Here, we present a protocol on how to determine the quantity and distribution of metals in a sample using synchrotron X-ray fluorescence. We focus on adherent cells, and describe the chemical fixation method to prepare this sample. We then describe how to mount and image the sample using synchrotron X-rays.
X-ray fluorescence imaging allows us to non-destructively measure the spatial distribution and concentration of multiple elements simultaneously over large or small sample areas. It has been applied in many areas of science, including materials science, geoscience, studying works of cultural heritage, and in chemical biology. In the case of chemical biology, for example, visualizing the metal distributions within cells allows us to study both naturally-occurring metal ions in the cells, as well as exogenously-introduced metals such as drugs and nanoparticles. Due to the fully hydrated nature of nearly all biological samples, cryo-fixation followed by imaging under cryogenic temperature represents the ideal imaging modality currently available. However, under the circumstances that such a combination is not easily accessible or practical, aldehyde based chemical fixation remains useful and sometimes inevitable. This article describes in as much detail as possible in the preparation of adherent mammalian cells by chemical fixation for X-ray fluorescent imaging.
X-ray fluorescence imaging allows for both the identity and quantity of elements present in a sample to be spatially resolved. Incident X-rays, of an energy selected to be greater than the electron binding energy of the heaviest element of interest, overcome the binding energy of inner-shell electrons to the nucleus1. This creates a ‘hole’ in the electron shell. As higher-energy electrons fall down into these holes, fluorescent X-rays are emitted whose wavelength is dependent on the energy separation of those orbitals. Since the energy spacing of the orbitals is characteristic of a given element, the X-ray fluorescence emission also has characteristic wavelengths, dependent on the element. It is this emission at a characteristic wavelength that allows the identification of the elements present. Calibration of the fluorescence intensity allows the quantitation of the elements present.
X-ray fluorescence microscopy (XFM) has become increasingly utilized, partly due to the development of very brilliant X-ray synchrotron sources, such as those at Spring-8 in Japan, the European Radiation Synchrotron Facility (ESRF) in France, and the Advanced Photon Source (APS) in the US2. These sources provide very high intensity X-ray beams. At the same time, improvements in X-ray optics, such as zone plate technology, allowed the focusing of these beams to sub-micron spots, albeit rather inefficiently3. With very high-intensity beams, even a relatively small amount of light that can be focused is sufficient to excite the endogenous metals in cells, producing signal that can be measured with currently available detector technology. Thus, studying the chemical biology of metals in the cell is one application in particular that makes use of many of the recent developments in this technique4-10.
There are many critical factors to be considered while applying XFM to investigate the elemental distribution and quantification of cultured mammalian cells or other biological samples. Firstly, the sample needs to be kept intact, both structurally and with respect to its elemental composition, in order for the measurement to be meaningful. Secondly, the sample must also be preserved in some way so that it is hardy to the radiation damage that can be caused by a focused X-ray beam. One way that a sample can meet both of these criteria at once is to be rapidly frozen into a vitreous, amorphous ice11,12. Rapid freezing is often achieved through various cryopreservation techniques such as plunge freezing or high pressure freezing13-16. It is generally accepted that cryopreservation preserves overall cellular architecture and chemical compositions in biological samples as close to native state as possible. Chemical fixation, on the other hand, due to the slow and selective penetration of fixatives into cells and tissues as well as subsequent changes in membrane permeability, may allow various cellular ions especially the diffusible ions such as Cl, Ca and K to be leached, lost or relocated, thus rendering investigation of these elements suboptimal17-19. Despite the clear advantage of cryo-fixation over chemical fixation in general, for adherent mammalian cells in particular, cryopreservation has various limitations20-23. The most obvious one is that not every research lab has easy access to cryopreservation instruments. Most current high pressure freezers or even plunge freezers are costly and owned only by a subset of cryo facilities, which may be far from where cells are incubated. The benefit of cryopreservation might be traded for the disadvantage of travel stress placed on the cells. Thus, while cryopreservation is surely the most rigorous way to preserve samples for X-ray fluorescence analysis, it is certainly not the most accessible to all researchers under all circumstances; nor is it always essential — if the metals of interest are tightly bound to fixable macromolecules, and the resolution at which the sample will be imaged is greater than the damage to the ultra-microstructure that might occur during drying. Mindful of the caveats24, chemical fixation and drying may be a suitable choice.
Other factors in a successful X-ray fluorescence imaging experiment include proper analysis. X-ray fluorescence imaging is fundamentally X-ray fluorescence emission spectroscopy combined with raster-scanning to provide spatial resolution. The X-ray fluorescence emission spectra collected contain a combination of overlapping emission peaks, background, and the elastic and inelastic scattering peaks of the incident beam. Software that enables the de-convolution of these contributions, and the fitting of the emission peaks, has been a critical development to this field25. Also, the development and commercial distribution of thin-film standards of known composition, used to calibrate fluorescence intensity relative to material quantity, has also been very important.
