This protocol describes the cryoAPEX method, in which an APEX2-tagged membrane protein can be localized by transmission electron microscopy within optimally-preserved cell ultrastructure.
Key cellular events like signal transduction and membrane trafficking rely on proper protein location within cellular compartments. Understanding precise subcellular localization of proteins is thus important for answering many biological questions. The quest for a robust label to identify protein localization combined with adequate cellular preservation and staining has been historically challenging. Recent advances in electron microscopy (EM) imaging have led to the development of many methods and strategies to increase cellular preservation and label target proteins. A relatively new peroxidase-based genetic tag, APEX2, is a promising leader in cloneable EM-active tags. Sample preparation for transmission electron microscopy (TEM) has also advanced in recent years with the advent of cryofixation by high pressure freezing (HPF) and low-temperature dehydration and staining via freeze substitution (FS). HPF and FS provide excellent preservation of cellular ultrastructure for TEM imaging, second only to direct cryo-imaging of vitreous samples. Here we present a protocol for the cryoAPEX method, which combines the use of the APEX2 tag with HPF and FS. In this protocol, a protein of interest is tagged with APEX2, followed by chemical fixation and the peroxidase reaction. In place of traditional staining and alcohol dehydration at room temperature, the sample is cryofixed and undergoes dehydration and staining at low temperature via FS. Using cryoAPEX, not only can a protein of interest be identified within subcellular compartments, but also additional information can be resolved with respect to its topology within a structurally preserved membrane. We show that this method can provide high enough resolution to decipher protein distribution patterns within an organelle lumen, and to distinguish the compartmentalization of a protein within one organelle in close proximity to other unlabeled organelles. Further, cryoAPEX is procedurally straightforward and amenable to cells grown in tissue culture. It is no more technically challenging than typical cryofixation and freeze substitution methods. CryoAPEX is widely applicable for TEM analysis of any membrane protein that can be genetically tagged.
Biological studies often include questions of resolving subcellular protein localization within cells and organelles. Immunofluorescence microscopy provides a useful low-resolution view of protein localization, and recent advances in super-resolution imaging are pushing the bounds of resolution for fluorescently tagged proteins1,2,3. However, electron microscopy (EM) remains the gold standard for imaging high-resolution cellular ultrastructure, though the labeling of proteins is a challenge.
Historically, several EM methods have been used to approach questions of ultrastructural protein localization. One of the most commonly utilized methods is immunoelectron microscopy (IEM), where antigen-specific primary antibodies are used to detect the protein of interest. EM signal is generated by the application of secondary antibodies conjugated with electron-dense particles, most commonly colloidal gold4,5. Alternately, antibodies conjugated with enzymes such as horse radish peroxidase (HRP) can be used to produce an electron-dense precipitate6,7,8. Two main approaches exist for IEM, termed pre-embedding and post-embedding labeling. In pre-embedding IEM, antibodies are introduced directly into cells, which necessitates light fixation and permeabilization of the cells9,10,11. Both steps can damage ultrastructure12,13. Development of significantly smaller antibodies consisting of an antibody Fab fragment conjugated with 1.4 nm nanogold allows very gentle permeabilization conditions to be used; however, nanogold is too small for direct visualization under TEM and requires additional enhancement steps to become visible14,15,16. In post-embedding IEM, antibodies are applied on thin sections of cells which have been fully processed by fixation, dehydration, and embedding in resin17. While this approach avoids the permeabilization step, preserving the epitope of interest throughout sample preparation is challenging18,19,20. The Tokuyasu method of light fixation followed by freezing, cryo-sectioning, and antibody detection provides improved epitope preservation21,22. However, the technical requirements of cryo-ultramicrotomy, as well as the sub-optimal contrast achieved in the cell, are disadvantages23.
