We developed a high-resolution fluorescence microscope in which fluorescent materials are directly excited using a focused electron beam. Electron beam excitation enables detailed observations on the nanometer scale. Real-time live-cell observation is also possible using a thin film to separate the environment under study from the vacuum region required for electron beam propagation. In this study, we demonstrated observation of cellular components by autofluorescence excited with a focused electron beam and performed dynamic observations of intracellular granules. Since autofluorescence is associated with endogenous substances in cells, this microscope can also be used to investigate the intrinsic properties of organelles.
© 2014 Optical Society of America
In biological cells, cellular functions emerge as a result of localization and dynamic interaction of molecules. To investigate the distribution and movement of such molecules in real time, several high-resolution microscopy techniques using light or electrons have been developed.
Light microscopy has the advantage of allowing noninvasive examination of living biological specimens. The spatial distribution of specific cellular components can be determined using fluorescence microscopy. Use of immunostaining with various fluorescent dyes, including potential sensitive dyes and pH- and ion-dependent fluorophores, enables investigation of cellular states and functions [1–5]. Autofluorescence analysis, in which cellular molecules are optically excited without the need for a fluorescent dye, is also useful for investigating the chemical composition of organelles. When the excitation and emission spectra of cellular molecules have been established , it can be used to identify specific components in cells. Cellular states can be investigated because physiological or pathological processes produce changes in the distribution and intensity of autofluorescence. The absence of any need for a fluorescent dye eliminates cell damage that may occur during labeling, which sometimes limits the types of observations that can be made on living biological specimens.
Scanning electron microscopy (SEM) is also a powerful tool for cell analysis because of its extremely high spatial resolution. In SEM analysis, several signals can be obtained, including secondary electrons, backscattered electrons, transmitted electrons, Auger electrons, X-rays, and cathodoluminescence (CL). Since CL is produced by loosely bound π-electrons in organic compounds under electron beam excitation , it can be used to investigate chemical bonding in specimens. CL analysis using a focused electron beam is more advantageous than conventional fluorescence microscopy [8–10] in that a higher spatial resolution can be achieved and it can be applied to a wider variety of substances because of the higher energy associated with accelerated electrons. However, CL analysis of living cells is limited because of the requirement of a vacuum environment and the need for sample preparation steps such as thin slicing, metal staining, and freezing.
Recent development of an electron-transparent membrane has made it possible to apply EM techniques for the analysis of hydrated specimens. This leads to various applications in biological analysis [11–15]. We have previously reported the development of a very-high-resolution direct electron-beam excitation assisted (D-EXA) optical microscope for dynamic observation of living cells . Fluorescent materials attached as labels to specific cellular molecules or endogenous fluorophores are directly irradiated with an electron beam, and the resulting CL is detected. Such electron beam excitation enables achievement of nanometer-scale resolution, beyond the diffraction limit of light.
In the present study, autofluorescence from HeLa cells in aqueous solutions was observed, and we demonstrated that intracellular granules and cytoskeletal structures can be seen without need for any staining. Dynamic movement of intracellular granules was observed using time-lapse imaging. Thus, the D-EXA microscope is extremely useful for autofluorescence studies of cells and for its ability to accomplish label-free imaging of intracellular structures.
2. Principle of autofluorescence imaging with the D-EXA microscope
Figure 1 shows the principle of live-cell imaging under electron beam excitation. Cells are directly cultured on an electron-transparent thin film and irradiated through the film using a focused electron beam without need for pretreatment such as fixation or staining. The autofluorescence produced by cellular components is emitted as CL, which is detected under atmospheric pressure. Use of a focused electron beam allows nanometer-scale resolution because the electron beam can be focused to a diameter of a few tens of nanometers after penetrating the film. The amount of electron scattering occurring in the film can be calculated by Monte Carlo simulation [16–18].
Images can be formed by raster scanning the electron beam, and a frame rate corresponding to real-time video can be achieved. Since some cellular substances emit CL, the D-EXA microscope enables label-free imaging in addition to qualitative chemical analysis using the CL spectrum.