This protocol provides a description of the preparation of adherent cells by chemical fixation and air drying. A vital step in this process is the growth of the cells on the silicon nitride windows, which often do not adhere well, making gentle rinsing in a particular fashion key to success.
1. Preparation of Instruments, Substrates, Culture Media and Dishes
2. Plating of Cells
3. Fixation of Cells
4. Sample Mounting and Imaging
The ability of X-ray fluorescence imaging to provide information about biological samples is contingent upon these samples being prepared in such a way that they are robust to radiation damage on the time-scale of the experiment, and yet their chemical and structural features are well preserved. In viewing the result of a sample that has been prepared as described above and imaged, it is possible to see that there is variation in the elements present — indicating that fixation preserved these aspects of the cell (Figure 1).
Conversely, looking instead at a sample where this process did not go well, and the buffer remains present during drying, extensive crystal formation from the molecules present creates structural damage to the cells and also interferes with the collection of the X-ray fluorescence spectra (Figure 2).
These images are generated by per-pixel fitting of the X-ray fluorescence spectrum at each point in the image. This means that each spectrum, collected at each pixel, has been individually analyzed. When viewing the panels in these images generated from the fitted data, the value assigned to each point in the image is the integrated sum of a Gaussian for the characteristic emission of a given element (e.g., iron) as best fits the X-ray fluorescence spectrum collected at that point. For example, a single pixel in the image has a value (number) for iron, which is the sum of the area under a Gaussian curve at the characteristic emission energy for iron, that fits and models the data for the emission spectrum of that point on the sample. The data is displayed with threshold values above each image (max and min) that assign the high- and low-ends of the color-table (at the bottom of each image) to specific values within the image.
Figure 1. X-ray Fluorescence Image of a Human SH-SY5Y Cell. Each panel in this image displays different information about the same cells. The first panel, labeled DIC, is the optical differential interference contrast micrograph of the cells. The following panels, left to right, are the phosphorus (P), sulfur (S), iron (Fe), and zinc (Zn) images of the same cells, showing their distribution over the area of the cell. The scale bar shown is 20 µm. Please click here to view a larger version of the figure.
Figure 2. X-ray Fluorescence Image of a Rat B103 Cell, prepared with poor or insufficient removal of buffers. Two cells are visible in the phosphorus (P) panel, and are slightly visible in the iron (Fe) and zinc (Zn) panels. However, the presence of crystallized buffer, visible as a smear of particulates through the center of the image, obscures most of the information. The scale bar shown is 20 µm. Please click here to view a larger version of the figure.
X-ray fluorescence imaging is useful in many fields, including geosciences, materials science, and chemical biology26-34. Advances in synchrotron X-rays, and their focusing, have produced very high-intensity beams. Focused X-ray beams sufficient to excite the endogenous metals in cells now exist, producing signal that can be measured with currently available silicon drift detector technology. And studying the chemical biology of metals in the cell is one application in particular that makes use of many of the recent developments in this area.
Yet, the study of biological materials using X-ray fluorescence microscopy also presents special challenges, relative to other materials, because they are hydrated. To prevent radiation damage, ideally the samples would be frozen in vitreous ice during the measurement. However, chemical fixation and air drying are much more accessible to the novice, and may be appropriate if the metals of interest are inertly bound to fixable macromolecules, and the resolution at which the sample will be imaged is greater than the damage to the ultra-microstructure that might occur during drying.
In considering substrates for adherent cell XFM studies, an ideal substrate needs to meet the following basic requirements: 1) support robust cell growth without alternating normal proliferative and phenotypic cell growth, 2) must not have X-ray fluorescence that overlaps with those of elements of interest, 3) be optically transparent for cell growth assessment under light microscope before XFM, and 4) be able to withstand any physical and chemical manipulation during culturing and subsequent fixation. Historically, different substrates, ranging from thin polycarbonate foils/films35-37, formvar coated gold transmission electron microscopy (TEM) grids38-42 and silicon nitride windows5,17,43-46, have been used to grow mammalian cells and used for X-ray fluorescence imaging. In comparison among the existing and potential substrates for imaging adherent mammalian cells by XFM, silicon nitride membrane (Si3N4) windows are most suitable due to their low fluorescence background and relatively simple elemental composition4-6,47,48. In addition Si3N4 windows support robust cell attachment and proliferation. Compared to other commonly used substrates such as TEM grids, Si3N4 windows also have a much larger uninterrupted imaging area which is a particular advantage when these windows are used for X-ray tomography.
