Differential effects of genomic context on the binding specificity of paralogous transcription factors

Duke University, Durham NC

Investigators

Abstract

Cells often use similar proteins to perform very different cellular functions. Focusing on regulatory proteins that dictate which genes are expressed in each cell (among the large repertoire of >20,000 human genes), this project will decipher mechanisms that allow similar proteins to interact with distinct regions of the genome in order to perform distinct functions. This is significant because most human proteins, including the transcription factors studied in this project, are closely-related members of large families. Over time, these proteins diversified in function, often with little change in the protein sequence. Our understanding of how closely-related transcription factors can bind to different regions of the genome is very limited. To address this gap in knowledge, this project will generate high-quality transcription factor-DNA binding data and models, and will analyze the data to study genetic mechanisms by which transcription factors target specific genomic regions. The data and models will be made available to the scientific community, and are expected to become a valuable resource for future studies of differences among related transcription factors. This research will expose postdoctoral fellows, graduate, undergraduate, high school, and middle school students to the genetic mechanisms used by human cells to regulate gene expression. Trainees will participate in data generation and analysis, as well as dissemination of results. Thus, they will be introduced to molecular and computational biology through practical studies of protein-DNA interactions. Trainees will also experience the power of interdisciplinary research, as they will combine computational modeling and biological experiments to address fundamental questions about human cells. In addition, by using biological questions as motivation to teach computational techniques, this project aims to attract female students of various age groups to STEM fields. Interactions between proteins and DNA are critical for many cellular processes, including the regulation of gene expression. This project focuses on how closely-related transcription factor proteins identify their unique DNA binding sites across the human genome, in order to regulate the expression of target genes. Interactions between transcription factors and DNA occur over short regions, oftentimes <10 nucleotides. However, the DNA binding sites always occur within a specific genomic context, which is very likely to influence the interactions with transcription factors. This project will investigate three mechanisms by which genomic context can influence binding of closely-related factors: homotypic clustering (i.e. the presence of additional binding sites for factors from the same protein family), heterotypic clustering (i.e. the presence of binding sites for factors from other families), and the presence of repetitive DNA sequence elements (which were recently shown to affect protein-DNA binding). The project will use high-throughput cell-free assays to quantitatively measure the binding levels of 21 human transcription factors (from 9 distinct proteins families) for tens of thousands of DNA sites in their native genomic sequence contexts. For each protein family, DNA sequences with different arrangements of binding sites and different repeat elements will be selected from the human genome and synthesized de novo on glass slides. Binding of related transcription factors to the selected genomic sequences will be measured quantitatively, on the slides, using genomic-context protein-binding microarray (gcPBM) assays. gcPBM data are quantitative, highly reproducible, and can detect differences in binding between closely related factors, which makes the gcPBM assay ideal for this project. The generated data will be used to develop computational models of protein-DNA binding specificity, and to identify differences in specificity between related factors. All models will take into account characteristics of the genomic context (homotypic clustering, heterotypic clustering, and repetitive DNA elements). The models will be validated computationally and experimentally. Finally, the project will assess the extent to which different mechanisms can explain differential genomic binding in the cell. By identifying genetic mechanism that contribute to the differential DNA binding of closely-related transcription factors, this project represents a significant step forward in understanding how these regulatory proteins select distinct genomic targets and perform distinct functions in human cells.