Nearest-neighbor non-additivity versus long-range non-additivity in TATA-box structure and its implications for TBP-binding mechanism

Abstract

TBP recognizes its target sites, TATA boxes, by recognizing their sequence-dependent structure and flexibility. Studying this mode of TATA-box recognition, termed 'indirect readout', is important for elucidating the binding mechanism in this system, as well as for developing methods to locate new binding sites in genomic DNA. We determined the binding stability and TBP-induced TATA-box bending for consensus-like TATA boxes. In addition, we calculated the individual information score of all studied sequences. We show that various non-additive effects exist in TATA boxes, dependent on their structural properties. By several criterions, we divide TATA boxes to two main groups. The first group contains sequences with 3-4 consecutive adenines. Sequences in this group have a rigid context-independent cooperative structure, best described by a nearest-neighbor non-additive model. Sequences in the second group have a flexible, context-dependent conformation, which cannot be described by an additive model or by a nearest-neighbor non-additive model. Classifying TATA boxes by these and other structural rules clarifies the different recognition pathways and binding mechanisms used by TBP upon binding to different TATA boxes. We discuss the structural and evolutionary sources of the difficulties in predicting new binding sites by probabilistic weight-matrix methods for proteins in which indirect readout is dominant.

Figure 1.

Phasing analysis of yTBPc-induced TATA-box bending. Shown are the relative mobilities of the bound DNA divided by the relative mobilities of the free DNA as a function of the linker length. The values shown are of one representative experiment (of 3–4 independent experiments). The line is from the best fit to a cosine function (44).

Figure 2.

Dissociation kinetics of yTBPc (27 nM) from consensus-like TATA-box variants embedded in hairpin constructs (0.4 nM). The number below each gel denotes the time after adding competitor DNA (1.76 μM).

Figure 3.

Plot of the fraction of molecules bound to consensus-like TATA-box variants at time (t) divided by the fraction of molecules bound at time zero is plotted as a function of time. The lines are from the best fit to a double exponential curve. Solid squares, MLP; solid circles, T₇A₈; solid down triangles, A₈; solid up triangles, T₈; solid diamonds, T₇; open squares, T₅; open circles, (TA)₄; open down triangles, T₅A₈, open up triangles, T₅T₈; open diamonds, T₅T₇. The shown experimental points are those from only one experiment, out of 3–6 independent experiments conducted with each DNA target. Hence, they may deviate slightly from the averaged values presented in Table 1.

Figure 4.

Dissociation kinetics experiments using methylated DNA targets. Left: double-stranded stem of DNA hairpin containing the MLP target with methylated cytosine residues (denoted by M). Right: stem of DNA hairpin containing the T₇ target with methylated cytosine residues. For other details see Figure 2.

Figure 5.

Comparison between the dissociation kinetics of yTBPc from methylated and non-methylated TATA boxes. Fraction of yTBPc molecules bound at time (t) divided by the fraction of molecules bound at time (0) is plotted as a function of time. Solid squares, MLP; open squares, methylated MLP; solid circles, T₇; open circles, methylated T₇. For other details see Figure 3.