The 2D NOESY sequence

Sign of diagonal Origin of crosspeak Sign of crosspeak 

By convention the diagonal is phased to be positive and the sign of all other signals are given relative to this. 

Selection of Xm for qualitative studies can be estimated from knowledge of longitudinal relaxation times, T). Often these are not known prior to running the experiment, so a quick measurement may assist here, or failing that an estimate based on previous knowledge. Precise measurements are little benefit since there will inevitably exist a spread of values within the molecule an a compromise value for Xn, is required in any case. For small molecules, a mixing 

Whilst a comprehensive discussion of this matter lies beyond the scope of this book, the principles involved are very briefly introduced here, the idea being that the reader should at least be able to follow discussions on structure calculations based on NOE measurements found in the chemical literature. The vast majority of work in this area has been applied to biomolecular structures and the protocols developed with macromolecules in mind, which will not translate directly to quantitative measurements in small molecules [25], although the general principles remain the same. 

The most used approach to distance measurements stems from equations 8.13 and 8.14 above and relies on a known reference distance, rxv. from which others may be calculated [26], and the assumption of uniform isotropic molecular tumbling. Recall the basic equation was  

F are 836w The variation of NOE and ZQC intensities during the NOESY mixing time. The ZQC contributions may be suppressed by making small random variations to Tm, whereas this has negligible effect on the NOE intensities. 

A more direct approach to the elimination of zero-quantum interference is via the use of swept-frequency inversion pulse applied in the presence of a field gradient, which destroys the zero-quantum contributions in a single transient. The application of this has already been described in Section 5.7.3 with a view to obtaining pure lineshapes in TOCSY spectra, and the operation of the filter is itself described in Section 10.6 and so will not be explored here. The implementation of this in the 2D NOESY sequence is illustrated in Fig. 8.37 and requires the incorporation of the filter within the usual mixing time, followed by a purging gradient. The effective suppression of zero-quantum interference is illustrated in the high-resolution NOESY spectra of Fig. 8.38 in which spectrum (a) shows clear anti-phase 

---

Figure 8.34. The 2D NOESY sequence and the associated coherence transfer pathway. The optional use of a pulsed field gradient (shown greyed) during the mixing time is described in the text.

Another important NMR experiment for monitoring catalyst performance is the 2D EXSY pulse sequence, which is equal to the 2D NOESY sequence (nuclear Over-hauser effect spectroscopy). The diagonal on these two-dimensional homonuclear spectra represents the ID spectrum. One wiU find two kinds of cross peaks in the... 

It should be noted that NOESY and ROESY pulse sequences also provide EXSY spectra, and therefore EXSY cross peaks may appear simultaneously in the 2D NOESY and ROESY spectra. EXSY cross peaks are always positive in both types of experiment, whereas dipolar cross peaks are negative in EXSY spectra independently of molecular weight and in NOESY spectra of small molecules. Therefore, in macromolecules the sign for NOESY and EXSY cross peaks is the same, and the two phenomena cannot be distinguished in NOESY experiments. In contrast, ROESY cross peaks have different sign from EXSY cross peaks and can be distinguished and even plotted selectively in ROESY experiments. These considerations are summarized in Table 8.3 for the reader s convenience. 

Exchange in solids can be studied by NOESY, except that the preparation period consists of a cross polarization sequence, rather than just a 90° pulse. Similar studies can also be carried out with 2H NMR by combining the 2D pulse sequence with the solid echo described in Chapter 7. 

In parallel with the ID NOESY sequences above, the 2D ROESY experiment also has its ID equivalent (in fact, this was the original ROE experiment [60]) and gradient-selected analogues [70-72] all of which incorporate selective excitation of the target spin. These can be derived from ID NOESY sequences by incorporation of a suitable spin-lock in place of the 90-Tm-90 segment of NOESY, and thus require no further elaboration. 

