Regulation of cAMP concentration by calmodulin-dependent cyclic nucleotide phosphodiesterase

Bovine brain contains two major calmodulin (CaM) dependent phosphodiesterase isozymes which are homodimeric proteins with subunit molecular masses of 60 and 63 kilodaltons (kDa), respectively. The 60-kDa subunit isozyme can be phosphorylated by cAMP-dependent protein kinase, resulting in a decrease in the enzyme affinity towards CaM. The phosphorylation is blocked by Ca2+ and CaM and reversed by the CaM-stimulated phosphatase (calcineurin). The 63-kDa subunit isozymes can also be phosphorylated, but in this case by a CaM-dependent protein kinase(s). This phosphorylation is also accompanied by a decrease in the isozyme affinity towards CaM and can be reversed by the CaM-dependent phosphatase. Analysis of the complex regulatory properties of the phosphodiesterase isozymes has led to the suggestion that fluxes of cAMP and Ca2+ during cell activations are closely coupled and that the CaM-dependent phosphodiesterase isozymes play key roles in this signal coupling phenomenon.     

Functions of the N-terminal region of cyclic nucleotide phosphodiesterase 3 (PDE 3) isoforms

The N-terminal portion of phosphodiesterase (PDE) 3 was arbitrarily divided into region 1 (amino acids 1-300), which contains a large hydrophobic domain with six predicted transmembrane helices, and region 2 (amino acids 301-500), with a smaller hydrophobic domain ( approximately 50 residues). To analyze these regions, full-length human (H)PDE3A and mouse (M)PDE3B and a series of N-terminal truncated mutants were synthesized in Sf9 cells. Activities of HPDE3A, H3A-Delta189, MPDE3B, and M3B-Delta196, which retained all or part of the hydrophobic domain in region 1, were recovered almost entirely in particulate fractions. H3A-Delta321 and M3B-Delta302, containing region 2, were recovered essentially equally in particulate and cytosolic fractions. H3A-Delta397 and H3A-Delta457, lacking both hydrophobic domains, were predominantly cytosolic. H3A-Delta510 and M3B-Delta604, lacking both regions 1 and 2, were virtually completely cytosolic. M3B-Delta196 eluted as a large aggregated complex during gel filtration. With removal of greater amounts of N-terminal sequence, aggregation of PDE3 decreased, and H3A-Delta607, H3A-Delta721, and M3B-Delta604 eluted as dimers. Truncated HPDE3A proteins were more sensitive than full-length HPDE3A to inhibition by lixazinone. These results suggest that the hydrophobic domains in regions 1 and 2 contain structural determinants important for association of PDE3 with intracellular membranes, as well for self-association or aggregation during gel filtration and sensitivity to a specific inhibitor.     

Regulation of cyclic nucleotide phosphodiesterase activity in human lung fibroblasts.

Cyclic nucleotide phosphodiesterase activity (3', 5'-cyclic-nucleotide 5'-nucleotidohydrolase, 3.1.2.17) was studied in homogenates of WI-38 human lung fibroblasts using 0.1--200 microgram cyclic nucleotides. Activities were observed with low Km for cyclic AMP(2--5 micron) and low Km for cyclic GMP (1--2 micron) as well as with high Km values for cyclic AMP (100--125 micron) and cyclic GMP (75--100 micron). An increased low Km cyclic AMP phosphodiesterase activity was found upon exposure of intact fibroblasts to 3-isobutyl-1-methylxanthine, an inhibitor of phosphodiesterase activity in broken cell preparations, as well as to other agents which elevate cyclic AMP levels in these cells. The enhanced activity following exposure to 3-isobutyl-1-methylxanthine was selective for the low Km cyclic AMP phosphodiesterase since there was no change in activity of low Km cyclic GMP phosphodiesterase activity or in high Km phosphodiesterase activity with either nucleotide as substrate. The enhanced activity due to 3-isobutyl-1-methylxanthine appeared to involve de novo synthesis of a protein with short half-life (30 min), based on experiments involving cycloheximide and actinomycin D. This activity was also enhanced with increased cell density and by decreasing serum concentration. Studies of some biochemical properties and subcellular distribution of the enzyme indicated that the induced enzyme was similar to the non-induced (basal) low Km cyclic AMP phosphodiesterase.     

Dissimilar cyclic nucleotide phosphodiesterase activities in subcellular fractions from normal and SV40-transformed WI-38 fibroblasts.

