Analysis of Lipid Droplet Content in Fission and Budding Yeasts using Automated Image Processing

1. Preparation of Solutions and Media

1. Prepare lipid staining solution.
   1. To prepare stock lipid staining solution dissolve 10 mg of BODIPY 493/503 in 10 mL of anhydrous DMSO (final concentration 1 mg/mL). Dissolve the whole content of a 10 mg BODIPY 493/503 vial to prevent loss of material during weighing.
      CAUTION: DMSO may pass through the skin. Wear appropriate personal protective equipment.
   2. Prepare working lipid staining solution by mixing 100 µL of the 1 mg/mL BODIPY 493/503 stock solution and 900 µL of anhydrous DMSO (final concentration 0.1 mg/mL).
   3. Aliquot the stock and working solutions, and store at -20 °C.
      NOTE: Dissolved BODIPY 493/503 is stable for several years at -20 °C. However, the solution has to be protected from moisture and light.
2. To prepare stock solution for cell boundary visualization, dissolve 25 mg of Cascade Blue dextran (whole vial) in 2.5 mL of deionized water (final concentration 10 mg/mL). Aliquot the stock solution and store at -20 °C protected from light.
3. To prepare microscope slide coating solution, dissolve 5 mg of soybean lectin in 5 mL of deionized water (final concentration 1 mg/mL). Aliquot the lectin solution and store at -80 °C.
   NOTE: The lectin solution is stable for several years at -80 °C. Aliquots currently at use may be stored at -20 °C.
4. Prepare cultivation media.
   1. To prepare 400 mL of complex YES cultivation medium for S. pombe and S. japonicus, dissolve 2 g of yeast extract and 0.1 g of SP supplements (if required for auxotrophic mutants) in 340 mL of deionized water in a 500 mL bottle and autoclave. Add 60 mL of 20% (w/v) of separately autoclaved or filter-sterilized glucose in aseptic conditions.
   2. To prepare 400 mL of defined EMM cultivation medium for S. pombe and S. japonicus, dissolve 4.9 g of EMM broth without dextrose in 360 mL of deionized water in a 500 mL bottle and autoclave. Add 40 mL of 20% (w/v) of separately autoclaved or filter-sterilized glucose in aseptic conditions.
      NOTE: For general guidelines on S. pombe and S. japonicus cultivation see²⁵ and²⁶, respectively.
   3. To prepare 300 mL of complex YPAD cultivation medium for Saccharomyces cerevisiae, dissolve 3 g of yeast extract, 6 g of peptone and 30 mg of adenine sulphate in 270 mL of deionized water in a 500 mL bottle and autoclave. Add 30 mL of 20% (w/v) of separately autoclaved or filter-sterilized glucose in aseptic conditions.
   4. To prepare 300 mL of defined minimal medium for S. cerevisiae, dissolve 2 g of yeast nitrogen base (without amino acids) in 270 mL of deionized water in a 500 mL bottle and autoclave. Add 30 mL of 20% (w/v) of separately autoclaved or filter-sterilized glucose in aseptic conditions.
      NOTE: For general guidelines on S. cerevisiae cultivation see²⁷.

2. Cell Cultivation

1. Growing S. pombe or S. japonicus to exponential or early stationary phase.
   1. In the morning, inoculate 5 mL of YES medium with fresh fission yeast biomass. Incubate at 32 °C with shaking (180 rpm) for several hours.
      NOTE: For all cultivations, use Erlenmeyer flasks having 10 times the volume of culture to ensure proper aeration. Some laboratories prefer to grow fission yeasts at 30 °C, but cultivation temperature of 32 °C results in shorter doubling times without detrimental effects to the cells, thus reducing the total time required to perform an experiment^25,28.
   2. In late afternoon of the same day (after at least 6 hours of cultivation), dilute the culture with fresh YES medium to a 10 mL final culture volume so that it reaches the desired optical density (OD) (or number of cells/mL) the following morning, and incubate at 32 °C with shaking (180 rpm). It is of advantage to know the doubling time of each used strain to accurately determine the dilution factor (use Equation 1).
      
