Biophysical Journal Volume 103 July 2012 163–164 163

Comments to the Editor

Electron Microscopy of Biological Specimens in Liquid Water

Mirsaidov et al. (1) have used a microfabricated, thin-window wet cell to examine aqueous biological specimensin the electron microscope. While the thin-window tech-nology used in their work represents a significant innova-tion, there is no reason to expect that the fundamentalconsequences of radiation damage could be changed bythis technology.

As the authors point out, earlier work had led to theconclusion that protein structures are damaged more rapidlyby exposure to the electron beam when in liquid water thanwhen in the frozen-hydrated state. The increase in tolerableradiation exposure that is provided by freezing proteincrystals has subsequently been confirmed in the context ofx-ray crystallography, as is reviewed in Nave and Garman(2) and Garman and Weik (3). Nevertheless, Mirsaidovet al. (1) now claim that the contrary is true for a crystallinebundle of actin filaments. Mirsaidov et al. also suggestthat radiation damage might be less in a liquid environmentthan in the solid (frozen-hydrated) state because radicalscould diffuse away from a protein rather than stay and dofurther damage. This suggestion is contrary to currentthinking, however, in which caging of radiolysis products,in the solid state, is thought to be important in retaininga close-to-original localization of atoms. Furthermore,it is believed that low-temperature immobilization of radi-cals produced in the surrounding solvent reduces theirdiffusion toward the protein, thereby reducing the contribu-tion that they make to secondary (chemical) damage of theprotein.

Perhaps there are other, more plausible alternatives thatcorrectly explain the measurements reported by Mirsaidovet al. If, for example, their specimens had inadvertentlybecome air-dried within the wet cell, the specimens couldhave become embedded in buffer, salt, or other nonvolatilesolutes, resulting in negatively stained specimens such asthose described by Massover and co-worker (4,5). In thiscase, it would not be surprising that the radiation toleranceof their sample would exceed that of frozen-hydratedspecimens. Indeed, the image shown in Fig. S1 C in theSupporting Material looks similar to that of a negativelystained specimen, the objects in Fig. 1, B and F, could beinterpreted as being air-dried samples, and specimen dryingduring irradiation (albeit at much higher electron exposures)

Submitted March 8, 2012, and accepted for publication May 17, 2012.

*Correspondence: rmglaeser@lbl.gov

� 2012 by the Biophysical Society

0006-3495/12/07/0163/2 $2.00

is described in the last paragraph of the SupportingMaterial. It thus would have been informative if the authorshad extended the measurements shown in Fig. 2 to deter-mine whether diffraction spots at a resolution of ~5 nm actu-ally last indefinitely, as they generally do for negativelystained specimens (6). A higher standard of proof is thusneeded to convincingly make the claim that structuraldamage occurs less rapidly in liquid water than in ice.For example, these studies should be extended to includethin catalase crystals in order to make a direct comparisonto earlier work with wet-hydrated and frozen-hydratedsamples.

The authors emphasize the goal of using wet specimens(as opposed to using frozen-hydrated, glucose-embedded,or other solute-embedded specimens) to observe proteindynamics with nanometer resolution. A point that needs tobe considered, however, is that enzymatic function ismuch more sensitive to radiation damage than is overallprotein structure (at the nanometer level of resolution).For example, very few proteins remain enzymatically activeafter receiving radiation doses of 100 Mrad (1 MGy) (7),whereas structural features can tolerate—in frozen-hydratedspecimens—up to 100 times that dose, depending uponthe resolution. Note that 100 Mrad ¼1010 erg depositedper gram, and 1 gray (Gy) ¼ 100 rad. Because the energydeposited per gram of protein is ~60 Mrad (0.6 MGy)when a single frame is recorded with an exposure of 10 elec-tron/nm2 (for 200 keV electrons), a case must be made thatbiological specimens can, at least in theory, remain activeover the course of recording two or more images at nano-meter resolution.

Robert M. Glaeser*

Lawrence Berkeley National Laboratory, University ofCalifornia, Berkeley, California

REFERENCES

1. Mirsaidov, U. M., H. Zheng, ., P. Matsudaira. 2012. Imaging proteinstructure in water at 2.7 nm resolution by transmission electron micros-copy. Biophys. J. 102:L15–L17.

2. Nave, C., and E. F. Garman. 2005. Towards an understanding ofradiation damage in cryocooled macromolecular crystals. J. SynchrotronRadiat. 12:257–260.

3. Garman, E. F., and M. Weik. 2011. Macromolecular crystallographyradiation damage research: what’s new? J. Synchrotron Radiat.18:313–317.

doi: 10.1016/j.bpj.2012.05.042

164 Comments to the Editor

4. Massover, W. H., and P. Marsh. 1997. Unconventional negative stains:Heavy metals are not required for negative staining. Ultramicroscopy.69:139–150.

5. Massover, W. H. 2008. On the experimental use of light metal salts fornegative staining. Microsc. Microanal. 14:126–137.

Biophysical Journal 103(1) 163–164

6. Glaeser, R. M. 1971. Limitations to significant information in biologicalelectron microscopy as a result of radiation damage. J. Ultrastruct. Res.36:466–482.

7. Kempner, E. S., and W. Schlegel. 1979. Size determination of enzymesby radiation inactivation. Anal. Biochem. 92:2–10.