalyzed activation of G-protein and G-protein-dependent activation of cGMP phosphodiesterase. The inactivation of G-protein (from light-scattering measurements) and that of phosphodiesterase (from measurements of cGMP hydrolysis) have been studied and compared in reconstituted systems containing various combinations of the proteins involved (rhodopsin, G-protein, phosphodiesterase, kinase, and arrestin). Our results show that rhodopsin kinase alone can terminate the activation of G-protein and that arrestin speeds up the process at a relative concentration similar to that reported in the rod (half-maximal effect at 50 nM for 4.4 microM rhodopsin). Measurements of rhodopsin phosphorylation under identical conditions show that in the presence of arrestin total metarhodopsin II inactivation is achieved when only 0.5-1.4 phosphates are bound per bleached rhodopsin, whereas in the absence of arrestin it requires binding of 12-16 phosphates per bleached rhodopsin. Phosphodiesterase activity can similarly be turned off by kinase, and the process is similarly accelerated by arrestin.     

cGMP suppresses GTPase activity of a portion of transducin equimolar to phosphodiesterase in frog rod outer segments. Light-induced cGMP decreases as a putative feedback mechanism of the photoresponse.

In rod photoreceptor cells, the light response is triggered by an enzymatic cascade that causes cGMP levels to fall: excited rhodopsin (Rho*)----rod G-protein (transducin, Gt)----cGMP-phosphodiesterase (PDE). This results in the closure of plasma membrane channels that are gated by cGMP. PDE activation by Gt occurs when GDP bound to the alpha-subunit of Gt (Gt alpha) is exchanged with free GTP. The interaction of Gt alpha-GTP with the gamma-subunits of PDE releases their inhibitory action and causes cGMP hydrolysis. Inactivation is thought to be caused by subsequent hydrolysis of Gt alpha-GTP by an intrinsic Gt-GTPase activity. Here we report that there are two portions of Gt in frog rod outer segments (ROS) expressing different rates of GTP hydrolysis: 19.5 +/- 3 mmol of Gt/mol of Rho, equivalent to that amount which participates in PDE activation, hydrolyzing GTP at a rate of approximately 0.6 turnover/s ("fast") and the remaining Gt (80.5 +/- 3 mmol/mol Rho) hydrolyzing GTP at a rate of 0.058 +/- 0.009 turnover/s. Fast GTPase activity is abolished in the presence of cGMP. This effect occurs over the physiological range of cGMP concentration changes in ROS, half-saturating at approximately 2 microM and saturating at 5 microM cGMP. cGMP-dependent suppression of GTPase is specific for cGMP; cAMP in millimolar concentration does not affect GTPase, while the poorly hydrolyzable cGMP analogue, 8-bromo-cGMP, mimics the effect. GTPase regulation by cGMP is not affected by Ca2+ over the concentration range 5-500 nM, which spans the physiological changes in cytoplasmic Ca2+ in rod cells. We suggest that the fast cGMP-sensitive GTPase activity is a property of the Gt that activates PDE. In this model, cGMP serves not only as a messenger of excitation but also modulates GTPase activity, thereby mediating negative feedback regulation of the pathway via PDE turnoff: a light-dependent decrease in cGMP accelerates the hydrolysis of GTP bound to Gt, resulting in the rapid inactivation of PDE.     

On a soluble system for studying light activation of rod outer segment cyclic GMP phosphodiesterase

Bovine rod outer segment (ROS) cyclic GMP phosphodiesterase (PDE) could be activated about 6-fold by light, an effect that could be simulated by isolated bleached rhodopsin. About 90% of PDE activity in ROS could be extracted with 10 mM Tris-HCl, pH 7.5, but light is ineffective in activating the soluble enzyme. However, bleached rhodopsin could activate it in the presence of a very low concentration of ATP, strongly suggesting the mediation of rhodopsin in the light activation of the enzyme in ROS. Direct evidence is presented to suggest that the phosphorylation of opsin (bleached rhodopsin) is unrelated to the activation of PDE by bleached rhodopsin and ATP. The reconstitution of the light activation of PDE in a soluble system presented here opens up a new direction to future investigations on the mechanism of light regulation of cyclic GMP levels in retina and its implication in the photoreceptor function.     

Effects of fluoride on retinal rod outer segment cGMP phosphodiesterase and G-protein

The effects of fluoride on ROS phosphodiesterase and G-protein have been studied using membrane-free extracts. When G-protein was present NaF, at millimolar concentrations, stimulated PDE activity however, in a G-protein free extract, cGMP hydrolysis was inhibited by high fluoride concentrations. Fluoride was also found to profoundly inhibit the ability of G-protein to bind a GTP analogue, GTP gamma S, both in the presence and absence of rhodopsin. Aluminium greatly modified these effects of fluoride on PDE and G-protein. The possibility that fluoride activates PDE through its effect on G-protein is discussed.     

