The 1 H NMR spectra of biological macromolecules fre-sequences. Excitation sculpting consists of applying the double-pulsed field-gradient spin-echo (DPFGSE) ( 14) se-quently exhibit regions of severe overlap, making it necessary to acquire spectra with the highest possible resolution quence -G 1 -S-G 1 -G 2 -S-G 2 -to a system with transverse magnetization, where S is some pulse or train of and to use two-dimensional (2D) or three-dimensional (3D) techniques to obtain unambiguous information. Several pulses and G 1 and G 2 are gradient pulses with unrelated amplitudes. In the original analysis of excitation sculpting groups have shown that in this situation substantial savings of spectrometer time can be made by replacing the 2D and ( 14), it was elegantly demonstrated that the net effect of this sequence is simply to return the transverse magnetization 3D experiments with their selective, one-dimensional (1D) counterparts (1-8). Such experiments use selective excita-to its starting place with its intensity scaled by a factor P 2 , where the frequency-dependent factor P is the degree to tion sequences in place of the indirect frequency dimensions of the analogous multidimensional experiments, giving a which longitudinal magnetization is inverted by the sequence S. Thus, the selectivity of the method depends solely on the reduction in the minimum recording time. This approach is attractive when only a restricted amount of spectral informa-inversion properties of S, and no frequency-dependent phase distortions are introduced, regardless of the specific choice tion is required, and when the sample gives adequate signalto-noise ratio with recording times that are less than the of S. It is therefore an extremely versatile technique and has already been used for solvent suppression (14, 17) and 2D minimum recording time of the full-dimensionality experiment, a situation which is more likely to occur when very heteronuclear experiments (18-21), as well as the abovementioned selective 1D experiments. high resolution is required in the indirect dimensions.
NMR studies of oligosaccharides often meet these condi-This article describes the use of excitation sculpting in the construction of 1D ROESY (Fig. 1b), TOCSY-ROESY tions, and in recent years, selective 1D equivalents of homonuclear ( 1 H) multidimensional experiments have been devel-(Fig. 1c), and ROESY-TOCSY (Fig. 1d) experiments, as well as the previously reported 1D TOCSY (Fig. 1a) experi-oped and widely used in structural studies of oligosaccharides (1,4,5,8,(9)(10)(11)(12)(13). For example, it has become ment, referred to here as XSROESY, XSTR, XSRT, and XSTOCSY, respectively. In the singly selective XSTOCSY common to obtain intraresidue assignments from selective 1D TOCSY (1,9, 10, 12) experiments, usually by recording and XSROESY experiments, one multiplet is selected by a single DPFGSE element. The doubly selective XSTR and spectra with long mixing times (about 140 ms) starting with selective excitation of an anomeric proton, resulting in co-XSRT experiments use two DPFGSE sequences and are thus 1D analogues of 3D experiments. Since the important prop-herence transfer to several of its scalar-coupled partners. Both intra-and interglycosidic connectivities can be deter-erty of the selective pulses used in the DPFGSE sequences is their inversion characteristics, there need not be any specific mined using 1D ROESY (10, 13) experiments with appropriate mixing times. Doubly selective 1D pulse sequences, phase relationship between one pair of DPFGSE pulses and another nor between either pair and the hard pulses. It is such as TOCSY-ROESY and TOCSY-NOESY, have also been used (4,11, 13). In all of these selective experiments thus possible to use different frequency sources for each pair of selective pulses and the hard pulses, making implementa-the first transfer is typically from a well-separated anomeric proton.
tion particularly straightforward. These experiments were used in the NMR analysis of an Recently, the so-called ''excitation sculpting '' technique (14) has been applied to 1D TOCSY experiments (15), as about 0.4 mM D 2 O solution of a trisaccharide fragment (Fig.
2), obtained from digestion and subsequent oxymercuration well as 1D COSY (15), NOESY (16), and RELAY (15) of chondroitin sulfate C, a glycosaminoglycan polysaccharide. The primary structure of this molecule had been de- โ To whom correspondence should be addressed.