Seven cyclic nucleotide phosphodiesterase gene families (PDE1-7) have been identified. PDE3 isoforms are characterized by their high affinity for cAMP and cGMP, their specific inhibition by certain drugs that increase myocardial contractility, relax airway and vascular smooth muscle, inhibit platelet aggregation and stimulate insulin secretion, and their rapid activation in response to insulin, IGF-1, IL-4, and agents that increase cAMP. The N-terminal portions of PDE3A and PDE3B isoforms contain a large hydrophobic region with 5-6 predicted transmembrane segments, and several downstream consensus phosphorylation sites. To study the role of the hydrophobic region in the membrane-association of PDE3, a human cardiac HPDE3A cDNA and several truncated recombinants (delta 190, delta 398, delta 511 which lack coding sequence for the first 190, 398, and 511 amino acids, respectively) and mouse adipocyte MPDE3B and delta 604MPDE3B cDNA, lacking coding sequence for the first 604 amino acids, were tagged at their C-terminal ends with the FLAG epitope, expressed transiently in COS-7 cells and localized by immunofluorescence microscopy. Both HPDE3A and MPDE3B showed strong staining of a reticular network which co-localized with the ER-marker BIP (not with alpha-tubulin) and intense perinuclear staining similar to that observed for Golgi proteins beta-COP and p58. A rat PDE3B recombinant permanently expressed in NIH 3006 fibroblasts exhibited reticular staining virtually identical to HPDE3A and MPDE3B in COS cells when detected with an anti-peptide antibody raised against the N-terminus of MPDE3B. delta 190HPDE3A (which includes part of the hydrophobic region) fluorescence was identical to that for PDE3A and PDE3B. delta 398HPDE3A (lacking the hydrophobic domain) fluorescence was more diffuse than intact HPDE3A with some reticular staining, indicating a difference in the distribution of this recombinant between membrane and cytosol. Staining of delta 511HPDE3A and MPDE3B delta 604 was diffuse, with little or no reticular staining, suggesting cytosolic localization of these recombinants. delta 511HPDE3A and delta 604MPDE3B activities were almost exclusively cytosolic when expressed in Sf9 cells. These results suggest that sequences in the hydrophobic domain plus additional downstream sequences are involved in the association of PDE3 with intracellular membranes. Future studies will attempt to further define the domains involved in and regulation of membrane targeting as well as effects of subcellular location on PDE3 regulation and function and the importance of PDE3 localization and compartmentalization in cell function. Although PDE3B may be involved in the regulation of several physiological processes, including insulin secretion and insulin action, little is known of its role in development. To understand more of PDE3B function, we are attempting to disrupt the mouse PDE3B gene by homologous recombination. An about 16 kb Sal1 genomic fragment containing mouse PDE3B 5' flanking sequences and exon 1 was cloned from a mouse genomic library. Both mouse and human PDE3B exon 1 encode the N-terminal portion of PDE3B that includes the putative membrane association domain and the downstream serine residue (analogous to rat PDE3B Ser 302) phosphorylated in response to insulin and isoproterenol in intact adipocytes. We are attempting to analyze promoter elements in the 5' flanking region of the MPDE3B gene using MPDE3B-Luciferase expression chimeras. A replacement vector targeting MPDE3B was constructed with about 6.0 kb Xba1 fragment (subcloned from the about 16 kb Sal1 genomic clone). We have recently successfully introduced the replacement vector into mouse ES cells by homologous recombination and then into pseudopregnant mice who have produced a number of chimeric agouti pups.
