DELMIC BV may profit from publication of the results of this study. A.P.J.E. is shareholder and employee of DELMIC BV. A.C.Z. is shareholder in DELMIC. P.K. and J.P.H. are shareholders and members of the advisory board to DELMIC. Vacuum-compatible immersion oil was supplied by DELMIC BV for this study. DELMIC BV is developing a commercial product based on the prototype microscope used in this study. This study was partly supported by NanoNextNL. There are no further patents, products in development or marketed products to declare. This does not alter the authors' adherence to all the PLOS ONE policies on sharing data and materials, as detailed online in the guide for authors.
Designed and constructed the integrated microscope: ACZ JPH. Constructed the optical microscope: NL ACZ ACN APJE JPH. Conceived and designed the experiments: NL ACZ PK JPH. Performed the experiments: NL ACZ ACN. Analyzed the data: NL PWV JPH. Contributed reagents/materials/analysis tools: PWV MSL JCH RAW. Wrote the paper: NL JPH.
Correlative light and electron microscopy (CLEM) is a unique method for investigating biological structure-function relations. With CLEM protein distributions visualized in fluorescence can be mapped onto the cellular ultrastructure measured with electron microscopy. Widespread application of correlative microscopy is hampered by elaborate experimental procedures related foremost to retrieving regions of interest in both modalities and/or compromises in integrated approaches. We present a novel approach to correlative microscopy, in which a high numerical aperture epi-fluorescence microscope and a scanning electron microscope illuminate the same area of a sample at the same time. This removes the need for retrieval of regions of interest leading to a drastic reduction of inspection times and the possibility for quantitative investigations of large areas and datasets with correlative microscopy. We demonstrate Simultaneous CLEM (SCLEM) analyzing cell-cell connections and membrane protrusions in whole uncoated colon adenocarcinoma cell line cells stained for actin and cortactin with AlexaFluor488. SCLEM imaging of coverglass-mounted tissue sections with both electron-dense and fluorescence staining is also shown.
Understanding cellular structure-function relations requires the complementary capabilities of both fluorescence and electron microscopy. Fluorescence microscopy (FM) visualizes individual proteins in color through the use of immunofluorescent or endogenous labeling
Typically, in CLEM research, inspection with FM and EM is performed on the two separate microscopes. In this way, both types of microscopy can be used at their full capabilities, including superresolution FM
The retrieval of ROI’s can be facilitated using external markers on the sample holder that allow definition of a universal coordinate system in both FM and EM
Integrated approaches, where an optical microscope is integrated in an EM vacuum chamber, offer a practical solution to several issues. This approach was pioneered in an SEM in the early work of Wouters
In the SEM, a high-NA optical microscope can be integrated into the vacuum chamber
For Simultaneous Correlative Light-Electron Microscopy (SCLEM), we use an integrated microscope where the objective lens is positioned in the vacuum chamber of a Scanning Electron Microscope (SEM), directly underneath the sample (see
(a) schematic lay-out for SCLEM, BSE: backscattered electrons, SE: secondary electrons, ETD: Everhard-Thornley detector, LED: light emitting diode, CCD: charge coupled device camera. (b) inside view of the integrated microscope for SCLEM showing optical objective lens in epi-configuration underneath sample holder and electron lens.
As can be seen in
Image formation in the SEM occurs through detection of either low-energy, secondary electrons (SE) or high-energy back-scattered electrons (BSE). Among other contrast mechanisms, SE imaging gives nanometer-scale detail of surface topography, while the BSE signal originates from a larger sample volume contrasting differences in atomic number or density. In our SCLEM setup, samples need to be mounted on a transparent substrate. The use of transparent conductive ITO-coated glass coverslides eliminates the need for a conductive over-coating of biological materials
The formation and growth of cellular extensions and protrusions, such as filopodia, lammelipodia, and invadopodia, plays a crucial role in cell motility and cell-cell signaling. These processes involve a wide variety of proteins. The role that these proteins play in the development and maturation of cellular topography, is an area of active research. SCLEM on uncoated, whole cells may serve as a powerful technique to investigate the role of protein localization as the SEM can record a detailed map of the network of cellular protrusions.