The use of genetically encoded tags eliminates many of the difficulties of IEM related to detection of the protein of interest. A variety of tags are available, including HRP, ferritin, ReAsH, miniSOG, and metallothionein24,25,26,27,28,29,30,31,32. Each of these has advantages over previous methods, but each also has drawbacks preventing widespread use. These drawbacks range from inactivity of HRP in the cytosol to the large size of the ferritin tag, light sensitivity of ReAsH, and small size and lack of compatibility with cellular staining of metallothionein. Recently, a protein derived from ascorbate peroxidase has been engineered as an EM tag, named APEX233,34. As a peroxidase, APEX2 can catalyze the oxidation of 3,3' diaminobenzidine (DAB), producing a precipitate that reacts with osmium tetroxide to provide local EM contrast with minimal diffusion from the protein of interest (less than 25 nm)33,35. Unlike traditional HRP-based methods, APEX2 is extremely stable and remains active in all cellular compartments33. Samples can be processed for TEM using traditional EM sample staining and methods that allow good visualization of the surrounding structures33,34,36. Because of its small size, stability, and versatility, APEX2 has emerged as an EM tag with great potential.
Many of the approaches discussed above either cannot be or have not yet been combined with the current state of the art in ultrastructural preservation, cryofixation and low-temperature freeze-substitution. Thus, they suffer from a lack of membrane preservation and/or cell staining to determine accurate protein localization. This necessarily limits the resolution and interpretation of the data that can be obtained. Cryofixation by high pressure freezing (HPF) involves rapid freezing of samples in liquid nitrogen at a high pressure (~2,100 bar), which causes vitrification rather than crystallization of aqueous samples, thus preserving cells in a near-native state37,38,39. HPF is followed by freeze substitution (FS), a low temperature (-90 °C) dehydration in acetone combined with incubation with typical EM stains such as osmium tetroxide and uranyl acetate. HPF and FS together provide a distinct advantage over traditional chemical fixation (a longer process which can lead to artefacts) and alcohol dehydration at room temperature or on ice (which can lead to extraction of lipids and sugars), and thus are desirable to combine with the best EM tags for protein detection.
One reason that HPF/FS has not been combined with APEX2 labeling is that light chemical fixation is a prerequisite for the peroxidase reaction, limiting the diffusion of the DAB reaction product. In APEX2 studies thus far, fixation and peroxidase reaction are followed by traditional EM methods for staining and alcohol dehydration33,36. However, it has been shown that following chemical fixation with HPF/FS provides a distinct advantage in preservation over traditional chemical fixation and alcohol dehydration alone40. The loss of ultrastructural integrity seen in traditional TEM samples appears less connected to fixation than to dehydration, which is typically done using alcohol at room temperature or on ice, and can lead to extraction of lipids and sugars40,41. To develop the cryoAPEX method, we hypothesized that chemical fixation and peroxidase reaction, followed by HPF and FS, would produce an optimal result in terms of ultrastructural preservation.
Here we present the cryoAPEX protocol, which combines APEX2 tagging with cryofixation and freeze substitution methods (Figure 1). This straightforward protocol consists of transfection of an APEX2-tagged protein of interest, chemical fixation of cells, and the peroxidase reaction. HPF and FS are then performed followed by typical resin embedding and thin sectioning. TEM imaging reveals excellent preservation of ultrastructure using this method. Additionally, high-resolution subcellular localization and spatial distribution of an endoplasmic reticulum (ER) lumenal protein were observed. This method is widely useful for detection of membrane protein localization within cells for electron microscopy analysis. In our hands, the method has worked successfully for a variety of cell lines grown in tissue culture, including HEK-293T (human embryonic kidney), HeLa (human cervical cancer), Cos7 (African green monkey kidney fibroblast), and BHK (baby hamster kidney). Detailed instructions are described below using HEK-293T cells.
1. Cell Culture and Transfection
2. Chemical Fixation and Peroxidase Reaction
3. High Pressure Freezing
4. Freeze Substitution
CAUTION: Use proper safety procedures and personal protective equipment when working with liquid nitrogen. Additionally, many of the chemicals utilized in step 4 are toxic, including tannic acid, osmium tetroxide, and uranyl acetate. These chemicals must be handled according to proper safety procedures and disposed of as hazardous chemical waste.
