Expansion microscopy (ExM) circumvents the diffraction limit of light microscopy by physically expanding the specimen four-fold in each dimension (Chen et al., 2015; Tillberg et al., 2016). Expansion is achieved by chemically anchoring proteins and other biomolecules directly to a hyper-swelling gel, followed by aggressive proteolysis to enable uniform swelling of the gel material. While other super-resolution approaches are not readily compatible with thick tissue slices and require specialized optics (Hell and Wichmann, 1994), fluorophores (Rust et al., 2006), or software (Gustafsson, 2000), ExM is compatible with any microscope (Gao et al., 2019; Zhang et al., 2016), including other super-resolution modalities (Gao et al., 2018; Halpern et al., 2017; Xu et al., 2019), and performs well in both cultured cells and thick tissue slices (Chen et al., 2015; Tillberg et al., 2016). Assuming sufficiently high labeling density, the resolution increase of ExM depends on the expansion factor of the gel recipe used. Recently, ExM variants have been described that seek to improve resolution by increasing the expansion factor. For example, iterative ExM (iExM) uses sequential embedding in multiple expansion gels to achieve 15× and greater expansion but requires a complex sequence of gel re-embedding, link cleaving, and fluorophore transfer (Chang et al., 2017), limiting its broad adoption.

The expansion factor of the gel itself can be improved by decreasing the concentration of crosslinker (Okay, 2009), usually bisacrylamide (bis), although this is generally at the expense of the mechanical integrity of the gel. For example, reducing the bis concentration in the original ExM recipe from 1.5 to 0.25 mg/mL produces an approximately nine-fold expanding gel (Chen et al., 2015, SF5), but these gels are too soft to hold their shape under the force of gravity. As a result, they are difficult to handle without breaking and display nonuniform expansion. This tradeoff of expansion versus gel mechanical integrity has not been explored in a quantitative or systematic way.

Another gel recipe variant, using a high concentration of the monomer dimethylacrylamide (DMAA), has enough crosslinking through side reactions and polymer chain entanglement that the crosslinker can be omitted entirely, producing ~10-fold expansion in one step (Truckenbrodt et al., 2018). This recipe has been used to expand cultured cells and thin cryosectioned tissue (Truckenbrodt et al., 2019), but reportedly requires rigorous degassing to remove oxygen prior to gelation, making it laborious to use. Moreover, expansion of thick tissue slices (>50 µm) has not been demonstrated using this method. Thus, a robustly validated and easily adoptable method that is compatible with multiple sample types and enables single-step expansion well over 4× without compromising gel integrity is lacking.

Here, we explored the expansion gel recipe space in a systematic manner, assessing the limits of single-round expansion using reagents and methods that would be familiar to labs already performing ExM. For any given choice of recipe parameters (monomer concentrations, gelation temperature, initiator concentration, etc.), varying the crosslinker alone yielded a family of recipes whose expansion factor and mechanical quality vary smoothly from high-expanding, mechanically unstable to low-expanding, tough gels. A range of crosslinker concentrations was tested for each family because the optimal crosslinker concentration may vary by family. From this exploration, we generated Ten-fold Robust Expansion Microscopy (TREx), an optimized ExM method that allows for robust 10-fold expansion in a single step. We show that TREx can be used to expand both thick tissue slices and adherent cells. It is compatible with antibodies and off-the-shelf small-molecule stains for total protein and membranes. Together, we show that TREx enables 3D nanoscopic imaging of specific structures stained with antibodies in combination with cellular ultrastructure.

To systematically explore the expansion recipe space, we developed a streamlined approach for synthesizing dozens of gel recipes and characterizing their mechanical quality in parallel. For every set of gel recipe parameters (component concentrations and gelation temperature, listed in Figure 1A), we define a recipe family as the set of recipes generated by varying the crosslinker (bisacrylamide) concentration. For each family, we tested five recipes with crosslinker concentrations log-spaced from 1000 to 10 µg/mL, plus one with zero crosslinker. We also included the original ExM recipe, with 0.15% (1500 µg/mL) crosslinker. For each recipe, we cast three gel specimens, expanded them fully in water, and measured the expansion factor (Figure 1B). We found that resistance to deformation under the force of gravity was a good proxy for the more subjective judgment of ease of gel handling. We measured gel deformation by placing a semicircular punch from each gel upright on its curved edge and allowing the gel corners to deflect under the force of gravity. We defined the deformation index as the vertical displacement of each gel corner, divided by the gel radius (Figure 1C), which ranges from 0 (for gels that do not deform) to 1 (for gels that deform freely under their own weight). We manually calibrated this measurement, finding that deformation indices between 0 and 0.125 corresponded to gels with excellent ease of handling, 0.125–0.25 corresponded to acceptable ease of handling, and anything higher than 0.25 was unacceptable. While not as theoretically informative as elastic modulus and yield strength measurements, this measurement can be repeated and extended to new gel recipes by any lab developing expansion methods, without access to specialized equipment. We plotted the deformation index for each gel as a function of its expansion factor (Figure 1D) to directly assess the tradeoff between expansion and mechanical quality.

Figure 1

Download asset Open asset

Subcellular imaging of specific proteins and cellular ultrastructure in thick brain slices

In electron microscopy, nonspecific stains for proteins and membranes are commonly used to provide structural detail at high spatial resolution. Recently, the use of nonspecific NHS ester protein stains and other small-molecule probes has been combined with ExM (M’Saad and Bewersdorf, 2020; Mao et al., 2020; Sim et al., 2021; Yu et al., 2020). Expansion allows visualization of intracellular detail in such densely stained samples, which would otherwise be too crowded to lead to meaningful contrast. These applications have the promise to bring together the advantages of light microscopy (specific staining using antibodies and volumetric imaging) with the advantage of seeing cellular context typically provided by electron microscopy. Because TREx reaches single-step expansion factors at which small-molecule stains are expected to be useful, we set out to explore this idea further.

We applied the inexpensive, green-emitting dye BODIPY-FL NHS (total protein stain; see Materials and methods) after expansion with TREx to demonstrate total protein distribution in thick (100 µm) slices of mouse brain cortex (Figure 2A, Figure 2—figure supplement 1, Figure 2—video 1). The neuropil region outside the cell somas contained a rich profusion of fibers and structures visible in sharp relief. The nucleus of each cell was easily identified, with especially strong staining in nucleoli-like structures. Surrounding each nucleus, the nuclear envelope could be identified, with particularly dense total protein stain on the side facing the nucleus. The nuclear envelope was punctuated by heavily stained spots that span the envelope, consistent with nuclear pore complexes (NPCs; Figure 2A, inset). Within the cytosol, several organelles were marked by either heavy inner staining with a dim border or weak inner staining.

