The crystal structure of AMP-bound PDE4 suggests a mechanism for phosphodiesterase catalysis

Cyclic nucleotide phosphodiesterases (PDEs) regulate the intracellular concentrations of cyclic 3',5'-adenosine and guanosine monophosphates (cAMP and cGMP, respectively) by hydrolyzing them to AMP and GMP, respectively. Family-selective inhibitors of PDEs have been studied for treatment of various human diseases. However, the catalytic mechanism of cyclic nucleotide hydrolysis by PDEs has remained unclear. We determined the crystal structure of the human PDE4D2 catalytic domain in complex with AMP at 2.4 A resolution. In this structure, two divalent metal ions simultaneously interact with the phosphate group of AMP, implying a binuclear catalysis. In addition, the structure suggested that a hydroxide ion or a water bridging two metal ions may serve as the nucleophile for the hydrolysis of the cAMP phosphodiester bond.     

Dissecting the cofactor-dependent and independent bindings of PDE4 inhibitors

Type 4 phosphodiesterases (PDE4s) are metallohydrolases that catalyze the hydrolysis of cAMP to AMP. At the bottom of its active site lie two divalent metal ions in a binuclear motif which are involved in both cAMP binding and catalysis [(2000) Science 288, 1822-1825; (2000) Biochemistry 39, 6449-6458]. Using a SPA-based equilibrium [(3)H]rolipram binding assay, we have determined that Mg(2+), Mn(2+), and Co(2+) all mediated a high-affinity (K(d) between 3 and 8 nM) and near stoichiometric (R)-rolipram binding to PDE4. In their absence, (R)-rolipram binds stoichiometrically to the metal ion-free apoenzyme with a K(d) of approximately 150 nM. The divalent cation dose responses in mediating the high-affinity rolipram/PDE4 interaction mirror their efficacy in catalysis, suggesting that both metal ions of the holoenzyme are involved in mediating the high-affinity (R)-rolipram/PDE4 interaction. The specific rolipram binding to the apo- and holoenzyme is differentially displaced by cAMP, AMP, and other inhibitors, providing a robust tool to dissect the components of metal ion-dependent and independent PDE4/ligand interactions. cAMP binds to the holoenzyme with a K(s) of 1.9 microM and nonproductively to the apoenzyme with a K(d) of 179 microM. In comparison, AMP binds to the holo- and apoenzyme with K(d) values of 7 and 11 mM, respectively. The diminished Mg(2+)-dependent component of AMP binding to PDE4 suggests that most of the Mg(2+)/phosphate interaction in the cAMP/PDE4 complex is disrupted upon the hydrolysis of the cyclic phosphoester bond, leading to the rapid release of AMP.     

Crystal structure of phosphodiesterase 4D and inhibitor complex(1)

Cyclic nucleotide phosphodiesterases (PDEs) regulate physiological processes by degrading intracellular second messengers, adenosine-3′,5′-cyclic phosphate or guanosine-3′,5′-cyclic phosphate. The first crystal structure of PDE4D catalytic domain and a bound inhibitor, zardaverine, was determined. Zardaverine binds to a highly conserved pocket that includes the catalytic metal binding site. Zardaverine fills only a portion of the active site pocket. More selective PDE4 inhibitors including rolipram, cilomilast and roflumilast have additional functional groups that can utilize the remaining empty space for increased binding energy and selectivity. In the crystal structure, the catalytic domain of PDE4D possesses an extensive dimerization interface containing residues that are highly conserved in PDE1, 3, 4, 8 and 9. Mutations of R358D or D322R among these interface residues prohibit dimerization of the PDE4D catalytic domain in solution.     

