Genetic mutation studies have revealed that abnormal expansion or contraction of certain triplet repeat (TR) sequences, such as CCG/CGG and CTG/CAG, is implicated in more than 10 neurodegenerative diseases.* 1 TR sequences are often found within or near relevant gene-coding regions16 and thus have promise as diagnostic probes and perhaps as targets for new genetic therapies for related hereditary diseases. To understand the possible mechanisms of dynamic mutations of TR sequences, we have initiated an investigation into the solution conformation of the six CXG and GXC triplets (X = A, C, G, T) using NMR methods. The presence of a stable structure associated with these TR sequences might play a role in disrupting the normal replication or transcription process.2Here we report a novel duplex motif formed by DNA CCG TR sequences in a neutral to basic pH range. The structure is characterized by NMR spectroscopy, electrophoresis, and NMRbased calculations. A series of CCG TR sequences (all sequences refer to DNA, unless otherwise noted), which include the homologous sequences CGCCG and (CCG)" (n = 2, 3, 4, and 5), the permutation isomers (CGC)2 and (GCC)2, and the linker sequence (CCG)2-xx-(CCG)2 (x = triethylene glycol linker), have been examined to supplement the results pertaining to the (CCG)2 sequence.NMR spectra of (CCG)2 in aqueous solutions containing 0.1 -0.2 M NaCl at pH 6.3-8.5 provide strong evidence for a distinct, stable Watson-Crick base paired structure. The complete !H and 31P assignments and a summary of the major spectral features of (CCG)2 are given in Tables SI and S2, respectively, in the supporting information. (CCGh forms a symmetrical duplex, which displays 'H and 31P resonances corresponding to a single strand of the hexamer (Figure SI, Table SI, supporting information).3 Twenty percent polyacrylamide nondenaturing gel electrophoresis shows that (CCG)2 and (CCG)3 duplexes migrate faster than the corresponding complementary hexamer and nanomer duplexes at 4 °C,4 excluding the possible presence of higher order structures. In the spectra recorded in 90% H2O (Figure S1A) G3 and G106 HN (imino) resonances at 13.04 and 13.48 ppm show an HN-HN NOE and HN-H2N (amino) NOEs (Figure SI A). These observations, especially the presence of the HN-HN NOE between G3 and G106 residues, are unequivocal evidence for the staggered alignment of (CCG)2 (Scheme 1).The 31P ID spectrum of (CCG)2 (Figure SIB) exhibits four well-resolved signals in a chemical shift range of 2.1 ppm [3'-P (ppm) of C2 -2.84; C5 -3.45; Cl -4.04; C4 -4.09; G3 -4.97], The 31P resonances of canonical DNA duplexes are usually detected in a ~0.6 ppm region.5 The large dispersion of31P resonances of (CCG)2 is diagnostic of an untwisted and extended duplex conformation as observed in several antibiotic-DNA complexes.6 Further evidence for a distorted (CCG)2 * To whom correspondence should be addressed.
