purine–purine pyrimidine–pyrimidine purine–pyrimidine. The propensity for DNA deformations to
occur at pyrimidine–purine steps is seen in a number of protein–DNA complexes, including those of the
IHF (Figure 10.6b) and TBP (Figure 10.5). The pyrimidine–purine step may adopt large roll or high twist
angles in protein–DNA complexes. Severe distortion of the DNA occurs in restriction enzyme/DNA com-
plexes, and here indirect readout may account for nearly half of the energy associated with cleavage speci-
ficity, corresponding to 25–30 kJ mol^1.^23 The detailed energy components of base stacking in the context
of protein–DNA complexes depend on the dielectric constant of the environment and the need to bury
protein–DNA interfaces away from the environment of the bulk solvent. Currently, it is hard to quantify
the energy associated with deformation of DNA or RNA within a protein–nucleic acid complex because
in such a complex the environment of the nucleic acid is different from bulk solvent. For instance, the
nucleic acid experiences a large change in dielectric constant as its protein partner surrounds it. This has a
large effect on the energies of any conformational adjustments in the nucleic acid.
10.4.6 Co-operativity through Protein–Protein and DNA–Protein Interactions
Many of the proteins, already described, bind to DNA elements as homo- or heterooligomers, and thus make
homotypicor heterotypicprotein–protein interactions, respectively. In such complexes, each subunit may
present a reading head that interacts with the DNA target. If properly arranged, the organization of binding
sites on the DNA can provide co-operativebenefits to binding, by simply helping to pre-organize the suc-
cessive binding of the other subunits. Even if the proteins pre-exist in solution, there is still an implicit co-
operative benefit because the matching of one subunit to its site pre-organizes the second site. This
decreases the entropic cost of binding, and is known as the chelate effect.^24
A typical control region for a eukaryotic gene is very complex and contains a myriad of binding sites
for different proteins. In many cases, such proteins form homotypic or heterotypic protein–protein inter-
actions, so that they mutually influence their DNA-binding affinity. Such co-operative effects might com-
pensate for the comparatively weaker binding and lower sequence discrimination of eukaryotic transcription
factors compared with prokaryotic repressors and activators. A complex organization that involves multiple
interacting proteins may also have kinetic benefits for gene regulation. For example, multi-component
assembly is involved in the pre-initiation transcription complexes of eukaryotes and archaea. Pre-initiation
complexes assist RNA polymerases to bind promoter sites, and are often composed of the TATA-element
binding protein (TBP) and the transcription factor TFIIB (Figure 10.11). Other components may be assem-
bled onto the exposed surface of the TBP, depending on the promoter. Because the complex is asymmet-
ric, it can specify the direction of transcription. How does this asymmetry arise?
The TBP protein has approximate twofold symmetry, so there are in principle two possible orientations
for TBP to bind with respect to the start site of transcription. TBP has a saddle-like shape that engages a
widened DNA minor groove, but clearly, TBP binding alone cannot specify the direction of transcription.
Transcription factor TFIIB binds asymmetrically to the TBP by recognition of one exposed surface and
makes a contact in the major groove on the 5side of the TATA element (at the ‘BRE’ recognition site)
through a helix-turn-helix motif (Figure 10.11). Thus, it is the combination of protein–protein interactions
between the TFIIB and the TBP and the protein–DNA interactions of the TFIIB that breaks the symmetry
of the complex and defines the direction of transcription.
Another example of protein–protein interactions is provided by the hormone and nuclear receptors
(Figure 10.2c). The DNA elements of steroid receptors usually comprise two hexameric elements that are
arranged as an inverted repeat with a three-nucleotide separation of half-sites (e.g.the idealized glucocor-
ticoid response element is 5-AGAACAxxxTGTTCT-3). The pseudo-dyad symmetry of the half-sites
aligns the proteins, which interact in a ‘head-to-head’orientation. The DNA-binding domains of these proteins
bind through favourable protein–protein interactions, which give rise to co-operativity. The two monomers
interact with the DNA through a dimerization interface, the proper alignment of which requires exact
spacing between the two monomer-binding sites. The two monomers would clash sterically if they were to
bind to a DNA sequence in which the half-site separation is either two or four bases.
402 Chapter 10