A Structural Study of Escherichia coli Cells Using an In Situ Liquid Chamber TEM Technology

Studying cell microstructures and their behaviors under living conditions has been a challenging subject in microbiology. In this work, in situ liquid chamber TEM was used to study structures of Escherichia coli cells in aqueous solutions at a nanometer-scale resolution. Most of the cells remained intact under electron beam irradiation, and nanoscale structures were observed during the TEM imaging. The analysis revealed structures of pili surrounding the E. coli cells; the movements of the pili in the liquid were also observed during the in situ tests. This technology also allowed the observation of features of the nucleoid in the E. coli cells. Overall, in situ TEM can be applied as a valuable tool to study real-time microscopic structures and processes in microbial cells residing in native aqueous solutions.


Introduction
High-resolution real-time observation of cells and cellular processes in their liquid environments is crucial for revealing cellular structures and functions. Electron microscopes (EMs) can easily achieve a nanometer level resolution [1]; therefore, EMs are efficient tools for observing cellular structures. However, traditional sampling procedures limit the application of EMs for real-time observation of liquid samples. Because of the high-vacuum operating conditions, extensive sample preparations (e.g., fixation, metal staining, plastic embedding and slicing [2], or freezing) are needed for biological TEM samples [3]. An in-depth understanding of cellular structures and intracellular processes requires realtime imaging of the entire biological object with high spatial resolutions in the native liquid environment. To allow realtime and fast observations of dynamic processes occurring in biological objects, the samples must be fully hydrated and nonfrozen. Environmental TEM uses a differential pump technology and therefore allows cellular studies under partially hydrated conditions [4]; however, this process is not conducted in pristine liquid environments [5,6].
In situ liquid chamber TEM has been developed recently for nanostructural systems; this enables the user to image and monitor biological samples in fully hydrated environments, thus providing real-time dynamic information [7]. In this method, two ultrathin electron-transparent window chips are used to construct a liquid cell chamber; the liquid sample is sealed between these chips. The in situ liquid cells can be used in standard TEM instruments [8]. A variety of samples and processes of nanomaterials have been studied with this technique, including nanomaterial reactions [9] and motion [10], nanocrystal depositions [11,12], and selfassembly of nanolipoprotein discs [13]. Biological cells have also been imaged in liquid with nanometer resolution, such as human leukocytes [14], COS7 fibroblast cells [15], yeast cells [16], K. pneumoniae CG43S3 cells [7], and the bacteria Deinococcus radiodurans (D. radiodurans) [17]. E. coli has also been investigated previously [18]. Typically, a specimen kit equipped with SiO 2 nanomembranes is applied in TEM imaging. This kit enables the observation of living E. coli [7]. E. coli has been reported to survive this TEM imaging process at 2.8 s under a 200 KeV electron beam [7]. Peckys et al. imaged E. coli [18] labeled with gold nanoparticles in liquid with scanning transmission electron microscopy (STEM). However, they only showed the overall morphology and viability of E. coli cells with limited resolution. The stability of the cells under the electron beam, the resolution, and the dynamic observation are key issues that need to be further addressed.
This study focuses on the fine structural analysis and dynamic observation of E. coli cells using in situ liquid chamber TEM. The integrity, stability, and viability of the cells under electron beams, which have not been examined in previous studies, are also examined.

Materials and Methods
Escherichia coli BL21(DE3) was grown in Luria-Bertani (LB) broths at 37 ∘ C for 12 h and then stored at 4 ∘ C before being used. The liquid chamber used for the in situ TEM was developed following a previously described procedure [9][10][11]18]. In brief, silicon chips with 50 nm thick Si 3 N 4 membranes (Ted Pella, Inc., CA, USA) were used as the electron transparent windows for the liquid chamber. The size of the windows was 0.5 × 0.5 mm. A JEOL JEM 2010 TEM (JEOL Ltd., Tokyo, Japan) was used for the in situ TEM and was operated under a 200 kV acceleration voltage. The beam intensity measured on the phosphorous screen is ∼50 pA/cm 2 . The E. coli cells used for fluorescence microscopy were suspended in 0.85% NaCl after growth medium was removed. The sample was then mixed with 2x Bacterial Viability Kit stock solution with an equal volume. LIVE/DEAD BacLight Bacterial Viability Kit was from life technologies (Eugene, Oregon, USA). Briefly, a drop of 0.6 L suspension was used in liquid chamber. The liquid chamber was imaged directly by fluorescence microscopy (Olympus IX51 inverted microscope, Tokyo, Japan) with exposure time: 67.08 ms. Green fluorescence image and corresponding red fluorescence image were merged together into one picture.

