The metabolism of leucine to isoamyl alcohol in yeast was examined by 13 C nuclear magnetic resonance spectroscopy. The product of leucine transamination, ␣-ketoisocaproate had four potential routes to isoamyl alcohol. The first, via branched-chain ␣-keto acid dehydrogenase to isovaleryl-CoA with subsequent conversion to isovalerate by acyl-CoA hydrolase operates in wild-type cells where isovalerate appears to be an end product. This pathway is not required for the synthesis of isoamyl alcohol because abolition of branched-chain ␣-keto acid dehydrogenase activity in an lpd1 disruption mutant did not prevent the formation of isoamyl alcohol. A second possible route was via pyruvate decarboxylase; however, elimination of pyruvate decarboxylase activity in a pdc1 pdc5 pdc6 triple mutant did not decrease the levels of isoamyl alcohol produced. A third route utilizes ␣-ketoisocaproate reductase (a novel activity in Saccharomyces cerevisiae) but with no role in the formation of isoamyl alcohol from ␣-hydroxyisocaproate because cell homogenates could not convert ␣-hydroxyisocaproate to isoamyl alcohol. The final possibility was that a pyruvate decarboxylase-like enzyme encoded by YDL080c appears to be the major route of decarboxylation of ␣-ketoisocaproate to isoamyl alcohol although disruption of this gene reveals that at least one other unidentified decarboxylase can substitute to a minor extent.In most eukaryotes, the catabolism of the branched-chain amino acids leucine, isoleucine, and valine has been well understood for many years (1). The first step is a transamination in which ␣-ketoglutarate accepts the amino group (from leucine, isoleucine, and valine) producing glutamate and ␣-ketoisocaproic acid, ␣-keto--methylvaleric acid and ␣-ketoisovaleric acid, respectively. Next is oxidative decarboxylation of the keto acids by branched-chain ␣-keto acid dehydrogenase to the corresponding acyl-CoA derivatives. Further steps yield, ultimately, acetyl-CoA and acetoacetate (from leucine), acetyl-CoA and propionyl-CoA (from isoleucine), and succinyl-CoA (from valine). All of these metabolites can enter the tricarboxylic acid (TCA) 1 cycle. It has been known for many years that yeasts do not operate the same metabolic routes because branched-chain amino acids can serve as the sole source of nitrogen but not carbon (2, 3). The predominant view, in a rather sparse literature, is that yeasts first use transamination but that decarboxylation of the keto acids proceeds via a "carboxylase" to an aldehyde that is then reduced in an NADH-linked reaction producing the appropriate "fusel" alcohol (2-4). This scheme is sometimes called the "Ehrlich pathway" to honor the originator of the ideas (5), which were slightly modified later (6). Acceptance of the so-called Ehrlich pathway is problematical for at least four reasons. First, the supposed pathway has never been proven to exist. Simply showing that e.g. radioactively labeled leucine is converted into isoamyl alcohol does not prove that the individual steps are those envisaged in ...