Modeling of the Water Network at Protein–RNA Interfaces

Water plays an important role in the mediation of biomolecular interactions. Thus, accurate prediction and evaluation of water-mediated interactions is an important element in the computational design of interfaces involving proteins, RNA, and DNA. Here, we use an algorithm (WATGEN) to predict the locations of interfacial water molecules for a data set of 224 protein-RNA interfaces. The accuracy of the prediction is validated against water molecules present in the X-ray structures of 105 of these complexes. The complexity of the water networks is deconvoluted through definition of the characteristics of each water molecule based on its bridging properties between the protein and RNA and on its depth in the interface with respect to the bulk solvent. This approach has the potential for scoring the water network for incorporation into the computational design of protein-RNA complexes.