STRUCTURE OF GENE 5 PROTEIN-OLIGODEOXYNUCLEOTIDE COMPLEXES AS DETERMINED BY H-1, F-19, AND P-31 NUCLEAR MAGNETIC-RESONANCE

1H NMR spectra at 270 MHz of gene 5 protein from bacteriophage fd and its complexes with tetra- and octadeoxynucleotides show that .apprx. 12 of the 37 aromatic protons of the protein undergo upfield shifts upon nucleotide binding. In the complex with d(pT)8, the upfield shifts of the aromatic protons average .apprx. 0.3 ppm, while in the d(pA)8 complex the same resonances (assigned to tyrosyl protons) shift upfield .apprx. 0.8 ppm. These are interpreted as ring currents shifts induced by stacking of the phenyl rings of 3 of the 5 tyrosyl residues with the bases of the nucleotides. 19F NMR of m-fluorotyrosyl gene 5 protein shows 5 separate resonances: 2 downfield from m-fluorotyrosine corresponding to buried tyrosyls and 3 near m-fluorotyrosine corresponding to surface tyrosyls. The latter (assigned to Tyr-26, -41 and -56, shown by chemical modification to be exposed to solvent) move upfield on nucleotide binding. The downfield 19F resonances are unaffected. The aromatic protons shifted upfield on nucleotide binding appear to be those of Tyr-26, -41 and -56. In contrast to tetra-, octanucleotide binding to gene 5 protein induces large changes in the 1H resonances of the -CH3 groups of the Val, Leu and Ile side chains. These may reflect conformational changes induced by protein-protein interactions between 2 monomers bound to the octanucleotide. 1H resonances of the .epsilon.-CH2 groups of the lysyl residues in the protein and the complexes with nucleotides are narrow with long T2 values, suggesting considerable rotational motion. .epsilon.-NH3+-phosphate interactions, if they occur, are on the surface of the complex and allow the .epsilon.-CH2 groups to retain considerable rotational freedom. 31P NMR of the bound nucleotides shows large decreases in T1 for the 3''-5'' diesters, but little chemical shift suggesting no unusual distortion of the nucletide backbone on binding to gene 5 protein. A 3-dimensional model of a gene 5 protein-octanucleotide complex was built based on predictions of the secondary structure from the amino acid sequence (87 AA) and tertiary folding dictated by known chemical and NMR features of the complex.