From recent findings about the indirect effect of metformin (MET) targeted on the respiratory chain complex I, we reconsidered this question and tried to determine the causality of any alteration at this enzymatic level using Xenopus laevis oocytes. Addition of MET (50 microM) reduced by 40% the rotenone-sensitive activity of complex I only in incubating intact oocytes but not in mitochondria isolated by differential centrifugation. The drug prior injected inside these cells had also no measurable effect. In spite of this and the weak binding of MET to the mitochondrial fraction, there was a fairly good correlation between the marked inhibitory action of MET on complex I and its progressive appearance within the oocyte cytoplasm. The intriguing observation that MET as a liposomal form was again able to exert its role when added directly to isolated mitochondria is in accordance with a membrane-mediated uptake and vesicular routing of MET. Furthermore, a temperature-dependent effect was clearly shown. At 4 degrees, oocytes failed to take up efficiently MET and accordingly its subsequent action on respiration was therefore lost. Likewise, MET transport was hindered and inhibition of complex I totally disappeared when a structural analog, asymmetrical dimethylarginine (ADMA), was placed together with MET either at an identical concentration or in excess. These data strongly support the view that MET may recognise some specific membranous sites, likely belonging to effector systems, before penetrating the cell in a bound state via an obscure endocytotic event which still has to be identified.