Pure-phase heteronuclear multiple-quantum spectroscopy using field gradient selection

Abstract

The fundamental advantages of using pulsed field gradients to improve multidimensional NMR experiments are now well established. Unwanted resonances, such as that of water in spectra of biomolecules, can easily be eliminated ( I ) , and coherences belonging to groups of nuclei of particular interest can be efficiently selected (2). Many experiments also inherently provide quadrature detection in the indirectly detected dimension because of the selection of a single coherence pathway (+2wr -P ~2) as opposed to a linear combination of pathways that results in amplitude modulation of a single quadrature channel (3). Less frequently addressed are problems associated with maintaining optimum resolution of cross peaks in 2D plots (4). Straightforward implementation of quadrature detection in the indirect dimension often produces phase-twisted peaks. This results because of the direct correlation of phases of coherence in one dimension with phases of magnetization in the other dimension (phase modulation of signal by t, evolution). For non-gradient-enhanced versions of 2D and 3D experiments these problems have traditionally been addressed using hypercomplex data acquisition in a manner such as that described by States et al. (5)) or time-proportional phase incrementation in a manner such as that described by Marion and Wtithrich (6). Neither of these procedures is directly applicable to gradient-enhanced spectroscopy because of selection of a single coherence pathway. Here we take a lead from a recent paper by Hurd et al. ( 7) that employs changes in the sign of gradient pulses to retrieve both positive and negative coherence pathways in a single experiment. The procedure described in that paper places high demands on gradient pulse length and may not, at this stage of hardware development, be generally applicable. However, a procedure using gradient changes in alternate t, points provides a simple alternative. Here, we acquire hypercomplex data for a heteronuclear multiple-quantum-coherence (HMQC) experiment. The data show pure amplitude modulation in both channels and transform to give pure-absorption signals. Moderate efficiency of acquisition is retained as illustrated with a 13C‘H natural-abundance HMQC data set on the trisaccharide Lewisx glucal (galactosyl-& ( 1,4) [ CX( 1,3 )-fucosyl] ghtcal). These data required just 36 min on 10 mg of the trisaccharide. Suppression of 12C lines is complete with just two pulses per t, point.