Triple-Helix Formation by Pyrimidine Oligonucleotides Containing Nonnatural Nucleosides with Extended Aromatic Nucleobases: Intercalation from the major groove as a method for recognizing C·G and T·A base pairs

Abstract

The sequence-specific recognition of double-helical DNA by oligonucleotide-directed triple helix formation is limited primarily to purine tracts. To identify potential lead compounds which are able to extend the sequence repertoire of triple helical complexes, we designed two carbocyclic nucleosides with nucleobases attached via amide bonds. N5-[(1R, 2S, 3R, 4R)-3-hydroxy-4-(hydroxymethyl)-2-methoxycyclopentyl]-2-{[(1H-pyrrol-2-yl)carbonyl]-amino}thiazole-5-carboxamide (L1) and 2-benzamido-N^5-[(1R, 2S, 3R, 4R)-3-hydroxy-4-(hydroxymethyl)-2-methoxycyclopentyl]thiazole-5-carboxamide (L2) were synthesized and incorporated into pyrimidine oligonucleotides. The 2-(trimethylsilyl)ethoxymethyl (SEM) protecting group for the 1H-pyrrole NH was found to be compatible with DNA solid-phase synthesis of pyrimidine Oligonucleotides. By quantitative DNase I footprinting analysis, both nonnatural nucleosides L1 and L2 showed preferential binding of pyrimidine over purine bases: L1/2·(C·G) ≈ L1/2·(T · A) > L1/2·(G·C) ≈ L1/2·(A · T). Comparison with the previously reported nonnatural nucleosides with extended aromatic nucleobases 1-(2-deoxy-β-d-ribofuranosyl)-4-(3-benzamidophenyl)-imidazole (D3) and N^4-[6-(benzamido)pyridin-2-yl]-2′-deoxycytidine (^(bz)M) suggests that the observed binding selectivity C · G ≈ T · A > G · C ≈ A · T for the nucleoside analogs L1, L2, D3, and ^(bz)M is derived from sequence-specific intercalation with preferential stacking of their nucleobases over pyrimidine · purine Watson-Crick base pairs.