Computational analysis of propensities of amino acids and nucleotides usage at protein-nucleic acid interfaces

The mechanism of protein-nucleic acid interaction is still not very clear, especially that of protein-RNA interaction. Therefore, with the increasing of available protein-nucleic acid complex structures in the protein data bank database, we have collected all the structures and then analyzed the rules controlling the recognition of residues by nucleotides using classical statistical methods. The results show the number of nucleotide-binding residues presents significant differences among each kind of protein - nucleic acid complex structures with different functions. Second, Arg is the most popular in both protein-DNA and protein-RNA interactions. Moreover, the size and orientation of the polarity of amino acids play important roles in determining whether they are combined with DNA/RNA molecules, and the steric hindrances formed by the side chain of amino acids appear to influence the course of recognizing residues by nucleotides. In addition, the choice of the type of amino acids for space-adjacent binding residues with cooperative interactions in protein-DNA complexes is significantly different with that of protein-RNA complexes, and it is also found that space-adjacent residues with cooperative interactions present frequently sequence-adjacent. Finally, specificity of amino acid usage for binding residues reduces while the threshold in defining of nucleotide-binding residues increases; however, there is no change in types of amino acid regardless of popular or unpopular ones.