A new nonhydrolyzable reactive cAMP analog, (Sp)-adenosine-3',5'-cyclic-S-(4-bromo-2,3-dioxobutyl)monophosphorothioate irreversibly inactivates human platelet cGMP-inhibited cAMP phosphodiesterase

Levels of cAMP that control critical platelet functions are regulated by cGMP-inhibited cAMP phosphodiesterase (PDE3A). We previously showed that millimolar concentrations of the hydrolyzable 8-[(4-bromo-2,3-dioxobutyl)thioadenosine 3',5'-cyclic monophosphate (8-BDB-TcAMP) inactivate PDE3A. We have now synthesized a nonhydrolyzable affinity label to probe the active site of PDE3A. The nonhydrolyzable adenosine 3',5'-cyclic monophosphorothioates, Sp-cAMPS and Rp-cAMPS, function as competitive inhibitors of PDE3A with K(i) = 47.6 and 4400 microM, respectively. We therefore coupled Sp-cAMPS with 1,4-dibromobutanedione to yield (Sp)-adenosine-3',5'-cyclic-S-(4-bromo-2,3-dioxobutyl)monophosphorothioate, [Sp-cAMPS-(BDB)]. Sp-cAMPS-(BDB) inactivates PDE3A in a time-dependent, irreversible reaction with k(max) = 0.0116 min(-1) and K(I) = 10.1 microM. The order of effectiveness of protectants in decreasing the rate of inactivation (with K(d) in microM) is: Sp-cAMPS (24) > Rp-cGMPS (1360), Sp-cGMPS (1460) > GMP (4250), AMP (10600), Rp-cAMPS (22170). These results suggest that the inactivation of PDE3A by Sp-cAMPS-(BDB) is a consequence of reaction at the overlap of the cAMP and cGMP binding regions in the active site.     

Effects of PDE4 inhibitors on lipopolysaccharide-induced priming of superoxide anion production from human mononuclear cells.

AIMS: Phosphodiesterase 4 (PDE4) inhibitors have been described as potent anti-inflammatory compounds, involving an increase in intracellular levels of cyclic 3'',5''-adenosine monophosphate (AMP). The aim of this study was to compare the effects of selective PDE4 inhibitors, rolipram and RP 73-401 with the cell permeable analogue of cyclic AMP, dibutyryl-cyclic AMP (db-cAMP) and the anti-inflammatory cytokine interleukin-10 (IL-10) on superoxide anion production from peripheral blood mononuclear cells preincubated with lipopolysaccharide (LPS). MAJOR FINDINGS: We report that, after incubation of the cells with LPS, a large increase in superoxide anion production was observed. Rolipram or RP 73-401 (10(-8) to 10(-5) M) induced significant reductions of fMLP-induced superoxide anion production in cells incubated with or without LPS. The db-cAMP (10(-5) to 10(-3) M) also elicited dose-dependent inhibitions of the fMLP-induced superoxide anion production. In contrast, IL-10 (1 or 10 ng/ml) did not elicit a reduction in fMLP-induced superoxide anion production in both conditions. PRINCIPAL CONCLUSION: These results suggest that the inhibitory activity of PDE4 inhibitors on fMLP-induced production of superoxide anion production is mediated by db-cAMP rather than IL-10.     

In addition to the SH3 binding region, multiple regions within the N-terminal noncatalytic portion of the cAMP-specific phosphodiesterase, PDE4A5, contribute to its intracellular targeting

