Lyso-PS (lyso-phosphatidylserine) has been shown to activate the G(i/o)-protein-coupled receptor GPR34. Since in vitro and in vivo studies provided controversial results in assigning lyso-PS as the endogenous agonist for GPR34, we investigated the evolutionary conservation of agonist specificity in more detail. Except for some fish GPR34 subtypes, lyso-PS has no or very weak agonistic activity at most vertebrate GPR34 orthologues investigated. Using chimaeras we identified single positions in the second extracellular loop and the transmembrane helix 5 of carp subtype 2a that, if transferred to the human orthologue, enabled lyso-PS to activate the human GPR34. Significant improvement of agonist efficacy by changing only a few positions strongly argues against the hypothesis that nature optimized GPR34 as the receptor for lyso-PS. Phylogenetic analysis revealed several positions in some fish GPR34 orthologues which are under positive selection. These structural changes may indicate functional specification of these orthologues which can explain the species- and subtype-specific pharmacology of lyso-PS. Furthermore, we identified aminoethyl-carbamoyl ATP as an antagonist of carp GPR34, indicating ligand promiscuity with non-lipid compounds. The results of the present study suggest that lyso-PS has only a random agonistic activity at some GPR34 orthologues and the search for the endogenous agonist should consider additional chemical entities.