The British Biophysical Society

Neutron scattering

Robert Gilbert

 

Neutrons are most aptly applied to the determination of hydration patterns and the structure of macromolecules and their complexes, and their interaction and dynamic activity.

Neutron crystal structure of lysozyme, revealing its highly organized hydration layers which possess unexpected arrangments of water molecules. Oxygen atoms are shown in red, deuterium (the heavy isotope of hydrogen) atoms are shown in yellow, and the backbone of the protein itself as a cyan cartoon.

 

Neutron scattering is not a mainstream technique in biomedical research. The interaction of neutrons with nuclei is very weak and consequently the length of time needed to collect a large amount of high resolution data using them is very long (for example, weeks of measuring time for neutron crystallography). Alternatively, it is often found that the amount of sample required is very large, or that the amount of information (e.g. spatial resolution) provided is very small. The interaction of neutrons with matter does, however, have special anomalous characteristics that makes them in some cases the ideal, if not only, way to carry out a particular experiment. Neutron scattering is never going to fill the niches occupied by X-ray crystallography, the pre-eminent structural technique, NMR, or electron microscopy - and I write as someone who spends his days determining structures by electron microscopy in a laboratory famed for its X-ray crystallography - but that doesn't change the fact they have their own special light to shed in the scientist's attempt to reconstruct nature.

Measuring scatter in different ways

The scattering of neutrons from the nuclei of molecules can be measured in a number of experimental geometries, for a range of different sample types, to yield very different kinds of information. Scattering neutrons from a solution of protein gives a scattering profile which is related to a spherically-averaged representation of the distribution of nuclear mass in the protein; scattering neutrons from a protein crystal gives a diffraction pattern which is related to the nuclear periodicities found within the crystal. In the first case it is possible to obtain a three-dimensional model of the protein, but by indirect methods in which an ex nihilo model is modified until it reproduces the scattering curve to some extent. In the second case a full atomic structure can be obtained, by methods similar to those employed in X-ray crystallography. In both these cases, the neutrons are treated as being scattered elastically i.e. there is no change in energy in the scattering event. Other experimental approaches which work like this are neutron lamellar diffraction, reflectivity, fibre diffraction, and small-molecule crystallography, which are again approaches which are used with X-rays. Other ways of obtaining information are from the ways neutrons change energy when they are scattered; this inelastic scattering is used to obtain dynamical information about the structures of samples.

What are neutrons? Where can I get them? Why use them?

Neutrons are fundamental particles which combine with protons to form atomic nuclei. They are generally made for experiments in one of two ways: either by firing a pulse of protons into a suitable target substance from which they are dislodged (a spallation source e.g. ISIS), or by collecting them from decay of a radioactive substance (a nuclear reactor source e.g. ILL). The interaction of neutrons with nuclei is in general affected by two factors. These are the isotopic scattering length b and the spin-dependent (or polarization-dependent) scattering length, B. The B component is averaged out in the absence of spin polarization and therefore does not contribute to the measured signal (there are a few significant biological experiments which have made use of spin-polarization). For ²H B = 5.7 fm while for ¹H, B = 58.24 fm; hence the scattering signals from the two isotopes of hydrogen are well differentiated. This allows the scattering from a hydrogenated component of a scattering species to be deconvoluted from the deuterated whole, permitting the measurement of scattering from components representing ~1% by mass of a particle.

With b the scattering lengths of ²H and ¹H are again very different, but in this instance they are reversed so that neutrons are scattered much more strongly from ²H than from ¹H. But the main reason why neutrons can be useful for the study of biological molecules is that the scattering lengths of ²H and ¹H are also opposite in sign (6.671 fm and -3.742 fm respectively) and between them almost encompass the range observed for all atoms. This means that by modulation of the ratio of ²H:¹H in a sample, the scattering from different kinds of macromolecule can be "matched out", i.e. their contrast is matched by the background scatter and so they do not contribute to the measured signal. For example, nucleic acid is matched out at ~65% ²H₂O, protein at ~42% ²H₂O, and lipids at ~13% ²H₂O. Additionally, the match point of a molecule or components of it may be manipulated by selective deuteration.

Neutron scattering techniques can be used to investigate structures at a range of resolutions. Small-angle neutron scattering (SANS) can provide information in the range 10 to 1000 Å; specular neutron reflection and off-specular diffraction yield data in the range ~5 to 1000 Å; lamellar diffraction can yield data at comparable resolutions to X-ray crystallograpy (albeit anisotropically), as of course can neutron single crystal diffraction. The resolution range for data obtained from neutron fibre diffraction is the same as from X-ray fibre diffraction.

Techniques

Neutron crystallography was a key technique in the early days of structural virology, where it was used to demonstrate the layered ordering of viral genomes under their capsids. Another area where it has made a significant impact is in providing low-resolution structures of detergent and lipid in crystals of membrane proteins. A final area where it has provided unexpected information is in showing the complex structures formed by water molecules at protein-solvent interfaces, where water is packed ~10% more densely than in solution, and forms strange cage-like hydrogen-bonded structures around bulky hydrophobic residues.

Neutron fibre diffraction has been used extensively to study the hydration of nucleic acids and the structures of filamentous viruses.

Small-angle neutron scattering and associated modeling has been used in determining dimensions and structures of a wide range of biological systems, including isolated proteins and their complexes, the ribosome, and lipid dispersions. It was used to show that hydration water at solvent:protein interfaces is densely packed.

Specular reflection and off-specular diffraction is particularly appropriate as a means of measuring the structures of membranes, the presence of periodic structures within them such as non-lamellar lipid phases or rafts, and the interaction between proteins and membranes. This last possiblity has barely been addressed but has the potential to combine the affinity measurements of a technique like SPR (BIAcore) with the structural measurements of electron microscopy.

 

Lamellar diffraction has proved useful in the structural and functional characterization of bacteriorhodopsin, which forms natural 2-dimensional crystals. For example, projection structures of different functional states of the molecule have been determined to show both structural rearrangements and changes in the hydration structure of the protein. In another case, the critical positioning of a single kind of glycolipid that mediates interactions within the bacteriorhodopsin crystal was determined.

Incoherent neutron scattering provides insights into the dynamical behaviour of systems, such as the pumping of protons by bacteriorhodopsin and the dynamic role of water in protein structure. Although a large amount of sample is required in experiments of this kind, the information obtained represents an insight into the functioning of molecular systems that could not be obtained in any other way.

Powder and single crystal diffraction of small molecules has been critical in revealing very high resolution information about drugs such as anisotropic temperature factors for hydrogen atoms.

Robert Gilbert is a Royal Society University Research Fellow and Group Leader in the Division of Structural Biology at the University of Oxford.