5. Resin Infiltration and Embedding
CAUTION: The resin used here (see Table of Materials) is toxic prior to polymerization, and should be handled with proper safety procedures and personal protective equipment. Any unpolymerized resin should be disposed of as hazardous chemical waste.
6. Sectioning
7. TEM Imaging
In order to compare the ultrastructural preservation using the cryoAPEX method with traditional fixation and dehydration, we prepared samples in which an endoplasmic reticulum membrane (ERM; ER membrane) peptide was tagged with APEX2 and transfected into HEK-293T cells. ERM-APEX2 localizes to the cytoplasmic face of the ER and remodels the ER structure into morphologically distinct structures known as organized smooth ER (OSER)34,42,43. OSER morphology includes regions of smooth, parallel, densely stacked membranes which serve as an optimal region to compare ultrastructural preservation. Preparation of the sample by traditional APEX methods resulted in clear labeling of OSER structures (Figure 2A–D). Upon inspection at high magnification, the stacked membranes appeared ruffled and non-uniform gaps were present between concentric membrane densities, indicating poor membrane preservation and lipid extraction (Figure 2D). The sample prepared by cryoAPEX also had clearly defined labeling of OSER structures; however, the membranes were smooth and parallel, and little to no lipid extraction was seen (Figure 2E–H). The results from cryoAPEX were of similar high quality preservation to those obtained from a sample which underwent HPF/FS without the additional chemical fixation and APEX2/DAB reaction steps (Figure 2I–K).
In addition to visually appreciable membrane preservation, the cryoAPEX method preserves the protein of interest such that aspects of protein distribution patterns may be observed in some cases. To illustrate this point, we used another ER-localized protein, huntingtin yeast interacting protein E (HYPE). HYPE is a membrane protein located on the luminal face of the ER membrane 44,45,46,47. HYPE-APEX2 constructs were overexpressed in HEK-293T cells. TEM analysis of 90 nm thin sections revealed that HYPE was present throughout the peripheral ER as well as the nuclear envelope (Figure 3A,B). Additionally, the HYPE density was able to be resolved into regularly spaced foci along the lumenal ER membrane (Figure 3C, arrows). HYPE distribution and foci were also visible in a sample prepared with traditional fixation and dehydration; however, extensive membrane disruption and extraction were present, making the sample suboptimal (Figure 3D,E).
To demonstrate the robust organellar specificity and applicability of the cryoAPEX method for a range of tagged proteins, we performed APEX2 labeling using three cellular markers. Mitochondria were labeled using mito-V5-APEX234. This marker of the mitochondrial matrix provided specific staining of mitochondria only (Figure 4A). Likewise, we assessed plasma membrane labeling using CAAX-APEX234, which produced distinct staining of the plasma membrane only (Figure 4C). No labeling was observed in intracellular organelles (Figure 4C). Additionally, we created a new construct as a marker for the Golgi lumen by fusing the first 118 amino acids of the mouse isoform of α-mannosidase with the APEX2 gene48. The resulting MannII-APEX2 was transiently transfected into cells which were subsequently prepared by the cryoAPEX method. Stained Golgi stacks were clearly distinguishable from the surrounding organelles (Figure 4B). Individual stacks, cisternae, and some vesicles were labeled, typical of Golgi staining (Figure 4B). Altogether, these markers demonstrate that the cryoAPEX method provides specific labeling of membrane proteins within various organelles at high enough resolution to distinguish them from surrounding sub-cellular structures.
