Establishing Live-Cell Single-Molecule Localization Microscopy Imaging and Single-Particle Tracking in the Archaeon Haloferax volcanii

In recent years, fluorescence microscopy techniques for the localization and tracking of single molecules in living cells have become well-established and are indispensable tools for the investigation of cellular biology and in vivo biochemistry of many bacterial and eukaryotic organisms. Nevertheless, these techniques are still not established for imaging archaea. Their establishment as a standard tool for the study of archaea will be a decisive milestone for the exploration of this branch of life and its unique biology. Here, we have developed a reliable protocol for the study of the archaeon Haloferax volcanii. We have generated an autofluorescence-free H. volcanii strain, evaluated several fluorescent proteins for their suitability to serve as single-molecule fluorescence markers and codon-optimized them to work under optimal H. volcanii cultivation conditions. We found that two of them, Dendra2Hfx and PAmCherry1Hfx, provide state-of-the-art single-molecule imaging. Our strategy is quantitative and allows dual-color imaging of two targets in the same field of view (FOV) as well as DNA co-staining. We present the first single-molecule localization microscopy (SMLM) images of the subcellular organization and dynamics of two crucial intracellular proteins in living H. volcanii cells, FtsZ1, which shows complex structures in the cell division ring, and RNA polymerase, which localizes around the periphery of the cellular DNA. This work should provide incentive to develop SMLM strategies for other archaeal organisms in the near future.


Supplementary Information
Establishing live-cell single-molecule localization microscopy imaging and single-particle tracking in the archaeon Haloferax volcanii Bartosz Turkowyd* 1 , Sandra Schreiber* 2 , Julia Wörtz 2 , Ella Shtifman Segal 3 , Moshe Mevarech 3 , Iain G. Duggin 4 , Anita Marchfelder# 2 , Ulrike Endesfelder# 1 Background sensitivity of single-molecule imaging Fluorescence microscopy studies can be compromised by autofluorescence background. While conventional fluorescence microscopy in wild-type red Haloferax cells is generally possible as already a few GFP together provide sufficient signal above background, single-molecule imaging is severely compromised, especially at high frame rates of dynamic imaging.
The background level is an important factor in the detection and localization of single molecules. If signals are not sufficiently brighter than the background, a large proportion of the weaker signals will remain undetected, and brighter signals that are still visible above the background are localized at lower precision, directly limiting the overall resolution of the super-resolved image.
In our measurements, WR806 cells significantly improve single-molecule imaging conditions for Haloferax: Supplementary Figure 2 shows that our fluorescent proteins yield on average ~ 200-250 photons per image (corresponding to 21.000 to 27.000 AD counts), typical numbers for single-molecule imaging. The pixel size is optimized for the detection and localization of single fluorescent spots and individual spots are measured on an area of about 3x3 pixels. Thus, the average signal of a single fluorescent protein is not much stronger than the average background of 1.300 AD counts per pixel in H119 cells (Figure 1b), which results a 9-pixel background signal of 11.700 AD counts (translates into ~ 107 photons for our setup). In contrast, WR806 cells are indistinguishable from the background outside cells and single fluorescent protein signals can be detected above background.

Sequences of fluorescent proteins
Dendra2 (E. coli opt)

Supplementary Figure 1. Cell phenotype varies under the production of different fluorescent proteins
All strains producing fluorescent proteins which showed efficient photoactivation/conversion (Figure 1c) showed similar phenotypes which are comparable to the parental strains H119 and WR806. In contrast, the strain producing mMaple3 shows strongly altered morphology. This and the fact that mMaple3 showed no fluorescence (Figure 1c) suggests improper protein folding that might cause increased cytotoxicity leading to a different growth phenotype.

Supplementary Figure 2. Dendra2Hfx performs superior to PAmCherry1Hfx.
(a) WR806 cells producing either FtsZ1-Dendra2Hfx or FtsZ1-PAmCherry1Hfx were mixed on the same agarose pad and imaged under different UV laser intensities. When 0.5 W/cm² of the UV light was applied, FtsZ1-Dendra2Hfx revealed good photoconversion which results in dense SMLM ring structures. In contrast, FtsZ1-PAmCherry1Hfx was not photoactivated. A read-out of FtsZ1-PAmCherry1Hfx became only possible when increasing the UV light intensity to 1.5 W/cm². (b) To provide an image quality comparable to FtsZ1-Dendra2Hfx when imaging FtsZ1-PAmCherry1Hfx, the UV light intensity had to be significantly increased. 10 W/cm² was barely sufficient to achieve the same quality as for the FtsZ1-Dendra2Hfx structure when applying 0.5W/cm². (c) Under imaging conditions as in (a), Dendra2Hfx emits on average more photons (median of 251 photons) than PAmCherry1Hfx (median of 217 photons) per fluorescent emission in a single imaging frame.