Molecular basis for the activity of the Box C/D snoRNP methylation enzyme. A combined solution-state NMR, solid-sate NMR and electron microscopy approach
Biochemistry
Final Report Abstract
Using the archaeal Box C/D sRNP, which, in contrast to the eukaryotic Box C/D snoRNP, can be reconstituted in vitro in an active form, we determined the solution-state structures of the half-loaded substrate D- and substrate D’-bound mono-RNPs. Instead of a single, welldefined conformational state, we found that fibrillarin exchanges between the substrate-bound and -unbound states, with the substrate D’-loaded complex displaying a higher population of the methylation-competent [on,off]-state than the substrate D-loaded RNP. Fittingly, the substrate D’-loaded RNP achieves higher levels of methylation. The existence of different equilibria of substrate-bound and unbound conformers of fibrillarin could not be detected by X-ray crystallography, which instead selects for the most ordered conformation. Our results suggest that the percentage of methylation-competent conformation is subtly tuned by the free energy difference between the active [on,off]- and inactive [off,off]-conformations (and possibly also by the kinetics of transition, on which our structural data at equilibrium do not provide any information). Recognition of the RNA ribose by fibrillarin is accompanied by a loss in entropy at the junction between the Nop5-NTD and the Nop5 coiled-coil and between the Box C/D (or Box C’/D’) RNA elements and the substrate–guide duplex. In addition, upon fibrillarin binding, the substrate–guide helical structure must deviate from the ideal A-form geometry, in order to adapt to the proteins. This is particularly evident at the 3’-end of the helix, where any base pair beyond the tenth is either melted or heavily distorted. These energetically costly events are compensated by the formation of contacts between fibrillarin and the RNA backbone, as well as by contacts between the Nop5-CTDs and fibrillarin and the two ends of the substrate–guide duplex. In addition, the MD simulations showed that the substrate D’–guide duplex is less stable at the substrate 5’ end than the substrate D–guide duplex. We hypothesize that the presence of an A–U base pair at this end favours turnover of substrate D’ by a sort of “zip” mechanism involving Nop5 helix α9’. In agreement with this, the zip-sR26 RNA, where the initial A–U base pair of the substrate D’– guide duplex is substituted by a G–C base pair, displays a notably reduced efficiency of methylation of substrate D’. In conclusion, methylation efficiency appears to be regulated by a complex interaction network depending on the substrate rRNA sequence beyond the methylation site and the C-G content of the substrate–guide duplex. We propose that, together with substrate turnover, the ability of different substrate-guide duplexes to generate different equilibria of [on,off]- and [off,off]-state conformers further modulates the level of methylation at distinct rRNA sites. When the difference in the free energies of the active and inactive enzyme states is small, the ratio between the populations of the active and inactive conformations provides a mechanism to tune the activity level. When the free energy difference is large and positive, the population of the methylation competent conformation becomes vanishingly small and the Box C/D RNP looses its capacity to catalyse methylation. This situation could be the basis for the repurposing of Box C/D complexes to functions unrelated to methylation. To calculate the structural models of the substrate D- and substrate D’-loaded RNPs we developed a novel hybrid structure calculations protocol that fits a combination of NMR and SAS data to an ensemble of conformations. The protocol proceeds in sequential steps, where the diversity of the structural ensembles is progressively increased while increasing the demand on the quality of the fit between predicted and experimental data. This protocol is particularly suited for multicomponent complexes, where the conformational heterogeneity affects the position of a defined subset of modules. We anticipate the methodology developed here to be generally applicable to modular enzymes undergoing conformational changes during catalysis.
Publications
- “Archaea box C/D enzymes methylate two distinct substrate rRNA sequences with different efficiency” RNA 2016 22, 764-772
A. Graziadei, P. Masiewicz, A. Lapinaite, T. Carlomagno
(See online at https://doi.org/10.1261/rna.054320.115)
