758
Biophysical Journal
Volume 110
February 2016
758–765
Article A Versatile High-Vacuum Cryo-transfer System for Cryo-microscopy and Analytics Sebastian Tacke,1,* Vladislav Krzyzanek,2 Harald Nu¨sse,1 Roger Albert Wepf,3 Ju¨rgen Klingauf,1 and Rudolf Reichelt1 1 Institute of Medical Physics and Biophysics, University of Mu¨nster, Mu¨nster, Germany; 2Institute of Scientific Instruments, The Czech Academy of Sciences, Brno, Czech Republic; and 3Scientific Center for Optical and Electron Microscopy, ETH Zu¨rich, Zu¨rich, Switzerland
ABSTRACT Cryogenic microscopy methods have gained increasing popularity, as they offer an unaltered view on the architecture of biological specimens. As a prerequisite, samples must be handled under cryogenic conditions below their recrystallization temperature, and contamination during sample transfer and handling must be prevented. We present a high-vacuum cryo-transfer system that streamlines the entire handling of frozen-hydrated samples from the vitrification process to low temperature imaging for scanning transmission electron microscopy and transmission electron microscopy. A template for cryo-electron microscopy and multimodal cryo-imaging approaches with numerous sample transfer steps is presented.
INTRODUCTION Preserving the native state until imaging is the final goal in sample preparation. Particularly for electron microscopy (EM), the conservation of the pristine architecture of hydrated samples is challenging. If not imaged in situ (1), two different strategies can be followed to prepare hydrated samples for EM: the conventional or the cryogenic routine. Conventional preparation protocols often include chemical fixation procedures as well as several staining steps, and samples are allowed to dry. The chemical processes are known to cause rearrangement or loss of cytoplasmic components, or agglomeration to membranes (2,3). Moreover, final staining with heavy elements offers just an indirect, replicalike illustration of the treated structures. In the 1960s to 1980s, Moor and Mu¨hlethaler (4) and Dubochet and McDowall (5) introduced the cryogenic preparation routine, which circumvents such drawbacks. Although Moor and Mu¨hlethaler (4) and Dubochet and McDowall (5) followed different postprocessing and imaging routines (freeze-etching versus direct imaging, respectively), the basic principle of both preparation procedures is the same: the physical arresting of the sample in a frozen-hydrated state, which is achieved by the vitrification of the sample within milliseconds, resulting in a sample embedded in amorphous ice, thereby avoiding structural changes due to ice crystal growth. The convincing structural preservation achieved in the case of isolated cell organelles (6,7), viruses (8), bacteria (9,10), eukaryotic cells (11) and other softmatter applications (12), and the technical improvements in sample preparation and handling (13–19), caused this
Submitted July 1, 2015, and accepted for publication January 20, 2016. *Correspondence: [email protected]
approach to spread out into other fields of microscopy like soft x-ray (20) and light microscopy (21,22), and enabled new hybrid approaches like correlative cryo-light and electron microscopy (23). However, the handling of vitrified samples is rather delicate. If imaged in the frozen-hydrated state, for example, the temperature of the specimen must be maintained below the glass transition temperature of water, 138 K (135 C) (5); for further details, see Capaccioli and Ngai (24) and Amann-Winkel et al. (25). As a consequence, the unprotected sample acts as a cold trap and is therefore susceptible to any kind of contamination. Samples that were partially or totally dehydrated, e.g., by freeze-drying, also require a gentle handling. Being dehydrated, the sample material turns hygroscopic and therefore must be transported in an anhydrous environment (3). Thus, the sample must be handled in liquid nitrogen (lN2), in a dry inert gas atmosphere or in a clean vacuum environment under controlled temperature conditions at all times. Most available transfer systems are either cryo-transfer systems or high-vacuum transfer systems. They typically require the installation of a specific cryo-stage inside the microscope, which is not feasible for in-lens systems, such as transmission electron microscopes or scanning transmission electron microscopes, due to the limited space in their specimen chambers. The latest electron microscopes are equipped with storage devices directly connected to the column, allowing samples to be stored and transferred to the microscope stage under high-vacuum environment at cryogenic conditions. However, transfer to other devices for further processing or analysis is still problematic, as the high-vacuum conditions have to be broken. Latest preparation protocols (17) and imaging techniques (26) include several transfer steps. The available
Editor: Andreas Engel. Ó 2016 by the Biophysical Society 0006-3495/16/02/0758/8
http://dx.doi.org/10.1016/j.bpj.2016.01.024
High-Vacuum Cryo-transfer System for EM
solutions are not sufficient and, as Rigort et al. (19) mention in their article: ‘‘Prefabrication of [.] frozen-hydrated specimen [.] involves a sequence of transfer and handling steps performed at cryogenic temperatures. Appropriate holders and transfer devices are therefore necessary to optimize the complex workflow. The whole procedure has to meet the essential requirements of any cryo-application, namely maintenance of temperature below the devitrification point of 135 C, avoidance of additional ice (frost) contamination, and prevention of any other potential damage’’.
In this study, a high-vacuum cryo-transfer system that satisfies all the aforesaid demands and overcomes the limitations of the existing systems is presented. The assembly was designed to streamline the handling of frozen-hydrated samples by offering a high turnover of specimens per working day combined with a safe storage capacity for prepared samples. Additionally, the system offers connectivity between different kinds of sample preparation devices or cryo-microscopes. It includes three parts: 1) sample cartridge, 2) storage device, and 3) high-vacuum cryo-shuttle (HVCS) (Fig. 1). The performance of the developed system is demonstrated on an in-lens cryo-scanning transmission electron microscope. To determine the quality of the transfer process, (dehydrated) Tobacco Mosaic virus (TMV) and amorphous carbon foil were transferred under cryogenic conditions. Before and after the transfer process, the mass of the TMV as well as the thickness of the carbon film were measured and compared. Here, any possible contamination would falsify the scattering characteristics of the sample material, and consequently cause an apparent increase in mass or thickness.
759 temperature-controlled cryo-stage, an anti-contaminator, and a docking-station for the high-vacuum cryo-shuttle (Fig.1 b). The entire apparatus is operated at high-vacuum conditions, which simplifies the construction because metal-to-metal flanges are not necessary and fluorocarbon O-rings (made of Viton; DuPont Performance Elastomers, Stow, OH) are sufficient to seal the chamber. To maintain a clean environment, the chamber is bakeable at 150 C and can be flooded by dry nitrogen before the insertion of the transport box. For vacuum generation, the chamber is equipped with a scroll pump (model No. XDS5; Edwards, Albany, NY) and a turbo molecular pump (model No. EXT75DX; Edwards). Being oil-free, the pump system guarantees an environment without carbonaceous contamination. The pump assembly achieves a vacuum environment of 107–108 mbar, which is monitored by a vacuum gauge (model No. ACS2000; Adixen Vacuum Products, Annecy, France). After flooding with dry nitrogen gas, the sample-stage is cooled well below the devitrification temperature by filling it with lN2 and afterwards, the transport box is mounted. Thereafter, the chamber is evacuated and the anti-contaminator (160 C) is moved above the cartridges containing the specimens to avoid ice contamination during the storage. The lN2 level of the reservoir as well as the temperature of the sample-stage are monitored and controlled by a homemade software tool written in LabVIEW (National Instruments, Austin, TX). Herewith, the system can operate independently as long as the lN2 supply is ensured. For details, see Fig. S1.
High-vacuum cryo-shuttle
High-vacuum cryo-transfer system
Once loaded in the storage device, the cartridges can be transferred to any other device equipped with a docking-station via the HVCS (Fig. 1 c). The construction of the HVCS was also kept as simple as possible. The main component is the three-way high-vacuum chamber, which is sealed by fluorocarbon O-rings (made of Viton; DuPont Performance Elastomers), getting a leakage rate of 109 cm3 mbar/s. To prevent ice contamination during the transfer, the chamber is equipped with an anti-contaminator. Directly connected to the cryogen vessel, the anti-contaminator reaches a temperature of ~180 C and therefore guarantees an optimal condensation efficiency. Moreover, the anti-contaminator is designed in such a way that the sample is almost completely surrounded by cold surfaces (see also inset of Fig. 1 c). For sample transfer under vacuum, the HVCS is mounted to the docking-station of the storage device. To ensure easy docking of the shuttle, the docking-station follows a plug-and-play concept. After fixing the HVCS, the cryo-stage of the storage device is tilted by 90 and the cartridge is screwed from the transport box onto the precooled transfer shaft of the HVCS. Thereafter, the shaft is retracted by the actuation magnets into the anti-contaminator in the vacuum chamber of the HVCS and brought in contact with the wall of the lN2 Dewar-vessel wall by magnets. For details, see Fig. S1.
Sample cartridge
Microscope
The concept of the sample cartridge utilized in this study is shown in Fig. 1 a. All parts were made of beryllium bronze or copper for optimal thermal conductivity. As indicated by the arrows, the functionality of the sample cartridge is based on an insert with an outer right-handed thread and an inner left-handed thread, with different pitch sizes. The outer thread is to attach the sample cartridge to the specific cryo-holder of the instrument. The inner thread is required for the dismounting procedure and transfer of the sample cartridge. After the vitrification process, the sample, prepared either on an EM-mesh grid or in a high-pressure freezer carrier, is mounted to the cartridge by fixing it with a clamping ring. Afterwards, the assembly is screwed to a transport box, which, itself, can be mounted to the storage device and can hold up to eight cartridges. For the transfer to the storage device, the transport box is filled with lN2 to guarantee a controlled environment and temperature. For details, see Fig. S1 in the Supporting Material.