The long cyclic AMP (cAMP)-specific phosphodiesterase isoform, PDE4A5 (PDE4A subfamily isoform variant 5), when transiently expressed in COS-7 cells, was shown in subcellular fractionation studies to be associated with both membrane and cytosol fractions, with immunofluorescence analyses identifying PDE4A5 as associated both with ruffles at the cell margin and also at a distinct perinuclear localisation. Deletion of the first nine amino acids of PDE4A5 (1) ablated its ability to interact with the SH3 domain of the tyrosyl kinase, LYN; (2) reduced, but did not ablate, membrane association; and (3) disrupted the focus of PDE4A5 localisation within ruffles at the cell margin. This deleted region contained a Class I SH3 binding motif of similar sequence to those identified by screening a phage display library with the LYN-SH3 domain. Truncation to remove the PDE4A5 isoform-specific N-terminal region caused a further reduction in membrane association and ablated localisation at the cell margin. Progressive truncation to delete the PDE4A long isoform common region and then the long isoform-specific UCR1 did not cause any further change in membrane association or intracellular distribution. However, deletion up to the super-short form splice junction generated an entirely soluble 'core' PDE4A species. We propose that multiple sites in the N-terminal noncatalytic portion of PDE4A5 have the potential to associate with intracellular structures and thus define its intracellular localisation. At least two such sites lie within the PDE4A5 isoform-specific N-terminal region and these appear to be primarily responsible for targeting PDE4A5 to, and organising it within, the cell margin; one is an SH3 binding motif able to interact with LYN kinase and the other lies within the C-terminal portion of the PDE4A5 unique region. A third membrane association region is located within the N-terminal portion of UCR2 and appears to be primarily responsible for targeting to the perinuclear region. Progressive N-terminal truncation, to delete defined regions of PDE4A5, identified activity changes occurring upon deletion of the SH3 binding site region and then upon deletion of the membrane association site region located within UCR2. This suggests that certain of these anchor sites may not only determine intracellular targeting but may also transduce regulatory effects on PDE4A5 activity.     

Regulation of AMP-activated protein kinase by cAMP in adipocytes: Roles for phosphodiesterases,protein kinase B,protein kinase A,Epac and lipolysis

AMP-activated protein kinase (AMPK) is an important regulator of cellular energy status. In adipocytes, stimuli that increase intracellular cyclic AMP (cAMP) have also been shown to increase the activity of AMPK. The precise molecular mechanisms responsible for cAMP-induced AMPK activation are not clear. Phosphodiesterase 3B (PDE3B) is a critical regulator of cAMP signaling in adipocytes. Here we investigated the roles of PDE3B, PDE4, protein kinase B (PKB) and the exchange protein activated by cAMP 1 (Epac1), as well as lipolysis, in the regulation of AMPK in primary rat adipocytes. We demonstrate that the increase in phosphorylation of AMPK at T172 induced by the adrenergic agonist isoproterenol can be diminished by co-incubation with insulin. The diminishing effect of insulin on AMPK activation was reversed upon treatment with the PDE3B specific inhibitor OPC3911 but not with the PDE4 inhibitor Rolipram. Adenovirus-mediated overexpression of PDE3B and constitutively active PKB both resulted in greatly reduced isoproterenol-induced phosphorylation of AMPK at T172. Co-incubation of adipocytes with isoproterenol and the PKA inhibitor H89 resulted in a total ablation of lipolysis and a reduction in AMPK phosphorylation/activation. Stimulation of adipocytes with the Epac1 agonist 8-pCPT-2′O-Me-cAMP led to increased phosphorylation of AMPK at T172. The general lipase inhibitor Orlistat decreased isoproterenol-induced phosphorylation of AMPK at T172. This decrease corresponded to a reduction of lipolysis from adipocytes. Taken together, these data suggest that PDE3B and PDE4 regulate cAMP pools that affect the activation/phosphorylation state of AMPK and that the effects of cyclic AMP on AMPK involve Epac1, PKA and lipolysis.     

Crystal structure of human phosphoribosylpyrophosphate synthetase 1 reveals a novel allosteric site

PRPP (phosphoribosylpyrophosphate) is an important metabolite essential for nucleotide synthesis and PRS (PRPP synthetase) catalyses synthesis of PRPP from R5P (ribose 5-phosphate) and ATP. The enzymatic activity of PRS is regulated by phosphate ions, divalent metal cations and ADP. In the present study we report the crystal structures of recombinant human PRS1 in complexes with SO4(2-) ions alone and with ATP, Cd2+ and SO4(2-) ions respectively. The AMP moiety of ATP binds at the ATP-binding site, and a Cd2+ ion binds at the active site and in a position to interact with the beta- and gamma-phosphates of ATP. A SO4(2-) ion, an analogue of the activator phosphate, was found to bind at both the R5P-binding site and the allosteric site defined previously. In addi-tion, an extra SO4(2-) binds at a site at the dimer interface between the ATP-binding site and the allosteric site. Binding of this SO4(2-) stabilizes the conformation of the flexible loop at the active site, leading to the formation of the active, open conformation which is essential for binding of ATP and initiation of the catalytic reaction. This is the first time that structural stabilization at the active site caused by binding of an activator has been observed. Structural and biochemical data show that mutations of some residues at this site influence the binding of SO4(2-) and affect the enzymatic activity. The results in the present paper suggest that this new SO4(2-)-binding site is a second allosteric site to regulate the enzymatic activity which might also exist in other eukaryotic PRSs (except plant PRSs of class II), but not in bacterial PRSs.     

