Proteins find binding sites by hopping from one site to the next rather than by sliding.


Arecent report defies the dogma for how sequence-specific DNA binding proteins find their final binding sites. The article, by Darren Gowers and Stephen Halford (University of Bristol, Bristol, UK), shows that proteins travel three-dimensional (3D), rather than one-dimensional (1D), routes to their resting sites.

In the old 1D, or sliding, model, proteins bind randomly to nonspecific DNA and then slide along as though on a track to their binding sites. By separating the nonspecific DNA and the binding site on interlinked rings of DNA called catenanes, the Bristol group now demonstrates that proteins can jump to their destinations.

In a catenane, a binding protein cannot interchange from one ring to the other without first dissociating from the DNA. However, restriction enzymes found their binding sites in the midst of extraneous nonspecific DNA just as efficiently whether the two were joined in one molecule or linked in a catenane. Thus, the authors suggest that proteins dissociate and rebind until they reach their specific sites. Searching was more efficient on supercoiled rather than on relaxed DNA, probably because the former compresses the volume of space that proteins must traverse.

Halford feels that the 3D model makes more sense than sliding for proteins that are not energy driven. “The main problem with the 1D model,” he says, “is that the protein has the same chance of moving left or right, so it stays in approximately the same place. But it can cover a bigger volume in three dimensions.” ▪


Gowers, B., et al. 2003. EMBO J. 22:1410–1418.