Recent developments in high-resolution solid-state NMR have made it possible to investigate the molecular structure of amorphous or polycrystalline solid samples in a manner similar to that routinely used in solution NMR (e.g. multidimensional correlation spectroscopy). These techniques require high spectral resolution, so that individual spectral lines can be identified and associated with specific nuclei and their local environment in the system under study, and some method of inducing observable interactions between nearby nuclei. Quantitation of these interactions provides information about internuclear distances, and hence molecular conformation. In solid-state NMR, these techniques have traditionally focused on low-g nuclei (e.g. 13C, 15N, 31P) in isotopically-enriched samples. Spectral resolution is increased by spreading the signal over two or more dimensions; nuclear spin connectivities between resolved resonances are then observed as cross-peaks in the multidimensional correlation spectrum. Driving the polarization transfer necessary to create crosspeaks in the spectrum requires interfering with the MAS-induced averaging of the dipolar couplings between low-g nuclei. This is often achieved by imposing a """"""""second motion"""""""" on the nuclear spin system via the application of a rotor-synchronized RF pulse train. Several of these dipolar recoupling pulse sequences have been proposed for inducing either heteronuclear (e.g. REDOR, FDR, RFDRCP) or homonuclear (e.g. Rotational Resonance, RFDR, DRAMA) dipolar interactions. The recently proposed MELODRAMA sequence overcomes much of the chemical shift dependence of the previous techniques, and so efficiently drives polarization transfer among essentially all parts of the spectrum. The technique reintroduces the couplings in the spin-locking interaction frame by rotor-synchronized 90~ phase shifts of an applied spin-locking field. It is an interaction frame analogue of previously proposed DRAMA experiments. Because the spin-locking fields both truncate and play the role of the chemical shift terms in the Hamiltonian, the technique is insensitive to the chemical shifts as well as chemical shift anisotropies of the coupled spins. Through numerical simulations of the one-dimensional magnetization exchange trajectory (equivalent to the mixing time dependence of the cross-peak intensity connecting two coupled spins in 2D spectra), the through-space distance between two spin nuclei can be estimated. Two-dimensional correlation spectra allow simultaneous analysis of the interactions among a greater number of spins. For example, in a 2D spectrum obtained using a relatively short mixing time (&1 ms), the strongest crosspeaks appear only between directly-bonded nuclei. As the mixing time increases (&2 ms), relatively smaller crosspeaks appear due to successive transfers from one neighbor to the next, and follow a """"""""negative-positive-negative"""""""" where the sign of the crosspeak alternates for each successive transfer. Given this property the assignment of the spectrum is straightforward. As the size of the molecule under study increases, the number of spectral lines in NMR spectra will also increase. Since the 13C resolution is only about 1 ppm on a 300MHz spectrometer and the range of 13C chemical shift dispersion is about 200 ppm, a one-dimensional 13C NMR spectrum can only resolve at most 200 different 13C nuclei. This resolution problem is reduced by obtaining higher-dimensional NMR spectra. Spectral overlap in any one dimension is often resolved in second, or higher, dimensions. As an example, we have recently demonstrated three-dimensional 15N-13C-13C chemical shift correlation experiments of uniformly labeled arginine and histidine. These spectra was obtained using relatively short dipolar mixing times, and the spectral assignment based on the crosspeak pattern is again straightforward.
Showing the most recent 10 out of 12 publications