The functionality of the HVCS was demonstrated by its use to transfer grids from the storage device to an in-lens, high-resolution scanning electron microscope (model No. S-5000; Hitachi, Tokyo, Japan). The electron microscope is equipped with a homemade annular-dark field (ADF) detector system and is operated in the transmission mode (27). Therefore, it may be denoted as a low-voltage scanning transmission electron microscope with the transfer conditions comparable to any other in-lens system such as the transmission electron microscopes on the market. The detector covers a solid angle ranging from 21 to 161 mrad. All images were taken at 30 keV under low-dose conditions (~300 e/nm2) with a nominal spot size of d ¼ 0.7 nm. For imaging at low temperatures, a modified cryo-holder (model No. CT3500; Gatan, Pleasanton, CA), which can be tilted by 90 to receive the sample cartridge, was utilized. Moreover, a TV camera was installed beside the docking-station, to monitor the transfer process to the cryo-holder. For details, see Fig. S2. The quantitative measurements described in the following sections require a meticulous calibration of the electron microscope. For both measurements the impinging flux of electrons and the fraction scattered toward the detector must be characterized precisely. The calibration of the microscope therefore includes the specification of a reference current to determine the impinging flux of electrons
MATERIALS AND METHODS
Storage device The design of the storage device was kept as simple as possible to ensure that it can easily be adopted by other laboratories. The essential components are the vacuum chamber in combination with vacuum pumps, a moveable and
Biophysical Journal 110(4) 758–765
760
Tacke et al.
a
b
c
FIGURE 1 Schematic of the high-vacuum cryo-transfer system. (a) Sample cartridge. The inner left-handed thread (point 1) corresponds to the thread of the specific holder (see also b, point 1 and c, point 3). The outer right-handed thread (point 2) fits to the thread of the copper front of the shaft (shown in the inset of b). The sample (point 3) is fixed to the cartridge by a clamping ring (point 4). The 3-mm-wide insert can hold either standard EM-mesh grids or a high-pressure freezer carrier. (Upper arrow) Direction of the impinging electrons in the microscope. (b) Storage device. Up to eight cartridges can be mounted in the transport box (point 1), which is fixed by a screw to the temperature-controlled cryo-stage (point 2) of the storage device. To avoid ice contamination during storage, an anti-contaminator (point 3) can be moved above the samples. The level of the lN2 reservoir (point 4) is monitored and automatically maintained. For sample removal, the cryo-stage is rotated by 90 (point 5) and the high-vacuum cryo-shuttle is connected to the plug-and-play docking-station (point 6). (Inset) Magnified view of the cryo-stage of the storage device and the mounted transport box. (c) High-vacuum cryo-shuttle (HVCS). For dismounting the cartridge, the magnetically linked shaft (point 1) of the HVCS is moved inside the chamber of the storage device. The front of the precooled shaft (point 2) is screwed to the cartridge (point 3). Once this is attached, the cartridge can be unscrewed from the transport box and inserted into the chamber of the HVCS (point 4). After the shaft has been retracted, a magnet ensures that the cartridge is connected to the wall of the Dewar-vessel of the HVCS (point 5), ensuring that it remains at a temperature well below the recrystallization temperature of 135 C. The anti-contaminator (point 6) and the high-vacuum environment guarantee a contamination- and artifact-free transfer. The HVCS is docked onto the counterpart of the docking-station (point 7) of the target device. Here, the sample cartridge is screwed to the cryo-holder of the target instrument. (Inset) Cross section of the vacuum chamber and the anti-contaminator of the HVCS. Scale bars , (a) 30 mm; (b) 180 mm; (c) 210 mm. during the measurement, and the detailed characterization of the electron detector. Here, not only the detection cone must be determined but also the counting efficiency must be precisely known (for details of the calibration, see Engel (28)). Due to the meticulous calibration of the microscope, we were able to perform absolute measurements. Consequently, no standards were utilized to calibrate the system.
Preparation of amorphous carbon The amorphous carbon films were prepared by evaporating carbon (model No. P-BBK-061-14; Baltic Pra¨paration, Niesgrau, Germany) onto a glass slide by electron beam evaporation, which was floated onto bidistilled water and spanned across standard 200-mesh copper grids covered with a holey carbon (for details, see Bozzola and Russel (29) and Mu¨ller et al. (30)). For the first thickness estimation, the evaporation flux of the carbon was monitored by a quartz crystal microbalance (model No. SQM-160; Inficon, Bad Ragaz, Switzerland).
Preparation of tobacco mosaic virus TMV strains were kindly provided by Dr. Ruben Diaz-Avalos (Grigorieff Lab, Janelia Research Campus, Ashburn, VA) and purified by Dr. Philippe Biophysical Journal 110(4) 758–765
Ringler (C-CINA; University of Basel, Basel, Switzerland) by gel filtration using Bio-Gel A-5m agarose beads (Bio-Rad, Hercules, CA; for details, see Mu¨ller et al. (30)). Before the final preparation, the EM-grids were rendered hydrophilic by glow-discharging for ~15 s. For lowvoltage-scanning transmission electron microscope investigations, TMV sample solution was mixed with ammonium acetate (100 mM) at a ratio of 1:5. Thereafter, 5 mL were placed on an EM-grid, prepared as described in the previous section, and left to sediment for 2 min. The grids were washed four times with 5–10 mL ammonium acetate (100 mM) and airdried.
Thickness determination EM offers the possibility to measure the sample thickness via the quantification of electron scattering (31,32). To reveal the relationship between thickness and measured signal, the scattering process toward the sample material of known density and composition is simulated by a Monte Carlo software tool (33–35) that was introduced by Reichelt and Engel (36), and extended by Krzyza´nek and Reichelt (37). Having simulated the scattering path of numerous electrons for a range of thicknesses, T, the software tool calculates the fraction of electrons Nsca/N0 that scatter toward the detector for each thickness. Here, Nsca represents the number of electrons counted by the ADF detector and
High-Vacuum Cryo-transfer System for EM
761
N0 the number of impinging electrons. Finally, the results of the simulation are fitted and herewith a function for the thickness determination is revealed:
Nsca T ¼ fMC : N0
(1)
Mass measurements
hmi Nsca ; N0 hsi
(2)
where hsi is the average partial cross section, hmi is the average atomic mass of the sample, and A is the irradiated area (pixel size). Clearly, this approach is only valid for thicknesses much lower than the mean-freepath length l (lprotein(30 keV) z 30 nm), because for thicker samples multiple scattering events dominate. The aforementioned thickness deter-
Cartridge Assembly
Pressure [mbar]
Temperature [°C]
a
M ¼ A T r;
(3)
where r is the density of the sample material. In this study, Eq. 3 was utilized for the mass determination.
Data analysis
The mass measurement was performed on the electron microscope following a technique introduced by Zeitler and Bahr (38), and extended by Wall and Hainfeld (39) and Engel (40). For the single scattering approximation, the mass in every pixel can be calculated by
M ¼ A
mination offers the opportunity to extend the applicable range. In this case, the mass, M, is calculated by
Storage Device
b
For the analysis of the data sets, the MATLAB software package MASDET (The MathWorks, Natick, MA) was used (for details, see Krzyza´nek et al. (41)) in which ADF images are converted into thickness/mass maps. Afterwards, the user identifies regions-of-interest (ROI) by marking them with boxes, typically 90 90 pixels. Each ROI is finally analyzed to determine the thickness distribution or mass-per-length (MPL). For each experiment, the result is represented by the mean and the uncertainty is represented by the standard deviation. TMV measurements were corrected for their dose-dependent mass loss, whereas in the case of the carbon foil the mass loss was neglected. For mass loss correction, several mass loss series were recorded during the experiment to fit each result to the initial mass value before imaging (zero dose). For more details concerning the mass loss correction, see Krzyza´nek et al. (41) and Fig. S3. The results of the
Microscope
Cryo high-vacuum shuttle
c
d
Imaging
e
re-crystallization constant temperature
lN
constant vacuum
lN
FIGURE 2 Transfer of vitrified specimens to the microscope. Individual columns show the most important transfer steps and their corresponding environment, including the temperature and pressure level. Note that the timescales of the different measurements are not equal. Important time frames are highlighted individually. Although vitrification of the sample material was not performed in this study, this preparation step is included in the following discussion because the mounting of the sampling cartridge follows this preparation step. (a) After vitrification, the cartridge is assembled in lN2. (b) Transport of the samples in the transport box filled with lN2 to the precooled cryo-stage of the storage device. (c) Selection of the sample cartridge and its transfer to another instrument utilizing the high-vacuum cryo-shuttle. (d) Mounting the sample cartridge onto the cryo-holder of the microscope or any other instrument equipped with a suitable holder. In this case, the sample cartridge was mounted on the cryo-holder (CT3500, Gatan) of the scanning electron microscope (HR-SEM, S-5000, Hitachi). (e) ADF image of TMV. During imaging, the vacuum and temperature conditions are constant. Additionally, the sample is surrounded by an anticontaminator to avoid ice contamination. Biophysical Journal 110(4) 758–765
762 different experiments were combined by weighting each data set according to its precision and the total error was calculated by applying Gaussian uncertainty propagation; for details, see Taylor (42).
