Solution structure of a GA mismatch DNA sequence, d(CCATGAATGG)2, determined by 2D NMR and structural refinement methods.

Abstract

GA mismatches in DNA have drawn attention because of their special repair mechanisms, stability, and variety of conformations. A symmetric 10-base oligodeoxyribonucleotide duplex, d(CCATGAATGG)2, containing two GA mismatches has been investigated by one- and two-dimensional multinuclear NMR and molecular refinement procedures to ascertain the conformational details of the 5'-pyrimidine-GA-purine-3' sequence. A molecular model established from the NMR results has a B-type right-handed helix with each of the bases retaining the normal anti-glycosidic torsional angles. Type I mismatched base pairs have GNH2-AN7 and GN3-ANH2 (edge-to-edge) hydrogen bonds, while type II base pairs have GN1H-AN1 and GO6-ANH2 (face-to-face) bonds. The conformation at the GA mismatch site has type I GA base pairs and an unusual cross-strand stacking of the adjacent G5 and A6 bases, which causes significant overwinding of the helix at the mismatch site. Unusual shifts of the 31P resonances suggest that the phosphate linkage between G5 and A6 is no longer in the low-energy BI conformation. One-dimensional imino and phosphorus NMR studies were carried out on a number of DNA sequences containing adjacent 5'-GA-3' mismatched base pairs to investigate the sequence dependence of the conformations and base-pairing types. Type I and type II conformations have very different imino proton and 31P NMR spectral patterns that can be used to classify any sequence with adjacent GA mismatches by base-pairing and conformational type. The NMR results indicate that the conformation selected is dictated completely by the flanking sequence: 5'-pyrimidine-GA-purine-3' sequences adopt the type I conformation, while 5'-purine-GA-pyrimidine-3' sequences have the type II conformation.