Figure 2 with 3 supplements see all

Download asset Open asset

We attempted to optimize protein retention, according to the total protein stain intensity, by reducing both protein anchoring and proteolysis compared with the original ExM. We tested a range of anchoring strengths by varying the concentration of the acryloyl-X SE (AcX) anchoring molecule applied prior to gelation. This was done in combination with two reduced disruption methods: proteinase K applied at one-tenth that of the original ExM method (Figure 2—figure supplement 1A, top row) and a high-temperature, protease-free denaturation treatment (Gambarotto et al., 2019; Ku et al., 2016; Tillberg et al., 2016; Zwettler et al., 2020) similar to that employed in Western blotting (Figure 2—figure supplement 1A, bottom row). The protease-free treatment enabled greater protein retention but at the cost of incomplete expansion. This could be offset through reduced AcX concentration, though this was not clearly superior to high AcX followed by proteolysis, indicating a general tradeoff between protein retention and gel expansion (see Materials and methods). We chose a hybrid approach with moderate AcX anchoring and low-concentration proteinase K digestion followed by high-temperature denaturation to proceed.

We next tested whether antibodies, applied to the tissue using a standard immunofluorescence procedure before embedding, were also retained in the TREx gel. We stained mouse brain cortex tissue for Bassoon (a marker for both excitatory and inhibitory presynaptic active zones), Homer (a marker for the excitatory postsynaptic apparatus), and VGAT (a vesicular GABA transporter in the presynaptic compartments of inhibitory synapses). After staining and anchoring with AcX, tissue was expanded with TREx and imaged by light sheet microscopy. Numerous putative excitatory synapses were observed at high density, with clearly separated Bassoon and Homer pre- and postsynaptic staining (Figure 2B and C, Figure 2—figure supplement 1B, Figure 2—video 2). Because of the excellent axial resolution, TREx allowed us to quantify the separation of Bassoon and Homer in 3D, regardless of the angle of the synapse with respect to the imaging plane (Figure 2D). We found an average separation of 1.17 µm ± 0.52 µm (mean ± SD, 583 synapses), which, when corrected for expansion, is consistent with previous reports in cultured neurons that estimated the synapse separation between 90 and 130 nm (Glebov et al., 2016; Wiesner et al., 2020). Compared with Bassoon and Homer, VGAT had a more extended staining pattern, consistent with the known distribution of synaptic vesicles throughout presynaptic boutons. Elaborately shaped compartments with dense VGAT staining were seen with multiple synaptic release sites marked by Bassoon (Figure 2C). As expected, these release sites were not paired with the excitatory postsynaptic marker Homer. In addition, spatial correlation was observed in excitatory synapses between Homer and Bassoon, reflecting the transsynaptic alignment reported previously (Hruska et al., 2018; Figure 2—figure supplement 1B). These results demonstrate the ability of TREx to preserve correct synaptic staining while enabling sub-diffraction limited imaging of large tissue sections.

Validation of expansion factor and deformation

Increasing the expansion factor from 4 to 10× could result in greater sensitivity of the expansion factor to local variation, for example, in protein dense complexes, resulting in less uniform expansion. To examine this, we explored the nanoscale isotropy of TREx by imaging NPCs, which have a highly stereotyped and well-characterized structure. The NPC constituent protein nucleoporin 96 (NUP96) appears in a ring structure with a 107 nm diameter (Thevathasan et al., 2019). NPCs have recently been explored as a reference structure for super-resolution microscopy methods, including ExM in combination with other super-resolution methods (Pesce et al., 2019; Thevathasan et al., 2019). For the conventional 4–5× expansion approach, this revealed that the diameter of the NPC was 14–29% smaller than expected from the macroscopic expansion of the gel. We used the NUP96-GFP homozygous knock-in cell line from Thevathasan et al., 2019 to study the quality of nuclear pore expansion using TREx with well-validated anti-GFP antibodies (Figure 3A). After expansion with TREx, individual NPCs were uniformly retained and clearly visible using diffraction-limited confocal microscopy (Figure 3B). An antibody against NUP153 similarly demonstrated individual NPCs but with less complete NPC coverage compared with the antibody stain against the NUP96-GFP tag (Figure 3C). The macroscopic gel expansion factor was 9.5×, suggesting an expected NPC size after expansion of 107 nm × 9.5 = 1.02 µm. We used a semi-automated approach to determining the diameter of 60 NPCs randomly chosen from three nonadjacent cells and found the size after expansion to be 939 nm ± 90 nm (mean ± SD) (Figure 3D). This is about 8% smaller than expected based on the macroscopic expansion of the gel and implies a local expansion factor of 8.8×, or 92% of the expected 9.5×. These data indicate that TREx offers more uniform local expansion compared to conventional ExM.

Figure 3

Download asset Open asset

Other reference structures for super-resolution microscopy include microtubules and clathrin-coated pits (CCPs), which have been extensively studied with various modalities including ExM (Bates et al., 2007; Chen et al., 2015). We fixed cells and stained for both tubulin and clathrin heavy chain and quantified the diameter of CCPs (Figure 3E and F). We found an average diameter of 1.16 ± 0.2 µm (mean ± SD), which, corrected for 10-fold expansion, is consistent with previously reported values (Chen et al., 2015; Jones et al., 2011). In samples with sufficiently high labeling densities (e.g., the periphery of COS7 cells), we could furthermore resolve the stereotyped M-profile across peripheral microtubules in COS7 cells, indicative of the hollow structure of microtubules. As expected for microtubules stained with primary and secondary antibodies (Mikhaylova et al., 2015), we observed a 450-nm peak-to-peak distance after expansion (Figure 3G).

Next, we quantified the measurement error introduced by nonuniform expansion by comparing antibody-stained microtubules imaged before expansion with 3D gSTED versus after expansion with confocal microscopy (Figure 3H–J), as described previously (Chen et al., 2015). Measurement lengths between pairs of points after expansion were compared to the distance expected given uniform expansion, and the average fractional deviation plotted as a function of measurement length (Figure 3I). For a large tiled acquisition of 42 fields of view from one gel (~650 × 750 µm after expansion), the measurement error was found to be a constant fraction (3.2% ± 1.7%) of the measurement length (Figure 3I). We used the similarity transform to calculate the overall expansion factor of the entire imaged area and found it to be 9.4×, consistent with the expansion expected for the whole gel. Together, these data show that TREx enables uniform single-step, 10-fold expansion that retains structural detail over large distances, in both cultured cells and thick tissue slices, with similar performance compared with the original 4× ExM.

