review nature methods��� |��� VOL.8��� NO.11��� |��� NOVEMBER���2011��� |��� 919 Accompanying the discovery of noncoding RNAs (ncRNAs) as abundant players in gene expression and regulation is the growing realization that most ncRNA sequences do not fold into a single native con- formation, but rather, ncRNAs sample many different conformations from their free-energy landscape1,2 to carry out their biological function. For example, many ncRNAs function as genetic switches by transitioning between entirely different secondary structural forms in response to a wide range of cellular stimuli3,4. Not only do the structural characteristics of the different states have to be optimized to carry out distinct func- tions, transitions between them have to occur at dedi- cated timescales and be triggered by specific cellular signals. Thus, a deep molecular understanding of how ncRNAs perform their functions requires insights into how RNA dynamically samples different conforma- tions along its energy landscape and how this land- scape is in turn modulated by cellular cues2. Beyond understanding function, there are many reasons motivating studies of RNA dynamics at atomic resolution. First, RNA is exploding in its importance as a drug target5, and a broader dynamic view of the Characterizing RNA dynamics at atomic resolution using solution-state NMR spectroscopy Jameson R Bothe1, Evgenia N Nikolova2, Catherine D Eichhorn2, Jeetender Chugh3, Alexandar L Hansen4���6 & Hashim M Al-Hashimi1,3 Many recently discovered noncoding RNAs do not fold into a single native conformation but sample many different conformations along their free-energy landscape to carry out their biological function. Here we review solution-state NMR techniques that measure the structural, kinetic and thermodynamic characteristics of RNA motions spanning picosecond to second timescales at atomic resolution, allowing unprecedented insights into the RNA dynamic structure landscape. From these studies a basic description of the RNA dynamic structure landscape is emerging, bringing new insights into how RNA structures change to carry out their function as well as applications in RNA-targeted drug discovery and RNA bioengineering. conformations populating the energy landscape1 is now widely recognized as essential for successfully implementing structure-based approaches in lead compound discovery and optimization2,6. Second, experimental data probing the dynamic aspects of RNA structure are urgently required for the contin- ued testing and improvement of computational force fields, which remain severely underdeveloped for nucleic acids as compared to proteins7. Third, a pre- dictive understanding of RNA dynamics will enable the design of RNA-based devices whose functionality often depends on dynamic transitions in structure. The free-energy landscape1 provides a unified and complete description regarding the dynamic proper- ties of RNA that are relevant for understanding func- tion (Fig. 1). The free energy is specified for every conformation that can be adopted by the RNA. The fractional population of a given conformer then depends on its relative free energy, whereas the rate at which two conformers interconvert depends on the free-energy barrier that separates them. Cellular cues perturb the free-energy landscape, diminishing bar- riers and/or stabilizing conformers that are otherwise 1Department of Chemistry, The University of Michigan, Ann Arbor, Michigan, USA. 2Chemical Biology Doctoral Program, The University of Michigan, Ann Arbor, Michigan, USA. 3Department of Biophysics, The University of Michigan, Ann Arbor, Michigan, USA. 4Department of Chemistry, The University of Toronto, Toronto, Ontario, Canada. 5Department of Biochemistry, The University of Toronto, Toronto, Ontario, Canada. 6Department of Molecular Genetics, The University of Toronto, Toronto, Ontario, Canada. Correspondence should be addressed to H.M.A.-H. (hashimi@umich.edu). Published oNliNe 28 oCtobeR 2011 doi:10.1038/NMeth.1735 �� 201 1 Nature America, Inc. All rights reserved. �� 201 1 Nature America, Inc. All rights reserved.