Figure 1: Schematic of the essential steps in the CryoAPEX protocol. (A) Cells are grown and transfected with an APEX2 plasmid. (B) Cells are pelleted and fixed with glutaraldehyde, followed by (C) incubation with DAB and hydrogen peroxide to produce the peroxidase reaction product. (D) The pellet is cryofixed by HPF, (E) freeze substituted with heavy metals and acetone, and (F) embedded in resin. Thin sections are collected on the microtome. (G) TEM imaging is performed and additional contrast may be added by post-staining. Please click here to view a larger version of this figure.
Figure 2: Comparison of OSER membrane preservation using traditional chemical fixation, cryoAPEX, and HPF/FS. The reorganized ER morphology in chemically fixed, DAB reacted ERM–APEX2-expressing cells that were processed via traditional chemical fixation and alcohol dehydration (A–D) or by cryoAPEX (E–H) was compared to ERM–APEX2 expressing cells that were cryofixed live and without the DAB reaction (I–K). The live cryofixed cells represent the best attainable ultrastructural preservation and serve here as the metric for evaluating membrane preservation obtained via the two APEX-based detection protocols (A–H). The evenly-spaced parallel lamellar stacking of the ER derived membranes obtained by cryoAPEX (exemplified in panels G and H), as opposed to the ruffled membranes obtained by traditional methods (panels C and D), highlights the superior membrane preservation obtained by cryoAPEX. This figure has been modified from Sengupta et al. 201948. Please click here to view a larger version of this figure.
Figure 3: Protein localization of an APEX2-tagged ER membrane protein can be resolved into periodic foci. (A) An image of a thin section of a HEK-293T cell expressing HYPE–APEX2 and processed by cryoAPEX reveals staining of the ER in a well-preserved (dense) cytoplasmic background (B,C). Higher magnification images of a small section of the peripheral ER (demarcated by yellow box in A and shown in B, with further magnification of red box in B shown in C) exhibits periodic foci of APEX2-generated density (B, red box and C, white arrowheads showing periodicity between the HYPE foci). (D) Image of a thin section of a cell expressing HYPE-APEX2 and prepared by traditional chemical fixation and dehydration shows specific staining of the cortical ER and the nuclear envelope (red arrows). (E) At a higher magnification, periodic HYPE-specific foci were apparent within stretches of the ER (yellow box and white arrow heads in the inset), despite extensive membrane disruption, indicated by red arrows. NE = nuclear envelope. This figure has been modified from Sengupta et al. 201948. Please click here to view a larger version of this figure.
Figure 4: Organelle markers show specificity of the signal obtained from APEX2-tagged proteins. APEX2-tagged protein constructs designed to localize to the mitochondrial matrix (mito-V5-APEX2; shown in A), or the Golgi lumen (α-mannII-APEX2; shown in B), or the plasma membrane (CAAX-APEX2; shown in C) were transiently expressed in HEK293 cells and the samples processed by cryoAPEX. Each construct yielded organelle specific densities. Magnified views of two sections (yellow or red boxes) from the cells expressing α-mannIIAPEX2 (panel B) or CAAX-APEX2 (panel C) are shown. This figure has been modified from Sengupta et al. 201948. Please click here to view a larger version of this figure.
The cryoAPEX protocol presented here provides a robust method to characterize the localization of membrane proteins within the cellular environment. Not only does the use of a genetically encoded APEX2 tag provide precise localization of a protein of interest, but the use of cryofixation and low-temperature dehydration provides excellent preservation and staining of the surrounding cellular ultrastructure. Combined, these approaches are a powerful tool for localizing a protein with high precision within its subcellular context.
The crux of the advancement of this method is the fact that the loss of ultrastructure experienced after preparation by traditional TEM methods comes primarily from the dehydration step rather than the fixation step40. It was previously believed that peroxidase-based methods were incompatible with HPF/FS because they require chemical fixation prior to the peroxidase reaction. To work around this, a protocol named CryoCHEM was recently developed in which samples are initially cryofixed, followed by rehydration and the peroxidase reaction49. This approach provides excellent target localization with significant improvements in sample staining and preservation. It has been shown to be useful for tissue samples and in cases where correlative fluorescence and electron microscopy is desired. In parallel to cryoCHEM, our method combines glutaraldehyde fixation with HPF and FS. CryoAPEX offers a streamlined protocol that works effectively even for small cellular samples.