TREx enables subcellular localization of proteins and cellular ultrastructure in cultured cells

We next explored the use of TREx for high-resolution imaging of specific proteins, NHS ester stains, and lipid membranes in cultured cells. For membranes, a custom-synthesized membrane probe compatible with the ExM process has been shown to visualize membranes in fixed brain tissue (Karagiannis et al., 2019). This probe relies on a peptide-modified lipid tail that intercalates in target membranes and provides opportunities for anchoring to the gel through D-lysines in its peptide sequence. We asked whether the commercially available membrane-binding probe mCLING could also be used for membrane staining and gel anchoring. mCLING has been developed as a fixable endocytosis marker consisting of a fluorophore and a short polypeptide group with one cysteine and seven lysines coupled to a palmitoyl membrane anchor (Revelo and Rizzoli, 2016). Due to the presence of multiple lysines, we hypothesized that mCLING would be compatible with standard ExM anchoring through AcX. While the standard protocol for mCLING delivery relies on active endocytosis in living cells, we tested whether mCLING would stain intracellular membranes more uniformly when added to fixed cells, which would have the added benefit of not perturbing intracellular membrane trafficking by long incubation in live cells. To test this, we fixed activated Jurkat T cells, incubated the fixed cells with mCLING overnight, and proceeded with the TREx protocol. We found that mCLING efficiently intercalates in both the plasma membrane and internal organelles and is retained following our standard anchoring procedure (Figure 4A, Figure 4—video 1).

Figure 4 with 3 supplements see all

Download asset Open asset

By carefully rendering the imaged volumes, we could, with one probe, both appreciate the ruffled morphology of the plasma membrane on top of the flattened part of the cell and visualize the organelle clustering typical of activated T cells (Figure 4A and B). As in electron microscopy, where distinct morphologies are used to identify organelles, we could clearly identify different organelles based on mCLING, suggesting that it could be used for automated segmentation of organelles. Indeed, we found that mitochondria could be readily segmented using a trainable Weka segmentation algorithm (Figure 4B; Arganda-Carreras et al., 2017). While the resolution of subcellular structures is limited by the density of mCLING moieties in the membrane, the efficiency of crosslinking to the gel, and the maximum expansion factor, we found that TREx allows sufficient single-step expansion to resolve individual mitochondrial cristae (Figure 4C), which are known to be as closely spaced as 70 nm (Stephan et al., 2019). Although mCLING is membrane impermeable in live cells (due to multiple positively charged amino groups), it readily stained fixed and unpermeabilized cells following extended incubation. Because this approach does not require labeling live cells and is expected to reduce differences in uptake efficiency between intracellular compartments, we used this approach in all subsequent experiments.

We next tested if mCLING could also be used to visualize membranes in more complex cell types. To test this, we used differentiated Caco-2 cells grown to form an epithelial monolayer. Using TREx, we could expand the entire monolayer and visualize membranes using mCLING (Figure 4D–H, Figure 4—video 2). The advantage of optical, volumetric imaging is underscored by the fact that we can easily render one dataset in several ways, either resembling scanning electron microscopy to highlight volumetric surface morphology (Figure 4D) or transmission electron microscopy to explore single planes in more detail (Figure 4E and F). For example, we were able to resolve the elaborate interdigitated cell-cell junctions that could previously only be clearly appreciated using electron microscopy (Drenckhahn and Franz, 1986), as well as resolve individual microvilli as hollow membrane protrusions within the dense brush border. To underscore the significant resolution increase of TREx compared to standard ExM, we incubated expanded TREx gels with solutions of increasing ionic strength to shrink the gel back to ~4.5× the size of the pre-expanded gel (Figure 4G, and Figure 4—figure supplement 1). When the 10× and 4.5× expanded gels were imaged, dense brush borders of differentiated cells could only be resolved in the 10× gel (Figure 4G). To validate the expansion factor, we quantified the diameter of individual microvilli as these have been thoroughly characterized with EM with a diameter of ~100 nm (Crawley et al., 2014). Indeed, we found an average diameter of 1.08 ± 0.16 µm (n = 12,339 from 12 cells, N = 3), which, corrected for an expansion factor of 10, is within 8% of the value from electron microscopy. Together, these data illustrate the robustness of TREx in expanding multiple cell types and show how the increased expansion factor combined with a commercially available membrane stain provides rapid volumetric insights into the elaborate membranous architecture of cells.

Previously, we used ExM to study the three-dimensional organization of microtubules in neurons and T cells (Hooikaas et al., 2020; Jurriens et al., 2021; Katrukha et al., 2021). In these experiments, cells were typically pre-extracted with detergent and glutaraldehyde to reduce background, followed by paraformaldehyde fixation (Tas et al., 2017). We reasoned that the increased expansion of TREx would dilute the soluble tubulin background, resulting in a relative boost in signal over background and eliminating the need for pre-extraction to enable high-resolution imaging of microtubules in combination with membranes. To test this, we fixed cells without pre-extraction, treated them with mCLING, stained for tubulin, and imaged the stained cells both before and after expansion with TREx (Figure 5A, second panel). Expanded cells retained high-quality anti-tubulin antibody signal exhibiting high contrast relative to the cytosolic background. We also observed increased detail in both mCLING and tubulin stains after expansion compared to before expansion, which was particularly apparent with side views of the same cell (Figure 5A, far right).

Figure 5 with 1 supplement see all

Download asset Open asset

We next fixed cells expressing GFP-Sec61β without pre-extraction, treated them with mCLING, stained for GFP and tubulin, and then proceeded with TREx (Figure 5B–D, Figure 5—video 1). Because small-molecule stains like mCLING do not rely on subsequent antibody amplification we reasoned that this decrease in linkage error should be reflected in the apparent size of subcellular structures. To test this, we quantified the diameter of endoplasmic reticulum (ER) tubules that were stained by both mCLING and GFP-Sec61β (Figure 5C). Indeed, we observed an average tubule diameter of 0.79 ± 0.1 µm for mCLING, which, corrected for 10-fold expansion, is consistent with the previously published range of 60–100 nm (Shibata et al., 2009; Shibata et al., 2006), while the anti-GFP signal from the same tubules is broadened to an average diameter of 0.93 ± 0.1 µm. Within the same dataset, the interplay between microtubules and ER (Figure 5D) in three dimensions could be resolved, which revealed how other membranous organelles were connected to both structures (Figure 5D). Thus, TREx facilitates high-resolution three-dimensional mapping of specific cytoskeletal and membranous structures in combination with markers that provide ultrastructural context.

