Efficient transcription of the adenovirus Ela-inducible E4 gene is mediated by an Ela-dependent enhancer in vivo. In vitro, the enhancer functions constitutively in the absence of Ela, conferring high levels of transcription on the E4 and heterologous promoters. Binding of a cellular transcription factor, E4F1, to two sites within the E4 enhancer and to one other functionally important site located directly upstream of the E4 TATA box is required for transcriptional activity. The relationship between the enhancer and the TATA box proximal site is further demonstrated by the fact that the E4 enhancer can be functionally substituted by two copies of the TATA box proximal site. These and other results suggest that the E4 promoter may be comprised solely of multiple E4F1 binding sites and a TATA box. In addition to the E4 promoter, E4F1 interacts with other adenovirus early promoters and thus may be involved in the co-ordinate expression of Ela-inducible early viral genes.

The sequence motif CGTCA is critical for binding of a group of cellular transcription factors (ATF, CREB, E4F, and EivF) and for activation of certain Ela-inducible and cyclic AMP (cAMP)-inducible promoters. We have tested different promoter elements containing the CGTCA motif (referred to here as ATF-binding sites) for the ability to function as Ela or cAMP response elements. The adenovirus E4 promoter and the cellular vasoactive intestinal peptide (VIP) promoter responded differently to Ela and cAMP, demonstrating that the activating potential of ATF-binding sites within these promoters is not equivalent. While particular ATFbinding sites were sufficient for the activity of both the E4 (Ela inducibility) and VIP (cAMP inducibility) enhancers, these two enhancers had contrasting effects on Ela-and cAMP-inducible transcription. Thus, the relationship between Ela-and cAMP-inducible transcription is not simply explained by the action of these two inducers through the same promoter elements.Transcriptional activation of eucaryotic genes often allows specific gene expression in response to a variety of physiological and viral inducing agents. Activation occurs, in part, through binding of cellular transcription factors to specific sites within the affected promoter (for reviews, see references 13, 33, 34, and 42) and through protein-protein interactions between activators (35, 38, 51). In this way genes can be targeted for activation according to the particular array of factor-binding sites present in their control regions. It is now apparent that this simple view is unable to explain certain aspects of structure-function relationships for inducible promoters. Of salient interest is the fact that multiple factors (or promoter elements) with similar or identical DNA-binding specificity (or sequence) have different activation potential (12,16,19,25,41,45,48). The subtle mechanisms underlying control of the above factors are central to understanding differential gene activation.Promoter elements containing the sequence motif CGTCA (referred to throughout this paper as ATF-binding sites) have been shown to be critical for Ela inducibility of certain adenovirus early viral promoters (3,7,22,26,27,44) and for cyclic AMP (cAMP)-mediated activation of many neuropeptide promoters (5,9,11,15,28,36,37,43,47). As is the case with other eucaryotic transcriptional control elements (see above), a complex set of activator proteins bind directly to ATF sites (7,10,19,22,31,36 matostatin promoter (36). Of the above polypeptides, CREB (which is likely to correspond to ATF-43) has been directly implicated in activation of the somatostatin promoter by cAMP (36, 50), and ATF (19) and EivF (7) have been shown to function as transcriptional activators in vitro. Beyond this, the functional relationship between polypeptides that bind to ATF sites is unknown. The inability to distinguish most (but not all [7,39]) ATF-binding sites containing the invariant CGTCA motif (2,7,20,23,27,31) together with the effect of mutations on binding of nucl...