Live Cell Electron Microscopy Is Probably Impossible Niels de Jonge*,†,‡ and Diana B. Peckys§ †
INM−Leibniz Institute for New Materials and ‡Department of Physics, Saarland University, D-66123 Saarbrücken, Germany Department of Biophysics, Saarland University, D-66421 Homburg/Saar, Germany
§
ABSTRACT: Electron microscopy of biological cells in liquid provides unique nanoscale information. A highly attractive idea is the capability to also study physiological processes of live cells with electron microscopy. However, this idea seems unrealistic because the minimal needed electron dose to obtain contrast is already many orders of magnitude above the lethal dose known to cause reproductive-cell death. We show here that claims of electron microscopy of viable cells in recent reports are based on a questionable interpretation of the used fluorescence live/dead assay. A practical alternative to study biological processes is correlative light and electron microscopy.
W
ith the availability of systems for electron microscopic studies of biological cells in their native liquid environment with nanoscale spatial resolution,1−5 the question has arisen if it is possible to image the ultrastructure of live cells or even certain processes of a life cell using these methods. Unfortunately, these aims seem unrealistic for reasons of radiation damage.3,6−8 The exposure to the electron beam leads to structural damage mostly due to the generation of reactive species, such as solvated electrons and H+ and OH− radicals, and the direct breakage of molecular bonds; cell death already occurs at a dose many orders of magnitude below even the lowest doses used in electron microscopy.6,8 Nonetheless, a recent paper by Kennedy et al.9 in ACS Nano claims that scanning transmission electron microscopy (STEM) of the physiology of live bacterial cells is possible using a radiation dose of 0.3 e−/Å2 per image, at least 3 orders of magnitude above the lethal dose known to cause reproductive-cell death of the imaged specimen, e.g., Escherichia coli bacteria or Bacillus subtillus spores.6,8 For reasons described here, the claim of viability in the paper lacks experimental evidence and may represent an incorrect interpretation of the experimental data. Another group reported imaging of live cells with transmission electron microscopy (TEM) earlier.10 It is important to critically discuss this topic to avoid false expectations toward the in itself promising capabilities of electron microscopy in liquid.11,12 An experimental configuration for STEM of cells in liquid is shown in Figure 1. Cells fully embedded in liquid are enclosed by thin electron-transparent windows protecting them from the vacuum in the electron microscope. The contrast in the STEM images originates from elastic scattering of the electrons and depends on the differences in atomic number (Z) and density between the object under observation and the liquid.3 Nanoscale information about eukaryotic cells with thicknesses of ∼5 μm can be obtained by using specific protein labels consisting of nanoparticles with high-Z cores.2,4 Yeast cells are thin enough so that it is possible to image the pristine structure with an order of magnitude better resolution13 than what is © 2016 American Chemical Society
Figure 1. Schematic of liquid-phase scanning transmission electron microscopy (STEM) of cells. Cells surrounded by liquid are contained between two silicon nitride (SiN) windows, transparent for electrons (and photons). The focused electron beam of the STEM scans a defined interior region of the specimen, while the scattered and transmitted electrons are used for dark-field detection.
achievable with conventional light microscopy, that is to say, without the help of fluorescent markers and super-resolution techniques.14 Bacteria are even thinner, so that nanometer resolution can readily be observed, in particular, when the cellular ultrastructure contains or is stained with a metal.9,15 Important biological information can be gained with these methods complementing high-resolution cryo-TEM studies mostly involving laboriously prepared thin sections.16 In order to understand what evidence is missing from the report claiming live-cell TEM,9 one should consider the characteristics that define life. One of the most important characteristics of a live cell is its capability to reproduce. A study claiming the observation of the physiology of a life cell should contain a positive test on reproduction or a test on any of the other well-known characteristics of life and could, for example, consist of the incubation of the cells in growth medium af ter Received: April 27, 2016 Published: October 25, 2016 9061
DOI: 10.1021/acsnano.6b02809 ACS Nano 2016, 10, 9061−9063
Letter to the Editor
www.acsnano.org
ACS Nano
Letter to the Editor
