Cyclic nucleotide phosphodiesterases: relating structure and function.

Cyclic nucleotide phosphodiesterases (PDEs) comprise a superfamily of metallophosphohydrolases that specifically cleave the 3',5'-cyclic phosphate moiety of cAMP and/or cGMP to produce the corresponding 5'-nucleotide. PDEs are critical determinants for modulation of cellular levels of cAMP and/or cGMP by many stimuli. Eleven families of PDEs with varying selectivities for cAMP or cGMP have been identified in mammalian tissues. Within these families, multiple isoforms are expressed either as products of different genes or as products of the same gene through alternative splicing. Regulation of PDEs is important for controlling myriad physiological functions, including the visual response, smooth muscle relaxation, platelet aggregation, fluid homeostasis, immune responses, and cardiac contractility. PDEs are critically involved in feedback control of cellular cAMP and cGMP levels. Activities of the various PDEs are highly regulated by a panoply of processes, including phosphorylation events, interaction with small molecules such as cGMP or phosphatidic acid, subcellular localization, and association with specific protein partners. The PDE superfamily continues to be a major target for pharmacological intervention in a number of medically important maladies.

Phosphorylation and activation of cGMP-dependent protein kinase by Src.

Using information obtained from experiments with peptide substrates of v-Src, a motif within the cGMP-binding domain of cGMP-dependent protein kinase (cGK) was identified as a potential phosphorylation site for v-Src. Here we show that the purified Ialpha isozyme of cGK is phosphorylated stoichiometrically and in a time-dependent manner by purified Src in vitro. The kinase activity of cGK is elevated approximately 4-fold (relative to autophosphorylated cGK) or 10-fold (relative to unphosphorylated cGK) upon tyrosine phosphorylation by Src. Phosphorylation of cGK by v-Src produces modest effects on the cGMP-binding properties and dissociation rates of cGK, and reduces the kact for cGMP. We hypothesize that the mechanism of activation may involve coupling of the cGMP binding domain to the catalytic domain.

High biochemical selectivity of tadalafil, sildenafil and vardenafil for human phosphodiesterase 5A1 (PDE5) over PDE11A4 suggests the absence of PDE11A4 cross-reaction in patients.

The physiological role of phosphodiesterase (PDE)11 is unknown and its biochemical characteristics are poorly understood. We have expressed human His-tagged PDE11A4 and purified the enzyme to apparent homogeneity. PDE11A4 displays K(m) values of 0.97 microM for cGMP and 2.4 microM for cAMP, and maximal velocities were 4- to 10-fold higher for cAMP than for cGMP. Given the homology between PDE11 and PDE5, we have compared the biochemical potencies of tadalafil (Cialis, Lilly-ICOS), vardenafil (Levitra, Bayer-GSK), and sildenafil (Viagra, Pfizer Inc.) for PDE11A4 and PDE5A1. PDE5A1/PDE11A4 selectivities are 40-, 9300-, and 1000-fold for tadalafil, vardenafil, and sildenafil, respectively. This suggests that none of these three compounds is likely to crossreact with PDE11A4 in patients.

Regulation of phosphoenolpyruvate carboxykinase gene transcription in H4IIE hepatoma cells: evidence for a primary role of the catalytic subunit of 3',5'-cyclic adenosine monophosphate-dependent protein kinase.

The purpose of these studies was to determine whether the catalytic subunit of cAMP-dependent protein kinase is involved in the regulation of P-enolpyruvate carboxykinase (PEPCK) gene transcription. Cyclic AMP analog pairs that preferentially stimulate either type I or type II protein kinase in a synergistic manner were used to compare regulation of mRNAPEPCK synthesis in H4IIE rat hepatoma cells with protein kinase activation in vitro. Type II protein kinase is predominant in H4IIE cells and analog pairs directed toward this isozyme resulted in a synergistic increase of mRNAPEPCK that was due to a corresponding enhancement of PEPCK gene transcription. When compared to a single analog the addition of a type II-directed analog pair reduced the total analog concentration required for maximal induction of transcription by about 30-fold. H4IIE cells have a small amount of type I kinase; pairs specific for this form of the enzyme were also effective, but to a lesser extent than those for the type II kinase. (Rp)-cAMPS, a cyclic nucleotide-dependent protein kinase antagonist, inhibited the agonist-induced increase of mRNAPEPCK in a concentration-dependent manner. The results indicate that the activation of PEPCK gene transcription by cAMP in H4IIE cells is mediated by cAMP-dependent protein kinase. Although the type II isozyme is primarily responsible, type I is also effective. These isozymes have identical catalytic subunits, hence this component presumably mediates the cAMP effect.

