Over the last decades obesity has become one of the mayor health burdens worldwide and the development of highly efficient drugs is of utmost importance. More than 40 different G-protein coupled receptors (GPCR) are expressed in different hypothalamic nuclei. They are essentially involved in regulation of food intake, body weight regulation and energy homeostasis, however, their detailed functions and impact on metabolism is not yet understood. So far targeting hypothalamic GPCRs in humans is not very satisfactory as drugs only lead to weight-loss of approximately 10 percent of body weight and have several side effects. One reason for side effects and the partial ineffectiveness might be that pharmacological interventions at one targeted GPCR should also affect interacting and therefore, functionally linked protein(s). Unfortunately, recognition of the full repertoire of interacting partners of GPCRs is not yet unraveled in the entirety. To determine new interacting partners for hypothalamic GPCRs we developed a hypothalamic library screen based on bimolecular fluorescence complementation. By screening for interaction partners of the melanocortin 3 and 4 receptor (MC3R and MC4R) we identified a variety of unexpected potential interactions (transmembrane and cytosolic proteins) that could link these receptors to e.g. glucose metabolism and inflammation on a molecular level. Several of these new interaction partners were confirmed by further experimental approaches and functional consequences of interaction were determined. Strikingly, we could demonstrate in vitro that the functionality of MC4R and MC3R is modulated indeed by so far unknown interacting proteins. Our study clearly points to a potentially much higher complexity of the melanocortin system in energy metabolism then previously assumed. We suggest that this has to be taken into consideration when targeting melanocortin receptors to combat obesity. Finally, under these aspects functional insights into GPCRs should be generally considered as incomplete and not be extrapolated easily to in vivo systems or pathophysiological situations without knowledge of specific interacting proteins.