Expansion Microscopy: Toward Nanoscale Imaging of a Diverse Range of Biomolecules

Abstract

Expansion microscopy (ExM) has become a powerful imaging tool for visualizing the nanoscale organization of protein and nucleic acid targets in cells and tissues using only a conventional microscope. Until recently, current ExM approaches have had limited applicability to imaging other biomolecules, such as lipids and small molecules. With the new TRITON probes reported by Wen et al. in this issue of ACS Nano, ExM can now be used to perform nanoscale imaging of the cytoskeleton and lipid membranes. In this Perspective, we offer a brief overview of recent developments in ExM, with a focus on biomolecule anchoring and labeling strategies that target a wide range of biomolecules to the water-swellable polymer formed in situ, a key step that ensures biomolecules or labels of interest are separated in space and can be resolved on a conventional microscope. In addition to these new advancements, we discuss challenges and future directions in this exciting field.

Optical microscopy has been an indispensable tool for imaging biological structures in biomedical research for over a hundred years. Conventional optical microscopes have a resolving power limited to the scale of a few hundred nanometers, dictated by the fundamental optical diffraction limit, which precludes the direct observation of ultra-fine biological structures, such as synapses, mitochondrial substructures, and nuclear pores. The development of far-field super resolution imaging techniques has provided researchers with various means to circumvent this physical limit.^1–3 However, conventional super-resolution imaging techniques often require specialized hardware and/or software, have slow acquisition times, and have limited imaging capabilities for samples of tissues and organs.⁴

Recently, Boyden’s group discovered an alternative approach for overcoming the optical diffraction limit: Rather than harnessing specialized optics or other optical techniques, expansion microscopy (ExM) relies on chemical methods to enlarge preserved biological samples in an even fashion, which in turn increases the effective resolution of an imaging system.⁵ Expansion microscopy protocols, which have been extensively reviewed elsewhere,^6–8 all follow a similar workflow, summarized in Figure 1. First, molecular anchors are either bound to or covalently attached to biomolecules and/or fluorescent labels, which facilitate incorporation of targeted molecules to the hydrogel in the following in situ polymerization step. Next, the tissue sample is incubated with a gelling solution containing monomers, such as sodium acrylate (SA) and acrylamide (AA), and a crosslinker, such as N-N′ -methylenebisacrylamide (BIS). The solution forms a dense and water-swellable polyelectrolyte hydrogel (sodium polyacrylate) (Figure 1, top right) via free radical polymerization. During this process, biomolecules with molecular anchors are chemically linked to the polymer network. Then, the sample is homogenized by either denaturation or enzymatic digestion, such as with proteinase-K, to enable biomolecule separation and expansion in water. Finally, when immersed in water, the polyelectrolyte hydrogel swells due to osmotic pressure, thus separating anchored biomolecules evenly in space. The sample expands typically 4.5-fold as large as the original size in pure water, enabling nanoscale features to be observed on conventional diffraction limited optical microscopes. It is possible to expand samples more than 10-fold by an iterative method,⁹ or by altering gel chemistry, such as using a solution with N,N-dimethylacrylamide (DMAA) and SA (Figure 1, bottom right).¹⁰

Figure 1.

Overview of expansion microscopy (ExM) workflow and commonly used gel chemistries. Left: biomolecules or labels of interest are covalently linked to anchoring agents. During in situ polymer synthesis, these anchoring agents are attached to the polymer matrix. Unanchored molecules are removed during mechanical homogenization. After addition of water, specimens can fully expand. Right: two commonly used gel chemistries for ExM. Top: 4.5-fold expansion gel⁵ composed of sodium acrylate (SA)- and acrylamide (AA)-based backbone crosslinked using N-N′ -methylenebisacrylamide (BIS). Bottom: 10-fold expansion gel¹⁰ composed of a SA and N,N-dimethylacrylamide (DMAA)-based backbone crosslinked with DMAA.

