Phosphate-binding sequences in nucleotide-binding proteins.

In the three-dimensional model of adenylate kinase, the phosphate-binding site for AMP and ATP has been identified [Pai, E.F. et al. (1977) J. Mol. Biol. 114, 37--45]. In this region one can distinguish a sequence glycine XXXX glycinelysine. The same sequence is found in many other mononucleotide-binding proteins including elongation factors and oncogenic P21 proteins. Dinucleotide-binding proteins display a pyrophosphate-binding unit with a glycine pattern different from that of mononucleotide-binding proteins. It has been found that P21 ras protein possesses a strand motif typical for (pyro)phosphate binding of a mononucleotide. A single mutation at position 12 can confer oncogenic activity on the protein. Based on the assumption that amino acid residues which are critical for function are preferentially conserved, we predict from the sequence that glycine residue 15 rather than residue 12 is important for (pyro)phosphate binding.