cGMP phosphodiesterase of retinal rods is regulated by two inhibitory subunits

Activation and solubilization of the retinal cGMP-specific phosphodiesterase by limited proteolysis. Role of the C-terminal domain of the beta-subunit

The cGMP-specific phosphodiesterase (PDE) of vertebrate retinal rod outer segments (ROS) is a peripheral enzyme activated in vivo by transducin. In vitro artificial activation can be achieved using trypsin. This was described as resulting from degradation of the inhibitory gamma subunit (2 copies/PDE molecule), leaving intact the alpha beta catalytic core. It was, however, observed that trypsin could induce the release of PDE (or solubilization) from the ROS membranes before its activation [Wensel, T. G. & Stryer, L. (1986) Proteins Struct. Funct. Genet. 1, 90-99]. Studying the time course of this solubilization, we were able to purify a trypsin-solubilized PDE still completely inhibited (i.e. with its two gamma subunits bound). The tryptic solubilization of PDE is therefore complete before any functional degradation of the gamma subunits occurs. It was recently suggested that this solubilization could coincide with the cleavage of a C-terminal fragment of the alpha subunit, which can be labeled by methylation of a terminal cysteine residue [Ong, O. C., Ota, I. M., Clarke, S. & Fung, B. K. K. (1989) Proc. Natl Acad. Sci. USA 86, 9238-9242]. We present the following evidence indicating that the C-terminus of the PDE beta subunit is mainly responsible for PDE anchorage to the ROS membrane. (a) The trypsin-solubilized PDE alpha beta gamma 2 has intact blocked N-termini. (b) It is still methylated on PDE alpha. (c) The C-terminus of PDE beta can also be labeled by methylation and its tryptic cleavage coincides well with the PDE solubilization. (d) Sequential cleavage of the alpha and beta polypeptides can also be detected by high-resolution gel electrophoresis: the first cleavage appears on the beta subunit and is completed when cleavage of the alpha subunit begins. The time course for cleavage of the gamma subunits appears to be slower than for the beta subunit and comparable to that of the alpha subunit. Upon longer trypsinization, a 70-kDa polypeptide appears which seems to be a degradation product of PDE beta. Gel-filtration analysis, however, shows that this 70-kDa fragment does not dissociate from the catalytic core.

Phosphodiesterase of cone photoreceptors from the lizard, Anolis carolinensis

Cone and rod photoreceptors utilize cyclic guanosine monophosphate (cGMP) in the light regulation of membrane polarization. The prototype for visual transduction is established for rod photoreceptors, which utilize a cascade of reactions to regulate a cyclic nucleotide phosphodiesterase (PDE) (EC 3.1.4.17) and thereby control the intracellular concentration of cGMP. Although cones appear to utilize a comparable cGMP cascade for their phototransduction, evidence exists that the PDE from cone photoreceptors may be different from that of rods. Dissociated cone photoreceptors, isolated retinas, and cone outer segments from the lizard, Anolis carolinensis, have been used to identify and characterize a PDE enzyme complex that shares several features in common with the rod outer segment (ROS) PDE complex. Immunoadsorption and sodium dodecyl sulfate-polyacrylamide gel electrophoresis have identified a subunit of lizard cone PDE that has an apparent electrophoretic mobility of 84 kDa and a subunit of lizard rod PDE that migrates at approximately 90 kDa. The lizard cone PDE complex is similar in size, extraction, activation, and immunological characteristics to the PDE complex of rod photoreceptors from lizard, bovine, and human retinas. The lizard cone PDE complex, and perhaps that from cone photoreceptors in general, differs from that of ROS in its chromatographic properties on anion-exchange resins. The sharing of physical and activation properties of the rod and cone PDE complex is compatible with the phototransduction process occurring by a similar mechanism in both cell types. The differences in light sensitivity and speed of response may be attributable to features of the individual proteins that form the PDE complexes of rods and cones or to other undisclosed features of the respective cascades.

Complete cDNA sequences of mouse rod photoreceptor cGMP phosphodiesterase alpha- and beta-subunits, and identification of beta'-, a putative beta-subunit isozyme produced by alternative splicing of the beta-subunit gene

We have characterized overlapping cDNA clones encoding cGMP phosphodiesterase (PDE) alpha- and beta-subunits of mouse retinal rod photoreceptors. The open reading frames predict an alpha-subunit of 100 kDa (856 residues), and a beta-subunit of 99 kDa (853 residues). Sequence analysis of two of twelve beta-subunit clones predicts the presence in the retina of an additional PDE, termed beta', which is generated by alternative splicing of the beta-subunit gene. beta' differs from beta only at the C-terminus being 55 residues shorter and lacking the Caax motif found at the C-termini of both the alpha- and beta-subunits. A 300 residue segment thought to contain the active site is present in the C-terminal half of alpha, beta and beta'.

