A variety of functional assays are available for agonist or antagonist screening of G protein-coupled receptors (GPCRs), but it is a priori not predictable which assay is the most suitable to identify agonists or antagonists of GPCRs with therapeutic value in humans. More specifically, it is not known how a given set of GPCR agonists compares in different functional assays with respect to potency and efficacy and whether the level of the signaling cascade that is analyzed has any impact on the detection of agonistic responses. To address this question, the authors used the recently cloned human S1P(5) receptor as a model and compared a set of 3 lipid ligands (sphingosine 1-phosphate [S1P], dihydro sphingosine 1-phosphate [dhS1P], and sphingosine) in 5 different functional assays: GTPgammaS binding, inhibition of adenylyl cyclase activity, mobilization of intracellular Ca(2+) via the FLIPR and aequorin technology, and MAP kinase (ERK1/2) activation. S1P induced agonistic responses in all except the ERK1/2 assays with EC(50) values varying by a factor of 10. Whereas dhS1P was identified as a partial agonist in the GTPgammaS assay, it behaved as a full agonist in all other settings. Sphingosine displayed partial agonistic activity exclusively in GTPgammaS binding assays. The findings suggest that assays in a given cellular background may vary significantly with respect to suitability for agonist finding and that ligands producing a response may not readily be detectable in all agonist assays.