Mice deficient in mitochondrial uncoupling protein (UCP) 1 are cold sensitive, despite abundant expression of the homologues, Ucp2 and Ucp3 (S. Enerbäck, A. Jacobsson, E. M. Simpson, C. Guerra, H. Yamashita, M.-E. Harper, and L. P. Kozak. Nature 387: 90-94, 1997). We have analyzed characteristics of mitochondrial proton leak from brown adipose tissue (BAT) of Ucp1-deficient mice and normal controls and conducted the first top-down metabolic control analysis of oxidative phosphorylation in BAT mitochondria. Because purine nucleotides inhibit UCP1 and because UCP2 and the long form of UCP3 have putative purine nucleotide-binding regions, we predicted that proton leak in BAT mitochondria from Ucp1-deficient mice would be sensitive to GDP. On the contrary, although control over mitochondrial oxygen consumption and proton leak reactions at state 4 are strongly affected by 1 mM GDP in mitochondria from normal mice, there is no effect in UCP1-deficient mitochondria. In the presence of GDP, the overall kinetics of proton leak were not significantly different between Ucp1-deficient mice and controls. In its absence, state 4 respiration in normal BAT mitochondria was double that in its presence. Leak-dependent oxygen consumption was higher over a range of membrane potentials in its absence than in its presence. Thus proton leak, potentially including that through UCP2 and UCP3, is GDP insensitive. However, our measurements were made in the presence of albumin and may not allow for the detection of any fatty acid-induced UCP-mediated leak; this possibility requires investigation.