DNA-binding requirements of the yeast protein Rap1p as selected in silico from ribosomal protein gene promoter sequences

Motivation: High-level transcriptional activation of most ribosomal protein (rp) genes in Saccharomyces cerevisiae is promoted by the global DNA-binding factor Rap1p. The creation of the complete database of yeast rp gene promoter sequences enabled us to develop a computational selection strategy aimed at acquiring detailed information concerning the DNA-binding specificity of Rap1p. Results: Rap1p sites in rp gene promoters are often found in duplicate, exhibiting strong preferences in both spacing and orientation. Using these preferences, a weight matrix was selected that represents the in vivo binding requirements of Rap1p. The resulting matrix renders the identification of functional Rap1p binding sites more accurate and allowed us to re-evaluate previous in vitro data. Tandemly arranged Rap1p binding sites appear to be typical for rp gene promoters and differ in preferred spacing from sites occurring in (sub)telomeric repeats. The preferred spacing that is found in duplicate Rap1p binding sites of rp gene promoters is restricted to a small window between 15 and 26 bp. This is proposed to reflect the borders within which binding co-operativity operates. The data presented clearly illustrate that computational selection strategies provide information that reaches beyond experimental data.