On the consequences of placing amino groups at the TBP-DNA interface. Does TATA really matter?

The TATA-box binding protein (TBP) belongs to a family of structural proteins involved in transcription in eukaryotic cells. TBP binds in the minor groove of DNA and recognizes specifically the consensus sequence: 5' TATAWAWR 3' (W = A or T). Recent reports show that the TATA-box is only present in 10% of all human polymerase II promoters. Therefore, TBP must bind frequently to low affinity DNA sequences, possibly with help of other transcription factors. In order to understand the intramolecular and intermolecular interactions that lead to the consensus sequence preferred by TBP, we use high resolution crystallographic structures of cognate TBP-DNA complexes as templates onto which 16 dinucleotide repeating sequence DNA oligomers were built. The binding free energy of each complex was calculated using the Molecular Mechanics/Poisson-Boltzmann Solvent Accessible (MM-PBSA) approximation. Parsing of the free energy components allowed us to identify the most important contributions to sequence selectivity: DNA deformation and the interaction energy between TBP residues and DNA bases, as expected. Surprisingly, poor interaction energies lead to larger deformation costs, suggesting strategies to improve affinity and selectivity. Local analysis of the TBP-DNA interface allowed us to build interaction and deformation energy tables that were used, without the need to fit their relative weights, to predict successfully both the consensus sequence for TBP, and relative binding affinities for a collection of TATA box variants.