3. Materials and methods
The D-EXA microscope is a combination of a scanning electron microscope and a fluorescence microscope [16,18], and open environment around specimens is available in this configuration [12,13,16,18]. A scanning electron microscope is used for excitation of specimens and scanning the electron beam to reconstruct images, and a fluorescence microscope is used for detection of CL excited with the electron beam. CL is detected using a photomultiplier tube (Hamamatsu Photonics K.K., H10721-110).
In the present study, a 50-nm thick silicon nitride (SiN) membrane window (Silson) was used as an electron-transparent film to support specimens. A 50 µm × 50 µm aperture was produced in the silicon substrate supporting the silicon nitride film. Biological cells were directly cultured on a SiN membrane fixed to a culture dish . The culture dish for the D-EXA microscope consisted of a glass dish, a SiN membrane, and a metal plate, as shown in Fig. 2. First, a SiN membrane was affixed to the metal plate, which was used to avoid a charging effect resulting from the electron beam irradiation. The metal plate had a hole in its center through which the electron beam was passed to excite specimens. A glass dish with a hole larger than the SiN membrane was then affixed to the metal plate. This dish was for containing the culture medium during incubation and observation. Epoxy resin was used for all fixation procedures. After the dish was sterilized for 21 min at 121°C in an autoclave (Super Clave FX-220, Hillson), the cells were incubated in Dulbecco’s Modified Eagle’s Medium (DMEM, Sigma) until they adhered to the film. After incubation, DMEM was removed and phosphate-buffered saline was placed in the dish before observations were made.
4. Observation results
The spatial resolution of our microscope depends on the excitation spot size formed by scattering of the electron beam. First, to evaluate the actual electron scattering in the substrate and the spot size used in the D-EXA microscope, 20-nm gold spheres (EMGC20, BBI solution) were observed using the D-EXA microscope and a conventional field emission scanning electron microscope (FE-SEM) (JSM 7001, JEOL). Figures 3(a) and 3(b) show the experimental setup for the D-EXA microscope and FE-SEM observations. In the D-EXA imaging, gold spheres under atmospheric pressure were observed through the SiN membrane. Specimens consisting of heavy elements were observed detecting the backscattered electrons (BE) through the film [12,13]. Figure 3(c) shows the BE image acquired using an acceleration voltage of 5 kV. The same area was observed under vacuum using FE-SEM. Figure 3(d) shows the secondary electron (SE) image of the gold spheres acquired using an acceleration voltage of 15 kV. In both images, gold spheres were visible as bright spots. Figure 3(e) and 3(f) show the line profile of each gold sphere, indicated with arrows. The intensity distribution was averaged with the width of the line, 5 pixels. Full widths at half maximum (FWHMs) of the fitting curves were about 63 and 18 nm. Figure 3(f) shows the actual size of the gold spheres, whereas FWHMs in Fig. 3(e) were broadened owing to electron scattering in the substrate.
Microscope images are formed as a convolution of the object function with the point-spread function (PSF) of the microscope. Given that the image function and the object function are the fitting Gaussian functions in Fig. 3(e) and 3(f) respectively, the FWHM of the PSF can be calculated as 60 nm. Thus, a spot size of several tens of nanometers is realized in the D-EXA microscope under an acceleration voltage of 5 kV after the electron beam penetrates the 50-nm thick SiN membrane. In this estimation, a BE image was employed for the image function. BE are scattered in the SiN membrane before they reach the detector. For more accurate evaluation, it is necessary to observe CL from phosphors smaller than a few tens of nanometers.