Binding of the gamma-subunit of retinal rod-outer-segment phosphodiesterase with both transducin and the catalytic subunits of phosphodiesterase

The gamma-subunit of retinal rod-outer-segment phosphodiesterase (PDE-gamma) is a multifunctional protein which interacts directly with both of the catalytic subunits of PDE (PDE alpha/beta) and the alpha-subunit of the retinal G (guanine-nucleotide-binding)-protein transducin alpha (T alpha). We have previously reported that the PDE gamma binds to T alpha at residue nos. 24-45 [Morrison. Rider & Takemoto (1987) FEBS Lett. 222, 266-270]. In vitro this results in inhibition of T alpha GTP/GDP exchange [Morrison, Cunnick, Oppert & Takemoto (1989) J. Biol. Chem. 264, 11671-11681]. We now report that the inhibitory region of PDE gamma for PDE alpha/beta occurs at PDE gamma residues 54-87. This binding results in inhibition of either trypsin-solubilized or membrane-bound PDE alpha/beta. PDE gamma which has been treated with carboxypeptidase Y, removing the C-terminus, does not inhibit PDE alpha/beta, but does inhibit T alpha GTP/GDP exchange. Inhibition by PDE gamma can be removed by T alpha-guanosine 5'-[gamma-thio]triphosphate (GTP[S]) addition to membranes. This results in a displacement of PDE gamma, but not in removal of this subunit from the membrane [Whalen, Bitensky & Takemoto (1990) Biochem. J. 265, 655-658]. These results suggest that low levels of T alpha-GTP[S] can result in displacement of PDE gamma from the membrane in vitro as a GTP[S]-T alpha-PDE gamma complex. Further activation by high levels of T alpha-GTP[S] occurs by displacement of PDE gamma from its inhibitory site on PDE alpha/beta, but not in removal from the membrane.

Subspecies of arrestin from bovine retina. Equal functional binding to photoexcited rhodopsin but various isoelectric focusing phenotypes in individuals

Arrestin (also named 48-kDa protein or S-antigen) binds to photoexcited and phosphorylated rhodopsin and thereby prevents activation of cGMP phosphodiesterase (EC 3.1.4.35) by transducin in retinal rods. We report here that retinal arrestin consists of several subspecies (isoelectric points between pH 5.5-6.2), which can be separated by FPLC anion-exchange chromatography and by FPLC chromatofocusing resulting in highly enriched individual subspecies. The entire heterogeneity pattern of arrestin is present in rod outer segments, independently of whether arrestin orginated from the outer or mostly from the inner segment of rod cells. The different subspecies show a similar binding behavior to photoexcited rhodopsin phosphorylated to various degrees and they quench the cGMP phosphodiesterase activity equally well. In the presence of rod outer segment membranes, arrestin is phosphorylated light-dependently by protein kinase C (0.2 mol phosphate/mol arrestin). This implies that the heterogeneity of arrestin is not primarily due to phosphorylation. Arrestin from different individuals exists as four isoelectric focusing patterns which occur with remarkably different frequencies in calf and cattle. The complexity of the IEF pattern does not increase with aging. Distinct subspecies of arrestin may reflect differences in their primary structure, or may result from differentially regulated post-translational modifications in individuals.

Retinal degeneration in the rd mouse is caused by a defect in the beta subunit of rod cGMP-phosphodiesterase

Mice homozygous for the rd mutation display hereditary retinal degeneration and the classic rd lines serve as a model for human retinitis pigmentosa. In affected animals the retinal rod photoreceptor cells begin degenerating at about postnatal day 8, and by four weeks no photoreceptors are left. Degeneration is preceded by accumulation of cyclic GMP in the retina and is correlated with deficient activity of the rod photoreceptor cGMP-phosphodiesterase. We have recently isolated a candidate complementary DNA for the rd gene from a mouse retinal library and completed the characterization of cDNAs encoding all subunits of bovine photoreceptor phosphodiesterase. The candidate cDNA shows strong homology with a cDNA encoding the bovine phosphodiesterase beta subunit. Here we present evidence that the candidate cDNA is the murine homologue of bovine phosphodiesterase beta cDNA. We conclude that the mouse rd locus encodes the rod photoreceptor cGMP-phosphodiesterase beta subunit.

