Extinction is defined as the loss of cell type-specific gene expression that occurs in somatic cell hybrids derived by fusion of cells with dissimilar phenotypes. To explore the basis of this dominant-negative regulation, we have studied the activities of the control elements of the liver-specific gene encoding tyrosine aminotransferase (TAT) in hepatoma/fibroblast hybrid crosses. We show that extinction in complete somatic cell hybrids is accompanied by the loss of activity of all known cell type-specific control elements of the TAT gene. This inactivity is the result of first, lack of expression of genes coding for the transcriptional activators HNF4 and HNF3[~ and HNF33,, which bind to essential elements of the enhancers; and second, loss of in vivo binding and activity of ubiquitous factors to these enhancers, including CREB, which is the target for repression by the tissue-specific extinguisher locus TSE1. Complete extinction of TAT gene activity is therefore a multifactorial process affecting all three enhancers controlling liver-specific and hormone-inducible expression. It results from lack of activation, rather than active repression, and involves both post-translational modification and loss of essential transcriptional activators.[Key Words: Extinction; tyrosine aminotransferase; chromatin; HNF3; HNF4; somatic cell hybrids] Received September 14, 1992; revised version accepted November 23, 1992.A particular cellular phenotype, characterized by the expression of distinct sets of genes, is thought to be determined by the action of positively and negatively acting regulatory proteins. This combinatorial control of gene activity involves the binding of multiple trans-acting factors to distinct cis regulatory elements. The differentiated state usually has to be actively maintained and continually transmitted (for review, see Blau 1992). Striking examples for this requirement of continual regulation are presented by the autoregulatory loops discovered in the maintenance of the muscle phenotype involving MyoD as a positive regulator (for review, see Weintraub et al. 1991) and by proteins of the polycomb group, among which are negative regulators required for differential expression of homeotic genes along the anteriorposterior axis of the Drosophila embryo (for review, see Paro 1990).Whereas many cell type-specific positive regulators have been characterized, the role of negatively acting regulatory proteins in specifying a particular cellular phenotype is less clear, although several examples of repressor-like activities have been described (for review,