We propose a dynamical model for nonspecific DNA-protein interaction, which is based on the "bead-spring" model previously developed by other groups, and investigate its properties using Brownian dynamics simulations. We show that the model successfully reproduces some of the observed properties of real systems and predictions of kinetic models. For example, sampling of the DNA sequence by the protein proceeds via a succession of three-dimensional motion in the solvent, one-dimensional sliding along the sequence, short hops between neighboring sites, and intersegmental transfers. Moreover, facilitated diffusion takes place in a certain range of values of the protein effective charge, that is, the combination of one-dimensional sliding and three-dimensional motion leads to faster DNA sampling than pure three-dimensional motion. At last, the number of base pairs visited during a sliding event is comparable to the values deduced from single-molecule experiments. We also point out and discuss some discrepancies between the predictions of this model and some recent experimental results as well as some hypotheses and predictions of kinetic models.

:The mechanical model based on beads and springs, which we recently proposed to study non-specific DNA-protein interactions [J. Chem. Phys. 130, 015103 (2009)], was improved by describing proteins as sets of interconnected beads instead of single beads. In this paper, we first compare the results obtained with the updated model with those of the original one and then use it to investigate several aspects of the dynamics of DNA sampling, which could not be accounted for by the original model. These aspects include the effect on the speed of DNA sampling of the regularity and/or randomness of the protein charge distribution, the charge and location of the search site, and the shape and deformability of the protein. We also discuss the efficiency of facilitated diffusion, that is, the extent to which the combination of 1D sliding along the DNA and 3D diffusion in the cell can lead to faster sampling than pure 3D diffusion of the protein. (#)