Finally, we tested whether TREx using general membrane stains could be combined with general protein stains and/or antibody stains. U2OS cells transfected with GFP-Sec61β were fixed, treated with mCLING, stained for GFP, and expanded with TREx followed by the NHS stain (Figure 5E). Because we performed the NHS stain after disruption, we used high-temperature denaturing disruption rather than proteolytic digestion. We found that secondary antibodies that had been used to stain GFP before gelation withstood this disruption step. We observed a clear degree of overlap between mCLING and NHS, especially in the dense perinuclear region, but we could also identify distinct features of each stain (Figure 5E, bottom). These results demonstrate that general stains for membranes and proteins can be combined with antibody-based labeling to reveal specific proteins in their ultrastructural context.

We developed TREx in order to expand biological specimens 10-fold in a single round of expansion, without specialized equipment or procedures. In developing this method, we established a framework for assessing gel recipes operating near this apparent limit of single-round expansion. We found that the mechanical performance of gel recipes, that is, resistance to deformation versus gel expansion factor, varies smoothly with changes in crosslinker. For all high-monomer (~3 molar total acrylamide and acrylate) gel recipe families, the relation between expansion factor and crosslinker concentration fell close to a common curve. The high radical initiation rate of family B enabled gelation without the inclusion of a crosslinker, suggesting that side reactions and polymer entanglement in these conditions create sufficient network crosslinks to form a gel. Gel deformation measurements plotted versus expansion factor, though less precise, also show high-monomer recipe families falling close to a common curve. Compared to the high-monomer families, the original ExM recipe family is less resistant to deformation for a given expansion factor and expands more for a given crosslinker concentration. The factor determining gel properties is not crosslinker concentration in the gel recipe per se, but rather the density of effective crosslinks formed between neighboring polymer chains in the gel. This suggests that the original, low-monomer recipe less efficiently incorporates crosslinker molecules as network crosslinks. This may be because the resulting lower rate of chain extension allows incorporated crosslinker molecules to be re-reacted by the same polymer chain before they can react with neighboring polymers.

While the expansion factor of the original ExM recipe can be tuned by varying the crosslinker concentration, it has been shown that increasing the monomer content is required to maintain nanoscale isotropy using centrioles as a convenient standard reference structure (Gambarotto et al., 2019). Considering gel quality versus expansion factor alone, the high-monomer recipe family B derived from the U-ExM recipe allows for a 10-fold expanding gel (at crosslinker concentration of 30 µg/mL) with a low deformation index of 0.13. However, the high radical initiation rate used in this family (5 mg/mL APS and TEMED) results in fast gelation. This increases the time sensitivity of mounting the specimen in the gelation chamber and adds an additional challenge for adapting the method to thick tissues, which require extended incubation in the gelation solution. Recipe families C and D solve this problem by reducing initiation rates at a slight expense of mechanical performance compared with family B. Like family B, family C has a high acrylate content, which might contribute to imaging artifacts due to shrinkage of the sample prior to gelation and inconsistent acrylate purity. Family D reduces the acrylate content by half while retaining similar mechanical performance to family C, especially in the 10-fold expansion regime. Finally, with family E we explored whether increasing the gelation temperature to 50°C would produce the improved mechanical performance of family B (through increased temperature-dependent radical initiation) without premature gelation at RT. However, we found that this reduced the expansion factor and increased susceptibility to experimental variation. Therefore, we based our TREx recipe on recipe family D. The exact crosslinker concentration that produces 10-fold expanding gels can vary due to differences in gelation chamber design, so we recommend that each lab test a range of crosslinker concentrations between 30-60 ppm using their choice of specimen, gelation chamber, and incubator.

We applied TREx to mouse brain tissue slices stained either for specific targets with antibodies, or for total protein distribution with NHS ester dyes. Single-round 10-fold expansion with TREx followed by total protein staining was sufficient to reveal densely packed axons and dendrites running through the neuropil, while individual organelles could be resolved within the neuronal soma. The nuclear envelope, along with presumptive NPCs, was also clearly resolved. The correct relative localizations of pre- and postsynaptic markers and presynaptic neurotransmitter vesicles stained with standard immunofluorescence were also retained in TREx-expanded tissue slices. For this characterization, we used a hybrid disruption approach incorporating reduced proteinase K digestion followed by high-temperature denaturation. We adopted this hybrid approach because we had noticed that sometimes the reduced protease treatment alone produces underexpanded nuclei while the protease-free treatment alone produced underexpanded synapses. In general, a higher degree of anchoring requires higher disruption strength to achieve full expansion on the macroscopic level. Further application-specific optimization may be beneficial, given the heterogeneity of biological tissue. For applications where maximizing total protein retention is not a priority, we recommend simply using a high concentration of proteinase K (e.g., 1:100 dilution, overnight).

We validated the nanoscale isotropy of TREx by comparing the size of well-characterized subcellular structures to values from literature. These include NPCs, ER tubules, CCPs, and microvilli. In the case of NPCs, the conventional 4× expansion approach yielded an NPC diameter 14–29% smaller than expected (Pesce et al., 2019; Thevathasan et al., 2019), suggesting that protein-dense complexes may resist full expansion. Using TREx, the average NPC diameter was 8% smaller than the expected value. Further optimization of anchoring and disruption conditions may improve expansion uniformity for protein-dense structures such as NPCs. The widths of CCPs, ER tubules, and microvilli measured with TREx are reasonable, given the variation in previously measured sizes of these structures or different staining and imaging modalities. Notably, we observed the broadening of apparent ER tubule width due to antibody linkage error compared to direct membrane labeling with mCLING.

In addition to nanoscale isotropy, we characterized the overall expansion isotropy by comparing microtubules before and after expansion, finding expansion-induced measurement errors on average 3.2% of a given measurement length. Although these errors do not saturate at sub-100 µm distances as they do for lower-expanding gels, the nanoscale fidelity over short length scales is in line with previous expansion methods (Chen et al., 2015; Tillberg et al., 2016) and accumulation of nanoscale deformation over longer distances is not a limiting factor for most biological applications. For experiments that would require lower long-distance deformations, we expect these deformations could be reduced through optimized gel handling and mounting.

