S. acidocaldarius DSM 639 (wild-type) or MW001 (uracil auxotrophic cloning strain) were grown in a shaking incubator at 75°C in Brock medium pH 2.9 supplemented with 0.1% N-Z-amine and 0.2% sucrose. MW001 cultures were supplemented with 4 μg/mL uracil. All cultures used for imaging were collected during exponential growth phase at an optical density at 600 nm (OD_600nm) of 0.1 to ~0.4.

All dyes in listed in Tables 1–3 were dissolved in DMSO and added to a final concentration of no greater than 0.1% (v/v). Note that much higher concentrations of DMSO (>10%) can be added to cultures without it interfering with growth. Staining was performed for either 5 min at either room temperature (23°C) or for 1 min at 75°C. The signal from dyes that stained cells over background during this time-frame did not greatly improve upon longer incubation times. Similarly, dyes that failed to label cells within 5 min at room temperature did not successfully stain cells over incubation times of up to 30 min. To settle on conditions for live imaging we tested candidate probes at different concentrations over different exposure times at room temperature. Probes that successfully stained cells at room temperature were then re-tested at the same concentrations at 75°C using different exposure times. Probes that successfully stained cells at 75°C were then tested for cytotoxicity (without illumination) by performing growth curves over 24 h in presence or absence of the dye at concentrations determined above. Only DiO(C6) was found to be cytotoxic at concentrations used for live cell imaging at either room temperature or 75°C. Note that because cells were not washing following labelling, it likely that the pool of dye present in the medium can exchange with the cellular pool – reducing the effects of photobleaching.

Live-cell imaging was performed at 75°C using the “Sulfoscope” chamber described in Pulschen et al. (2020), with modifications to the hardware described by Hurtig et al. (2023). Briefly, 25 mm coverslips were washed with EtOH and H₂O, then assembled into commercial Attofluor chambers (Invitrogen A7816). Chambers were filled with 300 μL of Brock medium and incubated at 75°C for at least 1 h or until the medium was dry. Afterwards, chambers were washed thoroughly with Brock medium, placed into the Sulfoscope chamber, and allowed to equilibrate to 75°C. Dyes detailed in Tables 1–3 were added to 5 mL of 75°C S. acidocaldarius cell culture (OD_600nm 0.15 to 0.3) immediately before imaging. For imaging, 400 μL of cell suspension (OD_600nm 0.15 to 0.3) was added into the chamber and immobilized using heated, semi-solid gelrite pads (0.6% Gelrite, 0.5× Brock medium pH 5, and a final concentration of 20 mM CaCl₂). Pads were prepared in the following way: ~15 ml molten Gelrite Brock medium solution was added to 9cm plastic petri dishes, and allowed to set at room temperature (~5 minutes). Half-moon shapes were then cut from the plate with a 7 mm diameter circle punch, placed onto 13 mm circular coverslips, and incubated at 75°C for 5 to 10 min in a bead bath. During this period of incubation, pads equilibrated to the imaging temperature and also dried slightly, causing the edges of the pad to curve downwards. Preheated pads were then placed in the chamber onto the cell suspension, such that the concave edge of the pad was in the center of the chamber. For the cell biological analysis, cells at the border of the immobilization pad were imaged since, in this area, cells are immobilized without being subjected to mechanical stress from the overlying Gelrite. Images were acquired on a Nikon Eclipse Ti2 inverted microscope equipped with a Yokogawa SoRa scanner unit and Prime 95B sCMOS camera (Photometrics). Imaging was performed with a 60× oil immersion objective (Plan Apo 60×/1.45, Nikon) using a custom formulated immersion oil for high temperature imaging (maximum refractive index matching at 70°C, n = 1.515 ± 0.0005; Cargille Laboratories), using the ×2.8 magnification of the SoRa unit (equivalent to a total magnification of ×168). Images were acquired using a 15 ms exposure time and 10% laser power at intervals of 15 s for 2 to 3 h. After acquisition, XY drift was corrected using the ImageJ plugin StackReg (Thevenaz et al., 1998).

