Open Access
Add to BibSonomyAbstract
Intracellular structures of HeLa cells are observed using a direct electron beam excitation-assisted fluorescence (D-EXA) microscope. In this microscope, a silicon nitride membrane is used as a culture plate, which typically has a low biocompatibility between the sample and the silicon nitride surface to prevent the HeLa cells from adhering strongly to the surface. In this work, the surface of silicon nitride is modified to allow strong cell attachment, which enables high-resolution observation of intracellular structures and an increased signal-to-noise ratio. In addition, the penetration depth of the electron beam is evaluated using Monte Carlo simulations. We can conclude from the results of the observations and simulations that the surface modification technique is promising for the observation of intracellular structures using the D-EXA microscope.
© 2015 Optical Society of America
1. Introduction
Silicon nitride membranes are used as culture plates in various microscopy technologies such as atmospheric scanning electron microscopy [1], scanning electron generation X-ray microscopy [2], scanning transmission electron microscopy [3], and electron beam excitation-assisted optical microscopy [4]. In atmospheric scanning electron microscopy, the silicon nitride membrane window is used as the bottom of an open culture dish to observe biological samples in the open environment [1], where the samples in the culture medium can be observed with high resolution.
Silicon nitride membranes are also used as a culture plate in the direct electron beam excitation-assisted fluorescent (D-EXA) optical microscope we have developed [5,6], which possesses a higher resolution than that allowed by the diffraction limit of conventional optical microscopes because the specimens are excited with a focused electron beam. In D-EXA, the silicon nitride membrane separates the atmosphere and the vacuum, and we can observe living specimens in the same conditions as those used in conventional optical microscopy. This is because vacuum conditions are not required on the sample side and the biological samples can be placed in atmospheric pressure. It is therefore possible to dynamically observe the activities of cell organelles and granules in real time with high spatial resolution using the D-EXA microscope.
Although silicon nitride membranes are widely used as a culture plate, its biocompatibility is not high, which prevents some cells from adhering strongly and proliferating well because of the low hydrophilicity of the surface. In this study, we modify the silicon nitride surface chemically by attaching carboxyl groups [7], after which we successfully culture HeLa cells on the modified surface. We find that cells on the modified silicon nitride surface spread more widely than those on an unmodified surface, and that the cellular adhesion force increases and the cells attach to the surface more strongly. Monte Carlo simulations. Monte Carlo simulations are used to calculate the penetration depth of the electron beam, and using the experimental conditions, to determine the optimal position of the cell structure from the silicon nitride membrane to produce a high-spatial-resolution image with strong cathodoluminescence. We present autofluorescence images of cells using the D-EXA microscope with the modified membrane, wherein the organelles and granules can clearly be observed. These results show that silicon nitride surface modification supports intercellular structure observation using electron beam excitation.
2. Principle of the D-EXA microscope
Figure 1 shows a schematic of the principle of the D-EXA microscope for living cell observation, wherein biological specimens are placed on a silicon nitride membrane and the electron beam is irradiated on the specimen through the membrane. Because the electrons are focused to a spot only a few tens of nanometers in diameter within the specimen, it is possible to achieve nanometric resolution. Emitted cathodoluminescence from the nanometric region is detected with a photomultiplier tube, and the image is constructed as the electron beam is raster scanned across the surface. The irradiation area of the electron beam is confined to a region tens of nanometers in diameter, although the electrons are scattered during propagation through the specimen, allowing the D-EXA microscope to achieve a resolution better than that allowed by the diffraction limit of light.
3. Surface modification to increase biocompatibility
To increase the hydrophilicity of the silicon nitride surface, carboxyl groups are covalently coupled on its surface. To achieve a monolayer formation of the carboxyl groups, 12-bromododecanoic acid (541397; Sigma Aldrich, St. Louis, MO, United States) is attached to the surface using an oxidation–reduction reaction by heating at 110°C with reflux. The surface density of the carboxyl groups on the silicon nitride surface is controlled to be in the range of 0.9–6.4 × 1013/cm2. The thickness of carboxylic monolayer is estimated around 15 nm which is deduced from molecule size of 12-bromododecanoic acid. The contact angle of water droplets on the membranes is then measured to evaluate the hydrophilicity of the surface, where it is seen that the contact angle decreases after the modification because the surface hydrophilicity is increased by the bounded carboxyl groups, and where control of the hydrophilicity of the silicon nitride surface is achieved by increasing the amount of bound carboxyl groups (data not shown). The procedure of the surface modification and evaluation of the hydrophilicity have been presented in [8].
