|
Many small DNA-binding regulatory proteins find their specific binding sites on DNA with the rates that are much faster than allowed by simple 3D diffusion in solution around DNA coil. It was suggested long ago that the reduction of diffusion space from 3D in bulk to 1D upon protein binding to DNA track is capable of accounting for these enhanced protein-DNA association rates. There are however some problems with theoretical models suggested in the literature that forced us to revisit this interesting phenomenon.
We have constructed a non-equilibrium model of protein binding in DNA coil and considered a facilitated location of targets on DNA. The model accounts for non-specific DNA-protein binding energy as well as for different values of 1D/3D protein diffusion coefficients. The search process is a combination of 3D and 1D diffusion steps. Using the standard procedure for the calculation of the mean first passage time along arbitrary energy landscape, we predicted that the final target search time can be up to 2 orders of magnitude faster than in Smoluchovskii 3D limit. Also, target search time was shown to be a non-monotonous function of both protein-DNA attraction strength and concentration of proteins in the bulk [1]. Some drawbacks of previous models of facilitated protein diffusion are improved due to the correct treatment of "correlation" diffusional terms. Then, we suggested a two-step mechanism for recognition of DNA-sliding proteins and DNA double helix. The first step is based on complementarity of DNA and protein charge interaction lattices. It occurs prior to DNA-protein recognition via hydrogen bond formation. The electrostatic recognition energy well obtained within the model near the charge-homologous DNA sequence is typically 3-10 kT in depth and it is capable to considerably slow down locally the protein diffusion. During the time of protein localization near the bottom of this well, the protein binding domains can rearrange themselves allowing easier/faster formation of hydrogen bonds with DNA bases. We have tested this assumption of charge complementarity performing detailed computational analysis of charge positioning along interfaces of many DNA-protein complexes from the Protein Data Bank [2]. We have observed that in particularly for large structural proteins, such as native nucleosome core particles, this charge sequence-specific zipper effect is very pronounced. Thus, the distribution of positive amino acids Lys and Arg on the protein surface underneath the bound DNA fragment is (evolutionary?) adjusted to provide a better fit for sequence specific charge pattern of DNA negative phosphates. This indicates the importance of DNA-protein sequence-specific electrostatic interactions for DNA-protein recognition. This fact was largely overlooked in the literature. The charge complementarity revealed here for DNA-protein complexes is reminiscent to that between aligned parallel DNA fragments in dense DNA assemblies as well as upon formation of dense protein-protein complexes. [1] A. G. Cherstvy, A. B. Kolomeisky, and A. A. Kornyshev, ´Protein-DNA interactions: reaching and recognizing the targetsĄ, J. Phys. Chem. B, 112 4741-4750 (2008). [2] A. G. Cherstvy, ´Positively charged residues in DNA-binding domains of structural proteins follow sequence-specific positions of DNA phosphate groupsĄ, J. Phys. Chem. B, 113 4242-4247 (2009). |
|