Non-specific vs. specific DNA binding free energetics of a transcription factor domain protein between search and recognition.

Abstract

Transcription factor (TF) proteins are key in genetic regulation by locating specific sites on the genome. It has been proposed that an optimized search can be achieved by protein alternating between 3D diffusion in the bulk and 1D diffusion along DNA in the facilitated diffusion model. While undergoing 1D diffusion, the protein can switch from a “search” mode for fast diffusion along non-specific DNA, to a “recognition” mode for protein stable binding to specific DNA. Though it was regarded that protein conformation transitions enable the search to recognition transitions, it was recently noticed that for a small TF domain protein, re-orientation of the protein on the DNA, without protein conformational changes, happens between the non-specific and specific DNA binding. We further conducted all-atom molecular dynamics simulations on the TF domain protein bound to a nonspecific or specific DNA site, to energetically confirm that the protein-DNA binding free energetics between the two systems is consistent with the “search” and “recognition”. Using both the Jarzynski equality and umbrella sampling methods pulling protein away from DNA, we obtained protein binding free energetics and a free energy difference of about 10 kBT between the two systems. However, the binding free energy difference estimated from experimental measurements was about 4 kBT on 15-bp DNA. We suggest that the discrepancy between the experimental and computational measurements can be due to DNA sequences flanking the 6-bp central specific and nonspecific binding sites. Such a proposal was also investigated through a highly simplified spherical protein model along with stochastic simulations, which well supported that flanking DNA sequences impact on overall protein dissociation kinetics and therefore on measuring binding affinity variations with specific or non-specific DNA in the central binding site.