A computer algorithm for testing potential prokaryotic terminators

Abstract

The nucleotide sequences of 30 factor-independent terminators of transcription with RNA polymerase from E. coli have been compiled and analyzed. The standard features - a stretch of thymine residues and a preceding dyad symmetry - are shared by most sequences, but there are striking exceptions which indicate that these features alone are not sufficient to describe these sites. In two thirds of the sequences the 3'-half of the dyad symmetry contains the pentanucleotide CGGG (G/C) or a close derivative; about one third have TCTG or a close derivative just downstream of the termination point. The TCTG -box might be implied in termination of stringently controlled operons of E. coli. An algorithm to locate terminators in templates of known nucleotide sequence has been constructed on the basis of correlation to the distribution of dinucleotides along the aligned signal sequences. The algorithm has been tested on natural sequences of a total length of about 11,500 N. It finds all known independent terminators and only a few other sites, including some of the rho-dependent and putative terminators.