High-aperture cryogenic light microscopy.

We took a number of considerations into account when we were designing the prototype high-aperture cryogenic microscope, the most important of which was ensuring the instrument was thermally stable in operation. As with any imaging instrument, a lack of thermal stability can lead to specimen drift during data collection, resulting in blurred images. Obviously, given the large thermal difference between the cryogen and ambient temperature there is more possibility of drift occurring in cryogenic imaging systems. Our other major design considerations were designing an optical system that performed robustly at cryogenic temperatures, and ensuring the immersion fluid was well matched to the refractive index of the cryogenic specimen and the optical system. In this section, we will describe the overall design and use of the microscope. The overall design of the cryogenic microscope is shown in a CAD model of the instrument in Fig. 1(A) and in greater detail in schematic form in Fig. 1(B) and (C) , with details of the specimen imaging compartment and the configuration of the major optical components in Fig. 1(D) and (E) . The actual instrument is shown in photographic form in Fig. 2(A–D) . As can be seen from these images, our initial ‘proof of concept’ prototype cryogenic microscope was built from relatively simple, readily available materials and components. To the point, it may appear that minimizing construction costs was a primary consideration. However, this was not the case. The materials used in the construction of the prototype were judiciously chosen based on their thermal properties and their performance in the assembled instrument. This fact, together with iteratively fine tuning the design, resulted in a prototype microscope that was thermally stable and capable of imaging reliably over extended periods of time. In operation, the entire microscope is cooled en bloc . Liquid nitrogen is introduced through a porthole and the microscope cooled by filling the reservoir to the appropriate level. There is no need to incrementally cool the system, or any of the optical or mechanical components. The liquid nitrogen reservoir is constructed from a 36 cm × 50 cm × 36 cm polystyrene box. The top plate of the instrument is constructed from a 1.3-cm thick aluminium plate. This forms the attachment surface for all components requiring a high degree of positional stability (e.g. the low-temperature objective lens and the specimen stage). The rotating platform is attached to a rotation stage by four 5-cm diameter fibreglass tubes. The platform is cooled by four 2.5-cm diameter aluminium rods ( Fig. 1(B-f) ), which extend down into the liquid nitrogen bath ( Fig. 1(B-d) ). This plate is maintained at a constant temperature of 90 K as long as the liquid nitrogen bath is in contact with the rods. The entire instrument was designed and built in-house by staff of the National Center for X-ray Tomography. The microscope and data collection are both controlled by Image-Pro (Media Cybernetics Inc., Bethesda, MD). Once cold, the microscope can be operated for an entire day with no frosting or drift. The specimen imaging compartment ( Fig. 1(D) and (E) ) is filled with immersion liquid, typically liquid propane generated by pumping gaseous propane from a standard laboratory cylinder into the imaging compartment using a 50-mL syringe. The design allows the propane level to be topped off if necessary between images. In typical operation, the propane cell is filled every hour. The temperature of the specimen imaging compartment is maintained by conduction through copper strips immersed in the liquid nitrogen reservoir ( Fig. 1(D) and (E) ). The temperature of the entire microscope remains constant since great care was taken to minimize hot/cold spots, as these could lead to the production of unstable thermal gradients. The objective lens forms one side of the imaging compartment, with glass forming the other three sidewalls (see Fig. 1(D) and (E) ). The objective lens used in the microscope is a standard, readily available optic that has been minimally modified. In our experience, objective lenses are much more robust towards being immersed in cryogen than one might expect. We have repeatedly kept lenses at cryogenic temperatures for extended periods (8–10 h) with no apparent ill effects. We have also subjected the entire optical system to a number of cycles of alternating between room temperature and cryogenic operation. Again, this had no deleterious effects on the microscope, or the optical system. The cryo light microscope can be configured with or without a tube