Expansion microscopy is essentially a specialized sample preparation method and, thus, has a unique attribute: ExM can expand the molecule classes available for imaging by adapting new probes and chemistries. The key is either to tag or to modify biomolecules of interest with a probe or a small molecule that binds to the polymer network during in situ polymerization (Figure 2A). The original ExM protocol used a custom trifunctional label composed of a DNA oligonucleotide conjugated with a fluorophore and a methacryolyl group capable of covalently linking to the polymer. In addition, it can hybridize with a complementary oligo attached to an affinity tag, such as an antibody.⁵ Since the original demonstration of ExM, multiple research groups have developed variants that enable the imaging of proteins, such as conventional antibodies or genetically encoded fluorophores,^10–14 and the imaging of nucleic acids, including DNA and RNA labeled with fluorescence in situ hybridization (FISH) probes.^14–17 These ExM variants have rapidly been adapted for imaging protein and nucleic acid identities and organization in a wide range of applications in biology and medicine, such as for visualizing pre- and postsynaptic proteins in Drosphilia,^18,19 imaging synaptic connections and nuclear-cytoskeletal organization in zebrafish,²⁰ observing herpes simplex virus-induced changes in chromatin distribution in cell culture,²² performing three-dimensional mapping of ryanodine receptors in isolated cardiomyocytes,²³ confirming outer mitochondrial membrane localization of a new genetically encoded construct,²⁴ examining human brain tissue from epileptic patients,²¹ and investigating podocyte foot processes in mouse and human kidney tissue,^14,25 among many other biomedical studies.

Figure 2.

Anchoring chemistries for expansion microscopy (ExM). (A) The original ExM protocol.⁵ Schematic of the trifunctional label (left) consisting of a methacryloyl group (blue), which is used to link covalently to the polymer network; a fluorophore; and an oligonucleotide to hybridize to an affinity tag. Example data (right) of ExM applied to YFP-expressing neurons with synapses stained for Bassoon (blue) and Homer1 (red) in mouse brain tissue. Scale bars: (left) 52.7, 42.5, and 35.2 mm (x,y,z); (top right) 13.5, 7.3, and 2.8 mm (x,y,z); (bottom right) 1 mm. Reproduced with permission from ref 5. Copyright 2015 The American Association for the Advancement of Science. (B) Direct protein anchoring. (i) Protein-retention expansion microscopy (ProExM)¹¹ schematic of AcX (left) used to link a protein to the polymer network covalently via a methacryloyl group (blue). Example data (right) of ProExM used in membrane-labeled Brainbow3.0 neurons in mouse hippocampus. Scale bar: 5 μm. Reproduced with permission from ref 11. Copyright 2016 Springer Nature. (ii) Methacrylic acid N-hydroxysuccinimidyl ester (MA-NHS) ExM¹² schematic of MA-NHS (left) used to link proteins covalently to the polymer network. Example data (right) of an expanded BS-C-1 cell stained for tyrosinated tubulin (green) and detyrosinated tubulin (magenta) compared to the pre-expansion image (top). Scale bar: 500 nm. Reproduced with permission from ref 12. Copyright 2016 Springer Nature. (iii) The magnified analysis of the proteome (MAP) protocol.¹³ Schematic of protein anchoring (left) using AA (blue) reacted with formaldehyde during hydrogel-tissue hybridization. Example data (right) of HeLa cells stained for alpha-tubulin. Scale bar: 10 μm. Reproduced with permission from ref 13. Copyright 2016 Springer Nature. (C) Nucleic acid anchoring. Schematic of expansion microscopy fluorescence in situ hybridization (ExFISH)¹⁵ (left) using AcX reacted with Label-IT to form LabelX, which covalently anchors RNA to the polymer network. Example data (right) of NEAT1 lncRNA imaged in a HeLa cell using single-molecule inexpensive fluorescence in situ hybridization (smiFISH) before expansion (top) and using ExFISH (bottom). Insets showing a NEAT1 cluster with smiFISH (left) and ExFISH (right). Scale bars: 2 μm (3.3×) (right) and 200 nm (3.3×) (left). Reproduced with permission from ref 15. Copyright 2016 Springer Nature. (D) Lipid anchoring. Schematic of a custom membrane probe, pGk5b, (left) used in mExM³⁵ to anchor lipid membranes to the polymer network. Example data (right) of pre- (left) and post-expanded (right). Scale bars: 10μm (top) and 5μm (bottom). Reproduced from ref 35, licensed under CC BY-NC-ND.