Characterization of the isolated cAMP-binding B domain of cAMP-dependent protein kinase.

A 14.4-kDa cAMP-binding fragment was generated during bacterial expression and purification of recombinant bovine cAMP-dependent protein kinase type I alpha regulatory subunit (RI alpha). The full-length RI alpha from which the fragment was derived contained a point mutation allowing its B domain to bind both cAMP and cGMP with high affinity while leaving its A domain highly cAMP selective. The NH2 terminus of the fragment was Ser-252, indicating that it encompassed the entire predicted B domain. Although the [3H]cAMP and [3H]cGMP exchange rates of the isolated B domain were increased relative to the B domain in intact RI alpha, the [3H]cAMP exchange rate was comparable to that of the B domain of full-length RI alpha containing an unoccupied A domain. A plasmid encoding only the isolated B domain was overexpressed in Escherichia coli, and a monomeric form of the B domain was purified that had identical properties to the proteolytically generated fragment, indicating that all of the elements for the high-affinity cAMP-binding B domain are contained within the 128 amino acid carboxyl terminus of the R subunit. Prolonged induction of the B domain in E. coli or storage of the purified protein resulted in the formation of a dimer that could be reverted to the monomer by incubation in 2-mercaptoethanol. Dimerization caused an approximate fivefold increase in the rate of cyclic nucleotide exchange relative to the monomer. The results show that an isolated cAMP-binding domain can function independently of any other domain structures of the R subunit.

Isolation and characterization of cDNAs encoding PDE5A, a human cGMP-binding, cGMP-specific 3',5'-cyclic nucleotide phosphodiesterase.

Human cGMP-binding, cGMP-specific 3',5'-cyclic nucleotide phosphodiesterase (PDE5A) cDNAs were isolated. A 3.1-kb composite DNA sequence assembled from overlapping cDNAs encodes an 875-amino-acid protein with a predicted molecular mass of 100012 Da (PDE5A1). Extracts prepared from yeast expressing human PDE5A1 hydrolyzed cGMP. This activity was inhibited by the selective PDE5 inhibitors zaprinast and DMPPO. PDE5A mRNA is expressed in aortic smooth muscle cells, heart, placenta, skeletal muscle and pancreas and, to a much lesser extent, in brain, liver and lung. A 5'-splice variant, PDE5A2, encodes an 833-amino-acid protein with eight unique amino acids at the amino terminus. PDE5A maps to chromosome 4q 25-27.

Phosphorylation of smg p21B/rap1B p21 by cyclic GMP-dependent protein kinase.

smg p21B/rap1B p21, a member of ras p21-like small GTP-binding protein superfamily, has been shown to be phosphorylated by cyclic AMP-dependent protein kinase (protein kinase A). We show here that this protein was also phosphorylated by cyclic GMP-dependent protein kinase (protein kinase G) in a cell-free system. The same serine residue (Ser179) in the C-terminal region was phosphorylated by both protein kinases G and A. The Km and Vmax values of smg p21B for protein kinase G were 5 x 10(-7) M and 4 x 10(-9) mol/min/mg, and those values for protein kinase A were 1 x 10(-7) M and 3 x 10(-8) mol/min/mg.

cAMP analogs used to study low-Km, hormone-sensitive phosphodiesterase.

The determination of cyclic nucleotide analog I50 values for phosphodiesterases is a relatively simple method to study interactions between the enzyme and cyclic nucleotide analogs. This approach allows a large number of derivatives to be tested for preliminary information concerning hydrolysis. To conclude that the I50 values is a measure of analog hydrolysis requires that the mechanism of inhibition of [3H]cAMP hydrolysis is competitive. It is possible that some analogs act as noncompetitive inhibitors. Provided the enzyme preparation is pure with respect to phosphodiesterases, the type of inhibition can be determined. When it is important to determine if an analog is hydrolyzed, the complementary method of measuring direct hydrolysis can be used. For the low-Km phosphodiesterase and the analogs studied here, relatively low I50 values are correlated with analog hydrolysis while relatively high I50 values are correlated with the absence of detectable hydrolysis. For analogs such as N6-benzoyl- and N6-monobutyryl-cAMP the method of determining I50 values provides information that is not obtainable by direct hydrolysis. For example, neither of these analogs appear to be hydrolyzed but N6-benzoyl-cAMP has a lower I50 value and therefore more readily interacts with the low-Km phosphodiesterase. This analog or other ones may be useful in cAMP analog affinity chromatography for purification of phosphodiesterases. The method for directly determining cAMP analog hydrolysis measures the disappearance of the substrate instead of appearance of the product. However, the method is very sensitive since some, but not all, cAMP analogs have lower activation constants than cAMP does. Therefore, analogs can be tested for hydrolysis at concentrations as low as 10-50 nM and small changes in analog concentration can be directed. The methods presented have not only provided information concerning the mechanisms and structural requirements for hydrolysis but the analog specificities for various phosphodiesterases can be used as one of the determinants of analog potency in intact cells. Furthermore, the correlation of analog I50 values as an indication of hydrolysis with the effects of insulin on analog-stimulated intact cell responses provides information concerning the mechanism of insulin action. Pitfalls. Since cAMP analog preparations may be contaminated with cAMP, it is advantageous to purify the analog before determining direct hydrolysis. A method using Sephadex G-25 is presented elsewhere in this volume.(ABSTRACT TRUNCATED AT 400 WORDS)

