15N,13C,2H-labelling is commonly referred to as triple labelling. The protein is produced by expression from bacteria which are grown on minimal medium supplemented with 15NH4Cl and 13C-glucose and using D2O instead of H2O. This will result in about 70-80% deuteration of the side-chains, as there is a certain amount of contaminating 1H present from the glucose. Higher levels of deuteration of around 95% can be achieved if 13C,2H-glucose is used. However, this is more expensive and in many cases the cheaper version is sufficient. Note that the NH groups are exchangeable. This means that they will back-exchange to 1H when the protein is purified in normal aqueous solution. In this way, many of the normal NH-based experiments can be carried out on triple-labelled protein.
Deuteration is mainly used for large proteins. On account of their slower tumbling, the relaxation properties become less favourable, and the spectral quality deteriorates – eventually the peaks become so broad, that no spectrum is detectable. By deuterating the protein and thus removing most 1H atoms (protons), the relaxation properties are improved again. Backbone and Cβ assignment of triply labelled proteins is possible using out-and-back versions of the HNCACB and HN(CO)CACB experiments. Note that the Cα and Cβ chemical shifts will be shifted by up to half a ppm or more from their values in 15N,13C-labelled protein due to the deuterium isotope effect.
More recently Otten et al. have shown using some new experiments that you can also use 15N,13C,2H-labelled protein made from protonated glucose to assign the methyl groups of Ala, Leu, Val, Ile, Met and Ala and then use these sites to help probe the structure and dynamics of the protein.
D.M. LeMaster (1990) Quart. Rev. Biophys. 23 133-174. (Link to Article)
R. Otten, B. Chou, K.D. Krewulak, H.J. Vogel and F.A.A. Mulder (2010) J. Am. Chem. Soc. 132 2952-2960. (Link to Article)