Characterization of vaccinia virus early promoters and evaluation of their informational content.

We have reported the isolation of cis-acting regulatory DNA sequences promoting expression of the herpes virus thymidine kinase gene in vaccinia virus recombinants. In this work we show that each of the inserts from recombinants VpT25, 28, 36 and 56 contains a vaccinia virus early promoter. The position of each of the early RNA start sites in the nucleotide sequence of these four vaccinia virus inserts was precisely mapped by an S1 nuclease mapping procedure. Among the four recombinants analysed only VpT56-infected cells also contained a substantial amount of a transcript with the same 5' end at late period. The insert present in VpT25 contained a new late RNA start site 50 nucleotides upstream from that of the early RNA. The four inserts were mapped on the vaccinia virus genome. We also localized the 5' end of the mRNA of a vaccinia virus host-range gene, whose DNA nucleotide sequence has recently been established. The 45 nucleotides preceding the RNA start site from most of 19 known vaccinia virus early promoters were found to be A + T-rich (at least 80%) and contained shorter A-rich (at least 60%) regions, beginning approximately 25 nucleotides upstream from the RNA start site. The information content, as expressed by the parameter Rsequence, of early vaccinia virus promoters revealed ten bits of information in the sequence of 28 nucleotides upstream from the early RNA start sites. Most of the information needed to locate an early promoter is contained within the nucleotide sequence upstream from an RNA start site. A consensus sequence consists of two blocks: the sequence AA(A/T)N(T/A)N(A/G)AAAANAANA starting at position -27 and the sequence (T/A)(C/T)N(A/T)T(A/G) starting at position -5. It was concluded that vaccinia virus early promoters may be characterized by an A + T-rich region of approximately 45 nucleotides preceding the RNA start site and include a specific 3'-terminal sequence of 28 nucleotides containing at least ten bits of information. A procedure for localizing putative early RNA start sites in nucleotide sequences is proposed.