Sequence-dependent DNA conformation and flexibility play a fundamental role in specificity of DNAprotein interactions. Here we quantify the DNA crookedness: a sequence-dependent deformation of DNA that consists on periodic bends of the base pair centers chain. Using molecular dynamics simulations, we found that DNA crookedness and its associated flexibility are bijective: unveiling a one-to-one relation between DNA structure and dynamics. This allowed us to build a predictive model to compute DNA stretching stiffness from solely its structure. Sequences with very little crookedness show extremely high stiffness and have been previously shown to form unstable nucleosomes and promote gene expression. Interestingly, the crookedness can be tailored by epigenetic modifications, known to affect gene expression. Our results rationalize the idea that the DNA sequence is not only a chemical code, but also a physical one that allows to finely regulate its mechanical properties and, possibly, its 3D arrangement inside the cell.
MAIN TEXTThe mechanism by which proteins interact with the genome with such extraordinary specificity is still an open question in biology. Since the discovery of the DNA double helix (dsDNA), it became clear the existence of a sequence dependent set of hydrogen bond donors and acceptors that are exposed in the major groove and specifically recognized by certain amino acids. However, there is increasing evidence that this mechanism is far from sufficient. In a number of DNA-protein complexes, DNA adopts a conformation that substantially deviates from the canonical B-form (1-3), suggesting a structural deformation or an exceptional flexibility intrinsic to the DNA sequence. Among the most-studied cases are sequencedependent DNA deformations -A-like structures, kinked base pair steps and A-tracts-that play an important role in transcription regulation (4-7). In parallel, the high sequence-dependent flexibility of DNA is used by several proteins to achieve binding specificity (1,8).However, many aspects of DNA flexibility have so far remained elusive. Indeed, It is not fully understood how a relatively stiff molecule, with a persistence length of P∼50 nm, is able to wrap around a histone octamer of ∼4 nm of radius. Even more intriguing is the fact that some sequences are hardly able to form stable nucleosomes, arguably as a consequence of a distinct conformation or mechanical properties (9, 10). The same question holds for other DNA-protein complexes, in particular, for many repressor systems where a highly bent loop is predicted in the DNA (11). These considerations, supported by recent findings on high bendability of DNA at short-length scales (12, 13) challenge the currently accepted wormlike chain model (WLC) and demand for an accurate description of sequence-dependent DNA flexibility at these scales.Using constant-force molecular dynamics (MD) simulations (14) we observed that the extension of the DNA changed from one sequence to another for molecules with the same number of base pairs. we...