Results and Discussion
To investigate the cellular fine structures of bacterial cells under aqueous conditions, in situ TEM was performed on an E. coli sample in the liquid chamber. Most of the cells appeared to be structurally robust against electron beam irradiation and preserved nanostructures during the TEM imaging. The E. coli cells in the in situ bright field TEM images (Figure 1 comparable to previous in situ liquid chamber TEM studies [7,18]. Furthermore, as seen in Figures 1(b)-1(d), our E. coli images show more details than other reports in the literature, with several cellular structures detected outside of the oval-rod shaped bacterial cells and a few cells showing dim dark nucleoid centers, which are indicated by the two green arrows in Figure 1(b). The fine structures outside of the E. coli cells in the in situ TEM images may be attributable to pili, which are displayed as gray halos around some of the bacterial cells (Figure 1(c)). Almost half of the E. coli cells showed pili, but the other cells had no apparent pili structures (Figure 1(d)). Whether the pili were observed might be related to the biological activity and cell life cycle of the bacteria. This conforms with the property of pili. It is possible that the resolution of in situ TEM might not be high enough to detect additional intracellular details of the E. coli cells, such as a clear nucleoid. However, images have been reported with better than 1 nm and even atomic resolutions using in situ liquid chamber TEM, which indicates that getting enough resolution should not be an obstacle for the technology. In our in situ TEM images of E. coli cells, pili were observed clearly, which are known to have diameters less than 10 nm. Resolutions better than 10 nm have been obtained as measured from the TEM images, which is more than enough to image a clear nucleoid if it exists. Thus, the lack of details for E. coli is not due to the resolution limit of in situ TEM but may be because of the stage of the cells during the observation. The absence of low contrast in the dark centers may suggest that the nucleoids in the E. coli cells under the fully hydrated environment were not as condensed as those cells that were dried or processed to be used in other imaging techniques.
With nanometer resolutions and the capability of observing cells in their fully hydrated living environments, in situ TEM is a more powerful tool than conventional EMs. However, the length of time that a cell can withstand the electron beam irradiation without being subjected to structural damage or cell death is also of concern. Currently, studies have only successfully irradiated cells for several seconds using in situ TEM, which has largely limited its applications to in situ monitoring of dynamic changes in the cells. In this study, E. coli cells demonstrated good structural stability. The cells were observed under the electron beam for extended time, and the majority of the cells did not show beaminduced structural damage. The cell viability was examined by using LIVE/DEAD BacLight Bacterial Viability Kit, and the observations are shown in Figure 2. The cells exposed under TEM high energy electron beam (Figure 2(a)) still emitted green fluorescence (which indicated alive cells) (Figure 2(b)), instead of red fluorescence (which indicated dead cells). The other cells also emitted green fluorescence but could not be observed clearly, possibly due to a variation of imaging depth. These results further verified the survival of the E. coli cells through the analysis.
In addition to the overall stability and integrity under the in situ TEM condition, several E. coli cells ruptured. In Figure 3(a), a few cells in the right area of the picture that are circled by red dotted lines were found to be lighter, suggesting that they were broken cells. To the right of the figure, we detected several such broken cells forming a vertical line. In the magnified image in Figure 3(b), one cell was broken from the top end. Compared to the intact cells, the ruptured cell showed a size extension at the broken site, which might partially be because of the unfolding of the cell membrane and the extruding pressure from the cytoplasm. The nonruptured end appeared to be smaller, possibly related to the loss of cytoplasm and pressure inside the cell. In Figure 3(c), in addition to the partially broken cell in the upper part, a completely broken cell is displayed in the lower part of the image. This cell showed a dim cell membrane with a larger diameter than a complete cell, and the cellular region within the membrane boundary has a lighter shade than an unbroken cell. In Figure 3(d), in addition to the cell to the right that was broken at one end, one E. coli cell to the left was broken from the middle, forming a pore. Figure 3(e) shows a lytic broken cell forming membrane fragments, and as a result the outline of the oval shape of the cell was not clear. Several studies have shown that E. coli cells become lytic when the cell membrane is damaged [19], releasing intracellular components [20]. The phenomenon we observed could be cell lysis induced by damage to the membrane. In previous studies, pores have been suggested to form at the early stages. Additionally, the osmotic pressure across the membrane might also induce membrane rupture and cause more dramatic damage to a bacterial cell [21]. The initiation of this rupture was not clear and may not necessarily be caused by the imaging because these cells might have already broken before being examined.
