The 3' enhancer of the T cell receptor s.chain (TCR~) gene directs the tissue-and stage-specific expression and V(D)J recombination of this gene locus. Using an in vitro system that reproduces TCRoL enhancer activity efficiently, we show that long-range promoter-enhancer regulation requires a T cell-specific repressor complex and is sensitive to DNA topology. In this system, the enhancer functions to derepress the promoter on supercoiled, but not relaxed, templates. We find that the TCRoL promoter is inactivated by a repressor complex that contains the architectural protein HMG I/Y. In the absence of this repressor complex, expression of the TCR~ gene is completely independent of the 3' enhancer and DNA topology. The interaction of the T cell-restricted protein LEF-1 with the TCR~ enhancer is required for promoter derepression. In this system, the TCR~ enhancer increases the number of active promoters rather than the rate of transcription. Thus, long-range enhancers function in a distinct manner from promoters and provide the regulatory link between repressors, DNA topology, and gene activity.[Key Words: TCR genes; transcription; enhancers; HMG I/Y; derepression; DNA topology] Received December 27, 1996; revised version accepted January 14, 1997.The widespread importance of long-range promoterenhancer interactions in tissue-specific and developmentally regulated gene expression has been documented extensively. Some well-known examples include the immunoglobulin, f~-globin, insulin, B-interferon, ~-fetoprotein, ovalbumin, and T-cell receptor genes, as well as the Drosophila alcohol dehydrogenase Adh, fushi tarazu, and ecdysone-regulated glue genes (for review, see Wasylyk 1988). These studies have yielded a wealth of information about enhancer-dependent transcription, but they have not provided mechanistic data to explain how these control elements actually work. For example, it is currently unknown how a gene must be packaged in the nucleus for natural enhancers to function. In this regard, the requirement for chromatin structure to generate long-range transcriptional regulation has been demonstrated in vitro using reconstituted DNA templates (Laybourn and Kadonaga 1992;Barton and Emerson 1994). It also remains to be determined how promoters communicate with distal enhancers. Several models have been proposed: DNA looping by the association of proteins bound at distal sites, DNA tracking by protein translocation, and long-range interactions affected by DNA topology (for review, see Ptashne 1986; Wang and 1Corresponding author.