Exploration of metal ion binding sites in RNA folds by Brownian-dynamics simulations

Background: Metal ions participate in the three-dimensional folding of RNA and provide active centers in catalytic RNA molecules. The positions of metal ions are known for a few RNA structures determined by X-ray crystallography. In addition to the crystallographically identified sites, solution studies point to many more metal ion binding sites around structured RNAs. Metal ions are also present in RNA structures determined by nuclear magnetic resonance (NMR) spectroscopy, but the positions of the ions are usually not revealed. Results: A novel method for predicting metal ion binding sites in RNA folds has been successfully applied to a number of different RNA structures. The method is based on Brownian-dynamics simulations of cations diffusing under the influence of random Brownian motion within the electrostatic field generated by the static three-dimensional fold of an RNA molecule. In test runs, the crystallographic positions of Mg2+ ions were reproduced with deviations between 0.3 and 2.7 Angstrom for several RNA molecules for which X-ray structures are available. In addition to the crystallographically identified metal ions, more binding sites for cations were revealed: for example, tRNAs were shown to bind more than ten Mg2+ ions in solution. Predictions for metal ion binding sites in four NMR structures of RNA molecules are discussed. Conclusions: The successful reproduction of experimentally observed metal ion binding sites demonstrates the efficiency of the prediction method. A promising application of the method is the prediction of cation-binding sites in RNA solution structures, determined by NMR.