The “superoxide radical” (O2•-) is a cytotoxic by-product of oxygen metabolism. In humans, about 2% of the oxygen taken in during breathing is transformed into O2•- instead of water. The amount of O2•- produced is increased in patients affected by neurodegenerative diseases such as Alzheimer’s, resulting in a worsening of these illnesses. Scientists are therefore looking for drugs to eliminate the superoxide radical.

Superoxide reductase (SOR) is an iron containing metalloenzyme that eliminates the superoxide radical in some microaerophilic bacteria and archaea. Although it differs from its counterpart in humans, superoxide dismutase (SOD), SOR carries out a simpler reaction than SOD. Understanding the chemical tricks used by this enzyme is not only of fundamental interest in the field of iron biochemistry, but could also open new avenues for developing drugs.

Kinetic crystallography [1] was used to study the catalytic cycle of SOR, in collaboration with the team of V. Nivière (CEA/iRTSV, Grenoble, France). A stroke of luck allowed us to produce a movie of SOR in action (Figure 72). In a single crystal of a SOR mutant that specifically slows down catalysis, three intermediate states were trapped by freezing the sample at a certain time delay after the reaction was triggered. The existence of the three states in one crystal can be explained by the presence of nominally identical active sites in the crystal asymmetric unit that are surrounded by different packing environments, creating slight differences between them.

Fig. 72: Kinetic crystallography of superoxide reductase allowed a tentative film of the reaction pathway to be produced.


Before attempting a mechanistic interpretation of the trapped states, it was necessary to evaluate their biological relevance. With this in mind, we developed the technique of in crystallo Raman spectroscopy at the Cryobench laboratory [2,3]. This laboratory, developed in a close collaboration between the ESRF and the Institut de Biologie Structurale, is dedicated to complementing X-ray crystallography with in crystallo UV-visible spectroscopy [4]. Applying complementary techniques to the same sample appears increasingly important in understanding the relationship between structure, dynamics and function in biological macromolecules [5]. The central instrument of the Cryobench is a microspectrophotometer that allows the analysis of nano-volume samples (in the liquid or crystalline state) in conditions identical to those used on X-ray crystallography beamlines (Figure 73). In addition to absorption and fluorescence, we implemented Raman spectroscopy to identify specific chemical bonds in protein crystals. Raman spectra collected in crystallo are often instrumental in identifying intermediate states building up in the crystal, and in making sure that these states are not altered by X-rays during collection of diffraction data.

Fig. 73: Intermediate states can be trapped in protein crystals using spectroscopic monitoring at the Cryobench.


In the case of SOR, the Raman spectra collected on crystals provided evidence that the trapped states corresponded to iron-peroxide species also seen in solution, and were not altered by a moderate exposure to synchrotron X-rays. Thus we could confidently use the ~2.0 Å structures of the three intermediates to propose a mechanism for SOR function.

Several hypotheses to explain SOR activity have been previously inferred from biochemical studies. The binding mode of O2•-, the role of key amino-acid residues, and the direct participation of a water molecule in catalysis were anticipated. The direct visualisation of these events now allows an understanding of how they fit together. We found that a key lysine residue moves around on the surface of the enzyme in order to grab a water molecule from the solvent. It then brings the water molecule into the enzyme active site, where it can donate a proton to the substrate bound in an “end-on” fashion (Figure 72). Puzzling questions remain, however, which will be addressed in future experiments. For example, when this lysine is mutated, the enzyme still functions, although at a slower rate: thus the mechanism suggested here might correspond to only one of several possible reaction pathways.

In conclusion, in crystallo Raman spectroscopy provides a complementary tool to X-ray crystallography, especially useful for researchers pursuing structural studies with mechanistic perspectives. The technique was used successfully to identify key iron-peroxide intermediates in the catalytic cycle of the non-heme enzyme superoxide reductase.



[1] D. Bourgeois and A. Royant, Curr. Opin. Struc. Biol. 15, 1 (2005).
[2] P.R Carey and J. Dong, Biochemistry 43, 8885 (2004).
[3] P. Carpentier, A. Royant, J. Ohana and D. Bourgeois, J Appl Cryst 40, 1113 (2007).
[5] T. De la Mora-Rey and C. Wilmot, Curr. Opin. Struc. Biol. 17, 580 (2007).

Principal publication and authors

G. Katona (a), P. Carpentier (a), V. Niviere (b), P. Amara (a), V. Adam (c), J. Ohana (a), N. Tsanov (a) and D. Bourgeois (a, c), Science 316, 449-53 (2007).
(a) Institut de Biologie Structurale, Grenoble (France)
(b) iRTSV, CEA (France)
(c) ESRF