In most studies, coupling of the beta2-adrenoceptor (beta2AR) to the stimulatory, heterotrimeric GTP-binding protein of adenylyl cyclase the (Gs) is studied indirectly by measuring adenylyl cyclase activation. The aim of this study was to establish a model system in which beta2AR-Gs interactions can be studied directly at the level of the G-protein. We expressed the beta2AR alone, in combination with the alpha-subunit of Gs (G(s alpha)), and as fusion protein with G(s alpha) (beta2AR-G(s alpha)) in Sf9 insect cells. The beta2AR expressed alone couples poorly to the endogenous G(s alpha)-like G-protein of Sf9 cells since no high-affinity agonist binding could be detected, and the effects of agonist and inverse agonist on adenylyl cyclase, high-affinity GTPase and guanosine 5'-O-(3-thiotriphosphate) (GTP[S]) binding were small. Beta2AR-G(s alpha) reconstituted high-affinity agonist binding and regulated adenylyl cyclase more effectively than the beta2AR co-expressed with a large excess of G(s alpha). In membranes expressing beta2AR-G(s alpha), highly effective agonist- and inverse agonist regulation of high-affinity GTP hydrolysis and GTP[S] binding was observed. In contrast, agonist and inverse agonist regulation of GTP hydrolysis and GTP[S] binding in membranes expressing beta2AR and G(s alpha) as separate proteins was difficult to detect. Our data show that the beta2AR interacts with G(s alpha) more efficiently when expressed as a fusion protein than when expressed with an excess of non-fused G(s alpha). The beta2AR-G(s alpha) fusion protein provides a very sensitive model system to study the regulation of Gs function by beta2AR agonists and inverse agonists directly at the level of the G-protein.