RESULTS AND DISCUSSION A general work flow is shown in Fig. 2. This can be separated into four different steps: vitrification, storage, transfer, and imaging. Although vitrification was not used in this study, it is included in the following discussion because it is an essential step in sample preparation (3,5). During vitrification and the subsequent cartridge assembly, the sample material and the equipment needed are covered with lN2 and therefore recrystallization and contamination can be avoided if the samples are handled carefully (Fig. 2 a). Enclosed by lN2, the transport box is fixed to the precooled sample-stage and the anti-contaminator (T z 160 C) is moved above the sample. Despite the evaporation of the lN2, the temperature of the sample is kept well below the devitrification temperature by active cooling of the cryostage (Fig. 2 b, point 1). After evaporation of the lN2, the vacuum chamber is evacuated and from this time point the sample is permanently in a high-vacuum environment of at least 1 106 mbar (Fig. 2 b, point 2). From here, the samples can be transferred to any instrument equipped with an HVCS docking-station. For the transfer, the sample cartridge is screwed onto the precooled shaft of the HVCS. As indicated in Fig. 2 c, point 3, the heat capacity of the shaft front is sufficient to keep the temperature stable until it is connected to the anti-contaminator of the HVCS, where the temperature stabilizes at ~160 C. For the transfer, the
FIGURE 3 Results of the Monte Carlo simulation. Simulations were performed according to the settings of the microscope (HR-SEM, S-5000, Hitachi). To see this figure in color, go online. Biophysical Journal 110(4) 758–765
Tacke et al.
HVCS must be decoupled from the vacuum system of the storage device (Fig. 2 c, point 4) and although it is just passively pumped by the lN2 cooled inner walls of the Dewar-vessel, a constant vacuum level of 5–6 107 mbar can still be provided. These contamination- and artifact-free conditions are maintained until the lN2 has evaporated (one filling lasts for 40 min). Connected to the microscope, or any other target device, the HVCS is actively pumped by the vacuum system of the microscope, and the vacuum level decreases to ~3 108 mbar during the experiment (Fig. 2 d). To transfer the sample to the cryo-holder of the microscope, the shaft is released from the anti-contaminator (Fig. 2 d, point 5). As active cooling immediately stops, the temperature of the sample cartridge rises until it is screwed onto a cryo-holder, e.g., in a cryo-microscope (Fig. 2 d, point 6; and see Movies S4 and S5). The observed temperature increase is ~10 C during a typical transfer process (in 5 min) and therefore, recrystallization of the amorphous ice can be excluded. During imaging, the temperature and vacuum conditions are maintained for the best structural preservation (Fig. 2 e). Although the high-vacuum conditions are maintained during the entire handling of the sample, the cooled sample may become contaminated because it acts as a condensation trap. To quantify the quality of the transfer process, the mass of TMV and the thickness of the amorphous carbon foil, spanning over the holey carbon film, were investigated. For this purpose, the relationship between thickness and measured signal fMC(Nsca/N0) must be revealed by a Monte Carlo simulation. The simulation was performed for thicknesses between 0 and 300 nm in steps of 0.1 nm utilizing the software package MONCA (37). For each thickness, the electron paths of 500.000 electrons were simulated and analyzed regarding their final scattering angle. Fig. 3 shows the results of the Monte Carlo simulation for carbon (density: 1.6 g/cm3, composition: C ¼ 100%), amorphous ice (density: 0.933 g/cm3, composition: H ¼ 67%, O ¼ 33%), and protein (density: 1.35 g/cm3, composition: H ¼ 49.2%, C ¼ 31.3, N ¼ 9.4%, O ¼ 10.1% (5,36)). By fitting the data, the relationship between signal and thickness is revealed and each pixel can be converted according to Eq. 1 into a thickness value or according to Eq. 3 into a mass value. With this tool in hand, the mass measurements as well as the thickness determinations were performed before and after the cryogenic transfer via HVCS. The idea behind these experiments was to determine a reference value before the transfer procedure and compare these references with the results obtained after the sample transfer under high-vacuum conditions at cryogenic temperatures. Any additional material that condenses onto the sample would increase its mass/ thickness and consequently the difference between the two experiments will show the degree of contamination. Both measurements can be performed on the same electron microscope, which reduces the external influences on the
High-Vacuum Cryo-transfer System for EM
results and simplifies the experimental procedure. Each experiment was repeated at least three times, resulting in ~1000 ROIs for each experiment. For each ROI the MPL was calculated and finally averaged. In the case of the thickness measurements, each pixel was considered to represent one ROI. Areas of carbon were identified and finally all values were averaged. All measurements were performed on the same EM-grid to ensure reproducible thickness and mass values. To keep the influence of mass loss as small as possible, each measurement was performed in different areas of the EM-grid. Herewith, we could avoid that a TMV particle or parts of the amorphous foil were imaged twice. For the first experiment, TMV was chosen as the specimen because its structure and mass distribution are well known (30,43,44) and it is typically utilized as a calibration/reference standard for the MPL measurements via EM (39). Moreover, TMV is the perfect indicator for any ice contamination because it turns hygroscopic if it is completely dehydrated. Air-dried TMV were prepared on EM-grids, assembled in the sample cartridge, and mounted to the sample-stage of the storage device. Thereafter, the samples were transferred at room temperature under highvacuum conditions to the microscope, where the mass of the virus was determined. The sample was transferred and imaged at room temperature in this first experiment to define the reference value for the two following experiments. The gained result of (130 5 5 kDa/nm) matches with the value specified in the literature (131 kDa/nm); see also Krzyza´nek et al. (41). In the second experiment, the mass was measured at (140 C) after the transfer at room temperature to exclude influences of the cryogenic imaging conditions. The result of (131 5 6) kDa/nm clearly shows that an influence of the cryo-conditions during imaging can be excluded. For the final measurement, the samples were precooled within the storage device to 140 C before the transfer process. Transferred within high-vacuum conditions and at cryogenic temperatures, the mass was again measured. The result of (134 5 3) kDa/nm shows on average a tendency for an increased MPL but because the reference value is still within the limits of the uncertainties, a significant contamination can be excluded. Fig. 4 a shows a representative image of a TMV particle and Fig. 4 b a histogram, typically revealed by the analysis of a data set. To test the contamination behavior of a nonhygroscopic specimen, the thickness of amorphous carbon covering the EM-grid was analyzed before and after the transfer via HVCS. Again, the results show no influence of transfer procedure on the thickness of amorphous carbon. The thickness of the carbon foil transferred at room temperature was (4.74 5 0.15) nm, whereas the thickness after transfer under cryogenic conditions was (4.69 5 0.17) nm. Therefore, the thickness measurements exclude the possibility of contamination. However, occasionally ice contamination was observed on some areas of the amorphous carbon. Fig. 5 a shows an example of an EM-grid where ice contamination was found.
763
a
b
FIGURE 4 (a) ADF image of TMV. Here, the sample was precooled to 140 C in the storage device and transferred at low temperature under high-vacuum conditions. Scale bar, 400 nm. (Inset) Magnified part of the 18-nm-wide TMV particle. (b) Histogram resulting from the analysis of 952 ROIs with a size of 90 90 pixels. This experiment yielded a MPL value of (131 5 6) kDa/nm after correction for beam-induced mass loss. To see this figure in color, go online.
As illustrated, ~8% of the grid was contaminated with ice. When contamination became visible, the thickness of the frost particles (leopard pattern) was on average ~3 nm; individual particles could reach a thickness of ~25 nm (Fig. 5, b and c). The majority of the EM-grid was contamination-free and in the areas where there was water deposition, it was thin and well defined, minimally influencing the image because ice layers scatter weakly and do not contribute to any diffraction contrast because water condenses preferentially in the amorphous state (45).
CONCLUSIONS Quantification of the transfer quality clearly shows that our high-vacuum cryo-transfer system is able to streamline the entire handling process for frozen-hydrated sample material by offering a contamination-free environment and the possibility to connect different devices (see Fig. S6). The connection of different instruments is a challenging task: each device has its own specific requirements that must be considered, and consequently, the design of a suitable sample cartridge was one of the key elements of this study. The presented sample cartridge will not only allow simple imaging but the geometry of the cartridge will theoretically also allow the use of the latest postprocessing techniques, e.g., focused-ion-beam milling (14,17), or imaging, e.g., cryoelectron tomography (26). Secondly, we present a solution that streamlines the workflow for frozen hydrated samples. Often, the pre- and postfabrication of vitrified samples is delicate, and well-prepared specimens are precious. Moreover, the cryo-instruments are typically oversubscribed so that samples must be studied during a single time slot. With the software-controlled storage device, numerous samples can be stored for days until imaging. In combination with the high-vacuum cryo-shuttle, a bidirectional transfer of sample material is possible, enabling specimens Biophysical Journal 110(4) 758–765
764
Tacke et al.
FIGURE 5 Assessment of ice contamination. (a) Secondary electron image of a transferred EM-grid. (Large red circle) Size of the EM-grid. (Small red circle) Area where ice contamination was found in this particular case. Comparison of the area framed by the dotted circle and the total available area yields a contamination degree of ~8%. (b) Close-up of ice contamination. This was formed during the transfer of the sample cartridge from the storage device to the microscope. (c) Thickness map of (b). Although ice was found, most of the grid was visibly contamination-free. This was confirmed by the background thickness determination for the carbon foil, as well as by the TMV analysis. Scale bars, (a) 400 mm and (b) 200 nm. To see this figure in color, go online.
to be reinvestigated by not occupying the cryomicroscopes. Consequently, the HVCS defuses the working process and streamlines the experimental procedure. Moreover, the high-vacuum cryo-shuttle offers the possibility to connect different devices and herewith to perform any kind of correlative experiment (see Fig. S7). Taken together, we believe that the presented high-vacuum cryo-transfer system is the first mobile transfer solution, to our knowledge, that not only unifies the benefits mentioned above, but is also applicable for in-lens systems like dedicated STEM and TEM. Therefore, the authors believe that the presented system could enable numerous new kinds of hybrid imaging approaches, like cryo-light and -electron microscopy techniques (23) and other analytical modalities. The authors look forward to building new collaborations and are pleased to facilitate the construction process of interested laboratories to find the optimal solution in every specific case. Therefore, interested readers should not hesitate to contact the authors if there are any questions regarding the presented transfer system.
we are pleased to acknowledge the modifications done by Gatan (cryoholder, CT3500) and Hitachi (HR-SEM, S-5000). This article is dedicated to Prof. Dr. Rudolf Reichelt who initiated this project. Unfortunately, he passed away on October 2, 2010, too early to see the final results. This project was funded by the Deutsche Forschungsgemeinschaft project No. RE782/11. V.K. acknowledges support under grant No. 14-20012S (from Grant Agency of the Czech Republic).