We further demonstrated the utility of TREx for the study of cell biology through combinations with several staining modalities in cultured cells, prepared in several culture formats. After a single round of expansion with TREx, the commercially available membrane stain mCLING was able to clearly resolve the internal structure of mitochondria and the detailed pattern of plasma membrane ruffling in activated Jurkat T cells. While these structures would be readily resolved with electron microscopy of mechanically sectioned cells, we were able to do so in the context of complete cells, enabling detection and automated segmentation of mitochondria clustered followed T cell activation. Caco-2 cells grown on permeable filters were also successfully stained with mCLING and expanded with TREx to reveal the detailed structure of epithelial microvilli and membrane interdigitations at the contacts of neighboring cells. These structures had previously been known from electron microscopy but were now imaged with ease in the context of entire cell monolayers. Successful application of TREx to filter-cultured Caco-2 cells further demonstrates the robustness of TREx because we had repeatedly failed to cleanly recover epithelial cultures using standard ExM. We speculate that this may be due to the increased robustness of TREx to interactions between the silicone filter material and the gelation process. This robustness is further demonstrated by its adoption in other biological systems (Gros et al., 2021) including in cultured neurons (Ozkan et al., 2021) and primary cultured human cells (Nijenhuis et al., 2021), and by its superior mechanical properties as measured by traditional materials characterization methods (Chen et al., 2021).

In U2OS cells, TREx retained anti-tubulin antibody stain with high efficiency, maintaining continuous microtubules with high signal-to-noise ratio after expansion. Moreover, we demonstrate that this specific labeling can be combined with total membrane labeling (mCLING) to reveal close appositions between microtubules and various organelles including ER and presumptive mitochondria. We also combined mCLING with total protein labeling (NHS ester dye), finding that the staining patterns were similar in their overall contours, albeit with some clear differences, such as the presence of presumptive NPCs in the NHS ester channel. It has been shown (Sim et al., 2021) that highly hydrophobic NHS ester dyes exhibit strong contrast for cytosolic organelles while highly hydrophilic NHS ester dyes strongly stain the nucleus. The moderate hydrophobicity dyes that we used (BODIPY-FL [Zanetti-Domingues et al., 2013] and Alexa Fluor 594 [Hughes et al., 2014]) exhibit both nuclear staining and contrast for cytosolic organelles. The strong overlap between NHS ester and mCLING stains is likely due to a combination of the reactivity of NHS esters towards unreacted lysines in the mCLING molecule and hydrophobic interactions that boost the reaction rate between the dye and proteins in membrane-rich organelles.

In summary, by systematically exploring the ExM recipe space, we established a novel recipe using standard ExM reagents that has been rapidly adopted by other labs. TREx allows for 10-fold expansion of both thick tissue slices and cells in a single expansion step and has applications in tissue and high-resolution subcellular imaging. Importantly, TREx of antibody-stained samples can be combined with off-the-shelf small-molecule stains for both membranes and total protein to localize specific proteins in their ultrastructural context.

All data generated or analyzed during this study are included in the manuscript and supporting file; Source data files are available at https://doi.org/10.6084/m9.figshare.19142264.

1. 1. Damstra HGJ
   2. Mohar B
   3. Eddison M
   4. Akhmanova A
   5. Kaptein LC
   6. Tillberg P

   (2022) figshare

   Source data for paper: Visualizing cellular and tissue ultrastructure using Ten-fold Robust Expansion Microscopy (TREx).

   https://doi.org/10.6084/m9.figshare.19142264

1. 1. Arganda-Carreras I
   2. Kaynig V
   3. Rueden C
   4. Eliceiri KW
   5. Schindelin J
   6. Cardona A
   7. Sebastian Seung H

   (2017) Trainable weka segmentation: a machine learning tool for microscopy pixel classification

   Bioinformatics 33:2424–2426.

   https://doi.org/10.1093/bioinformatics/btx180

   • PubMed
   • Google Scholar
2. 1. Bates M
   2. Huang B
   3. Dempsey GT
   4. Zhuang X

   (2007) Multicolor super-resolution imaging with photo-switchable fluorescent probes

   Science 317:1749–1753.

   https://doi.org/10.1126/science.1146598

   • PubMed
   • Google Scholar
3. 1. Berg S
   2. Kutra D
   3. Kroeger T
   4. Straehle CN
   5. Kausler BX
   6. Haubold C
   7. Schiegg M
   8. Ales J
   9. Beier T
   10. Rudy M
   11. Eren K
   12. Cervantes JI
   13. Xu B
   14. Beuttenmueller F
   15. Wolny A
   16. Zhang C
   17. Koethe U
   18. Hamprecht FA
   19. Kreshuk A

   (2019) ilastik: interactive machine learning for (bio)image analysis

   Nature Methods 16:1226–1232.

   https://doi.org/10.1038/s41592-019-0582-9

   • PubMed
   • Google Scholar
4. Conference

   1. Bogovic JA
   2. Hanslovsky P
   3. Wong A
   4. Saalfeld S

   (2016) 2016 IEEE 13th International Symposium on Biomedical Imaging (ISBI 2016)

   Robust registration of calcium images by learned contrast synthesis.

   https://doi.org/10.1109/ISBI.2016.7493463

   • Google Scholar
5. 1. Chang JB
   2. Chen F
   3. Yoon YG
   4. Jung EE
   5. Babcock H
   6. Kang JS
   7. Asano S
   8. Suk HJ
   9. Pak N
   10. Tillberg PW
   11. Wassie AT
   12. Cai D
   13. Boyden ES

   (2017) Iterative expansion microscopy

   Nature Methods 14:593–599.

   https://doi.org/10.1038/nmeth.4261

   • PubMed
   • Google Scholar
6. 1. Chen F
   2. Tillberg PW
   3. Boyden ES

   (2015) Expansion microscopy

   Science 347:543–548.

   https://doi.org/10.1126/science.1260088

   • Google Scholar
7. 1. Chen R
   2. Cheng X
   3. Zhang Y
   4. Yang X
   5. Wang Y
   6. Liu X
   7. Zeng S

   (2021) Expansion tomography for large volume tissue imaging with nanoscale resolution

   Biomedical Optics Express 12:5614–5628.

   https://doi.org/10.1364/BOE.431696

   • PubMed
   • Google Scholar
8. 1. Crawley SW
   2. Mooseker MS
   3. Tyska MJ

   (2014) Shaping the intestinal brush border

   The Journal of Cell Biology 207:441–451.

   https://doi.org/10.1083/jcb.201407015

   • PubMed
   • Google Scholar
9. 1. Drenckhahn D
   2. Franz H

   (1986) Identification of actin-, alpha-actinin-, and vinculin-containing plaques at the lateral membrane of epithelial cells

   The Journal of Cell Biology 102:1843–1852.

