Nuclear magnetic resonance spectroscopy
Nuclear magnetic resonance (NMR) spectroscopy (for the time being) is the second most common method of structure determination, providing ~7% of all entries in the PDB. It utilises the fact that some atomic nuclei are magnetically active and can emit radio frequency signals when placed in a strong external magnetic field (on the order of 10-20 Tesla, which is almost a million times stronger than the Earth’s magnetic field on the surface). Typical data collection may take 2-3 weeks for a small soluble protein, but can be substantially longer for larger systems.
The measurements in NMR spectroscopy are a number of different complex spectra that report, among other things, on the chemical environment for the magnetically active nuclei (most commonly 1H, 13C and 15N), on chemical bond connections between nuclei, and on short distances between specific atoms. These short distances constitute constraints for molecular dynamics (MD) simulation software, which attempt to satisfy as many of them as possible. The outcome of MD simulation is an ensemble of structures (usually 10-20) which, when combined, best satisfy the experimental data. The whole ensemble is deposited in the PDB.
On Line Database of Ensemble Representatives And DOmains (OLDERADO) provides analysis of clustering and domain composition for NMR structure ensembles.
In order to provide you with a real case scenario, we illustrate an example below (Figure 22):
Figure 22 shows a two-dimensional spectrum, where each peak corresponds to an N-H (amide) group and essentially labels a residue of the protein. The HSQC spectrum is therefore often called the “fingerprint” experiment, as each protein will have a unique pattern of peaks. The horizontal axis gives the chemical shifts of hydrogens, while the vertical – that of nitrogens. Chemical shift is a parameter which is very sensitive to the exact chemical environment of a particular atom, and can therefore act as a “label” or “reporter” for that atom. To acquire this experiment, the protein sample is usually enriched with the 15N isotope of nitrogen, which gives rise to more tractable spectra compared to the more common 14N isotope.