Tissue distribution of phosphodiesterase families and the effects of sildenafil on tissue cyclic nucleotides, platelet function, and the contractile responses of trabeculae carneae and aortic rings in vitro.

Sildenafil is a selective inhibitor of phosphodiesterase type 5 (PDE5), which has been shown to be a clinically effective treatment for erectile dysfunction. Its action results from increased levels of cyclic guanosine monophosphate (cGMP), which is normally degraded by PDE5. This cyclic nucleotide is a second messenger for nitric oxide, which is involved in the regulation of numerous functions, including vascular smooth muscle tone. In an attempt to better predict the effects of sildenafil on cardiovascular function, the distribution of PDE activity was determined with anti-PDE1 and anti-PDE5 antibodies in the human cardiac ventricle and saphenous vein, and in vitro studies were performed on the isolated human cardiac ventricle, corpus cavernosum, saphenous vein, and mesenteric artery as well as on rabbit aorta, dog coronary artery, dog trabecular tissue, and rabbit and human platelets. The major PDE activity in the human cardiac ventricle was shown to be calcium/calmodulin-dependent PDE1, but there was no detectable level of PDE5. In contrast, the human saphenous vein contained PDEs 1, 4, and 5, and the human mesenteric artery contained PDEs 1, 2, 3, 4, and 5. The distribution of PDE5 in the cardiovascular system is consistent with the observed pharmacodynamic and clinical effects of sildenafil. Sildenafil, unlike milrinone, a selective PDE3 inhibitor, had no effect on the isolated trabeculae carneae; this is consistent with the lack of PDE5 expression in cardiac myocytes. Sildenafil selectively increased cGMP levels in coronary vascular smooth muscle tissue but produced no change in cyclic adenosine monophosphate (cAMP) levels, which is consistent with the drug's selectivity for PDE5. In phenylephrine-contracted isolated rabbit aortic rings, sildenafil enhanced the relaxation induced by the nitric oxide donor glyceryl trinitrate, suggesting that sildenafil may potentiate the hypotensive effects of nitric oxide donor agents on the vasculature, an effect that has been observed clinically. Human platelets were found to contain PDE5, which was inhibited by 50% (IC50) by sildenafil at a concentration of 6.3 nM, consistent with the IC50 value in the corpus cavernosum. Sildenafil alone had no direct effect on platelet function, but it potentiated the in vitro antiaggregatory activity of sodium nitroprusside on rabbit and human platelets. The pharmacodynamic and adverse event profiles observed in clinical trials with sildenafil are consistent with the in vitro profile of the tissue distribution of PDE5 and its known mechanism of action as a selective inhibitor of PDE5.

A photoaffinity probe covalently modifies the catalytic site of the cGMP-binding cGMP-specific phosphodiesterase (PDE-5).

The cGMP-binding cGMP-specific phosphodiesterase (PDE-5) contains distinct catalytic and allosteric binding sites, and each is cGMP-specific. Cyclic nucleotide phosphodiesterase inhibitors, such as 3-isobutyl-1-methylxanthine (IBMX), are believed to compete with cyclic nucleotides at the catalytic sites of these enzymes, but the portion of PDE-5 that accounts for interaction of either of these inhibitors of the substrates themselves with the catalytic domain of the enzymes has not been identified. IBMX was derivatized to yield the photoaffinity probe 8([3-125I,-4-azido]-benzyl)-IBMX, which is referred to as 8(125IAB)-IBMX. This probe was incubated with partially purified recombinant bovine PDE-5. After UV irradiation and SDS-PAGE, a single radiolabeled band that coincided with the position of PDE-5 was visualized on the gel, and the photoaffinity labeling of PDE-5 was linear with increasing concentration of the 8(125IAB)-IBMX. Prominent Coomassie blue-stained bands other than PDE-5 were not labeled significantly. The photoaffinity labeling was progressively blocked by cGMP at concentrations higher than 10 microM, whereas cAMP or 5'-GMP exhibited only weak inhibitory effects. Other compounds that are believed to interact with the PDE-5 catalytic site, including IBMX, cIMP, and beta-phenyl-1,N2-etheno-cGMP (PET-cGMP), also inhibited the photoaffinity labeling in a concentration-dependent manner. The IC50 of PET-cGMP for inhibition of photoaffinity labeling was 10 microM, which compared favorably with an IC50 of 5 microM for inhibition of PDE-5 catalytic activity by this compound. It is concluded that the interaction of this photoaffinity probe with PDE-5 is highly specific for the catalytic site over the allosteric binding sites of PDE-5 and could prove useful in studies to map the catalytic site of PDE-5.

