cAMP is a powerful second messenger whose known general effector is protein kinase A (PKA). The identification of a cAMP binding protein, Epac, raises the question of its role in Ca(2+) signalling in cardiac myocytes. In this study, we analysed the effects of Epac activation on Ca(2+) handling by using confocal microscopy in isolated adult rat cardiomyocytes. [Ca(2+)](i) transients were evoked by electrical stimulation and Ca(2+) sparks were measured in quiescent myocytes. Epac was selectively activated by the cAMP analogue 8-(4-chlorophenylthio)-2'-O-methyladenosine-3',5'-cyclic monophosphate (8-CPT). Patch-clamp was used to record the L-type calcium current (I(Ca)), and Western blot to evaluate phosphorylated ryanodine receptor (RyR). [Ca(2+)](i) transients were slightly reduced by 10 microm 8-CPT (F/F(0): decreased from 4.7 +/- 0.5 to 3.8 +/- 0.4, P < 0.05), an effect that was boosted when cells were previously infected with an adenovirus encoding human Epac. I(Ca) was unaltered by Epac activation, so this cannot explain the decreased [Ca(2+)](i) transients. Instead, a decrease in the sarcoplasmic reticulum (SR) Ca(2+) load underlies the decrease in the [Ca(2+)](i) transients. This decrease in the SR Ca(2+) load was provoked by the increase in the SR Ca(2+) leak induced by Epac activation. 8-CPT significantly increased Ca(2+) spark frequency (Ca(2+) sparks s(-1) (100 microm)(-1): from 2.4 +/- 0.6 to 6.9 +/- 1.5, P < 0.01) while reducing their amplitude (F/F(0): 1.8 +/- 0.02 versus 1.6 +/- 0.01, P < 0.001) in a Ca(2+)/calmodulin kinase II (CaMKII)-dependent and PKA-independent manner. Accordingly, we found that Epac increased RyR phosphorylation at the CaMKII site. Altogether, our data reveal a new signalling pathway by which cAMP governs Ca(2+) release and signalling in cardiac myocytes.