Cyclic nucleotide phosphodiesterase (PDE) was demonstrated after the discovery of cAMP (Sutherland and Rall 1958). In most tissues, PDE exists in multiple forms which differ in subcellular localization, relative substrate specificity toward cAMP and cGMP, regulatory and immunological properties (Beavo 1995; Kakkar et al. 1999; Goraya and Cooper 2005). Most tissues examined have been shown to contain Ca2+ and calmodulin (CaM)-dependent cyclic nucleotide phosphodiesterase 1 (PDE1) and this enzyme has been intensively studied (Beavo 1995; Kakkar et al. 1999). Earlier, it was suggested that PDE1 consists of a single species (Wells and Hardman 1977), but later found to exist as a tissue-specific and immunological distinct enzyme (Sharma et al. 1984). With the rapidly expanding list of PDE1 isozymes, a new nomenclature was developed based on the primary structure of different PDEs (Beavo et al. 1994). It is now known that at least three different genes encode the PDE1 family; they are named PDE1A, PDE1B, and PDE1C. The complementary DNAs (cDNAs) encoding the brain 63 kDa PDE1 isozyme (PDE1B1), brain 60 kDa PDE1 isozyme (PDE1A2), heart PDE1 (PDE1A1), and 70 kDa PDE1 (PDE1C) were suggested to represent subfamily members. In addition, new splice variants of mouse and human PDEs (PDE1C1, PDE1C2, PDE1C3, PDE1C4, PDE1C5, PDE1A3, and PDE1B1) have been reported. They have been characterized in terms of their regulation by Ca2+, sensitivity to inhibitors, and tissue/cell-specific expression. The genes for different PDEs undergo tissue-specific alternative splicing that generates structurally and functionally diverse gene products.
Differential Stimulation of PDE1 by CaM and Ca2+
PDE1 can be distinguished from other forms of PDE by the stimulation with CaM in the presence of Ca2+. Although brain PDE1A2 and heart PDE1A1 isozymes are almost identical immunologically and kinetically in their properties (Sharma et al. 1984; Sharma and Kalra 1994), they are differentially activated by CaM (Sharma and Kalra 1994). It may be possible that differential affinity for CaM may reflect subtle differences in Ca2+ activation of these PDE1 isozymes. The differences in CaM affinity exhibited by these isozymes may be related to the relative concentration of CaM in these tissues. It has been suggested that the differential of CaM affinity is an important mechanism by which the regulatory action of CaM may be fine-tuned (Klee 1988). However, the physiological significance of differential CaM affinity requires further research in this area. It is noteworthy that CaM concentration in mammalian brain is approximately 10 times higher than in mammalian heart (Klee and Vanaman 1982). The PDE1A1 isozyme have a higher affinity for CaM than brain PDE 1A2 (Hansen and Beavo 1986; Sharma 1991). Similarly, the pig brain PDE1 has been shown to have a lower affinity for CaM than the isozymes from pig artery (Keravis et al. 1986). In addition, Ca2+ and CaM interact synergistically in activation of PDE1 isozymes (Huang et al. 1981). When the CaM concentration is increased, the Ca2+ concentration required for half-maximal activation is decreased. Such synergistic interactions have been repeatedly shown for various CaM-dependent enzymes. Although the physiological significance of the observed differential Ca2+ sensitivity of the PDE1 isozymes is not known, these studies suggest that the differential Ca2+ affinity of the tissue-specific isozymes may be a mechanism by which CaM regulatory reactions are adapted in the respective tissues. The lung PDE1 isozyme has the highest apparent affinity for CaM, since it contains CaM as a subunit (Sharma and Wang 1986a). A change in CaM concentration had no effect on the Ca2+ concentration of the lung enzyme, suggesting that this isozyme does not undergo a Ca2+-dependent reversible association with CaM. At present, the significance of CaM as a subunit is not known.
Kinetic Properties of PDE1 and Its Inhibitors
Originally, it was reported that PDE1 displays a much higher affinity for cGMP than for cAMP (Beavo 1995; Kakkar et al. 1999). The earlier studies from various laboratories reported marked differences in kinetic parameters; the discrepancy may be due to the purity of enzyme, as well as varying assay conditions. To understand the physiological role of PDE1 isozymes, purified PDE1 isozymes were used to determine kinetic parameters. PDE1 isozymes have a higher affinity toward cGMP than cAMP (Kakkar et al. 1999). However, PDE1B1 isozyme exhibits two- to threefold higher affinity for both the substrates, cAMP and cGMP, compared to PDE1A2 and PDE1A1 and a higher Vmax for cGMP and cAMP (Kakkar et al. 1999). It is interesting to note that PDE1A2 and PDE1A1 isozymes have very similar kinetic properties, whereas PDE1B1 isozyme is kinetically distinct from other PDE1 isozymes (Kakkar et al. 1999). However, PDE1C1 has approximately similar Km and Vmax for cAMP and cGMP (Vandeput et al. 2007).