Access to high pressure freezing and freeze substitution instruments is crucial to the cryoAPEX method. These instruments and skills are increasingly common in EM facilities. Even if HPF and FS equipment are not readily available, the chemically fixed sample is stable enough for a short amount of time to be transported modest distances40. We have found that samples can be stored after the DAB reaction at 4 °C for at least 48 hours prior to HPF without significant loss of quality. Another critical aspect of the cryoAPEX protocol is the inclusion of controls, which are essential for a robust experiment and convincing results. Samples prepared by transient transfection with efficiency less than 100% will contain negative control cells as well as labeled cells within the same sample. If using cell lines with stable expression of APEX2, a separate negative control should be prepared by transfection of cells with the non-APEX2 labeled protein of interest. Several constructs that can serve as organellar controls are available through Addgene, and published images are available in this and other publications that can be used for verification33,34,36,48. In-depth discussion of experimental design and verification of new APEX2 fusion constructs has been provided by Martell et al.36
While cryoAPEX is broadly useful for the detection of membrane proteins, some limitations exist. Although APEX2 is a small 28 kDa protein, some proteins may not be able to incorporate the tag33,34. APEX2 is not considered useful for labeling soluble proteins in the cytosol, due to the diffuse reaction product33,36. Additionally, the detection of small quantities of protein poses a challenge due to the presence of staining in the surrounding cell. Preparation by HPF and FS preserves cellular components which are extracted by traditional fixation and dehydration. This leads to overall darker staining in the cell, potentially competing with low levels of APEX2 labeling.
The cryoAPEX technique is widely applicable to many proteins, with a limited number of steps that may require optimization. First, due to individual variability among proteins, the protein expression level and/or DAB reaction time may need to be adjusted in order for the signal to be visualized above the background staining of the cell. Helpful information and protocols for validation of new APEX2 fusion constructs and optimization of the expression and DAB staining are provided by Martell et al.36 From a cellular staining perspective, the FS protocol and/or chemicals may need to be adjusted for optimal visualization of different organelle membranes, within different cell types, tissues, or organisms50,51,52,53,54. In our experience, the FS conditions presented here have worked well for a variety of mammalian cell lines.
The hybrid approach of cryoAPEX has the potential to be applied to many other genetic labeling techniques. Replacing traditional alcohol dehydration with HPF/FS is expected to greatly improve the ultrastructural preservation and the protein localization information. Utilizing sapphire discs as a cell substrate to fix cells as a monolayer improves the preservation of the cell periphery, including the cytoskeleton and cell-cell contacts. Minor modifications to the protocol would be required to use sapphire discs. APEX technology can be used to detect green fluorescent protein (GFP) tagged proteins via a GFP-binding peptide35. This indirect method of detection opens up the potential to utilize APEX technology for the myriad proteins already tagged with GFP. The recently-introduced split APEX2 will be advantageous for proximity and interaction studies55. Additionally, existing HRP-based methods can be combined with HPF/FS to improve cellular preservation. One example is fluorescent indicator and peroxidase for precipitation with EM resolution (FLIPPER), in which individual cell markers have been fused with both a fluorescent tag and HRP, providing lumenal markers for Golgi or ER56. Use of improved peroxidase substrates in place of DAB is also possible with this method, including substrates which are optimized for RNA labeling 57. CryoAPEX also provides in-cell labeling and ultrastructural preservation necessary for three dimensional analysis of protein distribution through electron tomography, and potentially at high volumes through SBF-SEM or FIB-SEM48,58.
Overall, CryoAPEX is a robust method with wide applicability. In principle, it can be applied to any membrane protein, whether within the lumenal space of an organelle, on the cytoplasmic face, within vesicles, on the cell's plasma membrane or even in the extracellular space. For this vast range of membrane proteins, the cryoAPEX method provides the potential to see the localization and distribution of a protein with accuracy within its subcellular context.
