Ultrastructure expansion microscopy in Trypanosoma brucei

The recently developed ultrastructure expansion microscopy (U-ExM) technique allows us to increase the spatial resolution within a cell or tissue for microscopic imaging through the physical expansion of the sample. In this study, we validate the use of U-ExM in Trypanosoma brucei measuring the expansion factors of several different compartments/organelles and thus verify the isotropic expansion of the cell. We furthermore demonstrate the use of this sample preparation protocol for future studies by visualizing the nucleus and kDNA, as well as proteins of the cytoskeleton, the basal body, the mitochondrion and the endoplasmic reticulum. Lastly, we discuss the challenges and opportunities of U-ExM.


Introduction
Trypanosoma brucei is a unicellular flagellated protozoan parasite and the causative agent of human African sleeping sickness and Nagana in cattle. The highly polarized parasite cell body is about 25 µm in length and is shaped by the subpellicular microtubule corset [1,2]. The T. brucei cell contains at least eight single-copy organelles, including the (i) flagellum that emanates from the (ii) flagellar pocket in the posterior region of the cell and is attached along the entire cell body by the (iii) flagellar attachment zone (FAZ) [3]. The axoneme of the flagellum nucleates at the (iv) basal body, which aside from the well-conserved microtubule organization also associates with the (v) microtubule quartet that has an orientation (posterior to anterior) opposite of the microtubule corset [4]. Additional single-copy organelles are the (vi) Golgi and the (vii) mitochondrion [5,6] with its (viii) single mitochondrial genome named the kinetoplast DNA (kDNA). The kDNA is the feature structure of the Kinetoplastea and in T. brucei is organized in a complex structure of approximately 5000 interlocked mini-and approximately 25 maxicircles [7,8]. Replication and segregation of the mitochondrial genome has been studied in some detail, and it is thought that more than 100 proteins are likely to be involved in these two processes [9,10]. In recent years, a number of studies have started to deconstruct the segregation machinery (tripartite attachment complex, TAC) connecting the basal body of the flagellum to the kDNA.
Expansion microscopy is based on a series of chemical steps to physically magnify the cell and visualize ultrastructures by optical microscopy (for details see [11][12][13]; figure 1; table 1 summarizes details of antibodies used in this study). In the first step, the precursor molecules formaldehyde and acrylamide are introduced to functionalize the cellular components allowing the subsequent linkage of swellable polyelectrolyte gel (sodium acrylate). The gel formation is accomplished analogously to polyacrylamide gels by using oxidizing reagents like tetramethylethylenediamine and ammonium persulfate (APS). After the gel is formed, the cell contents are denatured and thus homogenized through the use of sodium dodecyl sulfate (SDS) and heat. At this point, the sample is ready for the application of standard immunofluorescence techniques. Expansion microscopy has been applied to a wide range of organisms and scientific questions, including membrane biology in bacteria, neuronal development in fish and cell structure of protozoan parasites, to name just a few [11,[14][15][16].
We now provide data to validate this new sample preparation technique in T. brucei and give an overview of the opportunities and challenges with ultrastructure expansion microscopy (U-ExM) in the protozoan model system.  Figure 1. Schematic representation of the ultrastructure expansion microscopy concept in T. brucei. First, cells were settled at room temperature on coverslips. Then, the coverslips were incubated in a 24-well plate filled with a solution of formaldehyde and acrylamide in PBS. Cells were subsequently prepared for gelation by carefully transferring coverslips into a gel solution supplemented with APS and TEMED. After gelation, the cell components were denatured using heat and SDS, which was followed by first round of expansion and gel shrinkage in PBS. Then, cells were stained with primary and secondary antibodies followed by dialysing in water for the final expansion. royalsocietypublishing.org/journal/rsob Open Biol. 11: 210132 the mitochondrial genome segregation machinery (TAP110) relative to its interaction partner TAC102 [17]. In that study, the isotropicity was measured using the basal body, the kDNA and the nucleus as references. The expansion factors of those cellular compartments/complexes varied less than 10% between 3.61 and 3.86. In the current study, we were able to increase the expansion factor to 4.2-4.6 as measured by the expansion of the kDNA, nucleus and cytoskeleton maintaining the variation at about 10% (figure 2a-c). Aside from the measurements of cellular compartments, the isotropicity of the expansion becomes obvious in the maintenance of the overall cell shape when compared to the non-expanded trypanosomes ( figure 2a,b). The ability to image DNA in the nucleus and the mitochondrion of expanded cells will in future allow the precise localization of expressed genes, centromeric sequences or overall chromatin architecture. Thus, we evaluated the use of the minor groove AT-rich DNA binder DAPI in U-ExM in T. brucei. The mitochondrial DNA of T. brucei and for most Kinetoplastea is organized in a complex structure of interlocked minicircles and maxicircles. Depending on the imaging angle, this results in the DAPI stain to appear as a homogeneous cylinder or 'disc-like' structure (figure 3a,b) [18]. In the kDNA, the minicircles are parallel to each other and stretched out maximally thereby defining the height of the cylinder. As a consequence of this organization, the kDNA cylinder can be isotropically expanded in its diameter but not in its height (figure 3a). With U-ExM, we can visualize individual regions within the kDNA disc that have previously only been detected by electron microscopy. This includes the inner unilateral filament region, a DNA containing domain between the kDNA disc and the inner mitochondrial membrane (figure 3b) [19]. If the 'cloudy' DAPI signal at the basal body proximal face of the kDNA disc is made up of minicircle replication intermediates as could be predicted by the current replication model [9] or potentially maxicircle segments remains to be investigated (figure 3a,b, asterisk). Furthermore, during replication and prior to segregation, the kinetoplast assumes a conformation that from the side resembles a V. In this conformation, the DAPI signal seems strongly concentrated in the middle between the two newly developing kDNAs (late kDNA replication phase), which is likely a combination of the imaging angle and the replication of maxicircles that potentially occurs at the interphase of the two replicating kDNAs [20] (figure 3b).

