DNA sequences required for specific and efficient initiation of transcription at the polyoma virus early promoter

The 5'-flanking DNA sequences involved in the specific and efficient transcription of the polyoma virus early region have been investigated. Sequence requirements for efficient in vivo expression differed from those in vitro. Deletion of DNA located between 200 and 400 base pairs before the principal cap sites severely inhibited in vivo expression as measured by transformation ability, but did not affect in vitro transcription. Viable deletion mutants which lack the principal cap sites and the "TATA" box were very poor templates for in vitro transcription. Analysis of other deletion mutants in vitro demonstrated that no specific sequences more than 46 base pairs before the cap sites were important. Removal of the TATA box reduced in vitro transcriptional efficiency but did not alter the initiation sites. The synthesis of transcripts with abnormal 5' termini did not occur in vitro until sequence between the TATA box and the normal cap sites was also deleted. We further observed a nonspecific requirement for 90 to 100 base pairs of DNA 5' to the cap site for optimal transcription of DNA fragments in vitro.