The authors have nothing to disclose.
The protocol described here stems from a publication by Sengupta et al., Journal of Cell Science, 132 (6), jcs222315 (2019)48. This work is supported by grants R01GM10092 (to S.M.) and AI081077 (R.V.S.) from the National Institutes of Health, CTSI-106564 (to S.M.) from Indiana Clinical and Translational Sciences Institute, and PI4D-209263 (to S.M.) from the Purdue University Institute for Inflammation, Immunology, and Infectious Disease.
3,3'-Diaminobenzidine tetrahydrochloride hydrate | Sigma-Aldrich | D5637-1G | |
Acetone (Glass Distilled) | Electron Microscopy Sciences | 10016 | |
Beakers; Plastic, Disposable 120 cc | Electron Microscopy Sciences | 60952 | |
Bovine Serum Albumin | Sigma-Aldrich | A7906-100G | |
Cryogenic Storage Vials, 2 mL | VWR | 82050-168 | |
Dulbecco's Modified Eagle's Medium | Corning | 10-017-CV | |
Durcupan ACM Fluka, single component A, M epoxy resin | Sigma-Aldrich | 44611-500ML | |
Durcupan ACM Fluka, single component B, hardener 964 | Sigma-Aldrich | 44612-500ML | |
Durcupan ACM Fluka, single component C, accelerator 960 (DY 060) | Sigma-Aldrich | 44613-100ML | |
Durcupan ACM Fluka,single component D | Sigma-Aldrich | 44614-100ML | |
Embedding mold, standard flat, 14 mm x 5 mm x 6 mm | Electron Microscopy Sciences | 70901 | |
Embedding mold, standard flat, 14 mm x 5 mm x 4 mm | Electron Microscopy Sciences | 70900 | |
Fetal Bovine Serum; Nu-Serum IV Growth Medium Supplement | Corning | 355104 | |
Glass Knife Boats, 6.4 mm | Electron Microscopy Sciences | 71008 | |
Glass Knifemaker | Leica Microsystems | EM KMR3 | |
Glutaraldehyde 10% Aqueous Solution | Electron Microscopy Sciences | 16120 | |
HEK 293 Cells | ATCC | CRL-1573 | |
High Pressure Freezer with Rapid Transfer System | Leica Microsystems | EM PACT2 | Archived Product Replaced by Leica EM ICE |
Hydrogen Peroxide 30% Solution | Fisher Scientific | 50-266-27 | |
Lipofectamine 3000 Transfection Reagent | ThermoFisher Scientific | L3000015 | |
Membrane carrier for EM PACT2, 1.5 mm x 0.1 mm | Mager Scientific | 16707898 | |
Osmium Tetroxide, crystalline | Electron Microscopy Sciences | 19110 | |
Phosphate Buffered Saline (PBS) 20X, Ultra Pure Grade | VWR | 97062-950 | |
Plastic Capsules for AFS/AFS2, 5 mm x 15 mm | Mager Scientific | 16702738 | |
Slot grids, 2 x 1 mm copper with Formvar support film | Electron Microscopy Sciences | FF2010-Cu | |
Sodium Cacodylate Buffer, 0.2 M, pH 7.4 | Electron Microscopy Sciences | 102090-962 | |
Sodium Hydroxide, Pellet 500 G (ACS) | Avantor Macron Fine Chemicals | 7708-10 | |
Tannic Acid | Electron Microscopy Sciences | 21710 | |
Tissue Culture Dishes, Polystyrene, Sterile, Corning, 100 mm | VWR | 25382-166 | |
Ultra Glass Knife Strips | Electron Microscopy Sciences | 71012 | |
Ultramicrotome | Leica Microsystems | EM UC7 | |
Uranyl Acetate Dihydrate | Electron Microscopy Sciences | 22400 |