Total Internal Reflection Fluorescence (TIRF) microscopy is a powerful approach to observe structures close to the cell surface at high contrast and temporal resolution. We demonstrate how TIRF can be employed to study protein dynamics at the cortex of cell wall-enclosed bacterial and fungal cells.
TIRF microscopy has emerged as a powerful imaging technology to study spatio-temporal dynamics of fluorescent molecules in vitro and in living cells1. The optical phenomenon of total internal reflection occurs when light passes from a medium with high refractive index into a medium with low refractive index at an angle larger than a characteristic critical angle (i.e. closer to being parallel with the boundary). Although all light is reflected back under such conditions, an evanescent wave is created that propagates across and along the boundary, which decays exponentially with distance, and only penetrates sample areas that are 100-200 nm near the interface2. In addition to providing superior axial resolution, the reduced excitation of out of focus fluorophores creates a very high signal to noise ratios and minimizes damaging effects of photobleaching2,3. Being a widefield technique, TIRF also allows faster image acquisition than most scanning based confocal setups.
At first glance, the low penetration depth of TIRF seems to be incompatible with imaging of bacterial and fungal cells, which are often surrounded by thick cell walls. On the contrary, we have found that the cell walls of yeast and bacterial cells actually improve the usability of TIRF and increase the range of observable structures4-8. Many cellular processes can therefore be directly accessed by TIRF in small, single-cell microorganisms, which often offer powerful genetic manipulation techniques. This allows us to perform in vivo biochemistry experiments, where kinetics of protein interactions and activities can be directly assessed in living cells.
We describe here the individual steps required to obtain high quality TIRF images for Saccharomyces cerevisiae or Bacillus subtilis cells. We point out various problems that can affect TIRF visualization of fluorescent probes in cells and illustrate the procedure with several application examples. Finally, we demonstrate how TIRF images can be further improved using established image restoration techniques.
1. Sample Preparation
2. Image Acquisition
3. Representative Results
4. Representative Results
Figure 1. Adjusting parameters. A. Dirty vs. cleaned coverslip. Background signal arising from dust on the coverslip interfere with a potential GFP/RFP signal, cleaned coverslip has less background. B. Custom-made electrical heating controller showing set temperature (a), temperature at probe (b) and sample holder (c). C. Yeast protein Pma1GFP imaged with decreasing incidence angles (from top left to bottom right). Images show a sudden loss of signal at the critical angle and progressive decrease in structural information and signal to noise ratio. D. Uneven field of view. A dense suspension of fluorescent beads is imaged, showing that only part of the field of view is in focus. Scale bars in A, B: 2 μm, in C: 10 μm.
Figure 2. Double color imaging. A. Bleed through of Pil1RFP. Pil1RFP is excited with 488 nm and a 561 nm laser and fluorescence is recorded with a double band pass filter. Signal of Pil1RFP is weakly visible in the GFP channel. Scale bar 2 μm B. Typical steps to avoid bleed through in double color imaging. After imaging the cells with separate filter sets, they are deconvolved and aligned (using fluorescent beads) before being merged for analysis.
Figure 3. Representative Results. A. Comparison of GFPMbl distribution in B. subtilis using TIRF or regular epifluorescence (Epi) microscopy. Blue: Cell boundary from bright field image. Images represent average projections of a time series to indicate the area covered by motile MreB patches monitored with different imaging modes. B. Improved axial resolution and contrast of TIRF images taken of the transpeptidase PbpHGFP in B. subtilis. Arrows indicate septal staining that appears as ring in epifluorescence and as patches in TIRF. Exposure time: 100 ms. C. Comparison between epifluorescence and TIRF illumination of the TOR complex component Bit61 in S. cerevisiae. Bit61 is very weakly expressed9. Exposure time: 250 ms. D. Comparison of different image processing methods. Series showing raw TIRF image of Pma1GFP, subtraction of Gaussian or Median blurred images, as well as deconvolution using the maximum likelihood algorithm in Huygens. E. TIRF-FRAP of a single GFPMbl patch (above asterisk) moving across a B. subtilis cell. The kymograph of the non-recovering patch and the intensity profile along the kymograph (dotted line) rule out treadmilling as source of motion. F. TIRF-FRAP of a yeast cell expressing Pma1GFP. Note the diffusive closing of the gap in the cortical staining pattern. Scale bars: 2 μm. Time in seconds.
