Refinement of the solution structure of the DNA dodecamer 5'd(CGCGPATTCGCG)2 containing a stable purine-thymine base pair: combined use of nuclear magnetic resonance and restrained molecular dynamics.

The solution structure of the self-complementary dodecamer 5'd(CGCGPATTCGCG)2, containing a purine-thymine base pair within the hexameric canonical recognition site GAATTC for the restriction endonuclease EcoRI, is investigated by nuclear magnetic resonance spectroscopy and restrained molecular dynamics. Nonexchangeable and exchangeable protons are assigned in a sequential manner. A set of 228 approximate interproton distance restraints are derived from two-dimensional nuclear Overhauser enhancement spectra recorded at short mixing times. These distances are used as the basis for refinement using restrained molecular dynamics in which the interproton distance restraints are incorporated into the total energy function of the system in the form of effective potentials. Eight calculations are carried out, four starting from classical A-DNA and four from classical B-DNA. In all cases convergence to very similar B-type structures is achieved with an average atomic root mean square (rms) difference between the eight converged structures of 0.7 +/- 0.2 A, compared to a value of 6.5 A for that between the two starting structures. It is shown that the introduction of the purine-thymine mismatch does not result in any significant distortion of the structure. The variations in the helical parameters display a clear sequence dependence. The variation in helix twist and propeller twist follows Calladine's rules and can be attributed to the relief of interstrand purine-purine clash at adjacent base pairs. Overall the structure is straight. Closer examination, however, reveals that the central 5 base pair steps describe a smooth bend directed toward the major groove with a radius of curvature of approximately 38 A, which is compensated by two smaller kinks in the direction of the minor groove at base pair steps 3 and 9. These features can be explained in terms of the observed variation in roll and slide.