Calcium and the mechanism of light adaptation in vertebrate photoreceptors

Vertebrate photoreceptors transduce the absorption of light into a hyperpolarizing change in membrane potential. The mechanism of transduction is becoming fairly well understood and has been shown to occur via a G protein-coupled decrease in cyclic GMP. Attention is now turning to the way the enzymatic machinery in the outer segment of the photoreceptor cell is modulated during light adaptation. Recent studies show that light adaptation cannot occur if changes in the concentration of cytoplasmic free calcium in the outer segment are prevented, suggesting that calcium functions as a second messenger in sensitivity regulation.

Evidence for reduced binding of cyclic GMP to cyclic GMP phosphodiesterase in photoreceptors of mice heterozygous for the rd gene

The binding of radiolabelled cGMP to rod outer segment proteins has been investigated in mice, heterozygous for the recessive rd gene that leads to rod dysplasia. Two binding sites were detected, by Scatchard analysis, in a crude cGMP phosphodiesterase fraction, extracted with an EDTA wash from outer segments. Affinities were normal but the capacity of both was reduced 25-35%. Photoaffinity labelling with 3H-cGMP, followed by SDS PAGE and fluorography, suggested that cGMP PDE was the principal binding component in the extracts. If the finding reflects cGMP binding in situ, it might explain the 30-40% lower than normal level of cGMP found in the +/rd retina. Visual pigment has been regenerated in isolated normal and heterozygotic retinas by the application of active isomers of cis-retinal, and the time course of cGMP recovery to 'dark-adapted' levels monitored. The increase in the concentration of cGMP was significantly delayed as compared to that of rhodopsin. No differences in time course or kinetics of recovery were discerned between the two genotypes.

Subunit stoichiometry of retinal rod cGMP phosphodiesterase

The cyclic GMP phosphodiesterase of the retinal rod is composed of three distinct types of polypeptides: alpha (90 kDa), beta (86 kDa), and gamma (10 kDa). The gamma subunit has been shown to inhibit phosphodiesterase activity associated with alpha and beta. To investigate the subunit stoichiometry of the retinal phosphodiesterase, we have developed a panel of monoclonal and peptide antibodies that recognize individual phosphodiesterase subunits. By quantitative and immunochemical analysis of the purified subunits, we have shown that each phosphodiesterase molecule contains one copy each of alpha and beta subunit and two copies of gamma subunit. Moreover, gamma can be chemically cross-linked to both alpha and beta, but not to itself, suggesting that alpha and beta may each bind one gamma. The phosphodiesterase is fully activated when both copies of gamma were removed by proteolysis with trypsin. Upon recombination of the purified gamma subunit with the trypsin-activated phosphodiesterase containing alpha beta, the alpha beta gamma 2 stoichiometry is once again restored, with concomitant total inhibition of activity. Our results suggest that at least two activated transducin molecules are required to fully activate one molecule of phosphodiesterase in retinal rods.

Activation mechanism of retinal rod cyclic GMP phosphodiesterase probed by fluorescein-labeled inhibitory subunit

The cyclic GMP phosphodiesterase (PDE) of vertebrate retinal rod outer segments (ROS) is kept inactive in the dark by its gamma subunits and is activated following illumination by the GTP form of the alpha subunit of transducin (T alpha-GTP). Recent studies have shown that the stoichiometry of the inhibited holoenzyme is alpha beta gamma 2. T alpha-GTP and gamma act reciprocally. We have investigated the activation mechanism using fluorescein-labeled gamma subunit (gamma F) as a probe. gamma F containing a single covalently attached fluorescein was prepared by reaction of PDE with 5-(iodoacetamido)fluorescein and purification by reversed-phase high-pressure liquid chromatography (HPLC). gamma F, like native gamma, inhibits the catalytic activity of trypsin-activated PDE and transducin-activated PDE. Inhibition by gamma F was overcome by further addition of T alpha-GTP. gamma F binds very weakly to ROS membranes stripped of PDE and other peripheral membrane proteins. gamma F added to ROS membranes became incorporated into a component that could be extracted with a low ionic strength buffer. HPLC gel filtration showed that gamma F became part of the PDE holoenzyme. Incorporation occurred in less than 1 min in the presence of light and GTP, but much more slowly (t1/2 approximately 500 s) in the absence of GTP. This result indicates that transducin activates PDE by binding to the holoenzyme and accelerating the dissociation of gamma from the inhibitory sites. The binding of gamma F to trypsin-activated PDE alpha beta was monitored by steady-state emission anisotropy measurements and compared with PDE activity.(ABSTRACT TRUNCATED AT 250 WORDS)

