Analytical biochemistry

Weighing up protein folding

Article metrics

Labelling molecules by fast oxidation allows mass spectrometry to study protein folding at submillisecond time resolution. The method also brings a wealth of structural information about protein folding within reach.

When it comes to protein-folding studies, mass spectrometry can provide much structural information. But its time resolution has been insufficient to detect the fastest folding events, which occur on the microsecond timescale1. Chen et al.2 now report a solution to this problem in their study of the submillisecond folding of the barstar protein, published in the Journal of the American Chemical Society. They have married two techniques that could potentially reach microsecond resolution: laser temperature jumping (T-jumping), which initiates protein-folding reactions, and fast photochemical oxidation of the protein (FPOP), which allows mass spectrometry to monitor how far folding has progressed. Their technique should enable more structural information to be obtained from studies of protein-folding kinetics — crucial for developing the next generation of computational methods for simulating protein dynamics, and to allow more complex proteins and protein complexes to be studied experimentally.

There is currently great interest in fast-folding proteins. Computational simulations of protein folding have extended into the millisecond timescale, and can thus visualize the movements of proteins that fold in microseconds as they repeatedly fold and unfold3. This opens up the prospect of refining the force fields used in molecular dynamics simulations, so that the simulations can be made valid over even longer timescales. It also offers the hope of being able to correctly predict the biologically active structure of a protein starting from the unfolded state. More experimental data about fast-folding proteins are essential to realize these desirable goals. Fast-folding proteins are also of interest because they are predicted to undergo 'downhill' folding, in which no significant energy barrier is encountered. Further experimental confirmation of downhill folding would provide crucial evidence in support of an important mechanistic model of protein folding — the energy-landscape theory4.

Since the mid-1990s, resistive heating5 (in which a sample is warmed up by passing an electric current through it) and nanosecond laser T-jumps6 have been used in studies to initiate the refolding of proteins from their cold denatured states. Cold denaturation occurs at low temperatures when water molecules bind to hydrophobic amino acids that are normally buried inside proteins. The method generally used to rapidly probe refolding is to measure the fluorescence of tryptophan amino-acid residues in a protein. Tryptophan residues buried inside the protein fluoresce differently from those exposed to water, so that fluoresence serves as a global probe of folding. By contrast, detection techniques such as infrared spectroscopy show promise for acquiring localized structural information from proteins, and are fast enough to be combined with the laser T-jump method for initiating folding reactions.

Chen et al.2 now show that mass spectrometry could join the ranks of fast, structure-sensitive techniques for studying protein folding. The first step of their technique is to add hydrogen peroxide to a cold solution of a denatured protein (Fig. 1). By using a low concentration of peroxide at low temperature, the authors ensure that the protein does not rapidly react with the oxidizing agent. Next, a nanosecond laser T-jump is used to initiate refolding of the protein. After a set time delay, the sample is then irradiated with a nanosecond ultraviolet laser pulse, which breaks up some of the peroxide into hydroxyl radicals. These radicals exist for about a microsecond and efficiently oxidize solvent-exposed protein segments, changing the protein's mass. This is the FPOP step of the process7.

Figure 1: Improving the time resolution of mass spectrometry in protein-folding studies.

Chen et al.2 report a method that allows fast protein folding to be monitored using mass spectrometry. a, A solution of unfolded protein and hydrogen peroxide flows through a capillary tube into the reaction region. b, There, a T-jump — a laser pulse of wavelength 1,900 nanometres — initiates refolding of the protein. c, After a time delay, an ultraviolet laser pulse (wavelength 248 nm) breaks the peroxide into hydroxyl radicals. The radicals rapidly oxidize solvent-exposed amino acids in the protein (red dots become attached to the protein), increasing the protein's mass. The longer the delay, the more the protein is folded (exposing fewer amino acids to solvent), and so less mass is added. d, Finally, the protein sample is flushed out of the capillary, excess peroxide is removed and mass spectrometry is used to determine the extent of protein oxidation (and so of protein folding). (Figure based on an idea by Jiawei Chen.)