   https://doi.org/10.1083/jcb.102.5.1843

   • PubMed
   • Google Scholar
10. 1. Gambarotto D
    2. Zwettler FU
    3. Le Guennec M
    4. Schmidt-Cernohorska M
    5. Fortun D
    6. Borgers S
    7. Heine J
    8. Schloetel JG
    9. Reuss M
    10. Unser M
    11. Boyden ES
    12. Sauer M
    13. Hamel V
    14. Guichard P

    (2019) Imaging cellular ultrastructures using expansion microscopy (U-ExM)

    Nature Methods 16:71–74.

    https://doi.org/10.1038/s41592-018-0238-1

    • PubMed
    • Google Scholar
11. 1. Gao M
    2. Maraspini R
    3. Beutel O
    4. Zehtabian A
    5. Eickholt B
    6. Honigmann A
    7. Ewers H

    (2018) Expansion Stimulated Emission Depletion Microscopy (ExSTED)

    ACS Nano 12:4178–4185.

    https://doi.org/10.1021/acsnano.8b00776

    • PubMed
    • Google Scholar
12. 1. Gao R
    2. Asano SM
    3. Upadhyayula S
    4. Pisarev I
    5. Milkie DE
    6. Liu TL
    7. Singh V
    8. Graves A
    9. Huynh GH
    10. Zhao Y
    11. Bogovic J
    12. Colonell J
    13. Ott CM
    14. Zugates C
    15. Tappan S
    16. Rodriguez A
    17. Mosaliganti KR
    18. Sheu SH
    19. Pasolli HA
    20. Pang S
    21. Xu CS
    22. Megason SG
    23. Hess H
    24. Lippincott-Schwartz J
    25. Hantman A
    26. Rubin GM
    27. Kirchhausen T
    28. Saalfeld S
    29. Aso Y
    30. Boyden ES
    31. Betzig E

    (2019) Cortical column and whole-brain imaging with molecular contrast and nanoscale resolution

    Science 363:eaau8302.

    https://doi.org/10.1126/science.aau8302

    • PubMed
    • Google Scholar
13. 1. Glebov OO
    2. Cox S
    3. Humphreys L
    4. Burrone J

    (2016) Neuronal activity controls transsynaptic geometry

    Scientific Reports 6:1–11.

    https://doi.org/10.1038/srep22703

    • PubMed
    • Google Scholar
14. 1. Gros OJ
    2. Damstra HGJ
    3. Kapitein LC
    4. Akhmanova A
    5. Berger F

    (2021) Dynein self-organizes while translocating the centrosome in T-cells

    Molecular Biology of the Cell 32:855–868.

    https://doi.org/10.1091/mbc.E20-10-0668

    • PubMed
    • Google Scholar
15. 1. Gustafsson MGL

    (2000) Surpassing the lateral resolution limit by a factor of two using structured illumination microscopy

    Journal of Microscopy 198:82–87.

    https://doi.org/10.1046/j.1365-2818.2000.00710.x

    • PubMed
    • Google Scholar
16. 1. Halpern AR
    2. Alas GCM
    3. Chozinski TJ
    4. Paredez AR
    5. Vaughan JC

    (2017) Hybrid Structured Illumination Expansion Microscopy Reveals Microbial Cytoskeleton Organization

    ACS Nano 11:12677–12686.

    https://doi.org/10.1021/acsnano.7b07200

    • PubMed
    • Google Scholar
17. 1. Hell SW
    2. Wichmann J

    (1994) Breaking the diffraction resolution limit by stimulated emission: stimulated-emission-depletion fluorescence microscopy

    Optics Letters 19:780–782.

    https://doi.org/10.1364/ol.19.000780

    • PubMed
    • Google Scholar
18. 1. Hooikaas PJ
    2. Damstra HG
    3. Gros OJ
    4. van Riel WE
    5. Martin M
    6. Smits YT
    7. van Loosdregt J
    8. Kapitein LC
    9. Berger F
    10. Akhmanova A

    (2020) Kinesin-4 KIF21B limits microtubule growth to allow rapid centrosome polarization in T cells

    eLife 9:e62876.

    https://doi.org/10.7554/eLife.62876

    • PubMed
    • Google Scholar
19. 1. Hruska M
    2. Henderson N
    3. Le Marchand SJ
    4. Jafri H
    5. Dalva MB

    (2018) Synaptic nanomodules underlie the organization and plasticity of spine synapses

    Nature Neuroscience 21:671–682.

    https://doi.org/10.1038/s41593-018-0138-9

    • PubMed
    • Google Scholar
20. 1. Hughes LD
    2. Rawle RJ
    3. Boxer SG

    (2014) Choose your label wisely: water-soluble fluorophores often interact with lipid bilayers

    PLOS ONE 9:e87649.

    https://doi.org/10.1371/journal.pone.0087649

    • PubMed
    • Google Scholar
21. 1. Jones SA
    2. Shim SH
    3. He J
    4. Zhuang X

    (2011) Fast, three-dimensional super-resolution imaging of live cells

    Nature Methods 8:499–508.

    https://doi.org/10.1038/nmeth.1605

    • PubMed
    • Google Scholar
22. 1. Julio G
    2. Merindano MD
    3. Canals M
    4. Ralló M

    (2008) Image processing techniques to quantify microprojections on outer corneal epithelial cells

    Journal of Anatomy 212:879–886.

    https://doi.org/10.1111/j.1469-7580.2008.00898.x

    • PubMed
    • Google Scholar
23. 1. Jurriens D
    2. van Batenburg V
    3. Katrukha EA
    4. Kapitein LC

    (2021) Mapping the neuronal cytoskeleton using expansion microscopy

    Methods in Cell Biology 161:105–124.

    https://doi.org/10.1016/bs.mcb.2020.04.018

    • PubMed
    • Google Scholar
24. Preprint

    1. Karagiannis ED
    2. Seuk Kang J
    3. Won Shin T
    4. Emenari A
    5. Asano S
    6. Lin L
    7. Costa EK
    8. Grand Challenge Consortium I
    9. Marblestone AH
    10. Kasthuri N
    11. Boyden ES

    (2019) Expansion Microscopy of Lipid Membranes

    bioRxiv.

    https://doi.org/10.1101/829903

    • Google Scholar
25. 1. Katrukha EA
    2. Jurriens D
    3. Salas Pastene DM
    4. Kapitein LC

    (2021) Quantitative mapping of dense microtubule arrays in mammalian neurons

    eLife 10:e67925.

    https://doi.org/10.7554/eLife.67925

    • Google Scholar
26. 1. Ku T
    2. Swaney J
    3. Park JY
    4. Albanese A
    5. Murray E
    6. Cho JH
    7. Park YG
    8. Mangena V
    9. Chen J
    10. Chung K