Movie 1. Movie showing B. subtilis cells expressing an RFP fusion protein on a dirty coverslip. Note that RFP signal is hardly distinguishable from background noise. Cycle time 100 ms. Click here to view movie.
Movie 2. Movie of Pma1GFP with different incidence angles. Angle is altered from subcritical to critical angle, until signal is lost. Note, at subcritical angle internal signal is visible, causing noisy images until at critical angle just the protein at the PM is visible. Cycle time 200 ms. Scale bar: 2 μm. Click here to view movie.
Movie 3. Movie of yeast GFPRas2, showing alternating RAW and deconvolved TIRF images. GFPRas2 is a fast moving protein (T1/2 of 2s unpublished data), fast acquisition times are important to resolve the dynamic behavior. Cycle time 150 ms. Scale bar: 2 μm. Click here to view movie.
Movie 4. Double color movie of yeast proteins Fet3GFP and Pma1RFP. First 10 frames represent RAW images and last frames are deconvolved. This movie shows the importance of image processing for colocalization analysis. The blurry RAW images become sharpened, making an analysis possible. Cycle time 5 s. Scale bar: 2 μm. Click here to view movie.
TIRF microscopy is the technique of choice to image peripheral proteins. The low penetration depth of the evanescent field minimizes of out of focus light, which leads to a very good signal to noise ratio and allows data acquisition with high frame rates, or imaging of very weakly expressed proteins. The combination of high contrast and high frame rates makes TIRF microscopy a perfect tool for imaging spatio-temporal dynamics of cortically localized proteins. Interestingly the thick cell wall surrounding many microorganisms does not hamper imaging of peripheral proteins by TIRF. In fact, the actual generation of the evanescent very likely occurs at the interface between cell wall and plasma membrane5,10. Reflection of light that exits the cell might even lead to propagation of light along the cell wall, which might explain the large surface area that could be imaged in yeast cells8.
For optimal quality of TIRF images it is crucial to have a well calibrated microscopy system. The laser should be focused and aligned before each imaging session. Coverslips should be cleaned (Fig. 1A) and the cells have to be properly immobilized. For double color imaging spectral bleed through has to be taken into account or eliminated by using filter sets filters separate emission range. This will obviously come at the price of slower acquisition times due to mechanical filter change. Fast filter wheels or galvanometer driven illumination sources can circumvent this delay if necessary. Alternatively, the critical fluorophore (RFP) can be used on the weaker expressed protein in a pair, minimizing the level of signal interference.
Despite of the high contrast, images taken with a TIRF microscope contain significant noise. To remove this noise and further increase image contrast images can be processed with various filtering steps. The method with the best results in our experience is 2D deconvolution. This requires the experimental measurement of the PSF for each sample and microscope condition. However, this can be performed fairly easily and quickly using fluorescent beads mixed into the respective sample. For two-color experiments these beads can additionally be used for channel alignment.
We have demonstrated the advantages of using TIRF microscopy for imaging of cortical proteins in microorganisms. Combination of TIRF and image processing allows acquisition of images with high frame rates (<50 ms) and contrast and also enables visualization of weakly expressed proteins such as Bit61. An additional advantage of performing TIRF imaging on microorganism is that turgor pressure in these cells leads to flat plasma membranes, and minimizes topological effects to TIRF signals.
The authors have nothing to disclose.
This work was funded by the Max Planck society.
Name of the tool/reagent | Type | Company | Catalogue number |
Orbital Shaker | Tool | UniEquip | UNITWIST 3-D ROCKER SHAKER |
TIRF microscope | Till | Customized setup | |
Glass container | Tool | Vitlab | 340-232880353 |
Ceramic staining rack | Tool | Thomas Sci. | 8542E40 |
Concanavalin A | Reagent | Sigma | L7647 |
Coverslips #1(18 x 18 mm) | Microscope | Menzel Gläser | BB018018A1 |
Microscope Slides | Microscope | Menzel Gläser | AA00000102E |
Immersion Oil | Microscope | Zeiss | Immersol 518F |
Agarose | Reagent | Invitrogen | 16500-500 |
FluoSpheres | Reagent | Invitrogen | F8795 |