also within many intact vacuoles visible as distinct red fluorescent spots indicating the viability of the cells.17 For comparison, a fluorescence image recorded after electron beam irradiation is shown in Figure 2b. The specimen was transferred to the electron microscope after the recording of Figure 2a, and a low dose of the STEM image was recorded at the same location (Figure 2c). The electron dose amounted to only 0.2 e−/Å2, below the lowest dose used in the live-cell paper.9 Several features of the internal structure of the yeast are visible, such as the cell wall, secondary septa between two cells in the state of division, and darker areas indicating a lower density of material than the liquid. A second STEM image was also recorded at this location (Figure 2d). However, despite the low dose, already many structural changes are visible. The periplasm has widened for most of the cells (examples are indicated with dashed lines), implying that the cell interiors enclosed by the plasma membrane have shrunken. One cell wall appears broken at the arrow. Also, the cell with the broken cell wall reduced in size from 4.0 to 3.6 μm, measured at the location indicated. The changes caused by electron bean irradiation are also apparent from fluorescence microscopy (Figure 2b). The vacuole signature is lost, so that the cells are now filled with a homogeneous fluorescent signal. The color has changed, and some of the dye is leaking (for example, at the arrow) as a result of broken cell walls. These are the known characteristics for dead yeast cells observed using the FUN-1 dye.17 Since both STEM images already showed major structural changes, it can not be ruled out that the cells already died during the recording of the first image. Figure 2c thus represents a cell alive at the onset of STEM, but this should not be confused with live-cell imaging. The above experiment shows that the visibility of a fluorescence signal within cells using a live/dead assay does not prove at all that the cells are alive but may, in fact, occur also for dead cells for a certain period of time until full disintegration of the cell membranes occurs. One should thus critically consider the origin of the movements or changes of structures as seen in the published image series.9 Claiming electron microscopy of physiological processes from live organisms requires unambiguous evidence that the organisms were not already lethally damaged while acquiring the first image of the series. The observed dynamical effects9 could also have been caused by either or a combination of structural rearrangement after radiation damage,6 movements of the liquid, often seen during electron microscopy, for example, when gas bubbles are created,20 or simply by Brownian motion. Unless detection schemes are developed allowing at least 3 orders of magnitude lower electron doses, the ability of electron microscopy to study life cell processes at the nanoscale, although very attractive, probably remains wishful thinking. Of course, the possibility should not be ruled out altogether, but reports on this topic should include extensive control experiments. It is feasible to study cellular function by an alternative approach combining light and electron microscopy, whereby a process is followed with the light microscope and “snap shot” electron microscopy is then used to reveal molecular configurations.2,21,22 Chemical fixation, commonly used in light microscopy,23 can be applied between the two microscopy modalities since a cell is dead after electron microscopy anyway, and the practically achievable spatial resolution largely increases due to the higher stability of the structure.24
the electron microscopy experiment to observe both cell growth and division. Such test, however, is lacking altogether from the papers in question. Obviously a lipid vesicle containing a fluorescent dye is not alive. Yet, the authors claimed the absence of dye leaking from their objects under observation as key evidence of viability.9,10 The involved fluorescence live/dead assay is used in certain biological experiments as indication of an intact cell membrane17 but does not necessarily provide the evidence of viability and requires careful interpretation. For example, the cell walls may prevent leaking for a period of time after cell death. In fact, the cells were incubated with a highly toxic chemical staining agent (uranyl acetate) to improve contrast in one of the reports,9 which is known to fix and thus stabilize the lipid membranes in cells.18 Also the toxicity-related control experiments should be questioned, but that is beyond the scope of this report. What may happen with cells containing fluorescent dye upon electron beam irradiation is described in the following. In order to evaluate the effect of electron beam radiation on cells, a sample of live Schizosaccharomyces pombe yeast cells was imaged with liquid-phase STEM. In contrast to bacteria, yeast cells contain cell organelles that are visible with both light and electron microscopy, so that the effect of radiation can be more easily studied. The yeast cells were grown and harvested as described elsewhere.13 The fluorescent dye FUN-1 (Invitrogen, Carlsbad, CA) was used as the live/dead assay.17 A droplet containing live S. pombe cells was placed in a liquid enclosure for electron microscopy consisting of two microchips with electron-transparent windows.19 The loading procedure was completed within 1 min and was immediately followed by examination with fluorescence microscopy. The fluorescence image of Figure 2a shows that the dye is located in the cells and
