A strain (MAR33) of Saccharomyces cerevisiae containing a threonine deaminase [L-threonine hydrolyase (deaminating) EC 4.2.1.16] with decreased feedback sensitivity has been shown to have a specific activity of acetohydroxy acid synthetase higher than that of the parent strain (MD11) when both are grown on minimal medium. When strain MAR33 is grown on minimal medium supplemented only with isoleucine, the specific activity of the synthetase is reduced to that found in the parent strain. Another strain, D106-1A, contains a nonsense mutation in the middle of the gene for threonine deaminase. When this strain is grown on minimal medium containing appropriate supplements (which include a nonrepressing concentration of isoleucine), or on minimal medium supplemented with isoleucylglycine (which acts as a limiting source of isoleucine), acetohydroxy acid synthetase remains repressed. Leucine limitation causes partial derepression. With the reversion of the nonsense mutation, either intragenically or via a suppressor for the mutation, partial derepression of the synthetase returns. When D106-1A is diploidized with either M15, a mutant lacking the synthetase, or MD9, a strain containing the enzyme, normal, partially derepressed, values for this enzyme are found. This indicates that threonine deaminase is necessary for derepression, and that it possibly acts as an inducer.