Oxygen is critical to aerobic metabolism, but hyperoxia is cytostatic and cytocidal. The precise mechanisms involved in hyperoxic cell injury remain incompletely understood although there is substantial support for the possibility that hyperoxia increases the 'leak' of electrons from the mitochondrial electron transport chain and the resulting increased generation of reactive oxygen species (ROS) might explain the toxic effects of high oxygen. To evaluate the possible role of mitochondria in oxygen toxicity, we used cloned HeLa cells having no functional mitochondria (?° cells) as a model. In contrast to wild-type HeLa cells, ?° cells survive and grow in 80% O2 but are equally susceptible to oxidant-mediated killing. Two other strategies which diminish mitochondrial ROS generation (exposure of cultures to either carbonyl cyanide mchlorophenylhydrazone [CCCP] or chloramphenicol) also increase tolerance of wild-type HeLa cells to hyperoxia. Thus, three different maneuvers which minimize mitochondrial ROS production improve cell survival and growth under hyperoxia. This suggests that mitochondrial ROS production is primarily responsible for hyperoxic damage. Results of experiments with an oxygen tolerant strain of HeLa cells, which proliferates even under 80% O2, termed 'HeLa-80', lend further support to this general concept. In this oxygen tolerant cell line, antioxidant defenses are similar to the parent line of HeLa cells. In addition, these two cell lines are equally susceptible to killing by hydrogen peroxide and t-butyl hydroperoxide. However, under both 20% and 80% O2, intracellular ROS production is >2-fold higher in HeLa-20 cells compared to HeLa-80 cells as assessed by (1) dihydrodicholorfluorescein oxidation, (2) dihydroethidium oxidation, (3) hyperoxiamediated suppression of aconitase activity and (4) mitochondrial protein carbonyl content. Diminished ROS production may be related to the fact that the oxygen-tolerant HeLa-80 cells have significantly higher cytochrome c oxidase (COX) activity (~ 2.0 fold) than HeLa-20, and preferential blockade of this terminal complex of mitchondrial electron transport by treatment with n-methyl protoporphyrin, which selectively diminishes synthesis of heme-aa3 in COX, abrogates the tolerance of HeLa-80 cells to hyperoxia and increases ROS production in HeLa-80 cells under hyperoxia. In the oxygen tolerant HeLa cells, the elevated COX activity may be due to a >2-fold increase in COX Vb (a regulatory subunit of COX) whereas expression levels of three other COX subunits (COX I, II and IV) are very close to wild-type. Finally, preliminary results indicate a similar over-expression of COX activity in a unique strain of oxygen tolerant rats. Overall, our results suggest that it is