lens. This allows the magnification to be adjusted by either moving the camera or adjusting the tube lens layout, thus allowing for optimization of objective performance. The condenser system is relatively simple, however it is still possible to obtain a condenser with an NA > 1. With minor modification, a cryogenic immersion fluid could be used to increase the condenser aperture. Such a system is being developed for implementation of 4 pi light collection and imaging geometries. Multiple cameras can also be used. For example, we have found Qimaging Monochrome Retiga EXi model RET-EXi-F-M-12-C 1392 × 1040 pixel (Qimaging, Surrey, BC) fitted with a Sony CCD and a low light EMCCD from ANDOR Inc. (Belfast, Northern Ireland) with a back thinned E2V 500 × 500 cooled CCD model number I-XON 897 to work very well in this instrument. The excitation source can be either a multi-line AOTF modulated Laser source (633 nm, 543 nm, 488 nm and 457 nm) or a broadband Xenon Radio Frequency Plasma Source (Sutter Instruments (Novato, CA) model LAMBDA LS-XL). The instrument is sufficiently stable that it is possible to collect images with exposures as long as 7 s at ~250 nm resolution. Specimens are either flash frozen by rapidly plunging them into a second cell of cryogen, or cryo-transferred into the robotic specimen handling device ( Figs 1(B), 1(C) , 2(B) and 2(C) ). Specimens are prepared for imaging by mounting them on standard electron microscopy grids, cover slips, or placing them in thin walled glass capillaries. The latter mounting system has been used most frequently since this instrument was primarily designed for correlated light and soft X-ray microscopy studies (e.g. see Parkinson et al. , 2008 ). This specimen mount can be transferred between these two imaging modalities with ease. Capillaries are particularly suited to imaging ‘smaller’ cells, such as bacteria or yeast. The capillaries used for this purpose range in diameter from 3 to 15 microns. Even at the larger end of this size scale, the thermal conductivity is still sufficiently rapid that the specimen cools without the formation of potentially damaging ice crystals. Since the capillaries have a wall thickness ~150 nm, there is no significant aberration due to the difference in refractive index of the glass and the cryogenic immersion fluid. Our prototype high-aperture cryogenic microscope design is shown in Fig. 1(A–E) , and photographs of the functional microscope in Fig. 2(A–E) . The microscope was designed to allow specimens in a variety of mounts to be flash frozen and immediately imaged, or frozen and stored in the instrument for imaging at a later date (the design of this automated cryogenic specimen handling device is shown in Fig. 1(C) ). Having the opportunity to freeze and store a number of specimens at specific points in time also makes this technique well suited to time-dependent imaging studies. The prototype high-aperture cryogenic microscope was designed to operate in conjunction with a soft X-ray microscope (XM-2) at the National Center for X-ray Tomography (see http://ncxt.lbl.gov for further details). The new instrument was designed to function as a platform for specimen vitrification, short-term cryo-storage, correlated light imaging and cryo-transfer to the X-ray microscope for imaging. We have now designed, built and are in the processes of commissioning the next generation cryogenic microscope. The design of this new instrument incorporates a number of improvements that are based on our experience imaging cells using the prototype microscope. To facilitate our description of how the high-aperture cryogenic microscope is used, we will refer to the simple schematic of the microscope components in Fig. 1(B) and (C) . Specimens are manipulated using a 3-axis stepper motor stage located on top of the cryogenic microscope ( Fig. 1(B-k) ). This stage also contains a spring-driven plunge freezer that is used to rapidly vitrify specimens in a propane reservoir ( Fig. 1(B-c) ). Following vitrification, the specimen is moved from the liquid propane freezing reservoir into the liquid nitrogen temperature dry atmosphere of the cryo-microscope and then translated to the liquid propane imaging reservoir ( Fig. 1(B-g) ). Propane is a suitable low-temperature fluid with a high refractive index (1.33), and is therefore well matched to the 1.3 NA lens used in the microscope. The objective lens used for the current measurements is a modified hybrid constructed from a Fratelli Koristka lens 30× NA 1.00 and a Spencer Lens Company (Germany) 82× NA 1.33 oil lens. Finally, the rotating cryo-platform ( Fig. 1(B-f) ) is positioned to move the