REFERENCES 1. Dukes, M. J., R. Thomas, ., D. F. Kelly. 2014. Improved microchip design and application for in situ transmission electron microscopy of macromolecules. Microsc. Microanal. 20:338–345. 2. Pilhofer, M., K. Aistleitner, ., G. J. Jensen. 2014. Architecture and host interface of environmental chlamydiae revealed by electron cryotomography. Environ. Microbiol. 16:417–429. 3. Echlin, P. 1992. Low-Temperature Microscopy and Analysis. Plenum Publishing, New York. 4. Moor, H., and K. Mu¨hlethaler. 1963. Fine structure in frozen-etched yeast cells. J. Cell Biol. 17:609–628. 5. Dubochet, J., and A. W. McDowall. 1981. Vitrification of pure water for electron microscopy. J. Microsc. 124:3–4.
SUPPORTING MATERIAL
6. Takamori, S., M. Holt, ., R. Jahn. 2006. Molecular anatomy of a trafficking organelle. Cell. 127:831–846.
Five figures and two movies are available at http://www.biophysj.org/ biophysj/supplemental/S0006-3495(16)00101-6.
7. Nicastro, D., A. S. Frangakis, ., W. Baumeister. 2000. Cryo-electron tomography of Neurospora mitochondria. J. Struct. Biol. 129:48–56.
AUTHOR CONTRIBUTIONS
8. Jun, S., D. Ke, ., P. Zhang. 2011. Direct visualization of HIV-1 with correlative live-cell microscopy and cryo-electron tomography. Structure. 19:1573–1581. 9. Milne, J. L. S., and S. Subramaniam. 2009. Cryo-electron tomography of bacteria: progress, challenges and future prospects. Nat. Rev. Microbiol. 7:666–675.
R.R. and R.A.W. initiated the project; R.R., V.K., H.N., and S.T. designed the hardware components; S.T. performed the calibration of the microscope, all measurements, and data analysis; V.K. contributed the analytical tools; and H.N. developed the electronics. All authors wrote the article.
10. Koster, A. J., R. Grimm, ., W. Baumeister. 1997. Perspectives of molecular and cellular electron tomography. J. Struct. Biol. 120:276–308.
ACKNOWLEDGMENTS
11. Medalia, O., I. Weber, ., W. Baumeister. 2002. Macromolecular architecture in eukaryotic cells visualized by cryoelectron tomography. Science. 298:1209–1213.
We thank Martin Wensing from the Fine Mechanical Workshop, Ulrike Keller for help during preparation, Dr. Ruben Diaz-Avalos for kindly providing TMV, and Dr. Philippe Ringler for its purification. Moreover,
12. Parry, A. L., P. H. H. Bomans, ., S. C. G. Biagini. 2008. Cryo electron tomography reveals confined complex morphologies of tripeptide-containing amphiphilic double-comb diblock copolymers. Angew. Chem. Int. Ed. Engl. 47:8859–8862.
Biophysical Journal 110(4) 758–765
High-Vacuum Cryo-transfer System for EM
765
13. Al-Amoudi, A., L. P. O. Norlen, and J. Dubochet. 2004. Cryo-electron microscopy of vitreous sections of native biological cells and tissues. J. Struct. Biol. 148:131–135.
30. Mu¨ller, S. A., K. N. Goldie, ., A. Engel. 1992. Factors influencing the precision of quantitative scanning transmission electron microscopy. Ultramicroscopy. 46:317–334.
14. Marko, M., C. Hsieh, ., C. Mannella. 2007. Focused-ion-beam thinning of frozen-hydrated biological specimens for cryo-electron microscopy. Nat. Methods. 4:215–217. 15. Rubino, S., S. Akhtar, ., K. Leifer. 2012. A site-specific focused-ionbeam lift-out method for cryo transmission electron microscopy. J. Struct. Biol. 180:572–576.
31. Pfaff, M., E. Mu¨ller, ., D. Gerthsen. 2011. Low-energy electron scattering in carbon-based materials analyzed by scanning transmission electron microscopy and its application to sample thickness determination. J. Microsc. 243:31–39.
16. Hayles, M. F., D. A. M. de Winter, ., A. J. Verkleij. 2010. The making of frozen-hydrated, vitreous lamellas from cells for cryo-electron microscopy. J. Struct. Biol. 172:180–190. 17. Villa, E., M. Schaffer, ., W. Baumeister. 2013. Opening windows into the cell: focused-ion-beam milling for cryo-electron tomography. Curr. Opin. Struct. Biol. 23:771–777. 18. Severs, N. J. 2007. Freeze-fracture electron microscopy. Nat. Protoc. 2:547–576. 19. Rigort, A., F. J. B. Ba¨uerlein, ., J. M. Plitzko. 2010. Micromachining tools and correlative approaches for cellular cryo-electron tomography. J. Struct. Biol. 172:169–179. 20. Hagen, C., P. Guttmann, ., K. Gru¨newald. 2012. Correlative VIS-fluorescence and soft X-ray cryo-microscopy/tomography of adherent cells. J. Struct. Biol. 177:193–201.
32. Volkenandt, T., E. Mu¨ller, and D. Gerthsen. 2014. Sample thickness determination by scanning transmission electron microscopy at low electron energies. Microsc. Microanal. 20:111–123. 33. Ritchie, N. W. M. 2005. A new Monte Carlo application for complex sample geometries. Surf. Interface Anal. 37:1006–1011. 34. Drouin, D., A. R. Couture, ., R. Gauvin. 2007. CASINO V2.42: a fast and easy-to-use modeling tool for scanning electron microscopy and microanalysis users. Scanning. 29:92–101. 35. Demers, H., N. Poirier-Demers, ., D. Drouin. 2011. Three-dimensional electron microscopy simulation with the CASINO Monte Carlo software. Scanning. 33:135–146. 36. Reichelt, R., and A. Engel. 1984. Monte Carlo calculations of elastic and inelastic electron scattering in biological and plastic materials. Ultramicroscopy. 13:279–293.
21. Kaufmann, R., P. Schellenberger, ., K. Gru¨newald. 2014. Super-resolution microscopy using standard fluorescent proteins in intact cells under cryo-conditions. Nano Lett. 14:4171–4175.
37. Krzyza´nek, V., and R. Reichelt. 2003. MONCA: a new MATLAB package for Monte Carlo simulation of electron scattering in thin specimens in the energy range 10–200 keV. Microsc. Microanal. 9 (S03):110–111.
22. Weisenburger, S., B. Jing, ., V. Sandoghdar. 2014. Cryogenic colocalization microscopy for nanometer-distance measurements. ChemPhysChem. 15:763–770.
38. Zeitler, E., and G. F. Bahr. 1962. A Photometric procedure for weight determination of submicroscopic particles quantitative electron microscopy. J. Appl. Phys. 33:847–853.
23. Chang, Y. W., S. Chen, ., G. J. Jensen. 2014. Correlated cryogenic photoactivated localization microscopy and cryo-electron tomography. Nat. Methods. 11:737–739.
39. Wall, J. S., and J. F. Hainfeld. 1986. Mass mapping with the scanning transmission electron microscope. Annu. Rev. Biophys. Biophys. Chem. 15:355–376.
24. Capaccioli, S., and K. L. Ngai. 2011. Resolving the controversy on the glass transition temperature of water? J. Chem. Phys. 135:104504.
40. Engel, A. 1978. Molecular weight determination by scanning transmission electron microscopy. Ultramicroscopy. 3:273–281.
25. Amann-Winkel, K., C. Gainaru, ., T. Loerting. 2013. Water’s second glass transition. Proc. Natl. Acad. Sci. USA. 110:17720–17725. 26. Wolf, S. G., L. Houben, and M. Elbaum. 2014. Cryo-scanning transmission electron tomography of vitrified cells. Nat. Methods. 11: 423–428.
41. Krzyza´nek, V., S. A. Mu¨ller, ., R. Reichelt. 2009. MASDET—a fast and user-friendly multiplatform software for mass determination by dark-field electron microscopy. J. Struct. Biol. 165:78–87. 42. Taylor, J. R. 1996. An Introduction to Error Analysis. University Science Books, Herndon, VA.
27. Krzyzanek, V., and R. Reichelt. 2007. High-resolution scanning electron microscope for mass determination: progress in the development of a new nanoanalytical tool. Microsc. Microanal. 13 (S03):80–81.
43. Namba, K., and G. Stubbs. 1986. Structure of tobacco mosaic virus at ˚ resolution: implications for assembly. Science. 231:1401–1406. 3.6 A
28. Engel, A. 2009. Scanning transmission electron microscopy: biological applications. Adv. Imaging Electron Phys. 159:357–386.
44. Dı´az-Avalos, R., and D. L. Caspar. 1998. Structure of the stacked disk aggregate of tobacco mosaic virus protein. Biophys. J. 74:595–603.
29. Bozzola, J. J., and L. D. Russel. 1992. Electron Microscopy—Principles and Techniques for Biologists. Jones & Bartlett, Sudbury, MA.