Although rare, a dynamic process of cell damage has also been observed using in situ TEM. Within a minute, an ovalshaped E. coli was quickly damaged. The outer ring of the oval-shaped E. coli cell became deeper gray, getting darker and wider with time. The image of the E. coli cell became brighter with time (Figures 4(a)-4(d)). In the final image (Figure 4(f)), the contour of the E. coli was rod-like, smaller in size, and the brightness matched the background of the picture. In Figures 4(a)-4(e), approximately eight growing dark spots were also observed. The spots might originate from the dissolution of cellular debris. One reason for such coincidental damage might be the effect of radiation from beam-sample interactions. When high energy electrons pass Journal of Analytical Methods in Chemistry through an aqueous solution, the liquid system produces radicals and hydrated electrons, such as OH radicals [22,23]. These radicals and hydrated electrons could damage the biomolecules of the cells. To analyze the effect of radicals on E. coli cells, 10 mg/mL glucose was added to the medium as radical scavengers. In the liquid chamber containing glucose, most of the E. coli cells kept their native structures without damage. This chamber showed lower red fluorescence than the liquid chamber without glucose (shown in Figure S1 in Supplementary Material available online at http://dx.doi.org/10.1155/2015/829302). This indicated less dead cells in liquid chamber with glucose.
In the higher resolution images, pili were distinguished in the halo around the cells. Figures 5(a) and 5(b) show video frames taken from the labeled region in Figure 1(a). Partially because of the higher magnification, we obtained better images and can clearly detect the pili surrounding the cells. These pili observed in the current work resemble those previously reported using conventional imaging techniques [24,25]. Although the samples were in an isolated chamber, the liquid flow can still occasionally be observed in the sample. The liquid flow behavior was captured in the video. In Figure 5(c) (lower right part of the image), we detected a thicker liquid layer flowing into the cell, which got darkened and blurred. Figure 5(d) shows an image taken after the thick liquid covered the full imaging region, which produced an almost featureless picture because of the electron diffraction through the thick sample. Figures 5(e)-5(h) show magnified video frames taken from the labeled region in Figure 5(a); the arrow points to a pilus that changed shape with time. For the video of Figure 5 only, the imaging time was longer than 60 s, much longer than reported exposure time in the literature, and we did not see any structural damage to the cell. This makes a significant progress in dynamic observation of biosample using in situ liquid chamber TEM.
Generally, people believe there are big bubbles in the in situ liquid chamber [26,27], filling the majority of the space, leaving possibly only an ultrathin layer of liquid on the two windows, which was indicated in Figure 6. Thus the cells are not necessarily surrounded by liquid only but are surrounded by liquid plus vapor, which explains the good image resolution. Even when surrounded by vapor only, the surrounding vapor pressure is much larger than in an environmental TEM; thus the cells are more hydrated. Thus, it is easily understood that the liquid layer thickness can change with time due to liquid flow, resulting in the image resolution change, supporting the observation in Figures 5(a)-5(d) [18]. In addition to the motion of the pili and the liquid-cell interactions observed, TEM allows the study of various events, such as flagellar motions, nanoparticle-cell interactions, and even cell divisions. Therefore, this tool could be beneficial in numerous important studies investigating cell structures and processes.

Conclusions
In conclusion, in situ liquid chamber TEM was used to image E. coli cells in liquid environments. The in situ TEM results clearly displayed contours and fine cellular structures (such as pili) of the E. coli cells. Notably, the nucleoids in the E. coli cells in the fully hydrated environment appeared to be in noncondensed forms. Additionally, during our in situ TEM experiments, most of the E. coli cells were robust under the electron beam, without showing structural damage over time. The liquid chamber TEM allowed several fine structure and motion observations that were impossible with traditional  TEM or SEM analyses. Liquid chamber TEM is capable of in situ monitoring of significant biological processes, including cell rupture and cellular structural movement. A variety of cell damage types were also observed. Notably, liquid flow behavior was observed. Additionally, the motion of the pili was recorded for the first time using the liquid chamber TEM technology. Overall, the in situ liquid chamber TEM is a facile and efficient tool for E. coli analysis. This method provides reliable studies of the cellular structures and behaviors in their native growing environments.