Molecular basis of genetic regulation and robustness
Date
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
YU Faculty Profile
Abstract
Proper cell functioning depends on transcription factors (TFs) to control the spatio-temporal expression of its genetic material. Our current knowledge on where TFs bind and associate to regulate gene expression is incomplete. We developed a structure-based computational algorithm (TF2DNA) to identify binding specificities of TFs. Three-dimensional models of TFs are optimized in a molecular mechanics force field with all possible DNA sequences docked to them. Relative binding affinities are assessed using a knowledge-based potential. TF2DNA predictions were benchmarked against experimentally known TF binding motifs, showing a success rate of 82%. The method predicted 1,321 binding motifs (72.4%) from 1,825 predicted human TFs. The predicted human motifs were used to search promoter sequences and reconstruct the human regulatory network. The data is stored in a relational database that is available through a searchable web page (www.fiserlab.org/tf/). Gel shift assay experiments confirmed the TF2DNA prediction for the functionally uncharacterized T-cell leukemia homeobox 3 (TLX3) TF. The TLX3 motif identified genes enriched in functions related to hematopoiesis, tissue morphology, endocrine system and connective tissue development and function.;Gene regulatory networks (GRNs) show robustness to perturbations; however the underlying molecular mechanisms behind the acquisition of robustness remain poorly understood. We used a multi-tier model to integrate molecular sequence and structure information with network architecture and population dynamics. In the model, mutations in the DNA cause changes on two levels: (a) at the sequence level in individual binding sites (modulating binding specificity), and (b) at the network level (creating and destroying binding sites). The model was used to dissect the underlying mechanisms responsible for the evolution of robustness in GRNs. Results suggest that sparse architectures (represented by short promoter regions resembling more prokaryote-like organization), exploit a mixture of local-sequence and network-architecture level changes. At the local-sequence level, robustness evolves by preserving existent binding sites and avoiding the generation of new binding sites. Meanwhile, highly interconnected architectures (represented by long promoter regions, more typical for eukaryotic organisms), evolve robustness almost entirely via network level changes, deleting and creating binding sites that modify the network circuitry, thereby selecting alternative programs that are more robust to perturbations.