Broken cell preparations of WI-38 and SV40-transformed WI-38 (VA13) fibroblasts were used to compare the cyclic nucleotide phosphodiesterase activities of the two cell strains. The bulk of the cAMP or cGMP phosphodiesterase activity of WI-38 and VA13 homogenates was found in the 100,000 x g fibroblast supernatant fractions. WI-38 and VA13 soluble phosphodiesterase activities showed anomalous kinetic behavior with either cAMP or cGMP as the substrate. At low substrate concentrations, e.g., 0.1 muM, WI-38 supernatant fractions hydrolyzed cGMP much more rapidly than cAMP. At high substrate concentrations, e.g., 100muM, the same enzyme preparations degraded cAMP more than twice as fast as cGMP. In contrast, VA13 soluble phosphodiesterase activity catalyzed the hydrolysis of a wide range of cAMP and cGMP concentrations at similar rates. Phosphodiesterase activity in WI-38 supernatant fractions was generally more sensitive than that of the comparable VA13 enzyme activity to inhibition by MIX and papaverine. The cAMP phosphodiesterase activity of both WI-38 and VA13 supernatant preparations was decreased by cGMP in a concentration-dependent manner. cAMP was an effective inhibitor of cGMP hydrolysis by VA13 soluble phosphodiesterase activity. Yet, the cGMP phosphodiesterase activity of WI-38 supernatant fractions was only slightly reduced in the presence of cAMP. DEAE-cellulose chromatography of WI-38 and VA13 supernatant preparations revealed two major peaks of phosphodiesterase activity for each cell type. WI-38 peak I showed much greater activity with 1muM cGMP than with 1muM cAMP and appeared to be composed of two different phosphodiesterase activities. WI-38 peak Ia included phosphodiesterase activity which could be stimulated by boiled, dialyzed fibroblast homogenates while WI-38 peak Ib coincided with column fractions which contained most of the cyclic GMP hydrolytic activity. VA13 peak I phosphodiesterase activity was eluted from DEAE cellulose columns at the same ionic strength as WI-38 peak Ia and hydrolyzed these two substrates at nearly identical rates. This enzyme activity was also increased in the presence of boiled, dialyzed fibroblast preparations. Peak II phosphodiesterase activities from both WI-38 and VA13 fibroblasts were relatively specific for cAMP as the substrate. Phosphodiesterase activity with the properties of WI-38 peak Ib was not isolated from VA13 supernatant fractions. These results suggested that the dissimilar patterns of cAMP accumulation in WI-38 and VA13 cultures may be at least partially related to different phosphodiesterase activities in the normal and the transformed fibroblasts.     

Visual detection of AMP and real-time monitoring of cyclic nucleotide phosphodiesterase (PDE) activity in neutral aqueous solution. Chemosensor-coupled assay of PDE and PDE inhibitors

A phosphate chemosensor ensemble, [Zn(H-bpmp)(pyrocatechol violet)](+) is useful for a real-time assay of phosphodiesterases (PDEs) in a neutral aqueous solution. In addition, a simple and convenient screen and assay procedures for inhibitors of PDEs have been demonstrated using IBMX, a nonselective PDE inhibitor as an example.     

Cell-cycle-specific activity of cyclic nucleotide phosphodiesterase in Physarum polycephalum

The cell-cycle-related activities of the cAMP- and cGMP-dependent phosphodiesterases of Physarum polycephalum were assayed. The activities of plasmodial homogenate and of selected subcellular fractions were measured. The results suggested the presence of both cAMP- and cGMP-dependent phosphodiesterase in the isolated nuclei of P. polycephalum. In addition, they reveal that the cAMP- and cGMP-dependent phosphodiesterase activities of the subcellular fractions fluctuate throughout the cell cycle. The whole-cell homogenates exhibit no cell-cycle-related changes in the presence of 5 X 10(-4) M cGMP. Kinetic data suggest the presence of multiple phosphodiesterase activities in the homogenate and its particulate fractions for the cGMP-dependent enzyme. Multiple cAMP activities are also suggested for the particulate fractions. The Km values indicate that the substrate affinities of the phosphodiesterases from P. polycephalum are similar to those found previously in mammalian systems.     

