Finding. Human uncoupling protein 1 is not a carbon copy of the mouse protein. Expressed in mouse liver and read in isolated mitochondria, both proteins leak protons on their own: more substrate-supported respiration, lower membrane potential, no extra fatty acid required. Mouse UCP1 then behaves as the textbooks say, with a strong GDP brake and oleate reactivation. Human UCP1 barely listens to GDP, stays very fatty-acid responsive, and is braked by ATP at an apparent IC50 near 0.4 mM. Molecular dynamics put a GDP contact on mouse F88 that human S88 does not keep.
Why this paper matters
Almost every mechanistic sentence about UCP1 is a rodent sentence. If human BAT is going to be a metabolic therapy idea, that is a problem. This paper does the unglamorous, necessary experiment: put both proteins in the same mitochondrial background and ask whether the nucleotide rules travel.
They do not, not fully. Shared innate uncoupling means the human protein is a real protonophore in this assay, not a passenger. Divergent nucleotide control means GDP-centric rodent pharmacology will mis-rank human UCP1. ATP as the tighter human nucleotide, and ATP blunting oleate reactivation, is the regulatory rewrite.
What they actually measured
Adeno-associated viral expression in mouse liver, then isolated-mitochondria bioenergetics. Both UCP1s raise substrate-supported respiration and drop membrane potential without exogenous fatty acids. Mouse UCP1: potent GDP inhibition, oleate reactivation. Human UCP1: weak GDP, strong fatty-acid response, ATP IC50 ≈ 0.4 mM versus ≈ 1.4 mM for GDP, and ATP-desensitized oleate reactivation. OXPHOS capacity and OXPHOS protein levels stay intact, so the uncoupling is not a wrecked respiratory chain. MD: persistent GDP–F88 in mouse, absent at S88 in human; in-silico F88S weakens that GDP contact.
Liver is a feature and a limit. Feature: a low-UCP1 organelle background. Limit: not the brown-fat inner membrane, not adrenergic activation, not in vivo heat.
How to read the score
High seventies to low eighties. Head-to-head human versus mouse UCP1 with numbers and a residue hypothesis is rare and useful. Confidence is medium. IC50s are apparent, the F88S story is computational until someone swaps the residue in mitochondria, and liver is a heterologous chassis. Score 82 for the mitochondrial field; it changes how you cite “GDP-sensitive UCP1” in human sentences.
What to do with it
If you model human thermogenesis, BAT pharmacology, or UCP1 reconstitution, switch the default nucleotide to ATP and keep oleate in the model. Pull the titration curves and the MD contact table. Do not claim F88S is proven physiology. The directional implication is that human UCP1 is an innate uncoupler under ATP, not GDP, control.