Crystal structures of phosphodiesterases 4 and 5 in complex with inhibitor 3-isobutyl-1-methylxanthine suggest a conformation determinant of inhibitor selectivity

Cyclic nucleotide phosphodiesterases (PDEs) are a superfamily of enzymes controlling cellular concentrations of the second messengers cAMP and cGMP. Crystal structures of the catalytic domains of cGMP-specific PDE5A1 and cAMP-specific PDE4D2 in complex with the nonselective inhibitor 3-isobutyl-1-methylxanthine have been determined at medium resolution. The catalytic domain of PDE5A1 has the same topological folding as that of PDE4D2, but three regions show different tertiary structures, including residues 79-113, 208-224 (H-loop), and 341-364 (M-loop) in PDE4D2 or 535-566, 661-676, and 787-812 in PDE5A1, respectively. Because H- and M-loops are involved in binding of the selective inhibitors, the different conformations of the loops, thus the distinct shapes of the active sites, will be a determinant of inhibitor selectivity in PDEs. IBMX binds to a subpocket that comprises key residues Ile-336, Phe-340, Gln-369, and Phe-372 of PDE4D2 or Val-782, Phe-786, Gln-817, and Phe-820 of PDE5A1. This subpocket may be a common site for binding nonselective inhibitors of PDEs.     

Pyrazolopyridines as a novel structural class of potent and selective PDE4 inhibitors

Optimisation of a high-throughput screening hit resulted in the discovery of 4-(substituted amino)-1-alkyl-pyrazolo[3,4-b]pyridine-5-carboxamides as potent and selective inhibitors of Phosphodiesterase 4 (PDE4). Herein, we describe early SAR studies around this novel template highlighting preferred substituents and rationalization of SAR through X-ray crystal structures of analogues bound to the PDE4 active site. Pyrazolopyridine 20a was found to be a potent and selective PDE4 inhibitor which also inhibits LPS induced TNF-α production from isolated human peripheral blood mononuclear cells and has an encouraging rat PK profile suitable for oral dosing.     

Conformational conversion of PDE5 by incubation with sildenafil or metal ion is accompanied by stimulation of allosteric cGMP binding

Phosphodiesterase-5 (PDE5) is a dimer containing a cGMP-specific catalytic domain and an allosteric cGMP-binding subdomain (GAF A) on each subunit. PDE5 exhibits three conformational forms that can be separated by Native PAGE and are denoted as Bands 1, 2, and 3 in decreasing order of mobility. A preparation comprised mainly of Band 2 PDE5 was partially converted to Band 3 PDE5 by 1 h incubation with cGMP or the PDE5-specific inhibitors sildenafil, vardenafil, or tadalafil, but not with cAMP, milrinone (PDE3-specific), or rolipram (PDE4-specific). Band 2 PDE5 was converted almost entirely to Band 3 PDE5 by overnight incubation with sildenafil at 30 °C. This time-dependent conversion was accompanied by a 7-fold increase in allosteric cGMP-binding activity, suggesting that Band 3 PDE5 is a much more active form than Band 2 PDE5 for allosteric cGMP binding. Conversion of Band 2 PDE5 to Band 3 PDE5 occurred faster by pre-incubation with cGMP, which binds to both the allosteric and catalytic sites of PDE5, than with catalytic site-specific sildenafil. Overnight incubation of a Band 2/Band 3 PDE5 mixture with EDTA caused time-dependent conversion to Band 1 PDE5 (apoenzyme), and this conversion was accompanied by a 50% loss in cGMP-binding activity. After incubation with EDTA, addition of Mn⁺⁺ or Mg⁺⁺ caused reversion of Band 1 to a Band 2/Band 3 PDE5 mixture in which Band 3 PDE5 predominated. This reversion was accompanied by a 3-fold increase in allosteric cGMP-binding activity. The combination of results implied that physiological conversion of Band 2 to Band 3 PDE5 by cGMP and/or divalent metal ion occupancy of the catalytic domain would increase allosteric cGMP binding to the enzyme. This conversion would produce a greater negative feedback effect on cGMP action by increasing sequestration of cGMP at the allosteric cGMP-binding site of PDE5 and by increasing cGMP degradation at the catalytic site of the enzyme. This conversion would also increase PDE5 inhibitor binding to the enzyme.     