To enable imaging proteins using ExM, Boyden and co-workers reported protein-retention expansion microscopy (proExM),¹¹ an ExM method that uses succinimidyl ester of 6-((acryloyl)amino)hexanoic acid (acryloyl-X, SE; abbreviated as AcX) to modify amines on proteins with an acrylic group, which then covalently links proteins to the polymer backbone during the radical polymerization process (Figure 2B). Compared to the trifunctional oligo used by the original ExM protocol, proExM requires only commercially available reagents and can be applied to tissues and cells labeled with antibodies, streptavidin, or genetically encoded fluorophores, such as fluorescent proteins, thus greatly increasing the utility of ExM in general biological laboratories. The Vaughan lab independently developed related chemistries and demonstrated the application of methacrylic acid N-hydroxysuccinimidyl ester (MA-NHS) or glutaraldehyde to link fluorescent proteins and antibodies chemically to the polymer (Figure 2B).²⁶ Because MA-NHS is structurally similar to AcX, the MA-NHS protocol has comparable performance in both cultured cells and tissue slices. Glutaraldehyde was demonstrated on cultured cells, but not on intact tissues. Rather than using specialized chemical linkers, the magnified analysis of the proteome (MAP) protocol, developed by the Chung lab, uses modified fixation and homogenization steps.¹³ Magnified analysis of the proteome enables retention of proteins in the gel by adding a high concentration of acrylamide monomer during fixation, which reacts to protein-bound formaldehyde and forms a protein-bound acrylamide moiety. The acrylamide moiety can in turn participate in a free radical polymerization reaction and anchor proteins to the polymer backbone (Figure 2B). In addition, a high concentration of acrylamide monomer prevents protein-bound formaldehyde molecules’ crosslink with another protein, which enables enzyme-free homogenization using heat and a strong surfactant to denature proteins and to facilitate expansion in water. Because no protease is used, the epitopes are retained and can be labeled via immunostaining after expansion.

Researchers have also demonstrated nanoscale imaging of nucleic acids in ExM. Due to the large size of genomic DNA, most of the DNA molecules are retained in the hydrogel after expansion. By applying a set of small FISH probes, such as SureFISH from Agilent, Zhao et al. have demonstrated DNA FISH imaging in archival clinical samples.¹⁴ In contrast, RNA molecules tend to be washed away after expansion, and thus it is necessary to anchor either the RNA molecules or the RNA-binding probes to the polymer network. To address this issue, Chen et al. developed ExFISH for RNA imaging in ExM;¹⁵ ExFISH uses a custom reagent, LabelX, which can be easily synthesized by a one-step reaction with two commercially available reagents AcX and Label-IT amine (Figure 2C). LabelX contains a backbone for electrostatic interactions with nucleic acids and a reactive alkylating group that covalently links to any reactive heteroatom of nucleic acids (mainly guanine bases). As a result, RNA can be labeled and imaged post-expansion with improved resolution and quantification of RNA abundance. In addition, the RNA signals can be further amplified by coupling with hybridization chain reaction or can be multiplexed for many different RNAs via iterative single-molecule FISH steps.¹⁵ Tsanov et al. reported single-molecule inexpensive FISH in ExM by applying anchorable fluorescent FISH probes prior to expansion.¹⁶ This approach does not preserve RNA in the gel post-expansion, however, so it can only image targeted RNA molecules.

Despite increasingly broad use of ExM variants for imaging proteins and nucleic acids in biology and pathology,^20,27–34 imaging other biomolecule classes, such as lipids and small peptides, has remained technically challenging in ExM because, until recently, current anchoring reagents were not amenable to these molecule classes.

Lipids are a large class of biological molecules with diverse chemical structures in living organisms. The main structural component of plasma membranes and intracellular membranes in eukaryotic cells are glycerophospholipids. Other nonglyceride lipid components, such as sphingomyelin and cholesterol, are also present. The diverse chemical groups possessed by lipids impose challenges on the probe design. In order to facilitate lipid imaging in ExM, the probe design needs to consider the following features: (1) amphiphilicity, to enable partitioning into the lipid and diffusion in tissues; (2) a fluorophore that can survive the ExM process; and (3) a polymer-anchorable functional group to form a chemical bond with the polymer chains during the in situ formation of the water-swellable polymer, which is key for enabling expansion of the label. Following this principle, Karagiannis et al. demonstrated membrane ExM (mExM) (Figure 2D), where they developed a ExM-compatible lipophilic probe consisting of a backbone with ~5 D-lysines and a lipid tail on the amine terminus of the lysine chain.³⁵ The choice of D-lysine avoids degradation by proteinase-K during homogenization and enables binding of an anchoring reagent, such as AcX, to link the probe to the polymer network.

