Exploring the Hidden World of Conformational Penalties in Nucleic Acid Recognition

Abstract

Molecular recognition forms the fundamental physical mechanism underlying most biochemical processes occurring in cells. This process of preferential binding of biomolecules to others gives rise to several biological phenomena, such as enzyme catalysis of specific substrates, allosteric regulation, interaction of transcription factors with specific DNA sequences and the plethora of RNA-protein, RNA-ligand and other tertiary interactions, that play vital roles in processes such as riboswitch mediated gene expression, micro-RNA processing, ribosome assembly and RNA splicing. Through recent breakthroughs in x-ray crystallography and cryo-EM, we have now solved the structures of thousands of such biomolecular complexes, which have illuminated the importance of intermolecular interactions in recognition.However, most biomolecules do not fold into a single static structure, instead they form dynamic ensembles sampling different conformations with varying populations and lifetimes, as determined by their energetic landscapes. Such conformational changes are accompanied by an energetic cost or penalty, which contributes to the overall binding energetics. For nucleic acids, these conformational penalties could be significant, as many proteins melt stable duplexes before accessing single stranded regions to form specific interactions. Furthermore, sequence-dependent conformational energetics could also further enhance the specificity of these interactions. However, we currently lack robust experimental and computational approaches to determine ensembles and predict sequence dependent conformational penalties. Also largely unexplored, are the kinetic mechanisms that determine the rates of interconversion between these conformations in biomolecular reactions, thereby impacting cellular behavior and biological function. This dissertation develops new methods to determine the energetic landscapes of nucleic acid ensembles, using solution state nuclear magnetic resonance (NMR) spectroscopy and engineered mutations and chemical modifications. We recently developed a novel approach to determine RNA ensembles by combining NMR residual dipolar coupling measurements (RDCs) and chemical shifts with RNA structure prediction using Rosetta’s Fragment Assembly of RNA with Full-Atom Refinement (FARFAR), termed FARFAR-NMR. Using this approach, we were able to obtain an ensemble model for the dominant conformational state in HIV-1 TAR RNA, which enabled us to visualize the picosecond-to-nanosecond (ps-ns) inter-helical motions along the junctional motif in TAR. We further extended the approach by combining FARFAR-NMR with NMR relaxation dispersion (RD) measurements, which enabled us to capture the 3D structures of both the dominant ground state (GS) conformation as well as a lowly (~15 %) populated protonated excited state (ES) conformation with an exceptionally short lifetime of ~ 45 μs. We found that the apical loop of TAR zips up in the ES, sequestering key residues in the apical loop essential for intermolecular interactions with its protein binding partner, demonstrating the structural ability of the ES to disrupt molecular recognition. To further dissect the kinetic barriers separating the GS and the ES in TAR, we derived a kinetic framework for proton-coupled conformational changes in nucleic acids, combining pH dependent NMR dynamics measurements with chemical modifications to perturb the pKa or the conformational equilibrium. Using this framework, we were able to demonstrate that the conformational change in TAR follows the induced-fit pathway, with conformational change being the rate-limiting step. This work also revealed more generally how nucleic acid conformational landscapes, punctuated by states with distinct energetic stabilities, lifetimes, and pKa can be used to elicit complex kinetic responses to pH changes and chemical modifications, even in binding reactions involving the simplest ligand—the proton.