Direct Identification of Relayed Nuclear Overhauser Effects

The measurement of nuclear Overhauser effects (1) for determination of the threedimensional structure of mo lecules in solution is rapidly gaining popularity. The bui ldup rate of the NOE depends on r6, where I is the interproton distance. Measu.rement of the NOE buildup rates therefore provides a means to measure interproton distances. Most commonly, two-dimensional NMR experiments are used for measur ing the NOES in complicated spectra of macromolecules that are in the slow tumbling lim it (UT, > 1). To measure the true bui ldup rate, the 2D experiment has to be repeated fair a series of m ixing times, and the cross-peak intensity is mon itored as a function of this m ixing time (2-5). A complication commonly occurs if two protons, B and C, are close in space and a third proton, A, is close to B but distant from C. In this case, a cross peak can be observed between A and C which could be m istakenly interpreted as a direct NOE connectivity between A and C. Interpretation of the intensity of this cross peak generally will lead to an erroneous AC distance determination, i.e., to an incorrect structure. In principle, the shape of the NOE buildup curve could be used to distinguish direct from relayed NOES. However, to obtain a useful bui ldup curve it is especially important to measure the NOE for very short m ixing times. For these short m ixing times the cross peaks have very low intensities and consequently the sensitivity of this approach is poor. It is demonstrated here that direct and relayed NOES can readily be distinguished by using spin-locked NOE spectroscopy (6, 7). In the spin-locked NOE experiment the NOE effect is always positive and increases with slower mo lecular tumbling. Therefore, cross peaks due to direct NOE are always opposite in sign relative to the diagonal and consequently, cross peaks relayed via an intermediate nucleus will be in phase with the diagonal resonances. O f course, doubly relayed signals would be of a’pposite sign again but these resonances are usually too weak to be detected unless very long m ixing times are used. In 1D NOE difference spectra of small mo lecules (‘UT, < 1) this alternation of the sign of the NOE is commonly observed and is known a.s the “three-spin effect” (I).