A strategy for NMR structure determination

Abstract

Structure determination by two-dimensional NMR is frequently limited by spectral overlap in large macromolecules. The well-resolved peaks in 2D nuclear Overhauser effect spectra are used to determine a limited set of internuclear distances for three-dimensional structure determination. An alternative strategy that includes both resolved and overlapping peaks is to simulate the experimental NOESY spectra as a function of mixing time using a combination of measured and estimated peak volumes. When the experimental and simulated spectra match, the peak volumes can be directly converted to relaxation rates using a matrix approach that accounts for spin diffusion. Correlation times can be measured from the relaxation rates, and the rates can be converted into internuclear distances that can be used for structure determination. By simulation of the spectra at several mixing times, the uncertainties in the distances can be evaluated. This strategy is demonstrated for the eight-base-pair TaqI DNA restriction site d(CTTCGAAG) 2, where a combination of peak overlap and spin diffusion limits the measurement of internuclear distances by conventional methods.