Characterization of cyclic nucleotide phosphodiesterase activity in Phaseolus vulgaris

Cyclic nucleotide phosphodiesterase (3',5'-cyclic nucleotide nucleotidohydrolase, EC 3.1.4.17) activity isolated from Phaseolus vulgaris L. cv. Limberg seedlings was partially purified and characterized by fractional (NH₄)₂SO₄ precipitation, DEAE-cellulose chromatography, chromatography on 3',5'-cAMP-agarose, gel permeation chromatography and chromatofocusing. A crude enzyme preparation, a 30–65% (NH₄)₂SO₄ pellet, showed an acidic pH optimum. The enzyme activity was stimulated by imidazole and divalent cations such as Ca²⁺, Mg²⁺ and Mn²⁺, whereas NaF, PP_i and Fe³⁺ were inhibitory. Isobutylmethylxanthine had no significant effect on the plant enzyme. An M_I of 42 000 was estimated by gel permeation high performance liquid chromatography. By chromatography on 3',5'-cAMP-agarose a phosphodiesterase was resolved that produced 5'-AMP as sole reaction product.     

Effects of prostaglandins on cyclic nucleotide phosphodiesterase activity in the human term placenta

Slices of human full-term placentas, obtained by elective cesarean section, were incubated in the absence or presence of prostaglandins (PGs) and the cyclic AMP phosphodiesterase (cAMP PDE) activity was measured. PGE1 and PGI2 were shown to stimulate cAMP PDE activity. The effect of PGE1 is related to an increase in the Vmax of the low Km activity without alteration of this apparent Km. Several findings suggest that the cAMP PDE is activated by its own substrate; PGE1 and PGI2, promote an increase of cAMP formation which is observed before the cAMP PDE activation. Dibutyryl cAMP or theophylline also activate cAMP PDE. In contrast, PGF2 alpha does not influence either adenylate cyclase or AMP PDE. In addition, we found that the ability of the placenta to degrade cAMP, increases after parturition. PG levels are higher in the foeto-placental unit during labor, and a causal relationship between these two phenomena is possible. Our data supporting the concept of hormonal control of cAMP PDE is consistent with the hypothesis that an accelerated cAMP metabolism in placenta contributes to the maintenance of a constant equilibrium of the cyclic nucleotide levels in the foeto-placental unit.     

Cyclic nucleotide phosphodiesterase PDE1C1 in human cardiac myocytes

Isoforms in the PDE1 family of cyclic nucleotide phosphodiesterases were recently found to comprise a significant portion of the cGMP-inhibited cAMP hydrolytic activity in human hearts. We examined the expression of PDE1 isoforms in human myocardium, characterized their catalytic activity, and quantified their contribution to cAMP hydrolytic and cGMP hydrolytic activity in subcellular fractions of this tissue. Western blotting with isoform-selective anti-PDE1 monoclonal antibodies showed PDE1C1 to be the principal isoform expressed in human myocardium. Immunohistochemical analysis showed that PDE1C1 is distributed along the Z-lines and M-lines of cardiac myocytes in a striated pattern that differs from that of the other major dual-specificity cyclic nucleotide phosphodiesterase in human myocardium, PDE3A. Most of the PDE1C1 activity was recovered in soluble fractions of human myocardium. It binds both cAMP and cGMP with K(m) values of approximately 1 microm and hydrolyzes both substrates with similar catalytic rates. PDE1C1 activity in subcellular fractions was quantified using a new PDE1-selective inhibitor, IC295. At substrate concentrations of 0.1 microm, PDE1C1 constitutes the great majority of cAMP hydrolytic and cGMP hydrolytic activity in soluble fractions and the majority of cGMP hydrolytic activity in microsomal fractions, whereas PDE3 constitutes the majority of cAMP hydrolytic activity in microsomal fractions. These results indicate that PDE1C1 is expressed at high levels in human cardiac myocytes with an intracellular distribution distinct from that of PDE3A and that it may have a role in the integration of cGMP-, cAMP- and Ca(2+)-mediated signaling in these cells.     

Plasma membrane cyclic nucleotide phosphodiesterase 3B (PDE3B) is associated with caveolae in primary adipocytes

Caveolae, plasma membrane invaginations particularly abundant in adipocytes, have been suggested to be important in organizing insulin signalling. Insulin-induced activation of the membrane bound cAMP degrading enzyme, phosphodiesterase 3B (PDE3B) is a key step in insulin-mediated inhibition of lipolysis and is also involved in the regulation of insulin-mediated glucose uptake and lipogenesis in adipocytes. The aim of this work was to evaluate whether PDE3B is associated with caveolae. Subcellular fractionation of primary rat and mouse adipocytes demonstrated the presence of PDE3B in endoplasmic reticulum and plasma membrane fractions. The plasma membrane PDE3B was further analyzed by detergent treatment at 4 degrees C, which did not solubilize PDE3B, indicating an association of PDE3B with lipid rafts. Detergent-treated plasma membranes were studied using Superose-6 chromatography which demonstrated co-elution of PDE3B with caveolae and lipid raft markers (caveolin-1, flotillin-1 and cholesterol) at a Mw of >4000 kDa. On sucrose density gradient centrifugation of sonicated plasma membranes, a method known to enrich caveolae, PDE3B co-migrated with the caveolae markers. Immunoprecipitation of caveolin-1 using anti caveolin-1 antibodies co-immunoprecipitated PDE3B and immunoprecipitation of flag-PDE3B from adipocytes infected with a flag-PDE3B adenovirus resulted in co-immunoprecipitation of caveolin-1. Studies on adipocytes with disrupted caveolae, using either caveolin-1 deficient mice or treatment of adipocytes with methyl-beta-cyclodextrin, reduced the membrane associated PDE3B activity. Furthermore, inhibition of PDE3 in primary rat adipocytes resulted in reduced insulin stimulated glucose transporter-4 translocation to caveolae, isolated by immunoprecipitation using caveolin-1 antibodies. Thus, PDE3B, a key enzyme in insulin signalling, appears to be associated with caveolae in adipocytes and this localization seems to be functionally important.     

Reduction of cyclic AMP phosphodiesterase activity of several subcellular fractions of rat brain after pretreatment with phospholipase C

Cyclic AMP phosphodiesterase (PDE) activity was assayed in the plasma membrane, mitochondrial and microsomal fractions of rat brain. The specific activity of the enzyme was highest in the plasma membrane fraction followed by mitochondrial and then the microsomal fraction. Phosphodiesterase activity of all three fractions was reduced after pretreatment with lecithinase C (PCase) from Clostridium perfringens but less markedly affected by the pretreatment with sphingomyelinase (SMase) from human placenta. The PDE activity of the plasma membrane fraction was more sensitive to PCase treatment compared with the other two particulate fractions, which showed only a slight loss of activity. Temperature seemed to affect PDE activity of the plasma membrane. The enzyme was quite stable at 30 degrees C but its activity dropped by approximately 46% at 37 degrees C after 90 min of incubation. Pretreatment of the plasma membrane at 30 degrees C with PCase at a concentration of more than 5 U caused a marked loss of PDE activity and the decrease in activity reached a plateau at concentrations above 10 U.     

Mycobacterial phosphodiesterase Rv0805 is a virulence determinant and its cyclic nucleotide hydrolytic activity is required for propionate detoxification

Cyclic AMP (cAMP) signaling is essential to Mycobacterium tuberculosis (Mtb) pathogenesis. However, the roles of phosphodiesterases (PDEs) Rv0805, and the recently identified Rv1339, in cAMP homeostasis and Mtb biology are unclear. We found that Rv0805 modulates Mtb growth within mice, macrophages and on host-associated carbon sources. Mycobacterium bovis BCG grown on a combination of propionate and glycerol as carbon sources showed high levels of cAMP and had a strict requirement for Rv0805 cNMP hydrolytic activity. Supplementation with vitamin B12 or spontaneous genetic mutations in the pta-ackA operon restored the growth of BCGΔRv0805 and eliminated propionate-associated cAMP increases. Surprisingly, reduction of total cAMP levels by ectopic expression of Rv1339 restored only 20% of growth, while Rv0805 complementation fully restored growth despite a smaller effect on total cAMP levels. Deletion of an Rv0805 localization domain also reduced BCG growth in the presence of propionate and glycerol. We propose that localized Rv0805 cAMP hydrolysis modulates activity of a specialized pathway associated with propionate metabolism, while Rv1339 has a broader role in cAMP homeostasis. Future studies will address the biological roles of Rv0805 and Rv1339, including their impacts on metabolism, cAMP signaling and Mtb pathogenesis.     

