Proteins and Model Systems: Spectral Analyses

The first NMR spectra of proteins, obtained in the 1950s and 1960s, focused on the 1H nucleus because of its spin I = 1/2 nature, high natural abundance, and sensitivity.1 – 6 However, because of the small range of chemical shifts (due to folding) and the low spectral resolution available on 60-, 100-, and even 220-MHz instruments, there was relatively little spectral resolution available, and only a handful of individual proton resonances could be assigned, precluding use of computational methods to predict spectra from the structure. This problem of a small 1H chemical shift range was well known to organic chemists, and several groups, in particular those of Roberts,7 Sternlicht8, and Grant9, had embarked on the observation of 13C nuclei that, while having only a 1% natural abundance and low sensitivity, were known to have a ∼10× larger chemical shift range than 1H, opening up the intriguing prospect of 13C NMR of macromolecules, such as proteins. To get improved sensitivity (or at least, signal-to-noise ratios), these workers developed much larger (10–12 mm) sample probes (5 mm was, and still is, the standard for most solution samples); in addition, the advent of 1H-decoupling10 improved spectral resolution considerably. However, it took Ernst and Anderson’s development of the Fourier transform method11 to provide genuine sensitivity enhancements, opening the way to protein structure studies.