As a first illustration of the application of SCLEM, we immuno-labeled SW480 colon adenocarcinoma cell line cells for actin with phalloidin-Alexa488. Wide-field fluorescence allows for rapid identification of labeled cells and selection of a region of interest. In
(a) FM image of three adenocarcinoma cells actin labeled with Alexa488. The three cells are connected via tentacles and larger extrusions. Scalebar 5 µm. (b) SEM image of the boxed area in (a), showing detailed information on the connections between the cells. A dense network of tentacles and lamellae stretches between the upper and the right cell. Scalebar 3 µm(c) FM image of an extension connecting another two adenocarcinoma cells. Clear variations in actin concentration along the extrusion can be observed. (b) BSE image of the extrusion in (a). Red arrows mark areas with increased concentration of tentacles that occur before and after the thinner parts of the extrusion. Scale bar is 10 µm. (d, e) SE and BSE high-magnification images of the boxed areas in (b) showing a region rich in tentacles and small lamellar extrusions. Scale bars are 2 µm. (f) Fluorescence intensity profiles, normalized on the maximum, taken along the red and blue lines in (a). (g) Normalized SE intensity profile taken at the corresponding locations marked in (c).
In other cases, cell-cell connections were found to consist of larger extensions stretching several tens to hundreds of micrometers. In
Next, we labeled the SW480 adenocarcinoma cells for cortactin, again with Alexa488. Cortactin is involved in rearrangement of the actin network and as such important in the formation of filopodia, lammelipodia, and invadopodia. In
(a) After mounting the sample and vacuum pumping of the SEM chamber, a low-magnification image in SEM mode is taken to inspect surface coverage and position the sample stage. (b) An isolated cell is identified and the SEM focus is fine-tuned for high-magnification imaging (c) The fluorescence image is recorded after the marked cell was selected. Based on the spatial variations in cortactin distribution and the structural overview in (b), regions of interest are identified for high-magnification imaging in SEM mode. Blue, red, and yellow arrows indicate different type of regions with a local increase in cortactin density. Corresponding areas are also marked in the SEM images. (d) SE image recorded at 20keV of the region of interest identified in (c). The cortactin-rich areas marked with red arrows are directly neighboring regions with larger extrusions and high density of tentacles (e) BSE image recorded at 5keV. (f) SE image at 5keV reveals the details in surface topography. It can be clearly seen that the blue marked cortactin-rich regions located in the cell interior correspond to an increase in cell thickness. The cortactin-rich regions marked with yellow arrows surround a larger thin lamellar outgrowth with numerous extending tentacles. Typical time involved in such a procedure (sample mounting & pump down – a,b – c – d,e,f) amounts to 20–35 minutes (4 min –5 min –5 min –5–15 min).
The same areas are marked in the SEM images in
As mentioned above, one of the important results of SCLEM is that there is no need for specimen transfer and re-adjustment of a ROI to combine high-NA FM data with structural data retrieved with SEM. Correlative imaging is achieved without adding fiducial markers to either the specimen support or the sample itself. This greatly simplifies the experimental workflow for CLEM and allows a user to search for a new ROI directly after inspecting another one. As a demonstration, the total time involved in a typical inspection procedure, as with the cortactin-labeled cancer cells shown in
Often in the practice of FM, sample inspection is started with a low-magnification, low-NA objective lens to identify a ROI for high-resolution inspection. It is important to note that the field of view of the SEM easily extends millimeters squared and is thus much larger than that of the integrated high-NA FM. Thus, low-magnification SEM is well suited to perform a quick inspection of the sample, e.g. to analyze the surface coverage of cells (see
Clearly, prolonged exposure to the electron beam, such as after a high-magnification sub-cellular zoom-in, does lead to bleaching. The rate at which this occurs is dependent on electron energy, but also on the composition and thickness of the substrate and, importantly, the type of fluorophore
Thin tissue sections can be investigated with SCLEM after combined FM and EM staining. Several approaches have been reported that allow for EM staining while preserving fluorescence
(a) FM image of human skin tissue stained with DiIC18 fluorescence and uranyl acetate and osmium tetroxide for EM contrast. Scalebar 5 µm (b) BSE image of a selected region from (a), showing a cell nucleus not discernible in (a) (marked with a red arrow), and bundles of longitudinally and transversally cut collagen fibers. Scalebar is 5 µm (c, d) High-magnification images of the areas marked with (c) a red star, scalebar 1 µm, and (d) a yellow star, scalebar 2 µm.
