The activity of acetyl-CoA carboxylase, measured in various ways, was studied in 15000g extracts of rat liver hepatocytes and compared with the rate of fatty acid synthesis in intact hepatocytes incubated with insulin or glucagon. Hepatocyte extracts were prepared by disruption of cells with a Dounce homogenizer or by solubilization with 1.5% (v/v) Triton X-100. Sucrose-density-gradient centrifugation demonstrated that the sedimentation coefficient of acetyl-CoA carboxylase from cell extracts was 30-35S, regardless of the conditions of incubation or disruption of hepatocytes. Solubilization of cells with 1.5% Triton X-100 yielded twice as much enzyme activity (measured by [14C]bicarbonate fixation) in the sucrose-gradient fractions as did cell disruption by the Dounce homogenizer. Analysis by high-performance liquid chromatography of acetyl-CoA carboxylase reaction mixtures showed that [14C]malonyl-CoA accounted for 10-60% of the total acid-stable radioactivity, depending on the method for disrupting hepatocytes and on the preincubation of the 15000g extract, with or without citrate, before assay. Under conditions in which incubation of cells with insulin or glucagon caused an activation or inhibition, respectively, of acetyl-CoA carboxylase, only 25% of the acid-stable radioactivity was [14C]malonyl-CoA and enzyme activity was only 13% (control), 16% (insulin), and 57% (glucagon) of the rate of fatty acid synthesis. Under conditions when up to 60% of the acid-stable radioactivity was [14C]malonyl-CoA and acetyl-CoA carboxylase activity was comparable with the rate of fatty acid synthesis, there was no effect of insulin or glucagon on enzyme activity.