The ischemic activation of p38α mitogen-activated protein kinase (p38α-MAPK) is thought to contribute to myocardial injury. Under other circumstances, activation is through dual phosphorylation by MAPK kinase 3 (MKK3). Therefore, the mkk3^−/− murine heart should be protected during ischemia. In retrogradely perfused mkk3^−/− and mkk3^+/+ mouse hearts subjected to 30 minutes of global ischemia and 120 minutes of reperfusion, infarction/risk volume was similar (50±5 versus 51±4, P=0.93, respectively), as was intraischemic p38-MAPK phosphorylation (10 minutes ischemia as percent basal, 608±224 versus 384±104, P=0.43, respectively). This occurred despite undetectable activation of MKK3/6 in mkk3^−/− hearts. However, tumor necrosis factor (TNF)-induced p38-MAPK phosphorylation was markedly diminished in mkk3^−/− vs mkk3^+/+ hearts (percent basal, 127±23 versus 540±267, respectively, P=0.04), suggesting an MKK-independent activation mechanism by ischemia. Hence, we examined p38-MAPK activation by TAB1-associated autophosphorylation. In wild-type mice and mkk3^−/− mice, the p38-MAPK catalytic site inhibitor SB203580 (1 μmol/L) diminished phosphorylation during ischemia versus control (10 minutes ischemia as percent basal, 143±2 versus 436±96, P=0.003, and 122±25 versus 623±176, P=0.05, respectively) and reduced infarction volume (infarction/risk volume, 57±5 versus 36±3, P<0.001, and 50±5 versus 29±3, P=0.003, respectively) but did not alter TNF-induced activation, although in homogenates of ischemic hearts but not TNF-exposed hearts, p38-MAPK was associated with TAB1. Furthermore, adenovirally expressed wild-type and drug-resistant p38α-MAPK, lacking the SB203580 binding site, was phosphorylated when H9c2 myoblasts were subjected to simulated ischemia. However, SB203580 (1 μmol/L) did not prevent the phosphorylation of resistant p38α-MAPK. These findings suggest the ischemic activation of p38-MAPK contributing to myocardial injury is by TAB1-associated autophosphorylation.