To evaluate the biocompatibility of the modified surface, HeLa cells are directly cultured on the silicon nitride membranes, and Fig. 2 shows phase-contrast microscopy images of HeLa cells on the unmodified and modified membranes. The HeLa cells on the modified membrane are observed to spread wider than those on the unmodified membrane, and the density of carboxyl groups on the biocompatible surface is evaluated to be between 2.2 and 3.1 × 1013 cm−2.
Fig. 2 Phase-contrast microscopy image of HeLa cells (a) on a modified membrane and (b) on an unmodified membrane.
4. D-EXA microscopy observations
The autofluorescence of the HeLa cells are observed using a D-EXA microscope, where a comparison between the D-EXA and phase-contrast microscope images of fixed HeLa cells are shown in Fig. 3 for both the unmodified membrane and the modified membrane containing ~3.0 × 1013 carboxyl groups per cm2. The same area of the specimen is observed and the location of HeLa cells are matched in the D-EXA (Figs. 3(a)–3(b)) and phase-contrast (Figs. 3(c)–3(d)) microscopy images. The fine structures in the HeLa cells on the modified membrane are successfully observed by exciting autofluorescence of the cells using the D-EXA microscope. Intracellular structures of nuclei fiber-like structures and granules can clearly be seen with the modified membrane using D-EXA (Fig. 3(a)), though only a few intracellular granules can be seen using D-EXA with the unmodified membrane (Fig. 3(b)). This observation result clearly indicates that HeLa cells are able to stretch and adhere strongly on the modified membrane, thereby allowing the intracellular structures to approach the membrane surface. The intracellular structures in the penetration depth range of the electrons are therefore more clearly observed in these samples compared with those on the unmodified membrane.
Fig. 3 Images of fixed HeLa cells using D-EXA microscopy with (a) modified and (b) unmodified silicon nitride, and the same areas using phase-contrast microscopy with (c) modified and (d) unmodified silicon nitride. Granules (solid arrowheads) and fiber-like structures (hollow arrowheads) are indicated. The shafted arrow in (a) and (c) indicate the nucleus. The white squares in (a) and (b) are the areas evaluated in Fig. 4(a) and 4(b), respectively.
Figure 4 shows the line profiles of a granule in a HeLa cell on a modified membrane (Fig. 4(a)) and an unmodified membrane (Fig. 4(b)) imaged using D-EXA microscopy, where it can be seen that the maximum fluorescence intensity of the granule on the modified membrane is more than 10 times greater than that on the unmodified membrane. Additionally, the full width at half maximum of the granule on the modified membrane is 196.1 nm, while that on the unmodified membrane is 441.2 nm. This indicates that the granule on the modified membrane is being observed with higher resolution, which is because the electron beam is less scattered before it reaches the specimens. This reduced scattering is owing to the fact that the intercellular structures on the modified membranes are approaching nearer to the surface, thereby allowing higher resolution to be achieved.
Fig. 4 Photon number profiles of a granule in a HeLa cell on (a) modified and (b) unmodified silicon nitride. (Insets) Respective D-EXA images from which the line scan profiles are obtained.
5. Simulation results of electron scattering and transmission
We simulate the electron scattering using Monte Carlo simulations to evaluate the penetration depth in the cell, and the simulated electron trajectories can be seen in Fig. 5. For the simulations, we assume that the thickness of silicon nitride membrane is 50 nm and that the backside of membrane is filled with water. Additionally, the acceleration voltage is 4.8 keV, which is same as the observation condition, and the diameter of the focused electron beam is 2 nm.
The electrons are scattered when they encounter the silicon nitride, and their energy decreases with the repeated scattering. The distribution of electron beam at the silicon nitride surface is about 50 nm, and about 87% of the electrons lose all of their energy at a depth of 150 nm from the silicon nitride surface, which means that we can effectively excite the specimen in a region 150 nm from the silicon nitride surface. Therefore, to obtain a high-spatial-resolution image with strong cathodoluminescence, the structure should be placed no farther than 150 nm from the membrane surface.
