The Cat8p zinc cluster protein is essential for growth of Saccharomyces cerevisiae with nonfermentable carbon sources. Expression of the CAT8 gene is subject to glucose repression mainly caused by Mig1p. Unexpectedly, the deletion of the Mig1p-binding motif within the CAT8 promoter did not increase CAT8 transcription; moreover, it resulted in a loss of CAT8 promoter activation. Insertion experiments with a promoter test plasmid confirmed that this regulatory 20-bp element influences glucose repression and derepression as well. This finding suggests an upstream activating function of this promoter region, which is Mig1p independent, as delta mig1 mutants are still able to derepress the CAT8 promoter. No other putative binding sites such as a Hap2/3/4/5p site and an Abf1p consensus site were functional with respect to glucose-regulated CAT8 expression. Fusions of Cat8p with the Gal4p DNA-binding domain mediated transcriptional activation. This activation capacity was still carbon source regulated and depended on the Cat1p (Snf1p) protein kinase, which indicated that Cat8p needs posttranslational modification to reveal its gene-activating function. Indeed, Western blot analysis on sodium dodecyl sulfate-gels revealed a single band (Cat8pI) with crude extracts from glucose-grown cells, whereas three bands (Cat8pI, -II, and -III) were identified in derepressed cells. Derepression-specific Cat8pII and -III resulted from differential phosphorylation, as shown by phosphatase treatment. Only the most extensively phosphorylated modification (Cat8pIII) depended on the Cat1p (Snf1p) kinase, indicating that another protein kinase is responsible for modification form Cat8pII. The occurrence of Cat8pIII was strongly correlated with the derepression of gluconeogenic enzymes (phosphoenolpyruvate carboxykinase and fructose-1,6-bisphosphatase) and gluconeogenic PCK1 mRNA. Furthermore, glucose triggered the dephosphorylation of Cat8pIII, but this did not depend on the Glc7p (Cid1p) phosphatase previously described as being involved in invertase repression. These results confirm our current model that glucose derepression of gluconeogenic genes needs Cat8p phosphorylation and additionally show that a still unknown transcriptional activator is also involved.