1 mL DSM 639 culture in Brock medium was allowed to cool to room temperature before staining with the dyes listed in Tables 1–3, as described above. Imaging was performed on cells confined using a 1% low melt agarose pad. In brief: a 1% low melting temperature agarose (Sigma Aldrich, A9414) was prepared in MilliQ water by microwaving until the agarose was completely dissolved. 100 μL of molten agarose solution was pipetted onto a homemade spacer slide consisting of a glass slide with 4 layers of lab tape wrapped around either end. A second glass slide was then placed on top of the agarose, which was allowed to harden for 2–5 min at room temperature. Once the agarose had hardened, the spacer slide was removed and 10 μL of labelled cell suspension was added onto the agarose pad and allowed to dry fully before a 13 mm borosilicate coverslip was placed on top for imaging. Imaging was performed using the inverted microscopy set-up described above. Images were acquired with a NIKON 100x oil immersion objective (Apo TIRF 100x/1.49) and type F2 immersion oil (Nikon) in addition to the 2.8x magnification lens in the SoRA unit giving a total magnification of 280x. Z-stack images were acquired with a 0.22 μm step size (10 slices, covering ~2 μm) using an exposure time of 50 ms with laser power set to 10% of maximum.

For Stepwise ethanol fixation, 3 mL DSM 639 culture in Brock medium was added to 1.5 mL 4°C ethanol, incubated at 4°C for 10 min before adding a further 1.5 mL 4°C ethanol, incubated for 10 min, after which a final 4 mL of 4°C ethanol was added to a final concentration of 70%. For fixation in other buffer conditions, 3 mL of culture was spun for 3 min at 8000RPM in a table top centrifuge and resuspended in 3 mL of either Tris Buffer (25 mM Tris pH 7.4, 150 mM NaCl) or Citrate Buffer (25 mM Sodium Citrate, pH 3). For Instant ethanol fixation, 1 mL DSM 639 culture in Brock medium was added directly to 9 mL 77% 4°C ethanol. For Formaldehyde fixation 3 mL DSM 639 culture was spun for 3 min at 8000RPM in a table top centrifuge and resuspended in 1 mL 4% freshly prepared paraformaldehyde (PFA) in H₂O and incubated at room temperature with shaking for 10 min. Cells were then washed with 1 mL phosphate-buffered saline supplemented with 0.1% Tween 20 (PBST), resuspended in phospho-buffered saline supplemented with 0.1% Triton X-100 and incubated at room temperature with shaking for 10 min in order to permeabilize. Cells were then washed with 1 mL PBST before proceeding to immunolabelling. All samples were stored at 4°C before labelling and imaging.

Immunolabelling was performed as described by Hurtig et al. (2023). Briefly, 1 mL fixed cells was spun in a tabletop centrifuge (3 min, 8000RPM) after which the supernatant was discarded and cells were washed twice in 1 mL PBST supplemented with 3% bovine serum albumin (PBSTA) to remove all remaining fixative. Cells were resuspended in a final volume of 100 μL PBSTA supplemented with 5% fetal bovine serum (FBS) and primary antibodies (in this case a lab generated anti-CdvB: Tarrason Risa et al., 2020; Hurtig et al., 2023). Cells were incubated overnight at room temperature with 500 rpm agitation before washing with 1 mL PBSTA and resuspending in a final volume of 100 μL supplemented with secondary antibodies (either AlexaFluor-488 anti-rabbit (Thermo Fisher Scientific, A11034) or AlexaFluor 546 anti-rabbit (Thermo Fisher Scientific, A11035), 1:10,000) and 50 μg/mL Concanavalin A conjugated to Alexa Fluor 647 (Thermo Fisher Scientific, C21421). Cells were incubated for 3 h at room temperature and 500 rpm agitation after which cells were washed with 1 mL PBSTA and resuspended in a final volume of 1 mL supplemented with 3 μM DAPI (4′,6-diamidino-2-phenylindole; Thermo Fisher Scientific, 62,248). For imaging, Lab-Tek chambered slides (Thermo Fisher Scientific, 177437PK) were coated with 2% polyethyleneimine (PEI) at 37°C for a minimum of 30 min. Coated chambers were washed with Milli-Q water before 200 μL cell suspension was added per well and spun down for 1 h at 750 relative centrifugal force (RCF). Imaging was performed, as for live-cell imaging, at room temperature using an exposure time of 200 ms for detection of secondary antibodies and an exposure time of 500 ms for detection of DNA. Analysis and z-axis maximum projections were performed using ImageJ.