Early studies on inhibitors were carried out by using purified or partially purified PDE1. Therefore, it was not clear from previous studies which of the purified or partially purified PDE1 isozymes were used. Ginsenosides were found to be potent inhibitors of heart PDE1A1 and brain PDE1A2 but not of brain PDE1B1 (Kakkar et al. 1999). However, deprenyl (selegiline hydrochloride), an antiparkinsonian agent also inhibits brain PDE1A2 but is a poor inhibitor of brain PDE1B1 (Kakkar et al. 1999). In addition, amantadine only inhibits brain PDE1A2 isozyme but not brain PDE1B1, heart PDE1A1, and lung PDE1 isozymes (Kakkar et al. 1999). Since the inhibition of these isozymes is overcome by increasing the concentration of CaM, this suggests that these compounds act specifically and reversibly against the action of CaM. Therefore, these compounds may be valuable tools to investigate the diverse physiological roles of distinct PDE1 isozymes. Unlike other CaM-dependent enzymes, PDE1 has been suggested to be inhibited by dihydropyridine calcium antagonists, which act as direct vasodilator drugs rather than indirectly through their actions on CaM. In previous studies, it was not clear which of the specific PDE1 isozymes were used (Kakkar et al. 1999). The effect of dihydropyridine Ca2+ channel blockers felodipine and nicardipine on purified brain PDE1 isozymes has been examined. The results indicate that both brain isozymes are inhibited by felodipine and nicardipine by partial competitive inhibition, and these two Ca2+ antagonists appear to counteract each other (Sharma et al. 1997). The Ki values for felodipine (1.8 and 2.8 μM) and nicarpidine (2.3 and 5.8 μM) for PDE1A2 and PDE1B1, respectively, suggest that the two brain PDE1 isozymes have similar affinities for the Ca2+ antagonists, and both the isozymes bind felodipine slightly tighter than they bind nicarpidine. This study further demonstrated the existence of a specific site, distinct from the active site on PDE1 isozyme, which exhibits high-affinity binding of these drugs.
Interaction of the Ca2+ and cAMP Second Messenger Systems
All cells have the ability to identify and respond to changes in their environment. They recognize extracellular signals through specific cell membrane receptors. The stereospecific binding of the signal molecules to the receptors results in a series of rapid events which translate the external signal into specific cellular responses. Most of these intracellular reactions depend on second messenger systems, which are composed of second messenger molecules such as cAMP and Ca2+ and a host of enzymes, protein factors, and cell organelles which regulate the metabolism of and mediate the regulatory actions of the messenger molecules. In most cases, different second messenger systems undergo complex yet precise interactions, on one hand to achieve integrated regulation of cellular activities and on the other hand to terminate cell activities in an orderly manner. The interaction between the Ca2+ and cAMP second messenger system is among the most extensively studied. The two systems appear to interact at multiple levels and in either a synergistic or an antagonistic manner depending on the cellular process and/or the cell type.
The Ca2+ regulatory cascade (calcium signaling) is mediated by a Ca2+ binding protein, CaM, and involves a Ca2+-dependent, reversible association of CaM with its target protein(s) (Klee 1988). This results in changes in targeted protein activity. In the case of the cAMP second messenger system, cAMP-dependent protein kinase(s) is considered to be the essential molecule which regulates wide-ranging physiological effects. Upon binding of cAMP to a regulatory subunit, the subunits dissociate and two catalytically active subunits are released which catalyze phosphorylation of specific proteins. In addition, numerous cyclic nucleotide-binding proteins have been characterized. Firstly, the high-affinity allosteric binding sites on cGMP-stimulated phosphodiesterase constitute a family of homologous cAMP-binding sites on several proteins. Secondly, the cGMP and cAMP-binding sites on cyclic nucleotide-gated channels include the potassium channel on cardiac pacemaker cells (Fesenko et al. 1985). Initially, such channels were thought to be present only in the retina and cardiac pacemaker cells. However, these channels have since been identified in many other cell types as well. Recently, Epac (Eoac1, Epac2, exchange protein directly activated by cAMP) has been recognized as a genomic sequence with homology to cAMP-binding sites and GEFs (guanine nucleotide exchange factors) for Ras-like proteins. Originally, Epac1 and Epac2 were also identified as a cAMP-GEF1 and cAMP-GEF2, respectively (Kowasski et al. 1998).
