tte advent of two-dimensional Fourier transform methods triggered a veritable revolution in the practice of NMR spectroscopy (Jeener 1971; Aue et al. 1976). It presented the chemist with a clear pictorial representation of interactions between different chemical sites, displaying them in the form of "correlation peaks" on a contour map in two frequency dimensions. For the first time the structural chemist could use NMR spectra directly, without the recourse to tedious assignment techniques, such as double resonance, tte correlations between interacting chemical sites were plain to see on the new charts, almost as if touched by a magic wand. Chemistryjournals were soon full of two-dimensional spectra.

It was not long before the two-dimensional concept was extended to additional frequency dimensions. By concatenating two or more "evolution" intervals during which the nuclear spins moved according to different rules, and performing a Fourier transformation with respect to each evolution period in turn, the new technique generates a multidimensional spectrum, tte problem of display is solved by recording appropriate plane sections through the frequency-domain data matrix, tte availability of biochemical samples isotopically labelled in carbon-13 and nitro-gen-15 has initiated averitable explosion of applications of multidimensional NMR to the study of proteins, aided by the improved sensitivity and frequency dispersion of modern NMR spectrometers.

However there is a fly in this particular ointment - the experiments tend to be of long duration, often several days at a time, tte problem is that each and every evolution dimension has to be sampled systematically and independently, ttis puts an unwelcome brake on projects of really high dimensionality, or on experiments to monitor the time-dependence of the interactions. NMR spectroscopists are forced to make compromises with respect to resolution, so that fewer increments are used in the evolution dimensions, sometimes extending the time-domain data artificially by means of "linear prediction" algorithms, tte situation is clearly unsatisfactory; multidimensional NMR ties up an expensive spectrometer for long periods and reduces the throughput of samples.

Desperate measures are called for. ttis review outlines two radical new approaches to fast multidimensional NMR. But first we must accept a fundamental limitation on all experiments carried out in a very short time - sensitivity inevitably suffers because less energy is transferred from the spin system to the spectrometer receiver. Fortunately, in many high-field studies, particularly if a cryogenic receiver

Springer Series in Biophysics J.L.R. Arrondo and A. Alonso Advanced Techniques in Biophysics © Springer-Verlag Berlin Heidelberg 2006

coil is used, the signal-to-noise ratio it not a limiting factor. Speed is far more important. Two radically different approaches to the speed question are described in the following.

Was this article helpful?

0 0

Post a comment