      Where V_culture is the preculture volume needed for dilution, V_final is the total volume of the new culture (10 mL for standard cultivations), OD_final is the desired OD to be reached the following morning, OD_current is the currently measured OD of the preculture, t is the time of cell growth until harvesting, t_lag is duration of the lag phase (depends on laboratory conditions, needs to be empirically defined) and t_DT is the doubling time of the strain.
      NOTE: When exponential-phase cells are to be analyzed, do not let precultures reach the stationary phase as this dramatically alters cell physiology (including LD content) for several subsequent generations.
   3. In the morning of imaging day, if the culture reached slightly higher OD than required (in case of exponential-phase cells), dilute it with fresh YES and continue incubation for at least two more doubling times before staining of LDs. Otherwise proceed directly to staining (Section 3).
2. Growing S. cerevisiae to exponential and stationary phase.
   1. In the afternoon, inoculate 10 mL of YPAD medium with a small amount of fresh budding yeast biomass and incubate overnight at 30 °C with shaking (180 rpm).
   2. The morning of imaging day, dilute the culture to OD 0.1 in 10 mL of YPAD medium and grow to the required OD (e.g., OD 1 for exponential phase). Perform any culture dilutions as described in step 2.1.2. Proceed to staining (Section 3).

3. Lipid Droplet Staining

1. Prepare a microscope cover slip for each sample to be imaged. Spread 1 µL of slide coating solution onto a clean cover slip using the long side of a horizontally positioned pipette tip. Allow the coating solution to dry completely and store the cover slips in a dust-free environment.
   NOTE: Glass slides and coverslips can be cleaned prior to use if required. The cleaning procedure consists of washing with dishwashing detergent, rinsing with water, overnight soaking in 3% hydrochloric acid, and washing with distilled water. Cleaned slides and coverslips are stored in pure ethanol until use.
2. Measure the OD of cell culture or number of cells/mL, as required. Try to reach similar values among all tested strains to ensure comparable experimental conditions.
3. Pipette 1 mL of each cell culture to a 1.5 mL microcentrifuge tube. For S. cerevisiae only, add 5 µL of the slide coating solution, vortex briefly, and incubate at 30 °C with shaking for 5 min.
4. Add 1 µL of the lipid staining solution to each culture aliquot and vortex briefly. Then add 10 µL of the cell boundary visualization solution and vortex briefly.
   NOTE: Do not prepare pre-mixed solutions of both stains as this leads to fluorescence quenching of BODIPY 493/503.
5. Collect the cells by centrifugation (1,000 x g, 3 min, RT) and remove almost all supernatant (~950 µL). Resuspend the cells in the remaining supernatant.
6. Pipette 2 µL of the dense cell suspension on a lectin-coated cover slip and place onto a clean microscope slide. The cells should form a monolayer. Proceed to microscopy (Section 4) as quickly as possible to minimize artefacts in imaging; process maximum of two samples at a time.

