OBJECTIVES: In atrial myocytes, excitation-contraction coupling is controlled by large axial tubule junctions (AT-junctions) with the sarcoplasmic reticulum that activate rapid Ca2+ induced Ca2+ release. Previously, we identified a high phosphorylation level of junctional Ryanodine Receptor (RyR2) Ca2+ release channels, whereas non-junctional RyR2 clusters are low phosphorylated under basal conditions but can be activated by beta-adrenergic stimulation. Junctional RyR2 clusters are constitutively phosphorylated by protein kinase A at Serin2808 and Ca2+/calmodulin-dependent protein kinase II at Serin2814, while the subcellular mechanisms are yet unknown. Here, we aim to investigate the level of 3’,5’-cyclic adenosine monophosphate (cAMP) at axial tubule junctions by highly localized confocal FRET imaging in direct vicinity of RyR2.

RESULTS: Confocal immunofluorescence imaging revealed that junctin, a RyR2 binding protein of the sarcoplasmic reticulum, is highly colocalized with RyR2 at axial tubule junctions in atrial myocytes. Consequently, a new Epac1-based FRET biosensor coupled to junctin (Epac1-JNC) enabled us to monitor cAMP in direct vicinity of junctional RyR2. Co-immunoprecipitation confirmed the protein-protein interaction of Epac1-JNC and endogenous junctin with RyR2 in atrial myocytes, respectively. To contrast local cAMP levels at tubule junctions vs. subsurface Ca2+ release sites, we developed non-invasive confocal FRET imaging protocols for living atrial myocytes. Our data showed significantly decreased FRET ratios i.e. increased cAMP levels at AT-junctions (FRET ratio mean±SEM: 0.93±0.02) vs. subsurface (FRET ratio: 1.07±0.05; p<0.05) Ca2+ release sites under basal conditions. This difference was diminished by adenylyl cyclase inhibition (FRET ratio: AT-junctional 1.23±0.05 vs. subsurface 1.39±0.08; not significant), indicating compartmentalized cAMP pools at AT-junctions. Vice versa, beta-adrenergic stimulation with isoprenaline decreased FRET signals i.e. increased cAMP levels (FRET ratio: AT-junctional 0.85±0.05 vs. subsurface 1.0±0.05; not significant) and overcame cAMP compartmentalization. Moreover, confocal immunofluorescence imaging confirmed that adenylyl cyclase inhibition diminished junctional RyR2 phosphorylation at Serin2808 (median [25%-75%] RyR2-pS2808/RyR2 intensity: control 2.89 [2.32-4.12] vs. MDL 1.49 [1.11-2.0]; p<0.01) resulting in greatly reduced Ca2+ release (F/F0: control 9.12±0.58 vs. MDL 7.02±0.87; p<0.001). To assess the impact of constitutively increased cAMP levels to L-type Ca2+ currents, we combined FRET imaging and voltage-clamp which revealed a reduction in L-type Ca2+ channel current density shortly after cAMP levels decrease upon adenylyl cyclase inhibition (pA/pF: control -5.74±0.95 vs. MDL -1.46±0.26; p<0.05). Finally, while adenylyl cyclase V/VI expression levels are similar between atrial and ventricular myocytes, immunofluorescence imaging revealed adenylyl cyclase VI signals at AT-junctions in atrial myocytes.

CONCLUSIONS: Our data identify a compartmentalized cAMP nanodomain at junctional Ca2+ release units in atrial myocytes that is maintained by constitutively increased adenylyl cyclase activity. This cell-specific mechanism is required for intact excitation-contraction coupling and rapid junctional Ca2+ release in healthy atrial myocytes. These findings significantly expand our understanding of unique atrial physiology to identify therapeutic targets in atrial cardiomyopathy.