The longstanding question of the presence of mitochondria-bound polysomes has been recently revisited using new approaches. Genome-wide analyses provided evidence that many genes are actually translated on mitochondria-bound polysomes and GFP-labeling techniques have shown that, in vivo, the 3'UTR sequence of these genes contains signals which can target hybrid RNA molecules to the proximity of mitochondria. Evolutionary conservation of some of these signals will be presented. Interestingly, class I mRNA which are translated on free polysomes and class II mRNA which are translated on mitochondria-bound polysomes have, mostly, eukaryotic and prokaryotic origins respectively. Using ATP2, a typical prokaryotic-derived gene, as a model for class II mRNA, we showed that its 3'UTR sequence is essential both for a correct addressing of mRNA to mitochondria proximity and to a proper production of functional ATP synthases. These different observations suggest that prokaryotic-derived genes are, like the contemporary mitochondrial genes, translated near mitochondrial membranes. In both cases this locus specific translation process might be connected to a correct complex assembly program and the cases of ATP synthase and cytochrome c oxidase complexes will be considered in this respect.