Sinoatrial nodal cells (SANC) express Ca2+-activated adenylate cyclase (AC) isoforms that generate a high basal level of cAMP-mediated, protein kinase A (PKA)-dependent Ca2+ cycling protein phosphorylation, resulting in spontaneous rhythmic Ca2+ oscillations that ignite the surface membrane to generate rhythmic APs, i.e. pacemaker automaticity. Differences in basal phosphodiesterase (PDE) and AC activities and in PDE:AC activity within protein microenvironments in SANC are a potential mechanism for compartmental regulation of cAMP levels, leading to local differences in the effectiveness of cAMP signaling within these microdomains. The present study measured the AC activity and PDE activity in SANC in detergent-resistant microdomains (DRM) of SANC lysates to determine the PDE and AC activities in lipid raft gradients indexed by GM-1 and caveolin-3 immunolabeling. The microdomain Ca2+ dependence of AC activity, and the relative abundance of microdomain PDE types, based upon the effects of specific PDE inhibitors, were also determined. Under the conditions of our assay, PDE and AC activities are nearly identically matched in the fractions that contain higher densities of lipid raft markers. As the lipid raft density decreases below a threshold, the PDE:AC activity ratio becomes increased: from 10 to more than 200 fold, depending upon the Ca2+ milieu, while neither the relative extent of AC activation by different Ca2+ significantly varied among the different microenvironments, the Ca2+ milieu in lipid raft-rich fractions affects the matching of PDE to AC activities. The most optimal milieu for cAMP production and survival is when the Ca2+ is 1 microM, and the lipid raft marker density is high; when Ca2+ is reduced to 200 nM or heavily buffered by BAPTA, however, PDE activity exceeds that of AC, even in lipid raft-rich fractions. The very high PDE activity and lower AC activity in non-lipid raft domains favors degradation of cAMP, not only maintaining a reduced local level of cAMP production, but also creating a functional barrier for diffusion of cAMP from lipid rafts into non-lipid raft domains or into the cytosol. Most of the transcripts coding PDE catalytic subunits demonstrated significant difference in level of expression in SANC and LV. mRNA level of PDE 1A, 3B, 4B and 5A was higher in SANC cells then in LV cells. Moreover, expression level of PDE1A, 3B and 4B was significantly different not only between SANC and LV but also between RA and SANC (p<0.0001, p<0.05 and p<0.05 respectively). For the rest of the transcripts expression level in SANC was lower than in LV cells (PDE1C, 3A, 4A, 4D, 7A). We investigated if mRNA expression of differentially expressed PDE1A and PDE4B is correlated with protein expression. Due to lack of highly specific anti-rabbit antibodies, we created custom antibodies against these antigens. Subsequent Western blot analysis confirmed a pattern of PDE1A and PDE4B expression in SANC, LV and RA. Western blot analysis of rabbit cardiac tissues demonstrated presence of PDE3A isoform in a high amount in left ventricle and in less amounts in right atrium and sinoatrial node. Further Western Blot analyses revealed presence of PDE4B and PDE4D proteins in the lysates of rabbit sinoatrial nodal and left ventricle tissues. Co-Immunolabeling of PDE1A and the membrane bound Potassium/Sodium channel HCN4 in isolated rabbit SANC observed by Structured Illumination Microscopy (SIM) revealed PDE1A is found primarily just beneath the surface membrane region of the cell where it is likely to be in close proximity to AC and can function in the moderation of high basal cAMP levels. Biochemical experiments on cell lysates demonstrated that from cAMP-hydrolyzing subtypes of PDEs, PDE1, PDE2 and PDE4 are present in SANC. PDE1 represents the highest activity (43%) in our conditions. The most important PDE in LV are PDE3 and PDE4. PDE4 represents up to 50% of total PDE activity in LV. We found that inhibition of protein phosphatases in our experiments and increase in phosphorylation of cellular proteins causes activation of total PDE activity in both cell types and PDE1 and PDE4 in SANC. Addition of calmodulin dramatically increased PDE activity in the lysates of both cell types, revealing previously undetectable PDE1 in LV. We tried to elucidate microenvironment of PDEs within cells. In order to do this we performed several experiments using immunoprecipitation techniques. We found that mAKAP scaffolding protein co-precipitated PDE activity in rabbit VM lysates. We were able to immunoprecipitate PDE1 from rabbit SANC lysates but unfortunately the efficiency of this reaction was too low for further analysis of the co-immunoprecipitated proteins. Due to the restricted amount of sinoatrial tissue which we can get from rabbits, in order to improve the efficiency of data collection we decided to use the mouse cardiac cell line (HL1 cells) as a model of cells with pacemaker activity. In order to make sure that HL-1 cell line is an appropriate system we have performed QPCR detection of PDE1 subtypes and found distribution similar to that in rabbit heart atrial cells. We employed genetic manipulation in HL-1 cells (siRNA targeting PDE1A) to determine their effect on the spontaneous beating rate. On average in response to its mRNA silencing (n= 4) PDE1A expression level was reduced to 30 % of control and the beating rate was increased by 70%. We employed FRET (Fluorescence resonance energy transfer) technology to detect cAMP level in HL-1 cells to investigate the impact of PDE1A on intracellular cAMP level. FRET detection of cAMP in the single live beating cell revealed two-fold increase in response to IBMX treatment. In contrast, in HL-1 cells transfected with siRNA targeting PDE1A, we detected absence of cAMP increase in response to IBMX. These results indicate the important role of PDE1A in controlling intracellular level of cAMP and pacemaker function. Next we investigated the role of ACs in HL-1 beating activity. We measured the effect of siRNA-directed ADCY1 inhibition on HL-1 cell beating rate and cAMP level in cytoplasm and plasma membrane. Transfection of HL-1 cells with AC1 siRNA completely inhibited the IBMX-induced increase in cAMP production in cytoplasm and plasma membrane, and strongly inhibited the IBMX-induced increase in beating frequency of HL-1 cells. Pharmacological inhibition of both plasma membrane and cytoplasmic AC completely abolished the IBMX-induced increase of basal cAMP concentration as well. Inhibition of cytoplasmic AC resulted in complete cessation of spontaneous beating in HL-1 cells. Thus, as in adult sinoatrial nodal pacemaker cells, basal AC activity is required for the spontaneous beating in HL-1 cardiac cells. In experiments on Wild type (WT) and AC8-transgenic (AC8-TG) mice with cardiac-specific AC8 overexpression, as expected, we found that AC activity was significantly increased in AC8-TG SAN and LV tissue lysates. The activity was highly responsive to forskolin activation, and in LV tissue lysates it was highly concentrated in the membrane fraction. PDE activity in WT and AC8-TG groups was responsive to addition of Calcium and Calmodulin, and the level of detected activity highly correlated with the substrate concentration in the biochemical reaction. At the lowest level of cAMP substrate in the reaction, we found that total PDE reaction was significantly higher in AC8-TG LV lysates compare to WT group. Total PDE activity even at the lowest studied cAMP was higher than AC activity in both groups. Proteomics data analysis performed on WT and AC8-TG mice LVs demonstrated increase in the amount of PDE3A, PDE4A, PDE4D, PDE7B and PDE8A proteins in AC8-TG mice versus WT mice.
