A meeting of microscopic strengths

Integrated correlative light and electron microscopy (CLEM) offers the possibility to study the same area on a sample using both fluorescence and electron microscopy. One of the challenges associated with integrated CLEM (iCLEM) is the preparation of samples suitable for both FM and EM. During the past few years, correlative light and electron microscopy (CLEM) has gained in popularity as a research tool. This growing interest is because CLEM combines the strengths of fluorescence microscopy (FM) and electron microscopy (EM). FM is the ideal tool to collect functional information about specific components inside tissues, cells and organelles while EM offers substantially higher resolution and can provide detailed contextual structural information. FM can thus be used to point to regions of interest for subsequent higher resolution EM. Until recently, CLEM has been challenging, costly, time consuming and required a high level of operator expertise. Correlative methods normally require two distinctly different imaging setups which are traditionally located in separate facilities. The sample preparation methods for each tend to be incompatible. Due to the fundamental differences between microscopes, extra sample preparation steps are also usually necessary when switching from FM to EM. This often distorts the sample, hampering accurate correlation. In addition, it can be extremely challenging to relocate a region of interest originally identified with FM in EM since the information used to navigate in FM is not visible in EM, and this problem becomes more significant as the size of specimen increases. Integrated CLEM (iCLEM) overcomes most of these difficulties. By integrating fluorescence and scanning electron microscopy, the need to transfer between two different microscopes is eliminated. To get back to a region of interest becomes straightforward as the same area of the sample is observable with both microscopes. Furthermore, since the sample is not subjected to intermediate preparation steps, its conformation is guaranteed to be identical. Sample preparation for iCLEM has become a new research area with a limited number of protocols published to date. For non-integrated CLEM, excellent overviews of sample preparation methods exist^1-3. One of the difficulties of integrating sample preparation for FM and EM is that EM sample preparation protocols typically use heavy metal stains to introduce electron contrast. It is well known that these heavy metals can quench nearby fluorescence. Furthermore, EM requires vacuum compatible samples. As such, samples need to be dried which can influence the amount of fluorescence for hydration-sensitive dyes⁴. A recent publication has also shown that in SEM the amount of fluorescence can be influenced by variation of the vacuum pressure⁵. As iCLEM develops further as a powerful research tool, there will be a corresponding increase in the number of published iCLEM specific sample preparation protocols. Here, we present four different protocols for iCLEM. We have deliberately chosen to use examples with different types of samples and varied preparation techniques. Another consideration was to focus on (where possible) relatively simple protocols, thereby allowing many alterations and variations to be made. This is intended as an introduction into the possibilities of sample preparation for iCLEM. Our goal is to demonstrate that there are no fundamental limitations for integrated sample preparation. We also show that iCLEM can be extremely fast and, crucially, that it is possible to generate very accurate overlays with no additional image manipulation. Methods and Results We used the SECOM platform (DELMIC B.V.) for correlative imaging. The sample mounting procedure is therefore specifically designed for this platform. Samples are placed (or grown) directly on cover glasses coated with indium tin oxide (ITO). A thin coating of ITO is transparent to visible light and is conductive allowing imaging of uncoated biological samples in a scanning electron microscope (SEM)⁶. The results from each protocol are presented in Figure 1. More experimental detail may be found clicking here. Songbird brain Understanding synaptic connectivity is essential to extending our knowledge of neural mechanisms. The combination of EM, capable of resolving synaptic vesicles and post-synaptic densities, and fluorescent markers allows synapses observed in the EM to be associated with specific neuron types⁷. It is interesting to note that even though the EM staining used in this study quenched the initial fluorescence of the tracers, the tracer was able to be relabelled after sectioning using fluorescent antibodies, demonstrating that the protocol preserved antigenicity well enough to allow for on-section immunolabelling. Though this protocol is specific to neurological samples, it is a very interesting application and could be adapted for other applications. HeLa cells In this study, the goal was to investigate the distribution of the lipid diacylglycerol within cellular membranes ⁵. To do this, a protocol was developed that preserves GFP and mCherry fluorescence whilst retaining electron contrast in resin-embedded sections.  