The microscopy set-up used for live-cell imaging includes a heated cap and stage that functions to maintain a temperature of 75°C for several hours without dehydration (Pulschen et al., 2020). In addition, we recently added a Yokogawa SoRa unit to our confocal microscope (Azuma and Kei, 2015; Hurtig et al., 2023), enabling the resolution of discrete subcellular structures in ~1 μm diameter S. acidocaldarius cells. In previous live-imaging work, cells were stained with NileRed to mark the membrane and SybrSafe to label the DNA in order to visualize and characterize cell division in S. acidocaldarius (Pulschen et al., 2020). This is improved by using CellMask Deep Red Plasma Membrane Stain (CellMask) as a membrane stain (Figure 1B). CellMask provides a brighter and more specific membrane signal compared to NileRed, at both 75°C and room temperature (Figure 2A). Unlike NileRed, which inserts into the lipid core, CellMask is composed of a hydrophilic fluorophore attached to a lipophilic tail which inserts into the membrane. The improved resolution of the membrane signal relative to the cytoplasmic signal may therefore reflect the inability of CellMask to cross the archaeal membrane.

Importantly, under these conditions, control cells (MW001) labelled with CellMask were observed undergoing constriction over a period of ~12 min at a rate of ~0.1 μm/s (Hurtig et al., 2023), in line with speeds previously reported by Pulschen et al. (2020). This indicates that though CellMask is brighter and more specific than NileRed, it does not cause additional imaging stress. We did not observe significant phototoxicity over longer imaging periods of 2–3 h. Additional dyes that are compatible with imaging at 75°C, include the CellMask Orange Plasma Membrane Stain and Mitotracker Red CMXROS, although the latter also labels the cytoplasm (Table 1). Unfortunately, we were unable to identify a good label that could be used to image the proteinaceous surface layer at 75°C (see Table 2). A recent study however was able to visualize the cell contour of S. acidocaldarius by using a N-Hydroxysuccinimide (NHS)-ester functionalized Alexa Fluor dye to non-specifically label surface proteins in phospho-buffered saline at room temperature then returning cells into Brock medium for live-cell imaging at 75°C (Charles-Orszag et al., 2023).

While a range of dyes were found to strongly label the DNA of S. acidocaldarius cells at 75°C, the only DNA dye tested that proved compatible with long-term live cell imaging was SybrSafe (Table 3). Other dyes exhibited cytotoxicity and rapid bleaching. Nevertheless, the hardware improvements we have put in place allow for precise visualization of DNA morphology and dynamics using SybrSafe. Under these conditions, the S. acidocaldarius genome appears to be crescent shaped, lying close to the membrane on one side of the cell during interphase. Prior to division, it then compacts to form two foci that align with the plane of cell cleavage (Figure 1B). The organization of the genome and the membrane dynamics visualized by live cell imaging at 75°C using SybrSafe and CellMask was used to validate the other stains and probes described below.

While only a limited number of markers were found to be compatible with live imaging at 75°C, many more were found to be compatible with imaging at room temperature. In addition to the dyes described above, Mitotracker dyes, which have been used to visualize the membranes of euryarchaeota and DPANN (Maslov et al., 2018; Hamm et al., 2023), were compatible with staining the membrane of S. acidocaldarius cells in Brock culture medium at room temperature (Figure 2A and Table 1). Interestingly, eukaryotic type lipid stains such as BODIPY-Ceramide were also found to weakly stain the membrane of S. acidocaldarius cells, providing hope that other tools developed to visualize lipids in eukaryotes can be adapted to the study of the archaeal membrane. Long-chain carbocyanine dyes (e.g., DiI C18 etc), commonly used to visualize eukaryotic and bacterial membranes, did not stain the archaeal bounding membrane in culture medium or in Tris Buffer (pH 7.4). The short-chain carbocyanine DiO C6 was able to stain cells in Brock medium at room temperature, but proved to be cytotoxic at 75°C.

At room temperature, DNA could be stained with a number of STYO nucleic acid stains with different spectral properties (Figure 2B). While the signal at the start of imaging was good, these dyes suffered significant photobleaching when compared with SybrSafe. Nevertheless, these dyes could be used for labelling DNA in fixed cells, where photobleaching is less of a problem.