In addition to lipids, traditional small peptide probes, such as small peptide phalloidin for actin labeling,³⁶ α-bungarotoxin for acetylcholine receptor-specific labeling,³⁷ or glycans are also not suitable for direct anchoring to the polymer network. These small peptides and molecules lack primary amines or free thiol groups for decorating gel-anchorable moieties via conventional conjugation chemistries. As a result, these biomolecules are washed out during the expansion process.

Rather than targeting a specific class of biomolecule, one may ask if it is possible to develop an anchoring strategy to target multiple types of biomolecules to the polymer. Sun et al. recently reported click-ExM (Figure 3A) for imaging a range of biomolecules using 18 alkyne-containing labels constituted by DNA, RNA, proteins, lipids (Figure 3A, top right), glycans (Figure 3A, middle right), or small molecules (Figure 3A, bottom right).³⁸ These labels are anchored to the polymer network via click chemistry with azide-biotin, followed by streptavidin-biotin binding and sequential AcX/Glutaraldehyde treatment. With the diversity of clickable chemicals, Sun et al. demonstrated the universality of click-ExM on biomolecule imaging and its compatibility with signal-amplification techniques. One limitation of this approach is that it primarily relies on metabolic labeling and, therefore, click-ExM of some biomolecule classes, such as glycan and small molecules, cannot be applied to fixed tissues. In addition, the use of streptavidin-biotin limits the potential of simultaneously imaging multiple clickable labels.

Figure 3.

Strategies for universal biomolecule imaging in expansion microscopy (ExM). (A) Click-ExM.³⁸ Schematic of Click-ExM chemistry (top left), where cells metabolically labeled with alkynes are reacted with azide-biotin via click chemistry and stained using streptavidin-dye, then processed with the standard ExM workflow. Example data (right) of phospholipids (top, magenta), sialoglycans (middle, red), and small molecules (bottom, red) of cells treated with alkyne-choline, N-azidoacetylmannosamine (ManNAz), and azide-afatinaib, respectively. Scale bars: 10 μm (top, bottom), 2 μm (top, bottom inset), 50 μm (middle), 5 μm (middle, inset). Reproduced with permission from ref 38. Copyright 2020 Sun et al. (B) Multifunctional linking with trivalent anchoring (TRITON).³⁹ Schematic of the multifunctional probe design (top left) and an example TRITON probe (bottom left) consisting of a fluorescent reporter conjugated with a methacryloyl group and a reactive tetrafluorophenyl (TFP) ester for conjugation with amines such as B-phalloidin (for actin staining) or 1,2-distearoyl-sn-glycero-3-phos-phoethanolamine (DSPE) for lipid labeling. Example data (right) using TRITON probes for imaging cellular phospholipid membranes (pre-expansion top; post-expansion middle) through lipid conjugation to a fluorescent trivalent linker (Pacific Blue, DSPE) and α-tubulin stained via direct grafting of a DNA-conjugated secondary antibody and fluorescent oligo-based readout (post-expansion, bottom). Scale bars: 25 μm (top, middle), 5 μm (zoomed region top), 10 μm (zoomed region middle, bottom) and 5 μm (zoomed region bottom). Reproduced from ref 39. Copyright 2020 American Chemical Society.