The effect of the gamma-subunit of the cyclic GMP phosphodiesterase of bovine and frog (Rana catesbiana) retinal rod outer segments on the kinetic parameters of the enzyme

Rod-outer-segment cyclic GMP phosphodiesterase (PDE) (subunit composition alpha beta gamma 2) contains catalytic activity in alpha beta. The gamma-subunits are inhibitors. Removal of the gamma-subunits increases Vmax. without affecting the Km. The inhibitory effect of a single gamma-subunit (alpha beta gamma) on the Vmax. of alpha beta is much greater in bovine than in frog (Rana catesbiana) PDE. Bovine PDE in the alpha beta gamma 2 state has a Vmax. that is 2.6 +/- 0.4% of the Vmax. of alpha beta. The removal of one gamma-subunit to give alpha beta gamma results in a Vmax. 5.2 +/- 1% of that for maximal activity. Frog alpha beta gamma 2 has a Vmax. 10.8 +/- 2%, and alpha beta gamma has a Vmax. 50 +/- 18%, of the Vmax. of alpha beta. These data suggest that a single gamma-subunit can inhibit the catalytic activity of active sites on both alpha- and beta-subunits in bovine, but not in frog, rod-outer-segment PDE.

Bovine cone photoreceptor cGMP phosphodiesterase structure deduced from a cDNA clone

A full-length cDNA clone encoding the alpha' subunit of cGMP phosphodiesterase (PDE) from bovine cone photoreceptors was selected by probing a retinal library with a DNA fragment encoding the catalytic core of the rod cGMP PDE alpha subunit. Identity of the clone was confirmed by comparing its deduced sequence with cone PDE peptide sequences determined by Charbonneau et al. [Charbonneau, H., Prusti, R. K., LeTrong, H., Sonnenburg, W. K., Mullaney, P. J., Walsh, K. A. & Beavo, J. A. (1990) Proc. Natl. Acad. Sci. USA, pp. 288-292]. The cone PDE alpha' and the rod PDE alpha and beta subunits are encoded by distinct genes. cGMP PDE subunits share a common ancestry with cAMP PDEs and cyclic nucleotide-binding proteins. Sequence comparisons predict the presence of a catalytic core and possible secondary sites for noncatalytic cGMP binding. The presence of a C-terminal CAAX (Cys-aliphatic-aliphatic-Xaa) motif suggests the cone enzyme may be posttranslationally modified by proteolysis, methylation, and isoprenylation.

Transducin GTPase provides for rapid quenching of the cGMP cascade in rod outer segments

The role of transducin GTPase in rapid cGMP phosphodiesterase quenching was studied by simultaneous registration of GTP hydrolysis and phosphodiesterase activity in the same rod outer segments (ROS) preparation. The results thus obtained allow the conclusion that: (i) phosphodiesterase quenching coincides with transducin-bound GTP hydrolysis independently of ROS concentration; (ii) an increase in the ROS concentration results in the acceleration of cascade quenching due to the existence of a GTPase accelerating mechanism in ROS; (iii) approximation to physiological conditions (protein concentration, temperature) provides a transducin GTPase rate equal to 1-2 turnovers per second i.e., sufficiently high for satisfying the real rate of photoresponse reversion in dark-adapted rods.

Expression in bacteria of functional inhibitory subunit of retinal rod cGMP phosphodiesterase

The cGMP phosphodiesterase of vertebrate retinal rod outer segments plays a key role in visual transduction. A functionally active form of the inhibitory gamma subunit of the phosphodiesterase, which keeps the enzyme inactive in the dark, has been obtained in high yield from a synthetic gene expressed in Escherichia coli. A DNA sequence encoding the 87-residue bovine gamma subunit was chemically synthesized and assembled from 10 oligonucleotides. The synthetic gene was cloned into an expression vector that uses the promoter PL of lambda phage. E. coli was transformed with this vector, which encodes a fusion protein consisting of the first 31 residues of the lambda cII protein, a 7-residue joining sequence that is specifically cleaved at its C-terminal end by clotting protease factor Xa, and the 87-residue gamma subunit. The fusion protein was solubilized in 6 M urea and purified by ion-exchange chromatography on a CM-Sephadex column. The typical yield was 1 mg of fusion protein per liter of bacterial culture, which corresponds to the amount of gamma in about 2500 bovine retinas. Proteolytic cleavage of the fusion protein by factor Xa released a synthetic gamma with the same amino acid sequence as that of native gamma. Both fusion protein and synthetic gamma inhibited trypsin-activated phosphodiesterase with high affinity (Kd less than 100 pM). Likewise, both were as effective as native gamma in inhibiting transducin-activated phosphodiesterase in rod outer segment membranes. This inhibition was reversed by the activation of additional transducin. Thus, the N terminus of gamma is not intimately involved in interactions with either the catalytic subunits of the phosphodiesterase or the activated form of transducin. In contrast, a C-terminal deletion mutant terminating at residue 74 of gamma stimulated rather than inhibited the trypsin-activated enzyme. Thus, the C-terminal region of gamma is critical for inhibition of the phosphodiesterase.

