There is a futile cycle of pump and leak of protons across the mitochondrial inner membrane. The contribution of the proton cycle to standard metabolic rate is signi®cant, particularly in skeletal muscle, and it accounts for 20% or more of the resting respiration of a rat. The mechanism of the proton leak is uncertain: basal proton conductance is not a simple biophysical leak across the unmodi®ed phospholipid bilayer. Equally, the evidence that it is catalysed by homologues of the brown adipose uncoupling protein, UCP1, is weak. The yeast genome contains no clear UCP homologue but yeast mitochondria have normal basal proton conductance. UCP1 catalyses a regulated inducible proton conductance in brown adipose tissue and the possibility remains open that UCP2 and UCP3 have a similar role in other tissues, although this has yet to be demonstrated.

The mitochondrial Ca 2+ uniporter (MCU) has been characterized for several decades [1], but the molecule responsible for this transport activity is currently unknown. Therefore, a recent Nature Cell Biology paper by Trenker et al. [2] reporting a role for uncoupling proteins (UCP2/3) in mitochondrial Ca 2+ uniport generated much excitement in the field. Subsequently, the authors contend that Ca 2+ transport accounts for most physiologic effects assigned to UCPs [3].The defining characteristic of mitochondrial Ca 2+ uptake is that mitochondrial matrix Ca 2+ ([Ca 2+ ] m ) responds to changes in cytosolic Ca 2+ at a rate dependent on both the driving force (the electrochemical gradient comprising membrane potential (Δψ m ) plus Ca 2+ concentration gradient), and the transporter activity [4]. The steady-state free [Ca 2+ ] m depends on the balance between Ca 2+ uptake and efflux [4]. Trenker et al. report that UCP overexpression increased the free [Ca 2+ ] m attained upon agonist stimulation (Fig. 1c in [2]), apparently without effect on the rate of Ca 2+ uptake. When mitochondrial Ca 2+ efflux via the Na + /Ca 2+ exchanger was blocked using CGP37157, and the driving force for Ca 2+ entry (Δψ m ) was collapsed by oligomycin plus FCCP, agonist-stimulated free [Ca 2+ ] m was again higher in UCP overexpressing cells, with no discernible effect on Ca 2+ uptake rate (Fig. 1f in [2]). In such a system, where both Ca 2+ uptake and efflux are inactive, an effect of UCP expression on steady-state free [Ca 2+ ] m levels cannot be due to changes in Ca 2+ transport. Since overexpression of UCPs can sometimes result in a non-functional uncoupled phenotype thought to be due to protein misfolding [5], and the apparent lack of effect of UCP overexpression on Δψ m (supplement Fig. 1d ]), the significance of these data is also difficult to assess.Alternative to the MCU, the Na + /Ca 2+ or H + /Ca 2+ exchangers could serve a mitochondrial Ca 2+ uptake function, in which UCPs may play a role. However, the V max of these systems is 2 or 18 nmol Ca 2+ /min/mg protein for liver or heart mitochondria respectively, i.e. 2-3 orders of magnitude lower than the V max of the MCU (1700 nmol Ca 2+ /min/mg protein) [6,7]. Thus UCPs cannot significantly affect mitochondrial Ca 2+ uptake through such mechanisms.In Fig. 3 of [2], Ca 2+ uptake by isolated murine liver mitochondria was only partially inhibited by 100 nM ruthenium red (RR), although it is not clear whether crude RR was used (IC 50 ~1 µM for MCU activity) or its purified derivative Ru360 (IC 50 5 nM) [4,8]. Using this experimental system, a decrease in Ca 2+ uptake by liver mitochondria of Ucp2 −/− mice was observed relative to wild-type controls. However, liver mitochondria contain exceedingly small amounts of UCP2 (<1 ng UCP2/mg protein, vs. 25 or 75 ng UCP2/mg protein in mouse kidney or spleen mitochondria respectively; V. Azzu and M.D. Brand unpublished observations). Furthermore, the mitochondrial inner membrane surface area is 5 × 10 10 µm 2 / mg protein [9] and each µm 2 cont...