Rat adipose tissue cAMP-dependent protein kinase: a unique form of type II.

The rat adipose tissue cAMP-dependent protein kinase type II holoenzyme and regulatory (R) subunit were compared with type II from bovine heart and several other species and tissues. Adipose tissue type II was similar to the bovine heart type II by several criteria (S 20,W = 7.0, site 1 and site 2 dissociation rates for [3H]cAMP, rapid autophosphorylation and lack of MgATP inhibition of [3H]cAMP binding). However, some of its physical characteristics were similar to type I. The apparent molecular weight determined by SDS gel electrophoresis of the homogeneous adipose tissue R subunit was 51000 daltons compared to 49000 for type I and 53000-58000 for other type II R subunits. The adipose tissue holoenzyme eluted from DEAE-cellulose at an intermediate position between type I and bovine heart type II. The adipose tissue and bovine heart holoenzymes differed in several properties including Stokes radius (5.2 nm vs. 6.0 nm), calculated molecular weight (157000 vs. 181000 daltons) and frictional ratio (1.47 vs. 1.60). After autophosphorylation the adipose tissue R subunit, like type IIB forms from other species and tissues, did not shift to a higher apparent molecular weight on SDS gel electrophoresis like bovine heart type IIR subunit (a type IIA form). Even though the adipose tissue enzyme was quite similar to other type II forms in the kinetics of cAMP action, the cAMP binding sites could be shown to be different from them by the use of cAMP analogs. cAMP analogs modified at the N6 position of the adenine ring, such as N6-benzoyl-cAMP, had higher apparent Ka values for protein kinase activation for the adipose tissue enzyme than for the bovine and several other heart isozymes. cAMP analogs modified at the 8 carbon of the adenine ring showed positive cooperativity of activation for the adipose tissue enzyme but not for the bovine heart holoenzyme. The adipose tissue isozyme is the first type II form described to have a distinct kinetic characteristic.

Mechanisms of action of PDE5 inhibition in erectile dysfunction.

A spinal reflex and the L-arginine-nitric oxide-guanylyl cyclase-cyclic guanosine monophosphate (cGMP) pathway mediate smooth muscle relaxation that results in penile erection. Nerves and endothelial cells directly release nitric oxide in the penis, where it stimulates guanylyl cyclase to produce cGMP and lowers intracellular calcium levels. This triggers relaxation of arterial and trabecular smooth muscle, leading to arterial dilatation, venous constriction, and erection. Phosphodiesterase 5 (PDE5) is the predominant phosphodiesterase in the corpus cavernosum. The catalytic site of PDE5 normally degrades cGMP, and PDE5 inhibitors such as sildenafil potentiate endogenous increases in cGMP by inhibiting its breakdown at the catalytic site. Phosphorylation of PDE5 increases its enzymatic activity as well as the affinity of its allosteric (noncatalytic/GAF domains) sites for cGMP. Binding of cGMP to the allosteric site further stimulates enzymatic activity. Thus phosphorylation of PDE5 and binding of cGMP to the noncatalytic sites mediate negative feedback regulation of the cGMP pathway.

Single step isolation of sildenafil from commercially available Viagra tablets.

Sildenafil, the active ingredient in Viagra, has been purified from commercially available tablets. The purification, using Sephadex G25 chromatography under conditions of low ionic strength, is simple and inexpensive. Sildenafil purified according to this protocol has been characterized with respect to its IC50 for PDE5, its ultraviolet absorption profile, and by collision-induced dissociation fingerprinting, positive ion nanospray, and MALDI mass spectrometry. Tritated sildenafil (6 Ci/mmol) was prepared commercially using the sildenafil purified by this protocol and was verified to retain the potency of unlabeled sildenafil. This protocol and similar procedures will allow investigators to easily isolate sufficient amounts of sildenafil or other PDE5 inhibitors for conducting biochemical and in vitro studies of drug action.