Considerable controversy surrounds the role of protein kinase C (PKC) in ischemic preconditioning (PC). Previous studies have used pharmacological agents and/or measured total myocardial PKC activity; however, no information is available regarding the effects of PC on individual isoforms in vivo. We performed a comprehensive evaluation (using Western immunoblotting) of the expression and subcellular distribution of all 11 currently known PKC isoforms in the heart of conscious rabbits subjected to four different ischemic PC protocols known to induce early and/or late PC (one, three, or six cycles of 4-minute coronary occlusion [4'O]/4-minute reperfusion [4'R]; four cycles of 5-minute occlusion [5'O]/10-minute reperfusion [10'R]). Ten PKC isoforms (alpha, beta1/beta2, gamma, delta, epsilon, zeta, eta, iota, lambda, and mu) were found to be expressed in the rabbit heart. Quantitative immunoblotting demonstrated that as a subgroup, conventional PKCs (cPKCs) are more abundant than novel PKCs (nPKCs) (1445 versus 313 pg PKC/microg tissue protein, respectively) and that PKC alpha is the predominant isoform among the cPKCs (alpha, beta1, beta2, and gamma), representing 51% of this subgroup, and PKC epsilon is the most abundant among the nPKCs (delta, epsilon, zeta, and eta), accounting for 62% of this subgroup. None of the ischemic PC protocols examined caused appreciable changes in total PKC activity, in the subcellular distribution of total PKC activity, or in the subcellular distribution of PKC isoforms alpha, beta1/beta2, gamma, delta, zeta, iota, lambda, and mu. In contrast, all PC protocols caused significant translocation of PKC epsilon and PKC eta isoforms from the cytosolic to the particulate fraction. The particulate fraction of PKC epsilon increased in a dose-dependent fashion with the number of occlusion/reperfusion cycles performed, from 35+/-2% in the control group to 43+/-2% after one 4'O/5-minute reperfusion (5'R) cycle (P<.05), 52+/-2% after three cycles (P<.05 versus one cycle), and 66+/-3% after six cycles (P<.05 versus three cycles). The particulate fraction of PKC epsilon also increased, after four 5'O/10'R cycles, to 50+/-3% (P<.05 versus control). In contrast to PKC epsilon, the translocation of PKC eta was independent of the number of occlusion/reperfusion cycles performed. The particulate fraction of PKC eta increased from 67+/-3% in the control group to 84+/-2% after one 4'O/5'R cycle (P<.05), 84+/-2% after three 4'O/4'R cycles (P<.05), 86+/-3% after six 4'O/4'R cycles (P<.05), and 83+/-2% after four 5'O/10'R cycles (P<.05). When expressed as a percentage of control values, the increases in the particulate fraction of isoform epsilon were greater than those of isoform eta. The effects of 4'O without reperfusion were similar to those of one cycle of 4'O/5'R, indicating that 5'R did not attenuate isoform translocation. This is the first study to demonstrate PKC translocation after ischemic PC in vivo. The results indicate that in the conscious rabbit, ischemic PC causes selective translocation of the epsilon and eta isoforms without demonstrable changes in total myocardial PKC activity, implying that measurements of total PKC activity are not sufficiently sensitive to detect the involvement of PKC in PC. The results are consistent with the concept that the epsilon and eta isozymes play an important role in the genesis of ischemic PC in the conscious rabbit.