Regulation by insulin of cyclic nucleotide phosphodiesterase (PDE) from rat luteal cells

The effect of insulin on cyclic nucleotide phosphodiesterase (PDE) in rat luteal cells was studied. Cells were obtained from PMSG/hCG primed rats and further incubated or not with insulin. The hormone produced an increase of enzyme activity after a 10 min incubation of intact cells. Maximal stimulation was achieved at 0.2 nM of insulin. Two peaks of cyclic nucleotide phosphodiesterase activity were resolved after chromatography of cell cytosolic extracts on DEAE-cellulose. These peaks (I and II) were active with cAMP as substrate but only peak I was active with cGMP. The enzyme activity of both peaks was increased in cells treated with insulin. Phosphodiesterase activity in the two peaks show two kinetic components for cAMP hydrolysis, one of high affinity (Km 2-4 microM) and the other of low affinity (47-56 microM). Treatment of the cells with insulin produced a 2 to 8 fold increase of the Vmax of these peaks. In addition after stimulation with insulin, the activation of peak I phosphodiesterase by calmodulin was less effective.     

Purification and characterization of a human platelet cyclic nucleotide phosphodiesterase

A cyclic nucleotide phosphodiesterase was extensively purified from the 100000g supernatant fraction of human platelets. The purification was 2500-3000-fold with 30% recovery of activity. The enzyme was isolated by DEAE-cellulose chromatography followed by adsorption to blue dextran-Sepharose and elution with cAMP. The protein has a molecular weight of 140 000 as determined by gel filtration. On NaDodSO4-containing polyacrylamide gels the major band is at 61 000 daltons, suggesting that the enzyme may exist as a dimer in solution under nondenaturing conditions. The enzyme requires Mg2+ or Mn2+ for activity. The calcium binding protein calmodulin does not stimulate hydrolysis of cAMP by this enzyme. The purified enzyme hydrolyzes both cAMP and cGMP with normal Michaelis-Menten kinetics with Km values of 0.18 microM and 0.02 microM, respectively. The hydrolysis of cGMP, however, is only one-tenth as rapid as the hydrolysis of cAMP. Cyclic GMP does not stimulate cAMP hydrolysis but instead is a potent competitive inhibitor of cAMP hydrolysis. The enzyme is also competitively inhibited by the phosphodiesterase inhibitors papaverine, 3-isobutyl-l-methylxanthine, and dipyridamole. The enzyme did not cross-react with an antibody raised to a cAMP phosphodiesterase isolated from dog kidney, indicating that the enzymes are not immunologically related. The inhibition of cAMP hydrolysis by cGMP suggests a possible regulatory link between these two cyclic nucleotides. One of the roles of cGMP in platelets may be to potentiate increases in intracellular cAMP by inhibiting the hydrolysis of cAMP by this enzyme.     

Identification of overlapping but distinct cAMP and cGMP interaction sites with cyclic nucleotide phosphodiesterase 3A by site-directed mutagenesis and molecular modeling based on crystalline PDE4B

Cyclic nucleotide phosphodiesterase 3A (PDE3A) hydrolyzes cAMP to AMP, but is competitively inhibited by cGMP due to a low k(cat) despite a tight K(m). Cyclic AMP elevation is known to inhibit all pathways of platelet activation, and thus regulation of PDE3 activity is significant. Although cGMP elevation will inhibit platelet function, the major action of cGMP in platelets is to elevate cAMP by inhibiting PDE3A. To investigate the molecular details of how cGMP, a similar but not identical molecule to cAMP, behaves as an inhibitor of PDE3A, we constructed a molecular model of the catalytic domain of PDE3A based on homology to the recently determined X-ray crystal structure of PDE4B. Based on the excellent fit of this model structure, we mutated nine amino acids in the putative catalytic cleft of PDE3A to alanine using site-directed mutagenesis. Six of the nine mutants (Y751A, H840A, D950A, F972A, Q975A, and F1004A) significantly decreased catalytic efficiency, and had k(cat)/K(m) less than 10% of the wild-type PDE3A using cAMP as substrate. Mutants N845A, F972A, and F1004A showed a 3- to 12-fold increase of K(m) for cAMP. Four mutants (Y751A, H840A, D950A, and F1004A) had a 9- to 200-fold increase of K(i) for cGMP in comparison to the wild-type PDE3A. Studies of these mutants and our previous study identified two groups of amino acids: E866 and F1004 contribute commonly to both cAMP and cGMP interactions while N845, E971, and F972 residues are unique for cAMP and the residues Y751, H836, H840, and D950 interact with cGMP. Therefore, our results provide biochemical evidence that cGMP interacts with the active site residues differently from cAMP.     