Results and discussion
In T. brucei, the cytoskeleton is dominated by microtubules that can be visualized by an alpha-tubulin antibody. Aside from the subpellicular microtubules, the antibody also allows detection of the entire axoneme from the base of the flagellum to its tip, revealing its continuous attachment to the cell body as previously described [21] (figure 2a). By using higher zoom factors, we can visualize the pro-and mature basal bodies with their associated microtubule quartet in proximity to the kDNA (figure 2a). The basal body is the central organizer of the cell cycle specifically involved in the duplication of the Golgi and the mitochondrion, including its singular genome [4,6,22]. Although the precise number remains unknown, probably more than 100 proteins are involved in the biogenesis and structure of the basal body, royalsocietypublishing.org/journal/rsob Open Biol. 11: 210132 all concentrated in a small area. Thus, precise localization of the individual components will be key for understanding their role in basal body biogenesis. Centrins, for example, are calcium-binding proteins that are highly conserved in all centrosomes. The monoclonal antibody 20H5 that was raised against the centrin in Chlamydomonas binds to centrin 1 and 2 at the basal body and the bilobe region in T. brucei, [6] (figure 4a). Using U-ExM, we can now show that centrin1/2 U-ExM confirms that similar to the recent data from cryo-tomography SAS6 is in the centre of the cartwheel and protrudes at the proximal end of the basal body [24]. We measured the protrusion to be 170 and 210 nm at the pro-basal body and mature basal body, respectively (figure 4e). In non-expanded cells, this would translate to 40-50 nm that the cartwheel would protrude from the proximal end of the basal body (figure 4e). Another commonly used reagent to label the basal body is the antibody YL1/2 that binds to tyrosinated α-tubulin and also cross-reacts with the basal body protein TbRP2 [25,26]. With U-ExM and at high dilutions of the primary antibody, we find tyrosinated tubulin to be more or less equally distributed at the old and new basal body. At these concentrations, the previously described cross-reactivity of YL1/2 with TbRP2 cannot be detected (figure 4c, asterisks).
One question that we wished to address was whether membrane-bound organelles could be analysed using U-ExM in T. brucei. To test this, we localized the archaic protein translocase of the outer membrane (ATOM), which is part of the mitochondrial protein import machinery in the outer mitochondrial membrane, TAC102 which is a TAC protein inside the mitochondrion and the luminal ER component binding protein (BiP) [27][28][29]. In regular confocal  (e) Quantification of SAS6 protrusion at the pro-basal body (the left graph) and basal body (the right graph). Imagery was deconvolved using Huygens Professional and visualized using Imaris.
royalsocietypublishing.org/journal/rsob Open Biol. 11: 210132 microscopy, ATOM is continuously distributed throughout the entire mitochondrial membrane. In expansion microscopy, the distribution of the protein is visible in several hundred individual spots per cell, suggesting individual complexes can be visualized (figure 5a). The ER-resident protein BiP on the other hand seems to be more continuously distributed throughout the ER lumen even in U-ExM (figure 5b), which is in good agreement with its function as a chaperone. TAC102 that was previously shown to reside inside the mitochondrion in proximity to the kDNA disc can be visualized using a monoclonal antibody (figure 5c)

