Using fluorescence resonance energy transfer (FRET) microscopy, we investigate how heterotrimeric G proteins interact with G protein-coupled receptors (GPCRs). In the absence of receptor activation, the ␣2A adrenergic and muscarinic M4 receptors are present on the cell membrane as dimers. Furthermore, there is an interaction between the G protein subunits ␣, 1, and ␥2 and a number of GPCRs including M4, ␣2A, the adenosine A1 receptor, and the dopamine D2 receptor under resting conditions. The interaction between GPCRs and G␣ proteins shows specificity: there is interaction between the ␣2A receptor and Go, but little interaction between the ␣2A receptor and Gs. In contrast, the predominantly Gs-coupled prostacyclin receptor interacted with Gs, but there was little interaction between the prostacyclin receptor and Go. Inverse agonists did not change the FRET ratio, whereas the addition of agonist resulted in a modest fall. Our work suggests that GPCR dimers and the G protein heterotrimer are present at the cell membrane in the resting state in a pentameric complex.T wo opposing ideas are invoked to explain how membrane bound signaling proteins transfer information after activation. In the first, components in the membrane freely diffuse and interactions occur through ''collision coupling'' determined by diffusion. Historically, such mechanisms are thought to govern the interaction of G protein-coupled receptor (GPCR) with G protein and the interaction of G protein with downstream enzymes and ion channels. Signal amplification is a key feature of this mechanism (1-3). A second mechanism is the ''physical scaffolding'' hypothesis in which component proteins interact directly or indirectly with each other. The best example of this is the role of InaD in the Drosophila photoreceptor that scaffolds via PDZ domains the light-sensing GPCR rhodopsin, Ca 2ϩ influx TRP channels, phospholipase C, and protein kinase C (4). In principle, this is a way of generating fast activation, fast signal termination, and specificity. A variant of this hypothesis is the localization of proteins in membrane signaling microdomains such as caveolae and lipid rafts.The G protein-gated K ϩ channel (GIRK) was first identified in atrial myocytes. Channel activation occurs after binding of acetylcholine to muscarinic M2 receptors (5) and is responsible for slowing of the heart rate in response to vagal stimulation (6, 7). Analogous GIRK currents are present in neurons and neuroendocrine cells (8). Activation of native and cloned GIRK channels has been shown to involve a direct, membrane-delimited interaction with the G␥ subunit (9, 10). One critical point is that the activation occurs rapidly in both native and heterologous settings: complete channel activation can occur within 1 s of the addition of agonist (11-13). Such fast rates of signaling suggest that the components diffuse only small distances, if at all. From these considerations alone it is an appealing hypothesis to propose that the components may be physically scaffolded together. O...