An increasingly large fraction of proteins becomes folded as time passes after the T-jump, which means that more of the amino acids become buried within the proteins' interiors. Progressively fewer amino acids are therefore exposed to solvent as folding proceeds, and so less mass is added to the protein by FPOP as the time delay between the T-jump and the FPOP step increases. By performing a series of experiments in which the time delay is varied, and then measuring the mass of the resulting protein samples using mass spectrometry, protein folding can be tracked. In practice, the protein solution flowed through the laser set-up in a capillary tube. Chen et al. collected the oxidized samples, quenched any remaining peroxide using a chemical scavenger, and then performed mass spectrometry on the quenched samples in a separate step.

The authors2 used their T-jump–FPOP (TJFPOP) technique to study the refolding kinetics of denatured barstar as it adopts an intermediate conformation en route to the fully folded protein. The formation of the intermediate takes hundreds of microseconds, and had never before been observed using mass spectrometry techniques. The resulting spectra contained hundreds of peaks, and so the authors analysed only the centroid of the spectra. This simple approach limited the structural information that could be obtained, but the technique offers potentially much better time and structural resolution than was achieved in this proof-of-principle study. The ultimate time resolution of the FPOP step is about 1 microsecond, limited by the hydroxyl radicals' diffusion rate and lifetime. And as the authors point out2, if the oxidized protein samples were enzymatically degraded before mass spectrometry8, then analysis of the resulting fragments could pinpoint where oxidation had actually occurred. The TJFPOP technique could thus follow how amino acids in different parts of a protein become buried with time. Such a technique would be complementary to hydrogen-exchange mass-spectrometry methods, which track the formation of secondary structures in proteins by measuring how easily protons (H+ ions) are exchanged between water and amino acids in the proteins9.

The TJFPOP approach has some drawbacks in its current implementation. The method requires a T-jump for every kinetic data point, whereas fluorescence and infrared-detection methods collect data continuously after a single T-jump. The TJFPOP method thus introduces additional noise compared with the other techniques, and requires larger amounts of sample. Chen and colleagues' approach is also less suitable for studies at high temperatures, because the radical precursor (hydrogen peroxide) would itself react with the protein sample. But precursors more benign than peroxide can be developed.

On the plus side, a strong advantage of TJFPOP is that the whole experiment could easily be automated. But the real key to the utility of TJFPOP will be the development of suitable data-analysis techniques, to deconvolute the statistical mass distributions observed in the spectra and to analyse oxidation patterns of fragments. If these hurdles can be overcome, then the technique could provide truly massive amounts of detail about fast protein folding.


  1. 1

    Kubelka, J., Hofrichter, J. & Eaton, W. A. Curr. Opin. Struct. Biol. 14, 76–88 (2004).

  2. 2

    Chen, J., Rempel, D. L. & Gross, M. L. J. Am. Chem. Soc. 132, 15502–15504 (2010).

  3. 3

    Shaw, D. E. et al. Science 330, 341–346 (2010).

  4. 4

    Gruebele, M. C.R. Biol. 328, 701–712 (2005).

  5. 5

    Nölting, B., Golbik, R. & Fersht, A. R. Proc. Natl Acad. Sci. USA 92, 10668–10672 (1995).

  6. 6

    Ballew, R. M., Sabelko, J. & Gruebele, M. Proc. Natl Acad. Sci. USA 93, 5759–5764 (1996).

  7. 7

    Gau, B. C., Sharp, J. S., Rempel, D. L. & Gross, M. L. Anal. Chem. 81, 6563–6571 (2009).

  8. 8

    de Laureto, P. P., De Filippis, V., Di Bello, M., Zambonin, M. & Fontana, A. Biochemistry 34, 12596–12604 (1995).

  9. 9

    Tsui, V. et al. Protein Sci. 8, 45–49 (1999).

Download references

Author information

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Gruebele, M. Weighing up protein folding. Nature 468, 640–641 (2010) doi:10.1038/468640a

Download citation

Further reading


By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.