The discovery of a cold-labile cytosolic acetyl-CoA hydrolase of high activity in rat liver by Prass et al. [(1980) J. Biol. Chem. 255, 5215-5223] has questioned the importance of mitochondrial acetyl-CoA hydrolase for the formation of free acetate [Grigat et al. (1979) Biochem. J. 177, 71-79] under physiological conditions. Therefore this problem has been reevaluated by comparing various properties of the two enzymes. Cold-labile cytosolic acetyl-CoA hydrolase bands with an apparent Mr of 68000 during SDS/polyacrylamide gel electrophoresis, while the native enzyme elutes in two peaks with apparent Mr of 136000 and 245000 during gel chromatography in the presence of 2 mM ATP. The mitochondrial enzyme elutes under the same conditions with an apparent Mr of 157000. Under conditions where the cold-labile enzyme binds strongly to DEAE-Bio-Gel and ATP-agarose, the mitochondrial enzyme remains unbound. The cold-labile enzyme can be activated 14-fold by ATP, half-maximal activation occurring already at 40 microM ATP. AdoPP[NH]P, AdoPP[CH2]P and GTP have a similar though weaker effect. ADP as well as GDP can completely inhibit the cold-labile enzyme with 50% inhibition occurring for both nucleotides at about 1.45 microM. The binding of ATP and ADP is competitive. Acetyl phosphate and pyrophosphate have no effect on the activity of the cold-labile enzyme. The mitochondrial acetyl-CoA hydrolase is not affected by these nucleotides. CoASH is a strong product inhibitor (approximately equal to 80% inhibition at 40 microM CoASH) of the cold-labile enzyme, but only a weak inhibitor of the mitochondrial enzyme. Under in vivo conditions the activity of the cold-labile cytosolic acetyl-CoA hydrolase can be no more than 7% of the activity calculated for mitochondrial acetyl-CoA hydrolase under the same conditions. Accordingly the mitochondrial enzyme seems to be mainly responsible for the formation of free acetate by the intact liver, especially in view of the fact that the substrate specificity of the mitochondrial enzyme is much higher (activity ratios acetyl-CoA/butyryl-CoA 4.99 and 1.16 for the mitochondrial and the cold-labile enzyme respectively). Alloxan diabetes neither increased the activity of the cold-labile enzyme nor that of the mitochondrial enzyme. No experimental support has been found yet for the hypothesis that the acetyl-CoA hydrolase activity of the cold-labile enzyme represents the side-activity of an acetyl-transferase.