45. Angell, C. A. 2004. Amorphous water. Annu. Rev. Phys. Chem. 55:559–583.
Biophysical Journal 110(4) 758–765
Biophysical Journal
Volume 110
February 2016
758–765
Article A Versatile High-Vacuum Cryo-transfer System for Cryo-microscopy and Analytics Sebastian Tacke,1,* Vladislav Krzyzanek,2 Harald Nu¨sse,1 Roger Albert Wepf,3 Ju¨rgen Klingauf,1 and Rudolf Reichelt1 1 Institute of Medical Physics and Biophysics, University of Mu¨nster, Mu¨nster, Germany; 2Institute of Scientific Instruments, The Czech Academy of Sciences, Brno, Czech Republic; and 3Scientific Center for Optical and Electron Microscopy, ETH Zu¨rich, Zu¨rich, Switzerland
ABSTRACT Cryogenic microscopy methods have gained increasing popularity, as they offer an unaltered view on the architecture of biological specimens. As a prerequisite, samples must be handled under cryogenic conditions below their recrystallization temperature, and contamination during sample transfer and handling must be prevented. We present a high-vacuum cryo-transfer system that streamlines the entire handling of frozen-hydrated samples from the vitrification process to low temperature imaging for scanning transmission electron microscopy and transmission electron microscopy. A template for cryo-electron microscopy and multimodal cryo-imaging approaches with numerous sample transfer steps is presented.
INTRODUCTION Preserving the native state until imaging is the final goal in sample preparation. Particularly for electron microscopy (EM), the conservation of the pristine architecture of hydrated samples is challenging. If not imaged in situ (1), two different strategies can be followed to prepare hydrated samples for EM: the conventional or the cryogenic routine. Conventional preparation protocols often include chemical fixation procedures as well as several staining steps, and samples are allowed to dry. The chemical processes are known to cause rearrangement or loss of cytoplasmic components, or agglomeration to membranes (2,3). Moreover, final staining with heavy elements offers just an indirect, replicalike illustration of the treated structures. In the 1960s to 1980s, Moor and Mu¨hlethaler (4) and Dubochet and McDowall (5) introduced the cryogenic preparation routine, which circumvents such drawbacks. Although Moor and Mu¨hlethaler (4) and Dubochet and McDowall (5) followed different postprocessing and imaging routines (freeze-etching versus direct imaging, respectively), the basic principle of both preparation procedures is the same: the physical arresting of the sample in a frozen-hydrated state, which is achieved by the vitrification of the sample within milliseconds, resulting in a sample embedded in amorphous ice, thereby avoiding structural changes due to ice crystal growth. The convincing structural preservation achieved in the case of isolated cell organelles (6,7), viruses (8), bacteria (9,10), eukaryotic cells (11) and other softmatter applications (12), and the technical improvements in sample preparation and handling (13–19), caused this
Submitted July 1, 2015, and accepted for publication January 20, 2016. *Correspondence: [email protected]
approach to spread out into other fields of microscopy like soft x-ray (20) and light microscopy (21,22), and enabled new hybrid approaches like correlative cryo-light and electron microscopy (23). However, the handling of vitrified samples is rather delicate. If imaged in the frozen-hydrated state, for example, the temperature of the specimen must be maintained below the glass transition temperature of water, 138 K (135 C) (5); for further details, see Capaccioli and Ngai (24) and Amann-Winkel et al. (25). As a consequence, the unprotected sample acts as a cold trap and is therefore susceptible to any kind of contamination. Samples that were partially or totally dehydrated, e.g., by freeze-drying, also require a gentle handling. Being dehydrated, the sample material turns hygroscopic and therefore must be transported in an anhydrous environment (3). Thus, the sample must be handled in liquid nitrogen (lN2), in a dry inert gas atmosphere or in a clean vacuum environment under controlled temperature conditions at all times. Most available transfer systems are either cryo-transfer systems or high-vacuum transfer systems. They typically require the installation of a specific cryo-stage inside the microscope, which is not feasible for in-lens systems, such as transmission electron microscopes or scanning transmission electron microscopes, due to the limited space in their specimen chambers. The latest electron microscopes are equipped with storage devices directly connected to the column, allowing samples to be stored and transferred to the microscope stage under high-vacuum environment at cryogenic conditions. However, transfer to other devices for further processing or analysis is still problematic, as the high-vacuum conditions have to be broken. Latest preparation protocols (17) and imaging techniques (26) include several transfer steps. The available
Editor: Andreas Engel. Ó 2016 by the Biophysical Society 0006-3495/16/02/0758/8
http://dx.doi.org/10.1016/j.bpj.2016.01.024
High-Vacuum Cryo-transfer System for EM
solutions are not sufficient and, as Rigort et al. (19) mention in their article: ‘‘Prefabrication of [.] frozen-hydrated specimen [.] involves a sequence of transfer and handling steps performed at cryogenic temperatures. Appropriate holders and transfer devices are therefore necessary to optimize the complex workflow. The whole procedure has to meet the essential requirements of any cryo-application, namely maintenance of temperature below the devitrification point of 135 C, avoidance of additional ice (frost) contamination, and prevention of any other potential damage’’.
In this study, a high-vacuum cryo-transfer system that satisfies all the aforesaid demands and overcomes the limitations of the existing systems is presented. The assembly was designed to streamline the handling of frozen-hydrated samples by offering a high turnover of specimens per working day combined with a safe storage capacity for prepared samples. Additionally, the system offers connectivity between different kinds of sample preparation devices or cryo-microscopes. It includes three parts: 1) sample cartridge, 2) storage device, and 3) high-vacuum cryo-shuttle (HVCS) (Fig. 1). The performance of the developed system is demonstrated on an in-lens cryo-scanning transmission electron microscope. To determine the quality of the transfer process, (dehydrated) Tobacco Mosaic virus (TMV) and amorphous carbon foil were transferred under cryogenic conditions. Before and after the transfer process, the mass of the TMV as well as the thickness of the carbon film were measured and compared. Here, any possible contamination would falsify the scattering characteristics of the sample material, and consequently cause an apparent increase in mass or thickness.
759 temperature-controlled cryo-stage, an anti-contaminator, and a docking-station for the high-vacuum cryo-shuttle (Fig.1 b). The entire apparatus is operated at high-vacuum conditions, which simplifies the construction because metal-to-metal flanges are not necessary and fluorocarbon O-rings (made of Viton; DuPont Performance Elastomers, Stow, OH) are sufficient to seal the chamber. To maintain a clean environment, the chamber is bakeable at 150 C and can be flooded by dry nitrogen before the insertion of the transport box. For vacuum generation, the chamber is equipped with a scroll pump (model No. XDS5; Edwards, Albany, NY) and a turbo molecular pump (model No. EXT75DX; Edwards). Being oil-free, the pump system guarantees an environment without carbonaceous contamination. The pump assembly achieves a vacuum environment of 107–108 mbar, which is monitored by a vacuum gauge (model No. ACS2000; Adixen Vacuum Products, Annecy, France). After flooding with dry nitrogen gas, the sample-stage is cooled well below the devitrification temperature by filling it with lN2 and afterwards, the transport box is mounted. Thereafter, the chamber is evacuated and the anti-contaminator (160 C) is moved above the cartridges containing the specimens to avoid ice contamination during the storage. The lN2 level of the reservoir as well as the temperature of the sample-stage are monitored and controlled by a homemade software tool written in LabVIEW (National Instruments, Austin, TX). Herewith, the system can operate independently as long as the lN2 supply is ensured. For details, see Fig. S1.
High-vacuum cryo-shuttle
High-vacuum cryo-transfer system
Once loaded in the storage device, the cartridges can be transferred to any other device equipped with a docking-station via the HVCS (Fig. 1 c). The construction of the HVCS was also kept as simple as possible. The main component is the three-way high-vacuum chamber, which is sealed by fluorocarbon O-rings (made of Viton; DuPont Performance Elastomers), getting a leakage rate of 109 cm3 mbar/s. To prevent ice contamination during the transfer, the chamber is equipped with an anti-contaminator. Directly connected to the cryogen vessel, the anti-contaminator reaches a temperature of ~180 C and therefore guarantees an optimal condensation efficiency. Moreover, the anti-contaminator is designed in such a way that the sample is almost completely surrounded by cold surfaces (see also inset of Fig. 1 c). For sample transfer under vacuum, the HVCS is mounted to the docking-station of the storage device. To ensure easy docking of the shuttle, the docking-station follows a plug-and-play concept. After fixing the HVCS, the cryo-stage of the storage device is tilted by 90 and the cartridge is screwed from the transport box onto the precooled transfer shaft of the HVCS. Thereafter, the shaft is retracted by the actuation magnets into the anti-contaminator in the vacuum chamber of the HVCS and brought in contact with the wall of the lN2 Dewar-vessel wall by magnets. For details, see Fig. S1.
Sample cartridge
Microscope
The concept of the sample cartridge utilized in this study is shown in Fig. 1 a. All parts were made of beryllium bronze or copper for optimal thermal conductivity. As indicated by the arrows, the functionality of the sample cartridge is based on an insert with an outer right-handed thread and an inner left-handed thread, with different pitch sizes. The outer thread is to attach the sample cartridge to the specific cryo-holder of the instrument. The inner thread is required for the dismounting procedure and transfer of the sample cartridge. After the vitrification process, the sample, prepared either on an EM-mesh grid or in a high-pressure freezer carrier, is mounted to the cartridge by fixing it with a clamping ring. Afterwards, the assembly is screwed to a transport box, which, itself, can be mounted to the storage device and can hold up to eight cartridges. For the transfer to the storage device, the transport box is filled with lN2 to guarantee a controlled environment and temperature. For details, see Fig. S1 in the Supporting Material.