A nonhydrolyzable reactive cAMP analogue, (S(p))-8-[(4-bromo-2,3-dioxobutyl)thio]adenosine 3',5'-cyclic S-(methyl)monophosphorothioate, irreversibly inactivates human platelet cGMP-inhibited cAMP phosphodiesterase at micromolar concentrations

We previously showed that 8-[(4-bromo-2,3-dioxobutyl)thio]adenosine 3',5'-cyclic monophosphate inactivates cAMP phosphodiesterase (PDE3A); however, millimolar concentrations were needed to inactivate PDE3A because of ongoing hydrolysis. We have now synthesized a nonhydrolyzable reactive cAMP analogue, (S(p))-8-[(4-bromo-2,3-dioxobutyl)thio]adenosine 3',5'-cyclic S-(methyl)monophosphorothioate (S(p)-8-BDB-TcAMPSMe). S(p)-8-BDB-TcAMPSMe inactivates PDE3A in a time-dependent, irreversible manner, exhibiting saturation kinetics with a k(max) of (19.5 +/- 0.3) x 10(-3) min(-1) and a K(I) of 3.5 +/- 0.3 muM. To ascertain whether S(p)-8-BDB-TcAMPSMe reacts in the active site, nonhydrolyzable analogues of the substrate cAMP, or the competitive inhibitor cGMP, were included to protect against the inactivation of PDE3A. The order of effectiveness of protectants in decreasing the rate of inactivation (with K(d) values in micromolar) is as follows: S(p)-cAMPS (18) > R(p)-cGMPS (560) and S(p)-cGMPS (1260) > 5'-AMP (17 660), R(p)-cAMPS (30 110), and 5'-GMP (42 170). We docked S(p)-8-BDB-TcAMPSMe into PDE3A, based on the structural model of PDE3A-cAMP and the kinetic data from site-directed mutants. The S(p)-8-BDB-TcAMPSMe fits into the active site in the model. These results suggest that inactivation of PDE3A by the affinity reagent is a consequence of reaction at the overlap between cAMP and cGMP binding regions in the active site. S(p)-8-BDB-TcAMPSMe has proven to be an effective active site-directed irreversible cAMP affinity label for platelet PDE3A and can be used to identify amino acids in the active site of PDE3A as well as in other cAMP phosphodiesterases.     

Multiple conformations of phosphodiesterase-5: implications for enzyme function and drug development

Phosphodiesterase-5 (PDE5) is the target for sildenafil, vardenafil, and tadalafil, which are drugs for treatment of erectile dysfunction and pulmonary hypertension. We report here the crystal structures of a fully active catalytic domain of unliganded PDE5A1 and its complexes with sildenafil or icarisid II. These structures together with the PDE5A1-isobutyl-1-methylxanthine complex show that the H-loop (residues 660-683) at the active site of PDE5A1 has four different conformations and migrates 7-35A upon inhibitor binding. In addition, the conformation of sildenafil reported herein differs significantly from those in the previous structures of chimerically hybridized or almost inactive PDE5. Mutagenesis and kinetic analyses confirm that the H-loop is particularly important for substrate recognition and that invariant Gly(659), which immediately precedes the H-loop, is critical for optimal substrate affinity and catalytic activity.     

