A two-dimensional nuclear overhauser enhancement experiment using semiselective soft pulses, and its applications to proteins

Abstract

Suitably shaped soft pulses (1-3) have frequently been used in high-resolution NMR for the purposes of semiselective excitation. They can be employed to obtain onedimensional analogs of two-dimensional sequences (1, 4) and to record 2D spectra with reduced spectral width in the wI dimension (5). This approach also facilitates the recording of 3D NMR spectra in a reasonable measurement time (6, 7). Soft pulses, however, have rarely been used as mixing pulses. Only one example has been reported (8) in which preparation and mixing in a COSY experiment were achieved with soft pulses in order to restrict coherence transfer to connected transitions only, resulting in the appearance of cross peaks with a pattern of the E.COSY (9) or z-COSY (10) type. In this communication we describe a soft NOESY experiment consisting purely of semiselective soft pulses to selectively observe NOES between amide protons in a protein. Because the solvent signal is only minimally excited, soft NOESY spectra can be recorded with short mixing times (~80 ms) on samples containing up to 90% Hz0 even in the presence of a broad water signal. Because of a phenomenon known as radiation damping (11, 12), this is difficult to achieve with other techniques that do not employ presaturation of the water resonance. This is of practical importance in both protein and nucleic acid NMR, as it is often the case that presaturation leads to the disappearance of NH resonances via saturation transfer due to rapid exchange of NH protons with solvent. Further, the soft NOESY is especially suitable for the detection of cross peaks close to the diagonal since parallel transitions appear with much lower amplitudes. Finally, under suitable conditions, the soft NOESY experiment enables one to obtain higher resolution spectra in a shorter period of time for a fixed number of transients per increment than would be possible using a classical nonselective experiment (5). For various technical reasons, one may hesitate to furnish a pulse sequence completely with semiselective pulses. These include baseline distortions, huge phase gradients, and eventually distortions of the lineshape of the excited signals. More severe problems occur when soft pulses are applied to a large area of the spectrum as the excitation profile differs from that of a 90” pulse at positions far away from the carrier