On a functional level, there are at least three mechanisms by which the two messenger systems exert their mutual effects (Fig. 1). Firstly, CaM-stimulated enzymes are substrates of cAMP-dependent protein kinase, such as phosphorylase kinase and smooth muscle myosin light chain kinase. A second mechanism of Ca2+-CaM and cAMP interaction involves phosphorylation of common protein substrates by Ca2+-CaM and cAMP-dependent protein kinases. For example, glycogen synthase can be phosphorylated by cAMP-dependent protein kinase, by the CaM-dependent glycogen synthase kinase and phosphorylase kinase, as well as by at least four other kinases. These phosphorylations inactivate both muscle and liver glycogen synthases with the expectation that phosphorylation of the liver enzyme by CaM-dependent glycogen synthase kinase has no effect on enzyme activity. A third mode of interaction between the Ca2+ and cAMP signal systems involves the dephosphorylation of protein phosphatase inhibitor-1. Upon phosphorylation, protein phosphatase inhibitor-1 is active and inhibits the activity of protein phosphatase-1. When it is dephosphorylated by CaM-dependent protein phosphatase (calcineurin, CaN), protein phosphatase inhibitor-1 is inactivated, allowing protein phosphatase-1 to be active.
Regulation of PDE1 Isozymes by Phosphorylation
The main difference is that the brain PDE1A2 and heart PDE1A1 are substrates of cAMP-dependent protein kinase (Sharma and Wang 1985; Sharma 1991; Florio et al. 1994), and phosphorylation is inhibited by Ca2+ and CaM whereas brain PDE1B1 is phosphorylated by CaM-dependent protein kinase II in a Ca2+/CaM-dependent manner (Sharma and Wang 1986b). The phosphorylation of PDE1A2 and PDE1A1 is accompanied by a decrease in the isozymes affinity toward CaM and an accompanying increase in the Ca2+ concentrations required for the isozyme activation by CaM. PDE1A2 and PDE1B1 are phosphorylated by different protein kinases; however, both can be dephosphorylated by CaM-dependent protein phospatase. This dephosphorylation is accompanied by an increase in the affinity of the isozymes for CaM as well as decrease in the Ca2+ concentration for the activation of isozyme by CaM. It is interesting to note that PDE1C is phosphorylated by cAMP-dependent protein kinase and upon phosphorylation of PDE1C activity is inhibited with increased EC50 for CaM (Ang and Antoni 2002).
Role of PDE1A2 and PDE1B1 in the Regulation of cAMP Concentration
The multiple regulatory actions of the PDE1B1 isozyme are dependent on Ca2+ and CaM (Fig. 2b), whereas phosphorylation of PDE1A2 by cAMP-dependent protein kinase is inhibited by Ca2+ and CaM. Regulation of PDE1B1 by Ca2+ and CaM can occur by at least three mechanisms: (1) the PDE1B1 isozyme depends on Ca2+ and CaM for full activity, (2) it requires higher concentrations of Ca2+ for activation upon phosphorylation by Ca2+ and CaM-dependent protein kinase(s), and (3) the phosphorylation of PDE1B1 is reversed by the CaM-dependent protein phosphatase. Therefore, these three distinct CaM-dependent reactions can regulate PDE1B1 and can produce opposite effects on PDE1B1. Based on the findings from studies (Sharma and Wang 1985, 1986b), a working hypothesis has been proposed to indicate how these regulatory actions, separated temporally, could bring about meaningful interactions between the Ca2+ and cAMP signal systems during cell activation (Fig. 2b). It is postulated that adenylate cyclase and the CaM-dependent protein kinase(s) can be turned on at lower concentration of Ca2+ during cell activation, whereas the activation of the PDE1B1 and CaM-dependent protein phosphatase require higher concentration of Ca2+. The first two reactions which act in concert to increase cAMP concentration predominate at the early stage of surge in cytosolic Ca2+, whereas the other two reactions will reduce cAMP concentrations at the higher concentrations of the Ca2+ flux. For this proposed hypothesis, it is necessary that adenylate cyclase and the CaM-dependent protein kinase(s) be activated by lower concentrations of Ca2+ than the CaM-dependent protein phosphatase and PDE1B1 (Fig. 2b). It has been reported that brain adenylate cyclase is indeed activated by much lower concentration of Ca2+ than the PDE1. The validity of the working hypothesis depicted in Fig. 1b has been tested by purification and characterization of CaM-dependent protein kinase from bovine brain. Bovine brain contains two CaM-dependent protein kinases which were separated on a Sephacryl S-300 column (Zhang et al. 1993a). The high molecular weight 500 kDa protein kinase has been purified close to homogeneity. On the basis of its molecular mass, subunit size, and protein substrate specificity, the purified bovine brain CaM-dependent protein kinase is considered to belong to the CaM-dependent protein kinase II family. The phosphorylation of PDE1B1 by the CaM-dependent protein kinase II is dependent on the presence of Ca2+ and CaM, and after phosphorylation a further increase in Ca2+ concentrations is required for enzyme activation.