The functionality of the HVCS was demonstrated by its use to transfer grids from the storage device to an in-lens, high-resolution scanning electron microscope (model No. S-5000; Hitachi, Tokyo, Japan). The electron microscope is equipped with a homemade annular-dark field (ADF) detector system and is operated in the transmission mode (27). Therefore, it may be denoted as a low-voltage scanning transmission electron microscope with the transfer conditions comparable to any other in-lens system such as the transmission electron microscopes on the market. The detector covers a solid angle ranging from 21 to 161 mrad. All images were taken at 30 keV under low-dose conditions (~300 e/nm2) with a nominal spot size of d ¼ 0.7 nm. For imaging at low temperatures, a modified cryo-holder (model No. CT3500; Gatan, Pleasanton, CA), which can be tilted by 90 to receive the sample cartridge, was utilized. Moreover, a TV camera was installed beside the docking-station, to monitor the transfer process to the cryo-holder. For details, see Fig. S2. The quantitative measurements described in the following sections require a meticulous calibration of the electron microscope. For both measurements the impinging flux of electrons and the fraction scattered toward the detector must be characterized precisely. The calibration of the microscope therefore includes the specification of a reference current to determine the impinging flux of electrons
MATERIALS AND METHODS
Storage device The design of the storage device was kept as simple as possible to ensure that it can easily be adopted by other laboratories. The essential components are the vacuum chamber in combination with vacuum pumps, a moveable and
Biophysical Journal 110(4) 758–765
760
Tacke et al.
a
b
c
FIGURE 1 Schematic of the high-vacuum cryo-transfer system. (a) Sample cartridge. The inner left-handed thread (point 1) corresponds to the thread of the specific holder (see also b, point 1 and c, point 3). The outer right-handed thread (point 2) fits to the thread of the copper front of the shaft (shown in the inset of b). The sample (point 3) is fixed to the cartridge by a clamping ring (point 4). The 3-mm-wide insert can hold either standard EM-mesh grids or a high-pressure freezer carrier. (Upper arrow) Direction of the impinging electrons in the microscope. (b) Storage device. Up to eight cartridges can be mounted in the transport box (point 1), which is fixed by a screw to the temperature-controlled cryo-stage (point 2) of the storage device. To avoid ice contamination during storage, an anti-contaminator (point 3) can be moved above the samples. The level of the lN2 reservoir (point 4) is monitored and automatically maintained. For sample removal, the cryo-stage is rotated by 90 (point 5) and the high-vacuum cryo-shuttle is connected to the plug-and-play docking-station (point 6). (Inset) Magnified view of the cryo-stage of the storage device and the mounted transport box. (c) High-vacuum cryo-shuttle (HVCS). For dismounting the cartridge, the magnetically linked shaft (point 1) of the HVCS is moved inside the chamber of the storage device. The front of the precooled shaft (point 2) is screwed to the cartridge (point 3). Once this is attached, the cartridge can be unscrewed from the transport box and inserted into the chamber of the HVCS (point 4). After the shaft has been retracted, a magnet ensures that the cartridge is connected to the wall of the Dewar-vessel of the HVCS (point 5), ensuring that it remains at a temperature well below the recrystallization temperature of 135 C. The anti-contaminator (point 6) and the high-vacuum environment guarantee a contamination- and artifact-free transfer. The HVCS is docked onto the counterpart of the docking-station (point 7) of the target device. Here, the sample cartridge is screwed to the cryo-holder of the target instrument. (Inset) Cross section of the vacuum chamber and the anti-contaminator of the HVCS. Scale bars , (a) 30 mm; (b) 180 mm; (c) 210 mm. during the measurement, and the detailed characterization of the electron detector. Here, not only the detection cone must be determined but also the counting efficiency must be precisely known (for details of the calibration, see Engel (28)). Due to the meticulous calibration of the microscope, we were able to perform absolute measurements. Consequently, no standards were utilized to calibrate the system.
Preparation of amorphous carbon The amorphous carbon films were prepared by evaporating carbon (model No. P-BBK-061-14; Baltic Pra¨paration, Niesgrau, Germany) onto a glass slide by electron beam evaporation, which was floated onto bidistilled water and spanned across standard 200-mesh copper grids covered with a holey carbon (for details, see Bozzola and Russel (29) and Mu¨ller et al. (30)). For the first thickness estimation, the evaporation flux of the carbon was monitored by a quartz crystal microbalance (model No. SQM-160; Inficon, Bad Ragaz, Switzerland).
Preparation of tobacco mosaic virus TMV strains were kindly provided by Dr. Ruben Diaz-Avalos (Grigorieff Lab, Janelia Research Campus, Ashburn, VA) and purified by Dr. Philippe Biophysical Journal 110(4) 758–765
Ringler (C-CINA; University of Basel, Basel, Switzerland) by gel filtration using Bio-Gel A-5m agarose beads (Bio-Rad, Hercules, CA; for details, see Mu¨ller et al. (30)). Before the final preparation, the EM-grids were rendered hydrophilic by glow-discharging for ~15 s. For lowvoltage-scanning transmission electron microscope investigations, TMV sample solution was mixed with ammonium acetate (100 mM) at a ratio of 1:5. Thereafter, 5 mL were placed on an EM-grid, prepared as described in the previous section, and left to sediment for 2 min. The grids were washed four times with 5–10 mL ammonium acetate (100 mM) and airdried.
Thickness determination EM offers the possibility to measure the sample thickness via the quantification of electron scattering (31,32). To reveal the relationship between thickness and measured signal, the scattering process toward the sample material of known density and composition is simulated by a Monte Carlo software tool (33–35) that was introduced by Reichelt and Engel (36), and extended by Krzyza´nek and Reichelt (37). Having simulated the scattering path of numerous electrons for a range of thicknesses, T, the software tool calculates the fraction of electrons Nsca/N0 that scatter toward the detector for each thickness. Here, Nsca represents the number of electrons counted by the ADF detector and
High-Vacuum Cryo-transfer System for EM
761
N0 the number of impinging electrons. Finally, the results of the simulation are fitted and herewith a function for the thickness determination is revealed:
Nsca T ¼ fMC : N0
(1)
Mass measurements
hmi Nsca ; N0 hsi
(2)
where hsi is the average partial cross section, hmi is the average atomic mass of the sample, and A is the irradiated area (pixel size). Clearly, this approach is only valid for thicknesses much lower than the mean-freepath length l (lprotein(30 keV) z 30 nm), because for thicker samples multiple scattering events dominate. The aforementioned thickness deter-
Cartridge Assembly
Pressure [mbar]
Temperature [°C]
a
M ¼ A T r;
(3)
where r is the density of the sample material. In this study, Eq. 3 was utilized for the mass determination.
Data analysis
The mass measurement was performed on the electron microscope following a technique introduced by Zeitler and Bahr (38), and extended by Wall and Hainfeld (39) and Engel (40). For the single scattering approximation, the mass in every pixel can be calculated by
M ¼ A
mination offers the opportunity to extend the applicable range. In this case, the mass, M, is calculated by
Storage Device
b
For the analysis of the data sets, the MATLAB software package MASDET (The MathWorks, Natick, MA) was used (for details, see Krzyza´nek et al. (41)) in which ADF images are converted into thickness/mass maps. Afterwards, the user identifies regions-of-interest (ROI) by marking them with boxes, typically 90 90 pixels. Each ROI is finally analyzed to determine the thickness distribution or mass-per-length (MPL). For each experiment, the result is represented by the mean and the uncertainty is represented by the standard deviation. TMV measurements were corrected for their dose-dependent mass loss, whereas in the case of the carbon foil the mass loss was neglected. For mass loss correction, several mass loss series were recorded during the experiment to fit each result to the initial mass value before imaging (zero dose). For more details concerning the mass loss correction, see Krzyza´nek et al. (41) and Fig. S3. The results of the
Microscope
Cryo high-vacuum shuttle
c
d
Imaging
e
re-crystallization constant temperature
lN
constant vacuum
lN
FIGURE 2 Transfer of vitrified specimens to the microscope. Individual columns show the most important transfer steps and their corresponding environment, including the temperature and pressure level. Note that the timescales of the different measurements are not equal. Important time frames are highlighted individually. Although vitrification of the sample material was not performed in this study, this preparation step is included in the following discussion because the mounting of the sampling cartridge follows this preparation step. (a) After vitrification, the cartridge is assembled in lN2. (b) Transport of the samples in the transport box filled with lN2 to the precooled cryo-stage of the storage device. (c) Selection of the sample cartridge and its transfer to another instrument utilizing the high-vacuum cryo-shuttle. (d) Mounting the sample cartridge onto the cryo-holder of the microscope or any other instrument equipped with a suitable holder. In this case, the sample cartridge was mounted on the cryo-holder (CT3500, Gatan) of the scanning electron microscope (HR-SEM, S-5000, Hitachi). (e) ADF image of TMV. During imaging, the vacuum and temperature conditions are constant. Additionally, the sample is surrounded by an anticontaminator to avoid ice contamination. Biophysical Journal 110(4) 758–765
762 different experiments were combined by weighting each data set according to its precision and the total error was calculated by applying Gaussian uncertainty propagation; for details, see Taylor (42).
