A paradoxical microcirculatory constriction has been observed in heart of patients with ischemia secondary to coronary stenosis. Here, using the isolated mouse heart (Langendorff), we examined the mechanism of this response, assuming involvement of nitric oxide (NO) and endothelin-1 (ET-1) systems. Perfusion pressure was maintained at 65 mmHg for 70 min (Protocol 1), or it was reduced to 30 mmHg over two intervals, between the 20- and 40-min marks (Protocol 2) or from the 20-min mark onwards (Protocol 3). In Protocol 1, coronary resistance (CR) remained steady in untreated heart, while it progressively increased during treatment with the NO synthesis inhibitor L-NAME (2.7-fold) or the ET_A antagonist BQ610 (2.8 fold). The ET_B antagonist BQ788 had instead no effect by itself but curtailed vasoconstriction to BQ610. In Protocol 2, hypotension raised CR by 2.2-fold. This response was blunted by reactive oxygen species (ROS) scavengers (mannitol; superoxide dismutase plus catalase) and was converted into vasodilatation by L-NAME, BQ610, or BQ788. Restoration of normal pressure was followed by vasodilatation and vasoconstriction, respectively, in untreated and treated preparations. In Protocol 3, CR progressively increased with hypotension in the absence but not presence of L-NAME or BQ610. We conclude that coronary vasculature is normally relaxed by two concerted processes, a direct action of NO and ET-1 curtailing an ET_B2 mediated tonic vasoconstriction through ET_A activation. The negative feedback mechanism on ET_B2 subsides during hypotension and the ensuing vasoconstriction is ascribed to ET-1 activating ET_A and ET_B2 and reactive nitrogen oxide species originating from ROS-NO interaction.