Chapter 15

Protein NMR

Probably the most fascinating advances of NMR spectroscopy during the last decades have been made in the field of structural biology. It is now possible to assign the signals of nearly every single hydrogen, carbon and nitrogen nucleus in a protein with a typical mass of about 15–25 kDa containing hundreds of atoms. After the signal assignment it is possible to calculate a detailed structure of the protein using mainly NOE constraints, but also several other NMR-based structural parameters, e.g. ³J couplings. At the time of writing (2003) a new method is emerging, using residual dipolar couplings which are measured with the help of molecular alignment agents.

In a book entitled "Basic NMR Experiments" one might find these methods at first glance inappropriate. Certainly, this new chapter does not aim to address the experts in this field. However, it tries to introduce the novice into this complicated but exciting world of NMR spectroscopy which, as we had to find out, is seldom described in sufficient experimental detail. We have therefore assembled a series of 20 experiments which have been chosen so as to be able to record the standard data sets that are needed to obtain a full signal assignment and structure calculation for a medium-sized protein. Some methods are already older than 15 years, some are more recent. All experiments, however, are given in a current version, e.g. they are gradient-selected and some contain recent features such as frequency-swept shaped pulses.

All experiments are demonstrated with the same fully ¹³C- and ¹⁵N-labeled protein, human ubiquitin, which is commercially available and quite stable. Our sample has now lasted over four years in a sealed tube. The price of this sample is very high. However, considering the cost of a three-channel protein NMR spectrometer, these costs are indeed quite negligible. We strongly advise obtaining such a reference sample before starting work in this field of NMR. The possibilities of making errors in performing the pulse sequences described in this chapter are abundant, and the spectra obtained from proteins are usually very complex, so that these errors might be detected too late. The results for ubiquitin are well documented in the literature [6] and serve as a control for the experiments.

The 3D sequences shown in this chapter have almost all been performed on the same instrument, a Bruker DRX-600, vintage 1998, using a multinuclear inverse probe-head with an additional fixed coil for ¹³C and a z-gradient facility; thus the experimental set-up was always very similar and is given below. Of course, there are many other possible choices of hardware and software for this purpose, and probably no "professional" protein laboratory will perform the experiments just as described here. For example, the pulses in different frequency regions for ¹³C can be generated by two different ¹³C frequency sources and amplifiers or, as given in this chapter, by using phase-modulated shaped pulses via the same amplifier. Comparing pulse sequences from different expert groups demonstrates that everybody does it quite differ-