NMR Method Development

NMR of Larger Proteins

Faster relaxation of magnetization in larger proteins results in poor sensitivity and many broad overlapping peaks, degrading resolution even in 3D/4D NMR spectra. The cause is strong, long-range interactions between large magnetic dipoles in hydrogen nuclei. A solution is to replace sidechain hydrogen nuclei with deuterium, which has a much smaller nuclear magnetic dipole. However, too much deuteration will remove all the proton NMR signals, which provide most of the structural information, and an optimum level has to be found.

In the past, we have investigated different approaches, involving either random fractional deuteration to different levels or labelling particular residue types. In our studies of the p19INK4d protein, we made protein samples that were either completely deuterated or which had deuterated sidechains in all residues except those that are hydrophobic and expected to be buried in the core. These enabled the gathering of spectra of improved sensitivity and resolution and the rapid determination of a rough stucture, showing the global fold. Based on this structure, we were able to obtain the high resolution structure without the need to assign the 3D/4D NOESY spectra, which had heavily overlapping signals.

For more details see Smith et al., J. Biomol. NMR, 1996, 8, 360-68; Nietlispach et al., J. Am. Chem. Soc., 1996, 118, 407-15; Luh et al., Nature, 1997, 389, 999-1003.

In our current work, we are collaborating with other groups to develop new computational tools for structure determination of proteins based on NMR data recorded from such deuterated proteins. In particular, we are developing a new version of the ANSRS/CLOUDS approach developed by Per Kraulis and Miguel Llinas’s group. The CcpNMR version of ANSRS/CLOUDS allows ab-initio structure determination, without the assignment of NMR spectra, starting from a structure/model of a homologue of the protein of interest – the equivalent of molecular replacement in X-ray crystallography. We are also investigating the feasibility of using a Bayesian approach to structure determination called Inferential Structure Determination (ISD), which was developed in Michael Nilges group, to combine information from chemical shifts with other structural restraints.