In addition to testing DNA and membrane dyes, we also tested a host of markers that we hoped would label membrane proteins and the S-layer in live cells. Unfortunately, these did not stain cells in low pH Brock medium, Citrate Buffer, or Tris Buffer. An exception was CellBrite, which targets primary amines. Unfortunately, CellBrite only marked the cell periphery at a high pH (Figure 2C), conditions that compromise DNA and membrane organization and, likely, S. acidocaldarius viability. Interestingly, the cell content marker CellTracker CMFDA (5-chloromethylfluorescein diacetate) efficiently crossed the plasma membrane to provide a uniform labeling of the S. acidocaldarius cytoplasm in low pH Brock medium at both room temperature and 75°C.

In the absence of genetically encoded fluorescent proteins for use in hyperthermophiles, visualizing proteins and protein structures in these organisms currently relies on immunofluorescence. Unfortunately, as detergent and solvent permeabilization both greatly impact membrane integrity this precluded visualization of the membrane in fixed cells.

We have tested and optimized a range of fixation techniques for S. acidocaldarius. Our standard procedure for staining S. acidocaldarius cells employs a “Stepwise fixation” protocol in ethanol (Figure 3A, left). This involves adding 3 mL of culture to 1.5 mL of ethanol. After 10 min on ice a further 1.5 mL ethanol is added and after an additional 10 min on ice, the final volume is brought to 10 mL with ethanol, to a final concentration of 70% ethanol (Bernander and Poplawski, 1997; Han et al., 2017; Zhang et al., 2019). Importantly, this fixation protocol yields a DNA signal similar to that observed by live cell imaging. Cells fixed in this way, can also be effectively labelled with antibodies targeting ESCRT-III homologues to reveal division rings like those reported in previous studies (Samson et al., 2008, 2011; Tarrason Risa et al., 2020; Hurtig et al., 2023). In addition, we are able to use fluorescently conjugated Concanavalin A (ConA) to label glycosylated proteins and the cell contour in ethanol fixed cells. While this stain has been used as a proxy for the S-layer, which we know to be heavily glycosylated in S. acidocaldarius (Peyfoon et al., 2010), ConA likely also marks a range of other glycosylated membrane proteins. Note that ConA also causes cells to aggregate in a concentration dependent manner.

We attempted to streamline the two-step ethanol fixation protocol by adding 1 mL of culture directly to 9 mL of 77% ethanol, yielding a final concentration of 70% ethanol. Unfortunately, this “Instant fixation protocol” failed to preserve the structure of neither DNA nor division rings (Figure 3A, middle). While attempts to fix cells with formaldehyde (either paraformaldehyde, glutaraldehyde or a combination of the two) were able to fix DNA so that its organization resembled that seen during live cell imaging, it did not preserve division rings (Figure 3A, right). Formaldehyde fixation was able to preserve ConA labelling of the membrane/S-layer, however, the staining appeared weaker than when using the Stepwise ethanol fixation. Combining formaldehyde fixation with ethanol fixation by replacing the detergent permeabilization portion of the formaldehyde fixation protocol with the stepwise ethanol fixation protocol did not appear to preserve CdvB division ring structures (data not shown).

As the Stepwise ethanol fixation was the only method tested that faithfully preserved DNA organization and division rings, we used this as a starting point to test the impact of different buffer conditions on immunolabelling (Figure 3B). For this analysis, cells were spun down and resuspended in either a minimal Brock medium (mBrock: pH 5, no supplementation with NZ-amine or FeSO4), Tris or Citrate buffer, before being fixed using the Stepwise ethanol protocol described above. Interestingly, in both mBrock and Tris Buffer, CdvB protein structures appeared smoother and more continuous than when fixation was performed in Brock culture medium. However, DNA organization was entirely disrupted in Tris Buffer, in both fixed and live cell imaging at room temperature (Figure 3C). By contrast, fixation in Citrate buffer preserved both DNA and protein organization, but the signal was reduced in both cases compared to that observed following fixation in Brock medium. While the ESCRT-III signal appeared slightly more uniform in Citrate buffer, until the advent of thermostable GFP variants that work in Sulfolobus, it will not be possible to determine whether or not this represents the ring in its physiological state.