Compositional nonrandomness upstream of start codons in archaebacteria.

Since the regions directing transcriptional and translational initiation in archaebacteria are poorly characterized, the purpose of this study was to characterize them using measurements of nonrandomness upstream of start codons on eight fully sequenced archaebacterial genomes. Two distinctly different regions with conservation were identified. The location of the first corresponded well to the classical Shine-Dalgarno region (phi peak), and the other was located approximately 20-35 nucleotides upstream of start codons (alpha region), but both regions are not present in all strains. The composition of the region around the phi peak showed an overrepresentation of guanine, whereas composition in the alpha region had an overrepresentation of adenine and thymine. It is furthermore shown that the alpha region surprisingly is associated with start codon usage and other characteristics of the genes. The alpha region is likely to correspond to TATA-boxes and thereby indicates use of leaderless (or short-leadered) transcripts.