RESULTS AND DISCUSSION A general work flow is shown in Fig. 2. This can be separated into four different steps: vitrification, storage, transfer, and imaging. Although vitrification was not used in this study, it is included in the following discussion because it is an essential step in sample preparation (3,5). During vitrification and the subsequent cartridge assembly, the sample material and the equipment needed are covered with lN2 and therefore recrystallization and contamination can be avoided if the samples are handled carefully (Fig. 2 a). Enclosed by lN2, the transport box is fixed to the precooled sample-stage and the anti-contaminator (T z 160 C) is moved above the sample. Despite the evaporation of the lN2, the temperature of the sample is kept well below the devitrification temperature by active cooling of the cryostage (Fig. 2 b, point 1). After evaporation of the lN2, the vacuum chamber is evacuated and from this time point the sample is permanently in a high-vacuum environment of at least 1 106 mbar (Fig. 2 b, point 2). From here, the samples can be transferred to any instrument equipped with an HVCS docking-station. For the transfer, the sample cartridge is screwed onto the precooled shaft of the HVCS. As indicated in Fig. 2 c, point 3, the heat capacity of the shaft front is sufficient to keep the temperature stable until it is connected to the anti-contaminator of the HVCS, where the temperature stabilizes at ~160 C. For the transfer, the
FIGURE 3 Results of the Monte Carlo simulation. Simulations were performed according to the settings of the microscope (HR-SEM, S-5000, Hitachi). To see this figure in color, go online. Biophysical Journal 110(4) 758–765
Tacke et al.
HVCS must be decoupled from the vacuum system of the storage device (Fig. 2 c, point 4) and although it is just passively pumped by the lN2 cooled inner walls of the Dewar-vessel, a constant vacuum level of 5–6 107 mbar can still be provided. These contamination- and artifact-free conditions are maintained until the lN2 has evaporated (one filling lasts for 40 min). Connected to the microscope, or any other target device, the HVCS is actively pumped by the vacuum system of the microscope, and the vacuum level decreases to ~3 108 mbar during the experiment (Fig. 2 d). To transfer the sample to the cryo-holder of the microscope, the shaft is released from the anti-contaminator (Fig. 2 d, point 5). As active cooling immediately stops, the temperature of the sample cartridge rises until it is screwed onto a cryo-holder, e.g., in a cryo-microscope (Fig. 2 d, point 6; and see Movies S4 and S5). The observed temperature increase is ~10 C during a typical transfer process (in 5 min) and therefore, recrystallization of the amorphous ice can be excluded. During imaging, the temperature and vacuum conditions are maintained for the best structural preservation (Fig. 2 e). Although the high-vacuum conditions are maintained during the entire handling of the sample, the cooled sample may become contaminated because it acts as a condensation trap. To quantify the quality of the transfer process, the mass of TMV and the thickness of the amorphous carbon foil, spanning over the holey carbon film, were investigated. For this purpose, the relationship between thickness and measured signal fMC(Nsca/N0) must be revealed by a Monte Carlo simulation. The simulation was performed for thicknesses between 0 and 300 nm in steps of 0.1 nm utilizing the software package MONCA (37). For each thickness, the electron paths of 500.000 electrons were simulated and analyzed regarding their final scattering angle. Fig. 3 shows the results of the Monte Carlo simulation for carbon (density: 1.6 g/cm3, composition: C ¼ 100%), amorphous ice (density: 0.933 g/cm3, composition: H ¼ 67%, O ¼ 33%), and protein (density: 1.35 g/cm3, composition: H ¼ 49.2%, C ¼ 31.3, N ¼ 9.4%, O ¼ 10.1% (5,36)). By fitting the data, the relationship between signal and thickness is revealed and each pixel can be converted according to Eq. 1 into a thickness value or according to Eq. 3 into a mass value. With this tool in hand, the mass measurements as well as the thickness determinations were performed before and after the cryogenic transfer via HVCS. The idea behind these experiments was to determine a reference value before the transfer procedure and compare these references with the results obtained after the sample transfer under high-vacuum conditions at cryogenic temperatures. Any additional material that condenses onto the sample would increase its mass/ thickness and consequently the difference between the two experiments will show the degree of contamination. Both measurements can be performed on the same electron microscope, which reduces the external influences on the
High-Vacuum Cryo-transfer System for EM
results and simplifies the experimental procedure. Each experiment was repeated at least three times, resulting in ~1000 ROIs for each experiment. For each ROI the MPL was calculated and finally averaged. In the case of the thickness measurements, each pixel was considered to represent one ROI. Areas of carbon were identified and finally all values were averaged. All measurements were performed on the same EM-grid to ensure reproducible thickness and mass values. To keep the influence of mass loss as small as possible, each measurement was performed in different areas of the EM-grid. Herewith, we could avoid that a TMV particle or parts of the amorphous foil were imaged twice. For the first experiment, TMV was chosen as the specimen because its structure and mass distribution are well known (30,43,44) and it is typically utilized as a calibration/reference standard for the MPL measurements via EM (39). Moreover, TMV is the perfect indicator for any ice contamination because it turns hygroscopic if it is completely dehydrated. Air-dried TMV were prepared on EM-grids, assembled in the sample cartridge, and mounted to the sample-stage of the storage device. Thereafter, the samples were transferred at room temperature under highvacuum conditions to the microscope, where the mass of the virus was determined. The sample was transferred and imaged at room temperature in this first experiment to define the reference value for the two following experiments. The gained result of (130 5 5 kDa/nm) matches with the value specified in the literature (131 kDa/nm); see also Krzyza´nek et al. (41). In the second experiment, the mass was measured at (140 C) after the transfer at room temperature to exclude influences of the cryogenic imaging conditions. The result of (131 5 6) kDa/nm clearly shows that an influence of the cryo-conditions during imaging can be excluded. For the final measurement, the samples were precooled within the storage device to 140 C before the transfer process. Transferred within high-vacuum conditions and at cryogenic temperatures, the mass was again measured. The result of (134 5 3) kDa/nm shows on average a tendency for an increased MPL but because the reference value is still within the limits of the uncertainties, a significant contamination can be excluded. Fig. 4 a shows a representative image of a TMV particle and Fig. 4 b a histogram, typically revealed by the analysis of a data set. To test the contamination behavior of a nonhygroscopic specimen, the thickness of amorphous carbon covering the EM-grid was analyzed before and after the transfer via HVCS. Again, the results show no influence of transfer procedure on the thickness of amorphous carbon. The thickness of the carbon foil transferred at room temperature was (4.74 5 0.15) nm, whereas the thickness after transfer under cryogenic conditions was (4.69 5 0.17) nm. Therefore, the thickness measurements exclude the possibility of contamination. However, occasionally ice contamination was observed on some areas of the amorphous carbon. Fig. 5 a shows an example of an EM-grid where ice contamination was found.
763
a
b
FIGURE 4 (a) ADF image of TMV. Here, the sample was precooled to 140 C in the storage device and transferred at low temperature under high-vacuum conditions. Scale bar, 400 nm. (Inset) Magnified part of the 18-nm-wide TMV particle. (b) Histogram resulting from the analysis of 952 ROIs with a size of 90 90 pixels. This experiment yielded a MPL value of (131 5 6) kDa/nm after correction for beam-induced mass loss. To see this figure in color, go online.
As illustrated, ~8% of the grid was contaminated with ice. When contamination became visible, the thickness of the frost particles (leopard pattern) was on average ~3 nm; individual particles could reach a thickness of ~25 nm (Fig. 5, b and c). The majority of the EM-grid was contamination-free and in the areas where there was water deposition, it was thin and well defined, minimally influencing the image because ice layers scatter weakly and do not contribute to any diffraction contrast because water condenses preferentially in the amorphous state (45).
CONCLUSIONS Quantification of the transfer quality clearly shows that our high-vacuum cryo-transfer system is able to streamline the entire handling process for frozen-hydrated sample material by offering a contamination-free environment and the possibility to connect different devices (see Fig. S6). The connection of different instruments is a challenging task: each device has its own specific requirements that must be considered, and consequently, the design of a suitable sample cartridge was one of the key elements of this study. The presented sample cartridge will not only allow simple imaging but the geometry of the cartridge will theoretically also allow the use of the latest postprocessing techniques, e.g., focused-ion-beam milling (14,17), or imaging, e.g., cryoelectron tomography (26). Secondly, we present a solution that streamlines the workflow for frozen hydrated samples. Often, the pre- and postfabrication of vitrified samples is delicate, and well-prepared specimens are precious. Moreover, the cryo-instruments are typically oversubscribed so that samples must be studied during a single time slot. With the software-controlled storage device, numerous samples can be stored for days until imaging. In combination with the high-vacuum cryo-shuttle, a bidirectional transfer of sample material is possible, enabling specimens Biophysical Journal 110(4) 758–765
764
Tacke et al.
FIGURE 5 Assessment of ice contamination. (a) Secondary electron image of a transferred EM-grid. (Large red circle) Size of the EM-grid. (Small red circle) Area where ice contamination was found in this particular case. Comparison of the area framed by the dotted circle and the total available area yields a contamination degree of ~8%. (b) Close-up of ice contamination. This was formed during the transfer of the sample cartridge from the storage device to the microscope. (c) Thickness map of (b). Although ice was found, most of the grid was visibly contamination-free. This was confirmed by the background thickness determination for the carbon foil, as well as by the TMV analysis. Scale bars, (a) 400 mm and (b) 200 nm. To see this figure in color, go online.
to be reinvestigated by not occupying the cryomicroscopes. Consequently, the HVCS defuses the working process and streamlines the experimental procedure. Moreover, the high-vacuum cryo-shuttle offers the possibility to connect different devices and herewith to perform any kind of correlative experiment (see Fig. S7). Taken together, we believe that the presented high-vacuum cryo-transfer system is the first mobile transfer solution, to our knowledge, that not only unifies the benefits mentioned above, but is also applicable for in-lens systems like dedicated STEM and TEM. Therefore, the authors believe that the presented system could enable numerous new kinds of hybrid imaging approaches, like cryo-light and -electron microscopy techniques (23) and other analytical modalities. The authors look forward to building new collaborations and are pleased to facilitate the construction process of interested laboratories to find the optimal solution in every specific case. Therefore, interested readers should not hesitate to contact the authors if there are any questions regarding the presented transfer system.
we are pleased to acknowledge the modifications done by Gatan (cryoholder, CT3500) and Hitachi (HR-SEM, S-5000). This article is dedicated to Prof. Dr. Rudolf Reichelt who initiated this project. Unfortunately, he passed away on October 2, 2010, too early to see the final results. This project was funded by the Deutsche Forschungsgemeinschaft project No. RE782/11. V.K. acknowledges support under grant No. 14-20012S (from Grant Agency of the Czech Republic).
