Strong sequence patterns in eukaryotic promoter regions: potential implications for DNA structure.

1. Analysis of eukaryotic sequences reveals recurring trends in upstream regions. Oligomers composed of (G/C)n and (A/T)m blocks are preferentially flanked by (G/C)2 doublets on their 3' rather than on their 5' ends, that is (G/C)n(A/T)m(G/C)2 > (G/C)n+2(A/T)m. 2. These trends are stronger for larger n and smaller m. Additional trends are outlined below. 3. The trends are correlated with DNA structural parameters, in particular with twist and roll angles. 4. Generally, the trends hold if the base pair step joining the 5' (G/C)2 doublet to the (G/C)n (A/T)m oligomer is not undertwisted and is not strongly rolled into the major groove. 5. Other DNA parameters crucial for DNA-protein interactions are discussed as well.