Determination of cyclic nucleotide phosphodiesterase activity by high-performance liquid chromatography

A new assay for cyclic nucleotide phosphodiesterase activity by high-performance liquid chromatography with on-line radiochemical detection has been developed. The method is based on the measurement of 3H-labeled nucleoside monophosphates formed from cyclic nucleotides by the action of 3',5'-cyclic-nucleotide phosphodiesterase (PDE). The reaction products are determined from the incubation mixture after removal of the protein by injection of an aliquot into the liquid chromatograph. The detection limit with counting efficiency of 30% is 20 fmol of 3H-labeled product, which makes the method suitable for detection of low PDE activities.     

Identification of rat cyclic nucleotide phosphodiesterase 11A (PDE11A): comparison of rat and human PDE11A splicing variants.

We have isolated and characterized rat cyclic nucleotide phosphodiesterase (PDE)11A, which exhibits properties of a dual-substrate PDE, and its splice variants (RNPDE11A2, RNPDE11A3, and RNPDE11A4). The deduced amino-acid sequence of the longest form of rat PDE11A splice variant, RNPDE11A4, was 94% identical with that of the human variant (HSPDE11A4). Rat PDE11A splice variants were expressed in a tissue-specific manner. RNPDE11A4 showed unique tissue distribution distinct from HSPDE11A4, which is specifically expressed in the prostate. Rat PDE11A splice variants were expressed in COS-7 cells, and their enzymatic characteristics were compared. Although the Km values for cAMP and cGMP were similar for all of them (1.3-1.6 and 2.1-3.9 microM, respectively), the Vmax values differed significantly (RNPDE11A4 > RNPDE11A2 > RNPDE11A3). Human PDE11A variants also displayed very similar Km values and significantly different Vmax values (HSPDE11A4 > HSPDE11A2 > HSPDE11A3 > HSPDE11A1). The Vmax values of HSPDE11A4 for cAMP and cGMP were at least 100 times higher than those of HSPDE11A1. These observations indicate unique characteristics of PDE11A splicing variants.     

Membrane localization of cyclic nucleotide phosphodiesterase 3 (PDE3). Two N-terminal domains are required for the efficient targeting to, and association of, PDE3 with endoplasmic reticulum

Subcellular localization of cyclic nucleotide phosphodiesterases (PDEs) may be important in compartmentalization of cAMP/cGMP signaling responses. In 3T3-L1 adipocytes, mouse (M) PDE3B was associated with the endoplasmic reticulum (ER) as indicated by its immunofluorescent colocalization with the ER protein BiP and subcellular fractionation studies. In transfected NIH 3006 or COS-7 cells, recombinant wild-type PDE3A and PDE3B isoforms were both found almost exclusively in the ER. The N-terminal portion of PDE3 can be arbitrarily divided into region 1 (aa 1-300), which contains a large hydrophobic domain with six predicted transmembrane helices, followed by region 2 (aa 301-500) containing a smaller hydrophobic domain (of approximately 50 aa). To investigate the role of regions 1 and 2 in membrane association, we examined the subcellular localization of a series of catalytically active, Flag-tagged N-terminal-truncated human (H) PDE3A and MPDE3B recombinants, as well as a series of fragments from regions 1 and 2 of MPDE3B synthesized as enhanced green fluorescent (EGFP) fusion proteins in COS-7 cells. In COS-7 cells, the localization of a mutant HPDE3A, lacking the first 189 amino acids (aa) and therefore four of the six predicted transmembrane helices (H3A-Delta189), was virtually identical to that of the wild type. M3B-Delta302 (lacking region 1) and H3A-Delta397 (lacking region 1 as well as part of region 2) retained, to different degrees, the ability to associate with membranes, albeit less efficiently than H3A-Delta189. Proteins that lacked both regions 1 and 2, H3A-Delta510 and M3B-Delta604, did not associate with membranes. Consistent with these findings, region 1 EGFP-MPDE3B fusion proteins colocalized with the ER, whereas region 2 EGFP fusion proteins were diffusely distributed. Thus, some portion of the N-terminal hydrophobic domain in region 1 plus a second domain in region 2 are important for efficient membrane association/targeting of PDE3.