Homogeneous protein-protein interaction assays without the need of a separation step are an essential tool to unravel signal transduction events in live cells. We have established an isoform specific protein kinase A (PKA) subunit interaction assay based on bioluminescence resonance energy transfer (BRET). Tagging human Ralpha(I)-, Ralpha(II)-, as well as Calpha-subunits of PKA with Renilla luciferase (Rluc) as the bioluminescent donor or with green fluorescent protein (GFP2) as the energy acceptor, respectively, allows to directly probe PKA subunit interaction in living cells as well as in total cell extracts in order to study side by side PKA type I versus type II holoenzyme dynamics. Several novel, genetically encoded cAMP sensors and-for the first time PKA type I sensors-were generated. When C- and R-subunits are assembled to the respective holoenzyme complexes inside the cell, BRET occurs with a signal up to three times above the background. An increase of endogenous cAMP levels as well as treatment with the cAMP analog 8-Br-cAMP is reflected by a dose-dependent BRET signal reduction in cells expressing wild type proteins. In contrast to type II, the dissociation of the PKA type I holoenzyme complex was never complete in cells with maximally elevated cAMP levels. Both sensors dissociated completely upon treatment with 8-Br-cAMP after cell lysis, consistent with in vitro activation assays using holoenzymes assembled from purified PKA subunits. Interestingly, incubation of cells with the PKA antagonist Rp-8-Br-cAMPS leads to a significant BRET signal increase in cells expressing PKA type I or type II isoforms, indicating a stabilization of the holoenzyme complexes in vivo. Mutant RI subunits with reduced (hRIalpha-R210K) or abolished (hRIalpha-G200E/G324E) cAMP binding capability were studied to quantify maximal signal to noise ratios for the RI-BRET sensor. Utilizing BRET we demonstrate that PKA type II holoenzyme was rendered insensitive to beta-adrenergic receptor stimulation with isoproterenol when anchoring to the plasma membrane of COS-7 cells was disrupted by either using Ht31 peptide or by depletion of membrane cholesterol.