DNA sequence requirements for transcriptional initiator activity in mammalian cells

A transcriptional initiator (Inr) for mammalian RNA polymerase II can be defined as a DNA sequence element that overlaps a transcription start site and is sufficient for (i) determining the start site location in a promoter that lacks a TATA box and (ii) enhancing the strength of a promoter that contains a TATA box. We have prepared synthetic promoters containing random nucleotides downstream of Sp1 binding sites to determine the range of DNA sequences that convey Inr activity. Numerous sequences behaved as functional Inrs in an in vitro transcription assay, but the Inr activities varied dramatically. An examination of the functional elements revealed loose but consistent sequence requirements, with the approximate consensus sequence Py Py A+1 N T/A Py Py. Most importantly, almost every functional Inr that has been described fits into the consensus sequence that we have defined. Although several proteins have been reported to bind to specific Inrs, manipulation of those elements failed to correlate protein binding with Inr activity. The simplest model to explain these results is that all or most Inrs are recognized by a universal binding protein, similar to the functional recognition of all TATA sequences by the same TATA-binding protein. The previously reported proteins that bind near specific Inr elements may augment the strength of an Inr or may impart transcriptional regulation through an Inr.