Solution structure of the EcoRI DNA sequence: refinement of NMR-derived distance geometry structures by NOESY spectrum back-calculations.

Abstract

The solution structure of the self-complementary DNA duplex [d(CGCGAATTCGCG)]2, which contains the EcoRI restriction site sequence GAATTC at the center, has been studied by two-dimensional nuclear magnetic resonance spectroscopy. Time-dependent nuclear Overhauser effect spectra were used to obtain the initial cross-relaxation rates between 155 pairs of protons. These initial cross-relaxation rates were converted into interproton distances and entered into a distance (bounds) matrix. A distance geometry algorithm (DSPACE) was used to create embedded starting structures and to refine these structures until they showed good agreement with the distance matrix; symmetry constraints were included in the refinement procedure, making the two strands in the refined distance geometry structures virtually identical and significantly improving the agreement with the distance matrix. The NOESY spectrum for one of these distance geometry structures was then calculated from the explicit coordinates by numerically integrating all the z-magnetization transfer pathways among neighboring protons within a specified radius. Distances in this distance geometry structure that did not agree with the experimental NOESY time course were then adjusted accordingly. This process was iterated until a good agreement between calculated and experimental NOESY spectra was reached. The final structure, which generates good agreement with the experimental NOESY spectrum, displays kinks at the C3-G4 base step and at the A6-T7 base step that appear to be similar to those reported for the EcoRI restriction site DNA bound to its endonuclease. The solution structure is not the same as the crystal structure of this DNA duplex.