MCSS-based predictions of RNA binding sites

The diversity of RNA tertiary structures provides the basis for specific re cognition by proteins or small molecules. To investigate the structural bas is and the energetics which control RNA-ligand interactions, favorable RNA binding sites are identified using the MCSS method, which has been employed previously only for protein receptors. Two different RNAs for which the st ructures have been determined by NMR spectroscopy were examined: two struct ures of the TAR RNA which contains an arginine binding site, and the struct ure of the 16S rRNA which contains an aminoglycoside binding site (paromomy cin). In accord with the MCSS methodology, the functional groups representi ng the entire ligand or only part of it (one residue in the case of the ami noglycosides) are first replicated and distributed with random positions an d orientations around the target and then energy minimized in the force fie ld of the target RNA. The Coulombic term and the dielectric constant of the force field are adjusted to approximate the effects of solvent-screening a nd counterions. Optimal force field parameters are determined to reproduce the binding mode of arginine to the TAR RNA. The more favorable binding sit es for each residue of the aminoglycoside ligands are then calculated and c ompared with the binding sites observed experimentally. The predictability of the method is evaluated and refinements are proposed to improve its accu racy.