Earlier, it was postulated (Fig. 2b) that the CaM-dependent protein kinase is activated by CaM at much lower concentrations of Ca2+ than the CaM-dependent protein phosphatase and PDE1B1. However, this suggestion was not supported when the dose-dependent activation of PDE1B1 by Ca2+ was compared with the purified CaM-dependent protein kinase II at identical concentrations of CaM. The results suggest that CaM-dependent protein kinase II and the PDE1B1 have similar Ca2+ concentration dependence at identical concentrations of CaM (Zhang et al. 1993a). However, the observation that the CaM-dependent protein kinase II becomes Ca2+-independent and fully activated upon autophosphorylation suggests an alternative mechanism (Zhang et al. 1993b). Therefore, the CaM-dependent protein kinase may use the autophosphorylation reaction to override its requirements for higher concentrations of Ca2+.
Bovine brain CaM-dependent protein kinase II is autophosphorylated rapidly in the presence of Ca2+ and CaM; however, it is converted into a Ca2+-independent protein kinase, that is, the phosphorylation of PDE1B1 isozyme by autophosphorylated CaM-dependent protein kinase II becomes Ca2+-independent. It is postulated that the autophosphorylation reaction can be used to achieve the required temporal separation of the activation of protein kinase from that of phosphatase and/or PDE1B1. Therefore, upon very brief exposure to high concentrations of Ca2+, the CaM-dependent protein kinase II becomes active and insensitive to subsequent increases in Ca2+ concentrations, whereas the activation of PDE1B1 requires the continued presence of high concentrations of Ca2+.
In summary, during the early stage of cell activation, initial increase in cAMP and Ca2+ causes a temporary suppression of PDE1B1 isozyme activity to maintain the rise in cAMP concentration. As Ca2+ concentration in the cell is subsequently elevated, CaM-dependent protein phosphatase is activated to reverse the phosphorylation of PDE1B1 isozyme and reactivate PDE1B1. Since CaM-dependent protein phosphatase also dephosphorylates protein phosphatase inhibitor-I to cause reactivation of protein phosphatase-I, autophosphorylation of CaM-dependent protein kinase II is also reversed. Therefore, rephosphorylation of PDE1B1 isozyme will no longer occur as Ca2+ concentration subsides in the cell.
The main feature of the working hypothesis for the regulation of PDE1 isozymes is that the multiple regulatory actions exerted by second messengers on a single PDE1 isozyme are temporally separated. This is possible because the concentrations of both cAMP and Ca2+ undergo continuous change during cell activation. When the cell Ca2+ flux changes, cAMP flux will change accordingly. The transitory elevations of two messengers during cell activation are, therefore, coupled to each other with PDE1 isozymes playing key roles in this signal coupling phenomenon.
One of the most intensively studied cyclic nucleotide phosphodiesterase enzymes is the calmodulin-dependent cyclic nucleotide phosphodiesterase (PDE1) which is stimulated by the binding of Ca2+ and calmodulin. The earlier notion that PDE1 consists of a single species has been surpassed and it is clear that PDE1 exists as different isozymes, namely, PDE1A, PDE1B, PDE1C, and their various splice variants. These isozymes are regulated by multiple second messenger–dependent regulatory systems. One of the main features of the findings is that during cell activation, the various regulatory activities are temporally separated. Therefore, a temporal separation of second messenger–dependent reactions can be considered as natural consequences of dynamic fluxes of the messengers. These studies indicate that the activity of PDE1 is precisely regulated by cross talk between the Ca2+ and cAMP-signaling pathways.
Despite substantial progress, it is important to acknowledge that the understanding of the functions of PDE1 in disease states is far from complete. In the future, it is important to identify adaptive responses where PDE1 isozyme’s activity and expression are changed and the change in levels of proteins which regulate and target PDE1 isozymes.
- Beavo JA, Conti M, Healip RJ. Multiple cyclic nucleotide phosphodiesterases. Mol Pharmacol. 1994;75:399–405.Google Scholar
- Fesenko EE, Kolesnikov SS, Lyubarsky AL. Induction of cyclic GMP of cationic conductance in plasma membrane of retinal rod outer segment. Nature. 1985;313:310–3.Google Scholar
- Klee CB. Interaction of calmodulin with Ca2+ and target proteins. In: Cohen P, Klee CB, editors. Molecular aspects of cell regulation. 5 New York: Elsevier Science; 1988. p. 35–56.Google Scholar
- Zhang GY, Wang JH, Sharma RK. Purification and characterization of bovine brain calmodulin-dependent protein kinase II. The significance of autophosphorylation in the regulation of 63 kDa calmodulin-dependent cyclic nucleotide phosphodiesterase isozyme. Mol Cell Biochem. 1993a;122:159–69.PubMedCrossRefGoogle Scholar