Kinetics, binding constant, and activation energy of the 48-kDa protein-rhodopsin complex by extra-metarhodopsin II

We have found that the 48-kDa protein (or S-antigen 48k) of the rod photoreceptor enhances the light-induced formation of the photoproduct metarhodopsin II (MII) from prephosphorylated rhodopsin. The effect is analogous to the known enhancement of MII (extra-MII) that results from selective interaction of MII with G-protein. We have determined some parameters of the MII-48k interaction by measuring the extra-MII absorption change induced by the 48-kDa protein. The amplitude saturation yields a dissociation constant for the MII-48k complex on the order of 50 nM. At the technical limit of these measurements, 13.7 degrees C and 12 microM 48-kDa protein, we find a rate of 2.3 s-1 for formation of the 48k-MII complex. Extrapolation of these values to cellular conditions yields an occupation time of phosphorylated MII by 48k less than 200 ms. This is short compared to estimated rates of phosphorylation. The temperature dependence of the MII-48k formation rate is very high (Q10 for 5 degrees C/15 degrees C = 9-10). The related Arrhenius activation energy (165 kJ mol-1) is correspondingly high and indicates a considerable transient chemical change during the binding process.     

Light activation of one rhodopsin molecule causes the phosphorylation of hundreds of others. A reaction observed in electropermeabilized frog rod outer segments exposed to dim illumination

A rhodopsin phosphorylation reaction that occurs with high-gain is observed if measurements are made in electropermeabilized frog rod outer segments (ROS) stimulated by a dim flash of light in the operating range of the photoreceptor. Flashes of light exciting 1000 or fewer of the 3 x 10(9) rhodopsins present/ROS results in the incorporation of 1400 phosphates from ATP into the rhodopsin pool for each excited rhodopsin (Rho*). This amplification decreases with increasing light intensity, falling most sharply after each disk has absorbed one photon. The high-gain reaction is lost if the ROS are broken into vesicles by shearing, leaving a low-gain rhodopsin phosphorylation characterized in previous studies using brighter illumination. The high-gain but not the low-gain phosphorylation appears to be regulated by G-protein and by calcium levels in the range over which intracellular calcium changes when rod photoreceptors are illuminated. Kinetic measurements made on the phosphorylation observed at higher light intensities shows that it initially occurs rapidly enough for a role in terminating the photoresponse. The high-gain phosphorylation observed at lower light intensities may play a global role in regulating light-adaptation of the rod photoreceptor, and its existence suggests that a search for a similar high-gain modification in systems using the homologous beta-adrenergic or muscarinic acetylcholine receptors might be rewarding.     

Contribution of the guanosinetriphosphatase activity of G-protein to termination of light-activated guanosine cyclic 3',5'-phosphate hydrolysis in retinal rod outer segments

Light activation of GTP binding to G-protein and its eventual hydrolysis are hypothesized to lead to activation and inactivation of cGMP phosphodiesterase (PDE) in vertebrate rod disk membranes (RDM). However, the reported GTPase rate of 3 per minute is too slow to account for the observed rapid inactivation of PDE. Our investigations on GTPase activity showed that RDM isolated in the dark have considerable dark GTPase activity, which is enhanced by light. In dark and light, the enzyme exhibits biphasic substrate dependence with two Km's for GTP of 2-3 and 40-80 microM at 22 degrees C and less than 1 and 10-25 microM at 37 degrees C. The Km's were not influenced by light. On the basis of G-protein content of the RDM, the Vmax's for the two activities at 37 degrees C in light are 4-5 and 20-30 GTPs hydrolyzed per minute per G-protein. RDM washed free of soluble and peripheral proteins do not have measurable GTPase activity in the dark or light. Purified G-protein alone also did not turn over GTP, apparently because bleached rhodopsin is required for it to bind GTP. Reconstitution of washed membranes with purified G-protein restores both the low- and high-Km GTPase activities. Inactivation of G-protein as measured by PDE turnoff and dissociation signal recovery is found to be faster at higher than lower [GTP], consistent with the observation that the higher GTPase activity associated with the higher Km alos resides in the G-protein.(ABSTRACT TRUNCATED AT 250 WORDS)     

Characterization of rhodopsin kinase purified from bovine rod outer segments

Rhodopsin kinase was purified by sequential chromatography on DEAE-cellulose and blue-Sepharose. Kinase activity co-purified with a 62-kDa polypeptide, which bound light-dependently in the absence of ATP to purified vesicle-reconstituted rhodopsin. Purified rhodopsin kinase is free of any detectable arrestin or the retinal G-protein. Rhodopsin kinase is autophosphorylated on serine residues which is unaffected by the presence of bleached rhodopsin and results in a transition in molecular mass to 64 kDa. Autophosphorylation of the kinase did not appear to alter the overall rate of rhodopsin phosphorylation or the apparent KM (0.6 microM) for purified reconstituted rhodopsin. Peptides corresponding to sequences within opsin loops 3-4 and 5-6 and the COOH terminus inhibited kinase phosphorylation of bleached rhodopsin, suggesting at least three potential sites to account for the stable high affinity binding of rhodopsin kinase to the bleached photoreceptor molecule that are at least in part distinct from the substrate sites for phosphorylation. These sequences are similar to those proposed for receptor recognition of G-proteins and indicate that the domains involved in light-dependent binding of rhodopsin kinase and retinal G-protein are similar or overlapping.