Figure 2. Investigation of the viability of S. pombe cells before and after liquid-phase STEM. (a) Fluorescence image recorded prior to STEM showing red FUN-1 dye accumulated in vesicles, indicating that the cells were alive. (b) Fluorescence image recorded 5 min after electron beam exposure. The cells were not viable after the recording of two STEM images as can be seen from the fluorescence signature. (c) First liquid-phase STEM image recorded at this location, representing the pristine ultrastructure. (d) Second liquid STEM image recorded 1 min after the recording of (c). The gradient in contrast over the STEM images originates from the variation in liquid thickness as a function of position caused by bulging of the SiN windows. 9062
DOI: 10.1021/acsnano.6b02809 ACS Nano 2016, 10, 9061−9063
ACS Nano
Letter to the Editor
(6) Reimer, L.; Kohl, H. Transmission Electron Microscopy: Physics of Image Formation; Springer: New York, 2008. (7) Matricardi, V. R.; Moretz, R. C.; Parsons, D. F. Electron Diffraction of Wet Proteins: Catalase. Science 1972, 177, 268−270. (8) Isaacson, M. S. Specimen Damage in the Electron Microscope. In Principles and Techniques of Electron Microscopy; Hayat, M. A., Ed.; Van Nostrand: New York, 1977; pp 1−78. (9) Kennedy, E.; Nelson, E. M.; Tanaka, T.; Damiano, J.; Timp, G. Live Bacterial Physiology Visualized with 5 Nm Resolution Using Scanning Transmission Electron Microscopy. ACS Nano 2016, 10, 2669−2677. (10) Liu, K. L.; Wu, C. C.; Huang, Y. J.; Peng, H. L.; Chang, H. Y.; Chang, P.; Hsu, L.; Yew, T. R. Novel Microchip for in situ Tem Imaging of Living Organisms and Bio-Reactions in Aqueous Conditions. Lab Chip 2008, 8, 1915−1921. (11) de Jonge, N.; Ross, F. M. Electron Microscopy of Specimens in Liquid. Nat. Nanotechnol. 2011, 6, 695−704. (12) Ross, F. M. Opportunities and Challenges in Liquid Cell Electron Microscopy. Science 2015, 350, aaa9886. (13) Peckys, D. B.; Mazur, P.; Gould, K. L.; de Jonge, N. Fully Hydrated Yeast Cells Imaged with Electron Microscopy. Biophys. J. 2011, 100, 2522−2529. (14) Lippincott-Schwartz, J.; Manley, S. Putting Super-Resolution Fluorescence Microscopy to Work. Nat. Methods 2009, 6, 21−23. (15) Woehl, T. J.; Kashyap, S.; Firlar, E.; Perez-Gonzalez, T.; Faivre, D.; Trubitsyn, D.; Bazylinski, D. A.; Prozorov, T. Correlative Electron and Fluorescence Microscopy of Magnetotactic Bacteria in Liquid: Toward in vivo Imaging. Sci. Rep. 2014, 4, 6854−6851−6858. (16) Kourkoutis, L. F.; Plitzko, J. M.; Baumeister, W. Electron Microscopy of Biological Materials at the Nanometer Scale. Annu. Rev. Mater. Res. 2012, 42, 33−58. (17) Millard, P. J.; Roth, B. L.; Thi, H. P.; Yue, S. T.; Haugland, R. P. Development of the Fun-1 Family of Fluorescent Probes for Vacuole Labeling and Viability Testing of Yeasts. Appl. Environ. Microbiol. 1997, 63, 2897−2905. (18) Glauert, A. M.; Lewis, P. R. Biological Specimen Preparation for Transmission Electron Microscopy; Portland Press: London, 1998. (19) Ring, E. A.; de Jonge, N. Microfluidic System for Transmission Electron Microscopy. Microsc. Microanal. 2010, 16, 622−629. (20) Woehl, T. J.; Jungjohann, K. L.; Evans, J. E.; Arslan, I.; Ristenpart, W. D.; Browning, N. D. Experimental Procedures to Mitigate Electron Beam Induced Artifacts During in situ Fluid Imaging of Nanomaterials. Ultramicroscopy 2013, 127, 53−63. (21) Liv, N.; van Oosten Slingeland, D. S.; Baudoin, J. P.; Kruit, P.; Piston, D. W.; Hoogenboom, J. P. Electron Microscopy of Living Cells During in situ Fluorescence Microscopy. ACS Nano 2016, 10, 265− 273. (22) Hirano, K.; Kinoshita, T.; Uemura, T.; Motohashi, H.; Watanabe, Y.; Ebihara, T.; Nishiyama, H.; Sato, M.; Suga, M.; Maruyama, Y.; Tsuji, N. M.; Yamamoto, M.; Nishihara, S.; Sato, C. Electron Microscopy of Primary Cell Cultures in Solution and Correlative Optical Microscopy Using Asem. Ultramicroscopy 2014, 143, 52−66. (23) Tanaka, K. A.; Suzuki, K. G.; Shirai, Y. M.; Shibutani, S. T.; Miyahara, M. S.; Tsuboi, H.; Yahara, M.; Yoshimura, A.; Mayor, S.; Fujiwara, T. K.; Kusumi, A. Membrane Molecules Mobile Even after Chemical Fixation. Nat. Methods 2010, 7, 865−866. (24) Peckys, D. B.; Korf, U.; de Jonge, N. Local Variations of Her2 Dimerization in Breast Cancer Cells Discovered by Correlative Fluorescence and Liquid Electron Microscopy. Sci. Adv. 2015, 1, e1500165.