Ultrastructure expansion microscopy challenges and opportunities
U-ExM offers several technical advantages for visualizing ultrastructures, but it also faces significant challenges. One of the main chemicals used in U-ExM is sodium acrylate. Its quality varies from batch to batch, and impurities negatively affect polymerization and ultimately the expansion factor. Another key challenge in expansion microscopy is the imaging process. The transparent aqueous gel starts shrinking after some time due to laser power and the evaporation of water out of the gel, which results in shifted images. Furthermore, sample storage is not as simple as for immunofluorescence slides. While immunofluorescence slides can be kept at four degrees for a long period of time, gel quality drops rapidly. It should also be noted that U-ExM requires significantly more antibody than typical immunofluorescence staining. Also, we have experienced potentially due to the denaturation of epitopes that not all antibodies that work in regular immunofluorescence microscopy can be used in U-ExM. U-ExM enables super-resolution imaging through the improvement of axial and lateral effective resolution by an isotropic increase in the distance between individual molecules to be analysed. Correspondingly, expansion by a factor of four increases resolution from around 200 nm to less than 50 nm using regular confocal microscopy and significantly further by super-resolution techniques. In future, this will allow us to precisely visualize and characterize biomolecular structures that are in close proximity to each other such as components within the basal body and nuclear pore complexes. In contrast with other high-resolution imaging techniques, U-ExM is accessible without the need of special equipment. In combination with protease treatment, the technique will likely allow improved intra-molecular resolution leading to in situ structure analysis, while using established reagents. Many antibodies and stains used for regular immunofluorescence microscopy show good results also in U-ExM. Furthermore, by tuning the sample functionalization steps and increasing the expansion factor, resolution in the singledigit nanometer range seems feasible. In summary, U-ExM has developed into a very powerful tool bridging the gap between light and electron microscopy.

Mounting and image acquisition
After the final expansion, gel size was measured with a caliper to calculate the fold expansion. The gel was then cut with a razor blade in pieces that fit in a 36 mm metallic chamber for imaging. The piece of gel was then mounted on a 24 mm round poly-D-lysine functionalized coverslip, already inserted in the metallic chamber and gently pressed with a brush to ensure adherence of the gel to the coverslip. Confocal microscopy was performed on a Leica TCS SP8 using a 63x 1.4 NA oil objective, with the following parameters: z step size at 0.3 μm interval with a pixel size of 35 nm. LAS X software (Leica Microsystems), Huygens Professionals and ImageJ were used to analyse the images.

Calculation of expansion from ultrastructure expansion microscopy
The expansion factor was determined by comparing the ratio of expanded cell length, kDNA and nucleus to non-expanded cell length, kDNA and nucleus. For unexpanded kDNA and nucleus measurements, n = 20 cells from immunofluorescence imagery were analysed. The length of the kDNA was determined by measuring the maximum length of the kDNA observed in each cell. The diameter of the nucleus was determined by measuring the widest diameter observed in each cell. The length of the cell structure was determined by measuring the cell length from anterior to posterior end in each cell.
Data accessibility. Data are all in the manuscript. Original imagery is available on request.