REFERENCES 1. Dukes, M. J., R. Thomas, ., D. F. Kelly. 2014. Improved microchip design and application for in situ transmission electron microscopy of macromolecules. Microsc. Microanal. 20:338–345. 2. Pilhofer, M., K. Aistleitner, ., G. J. Jensen. 2014. Architecture and host interface of environmental chlamydiae revealed by electron cryotomography. Environ. Microbiol. 16:417–429. 3. Echlin, P. 1992. Low-Temperature Microscopy and Analysis. Plenum Publishing, New York. 4. Moor, H., and K. Mu¨hlethaler. 1963. Fine structure in frozen-etched yeast cells. J. Cell Biol. 17:609–628. 5. Dubochet, J., and A. W. McDowall. 1981. Vitrification of pure water for electron microscopy. J. Microsc. 124:3–4.
SUPPORTING MATERIAL
6. Takamori, S., M. Holt, ., R. Jahn. 2006. Molecular anatomy of a trafficking organelle. Cell. 127:831–846.
Five figures and two movies are available at http://www.biophysj.org/ biophysj/supplemental/S0006-3495(16)00101-6.
7. Nicastro, D., A. S. Frangakis, ., W. Baumeister. 2000. Cryo-electron tomography of Neurospora mitochondria. J. Struct. Biol. 129:48–56.
AUTHOR CONTRIBUTIONS
8. Jun, S., D. Ke, ., P. Zhang. 2011. Direct visualization of HIV-1 with correlative live-cell microscopy and cryo-electron tomography. Structure. 19:1573–1581. 9. Milne, J. L. S., and S. Subramaniam. 2009. Cryo-electron tomography of bacteria: progress, challenges and future prospects. Nat. Rev. Microbiol. 7:666–675.
R.R. and R.A.W. initiated the project; R.R., V.K., H.N., and S.T. designed the hardware components; S.T. performed the calibration of the microscope, all measurements, and data analysis; V.K. contributed the analytical tools; and H.N. developed the electronics. All authors wrote the article.
10. Koster, A. J., R. Grimm, ., W. Baumeister. 1997. Perspectives of molecular and cellular electron tomography. J. Struct. Biol. 120:276–308.
ACKNOWLEDGMENTS
11. Medalia, O., I. Weber, ., W. Baumeister. 2002. Macromolecular architecture in eukaryotic cells visualized by cryoelectron tomography. Science. 298:1209–1213.
We thank Martin Wensing from the Fine Mechanical Workshop, Ulrike Keller for help during preparation, Dr. Ruben Diaz-Avalos for kindly providing TMV, and Dr. Philippe Ringler for its purification. Moreover,
12. Parry, A. L., P. H. H. Bomans, ., S. C. G. Biagini. 2008. Cryo electron tomography reveals confined complex morphologies of tripeptide-containing amphiphilic double-comb diblock copolymers. Angew. Chem. Int. Ed. Engl. 47:8859–8862.
Biophysical Journal 110(4) 758–765
High-Vacuum Cryo-transfer System for EM
765
13. Al-Amoudi, A., L. P. O. Norlen, and J. Dubochet. 2004. Cryo-electron microscopy of vitreous sections of native biological cells and tissues. J. Struct. Biol. 148:131–135.
30. Mu¨ller, S. A., K. N. Goldie, ., A. Engel. 1992. Factors influencing the precision of quantitative scanning transmission electron microscopy. Ultramicroscopy. 46:317–334.
14. Marko, M., C. Hsieh, ., C. Mannella. 2007. Focused-ion-beam thinning of frozen-hydrated biological specimens for cryo-electron microscopy. Nat. Methods. 4:215–217. 15. Rubino, S., S. Akhtar, ., K. Leifer. 2012. A site-specific focused-ionbeam lift-out method for cryo transmission electron microscopy. J. Struct. Biol. 180:572–576.
31. Pfaff, M., E. Mu¨ller, ., D. Gerthsen. 2011. Low-energy electron scattering in carbon-based materials analyzed by scanning transmission electron microscopy and its application to sample thickness determination. J. Microsc. 243:31–39.
16. Hayles, M. F., D. A. M. de Winter, ., A. J. Verkleij. 2010. The making of frozen-hydrated, vitreous lamellas from cells for cryo-electron microscopy. J. Struct. Biol. 172:180–190. 17. Villa, E., M. Schaffer, ., W. Baumeister. 2013. Opening windows into the cell: focused-ion-beam milling for cryo-electron tomography. Curr. Opin. Struct. Biol. 23:771–777. 18. Severs, N. J. 2007. Freeze-fracture electron microscopy. Nat. Protoc. 2:547–576. 19. Rigort, A., F. J. B. Ba¨uerlein, ., J. M. Plitzko. 2010. Micromachining tools and correlative approaches for cellular cryo-electron tomography. J. Struct. Biol. 172:169–179. 20. Hagen, C., P. Guttmann, ., K. Gru¨newald. 2012. Correlative VIS-fluorescence and soft X-ray cryo-microscopy/tomography of adherent cells. J. Struct. Biol. 177:193–201.
32. Volkenandt, T., E. Mu¨ller, and D. Gerthsen. 2014. Sample thickness determination by scanning transmission electron microscopy at low electron energies. Microsc. Microanal. 20:111–123. 33. Ritchie, N. W. M. 2005. A new Monte Carlo application for complex sample geometries. Surf. Interface Anal. 37:1006–1011. 34. Drouin, D., A. R. Couture, ., R. Gauvin. 2007. CASINO V2.42: a fast and easy-to-use modeling tool for scanning electron microscopy and microanalysis users. Scanning. 29:92–101. 35. Demers, H., N. Poirier-Demers, ., D. Drouin. 2011. Three-dimensional electron microscopy simulation with the CASINO Monte Carlo software. Scanning. 33:135–146. 36. Reichelt, R., and A. Engel. 1984. Monte Carlo calculations of elastic and inelastic electron scattering in biological and plastic materials. Ultramicroscopy. 13:279–293.
21. Kaufmann, R., P. Schellenberger, ., K. Gru¨newald. 2014. Super-resolution microscopy using standard fluorescent proteins in intact cells under cryo-conditions. Nano Lett. 14:4171–4175.
37. Krzyza´nek, V., and R. Reichelt. 2003. MONCA: a new MATLAB package for Monte Carlo simulation of electron scattering in thin specimens in the energy range 10–200 keV. Microsc. Microanal. 9 (S03):110–111.
22. Weisenburger, S., B. Jing, ., V. Sandoghdar. 2014. Cryogenic colocalization microscopy for nanometer-distance measurements. ChemPhysChem. 15:763–770.
38. Zeitler, E., and G. F. Bahr. 1962. A Photometric procedure for weight determination of submicroscopic particles quantitative electron microscopy. J. Appl. Phys. 33:847–853.
23. Chang, Y. W., S. Chen, ., G. J. Jensen. 2014. Correlated cryogenic photoactivated localization microscopy and cryo-electron tomography. Nat. Methods. 11:737–739.
39. Wall, J. S., and J. F. Hainfeld. 1986. Mass mapping with the scanning transmission electron microscope. Annu. Rev. Biophys. Biophys. Chem. 15:355–376.
24. Capaccioli, S., and K. L. Ngai. 2011. Resolving the controversy on the glass transition temperature of water? J. Chem. Phys. 135:104504.
40. Engel, A. 1978. Molecular weight determination by scanning transmission electron microscopy. Ultramicroscopy. 3:273–281.
25. Amann-Winkel, K., C. Gainaru, ., T. Loerting. 2013. Water’s second glass transition. Proc. Natl. Acad. Sci. USA. 110:17720–17725. 26. Wolf, S. G., L. Houben, and M. Elbaum. 2014. Cryo-scanning transmission electron tomography of vitrified cells. Nat. Methods. 11: 423–428.
41. Krzyza´nek, V., S. A. Mu¨ller, ., R. Reichelt. 2009. MASDET—a fast and user-friendly multiplatform software for mass determination by dark-field electron microscopy. J. Struct. Biol. 165:78–87. 42. Taylor, J. R. 1996. An Introduction to Error Analysis. University Science Books, Herndon, VA.
27. Krzyzanek, V., and R. Reichelt. 2007. High-resolution scanning electron microscope for mass determination: progress in the development of a new nanoanalytical tool. Microsc. Microanal. 13 (S03):80–81.
43. Namba, K., and G. Stubbs. 1986. Structure of tobacco mosaic virus at ˚ resolution: implications for assembly. Science. 231:1401–1406. 3.6 A
28. Engel, A. 2009. Scanning transmission electron microscopy: biological applications. Adv. Imaging Electron Phys. 159:357–386.
44. Dı´az-Avalos, R., and D. L. Caspar. 1998. Structure of the stacked disk aggregate of tobacco mosaic virus protein. Biophys. J. 74:595–603.
29. Bozzola, J. J., and L. D. Russel. 1992. Electron Microscopy—Principles and Techniques for Biologists. Jones & Bartlett, Sudbury, MA.
45. Angell, C. A. 2004. Amorphous water. Annu. Rev. Phys. Chem. 55:559–583.
Biophysical Journal 110(4) 758–765