In the SEM image in
The method of SCLEM removes the need to retrieve a ROI as the alignment between SEM and FM optical axes is fixed while the sample is translated through focus. Thus, the SCLEM time for identification and inspection of a ROI is on the order of few tens of minutes, in which a user can move forth and back arbitrarily between the different SEM and FM detectors. In this way, a sample can be quickly scanned for ROI’s in either SEM or FM mode of operation and a large number of ROI’s can be investigated in a short time compared to CLEM operation on the two microscopes. In addition, issues involved in sample transfer between the microscopes, such as contamination risk, are removed from the workflow. Obviously, this also means that the sample has to be prepared to render contrast in both FM and SEM mode of operation. As we have illustrated with examples this can be done by either performing double staining, or by inspection of whole, uncoated cells with only fluorescent labeling. Alternatively, labeling with dual-contrast probes, like semiconductor quantum dots
The development of probes and preparation protocols for correlative research has emerged in recent years. Watanabe
The surface topography of entire cells can be inspected with SEM without the need for EM staining or even conductive coating of the sample. Cells can be cultured directly on glass substrates that have a transparent, conductive ITO coating, as demonstrated by others
SEM inspection of whole cells probes cellular surface structures important in cell motility and cellular signaling, such as tentacles, lammelipodia, filopodia and cell-cell connections. As illustrated in this work, SCLEM can quantitatively correlate protein distributions to densities and sizes of such surface features. As the electron beam penetrates, depending on electron energy, for several micrometers into the sample, investigation of sub-membrane structures could also be possible, albeit at progressively lower resolution. This would then require incorporation of an EM stain that generates BSE or SE contrast, like in our example of tissue sections. Still, due to scattering of the probe beam, high-resolution imaging would be limited to about 100 nm below the surface.
In the presented SCLEM set-up fluorescence microscopy is performed with a wide-field optical microscope. The low axial resolution of the wide-field microscope does not play a role in the investigation of sections or the thinner progressing or retracting parts of a cell. For samples with a thickness of a micrometer or more, the fluorescence signal may need to be optically sectioned in order to establish a correlation with the SEM signal that originates from the upper part of the sample. As most optical components, such as filters, source and detector, are placed outside the SEM vacuum chamber, illumination and detection paths can be easily adjusted or expanded without the need for vacuum-compatible components. Confocal filtering could in principle be achieved through the insertion of a pinhole. With the use of high-NA immersion objectives optical sectioning at sub-micrometer resolution should be possible. We note that also phase shaping to correct for aberrations due to refractive index differences in thick samples could be possible through the insertion of a spatial light modulator or related optics.