磷酸二酯酶5（PDE5）在人增生前列腺移行带中的表达及分布

目的:探讨磷酸二酯酶5在人前列腺移行带组织中的表达与分布。方法:应用免疫组化SP法,选择前列腺增生患者34例,检测PDE5表达及分布。结果:34例标本中PDE5均呈阳性表达。PDE5在前列腺移行带的腺组织上皮普遍呈中等表达,间质部分弱表达。结论:PDE5在人增生前列腺组织中分布,但分布不均,表明不同PDE5可能行使不同的功能。     

Partial reconstitution of photoreceptor cGMP phosphodiesterase characteristics in cGMP phosphodiesterase-5

Photoreceptor cGMP phosphodiesterases (PDE6) are uniquely qualified to serve as effector enzymes in the vertebrate visual transduction cascade. In the dark-adapted photoreceptors, the activity of PDE6 is blocked via tight association with the inhibitory gamma-subunits (Pgamma). The Pgamma block is removed in the light-activated PDE6 by the visual G protein, transducin. Transducin-activated PDE6 exhibits an exceptionally high catalytic rate of cGMP hydrolysis ensuring high signal amplification. To identify the structural determinants for the inhibitory interaction with Pgamma and the remarkable cGMP hydrolytic ability, we sought to reproduce the PDE6 characteristics by mutagenesis of PDE5, a related cyclic GMP-specific, cGMP-binding PDE. PDE5 is insensitive to Pgamma and has a more than 100-fold lower k(cat) for cGMP hydrolysis. Our mutational analysis of chimeric PDE5/PDE6alpha' enzymes revealed that the inhibitory interaction of cone PDE6 catalytic subunits (PDE6alpha') with Pgamma is mediated primarily by three hydrophobic residues at the entry to the catalytic pocket, Met(758), Phe(777), and Phe(781). The maximal catalytic rate of PDE5 was enhanced by at least 10-fold with substitutions of PDE6alpha'-specific glycine residues for the corresponding PDE5 alanine residues, Ala(608) and Ala(612). The Gly residues are adjacent to the highly conserved metal binding motif His-Asn-X-X-His, which is essential for cGMP hydrolysis. Our results suggest that the unique Gly residues allow the PDE6 metal binding site to adopt a more favorable conformation for cGMP hydrolysis.     

Structural requirements for the binding of AMP to the allosteric site of NAD-specific isocitrate dehydrogenase from bakers' yeast

The specificity of yeast NAD-specific isocitrate dehydrogenase for the structures of the allosteric effector 5'-AMP was examined with analogues modified in the purine ring, pentosyl group, and 5'-phosphate group. An unsubstituted 6-amino group was essential for activation as was the phosphoryl group at the 5'-position. Activity was retained when an oxygen function of the 5'-phosphoryl was replaced by sulfur (Murry & Atkinson, 1968) or by nitrogen (phosphoramidates). 2-NH2-AMP, 2-azido-AMP, and 8-NH2-AMP were active; 8-azido-AMP and 8-Br-AMP were inactive. The configuration or nature of substituents about carbons 2' and 3' of the pentosyl portion of AMP was not critical for allosteric activation since AMP analogues containing, e.g., 2',3'-dideoxyribose or the bulky 2',3'-O-(2,4,6-trinitrocyclo-hexadienylidene) substituent (TNP-AMP) were active. TNP-AMP was bound to the enzyme with fluorescence enhancement and had an S0.5 for activation similar to the S0.5 for AMP. Positive effector activity was decreased when the pentosyl moiety of 5'-AMP was replaced by the six-membered nitrogen-containing morpholine group, indicating that the pentosyl group may be critical as a spacer for the proper geometry of binding to enzyme at the 6-amino and 5'-phosphoryl groups of 5'-AMP. A comparison of molecular models of 5'-AMP with 8,5'-cycloAMP suggests that the species of 5'-AMP required for binding to the enzyme contains the purine and ribose moieties in an anti conformation and positioning of the 5'-phosphate trans with respect to carbon 4'.(ABSTRACT TRUNCATED AT 250 WORDS)     

Cyclic nucleotide phosphodiesterases of Hyalophora cecropia silkmoth fat body.

