Structure of a C-rich strand fragment of the human centromeric satellite III: A pH-dependent intercalation topology

Repetitive DNA sequences may adopt unusual pairing arrangements. At acid to neutral pH, cytidine-rich DNA oligodeoxynucleotides can form the i-motif structure in which two parallel-stranded duplexes with C Ct pairs are intercalated head-to-tail. The i-motif may be formed by multimeric associations or by intra-molecular folding, depending on the number of cytidine tracts, the nucleotide sequences between them, and the experimental conditions. We have found that a natural fragment of the human centromeric sat ellite III, d(CCATTCCATTCCTTTCC), can form two monomeric i-motif structures that differ in their intercalation topology and that are favored at pH values higher (the eta -form) and lower (the lambda -form) than 4.6. The change in intercalation may be related to adenine protonation in the loops. We studied the uridine derivative methylated on the first cytidine base, d(5mCCATTCCAUTCCUTTCC), whose proton spectrum is better resolved. The intercalation topologies are (C7.C17)/(5mC1.C11)/(C6.C16)/(C2.C12) for form lambda and (5mC1.C11)/(C7.C17)/(C2.C12)/ (C6.C16) for form eta. We have solved the structure of the eta -form, and we present a model for the h-form. The switch from eta to lambda involves disruption of the i-motif. In both forms, the central AUT linker crosses the wide groove, and the first and the third linkers loop across the minor grooves. The i-motif core is extended in the eta -form by the inter-loop reverse Watson-Crick A3.U13 pair, whose dissociation constant is around 10(-2) at 0 degreesC, and in the lambda -form by the interloop T5.T15 pair. In contrast, d(5mCCATTCCTTACCTTTCC) folds into a pH-independent structure that has the same intercalation topology as the h-form. The i-motif core is extended below by the interloop T5.T15 pair and closed on top by the T8.A10 pair. In contrast, d(5mCCATTCCTTACCTTTCC) folds into a pH-independent structure that has the same intercalation topology as the h-form. The i-motif core is extended below by the interloop T5.T15 pair and closed on top by the T8.A10 pair. Thus, the C-rich strand of the human satellite III tandem repeats, like the G-rich strand, can fold into various compact structures. The relevance of these features to centromeric function remains unknown. (C) 2001 Academic Press.