    (2016) Multiplexed and scalable super-resolution imaging of three-dimensional protein localization in size-adjustable tissues

    Nature Biotechnology 34:973–981.

    https://doi.org/10.1038/nbt.3641

    • PubMed
    • Google Scholar
27. 1. Mao C
    2. Lee MY
    3. Jhan JR
    4. Halpern AR
    5. Woodworth MA
    6. Glaser AK
    7. Chozinski TJ
    8. Shin L
    9. Pippin JW
    10. Shankland SJ
    11. Liu JTC
    12. Vaughan JC

    (2020) Feature-rich covalent stains for super-resolution and cleared tissue fluorescence microscopy

    Science Advances 6:eaba4542.

    https://doi.org/10.1126/sciadv.aba4542

    • PubMed
    • Google Scholar
28. 1. Mikhaylova M
    2. Cloin BMC
    3. Finan K
    4. van den Berg R
    5. Teeuw J
    6. Kijanka MM
    7. Sokolowski M
    8. Katrukha EA
    9. Maidorn M
    10. Opazo F
    11. Moutel S
    12. Vantard M
    13. Perez F
    14. van Bergen en Henegouwen PMP
    15. Hoogenraad CC
    16. Ewers H
    17. Kapitein LC

    (2015) Resolving bundled microtubules using anti-tubulin nanobodies

    Nature Communications 6:1–7.

    https://doi.org/10.1038/ncomms8933

    • PubMed
    • Google Scholar
29. 1. M’Saad O
    2. Bewersdorf J

    (2020) Light microscopy of proteins in their ultrastructural context

    Nature Communications 11:3850.

    https://doi.org/10.1038/s41467-020-17523-8

    • PubMed
    • Google Scholar
30. Preprint

    1. Nijenhuis W
    2. Damstra HGJ
    3. van GE
    4. Iwanski MK
    5. Praest P
    6. Soltani ZE
    7. van GM
    8. Brunsveld JE
    9. de KT
    10. Rodenburg LW
    11. de JD
    12. Raeven HHM
    13. Spelier S
    14. Amatngalim GD
    15. Akhmanova A
    16. Nijhuis M
    17. Lebbink RJ
    18. Beekman JM
    19. Kapitein LC

    (2021) Optical Nanoscopy Reveals SARS-CoV-2-Induced Remodeling of Human Airway Cells

    bioRxiv.

    https://doi.org/10.1101/2021.08.05.455126

    • Google Scholar
31. Book

    1. Okay O

    (2009) Hydrogel Sensors and Actuators

    In: Gerlach G, editors. General Properties of Hydrogels In. Berlin, Heidelberg: Springer. pp. 1–14.

    https://doi.org/10.1007/978-3-540-75645-3

    • Google Scholar
32. 1. Ozkan N
    2. Koppers M
    3. van Soest I
    4. van Harten A
    5. Jurriens D
    6. Liv N
    7. Klumperman J
    8. Kapitein LC
    9. Hoogenraad CC
    10. Farías GG

    (2021) ER - lysosome contacts at a pre-axonal region regulate axonal lysosome availability

    Nature Communications 12:1–18.

    https://doi.org/10.1038/s41467-021-24713-5

    • PubMed
    • Google Scholar
33. 1. Pesce L
    2. Cozzolino M
    3. Lanzanò L
    4. Diaspro A
    5. Bianchini P

    (2019) Measuring expansion from macro- to nanoscale using NPC as intrinsic reporter

    Journal of Biophotonics 12:e201900018.

    https://doi.org/10.1002/jbio.201900018

    • PubMed
    • Google Scholar
34. 1. Revelo NH
    2. Rizzoli SO

    (2016) The Membrane Marker mCLING Reveals the Molecular Composition of Trafficking Organelles

    Current Protocols in Neuroscience 74:ns0225s74.

    https://doi.org/10.1002/0471142301.ns0225s74

    • PubMed
    • Google Scholar
35. 1. Rust MJ
    2. Bates M
    3. Zhuang XW

    (2006) Sub-diffraction-limit imaging by stochastic optical reconstruction microscopy (STORM)

    Nature Methods 3:793–795.

    https://doi.org/10.1038/nmeth929

    • PubMed
    • Google Scholar
36. 1. Shibata Y
    2. Voeltz GK
    3. Rapoport TA

    (2006) Rough Sheets and Smooth Tubules

    Cell 126:435–439.

    https://doi.org/10.1016/J.CELL.2006.07.019

    • Google Scholar
37. 1. Shibata Y
    2. Hu J
    3. Kozlov MM
    4. Rapoport TA

    (2009) Mechanisms Shaping the Membranes of Cellular Organelles

    Annual Review of Cell and Developmental Biology 25:329–354.

    https://doi.org/10.1146/ANNUREV.CELLBIO.042308.113324

    • Google Scholar
38. Preprint

    1. Sim J
    2. Park CE
    3. Cho I
    4. Min K
    5. Lee JS
    6. Chong Y
    7. Kim J
    8. Kang JS
    9. Piatkevich KD
    10. Jung EE
    11. Kwon SK
    12. Yoon YG
    13. Boyden ES
    14. Chang JB

    (2021) Whole-ExM: Expansion Microscopy Imaging of All Anatomical Structures of Whole Larval Zebrafish

    bioRxiv.

    https://doi.org/10.1101/2021.05.18.443629

    • Google Scholar
39. 1. Stephan T
    2. Roesch A
    3. Riedel D
    4. Jakobs S

    (2019) Live-cell STED nanoscopy of mitochondrial cristae

    Sci Reports 91:1–6.

    https://doi.org/10.1038/s41598-019-48838-2

    • Google Scholar
40. 1. Tas RP
    2. Chazeau A
    3. Cloin BMC
    4. Lambers MLA
    5. Hoogenraad CC
    6. Kapitein LC

    (2017) Differentiation between Oppositely Oriented Microtubules Controls Polarized Neuronal Transport

    Neuron 96:1264–1271.

    https://doi.org/10.1016/j.neuron.2017.11.018

    • Google Scholar
41. 1. Thevathasan JV
    2. Kahnwald M
    3. Cieśliński K
    4. Hoess P
    5. Peneti SK
    6. Reitberger M
    7. Heid D
    8. Kasuba KC
    9. Hoerner SJ
    10. Li Y
    11. Wu Y-L
    12. Mund M
    13. Matti U
    14. Pereira PM
    15. Henriques R
    16. Nijmeijer B
    17. Kueblbeck M
    18. Sabinina VJ
    19. Ellenberg J
    20. Ries J