A full description and details of the different embedding media that were used is described in the original research article⁵. One of the interesting findings is that the authors argue that the use of a quick freeze substitution protocol⁸ might be essential to preserve the fluorescence of GFP for iCLEM. Zebrafish Heavy metal staining clearly influences fluorescence so we decided to experiment with sample preparation protocols without any. In this way, the fluorescence signal is optimally preserved. The adverse effect is that the level of electron contrast is significantly reduced. Nevertheless, it is clear that there is still a considerable amount of contrast in EM mode. The cell membranes, however, are not visible. Subsequent experiments are being performed with low amounts of osmium tetroxide. In the FM mode, the GFP and E2-Crimson signals were easily detectable. It is noted that the red structures in Figure 1 are actually pigment cells and not due to specific labelling. The mCherry signal was present in other parts of the embryo. HUVEC Human umbilical vein endothelial cells (HUVEC) contain rod-like storage granules called Weibel-Palade bodies which contain Von Willebrand factor (VWF). These organelles play an important role in blood coagulation. Here, the goal was to image these rod-like structures in the thin parts of the cell where they can be seen under the cell membrane using the SEM. We used a very fast sample preparation protocol where fixation, immunolabeling, dehydration and correlative imaging were performed in one day. Since whole cells typically display good enough contrast when imaged at low accelerating voltages, no additional EM staining was used⁹. The fluorescent signal was preserved remarkably well, and after storing the dried sample in a refrigerator for a month, the samples still displayed enough fluorescence for imaging. [caption id="attachment_40687" align="alignright" width="400"] Correlative light and electron micrographs using the SECOM platform (DELMIC B.V., Delft) installed on a Quanta 250 FEG (FEI Company, Eindhoven). 1st row: fluorescence image. 2nd row: scanning electron micrograph. 3rd row: overlay of FM and EM. Columns 1 to 4: projection neurons in songbird brain, HeLa cell expressing GFP-C1, Zebrafish and human umbilical vein endothelial cells labelled for Von Willebrand factor. Columns 1, 3 and 4: EM imaging using secondary electron detector and FM imaging with Nikon Plan Apo 60x /0.95 lens, multicolour LED light engine, Clara CCD camera (Andor Technology, Belfast). Column 2: EM imaging using the vCD backscatter detector and FM imaging with Nikon Plan Apo 100x /1.40 oil immersion lens using vacuum compatible immersion oil, laser light source, Zyla sCMOS camera (Andor Technology, Belfast).[/caption] We have illustrated different sample preparation possibilities for iCLEM, each of which uses a different approach on a variety of samples. These methods demonstrate that it is possible to find an integrated sample preparation solution for a wide range of applications. The neurology application example provides a good demonstration for on-section immunolabelling of resin embedded material, whilst the HeLa cell example shows that it is possible to retain GFP fluorescence in resin using freeze-substitution. It is clear that the protocol used for Zebrafish is very much a work in progress with many opportunities for improvement. Nevertheless, we included this protocol together with that for HUVECS since each protocol shows that even without additional contrast enhancement, the level of detail available using EM is sufficient. Furthermore, these protocols demonstrate the potential of straightforward sample preparation methods for iCLEM. Depending on the specific research question of interest, these protocols can be modified or extended to deliver valuable results. Although fluorescence preservation and intensity is clearly influenced by the restraints of an integrated approach, we have shown that a suitable balance between EM and FM contrast can be found. Where this balance lies and how it can be achieved will obviously vary for each experiment, and as such, it is advisable to optimise sample preparation protocols for each application. The main advantages of integrated preparation methods are the absence of intermediate specimen preparation steps when moving from FM to EM, and the high degree of correlation accuracy with minimal or no image manipulation over both small and large fields of view. As such, integrated solutions offer a streamlined imaging workflow which is both faster and more accurate. Acknowledgements Songbird brain samples were kindly provided by Dr Thomas Templier and Prof Dr Richard H R Hahnloser, University of Zurich and ETH Zurich. Sample preparation and imaging of transfected HeLa cells was performed by Dr Christopher J Peddie and Dr Lucy M Collinson, Cancer Research UK London Research Institute. Zebrafish samples were provided by Rohola Hosseini and Gerda Lamers, Leiden University. We gratefully acknowledge the help of Marjon J. Mourik, Leiden University Medical Center, with the sample preparation of HUVECs. References 1. Robert Kirmse and Eric Hummel. Correlative Microscopy Protocols. Carl Zeiss Microscopy GmbH, June 2013. Web. 28 April 2014 2. Published Sample Preparation Protocols. FEI Company. Web. 28 April 2014 3. Thomas Müller-Reichert and Paul Verkade, eds. Correlative Light and Electron Microscopy. Vol. 111. Academic Press, 2012. 4. Matthia A. Karreman, et al. “Optimizing immuno-labeling for correlative fluorescence and electron microscopy on a single specimen.” Journal of structural biology 180.2 (2012): 382-386. 5. Christopher J. Peddie, et al. “Correlative and integrated light and electron microscopy of in-resin GFP fluorescence, used to localise diacylglycerol in mammalian cells.” Ultramicroscopy (2014). 6. Helma Pluk, et al. “Advantages of indium–tin oxide coated glass slides in correlative scanning electron microscopy applications of uncoated cultured cells.” Journal of microscopy 233.3 (2009): 353-363. 7. Daniele Oberti, Moritz A. Kirschmann, and Richard HR Hahnloser. “Correlative microscopy of densely labeled projection neurons using neural tracers.” Frontiers in neuroanatomy 4 (2010). 8. Kent L. McDonald and Richard I. Webb. “Freeze substitution in 3 hours or less.” Journal of microscopy 243.3 (2011): 227-233. 9. Nalan Liv. “Protocol for Simultaneous Correlative Light Electron Microscopy with High Registration Accuracy” PhD Thesis (2014). Authors Lennard Vortman & Sander den Hoedt of DELMIC B.V., and Jacob Hoogenboom of Delft University of Technology