We equipped the fluorescence microscope with a high-NA 100× objective lens using vacuum-compatible immersion oil. The possibility to use a high-NA objective lens with coverglass-mounted samples means that total internal reflection microscopy, and superresolution techniques like PALM, could be used directly in a SCLEM experiment. For superresolution microscopy, with protein localization at a few tens of nanometer resolution, the precise positioning of proteins with respect to the ultrastructure becomes increasingly important
The registration between EM and FM images is a major challenge in CLEM. In the microscope we have used for our SCLEM experiments, the axial alignment between both modalities is within 10 µm. For the examples shown in this work this gives us, in combination with endogenous markers present in the sample, sufficient registration to identify and examine a ROI with both modalities. High accuracy determination could be carried out using conventional techniques, such as the use of fiducial markers
SCLEM relies on the possibility to perform both electron and optical microscopy simultaneously. We have observed that low-magnification SEM imaging at 20keV does not lead to a visible degradation of sample fluorescence. This gives us the possibility to perform wide field of view SEM inspection prior to FM investigation. Interestingly, SCLEM brings the possibility to study bleaching induced by electron-beam exposure in a quantitative and dynamic way by recording the fluorescence signal as a function of electron dose. This would not only provide a novel way of analyzing electron-induced reactions in molecules, but would also enable one to study the electron-stability of organic fluorophores and fluorescent proteins. The initial results on cathodoluminescence bleaching of organic fluorophores reported by Niitsuma
In conclusion, the method of SCLEM offers a fast and easy method for correlative microscopy. The same area of the sample can be illuminated by both light and electron microscope at the same time. This removes complications related to retrieval of regions of interest or the definition of fiducial markers from the correlative workflow. Inspection times are reduced to the order of minutes, there is no risk of sample contamination or damage as a result of transfer between microscopes, and a user can switch between both modalities during inspection of a region of interest. Importantly, large areas can be inspected without re-evaluation of the overlay between both images and without the need for stitching images from different areas.
We have demonstrated SCLEM with a high-NA objective lens, which allows for quantitative fluorescence microscopy in correlation to cellular ultrastructure. Equivalently, SCLEM could be performed with a large field-of-view low-NA objective lens if fluorescence labeling is solely used as a marker to track rare events suitable for EM investigation. The described implementation of SCLEM with a high-NA objective lens could be used with different optical modalities, including superresolution microscopy. We have shown SCLEM on coverglass-mounted tissue sections, as well as on whole, uncoated cells without any EM-specific staining. In the latter case, protein distributions measured in fluorescence can be correlated to the growth and size of extrusions and protrusions of the cell membrane. Thus, SCLEM could be a valuable method in the investigation of cell motility and cell-cell signaling. The ease of use and versatility of SCLEM may enable the widespread application of quantitative correlative microscopy in biology and biomedicine.
All imaging experiments were done on in-house developed optical microscope integrated in a commercial SEM (Quanta™ 200 FEG microscope (FEI, Eindhoven, The Netherlands)) as described above. SEM images were made at standard high-vacuum settings with varying acceleration voltages and different magnifications as stated in the manuscript. An Everhart-Thornley detector and a solid-state backscatter detector were used for SE and BSE detection, respectively.
Fluorescence imaging was done at room temperature using the custom made epi-fluorescence microscope which has an objective lens mounted just beneath the sample holder in the SEM chamber. The epifluorescence microscope was equipped with a 470 nm LED light source (Thorlabs M470L2-C), a CCD camera (Photometrics CoolSNAP, Tucson, Arizona, USA) and an 100X 1.4 NA objective lens (Nikon CFI Plan Apochromat VC 100x). The objective lens was tested for vacuum compatibility prior to first use. The light from the LED source passes through a collimator lens (Thorlabs LED collimator for Nikon microscopes), a planoconvex lens to focus the beam in the back-focal plane of the objective, a band-pass filter (Newport Spectra-Physics 10XM20-485), a dichroic mirror (Semrock FF506-Di03), and then through a 10 mm thick, 50 mm diameter, 425–675 nm anti-reflection coated BK7 glass window (CVI Melles Griot) into the SEzM vacuum chamber. The detection path further consists of a long-pass filter (Semrock BLP01-488R), and a standard Nikon 1X tube lens. Vacuum-compatible immersion oil was supplied by DELMIC BV (Delft, the Netherlands).
