Abstract

Gene expression and regulation rely on an apparently finely tuned set of reactions between some proteins and DNA. Such DNA-binding proteins have to find specific sequences on very long DNA molecules and they mostly do so in the absence of any active process. It has been rapidly recognized that, to achieve this task, these proteins should be efficient at both searching (i.e., sampling fast relevant parts of DNA) and finding (i.e., recognizing the specific site). A two-mode search and variants of it have been suggested since the 1970s to explain either a fast search or an efficient recognition. Combining these two properties at a phenomenological level is, however, more difficult as they appear to have antagonist roles. To overcome this difficulty, one may simply need to drop the dichotomic view inherent to the two-mode search and look more thoroughly at the set of interactions between DNA-binding proteins and a given DNA segment either specific or nonspecific. This chapter demonstrates that, in doing so in a very generic way, one may indeed find a potential reconciliation between a fast search and an efficient recognition. Although a lot remains to be done, this could be the time for a change of paradigm.