    (2019) Nuclear pores as versatile reference standards for quantitative superresolution microscopy

    Nature Methods 16:1045–1053.

    https://doi.org/10.1038/s41592-019-0574-9

    • PubMed
    • Google Scholar
42. 1. Tillberg PW
    2. Chen F
    3. Piatkevich KD
    4. Zhao Y
    5. Yu C-CJ
    6. English BP
    7. Gao L
    8. Martorell A
    9. Suk H-J
    10. Yoshida F
    11. DeGennaro EM
    12. Roossien DH
    13. Gong G
    14. Seneviratne U
    15. Tannenbaum SR
    16. Desimone R
    17. Cai D
    18. Boyden ES

    (2016) Protein-retention expansion microscopy of cells and tissues labeled using standard fluorescent proteins and antibodies

    Nature Biotechnology 34:987–992.

    https://doi.org/10.1038/nbt.3625

    • PubMed
    • Google Scholar
43. 1. Truckenbrodt S
    2. Maidorn M
    3. Crzan D
    4. Wildhagen H
    5. Kabatas S
    6. Rizzoli SO

    (2018) X10 expansion microscopy enables 25-nm resolution on conventional microscopes

    EMBO Reports 19:e45836.

    https://doi.org/10.15252/embr.201845836

    • PubMed
    • Google Scholar
44. 1. Truckenbrodt S
    2. Sommer C
    3. Rizzoli SO
    4. Danzl JG

    (2019) A practical guide to optimization in X10 expansion microscopy

    Nature Protocols 14:832–863.

    https://doi.org/10.1038/s41596-018-0117-3

    • PubMed
    • Google Scholar
45. 1. Wiesner T
    2. Bilodeau A
    3. Bernatchez R
    4. Deschênes A
    5. Raulier B
    6. De Koninck P
    7. Lavoie-Cardinal F

    (2020) Activity-Dependent Remodeling of Synaptic Protein Organization Revealed by High Throughput Analysis of STED Nanoscopy Images

    Frontiers in Neural Circuits 14:57.

    https://doi.org/10.3389/fncir.2020.00057

    • PubMed
    • Google Scholar
46. 1. Xu H
    2. Tong Z
    3. Ye Q
    4. Sun T
    5. Hong Z
    6. Zhang L
    7. Bortnick A
    8. Cho S
    9. Beuzer P
    10. Axelrod J
    11. Hu Q
    12. Wang M
    13. Evans SM
    14. Murre C
    15. Lu LF
    16. Sun S
    17. Corbett KD
    18. Cang H

    (2019) Molecular organization of mammalian meiotic chromosome axis revealed by expansion STORM microscopy

    PNAS 116:18423–18428.

    https://doi.org/10.1073/pnas.1902440116

    • PubMed
    • Google Scholar
47. 1. Yu CCJ
    2. Barry NC
    3. Wassie AT
    4. Sinha A
    5. Bhattacharya A
    6. Asano S
    7. Zhang C
    8. Chen F
    9. Hobert O
    10. Goodman MB
    11. Haspel G
    12. Boyden ES

    (2020) Expansion microscopy of C. elegans

    eLife 9:e46249.

    https://doi.org/10.7554/eLife.46249

    • Google Scholar
48. 1. Zanetti-Domingues LC
    2. Tynan CJ
    3. Rolfe DJ
    4. Clarke DT
    5. Martin-Fernandez M

    (2013) Hydrophobic fluorescent probes introduce artifacts into single molecule tracking experiments due to non-specific binding

    PLOS ONE 8:e74200.

    https://doi.org/10.1371/journal.pone.0074200

    • PubMed
    • Google Scholar
49. 1. Zhang YS
    2. Chang JB
    3. Alvarez MM
    4. Trujillo-de Santiago G
    5. Aleman J
    6. Batzaya B
    7. Krishnadoss V
    8. Ramanujam AA
    9. Kazemzadeh-Narbat M
    10. Chen F
    11. Tillberg PW
    12. Dokmeci MR
    13. Boyden ES
    14. Khademhosseini A

    (2016) Hybrid Microscopy: Enabling Inexpensive High-Performance Imaging through Combined Physical and Optical Magnifications

    Scientific Reports 6:22691.

    https://doi.org/10.1038/srep22691

    • PubMed
    • Google Scholar
50. 1. Zwettler FU
    2. Reinhard S
    3. Gambarotto D
    4. Bell TDM
    5. Hamel V
    6. Guichard P
    7. Sauer M

    (2020) Molecular resolution imaging by post-labeling expansion single-molecule localization microscopy (Ex-SMLM)

    Nature Communications 11:111.

    https://doi.org/10.1038/s41467-020-17086-8

    • PubMed
    • Google Scholar

Author details

1. Hugo GJ Damstra

   Cell Biology, Neurobiology and Biophysics, Department of Biology, Faculty of Science, Utrecht University, Utrecht, Netherlands

   Contribution

   Formal analysis, Investigation, Methodology, Validation, Visualization, Writing – original draft, Writing – review and editing

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0003-0847-609X

2. Boaz Mohar

   Janelia Research Campus, HHMI, Ashburn, United States

   Contribution

   Formal analysis, Investigation, Methodology, Visualization, Writing – original draft, Writing – review and editing

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0002-8613-2869

3. Mark Eddison

   Janelia Research Campus, HHMI, Ashburn, United States

   Contribution

   Formal analysis, Investigation, Methodology, Writing – review and editing

   Competing interests

   No competing interests declared

4. Anna Akhmanova

   Cell Biology, Neurobiology and Biophysics, Department of Biology, Faculty of Science, Utrecht University, Utrecht, Netherlands

   Contribution

   Funding acquisition, Supervision, Writing – review and editing

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0002-9048-8614

5. Lukas C Kapitein

   Cell Biology, Neurobiology and Biophysics, Department of Biology, Faculty of Science, Utrecht University, Utrecht, Netherlands

   Contribution

   Conceptualization, Formal analysis, Funding acquisition, Investigation, Project administration, Supervision, Visualization, Writing – original draft, Writing – review and editing

   For correspondence

   l.kapitein@uu.nl

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0001-9418-6739

6. Paul W Tillberg

   Janelia Research Campus, HHMI, Ashburn, United States

   Contribution

   Conceptualization, Formal analysis, Investigation, Methodology, Resources, Software, Supervision, Validation, Writing – original draft, Writing – review and editing

   For correspondence

   tillbergp@janelia.hhmi.org

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0002-2568-2365

• 16,129

  views

• 1,733

  downloads

• 101

  citations

Views, downloads and citations are aggregated across all versions of this paper published by eLife.