Development and application of a quantitive model for HIV-1 transcriptional activation driven by TAR RNA conformational dynamics

Project Summary Many processes essential for HIV-1 viral replication are driven by the association of regulatory RNA elements in the retroviral genome and host/viral proteins that form biologically active complexes. Despite advances in solving the 3D structures of RNA-protein complexes and in measuring RNA-protein interactions in vitro and in cells, our current understanding of RNA-protein interactions is qualitative and not quantitative, descriptive, not predictive. Attaining a quantitative and predictive understanding is necessary to reveal the forces and conformational states driving viral processes and to fully define the landscape of opportunities for therapeutic intervention. The overarching hypothesis of this proposal is that the cellular transcriptional activity of HIV-1 TAR RNA can be predicted from its sequence based on its conformational propensity to form the binding-competent conformation and its affinity for the transactivator protein Tat and the human super elongation complex (SEC) Tat:SEC. The project will (i) develop a suite of technologies to obtain experimental data on RNA ensembles, RNA-protein interactions, and cellular activity quantitatively in high throughput over a large and common expanse of RNA sequence and structure space (ii) closely integrate NMR data with computational molecular dynamics (MD) simulations and empirical RNA structure prediction algorithms (FARFAR) to determine RNA ensembles free and bound to proteins and to test and guide refinement of the computational models through a community wide effort; and (iii) test and refine a thermodynamic model predicting cellular function that dissects TAR•Tat:SEC binding energetics into contributions from intermolecular contacts and conformational propensities. From a common library of 1000s of TAR RNA variants, Aim 1 will determine conformational propensities and measure binding affinities to Tat and Tat:SEC across solution conditions and measure transcriptional activation in cells and with Tat concentration varied. The data will be used to develop a quantitative and predictive model for cellular transcriptional activation based on in vitro measurements and iteratively refine the model. Aim 2 will integrate NMR data with MD simulations and FARFAR; determine atomic-resolution ensembles for 20 TAR variants, free and bound to the Tat RNA binding domain; use the ensembles to define the bound conformational states and refine conformational propensities; identify strengths and weaknesses of MD and FARFAR; and develop and test a new method (FARFAR-CS) for determining RNA ensembles and use it to refine propensities for 100s of TAR variants. Aim 3 will extend the model to include alternative secondary structure propensities, use NMR experiments to measure these propensities for 100s of TAR variants, and extend the model to include binding of 20 small molecules and competition with 7SK RNA for 1000s of RNA variants. When completed, this project will make it possible to quantitatively predict cellular transcriptional activity from TAR sequences, will reveal the profound contribution of conformational propensities to RNA-protein binding, and will provide a roadmap for future efforts that link biochemical and biophysical properties to molecular behavior and function in cells.