A spot size of several tens of nanometers was also demonstrated by observation of zinc oxide (ZnO) nanoparticles smaller than 50 nm (Sigma-Aldrich; 677450). ZnO is well known as a brilliant emitter of photoluminescence and CL [19,20] and can be applied for bio-imaging . Figure 4(a) shows a secondary electron image of ZnO nanoparticles acquired with an FE-SEM. ZnO nanoparticles were dispersed on the substrate and dried. Figure 4(b) shows a D-EXA image of an isolated ZnO nanoparticle in aqueous solution. The acceleration voltage is 5 kV. A ZnO nanoparticle is visible as a bright spot. A line profile between the arrows is shown in Fig. 4(c). The intensity distribution was averaged with the width of the line, 10 pixels. The FWHM of the Gaussian fitting curve is about 57 nm. The average signal-to-noise ratio (SNR) was 8.52. The SNR was determined from the ratio of the peak signal height and the standard deviation of the background signals. The peak signal height is the difference between the maximum signal and the average of the background signal. The average and standard deviation of the background signal were determined at horizontal positions between 0 and 500 nm in 5 areas. According to the Rose criterion [11,22], an SNR larger than 5 is required for reliable identification of an image object. Given that the SNR of this image is 8.52, larger than 5, the observation of ZnO nanoparticles indicated that a spot size of the electron beam was smaller than 60 nm after penetrating the SiN membrane in this experiment.
For investigating structures excited with the electron beam, fixed HeLa cells were observed at an electron acceleration voltage of 5 kV in phosphate-buffered saline (PBS) solution. Cells adhering to the SiN membrane were fixed with a 1% glutaraldehyde solution. Figure 5 shows a comparison of images of fixed HeLa cells recorded under a phase-contrast microscope and the D-EXA microscope. In the phase-contrast image in Fig. 5(a), intracellular granules appear as small spots, as shown with dark arrows. Figure 5(b) shows the corresponding autofluorescence image, in which the intracellular granules observed in Fig. 5(a) are clearly observed as bright spots. Since cell membranes also emit autofluorescence, the outlines of the cells can be observed. This image was acquired with a probe current of 2.056 nA.
Autofluorescence observations were also made on cytoskeleton structures in fixed HeLa cells. Before the observations, the cells were fixed using a 1% glutaraldehyde solution and treated with a 0.25% Triton X in PBS solution . Figure 6(a) shows an autofluorescence image obtained under UV excitation using a conventional epifluorescence microscope. The autofluorescence is emitted from structures in the nucleus and cell membrane. The same specimen was then imaged at high magnification using the D-EXA microscope at a probe current of 1.23 nA, as shown in Fig. 6(b). Filamentous structures in the cytoskeleton are clearly observed without the need for any staining. These structures were not distinguishable with the conventional fluorescence microscope. In the D-EXA microscope, observation depth is limited by the penetration depth of the incident electrons, which is calculated as approximately 600 nm in the water layer by Monte Carlo simulation, at an acceleration voltage of 5 kV. Thus, the autofluorescence is excited only in the thin region close to the substrate, and highly sensitive imaging of the membrane structure is made possible in ways similar to a total internal reflection fluorescence microscopy. These results show that the D-EXA microscope is capable of label-free imaging of intracellular structures in living cells.
The dynamic movement of intracellular granules was investigated using time-lapse imaging. The time series was acquired with an image size of 1024 × 1024 pixels, a probe current of 895 pA, a frame rate of 0.05 fps, and an acceleration voltage of 5 kV. The pixel size was calculated as approximately 50 nm. In the area of 24.3 µm × 18.7 µm in the acquired images, the dynamic behavior of the granules is shown in the video file (Media 1) and Fig. 7 shows time-series pseudo-color images from the video data. The video was played back at 8 fps, representing a speedup of 160 times. In Fig. 7, it can be clearly observed that the granules move over time. From the beginning of the observations, the granules began to aggregate, as indicated by the dotted outlines at 20 and 680 s. Because the electron beam was focused in the region of the specimen close to the substrate surface and the excitation energy decreased with the distance from the surface owing to scattering in the specimen, only the specimens near the surface are observed at an acceleration voltage of 5 kV . Intracellular granules, indicated by the solid arrowhead, the open arrowhead, and the white arrow, appeared during the observation. These are granules moving to and near the surface. The reduction in signal intensity of these particles is mainly due to the motion of the granules away from the substrate film . These results clearly demonstrate that the D-EXA microscope can be used to observe the dynamic behavior of intracellular granules.
