On the measurement of 15N-{1H} nuclear Overhauser effects.

Accurate quantification of the 15N-{1H} steady-state NOE is central to current methods for the elucidation of protein backbone dynamics on the fast, sub-nanosecond time scale. This experiment is highly susceptible to systematic errors arising from multiple sources. The nature of these errors and their effects on the determined NOE ratio is evaluated by a detailed analysis of the spin dynamics during the pair of experiments used to measure this ratio and possible improvements suggested. The experiment that includes 1H irradiation, is analyzed in the framework of Average Liouvillian Theory and a modified saturation scheme that generates a stable steady-state and eliminates the need to completely saturate 1H nuclei is presented. The largest source of error, however, in 1H-dilute systems at ultra-high fields is found to be an overestimation of the steady-state NOE value as a consequence of the incomplete equilibration of the magnetization in the so-called "reference experiment". The use of very long relaxation delays is usually an effective, but time consuming, solution. Here, we introduce an alternative reference experiment, designed for larger, deuterated systems, that uses the fastest relaxing component of the longitudinal magnetization as a closer approximation to the equilibrium state for shorter relaxation delays. The utility of the modified approach is illustrated through simulations on realistic spin systems over a wide range of time scales and experimentally verified using a perdeuterated sample of human ubiquitin.