The metabolic control of respiration is still poorly understood, due mainly to the lack of suitable approaches for studying it in vivo. Experiments on isolated mammalian mitochondria have indicated that a relatively small fraction of each of several components of the electron transport chain is sufficient to sustain a normal O 2 consumption rate. These experiments, however, may not ref lect accurately the in vivo situation, due to the lack in the mitochondrial fraction of essential cytosolic components and to the use of excess of substrates in the in vitro assays. An approach is described here whereby the control of respiration by cytochrome c oxidase (COX; EC 1.9.3.1) was analyzed in intact cultured human osteosarcoma 143B.TK ؊ cells and other wild-type cells and in mitochondrial DNA mutation-carrying human cell lines. Surprisingly, in wild-type cells, only a slightly higher COX capacity was detected than required to support the endogenous respiration rate, pointing to a tighter in vivo control of respiration by COX than generally assumed. Cell lines carrying the MERRF mitochondrial tRNA Lys gene mutation, which causes a pronounced decrease in mitochondrial protein synthesis and respiration rates, revealed, in comparison, a significantly greater COX capacity relative to the residual endogenous respiration rate, and, correspondingly, a higher COX inhibition threshold above which the overall respiratory f lux was affected. The observed relationship between COX respiratory threshold and relative COX capacity and the potential extension of the present analysis to other respiratory complexes have significant general implications for understanding the pathogenetic role of mutations in mtDNA-linked diseases and the tissue specificity of the mutation-associated phenotype.The rapid accumulation of knowledge concerning mitochondrial diseases, especially those caused by mitochondrial DNA (mtDNA) mutations, has stimulated in recent years a strong interest in the metabolic control of oxidative phosphorylation (OXPHOS). In particular, the discovery of threshold effects in the capacity of a mtDNA mutation to produce an OXPHOS defect in the presence of varying amounts of wild-type mtDNA has called attention to the degree of control that a particular step exerts in the OXPHOS pathway. It has been emphasized that application of the metabolic control theory (1, 2) to the study of mitochondrial metabolism can be a valuable approach for determining the level of control exerted by different OXPHOS steps on the rate of mitochondrial respiration (3), and for identifying and quantifying enzymatic defects caused in the OXPHOS machinery by mitochondrial or nuclear DNA mutations (4-7). Recently, this approach has been applied to isolated rat tissue mitochondria, by using increasing concentrations of inhibitors of complex I (rotenone), complex III (myxothiazol or antimycin A), or complex IV (potassium cyanide) to mimic the effects of mutations affecting these complexes (4-7). These experiments have produced results suggesting that ...