In the cardiac myocyte, β1 adrenoceptors (AR) couple with Gs to produce cAMP signals that propagate throughout the cell, resulting in PKA phosphorylation of targets at the sarcolemma (L type Ca2+ channel), sarcoplasmic reticulum (phospholamban; PLB, ryanodine receptor; RyR) and myofilaments (troponin I; TnI). β2 AR also couple with Gs, but concomitant coupling with Gi limits the magnitude and spread of the cAMP signal. We have previously shown that caveolae, invaginated lipid rafts, are responsible for Gi-dependent restriction of β2 AR cAMP signalling (1). Here we determine the extent of caveolar control of this pathway. Adult rat ventricular myocytes were isolated enzymatically; some cells were treated with 1 mM methyl-β-cyclodextrin (MBCD) to disrupt caveolae. β2 AR stimulation was achieved with 10 µM zinterol in the presence of 300 nM CGP20712A. [Ca2+]i transients and shortening were measured in fura 2-loaded cells; ICa,L was recorded in ruptured patch configuration; cAMP was indexed using FRET-based genetically encoded sensors for cAMP (2) in tethered PKA domains (using a PKA II-based probe) and throughout the cytosol (using an Epac2-based probe); protein phosphorylation was measured by Western blotting. Following β2 AR stimulation, control cells showed no increase in the amplitude of ICa,L, [Ca2+]i transient, shortening or PKA probe response (P>0.05 vs baseline; n=3-12 cells; Student’s t-test). The Epac probe response increased by 4.3 ± 0.8 ΔR/R0 (n=6). No increase in phosphorylation of Ser16 PLB, Ser2809 RyR, or Ser23/24 TnI was detected following β2 AR stimulation in these cells (P>0.05 vs baseline; cells from n=5-6 hearts). By contrast, in MBCD-treated cells, β2 AR stimulation resulted in a significant increase in [Ca2+]i transient amplitude (21.1 ± 4.1% over baseline) and shortening (70.2 ± 9.7% over baseline), a marked lusitropic effect (P<0.05; n=12-20), but no change in ICa,L (P>0.05; n=4). In these cells, β2 AR stimulation produced an Epac probe response identical to controls (4.2 ± 0.5 ΔR/R0; n=6) but also a PKA probe response (4.7 ± 1.4 ΔR/R0; n=13) not observed in controls. MBCD-treated cells showed a 5-fold increase in phosphorylation of PLB (P<0.05) but no change in phosphorylation of RyR or TnI (P>0.05; n=5-6). Our data show that inotropic and lusitropic responses to β2 AR stimulation seen in caveolae-disrupted cells result from preferential access of the cAMP-dependent signal to PLB. Others have recently shown that cAMP levels close to PLB are specifically limited by a Gi-Gβγ-PI3K-PDE4 signalling cascade (3). This finding is consistent with our conclusion that caveolae selectively control β2 AR-mediated cAMP signalling in a PKA II sub-compartment of the sarcoplasmic reticulum containing PLB.