Colorectal cancer (CRC) cell line SW480 (ATCC, UK) were maintained in Dulbecco’s Modified Eagles Medium (DMEM) from Gibco Invitrogen, supplemented with penicillin (50 U/ml) and streptomycin (50 µg/ml) and 10% fetal calf serum(FCS). CRC cell line HCT116 SMAD4−/− cell line used for Cortactin labeling (obtained from Dr. B. Vogelstein - John Hopkins, Baltimore)
ITO-coated microscope slides (thickness #1, 22×22 mm with 8–12 Ωsq−1 or 22×40 mm with 70–100 Ω sq−1; SPI Supplies, West Chester, PA, U.S.A.) were washed with ethanol and water, placed in 12-well tissue culture dishes with the conductive side upwards and washed with culture medium. The cells were 2x times washed with Phosphate Buffered Saline (PBS), then trypsinized and seeded onto the ITO coated glass slides as 2 mL per well. Cells were cultured for 16–24 h at 37°C. Cells grown on ITO-coated glass at a confluency of 50%, were then washed twice with PBS containing 0.5 mM MgCl2, fixed for 10 minutes with a mixture of 2.5% paraformaldehyde and 1.25% glutaraldehyde in PBS, pH 7.4. Samples were washed 3 times with PBS after fixation.
Staining actin with phalloidin (Alexa Fluor 488 phalloidin; Invitrogen, Carlsbad, CA) was performed according to manufacturer's instructions. 5 µL 6.6 µM stock solution was diluted into 200 µL PBS for each coverslip and 1% bovine serum albumin (BSA) was added to the staining solution to reduce nonspecific background staining. The staining was carried on for 30 minutes at room temperature and afterwards samples were washed 3 times with PBS.
For immunolabeling of cortactin, cells were pre-incubated with PBS with 1% BSA and 0.1% Triton for 10 min, then incubated with the primary antibody in PBS/BSA/Triton for 1hr at dilution 1∶200 at room temperature. Cells were washed 3 times with PBS containing 1% BSA and 0,1% Triton. The cells were then incubated with the secondary antibody dissolved 1∶200 in PBS/1% BSA/0,1% Triton for 30 minutes at room temperature and then washed again 3 times with PBS containing 1% BSA and 0,1% Triton. The primary antibody used was Anti-Cortactin (p80/85) (mouse), clone 4F11(Millipore, MA, USA) and the secondary antibody was Alexa fluor 488 goat anti-mouse IgG (H+L) (Invitrogen,NY, USA). After labeling the samples were 3 times washed with dH2O and left in dH2O at 4°C overnight to remove any remaining salt residue from the sample.
The samples were air dried. Before imaging, conductive carbon tape was used to connect the upper, ITO-coated side of the microscope slides holding the sample to the sample holder of the SCLEM platform.
One of the authors (R. A. W.) took samples of human skin from his own arm using standard 2 mm biopsy punches. The tissue samples were high-pressure frozen, freeze-substituted in acetone, and embedded in HM20. During freeze-substitution it was stained with osmium tetroxide, uranyl acetate, and DiIC18. Freeze-substitution was performed as follows: 27 hours at −90°C, temperature rise to −60°C at 10°C/hour, 6 hours at −60°C, temperature rise to −40°C at 10°C/hour, 5 hours at −40°C. Then the stains were washed out and infiltration was started with HM20 (30% and 70% in ethanol, and then 100% overnight). Polymerization was done with UV-light at −40°C for 3 days. 100nm sections were cut and transferred to ITO-coated thickness #1 glass cover slides. Before imaging, they were connected to the sample holder of the SCLEM platform with conductive carbon tape.
We would like to thank Sjoerd Stallinga and Sander den Hoedt for helpful discussions, Ruud van Tol, Frans Berwald, Ger Schotte, and Cor Barends for technical support.