4. Setting up the Microscope and Imaging

1. Optimize imaging conditions.
   NOTE: Setting up the microscope requires long exposures to strong light sources that could cause damage to the sample and skew results. Therefore, set up the imaging conditions using a dedicated sample slide that will not be further used for LD quantification.
   1. Focus on the cells using phase contrast or differential interference contrast (DIC).
      NOTE: Phase contrast or DIC images may be taken for reference, but they are not used during the automated image analysis step.
   2. Set z-stack settings to span the whole cell volume. The total vertical distance depends on the cell size; the number of optical slices depends on the numerical aperture of the objective (point spread function in z-axis). Set the focus to move relative to the central focal plane.
      NOTE: The optimal number of slices is often set by the microscope control software and does not need to be calculated manually. The typical cell widths are 3-5 µm for S. pombe, 4-7 µm for S. japonicus, and 3-7 µm for S. cerevisiae.
   3. To image LDs, set light intensity and exposure time in the green channel (excitation and emission maxima of BODIPY are 493 and 503 nm, respectively).
      NOTE: BODIPY 493/503 is a very bright fluorochrome; however, it may get bleached rapidly with overly strong light intensity. Moreover, LDs are mobile in live cells, thus minimize exposure time and capture the full green-channel z-stack first (before switching to the blue channel) to prevent blurring artifacts. Also, take into account the linear range of the camera for signal intensity to avoid saturated pixels.
   4. To image cell boundaries, set light intensity and exposure time in the blue channel (excitation and emission maxima of Cascade Blue dextran are 400 and 420 nm, respectively).
      NOTE: Signal intensity in the blue channel is required for image segmentation, but it is not used for LD quantification itself. Therefore, optimal settings in this channel are not crucial for analysis.
   5. If possible, create an automated experimental workflow in the microscope control software to facilitate imaging of multiple samples under standardized conditions.
2. Once imaging conditions have been optimized, image samples to be used for quantification. Focus on the cells and image them in green and blue channels as described in Step 4.1.
   NOTE: All images must be acquired using the same settings to allow comparison between samples. Image multiple fields of view per sample to obtain robust, representative data.
3. Save the blue and green channel z-stack images as 16-bit multi-layer TIFF files (i.e., two files per field of view). Include words “green” or “blue” in the corresponding file names. Proceed with image analysis (Section 5).

5. Image Analysis

1. Visually check the quality of acquired images.
   1. Open microscopic images in ImageJ^29,30 or other suitable image analysis software.
   2. Remove any image stacks containing a considerable number of cells that moved during acquisition (and thus created blurring artifacts).
   3. Remove any image stacks containing highly fluorescent non-cell particles in the blue channel (e.g., dirt on microscope slide or cover slip, impurities in cultivation medium).
      NOTE: Very bright non-cell objects in the blue channel may create cell detection artifacts or may interfere with detection of cells in their vicinity.
   4. Remove any image stacks containing a large proportion of dead cells (i.e., cells with increased blue fluorescence compared to live cells).
      NOTE: While the presence of a small proportion of dead cells in the sample is typically not a problem and these cells are automatically discarded during analysis, some dead or dying cells may occasionally be recognized as live cells by the segmentation algorithm and thus skew the reported results.
2. Analyze images in the MATLAB software.
   1. Create a main folder and copy all MATLAB scripts to this location.
   2. Create a sub-folder (“pombe”, “cerevisiae” or “japonicus”) and copy input TIFF image files to this location.
   3. Start MATLAB, open script MAIN.m and run it. In the menu select the yeast species to be analyzed and start image processing.
      NOTE: Some of the parameters required for cell and LD detection are pre-set for the particular species, others are determined automatically during image processing. The pre-set values were determined empirically and depend on several factors such as objective magnification, camera type and sensitivity, and imaging settings. If required, users may edit the script files to change the organism-specific presets to better reflect their experimental setup. Namely, during cell recognition acceptable object sizes are given by the “minArea” and “maxArea” parameters, and the minimum fraction of filled volume within the object boundaries is given by the “Solidity” parameter. For LD recognition, the brightness threshold is given by the “th” parameter (its value is affected mostly by image bit depth and fluorescence signal intensity), and maximum acceptable LD size is given by the “MaxArea” parameter.
   4. Inspect and process the output files as required using a spreadsheet editor or statistical package; the workflow produces semicolon-separated CSV files, and segmented TIFF files with detected cell objects and LDs.
      NOTE: The workflow segments images into background and cell objects, where each cell object may be composed of multiple adjacent cells. Therefore, the output in “xxxx_cells.csv” files does not represent single-cell data and should only be used to calculate per-unit-of-cell-volume metrics.