Diminished coronary reactive hyperaemia (RH) is associated with high risk of cardiovascular events. Adenosine, acting through its four receptor subtypes, plays important roles in RH. However the individual role of each receptor subtype in RH and underlying mechanisms remains unknown. Using gene-deleted mice combined with pharmacological approaches, we observed a decreased RH (similar to 25% decrease from wide type) in Langendorff-perfused hearts from A(2A) AR knockout (KO) or WT mice pretreated with SCH58261(2-(2-Furanyl)-7-(2-phenylethyl)-7H-pyrazolo[4,3-e][1,2,4] triazolo[1,5-c] pyrimidin-5-amine), a selective A2A AR antagonist. In contrast, deletion of A(1) AR augmented RH (similar to 30% increase from WT) which was abolished by A(2A) AR blockade. No difference was observed in RH between A(2B), A(3) KO and WT mice. Catalase, an enzyme that breaks down H2O2, significantly attenuated RH in WT but had no effect in A(2A) KO mice or SCH58261-treated WT mice. Fluorescent imaging on isolated coronary arteries illustrated that adenosine induced a higher magnitude of H2O2 production (2.7 +/- 0.54 vs. 1.5 +/- 0.17 times baseline) in A(1) KO compared with WT mice which was blocked by SCH58261. Glibenclamide, an ATP-sensitive potassium channel (K-ATP) blocker, also attenuated RH, but to a higher extent than catalase. Of importance, SCH58261, catalase, and glibenclamide all abolished the augmented RH in A(1) KO mice. Taken together, our results suggest that A(1) and A(2A) ARs counteractively modulate RH via modulating H2O2 production and subsequent opening of K-ATP channels in smooth muscle cells (SMCs) of coronary artery, and deletion of A1 AR leads to a removal of A(2A) AR inhibition, thus resulting in an enhanced H2O2 production, opening of KATP channels, and an increased RH.