An NMR study of d(CTACTGCTTTAG).d(CTAAAGCAGTAG) showing hydration water molecules in the minor groove of a TpA step

Hexamer oligonucleotide topology and assembly under solution phase NMR and theoretical modeling scrutiny

The entire family of noncomplementary hexamer oligodeoxyribonucleotides d(GCXYGC) (X and Y = A, G, C, or T) were assessed for topological indicators and equilibrium thermodynamics using a priori molecular modeling and solution phase NMR spectroscopy. Feasible modeled hairpin structures formed a basis from which solution structure and equilibria for each oligonucleotide were considered. ¹H and ³¹P variable temperature-dependent (VT) and concentration-dependent NMR data, NMR signal assignments, and diffusion parameters led to d(GCGAGC) and d(GCGGGC) being understood as exceptions within the family in terms of self-association and topological character. A mean diffusion coefficient D(298 K) = (2.0 ± 0.07) × 10⁻¹⁰ m² s⁻¹ was evaluated across all hexamers except for d(GCGAGC) (D(298 K) = 1.7 × 10⁻¹⁰ m² s⁻¹) and d(GCGGGC) (D(298 K) = 1.2 × 10⁻¹⁰ m² s⁻¹). Melting under VT analysis (T(m) = 323 K) combined with supporting NMR evidence confirmed d(GCGAGC) as the shortest tandem sheared GA mismatched duplex. Diffusion measurements were used to conclude that d(GCGGGC) preferentially exists as the shortest stable quadruplex structure. Thermodynamic analysis of all data led to the assertion that, with the exception of XY = GA and GG, the remaining noncomplementary oligonucleotides adopt equilibria between monomer and duplex, contributed largely by monomer random-coil forms. Contrastingly, d(GCGAGC) showed preference for tandem sheared GA mismatch duplex formation with an association constant K = 3.9 × 10⁵M⁻¹. No direct evidence was acquired for hairpin formation in any instance although its potential existence is considered possible for d(GCGAGC) on the basis of molecular modeling studies.

NMR evidence for base dynamics at all TpA steps in DNA

NMR evidence is presented indicating that the exceptional conformational dynamics found at TpA steps in DNA is general to all immediate sequence contexts. One easily tractable NMR parameter that is sensitive to TpA base dynamics is the resonance linewidth of the TpA adenine H2 proton. This resonance experiences a temperature-dependent broadening due to conformational dynamics. Unusual dynamics at TpA steps were originally observed in the sequence context (T)pTpTpApAp(A). We have since shown that the evidence for TpA dynamics persists when either the thymine preceding the TpA step or the adenine following the TpA step is preserved [McAteer et al., Nucleic Acids Res. 23, 3962-3966 (1995)]. Here, in order establish whether or not exceptional TpA dynamics occurs in all DNA sequence contexts, we investigated a series of DNA sequences of the form GCNaTANbNbTANaGC, where N=A,T,C,G. In this family of sequences, all 16 possible immediate sequence context environments of the form NaTANb were examined using 10 DNA sequences. Our NMR results show that the TpA adenine H2 resonance contains a temperature dependent excess linewidth indicative of dynamics in all 16 sequence context environments. By studying a complete set of sequence contexts, it was possible to recognize trends relating resonance parameters and sequence environment. For example, the magnitude of the maximum linewidth is largely determined by the identity of the nucleotide following the TpA step and the magnitude of the linewidth maximum is moderately correlated (r=0.56) with the temperature of the linewidth maximum. The physical basis for these correlations is discussed.

A DNA hairpin with a single residue loop closed by a strongly distorted Watson-Crick G x C base-pair

Our previous NMR and modeling studies have shown that the single-stranded 19mer oligonucleotides d(AGCTTATC-ATC-GATAA GCT) -ATC- and d(AGCTTATC-GAT-GATAAGCT) -GAT- encompassing the strongest topoisomerase II cleavage site in pBR322 DNA could form stable hairpin structures. A new sheared base-pair, the pyrimidine-purine C x A, was found to close the single base -ATC- loop, while -GAT- displayed a flexible loop of three/five residues with no stabilizing interactions. Now we report a structural study on -GAC-, an analog of -GAT-, derived through the substitution of the loop residue T by C. The results obtained from NMR, non-denaturing PAGE, UV-melting, circular dichroism experiments and restrained molecular dynamics indicate that -GAC- adopts a hairpin structure folded through a single residue loop. In the -GAC- hairpin the direction of the G9 sugar is reversed relative to the C8 sugar, thus pushing the backbone of the loop into the major groove. The G9 x C11 base-pair closing the loop is thus neither a sheared base-pair nor a regular Watson-Crick one. Although G9 and C11 are paired through hydrogen bonds of Watson-Crick type, the base-pair is not planar but rather adopts a wedge-shaped geometry with the two bases stacked on top of each other in the minor groove. The distortion decreases the sugar C1'-C1' distance between the paired G9 and C11, to 8 A versus 11 A in the standard B-DNA. The A10 residue at the center of the loop interacts with the G9 x C11 base-pair, and seems to contribute to the extra thermal stability displayed by -GAC- compared to -GAT-. Test calculations allowed us to identify the experimental NOEs critical for inducing the distorted G.C Watson-Crick base-pair. The preference of -GAC- for a hairpin structure rather than a duplex is confirmed by the diffusion constant values obtained from pulse-field gradient NMR experiments. All together, the results illustrate the high degree of plasticity of single-stranded DNAs which can accommodate a variety of turn-loops to fold up on themselves.