Nuclear magnetic resonance (NMR) spectroscopy 8/1/03. By Richard Twyman Nuclear magnetic resonance is used to determine the structure of molecules including small, soluble proteins.

Nuclear magnetic resonance describes the way some atomic nuclei 'wobble' in a magnetic field when exposed to radio waves. The frequency of radio waves required to cause resonance in different nuclei depends on the way atoms are arranged in a molecule. The analysis of a molecule's NMR spectrum therefore allows a structural model to be built. NMR spectroscopy is used for the structural analysis of proteins and other molecules.

How does it work?

The nuclei of some atoms behave like tiny magnets. Any small magnet will align itself in an external magnetic field. For example, a compass needle will align with the Earth's magnetic field and point North. Atomic nuclei can behave in the same way.

If you move a compass needle with your finger and force it to point South, it will be unstable. It will soon reorient itself in the magnetic field and point North once again. Similarly, atomic nuclei can be reoriented in a magnetic field by supplying energy in the form of radio waves, but will soon return to the most stable orientation.

At just the right frequency of radio waves, a nucleus will flip backwards and forwards between the stable and unstable states. This wobbling effect is called nuclear magnetic resonance (NMR).

The radio frequency required to achieve magnetic resonance depends on the type of atom, what other atoms it is connected to and what other atoms are nearby. By analysing the NMR spectrum of a molecule (a graph showing NMR in a molecule over a range of radio wavelengths) it is therefore possible to deduce the three-dimensional structure of that molecule.

How is it used?

As with X-ray crystallography, one of the th main applications of NMR spectroscopy in biology is to determine the structure of proteins and other molecules. NMR is particularly suitable for the analysis of small, soluble proteins. Structures solved by NMR are generally presented as an ensemble of models, all of which satisfy the available data.