As for the 2D NOESY data described above, spectra are vulnerable to interference from undesirable zero-quantum contributions between coupled spins this is apparent, for example, in trace (c) of Fig. 8.42 between H4 and H5. This can be especially problematic when shorter mixing times are employed, as in the generation of NOE build-up curves [35], and again the inclusion of the swept-frequency/gradient zero-quantum filter should prove beneficial. The complete selective ID NOESY sequence incorporating this and employing the optimised DPFGSE selection procedure is presented in Fig. 8.43, and an illustration of the improvements provided by the filter may be seen in Fig. 8.44 where the removal of unwelcome anti-phase dispersive contributions between coupled spins is apparent in (b). 

In addition to the usual evolution time, t, this sequence contains a fixed mixing time during which the NOE information is generated. The 2D NOESY spectrum consists of a diagonal, plus cross-peaks at the chemical shifts which demonstrate an NOE. 

Figure 4 (A) The 2D NOESY pulse sequence, which uses three 90° pulses. The times U and fs are the evolution and detection periods, respectively is the mixing time during which only longitudinal magnetization is retained, either by gradients or by cycling the phases (01,02.03) of the pulses. (B) Schematic NOESY spectrum, showing that in such spectra the NOEs are manifested as cross-peaks between the various spins, the resonances of which lie on the diagonal.

Fig. 17.7 2D NOESY experiment with 13C, 15N fil- The sequence uses double filters for 13C, single ter in both H dimensions, selecting intramolecu- filters for 15N, and additional 13C, 15N purge lar NOEs within the unlabeled components [22]. pulses during the NOE mixing time ( ).

The pioneering work in this field, a two-dimensional relayed-NOE experiment proposed by Wagner [7], was quickly followed by the appearance of several related NMR techniques [8-17]. Application of isotropic mixing during the J-transfer period yielded the 2D TOCSY-NOESY [11, 15] and NOESY-TOCSY [12, 14] experiments. When spin-lock sequences were applied to both J and NOE-transfers, the 2D TOCSY-ROESY and ROESY-TOCSY experiments [10, 16, 17] emerged. 

Fig. 8.2. Some of the most common 2D pulse sequences that can be employed using a proper choice of parameters to record 2D spectra of paramagnetic molecules (A) NOESY, (B) ROESY, (C) COSY, (D) ISECR COSY, (E) zero-quantum (double quantum) COSY, (F) TOCSY, (G) HMQC, (H) HSQC. Sequences (A), (B) and (F) are also used to obtain EXSY spectra. SL indicates a soft spin-lock sequence, while MLEV17 indicates a train of spin-locking hard pulses that optimizes the development of J/j coupling. In the reverse heteronuclear experiment (G) the upper and lower levels refer to H and heteronucleus, respectively. The phase cycles are not indicated. For clarity of discussion, all initial pulses can be thought to be applied along the y axis, in such a way that the coherence after the first 90° pulse is always along x. ...

Strip plots can only be constructed when the crosspeaks have already been assigned in the 2D HSQC spectrum. In a 15N-labeled protein, sequence-specific assignments come from sequential NOE (a,N, /3,N and N,N) crosspeaks located in the 3D HSQC-NOESY spectrum. The walk through the protein backbone is done in the same way as with unlabeled proteins, except that overlap in NOESY spectra is greatly reduced by spreading the crosspeaks out in the 15N dimension of a 3D spectrum. 

A 2002 review by Reynolds and Enriquez describes the most effective pulse sequences for natural product structure elucidation.86 For natural product chemists, the review recommends HSQC over HMQC, T-ROESY (transverse rotating-frame Overhauser enhancement) in place of NOESY (nuclear Over-hauser enhancement spectroscopy) and CIGAR (constant time inverse-detected gradient accordion rescaled) or constant time HMBC over HMBC. HSQC spectra provide better line shapes than HMQC spectra, but are more demanding on spectrometer hardware. The T-ROESY or transverse ROESY provides better signal to noise for most small molecules compared with a NOESY and limits scalar coupling artefacts. In small-molecule NMR at natural abundance, the 2D HMBC or variants experiment stands out as one of the key NMR experiments for structure elucidation. HMBC spectra provide correlations over multiple bonds and, while this is desirable, it poses the problem of distinguishing between two- and three-bond correlations. 

---