An estimate of the number of G regulator proteins activated per excited rhodopsin in living Limulus ventral photoreceptors

Previous work by others on Limulus photoreceptors has shown that application of a variety of guanine nucleotide-binding regulatory protein (G protein) activators produces discrete waves of depolarization similar to those generated by single photos, but smaller in size. We investigated whether these events might originate at a site other than the G protein. Initiation of the events did not depend on the state of the visual pigment, suggesting that the events do not originate at the pigment level. The events could be blocked by the G-protein blocker guanosine 5'-[beta-thio]diphosphate (GDP[betaS]) and thus support the conclusion that these discrete events are due to the activation of G protein itself. Quantitative measurements indicate that the average size of these events is approximately 8 times smaller than that evoked by single photons under the same conditions. Given certain reasonable assumptions, these results imply that the gain of the first stage of transduction in vivo is approximately 8, a value considerably lower than that measured in vitro in vertebrate rods (gain, 100-500). Furthermore, independent evidence for a low first-stage gain in Limulus is derived from the observation that GDP[betaS] barely affects the size of the response to single photons, but greatly reduces the probability that a photon evokes a response. These results can be explained if rhodopsin normally activates only a few G proteins.

Comparison of the phosphodiesterase inhibitory subunit interactions of frog and bovine rod outer segments

The rod outer segments of the bovine and frog retina possess a cyclic GMP phosphodiesterase (PDE) that is composed of two larger subunits, alpha and beta (P alpha beta), which contain the catalytic activity and a smaller gamma (P gamma) subunit which inhibits the catalytic activity. We studied the binding of P gamma to P alpha beta in both the bovine and frog rod outer segment membranes. Analysis of these data indicates that there are two classes of P gamma binding sites per P alpha beta in both species. The activation of PDE by the guanosine 5'-[gamma-thio]triphosphate form of the alpha subunit of transducin, T alpha.GTP gamma S, was also studied. These data indicate that the two classes of P gamma binding sites contribute to the formation of two classes of binding sites for T alpha.GTP gamma S. We demonstrate solubilization of a portion of the P gamma by T alpha.GTP gamma S in both species. There is also present, in both species, a second class of P gamma which is not solubilized even when it is dissociated from its inhibitory site on P alpha beta by T alpha.GTP gamma S. The amount of full PDE activity which results from release of the solubilizable P gamma is about 50% in the frog PDE but only approx. 17% in the bovine PDE. We also show that activation of frog rod outer segment PDE by trypsin treatment releases the PDE from the membranes. This type of release by trypsin has already been demonstrated in bovine rod outer segments [Wensel & Stryer (1986) Proteins: Struct. Funct. Genet. 1, 90-99].

Light-induced losses and dark recovery rates of guanosine 3',5'-cyclic monophosphate in rod outer segments of intact amphibian photoreceptors

We used an apparatus in which pieces of dark-adapted amphibian retinas (Rana pipiens, Bufo marinus) obtained under infrared illumination were exposed to precise intervals of 500-nm illuminations, and then frozen by contact of their outer segment surface with a liquid helium-cooled copper mirror. Sections of the frozen outer segment layer were obtained in a cryostat and then assayed for total extractable cyclic 3',5'-guanosine monophosphate (cGMP). Significant losses of cGMP with respect to the dark level were evident as early as 60 ms after light onset. With dim subsecond illuminations these losses were surprisingly large, which suggests a previously underestimated magnification in the cGMP cascade, or a transient substantial inhibition of guanylate cyclase activity in combination with increased cyclic GMP phosphodiesterase activity. Within the subsecond period, significant losses that were proportional to light intensity (2-log-unit range) and duration (60-550 ms) were generally not evident. However, losses significantly proportional to these factors became evident with durations of 1 s or longer. When pieces of retina were first illuminated (10 or 60 ms), then held in darkness for increasing periods before freezing, we observed a continuous loss of cGMP during the early postillumination dark period, followed by a recovery of the total cGMP level. The times for recovery to the preillumination level appear to be significantly longer than times reported for the recovery of the photoreceptor membrane potential after similar light exposures.