Two-dimensional proton NMR of gene 5 protein indicates that only two aromatic rings interact significantly with oligodeoxynucleotide bases

Publisher Website
Google Scholar
Abstract
The interactio of gene 5 protein (G5P) with oligodeoxynucleotides is investigated by 1H NMR methods, principally two-dimensional nuclear Overhauser effect spectroscopy (NOESY). Aromatic resonances of G5P are specifically assigned from crystallographic data, while the low-field resonances of nucleotides are assigned with sequential or other procedures. Chemical shift changes that accompany binding of d(pA)4, d(A)4, d(pT)4, and d(pA)8, combined with specific protein-nucleotide nuclear Overhauser effects (NOEs) obtained from Noesy spectra, suggest that Phe-73 and Tyr-26 are the only aromatic residues that stack significantly with nucleotide bases. Chemical shift data also imply a role for Leu-23, though this has not been confirmed with intermolecular NOEs. Binding of all four oligonucleotides caused marked upfield movements (0.1-0.6 ppm) of G5P NOESY cross peaks belonging to Tyr-26, Leu-28, and Phe-73. Most other G5P spin systems, notably those of Tyr-34 and Tyr-41, do not appear to be significantly affected. In the d(pA)4-G5P complex an intermolecular NOE is observed between Tyr-26 and H1'' of Ade-1, while Phe-73 has NOEs with the H2, H8, and H1'' protons of Ade-2 and -3. Intermolecular NOEs seem to follow a similar pattern in the partially cooperative d(pA)8-G5P complex, though specific nucleotide resonance assignments are not possible in this case. Binding causes small chemical shift changes for the bases resonances in adenylyl nucleotides, suggesting that there is some, but not complete, unstacking of the bases. In the d(pA)8 complex, the upfield ring current shifts of the Tyr-26 and Phe-73 resonances become greater than those caused by d(pA)4, implying that cooperative interactions result in tighter complex. The view of the G5P-oligonucleotide interaction that emerges from this study is different from one previously proposed from model-building exercise [Brayer, G. D., and McPherson, A. (1984) Biochemistry 23, 340-349]. The major point of difference concerns the role of Tyr-41, which has been postulated to stack with nucleotide bases. Our NMR data provide no support for this proposal. Rather, a comparison of the NMR data with the X-ray structure of native G5P suggests that Tyr-26, Leu-28, and Phe-73'' form the major components of a dominant nucleotide interaction site that lies near a phosphate-binding electropositive cluster formed by the side chains of Arg-16, Arg-21, Lys-24, and Lys-46. Such an interaction mode creates a shorted nucleotide path across the protein surface than previously proposed. At present it is unclear whether a shorter nucleotide path is relevant to the much larger G5P-fd DNA complex, but it may facilitate "inside" packing of DNA in the macromolecular superhelix, reconciling previous neutron-scattering data with a structural model.

This publication has 26 references indexed in Scilit: