Carbon monoxide (CO) produced by heme oxygenase (HO)-1 and HO-2 or released from the CO-donor, tricarbonyldichlororuthenium (II) dimer (CORM-2) causes vasodilation, with unknown efficacy against stress-induced gastric lesions. We studied whether pretreatment with CORM-2 (0.1-10 mg/kg oral gavage (i.g.)), RuCl₃ (1 mg/kg i.g.), zinc protoporphyrin IX (ZnPP) (10 mg/kg intraperitoneally (i.p.)), hemin (1-10 mg/kg i.g.) and CORM-2 (1 mg/kg i.g.) combined with N(G)-nitro-l-arginine (l-NNA, 20 mg/kg i.p.), 1H-[1,2,4]oxadiazolo[4,3-a]quinoxalin-1-one (ODQ, 10 mg/kg i.p.), indomethacin (5 mg/kg i.p.), SC-560 (5 mg/kg i.g.), and celecoxib (10 mg/kg i.g.) affects gastric lesions following 3.5 h of water immersion and restraint stress (WRS). Gastric blood flow (GBF), the number of gastric lesions and gastric CO and nitric oxide (NO) contents, blood carboxyhemoglobin (COHb) level and the gastric expression of HO-1, HO-2, hypoxia inducible factor 1α (HIF-1α), tumor necrosis factor α (TNF-α), cyclooxygenase (COX)-2 and inducible NO synthase (iNOS) were determined. CORM-2 (1 mg/kg i.g.) and hemin (10 mg/kg i.g.) significantly decreased WRS lesions while increasing GBF, however, RuCl₃ was ineffective. The impact of CORM-2 was reversed by ZnPP, ODQ, indomethacin, SC-560 and celecoxib, but not by l-NNA. CORM-2 decreased NO and increased HO-1 expression and CO and COHb content, downregulated HIF-1α, as well as WRS-elevated COX-2 and iNOS mRNAs. Gastroprotection by CORM-2 and HO depends upon CO's hyperemic and anti-inflammatory properties, but is independent of NO.