In this issue of ACS Nano, Wen and co-workers report a trifunctional chemical probe, called trivalent anchoring or TRITON, that enables ExM imaging of a wide range of biomolecules in preserved samples, including lipids and small cytostatic peptides.³⁹ As illustrated in Figure 3B, the TRITON design features three functional handles: a biomolecule targeting moiety, a fluorescent reporter unit, and an acryloyl group for covalently attaching to the polymer network. This design is somewhat like the trifunctional oligo probes used in the original ExM but uses a much smaller scaffold. The TRITON probe enables direct, covalent grafting of a wide range of target molecules and fluorophores to the hydrogel network. To enable lipid imaging, the authors synthesized and use a TRITON probe carrying 1,2-distearoyl-sn-glycero-3-phosphoethanolamine to stain phospholipid bilayers and to visualize the ultra-fine structure of membranes in expanded cultured cells (Figure 3B, top right). In addition, the authors demonstrated the ability to image actin using a TRITON variant conjugated with the small cytostatic peptide phalloidin (Figure 3B, bottom right). Previously, actin imaging in ExM was only possible with genetically encoded probes, such as Actin-mRuby2.¹¹ The TRITON probe further enhances the capability of ExM to image endogenous actin. Finally, the authors also demonstrated post-expansion labeling by conjugation of a TRITON DNA tag to an antibody for post-expansion oligonucleotide hybridization. This method enables researchers to use cyanine reporter dyes as labels with ExM, which otherwise are oxidized during the polymerization process. Another advantage of post-expansion labeling is that it may facilitate highly multiplexed imaging by iterative rounds of labeling and imaging steps. Although the authors acknowledged another post-expansion labeling method by Shi and co-workers,⁴⁰ it is worth noting that there are other post-expansion labeling methods, such as MAP¹³ and immuno-SABER³⁴ for protein targets and ExFISH¹⁵ for nucleic acids.

OUTLOOK AND FUTURE CHALLENGES

Expansion microscopy is an emerging, powerful, easy-to-use, and low-cost nanoscale imaging technique that is expected to speed up the democratization of super-resolution optical imaging in biology and medicine. Recent advances on ExM-compatible biomolecule labeling methods, such as TRITON,³⁹ mExM,³⁵ and click-ExM,³⁸ have greatly expanded the types of targetable biomolecules beyond proteins and nucleic acids. Together, these new developments in ExM may enable previously inaccessible studies on a wide range of biological contexts, such as cell–cell communication, membrane signaling, metabolic pathways and cellular transport.

Yet, work on broadening the capabilities of ExM has only just begun. Many challenges must be overcome before establishing ExM as a routine nanoscale biomolecule imaging toolset that is accessible to most biological laboratories. For example, recently developed ExM variants TRITON and click-ExM have only been demonstrated in preserved cell culture. We expect to see further assessment and optimization of these new ExM variants in preserved tissues. In addition, the utility of these new ExM variants can be limited due to the need for synthesizing customized probes. The development of new derivatives with simpler chemistries and commercially available reagents would improve accessibility of these methods for regular biological laboratories that do not have expertise in organic chemistry. Moreover, the 4.5 × gel chemistry is still the primary choice for these new methods; therefore, the resolution is currently limited to ~65 nm with a conventional microscope. We may see further development of TRITON and other new ExM probes based on iterative ExM (which expanded a specimen twice⁹) or ×10 ExM (which uses a water-swellable hydrogel with 10× expansion¹⁰), which may achieve nanoscale optical imaging with lateral resolutions of 10–20 nm, bringing the performance of ExM one step closer to that of electron microscopy. In addition to resolution improvement, it may also be possible to develop ExM variants for highly multiplexed imaging of lipids and other small molecules with nanoscale resolution. For instance, a set of TRITON probes conjugated with barcode oligonucleotides can be designed to target a range of biomolecules and to label each of them with a unique barcode that can be read out after expansion, thus facilitating highly multiplexed nanoscale in situ profiling of biomolecules in various samples.

Suggested pull quotes.

Expansion microscopy can expand the molecule classes available for imaging by adapting new probes and chemistries.

In this issue of ACS Nano, Wen and co-workers report a trifunctional chemical probe, called trivalent anchoring or TRITON, that enables ExM imaging of a wide range of biomolecules in preserved samples, including lipids and small cytostatic peptides.

ABBREVIATIONS

AA

acrylamide

AcX

6-((acryloyl)amino)hexanoic acid

BIS

N-N-methylenebisacrylamide

DMAA, N

N-dimethylacrylamide

ExFISH

expansion microscopy fluorescence in situ hybridization

ExM

expansion microscopy

FISH

fluorescence in situ hybridization

immuno-SABER

immunostaining with signal amplification by exchange reaction

MA-NHS

methacrylic acid N-hydroxysuccinimidyl ester

MAP

magnified analysis of the proteome

mExM

membrane expansion microscopy

proExM

protein-retention expansion microscopy

SA

sodium acrylate

SABER-FISH

signal amplification by exchange reaction fluorescence in situ hybridization

smiFISH

single-molecule inexpensive fluorescence in situ hybridization

TRITON

trivalent anchoring