5. Discussion and conclusions
We have demonstrated label-free live-cell imaging using the D-EXA microscope, in addition to dynamic observations of intracellular fine structures. The spatial resolution is mainly determined by electron scattering in the film and the specimen being bio-imaged. Based on the results of Monte Carlo simulations of scattering in water , a higher acceleration voltage may lead to higher spatial resolution. For autofluorescence imaging, the spatial resolution is also expected to be reduced as a result of the low image contrast due to homogeneous emission from the membrane, as shown in Fig. 5. To overcome this problem, reduction of such background signals is necessary.
Electron scattering can also damage specimens by energy transfer. The average electron dose used for the dynamic observation of intracellular granules was calculated as approximately 42 electrons/nm2 lower than the numbers of electrons used for imaging of living COS-7 cells in liquid STEM at an acceleration voltage of 200 kV . However, after acquisition of 1 or 2 images, cells can be subjected to photobleaching or can undergo morphological changes such as shrinkage. Since the cell membrane adhered to the SiN substrate, it was severely damaged by the electron beam bombardment, and photobleaching was observed as shown in Fig. 7. After long-term observations, cells appear dead and some cells detach from the substrate. To examine the intrinsic movements associated with cellular activity and interactions between molecules, it will be necessary to optimize electron beam irradiation conditions such as dwell time, probe current, acceleration voltage, and energy density. The first step toward achieving this is to clarify the relationship between cell damage and irradiation conditions.
In the present study, higher fluorescence intensity was obtained from intracellular granules than from other areas of cells. We attribute this result to the presence of a higher concentration of fluorescence substances in such granules. Future investigations will aim to identify these substances emitting CL. There have been several reports of the identification of organic substances based on CL wavelength [25,26].
In summary, the D-EXA microscope has high potential for CL analysis of living cells. It not only enables observations of specific molecules with high spatial resolution but also provides information about chemical compositions in living cells and wet materials under various conditions. This capability is important in many fields of science.
References and links
1. D. L. Taylor and Y.-L. Wang, Methods in Cell Biology vol. 30 (Academic Press Inc., San Diego, 1989).
2. M. Zochowski, M. Wachowiak, C. X. Falk, L. B. Cohen, Y. W. Lam, S. Antic, and D. Zecevic, “Imaging membrane potential with voltage-sensitive dyes,” Biol. Bull. 198(1), 1–21 (2000). [CrossRef] [PubMed]
3. R. A. Gottlieb, J. Nordberg, E. Skowronski, and B. M. Babior, “Apoptosis induced in Jurkat cells by several agents is preceded by intracellular acidification,” Proc. Natl. Acad. Sci. U.S.A. 93(2), 654–658 (1996). [CrossRef] [PubMed]
4. J. P. Y. Kao, “Practical aspects of measuring [Ca2+] with fluorescent indicators,” in Methods in Cell Biology vol. 40, R. Nuccitelli, ed. (Academic Press, Inc., San Diego, 1994).
5. S. L. Sensi, D. Ton-That, P. G. Sullivan, E. A. Jonas, K. R. Gee, L. K. Kaczmarek, and J. H. Weiss, “Modulation of mitochondrial function by endogenous Zn2+ pools,” Proc. Natl. Acad. Sci. U.S.A. 100(10), 6157–6162 (2003). [CrossRef] [PubMed]
7. M. De Mets, “Cathodoluminescence of organic chemicals,” in Principles and Techniques of Scanning Electron Microscopy vol. 2, M.A. Hayat, ed. (Van Nostrand Reinhold, NewYork, 1974).