920��� |��� VOL.8��� NO.11��� |��� NOVEMBER���2011��� |��� nature methods review unfavorable, and thereby redistribute the conformer populations to effect specific biological outcomes. Although rich in informa- tion, the free-energy landscape is very complex and cannot be measured experimentally. Fortunately, important insights into biological function can be obtained by focusing on a subset of conformers that populate minima along the energy landscape. Studies increasingly show that such ���low-hanging fruit��� conform- ers are often the ones that are stabilized by cellular cues to carry out biological function8���10. These more appreciably populated conformers are also more amenable to experimental characteriza- tion using spectroscopic techniques that probe dynamic fluctua- tions in structure along the energy landscape. We refer to this partial energy landscape as the ���dynamic structure landscape���. Among many techniques that are now being developed to study RNA dynamics, solution-state NMR spectroscopy, which has also contributed considerably toward the characterization of protein dynamics11, has a unique role (Fig. 1). First, NMR spectro- scopy can be used to measure dynamics at atomic resolution, comprehensively for sugar, base and backbone moieties across different residues. Second, multiple interactions can be measured at a given site to deduce structural, kinetic and thermodynamic characteristics of not one but many motional modes occurring at different timescales. Third, NMR spectroscopy has broad sen- sitivity to motions spanning picosecond to second and longer timescales and can be used to characterize very subtle changes in conformation, including those involving minutely populated conformers (on the order of ~10���7%) that have exceptionally short lifetimes (on the order of nanoseconds). Last but not least, NMR spectroscopy is a powerful approach for exploring how the dynamic structure landscape is modulated by cellular cues, and time-resolved methods can be used to follow these perturbations in real time. Here we review solution-state NMR spectroscopy methods for studying RNA dynamics and highlight some of the new insights that have been obtained regarding the RNA dynamic structure landscape and its relationship to function. nuclear spin interactions used to study rna dynamics The basic NMR experiment can be simplistically described as follows. Nuclei behave as tiny magnets, and because of the quan- tization of the nuclear spin angular momentum, they align either parallel (�� state) or antiparallel (�� state) relative to the applied magnetic field. As the parallel (or antiparallel, depending on the nucleus) alignment is energetically more favorable, a net bulk magnetization over an ensemble of nuclei builds up parallel to the magnetic field. Radiofrequency pulses are then used to rea- lign this bulk magnetization along a direction perpendicular to the magnetic field. The bulk magnetization then precesses about the magnetic field at a characteristic NMR frequency called the ���chemical shift��� and gives rise to a detectable oscillating magnetic field. This non-equilibrium magnetization ultimately relaxes back to the equilibrium, parallel state. The time-domain spectrum is Fourier-transformed to yield the standard frequency-domain NMR spectrum, in which unique signals at characteristic chemi- cal shift frequencies are observed for different types of nuclei. For nucleic acid applications, one is typically interested in the NMR-active nuclei 1H, 13C, 15N, 2H and 31P. The 13C, 15N and 2H isotopes can be introduced during synthesis, typically by using labeled NTPs via in vitro transcription reactions (Fig. 1). The NMR chemical shift (Box 1) is proportional to the energy gap between the �� and �� states and the static applied magnetic field. However, electronic clouds surrounding nuclei can ���shield��� or ���deshield��� nuclei from the external magnetic field by variable amounts dependent on the electronic structure. This leads to a wide range of chemical shifts that makes it possible to measure dynamics with site-specific resolution. Furthermore, changes in chemical shift owing to local fluctuations in the electronic envi- ronment form the basis for relaxation dispersion experiments to measure exchange processes occurring at microsecond to millisecond timescales, ZZ-exchange spectroscopy to measure motions at millisecond to second timescales and time-resolved NMR spectroscopy experiments to follow transitions occurring at timescales longer than a few seconds. In an NMR spectroscopy experiment, interactions such as dipolar coupling and chemical shift anisotropy (CSA) (Box 1) modulate the effective field experienced by a given nucleus in an orientation-dependent manner, thereby perturbing the energy gap between the �� and �� states and the observed NMR spectrum. Many NMR spectroscopy experiments that probe dynamics take advantage of these so-called ���anisotropic��� interactions. In solution, the orientation dependence of anisotropic interactions gives rise to a time-dependent fluctuating field (Fig. 2a), which in turn influences the rate at which the magnetization relaxes back to equilibrium. The contribution to so-called transverse relaxation (R2) is encoded within the linewidth of the NMR signal, with faster motions leading to narrower lines. Spin relaxation measure- ments take advantage of these effects to probe internal motions of bond vectors at picosecond to tens of nanosecond timescales. Librations Interhelical motions Excited states Secondary structural transitions Free energy R1�� ZZ exchange CPMG Residual dipolar couplings Time-resolved NMR H-D exchange NMR spectroscopy experiment 10���12 10���9 10���6 10���3 100 103 106 Timescale of motion (s) ��s ps���ns ms s Probes O OH OH H H HO O P OH O H exchange 2���6 kcal mol���1 9���12 kcal mol���1 13���16 kcal mol���1 17 kcal mol���1 G N N N O N N H H H H C H N N O N H H H U N O O N H H H A N N N N N H H H H Spin relaxation N H N N H N N N H N N N N NN N N N OO O N N H H H N N N H H N N N N N N N H NN N N N N N N O O NH2H22 N N H N N N N N N N N N N N N N N N N N N N N N N N N N N N N N N N N H H Figure 1 | NMR spectroscopy techniques and site-specific probes for characterizing motional modes that carry RNA structure along various regions of the dynamic structure landscape. In the dynamic structure landscape (top), transition free energies corresponding to typical timescales of interconversion were estimated using transition state theory at 25 ��C. In NMR spectroscopy experiments (middle), solid lines indicate the timescales at which each NMR spectroscopy experiment is optimally suited for, and dashed lines indicate timescales that are difficult to probe. For probes (bottom), nuclei most commonly used for RNA dynamics measurements: protonated carbons (blue), imino nitrogens (green) and backbone phosphorus (red). �� 201 1 Nature America, Inc. All rights reserved. �� 201 1 Nature America, Inc. All rights reserved.