Book of Abstracts: Albany 2007
June 19-23 2007
Structural Insights into Gene Regulation: The Role of Sequence-Specific DNA Structure in Protein-DNA Recognition
Gene expression is regulated by protein binding to DNA in a sequence-specific manner. The structural principles of protein-DNA recognition have been extensively studied in a protein-centric approach (1) and more recently based on protein-DNA interfaces (2). The role of DNA structure in protein-DNA recognition is still not fully understood. DNA is flexible in comparison to more condensed proteins. It has been noted earlier that DNA structure is a function of dinucleotide geometry (3). In addition, DNA conformation is determined by its sequence environment and, if bound, by the interaction with its protein partners. Thus, indirect readout is associated with intrinsic DNA structure and direct readout accompanied by protein-induced conformational changes (4).
The talk will address the crucial role of sequence-specific DNA structure and flexibility in protein-DNA recognition. An all-atom Monte Carlo (MC) simulation approach (5) based on a simplified energy function has been applied to conformational sampling of DNA. Starting from standard A- and B-DNA conformations, our DNA structure prediction method explores the intrinsic contribution of DNA sequence to protein-DNA recognition. Due to their biological significance, we exemplify our search for a DNA-centric structural recognition code based on transcription factor binding sites.
The DNA binding affinity of papillomavirus E2 proteins depends on DNA base pairs not involved in direct readout. Our MC studies have shown in accordance with high-resolution X-ray data that the high affinity binding site of the cancer-associated human papillomavirus is intrinsically bent (see figure, top left, for representative conformational state) (6, 7). However, bending is not seen for the binding site of the bovine papillomavirus (8). Another group of transcription factors are homeodomain-containing proteins of the morphology-regulating Hox protein family. New X-ray data show that a Hox protein-cofactor heterodimer binds high- and low-specificity DNA sites in a different binding geometry. The profile of the DNA minor groove was identified as an intrinsic structural feature in MC simulations (9). Here, groove narrowing appears where protein side chains are inserted probably assisting direct readout. For the DNA recognition by the E. coli Sigma-E transcription factor, which regulates cellular stress responses, the minor groove profile was also shown to be of intrinsic origin. Base pairs not involved in direct readout were identified by genetic screening experiments to be required for binding. MC based computational mutagenesis experiments have proven that those base pairs reinforce an intrinsic minor groove profile as seen in the X-ray structure (10).
We revealed bending and minor groove profile as intrinsic recognition code elements, which play a major role in both direct and indirect readouts. The discussed global structural features are reasoned by sequence-specific local structural parameters of DNA. In order to extend our understanding of a DNA-centric structural recognition code for transcription factors in higher oligomeric states, we are currently investigating recognition elements of DNA binding to the tumor suppressor protein p53 (11).
References and Footnotes
1Howard Hughes Medical Institute, Department of Biochemistry and Molecular Biophysics, Columbia University, New York, USA