KOZAK序列Compilation and analysis of sequences upstream from the translational start site in eukaryo.pdfVIP

Volume 12 Number 2 1984 Nucleic Acids Research Compilation and analysis of sequences upstream from the translational start site in eukaryotic mRNAs Marilyn Kozak Department of Biological Sciences, University of Pittsburgh, Pittsburgh, PA 15260, USA Received 15 November 1983; Accepted 21 November 1983 ABSTRACT 5-Noncoding sequences have been tabulated for 211 messenger RNAs from higher eukaryotic cells. The 5-proximal AUG triplet serves as the initiator codon in 95% of the mRNAs examined. The most conspicuous conserved feature is the presence of a purine (most often A) three nucleotides upstream from the AUG ini- tiator codon; only 6 of the mRNAs in the survey have a pyrimidine in that posi- tion. There is a predominance of C in positions -1, -2, -4 and -5, just upstream from the initiator codon. The sequence CCGACCAUG(G) thus emerges as a consensus sequence for eukaryotic initiation sites. The extent to which the ribosome bind- ing site in a given mRNA matches the -1 to -5 consensus sequence varies: more than half of the mRNAs in the tabulation have 3 or 4 nucleotides in common with the CCACC consensus, but only ten mRNAs conform perfectly. INTRODUCTION Two years ago I prepared a compilation of the then-available 5-noncoding sequences of eukaryotic mRNAs (Curr. Topics Microbiol. Immunol. 93, 81-123, 1981). Apart from a bevy of globin and histone mRNAs, only 32 other cellular mRNA se- quences were known at that time. In contrast, there are 166 cellular mRNAs in the present compilation, not counting the globins and histones. I have excluded the mRNAs of lower eukaryotes and viruses, only to keep the survey manageable. One of my objectives was to determine whether certain patterns noted in the earlier compilation would be evident with this larger, more diversified set of sequences