Importance of Accurate DNA Structures in Solution: The Jun–Fos Model

Understanding the recognition of DNA sequences by proteins requires an accurate description of the structural dynamics of free DNA, especially regarding indirect readout. This involves subtle sequence-dependent effects that are difficult to characterize in solution. To progress in this area, we applied NMR and extensive simulations to a DNA sequence relevant to the Jun–Fos system. The backbone and base behaviors demonstrate that unrestrained simulations with major force fields (Parm98, Parmbsc0, and CHARMM27) are not reliable enough for in silico predictions of detailed DNA structures. More realistic structures required molecular dynamics simulations supplemented by NMR restraints. A new methodological element involved restraints inferred from the phosphate chemical shifts and from the phosphate dynamics. This provided a detailed and dynamic view of the intrinsic properties of the free DNA sequence that can be related to its recognition, by comparison with a relevant DNA–protein complex. We show how to exploit the relationship between phosphate motions and helicoidal descriptors for structure determination toward an accurate description of DNA structures and dynamics in solution.