9. P. J. Fisher, W. S. Wessels, A. B. Dietz, and F. G. Prendergast, “Enhanced biological cathodoluminescence,” Opt. Commun. 281(7), 1901–1908 (2008). [CrossRef]
11. N. de Jonge, N. Poirier-Demers, H. Demers, D. B. Peckys, and D. Drouin, “Nanometer-resolution electron microscopy through micrometers-thick water layers,” Ultramicroscopy 110(9), 1114–1119 (2010). [CrossRef] [PubMed]
12. H. Nishiyama, M. Suga, T. Ogura, Y. Maruyama, M. Koizumi, K. Mio, S. Kitamura, and C. Sato, “Reprint of: Atmospheric scanning electron microscope observes cells and tissues in open medium through silicon nitride film,” J. Struct. Biol. 172(2), 191–202 (2010). [CrossRef] [PubMed]
13. M. Suga, H. Nishiyama, Y. Konyuba, S. Iwamatsu, Y. Watanabe, C. Yoshiura, T. Ueda, and C. Sato, “The Atmospheric Scanning Electron Microscope with open sample space observes dynamic phenomena in liquid or gas,” Ultramicroscopy 111(12), 1650–1658 (2011). [CrossRef] [PubMed]
14. O. Cohen, R. Beery, S. Levit, J. Ilany, I. Schwartz, M. Shabtai, D. Anaby, D. Cohen, R. Alfici, A. Czerniak, and A. Karasik, “Scanning electron microscopy of thyroid cells under fully hydrated conditions-a novel technique for a seasoned procedure: a brief observation,” Thyroid 16(10), 997–1001 (2006). [CrossRef] [PubMed]
15. U. M. Mirsaidov, H. Zheng, Y. Casana, and P. Matsudaira, “Imaging protein structure in water at 2.7 nm resolution by transmission electron microscopy,” Biophys. J. 102(4), L15–L17 (2012). [CrossRef] [PubMed]
16. Y. Nawa, W. Inami, A. Chiba, A. Ono, A. Miyakawa, Y. Kawata, S. Lin, and S. Terakawa, “Dynamic and high-resolution live cell imaging by direct electron beam excitation,” Opt. Express 20(5), 5629–5635 (2012). [CrossRef] [PubMed]
17. D. C. Joy, Monte Carlo Modeling for Electron Microscopy and Microanalysis (Oxford Univ. Press, New York, 1995).
19. B. K. Meyer, H. Alves, D. M. Hofmann, W. Kriegseis, D. Forster, F. Bertram, J. Christen, A. Hoffmann, M. Straßburg, M. Dworzak, U. Haboeck, and A. V. Rodina, “Bound exciton and donor–acceptor pair recombinations in ZnO,” Phys. Status Solidi 241(2), 231–260 (2004). [CrossRef]
20. M. R. Phillips, O. Gelhausen, and E. M. Goldys, “Cathodoluminescence properties of zinc oxide nanoparticles,” Phys. Status Solidi 201(2), 229–234 (2004). [CrossRef]
21. Y. L. Wu, C. S. Lim, S. Fu, A. I. Y. Tok, H. M. Lau, F. Y. C. Boey, and X. T. Zeng, “Surface modifications of ZnO quantum dots for bio-imaging,” Nanotechnology 18(21), 215604 (2007). [CrossRef]
22. A. Rose, Advances in Electronics vol.1 (Academic Press Inc. New York, 1948).
23. C. A. Schnaitman, “Solubilization of the cytoplasmic membrane of Escherichia coli by Triton X-100,” J. Bacteriol. 108(1), 545–552 (1971). [PubMed]
25. T. Hürter, W. Bröcker, and K. A. Hossmann, “Evaluation of vasogenic edema in experimental brain tumors by cathodoluminescence and fluorescence microscopy,” Histochemistry 72(2), 249–254 (1981). [CrossRef] [PubMed]
26. T. Nakano, T. Fujimoto, H. Koike, and K. Ogawa, “Application of analytical color fluorescence electron microscopy to biomedical field: I. vitamin A ester in rat retina,” Acta Historiem. Cytochem. 23(6), 753–767 (1990). [CrossRef]