On the Relative Ability of Centromeric GNA Triplets to Form Hairpins versusSelf-paired Duplexes

Abstract

While tandem repeats of the human centromere DNA pentamer sequence T GRAA form stable “self-complementary” [T GRAAT GRAA] 2duplexes (R=G or A) containing the GA-bracketed unpaired purine stack motif, their phase-shifted variants NAAT GNAAT G(N=A, G, C, T) were found to exist in solution as an equilibrium mixture of a duplex containing the GA-bracketed unpaired stack motif and a hairpin containing a single-residue loop closed by a sheared G·A pair. The stability of the hairpin form relative to duplex form of GNA triplets was found to be GCA>GAA/GTA>>GGA, with the CAAT GCAAT Gsequence mostly in the hairpin form and the GAAT GGAAT Gsequence mostly in the [GAAT GGAATG] 2duplex form. The chemical shifts of the H1′ and H4′ protons of the central N residue in GNA triplets were found to differ markedly in the duplex and hairpin forms and are diagnostic indicators of which conformation the oligonucleotide adopts. Comparison between the structures of the G·A-closed C loop motif and the G·A-bracketed unpaired G-stack [GGA] 2motif reveals remarkably similar stacking by the loop C residue and the intercalated G residue on the adjacent sheared G·A pair. The anomalous upfield chemical shifts of the H1′ and H4′ protons in [GGA] 2motifs and the H4′ proton in GCA loops, and the different sugar conformations in these two motifs, can be explained by interstrand versusintrastrand stacking of the central (G or C) deoxyribose with the adenine base. Based on these studies, a DNA sequence GT GGAAT GGAAT GGAAC was designed and shown to form a duplex containing three [GGA] 2motifs, while its 9G→9C analog GT GGAAT GCAAT GGAAC was found to adopt a stable hairpin containing a (GGA) 2motif in the stem and a G·A-closed single C-loop.