Two soluble forms of 3':5'-cyclic-nucleotide phosphodiesterase (o':5'-cyclic-nucleotide 5'-nucleotidohydrolase, EC 3.1.4.17) were found in the larval fat body of the silkmoth Hyalophora cecropia. These differ in elution profile on Sephadex G-200, solubility in ammonium sulfate, metal ion requirements and kinetic properties. Phosphodiesterase I has Km values of 11 muM and 1.8 muM for cyclic AMP and cyclic GMP, respectively, has 5-fold greater maximal activity with cyclic AMP than with cyclic GMP, and is activated by Mg2+ and Co2+, and inhibited by EDTA. phosphodiesterase II has Km values of 625 muM and 125 muM for cyclic AMP and cyclic GMP, respectively, has similar maximal activity with both substrates, and is not activated by divalent metal ions or inhibited by EDTA. Cyclic nucleotides and methylxanthines competitively inhibit both enzymes. Phosphodiesterase is found in both soluble and particulate fractions of homogenates. Total activity is highest during the larval stage of the insect, drops markedly following pupation, and rises again during pharate adult development.     

In resting COS1 cells a dominant negative approach shows that specific, anchored PDE4 cAMP phosphodiesterase isoforms gate the activation, by basal cyclic AMP production, of AKAP-tethered protein kinase A type II located in the centrosomal region

We employ a novel, dominant negative approach to identify a key role for certain tethered cyclic AMP specific phosphodiesterase-4 (PDE4) isoforms in regulating cyclic AMP dependent protein kinase A (PKA) sub-populations in resting COS1 cells. A fraction of PKA is clearly active in resting COS1 cells and this activity increases when cells are treated with the selective PDE4 inhibitor, rolipram. Point mutation of a critical, conserved aspartate residue in the catalytic site of long PDE4A4, PDE4B1, PDE4C2 and PDE4D3 isoforms renders them catalytically inactive. Overexpressed in resting COS1 cells, catalytically inactive forms of PDE4C2 and PDE4D3, but not PDE4A4 and PDE4B1, are constitutively PKA phosphorylated while overexpressed active versions of all these isoforms are not. Inactive and active versions of all these isoforms are PKA phosphorylated in cells where protein kinase A is maximally activated with forskolin and IBMX. By contrast, rolipram challenge of COS1 cells selectively triggers the PKA phosphorylation of recombinant, active PDE4D3 and PDE4C2 but not recombinant, active PDE4A4 and PDE4B1. Purified, recombinant PDE4D3 and PDE4A4 show a similar dose-dependency for in vitro phosphorylation by PKA. Disruption of the tethering of PKA type-II to PKA anchor proteins (AKAPs), achieved using the peptide Ht31, prevents inactive forms of PDE4C2 and PDE4D3 being constitutively PKA phosphorylated in resting cells as does siRNA-mediated knockdown of PKA-RII, but not PKA-RI. PDE4C2 and PDE4D3 co-immunoprecipitate from COS1 cell lysates with 250 kDa and 450 kDa AKAPs that tether PKA type-II and not PKA type-I. PKA type-II co-localises with AKAP450 in the centrosomal region of COS1 cells. The perinuclear distribution of recombinant, inactive PDE4D3, but not inactive PDE4A4, overlaps with AKAP450 and PKA type-II. The distribution of PKA phosphorylated inactive PDE4D3 also overlaps with that of AKAP450 in the centrosomal region of COS1 cells. We propose that a novel role for PDE4D3 and PDE4C2 is to gate the activation of AKAP450-tethered PKA type-II localised in the perinuclear region under conditions of basal cAMP generation in resting cells.