Sequence-dependent DNA structure. The role of base stacking interactions.

The sequence-dependent structure of DNA is analysed on the basis of the energetics of the base stacking (pi-pi) interactions. The conformational preferences of the ten possible base-pair steps in double-helical DNA have been calculated and compared with experimental data from X-ray fibre diffraction, X-ray crystal structures and gel-running experiments. The calculations account for many features of sequence-dependent DNA structure, including polymorphism in DNA, the lack of polymorphism in RNA, the structure of Z-DNA, bistability in pyrimidine-purine (YR) steps, the origin of propeller twist and buckle and the role of TATA sequences at the sites of origin of replication. The computational model used specifically allows for the charge distribution associated with the out-of-plane pi-electron density of the bases. The results obtained are rationalized on the basis of the shapes and charge distributions of the bases. Calladine's cross-strand steric clashes at pyrimidine-purine (YR) and CX/XG steps are reproduced. In AX/XT steps, same-strand steric clashes occur between the thymine methyl group and the 5'-neighbouring sugar. They are the cause of the large negative propeller twist observed in A.T base-pairs. Steric clashes between the thymine methyl group and the 5'-neighbouring base block A-DNA conformations in AX/XT steps. Electrostatic interactions between partial atomic charges are most important for C.G base-pairs which are highly polarized. They lead to strong preferences for positive slide and negative slide conformations in CG and GC steps, respectively. This combination can be accommodated in poly(CG) by left-handed Z-DNA. Many of the more subtle sequence-dependent effects are caused by electrostatic interactions between the partial atomic charges on one base-pair and the out-of-plane pi-electron density on another. The effect is most important in CX-XG steps and leads to bistability. In general, electrostatic interactions cause non-zero slide. This is opposed by the hydrophobic effect which favours the zero slide B-type conformation. Thus, B-DNA is observed at high water content in fibres and electrostatic interactions force high or low slide A or C-DNA conformations at low water content. If two juxtaposed steps have very different conformational preferences, this incompatibility can lead to unusual structures such as Z-DNA or strain such as in TATA sequences.