© 2012 Nature America, Inc. All rights reserved. NATURE CHEMICAL BIOLOGY | ADVANCE ONLINE PUBLICATION | www.nature.com/naturechemicalbiology 1 ARTICLE PUBLISHED ONLINE: 27 MAY 2012 | DOI: 10.1038/NCHEMBIO.961 B eyond the classical view of GPCRs working linearly as on- off switches that simultaneously trigger all possible cellular functions mediated by a specific receptor, the notion of ligand-directed or ligand-biased signaling emerged recently. According to this concept, different ligands stabilize distinct receptor states, ‘selecting’ specific effector activation and fine- tuning cellular responses 1 . This notion of ‘ligand-directed traffick- ing’ emphasizes the function of receptors as filters capable of generating textured responses to ligands, thus leading to the con- cept of “pluridimensional efficacy” 2 . In keeping with this concept, one challenging task that remains is the optimization of highly sensitive pharmacological assays to fully capture the nuances of ligand activity. Given the pleiotropic signal- ing properties of GPCRs, this complexity needs to be analyzed from different points of view to be globally understood. Furthermore, crosstalk between different signaling pathways and regulatory events taking place at different levels of individual pathways makes analysis more complicated the further downstream one looks from receptor activation at the plasma membrane. One strategy to over- come these limitations would be to explore the initial step common to all GPCR-family signaling pathways: activation of heterotrimeric G proteins. Numerous studies have analyzed GPCR ligands in search of possible G protein–biased agonists, but most relied on indirect evaluations of G protein involvement by measuring downstream G protein–effector activities 3 . Moreover, although GPCRs can couple multiple G protein isoforms, direct and accurate evaluation of selective ligand efficacy for different G protein isoforms remains elusive. To date, [ 35 S]GTPγS binding is the most widely used method to directly analyze G protein activity, but it suffers from poor sensiti- vity, is restricted mainly to analysis of the Gα i/o family and cannot distinguish among Gα i/o isoforms 4 . Improvements have been made with immunocapture of G proteins of interest after [ 35 S]GTPγS bind- ing 4 , although a lack of specific antibodies remains an impediment. Another difficulty with this assay is the use of purified cell mem- branes, which can exclude potential G protein or receptor regula- tors influencing ligand efficacy. Other means have been developed to bypass such limitations, including receptor–G protein fusions, chimeras and biophysical approaches such as surface plasmon resonance 4 . However, there is still no tool available to accurately measure the activation profile of G protein isoforms. Using new bioluminescence resonance energy transfer 2 (BRET)- based biosensors to directly measure activation of all G protein fam- ilies in living cells, we found that the well-known β-arrestin–biased agonist SII 5–7 acted as a partial agonist on G protein activation at angiotensin II (Ang II)-receptor type 1A (AT 1A -R). Notably, dissec- tion of downstream effector activation revealed that SII acts through a mechanism totally different from that of Ang II and not simply via a β-arrestin bias. Our results demonstrate for the first time, to our knowledge, the notion that a biased agonist does not automatically mimic part of the physiological agonist but rather can act as a new agonist with unique signaling output. RESULTS Development of BRET-based G protein activation biosensors We have previously described a direct BRET sensor that allows real-time detection of receptor-mediated G protein activation in living cells by measuring interactions between heterotrimeric Gαβγ subunits. For that purpose, the Renilla reniformis luciferase (RLuc) BRET energy donor was inserted after amino acid 91 in the helical domain of Gα i (Gα i1 -91RLuc) 8 . When we measured the interaction of this BRET donor with the energy acceptor, 1 Institut des Maladies Métaboliques et Cardiovasculaires, Institut National de la Santé et de la Recherche Médicale, Université Toulouse III Paul Sabatier, Toulouse, France. 2 Department of Cellular and Molecular Biology, Pierre Fabre Research Institute, Castres, France. 3 Laboratory for Molecular Cardiology, Danish National Research Foundation Centre for Cardiac Arrhythmia, Department of Diabetes NBEs and Obesity Biology, University of Copenhagen, Copenhagen, Denmark. 4 Department of Pharmacology, Centre Hospitalier Universitaire de Toulouse, Toulouse, France. 5 Present address: Novo Nordisk A/S, Novo Nordisk Park, GLP-1 & Obesity Biology Section, Måløv, Denmark. 6 Deceased. *e-mail: celine.gales@inserm.fr Deciphering biased-agonism complexity reveals a new active AT 1 receptor entity Aude Saulière 1 , Morgane Bellot 1 , Hervé Paris 1,6 , Colette Denis 1 , Frédéric Finana 2 , Jonas T Hansen 3 , Marie-Françoise Altié 1 , Marie-Hélène Seguelas 1 , Atul Pathak 1,4 , Jakob L Hansen 5 , Jean-Michel Sénard 1,4 & Céline Galés 1 * Functional selectivity of G protein–coupled receptor (GPCR) ligands toward different downstream signals has recently emerged as a general hallmark of this receptor class. However, pleiotropic and crosstalk signaling of GPCRs makes functional selectivity difficult to decode. To look from the initial active receptor point of view, we developed new, highly sensitive and direct bioluminescence resonance energy transfer–based G protein activation probes specific for all G protein isoforms, and we used them to evaluate the G protein–coupling activity of [ 1 Sar 4 Ile 8 Ile]-angiotensin II (SII), previously described as an angio- tensin II type 1 (AT 1 ) receptor–biased agonist that is G protein independent but b-arrestin selective. By multiplexing assays sensing sequential signaling events, from receptor conformations to downstream signaling, we decoded SII as an agonist stabilizing a G protein–dependent AT 1A receptor signaling module different from that of the physiological agonist angiotensin II, both in recombinant and primary cells. Thus, a biased agonist does not necessarily select effects from the physiological agonist but may instead stabilize and create a new distinct active pharmacological receptor entity.