6. Conclusions
We demonstrate the modification of a silicon nitride membrane using carboxyl groups, where HeLa cells are seen to spread very well on the membrane containing between 2.2 and 3.1 × 1013 carboxyl groups per cm2. The presence of HeLa cells on the modified silicon nitride surface with the D-EXA microscope is also demonstrated, and the autofluorescence of intercellular structures such as nuclei and fibers are imaged clearly and with high resolution. In the D-EXA microscope, we can obtain the highest spatial resolution near the surface of a silicon nitride membrane because the electron beam is scattered and broadens after entering the atmospheric pressure on the opposite side of the membrane. Because the cells spread wider on the modified membrane, the intracellular structures effectively approach the surface and are excited with high efficiency. Our surface modification technique is therefore promising for the observation of intracellular structures using the D-EXA microscope.
We observed the auto-fluorescence of the intracellular structures with the D-EXA microscope. As an important next research topic, we need to measure the spectrum of auto-fluorescence of cells, and the biomolecules excited by the D-EXA microscope can be identified with comparison of the cathodoluminescence spectra of biological molecules [9,10]. The damages of carboxylic monolayer film with the irradiation of electron beam have not been observed during our experiments. The detailed evaluation of the film damage is required for the optimization of observation condition.
Acknowledgments
This work has been supported by JST-CREST (Japan Science and Technology Agency, Core Research for Evolutionary Science and Technology).
References and links
1. H. Nishiyama, M. Suga, T. Ogura, Y. Maruyama, M. Koizumi, K. Mio, S. Kitamura, and C. Sato, “Atmospheric scanning electron microscope observes cells and tissues in open medium through silicon nitride film,” J. Struct. Biol. 169(3), 438–449 (2010). [CrossRef] [PubMed]
2. T. Ogura, “Direct observation of unstained wet biological samples by scanning-electron generation X-ray microscopy,” Biochem. Biophys. Res. Commun. 391(1), 198–202 (2010). [CrossRef] [PubMed]
3. N. de Jonge, D. B. Peckys, G. J. Kremers, and D. W. Piston, “Electron microscopy of whole cells in liquid with nanometer resolution,” Proc. Natl. Acad. Sci. U.S.A. 106(7), 2159–2164 (2009). [CrossRef] [PubMed]
4. W. Inami, K. Nakajima, A. Miyakawa, and Y. Kawata, “Electron beam excitation assisted optical microscope with ultra-high resolution,” Opt. Express 18(12), 12897–12902 (2010). [PubMed]
5. 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]
6. Y. Nawa, W. Inami, A. Miyake, A. Ono, Y. Kawata, S. Lin, and S. Terakawa, “Dynamic autofluorescence imaging of intracellular components inside living cells using direct electron beam excitation,” Biomed. Opt. Express 5(2), 378–386 (2014). [CrossRef] [PubMed]
7. F. Cattaruzza, A. Cricenti, A. Flamini, M. Girasole, G. Longo, A. Mezzi, and T. Prosperi, “Carboxylic acid terminated monolayer formation on crystalline silicon and silicon nitride surfaces,” J. Mater. Chem. 14(9), 1461–1468 (2004). [CrossRef]
8. Y. Masuda, W. Inami, A. Miyakawa, and Y. Kawata, “Cell culture on hydrophilicity-controlled silicon nitride surfaces,” World J. Microbiol. Biotechnol.under review.
9. G. Nichols, “Applications of cathodoluminescence spectroscopy and imaging in the characterisation of pharmaceutical materials,” Eur. J. Pharm. Sci. 45(1-2), 19–42 (2012). [CrossRef] [PubMed]
10. P. V. C. Hough, W. R. McKinney, M. C. Ledbeter, R. E. Pollack, and H. W. Moos, “Identification of biological molecules in situ at high resolution via the fluorescence excited by a scanning electron beam,” Proc. Natl. Acad. Sci. USA 73(2), 317–321 (1976). [CrossRef] [PubMed]
