An RNA Structural Ensemble Derived by Combining MD and Elongated NMR RDCs Provides Evidence for Adaptive Recognition via Conformational Selection

Abstract

We report implementation of a general strategy that combines molecular dynamics (MD) simulations and “elongated” NMR residual dipolar couplings for constructing atomic resolution dynamical ensembles of RNA with timescale sensitivity extending up to milliseconds. The ensemble of HIV-1 transactivation response element (TAR) constructed using this approach reveals spatially correlated inter-helical motions similar to those reported recently. Bulge residue U23 stacks onto A22, and together they undergo limited motions with U23 occassionaly looping out. Greater mobility is observed for the looped out bulge reidues C24 and U25. The corresponding ensemble obained for HIV-2 TAR in which bulge residue C24 is omitted, shows a significant reduction in the local mobility at A22, U23, and U25 which is accompanued by a reduction in the amplitude of inter-helical motions. The HIV-2 TAR ensmemble exhibits even more pronounced correlations between twisting motions around the two helices, indicating that these spatial correlations originate in part from having helices exert a pull on the other during to the finite length of the linker. The HIV-1 TAR ensemble contains conformers with local binding pockets and global inter-helical orientations similar to those observed in seven distinct ligand bound TAR conformations, supporting the notion that adaptive recognition occurs via ‘conformational selection’.