The fragile X syndrome single strand d(CGG)n nucleotide repeats readily fold back to form unimolecular hairpin...

Expansion of a d(CGG)n run within the 5'-untranslated region of the X-linked human gene FMR1 blocks FMR1 transcription, delays its replication, and precipitates the fragile X syndrome. We showed previously that d(CGG)n tracts aggregate into interstrand tetrahelical complexes (Fry, M., and Loeb, L. A. (1994) Proc. Natl. Acad. Sci. U. S. A. 91, 4950-4954). Here we show that these sequences also form under physiological conditions in in vitro unimolecular hairpin structures. Folding is demonstrated by temperature-dependent mobility of d(CGG)n oligomers in a nondenaturing polyacrylamide gel, by UV-hyperchromicity of thermally denaturing oligomers, and by UV cross-linking of compact forms of d(CGG)n runs interspersed by thymidine clusters. That the compact d(CGG)n structures are unimolecular is suggested by their zero-order kinetics of formation. Diethyl pyrocarbonate modification reveals a single, 4-5 residue-long central or epicentral unpaired loop in folded d(CGG)n oligomers. The position of this loop remains unchanged by insertion of thymidine clusters into 15- or 33-mer d(CGG) tracts as indicated by KMnO4 probing of unpaired thymidines. The presence of a single loop in folded d(CGG)n oligomers and the accessibility of every guanine to dimethyl sulfate modification suggest that they are hairpin and not tetraplex structures. Modeling indicates that different d(CGG)n hairpins are stabilized by guanine-guanine Hoogsteen hydrogen bonds or by Hoogsteen and Watson-Crick bonds. If formed in vivo, d(CGG)n hairpins could entail slippage and trinucleotide expansion during replication and could obstruct FMR1 gene transcription and replication.