Protein analysis by microPIXE

Geoff Grime and Elspeth Garman have pioneered a novel technique for determining the heavy metal constituents found in some proteins which are often hard to determine via other methods.

Getting your sample measured

We can measure all kinds of samples from liquid to crystal, read through our frequently asked questions to find out how we can help you.

Get in contact

If you're interested in getting a sample measured through us then please email Professor Roger Webb.

All samples

All elements in the periodic table heavier than neon (mass 20). For sodium measurements, a special thin filter must be used on the X-ray detector, so make it clear from the beginning if you want quantitative sodium data.

1 part in 106 of the dry weight of the sample. For example this means that 1 zinc atom in a 100kDa protein can be detected.

An accuracy of between 10 and 20 per cent is usually achieved, depending on what elements are present and how much of that element is present.

Only one such sample has been presented so far! As the number of sulphurs is used to get the number of atoms of the element of interest (X) from the ratio of the sulphur/X concentrations measured, it is not possible to obtain accurate quantitation if there is no sulphur in the sample. Absolute concentration measurements are very difficult, and we have not attempted them. For the sample we have already analysed which had no sulphur, we were able to identify the metal present but not give a stoichiometric ratio for it with the protein.

Yes, as the phosphorus acts as the internal standard to give accurate stoichiometric ratios.

Samples are mounted on 2 micron thick mylar or 4 micron thick polypropylene film stretched and stuck across a hole in a standard aluminium holder. Liquid samples are pipetted gently on to the film and allowed to dry naturally. Crystal samples are usually transferred onto the film using a cryo-loop. Some training in preparing the holders is required, as they must be kept clean.

The mylar film has traces of calcium and phosphorus in a known ratio. If the protein has either calcium or phosphorus (e.g. from bound DNA) in it, the background can be deconvoluted. However, this increases the error, and if the accuracy is critical, another backing may be required. Polypropylene is clean of any trace elements, and it is the film we are currently using most of the time.

Each sample takes around 30 to 40 minutes to measure (e.g.1 map and 4 points of around 5 minutes each). The spectra are then analysed with the software programme Dan32, which takes around an hour or two, depending on how difficult it is to fit the Rutherford Backscattered proton spectra to get the matrix composition and thickness.

Preparation and setup: £300 (dimensions of sample holders and list of appropriate backing films will be provided if the user prefers to prepare the samples at home and send them ready for beam irradiation). Ion beam irradiation and analysis: £300 per sample.

There are two publications describing this method of analyzing proteins:

  • Leaving no element of doubt: analysis of proteins using microPIXE. Elspeth Garman. Structure (1999) 7, R291-299
  • Elemental analysis of proteins by microPIXE. Elspeth Garman and Geoff Grime. Progress in Biophysics and Molecular Biology (2005) 89/2, 173-205.

Liquid samples

Only a very low volume is required: 0.3 microlitres is plenty.

This is a really important question. The necessary concentration in mg/ml depends on ratio of the total number of amino acids to the number that contain sulphur. We have established a reliable empirical limit for this and plotted on a graph as a guide to users. It can be found in Progress in Biophysics and Molecular Biology (2005) 89/2, p. 193, Figure 6 (above), and should always be checked before mounting any samples. It is a waste of time and money to attempt to measure samples lying below the line on the scatter plot.

Yes, it is crucial that the buffer contains NO sulphur or chlorine (please see next question).

Common buffers used in protein preparation which contain sulphur and are thus unsuitable are: BES, DDT, HEPES, MES, MOPSO, PIPES. Suitable buffers are: EDTA, TRIS (make sure you don’t use HCl during the preparation, see question below), Tricine, glycine, and TEA. A full list can be found in Table 4 of Garman and Grime, Progress in Biophysics and Molecular Biology, (2005) 89/2, 173-205.

Yes, as this will give an intense chlorine peak next to the sulphur peak. This gives a shoulder under the sulphur peak and thus increases the % error on the measurement. An excellent substitute is KBr or NaBr, and we have been successful in a number of cases using it. KBr gives a convenient internal check of the systematics, since the ratio should be 1:1 from the spectrum analysis.

Crystal samples

It can be really awful: microPIXE measures the material in the beam, it does not care about diffraction quality, so as long as there is some identifiable object to analyse, it does not matter whether or not it is highly ordered.

Anything over 50 microns is fine. The smallest we have ever measured was 10 microns square and about 2 microns thick (it turned out to be zinc chloride…).

This depends on whether the crystallisation buffer contains sulphur. If is does, the crystal should be washed 5 or 6 times in milliQ (NOT tap water which adds chlorine to the buffer) by means of a cryoloop. It does not matter if the crystal disintegrates, as long as some of it gets onto the film. If the crystal was washed, it is safest to carry out PIXE measurements on both a washed and an unwashed crystal.

Contact us

Find us


Ion Beam Centre
Advanced Technology Institute
University of Surrey