METHODS Assembly of the Cells in the Microfluidic Chamber. The liquid-phase STEM system contained a microfluidic chamber assembled from two silicon microchips with electron-transparent windows of silicon nitride (Protochips Inc., NC).19 The windows had dimensions of 50 × 400 μm and were 50 nm thick. A sample with cells in liquid was placed between the windows, thus protecting them from the vacuum in the electron microscope. Prior to electron microscopy, fluorescence images were recorded of the microfluidic chamber assembled in the liquid specimen holder. Details of the loading procedure are described elsewhere.13 Fluorescence Microscopy. High-quality fluorescence images were obtained by using a 60×, 1.0 NA water immersion lens on an inverted microscope (Nikon, Japan, TS100). Water was placed between the lens and the lower silicon nitride window. A control experiment showed that yeast cells remained viable during fluorescence microscopy for extended periods of time.13 Liquid-Phase STEM. Low dose liquid-phase STEM was accomplished by adjusting the microscope settings at a sample region away from studied region in order to minimize the electron dose with a procedure described elsewhere.13 Gold nanoparticles at the windows served as feducial markers for the alignment. The STEM (CM200 TEM/STEM, Philips/FEI Company, OR) was operated at 200 keV beam energy, a convergence semiangle of 5.6 mrad, an electron probe current of 0.22 nA, a pixel dwell time of 10 μs, and a pixel size of 25 nm (magnification 4800). Assuming a homogeneous illumination of a pixel, the electron dose thus amounted to 0.2 e−/Å2. The annular darkfield detector was operated at a semiangle of 70 mrad. The image noise was filtered using a convolution filter with a kernel of (1, 1, 1; 1, 3, 1; 1, 1 1) in ImageJ (NIH). The gamma level was set to 0.75.
AUTHOR INFORMATION Corresponding Author
*E-mail:
[email protected]. Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS STEM images were recorded at the SHaRE User Facility, sponsored by the Division of Scientific User Facilities, Office of Basic Energy Sciences, U.S. N.dJ. is grateful to K.L. Gould for providing the yeast sample, to Protochips Inc. for providing the liquid-phase STEM system, and to E. Arzt for his support through INM. This research was supported in part by National Institutes of Health Grant 1R43EB008589 (to S. Mick). REFERENCES (1) Thiberge, S.; Nechushtan, A.; Sprinzak, D.; Gileadi, O.; Behar, V.; Zik, O.; Chowers, Y.; Michaeli, S.; Schlessinger, J.; Moses, E. Scanning Electron Microscopy of Cells and Tissues under Fully Hydrated Conditions. Proc. Natl. Acad. Sci. U. S. A. 2004, 101, 3346−3351. (2) de Jonge, N.; Peckys, D. B.; Kremers, G. J.; Piston, D. W. Electron Microscopy of Whole Cells in Liquid with Nanometer Resolution. Proc. Natl. Acad. Sci. U. S. A. 2009, 106, 2159−2164. (3) Peckys, D. B.; de Jonge, N. Liquid Scanning Transmission Electron Microscopy: Imaging Protein Complexes in Their Native Environment in Whole Eukaryotic Cells. Microsc. Microanal. 2014, 20, 346−365. (4) Nishiyama, H.; Suga, M.; Ogura, T.; Maruyama, Y.; Koizumi, M.; Mio, K.; Kitamura, S.; Sato, C. Atmospheric Scanning Electron Microscope Observes Cells and Tissues in Open Medium through Silicon Nitride Film. J. Struct. Biol. 2010, 169, 438−449. (5) Park, J.; Park, H.; Ercius, P.; Pegoraro, A. F.; Xu, C.; Kim, J. W.; Han, S. H.; Weitz, D. A. Direct Observation of Wet Biological Samples by Graphene Liquid Cell Transmission Electron Microscopy. Nano Lett. 2015, 15, 4737−4744. 9063
DOI: 10.1021/acsnano.6b02809 ACS Nano 2016, 10, 9061−9063