cryogenic condenser lens ( Fig. 1(B-h) ) opposite to the high-aperture objective ( Fig. 1(B-i) ). The bright field illumination is provided by a halogen lamp ( Fig. 1(B-a) ) coupled to the cryo-storage unit by a fibre light guide aimed at a plane mirror aligned with the condenser lens, both of which rotate with the cryo-platform. Bright field light microscopy is first used to locate samples of interest. In the first instance, the CCD detector ( Fig. 1(B-j) ) is moved closer to the objective ( Fig. 1(B-i) ). This increases the field of view and reduces the magnification. Once a specimen of interest is found, the detector is moved back to a position that produces the desired field of view and magnification. High-aperture multi-wavelength fluorescence microscopy can then be performed on the sample by moving the appropriate dichroic filter into position and switching the illumination source. Illumination for fluorescence measurements is provided by a liquid light guide coupled xenon light source. The microscope is outfitted with a sliding holder containing three positions for dichroic filter cubes. These are manually positioned to select the desired excitation dichroic and emission filters (for clarity, the filter block assembly is not shown in the schematic; it is located immediately in front of the objective ( Fig. 1(B-i) ). Once an adequate fluorescence image has been obtained, the specimen can be transferred to a storage position in the robotic specimen handling instrumentation (shown in Fig. 1(C-l) ), to await further imaging in the cryogenic microscope or be cryo-transferred to the soft X-ray microscope. The system allows for vitrification, imaging and storage of up to 15 samples. The unit is thermally efficient, and only requires refilling with liquid nitrogen every 6 h. If long-term storage is required, the samples can be cryo-transferred into a long-term liquid nitrogen Dewar. Although this instrument is designed as a stand-alone microscope, it would be possible with minimal redesign effort to adapt it for a standard commercial microscope. We recently coupled it with our Zeiss LSM 510 NLO imaging system.

Specimen preparation for cryogenic microscopy

E. coli bacteria were transformed with a pET-23b vector containing a C-terminal His-Tag sequence (EMD Chemicals, Inc., Gibbstown, NJ) containing six histidines plus the coding sequence for yellow fluorescent protein. Cells were grown to mid-log phase in LB medium and expression of the plasmid induced by adding IPTG to a final concentration of 1 mM. After 2 h, the resultant over-expression of YFP caused the cells to appear yellowish green. The expressed YFP construct has no bacterial function, and was distributed throughout the cell, with the exception of a single region that is likely an inclusion body, comprised of mis-folded protein molecules. Since yellow fluorescent protein has to fold properly to fluoresce, this explains the lack of YFP emission signal. This explanation also correlates with the high density of this region when cells are imaged using soft X-ray tomography shown in Fig. 3(D) and (E) . Wild-type S. pombe cells (strain #972 h) were grown with rotary shaking at 30°C in YES media (yeast extract + adenine + casamino acids) supplemented with leucine, histidine, uracil and dextrose. Vacuoles were stained by the addition of 5-chloromethyl fluorescein diacetate (CMFDA; Molecular Probes ® , Invitrogen, Carlsbad, CA) to a final concentration of 2.5 μM in growth media. NIH/3T3 cells (ATCC) were maintained in DME (ATCC) supplemented with 10% bovine calf serum (ATCC) then plated at a density of 25,000 cells/cm 2 on poly-L-lysine-coated glass cover slips and allowed to attach for 16–24 h. Cells were then fixed with 4% formaldehyde and 0.1% glutaraldehyde (30 min at room temperature) then permeabilized with 0.1% saponin. For actin labelling, cells were incubated in a 1:40 solution of ALEXA-488-conjugated phalloidin (Invitrogen) for 60 min at room temperature. Photo-bleaching experiments The microscope was set up to continuously record fluorescence images of bacteria expressing YFP that had been mounted in a thin-walled capillary. Measurements were collected at room temperature every 50 milliseconds for a total cumulative exposure time of 10 min. The total fluorescence intensity was integrated in each image and plotted against time. The experiment was repeated using a similar specimen and the same data collection parameters, but with the microscope operated at cryogenic temperature. Soft X-ray tomography of E. coli Soft X-ray tomography of E coli cells was carried out using the instruments and methods described in Le Gros et al. (2005) .