Si3N4 windows are commercially available from a limited number of vendors. Working with silicon nitride windows is an acquired skill. The windows are very fragile, and require great care to not create any twisting forces while handling them that might shatter them, or to drop them. Handling them by the edges using reverse tweezers is a popular approach. Most windows have thicknesses ranging from 30 nm to 500 nm and a width from 1 x 1 mm to 5 x 5 mm. In general, the thicker the membrane, the more robust they are to all kinds of handling stress. However, with increased thickness and less optical transparency, they will increase the background emission from the specimen. The 200 nm thick membranes are quite ideal. They have minimal impact on the measurement, yet are sturdy enough to be handled without much breakage.
Although many cell types tested can initially attach and robustly grow on the sterile Si3N4 windows, cells grown on silicon nitride windows are in general much more easily agitated than cells on traditional cell culture plates. They become easily detached, aggregated or rounded up even during routine media changes. We found pre-washed windows often became more hydrophilic; cells attach better and had less chance to aggregate together. The steps described in the protocol above delineate a gentle-washing approach to keeping the cells on the window. Some other steps that may be taken include coating the windows with poly-lysine, or laminin, just as one might coat a coverslip.
Another aspect of this method that can be the success or failure of a given experiment is agitation of the samples. In observing cells’ growth on plastic culture vessels, the normal range of agitation resulting from pipetting and plate transportation doesn’t seem to cause much harm. However, for cells grown on the silicon nitride windows, extra care during media change and plate transportation is needed. For some easily agitated cells such as mouse fibroblast cells NIH/3T3, the culture plates containing windows need to be handled carefully. They have to be carefully taken out from the incubator and gently set on the microscope stage or cabinet surface. Not necessarily every little physical shock will disturb the cells, but the risk of cell detachment increases without extra care. In addition, different batches of fetal bovine serum and silicon nitride windows may have certain effects on the attachment of cells on the window. Some sources of serum or batches of windows seem easier to work with, i.e., without seeing too much disturbance by similar care. So, some trial and error for the cell line of interest may be advised.
With developments in cryogenic stages for X-ray fluorescence microprobes, as well as new research in the preparation of frozen hydrated biological samples, it is likely that sample preparation in the future will much more increasingly include frozen hydrated specimens. Many of the same challenges in working with the silicon nitride windows, and retaining cell adhesion exist there as well. Yet, mastering this technique remains a very good place to start in developing skills before attempting cryopreservation, and many times, is a perfectly suitable way to image samples.
The authors have nothing to disclose.
The authors acknowledge Stefan Vogt for his assistance in the fitting of the representative data shown in this paper, and helpful discussions. The authors also acknowledge Chris Jacobsen for his support to Q. J.
Use of the Advanced Photon Source, beamlines 2-ID-E and 8-BM-B, at Argonne National Laboratory was supported by the U. S. Department of Energy, Office of Science, Office of Basic Energy Sciences, under Contract No. DE-AC02-06CH11357.
silicon nitride windows | Silson Ltd/J B J Business Park/Northampton Rd, Northampton NN7 3DW, United Kingdom | No part numbers available. Order by size. Membrane size: 1.5 mm x 1.5 mm. Thickness 500 nm. Frame size: 5 mm x 5 mm. Frame thickness: 200 µm | Alternate source: SPI Supplies / Structure Probe, Inc.West Chester, PA |
reverse tweezers | Electron Microscopy Sciences, P.O. Box 550, 1560 Industry Road, Hatfield, PA 19440, Tel: 215-412-8400, Toll Free: 800-523-5874, Fax: 215-412-8450 | 78520-5X | EMS 5X, NC – Ultra Fine Tweezers |
rubber grid mat | Electron Microscopy Sciences, P.O. Box 550, 1560 Industry Road, Hatfield, PA 19440, Tel: 215-412-8400, Toll Free: 800-523-5874, Fax: 215-412-8450 | 71170 | Round Grid Mat |
acetic acid | Sigma-Aldrich, 3050 Spruce St., St. Louis, MO 63103, Tel: 800-325-3010, Fax: 800-325-5052 | 338826 | trace metals grade concentrated acetic acid |
PIPES buffer | Sigma-Aldrich, 3050 Spruce St., St. Louis, MO 63103, Tel: 800-325-3010, Fax: 800-325-5052 | P6757 | solid PIPES buffer |
formaldehyde stock solution | Electron Microscopy Sciences, P.O. Box 550, 1560 Industry Road, Hatfield, PA 19440, Tel: 215-412-8400, Toll Free: 800-523-5874, Fax: 215-412-8450 | RT 17113 | 10 x 10mL ampules of 20% aqueous paraformaldehyde |