Chapter 1
Phosphodiesterases and Cyclic Nucleotide Signaling in the CNS

Marco Conti and Wito Richter

Center for Reproductive Sciences, Department of Obstetrics, Gynecology and Reproductive Sciences, School of Medicine University of California, San Francisco San Francisco CA USA

Introduction

Discovery of PDEs, Historical Perspectives, and Progress in Understanding the Complexity of PDE Functions

Soon after the discovery of the second messenger cAMP by Sutherland and Rall [1], it was observed that cyclic nucleotides are unstable in tissue extracts. This observation paved the way for the identification of the enzymatic activities responsible for their destruction [1]. Sutherland and coworkers correctly attributed this activity to a Mg2+-dependent, methylxanthine-inhibited enzyme that cleaves the cyclic nucleotide phosphodiester bond at the 3′-position, hence the name phosphodiesterase (PDE) (Figure 1.1). With the discovery of cGMP and the improvement of protein separation protocols [2], it also became apparent that multiple PDE isoforms with different affinities for cAMP and cGMP and sensitivity to inhibitors coexist in a cell (Figure 1.1). Only with the application of protein sequencing and molecular cloning techniques has it been realized that 21 genes code for PDEs in humans and that close to 100 proteins are derived from these genes, forming a highly heterogeneous superfamily of enzymes (Figures 1.1 and 1.2) [3].

Figure 1.1 Timeline of the major discoveries related to the field of phosphodiesterases.

Figure 1.2 The domain organization of the different families of phosphodiesterases. Domains are depicted as “barrels” connected by “wires” indicating linker regions. Phosphorylation sites are shown as red circles with the respective kinase phosphorylating this site listed above. PDEs are composed of a C-terminal catalytic domain (shown in red) and distinct regulatory domains at the N-terminus. These include Ca2+/calmodulin (CaM)-binding sites (PDE1), GAF domains that function as cAMP or cGMP sensors (PDE2, PDE5, PDE6, PDE10, and PDE11), the UCRs that include a phosphatidic acid (PA)-binding site in PDE4, and the PAS domain (PDE8). The inhibitory gamma subunit of PDE6 is indicated as a yellow ellipse. Domains functioning as targeting sequences by mediating membrane–protein or protein–protein interactions are indicated as red striated barrels and the transmembrane (TM) domains of PDE3 are indicated in blue. The number of PDE genes belonging to each PDE family is indicated in parentheses beside the PDE family name.

Although PDEs were implicated early on in the control of intracellular levels of cAMP and cGMP and the termination of the neurotransmitter or hormonal signal, 30 additional years of research have been necessary to understand that PDEs are not simply housekeeping enzymes. The activity of PDEs is finely regulated by a myriad of regulatory loops and integrated in a complex fashion with the cyclic nucleotide signaling machinery and other signaling pathways. Blockade of PDE activity does not exclusively lead to an increase in cyclic nucleotides and a gain of function, as one would predict from the removal of cyclic nucleotide degradation. On the contrary, complex changes in cellular responses are associated with PDE inhibition, often causing loss of function, as documented by the phenotypes of natural mutations or engineered inactivation of the PDE genes [4–7]. These findings imply that PDEs and their regulation are indispensable to faithfully translate extracellular cues into appropriate biological responses. Indeed, in neurons as in other cells, the biological outcome of activation of a receptor is defined by the multiple dimensions of the cyclic nucleotide signal. This specificity of the response depends on the changes in concentration of the cyclic nucleotide, the time frame in which these changes occur, and the subcellular locale in which the nucleotides accumulate. Because cyclic nucleotide accumulation is dependent on the steady state of cAMP/cGMP production as well as hydrolysis, degradation by PDEs is a major determining factor of all three dimensions of the cyclic nucleotide signal.

In spite of seemingly comparable enzymatic functions, each of the several PDEs expressed within a cell appears to serve unique roles. This view is paradoxical because it implies, as fittingly summarized by L.L. Brunton, that “Not all cAMP has access to all cellular PDEs” [8]. As an extension of this concept, a PDE may play critical functions in a cell even if it represents a minor fraction of the overall hydrolytic activity, a view with considerable impact on pharmacological strategies targeting PDEs. The discovery of macromolecular complexes involving PDEs has confirmed this concept and added a new dimension to the function of these enzymes in signaling. In those complexes in which they are associated with cyclic nucleotide targets, it is likely that PDEs play an essential role in controlling or limiting the access of cyclic nucleotides to their effectors. Since protein kinase A (PKA), protein kinase G (PKG), GTP exchange protein activated by cAMP (EPAC), and cyclic nucleotide-gated (CNG) channels are tethered to specific subcellular compartments, PDEs likely contribute to the compartmentalization of cyclic nucleotide signaling and to the spatial dimension of the signal. PDEs may also have scaffolding properties within these complexes, opening the possibility that PDEs serve functions beyond their catalytic activity and that a dynamic formation and dissolution of these complexes may contribute to the allosteric regulation of PDE activities.

The PDE Superfamily

After several more PDE genes were discovered through homology screening of nucleotide sequence databases between 1996 and 2000 (PDE8, PDE9, PDE10, and PDE11), the completion of the Human Genome Project in 2001 eventually established that there are 21 PDE genes in humans [9]. Orthologs of all 21 genes are encoded in the genomes of rats and mice and might be present in the same number in other mammals. Metazoan model organisms such as Caenorhabditis elegans or Drosophila melanogaster express orthologs of some, but usually not all of the mammalian PDEs [3]. Based upon their substrate specificities, kinetic properties, inhibitor sensitivities, and, ultimately, their sequence homology, the 21 mammalian PDE genes are subdivided into 11 PDE families, each consisting of 1 to a maximum of 4 genes (Table 1.1). Most PDE genes are expressed as a number of variants through the use of multiple promoters and alternative splicing. The PDE6 genes, with only 1 transcript per gene reported, and PDE9A, for which more than 20 putative variants have been proposed, represent the extremes in the number of variants generated from individual genes. Together, close to 100 PDE proteins are generated in mammals, each likely serving unique cellular functions.

Table 1.1 The Properties of the Mammalian PDE Genes