Extending the limits of the selective 1D NOESY experiment with an ...

204 Communication / Journal of Magnetic Resonance 171 (2004) 201–206system, provided that it is sufficiently well-resolved fromthe rest of the peaks. In practice, the isotropic mixingperiod is optimized so as to maximize coherence transferto the target spin for the NOESY mixing period.The 1D 1 H spectrum of lasalocid used in this study isshown in Fig. 2 along with its molecular structure and1 H chemical shift assignments. The 1D 1 H spectrumshows substantial overlap, with only a small numberof well-resolved peaks in the region d > 2.4 ppm. Todemonstrate the efficiency of the new ZQF, Fig. 3 comparesthe selective 1D TOCSY spectra of proton H 14 obtainedusing the pulse sequence of Fig. 1B, with andFig. 3. Selective 1D TOCSY spectra of lasalocid-{H 14 } obtained withthe pulse sequence of Fig. 1B, with (lower panel) and without (lowerpanel) the ZQFs in place. A 20 ms Gaussian 180° pulse was used in theDPFGSE for the selective excitation of H 14 . The isotropic mixingperiod was 25 ms using the DIPSI-2 sequence [30] and a 6.25 kHz B 1field, which generated optimal magnetization transfer to H 15 . Whenthe ZQFs are omitted (upper panel), the spectrum shows significantanti-phase distortions due to ZQ interference. These distortions areessentially absent when the ZQFs are included (lower panel). A 30 or50 ms constant-adiabaticity WURST inversion pulse [24] with abandwidth of 20 or 30kHz was used in each of the ZQFs, respectively.The gradients were optimized accordingly as described by Thrippletonand Keeler [22].without the ZQFs. In the case where no ZQ filtrationis applied (upper panel), the anti-phase contributionsby ZQC to the multiplet patterns are evident. Theseanti-phase distortions are largely absent when two ZQFswith appropriately adjusted durations and gradientstrengths are applied (lower panel). The apparentlyhigher resolution exhibited by H 15 in the absence ofZQ filtration (upper panel) is only a deception attributableto anti-phase components introduced by ZQ interference,as can be confirmed by comparison with the 1D1 H spectrum.The spin system consisting of protons H 14 ,H 15 ,H 16 ,H 17a , and H 17b provides an excellent example for theSTEP–NOESY experiment shown in Fig. 1C. In the1D 1 H spectrum (Fig. 2), with the only exception ofH 14 , all of these peaks show at least some degree of overlap,rendering them unsuitable for direct selective excitation.With the additional spectral editing step providedby the STEP function, it is now straightforward to selectivelyexcite each one of the resonances and observe itsNOE correlations.The doubly selective 1D STEP–NOESY spectrum of{H 15 {H 14 }} obtained with the pulse sequence of Fig. 1Cis shown in Fig. 4A, where the shorthand notation{H b {H a }} denotes H a as the target spin for the STEPfunction, and H b for the NOE mixing period. All ofthe NOE correlations can be assigned in a straightforwardmanner. However, there still exist significant distortionsto the multiplets of those spins scalar coupledto H 15 . This is better exemplified in the NOE buildupof H 17a , shown in Fig. 4B. The anti-phase distortionsare evident for shorter mixing times, while for intermediateto longer NOE mixing times, such distortions areseemingly absent. However, the H 17a peak appears tobe a doublet, while all other available evidences (e.g.,its coupling to both H 17b and H 15 ) suggest that it shouldbe a doublet of a doublet.As pointed out by Shaka and co-workers [26,27],such distortions are not the result of ZQC leakagethrough the ZQFs. Instead, their origin is a phenomenonknown as spin-state mixing, which is encounteredin strongly scalar coupled spin systems. Even thoughthese distortions are generally rather small, they pose asignificant problem in the NOE experiment, since thesignals of interests are usually very weak. These authorsfurther demonstrated that, by subtracting a referencespectrum with ‘‘zero’’ mixing time, these distortionscan be largely removed. This results in a double differencespectrum. The same principle can be applied tothe STEP–NOESY experiment. Fig. 4C shows the doubledifference STEP–NOESY spectrum of {H 15 {H 14 }}.Note that there is a slight reduction in the signal-tonoisecompared to the single difference spectrum ofFig. 4A. However, the correct multiplet patterns arenow restored for all of the spins coupled to H 15 . Forcomparison, the distortion-free NOE buildup of H 17a