If I understood your question correctly you are describing a process defined as "indirect readout" (or less frequently shape readout. Here is a good starting review on the topic: Rohs et al. 2010. Origins of specificity in protein-DNA ecognition. Annu Rev Biochem.).

Indirect readout is the recognition of a DNA sequence through the conformational properties of the binding site (any conformation different from the B-DNA

may be recognized by its peculiar local 3D structure or by its structural properties (i.e flexibility).

Sequences of the dsDNA not involved in specific interactions (i.e. bonds with the DNA bases) or non-specific interactions (i.e bonds with the phosphate ackbone of the dsDNA) may in fact play a crucial role in DNA recognition by a given protein.

A very good example in literature is that of E2 protein from Human Papilloma Virus (HPV) (Zang et al. 2004. Predicting indirect readout effects in protein-DNA interactions)

E2 is a homodimer where each monomer inserts an alpha-helix in the major groove of an "ACCG" sequence (direct readout), holding the DNA in an arc and leaving the spacer AATT uncontacted by the E2 protein. This spacer is responsable for the indirect readout effect since it is not bound by E2 but its flexibility (AT-rich) allows for the DNA to be bent "around" E2.(AT-rich) allows for the DNA to be bent "around" E2.Since E2 induces a strong bend in the binding site upon binding to it, the higher the bendability of this region the higher the affinity and the strenght of the interaction.

Indirect readout is so important that some authors suggest also that the binding site for a site-specific DNA-binding protein should always be given together with the flanking regions when these are responsable for indirect readout effects (Linda Jen-Jacobson. 1997 Protein—DNA recognition complexes: Conservation of structure and binding energy in the transition state. Biopolymers)("...Our work on EcoRI endonuclease indicates that the three flanking base pairs on either side of the recognition site may alter the specific binding constant by as much as 500-fold" - see also Fig. 9 of this paper).

In your case the effect of the indirect readout would be further related to the binding of the cofactor.

Expecially in Eukariotes transcription factors bind in combinations (some do not even contact the DNA - these are defined protein cofactors - but partecipate in the final assembly) (Moretti and Ansari. 2008. Expanding the specificity of DNA targeting by harnessing cooperative assembly. Biochimie).

Finally remember the process of the "induced fit" where the conformation of a given DNA-binding protein bound to the cognate binding site could be different

from that in the unbound state. Indirect readout effects allowing the transcription factor to bind by induced fit may allow stronger binding of the cofactor through allosteric mechanism (the conformation of the transcription factor is not rigid and upon binding to the DNA it may slighly change to be bound better by the cofactor eventually)(Jeffrey et al. 1998. Allosteric effects of DNA on transcriptional regulators

Nature)(..."Hence DNA acts as an allosteric ligand whose binding alters the regulator's affinity for other ligands, such as coactivators or corepressors. If the regulator assumes distinct conformations on binding to different genomic sites, then the DNA sequence of the response element can specify which surfaces are available to contact target factors, and thereby determine how the protein regulates transcription at that site (Fig. 1a, b). This mechanism is versatile and adaptable; for example, transcriptional regulators could be 'tethered' through protein–protein contacts with DNA-bound factors, without themselves contacting DNA. In such cases, the bound protein would serve as the allosteric effector).

I hope this may help.

Massimiliano.

It was really helpful Massiniliano! Thanks alot.

Hi Nima thanks for your comments (if you think you are welcome vote up the answer but it is not fundamental actually. I am happy if I can help and when I find the answer to my doubts).

I have struggled a lot in the past to understand this process (in our case the flanking regions were dictating the affinity of binding and in a similar way to E2 the only requirement for the spacer was to be AT-rich - the paper PubMed ID is PMID: 22492712 eventually).

Max.

Hi Nima,

I made a mistake in the PMID actually (the correct one is this: PMID:19376860).

Anyway there are many other papers on the subject which are very good like those cited in the paper from Zang et al. 2004.

Thanks again and regards,

Max.