4D Microscopy of Yeast

This protocol describes the analysis of fluorescently labeled intracellular compartments in budding yeast using multi-color 4D (time-lapse 3D) confocal microscopy. The imaging parameters are chosen to capture adequate signals while limiting photodamage. Custom ImageJ plugins allow labeled structures to be tracked and quantitatively analyzed.

This method tracks structures in yeast cells in three dimensions over many minutes, allowing us to study the dynamics of intracellular organelles and compartments. 4D imaging ensures that we can draw reliable conclusions about intracellular dynamics, particularly for structures or markers that may be transient. Demonstrating the procedure will be Natalie Johnson, a postdoc from my laboratory.

Begin by growing an overnight culture of the yeast strain of interest in five milliliters of non-fluorescent synthetic defined, or NSD, medium, in a 15 milliliter baffled flask with good aeration, at 23 degrees Celsius. Three to four hours before the analysis, dilute the logarithmic phase yeast culture in fresh NSD medium, so that the final optical density at 600 nanometers, or OD600, will be 0.5 to 0.8 at the time of imaging. At least one hour before the culture is ready, centrifuge an aliquot of a two milligrams per milliliter Concanavalin A solution for five minutes at full speed, to pellet any particulates, and add 250 microliters of the supernatant into a clean 35 millimeter glass bottom microscopy dish.

After 15 minutes, wash the dish two to three times with two milliliters of distilled water per wash, adding 250 microliters of the yeast culture to the Concanavalin A-coated dish after it is dried. Wait 10 minutes to allow the cells to adhere, before gently washing the dish two to three more times with two milliliters of fresh NSD medium. Then cover the cells with two milliliters of fresh NSD.

To image the yeast cells, select a 63X oil immersion lens with a numerical aperture of at least 1.4 on a confocal light microscope. Here, a 100X objective lens can also be used instead of a 63X objective lens. Then, place the dish on the objective lens, spotted with immersion oil.

From the dropdown menu of the Acquisition Mode tab, select xyzt. Under the XY tab, format the frame size to 256 width by 128 height. Use the maximum scan speed, which is typically on the order of eight kilohertz.

Turn on bidirectional X scanning if it is available, and adjust the zoom factor to nine, which will result in a pixel size of approximately 80 nanometers. If a 100X objective is used, adjust the zoom factor accordingly to maintain a pixel size of approximately 80 nanometers. Then set the line accumulation to four or six.

Set the pinhole to 1.2 Airy units, and turn on the white light laser, if available. Then set the excitation wavelength and percent laser power for each fluorescence channel. If available, enable the use of photon counting mode under the configuration tab by deselecting maximum integration time.

For each fluorescence channel, assign a high sensitivity detector, set the emission wavelength range, and turn on photon counting mode, if available. Set the time gating window to 0.6 to 10 nanoseconds for each fluorescence channel, to avoid capturing reflected light from the glass dish. Next, turn on bright field imaging, and select low sensitivity detector for the data collection.

Turn on live imaging mode, and turn up the gain in the bright field channel until the cells are clearly visible. If in photon counting mode, change the range of gray values from manual to automatic to view the fluorescent signal. Set the Z-stack to image the entire volume of yeast cells and specify the directionality of the imaging such that down moves toward the cover slip.

Set the Z-step interval to 0.25 to 0.35 micrometers to obtain about 20 to 25 optical sections per Z-stack, and turn on the Galvo Flow if it is available. For a typical movie, set the time interval between Z-stacks to two seconds, and set the movie duration to five to 10 minutes. Then save the movie as a LIF file.

For movie deconvolution, launch an appropriate deconvolution software program that uses the classic maximum likelihood estimation algorithm, and open the data series. Select deconvolution wizard and parameter editor to confirm that the imaging parameters are detected and correctly displayed. Change the embedding medium refractive index's value to 1.4 to approximate the yeast cytoplasm.

Then use the editor to estimate the cover slip position. Select Set all verified and click Accept, and select Enter wizard. Click the next arrow to bypass the point spread function selection and cropping pre-processing stages, and proceed through the deconvolution wizard for each fluorescence channel.

Select the compute logarithmic vertical mapping function and inspect the raw data fluorescence intensity profile. Set manual as the mode for the background estimation, enter a background value, and click Accept. Leave the maximum iterations value at 40 and enter an estimated signal to noise ratio.

Then turn off the bleach correction, and click Deconvolve. Select Accept to next channel in the deconvolution result window, if the noise is sufficiently removed without eliminating the genuine fluorescence from the dim structures. Once all of the fluorescent channels have been satisfactorily deconvolved, click All done and arrange the red channel first, followed by the green, blue, and bright field channels, for subsequent editing in ImageJ.

Then save the image sequence as an eight bit TIF file and select one file per channel and contrast stretch as the conversion mode. For bleach correction, import the deconvolved image sequences into ImageJ and click Image, Hyperstacks, and Stack to Hyperstack to convert the images into a hyperstack. Select xyzct from the dropdown menu and enter the number of channels, Z-stack slices, and timeframes.

To correct the fluorescence channels for photo bleaching, select Image, Color, Split Channels, and for fluorescent channels, select Plugins, EMBLtools, Bleach Correction, and Exponential Fit. Then select Image, Color, and Merge Channels to merge the bright field and bleach corrected fluorescence channels into a hyperstack, and save the deconvolved and bleach corrected hyperstack for subsequent movie generation and editing as an eight bit TIF file. To convert the deconvolved and bleach corrected data set to a scaled montage, select Plugins, IJ_Plugins, and Make Montage Series, and select the relevant eight bit hyperstack.

Click Open and OK to accept that all of the slices will be used to create the montage, and click OK again to accept the suggested scale factor value. Save the original montage as an eight bit TIF file and select Plugins, IJ_Plugins, Montage Series to Hyperstack and OK, to create a 4D hyperstack from the original montage that includes all of the timeframes. To generate the original average projected movie, first select Plugins, IJ_Plugins, Project Hyperstack, and OK, to accept the ZProjection default parameters.

Save the average projected movie as an eight bit TIF file and inspect the movie to identify individual structures that can be tracked for the duration of their labeling period. Identifying a structure that can be reliably tracked for the duration of the movie, and isolating that structure for analysis, are the most critical steps of the procedure. Then follow instructions in the plugin user guide, to isolate structures of interest by editing the montage.

Create a 4D hyperstack from the edited montage for the period including the structures of interest, and save the edited hyperstack. To quantify the fluorescence intensity over time for the isolated structure in the edited 4D hyperstack, select Plugins, IJ_Plugins, and Analyze Edited Movie, input the time interval between Z-stacks, and click OK.To create a final movie of the isolated structure, select Plugins, IJ_Plugins, and Project Hyperstack, specify the Z-stack slices that should be included, select Average Intensity as the Projection type, and click OK.To create a 4D hyperstack with the original data over the edited data, select plugins, IJ_Plugins, Merge Two Hyperstacks, and Place first above second. To generate a movie from the 4D hyperstack with the original data over the edited data, select Plugins, IJ_Plugins, and Project Hyperstack, specify the Z-stack slices that should be included, select Average Intensity as the projection type, and click OK.Here, the first frame of a Z-stack projection of raw data can be compared to the same deconvolved and bleach corrected data.

These frames from the same deconvolved and bleach corrected movie show two cisternae that were analyzed, which first label with the green Vanadate resistance glycosylation protein 4 or Vrg4 marker, and later with the red Secretory 7 or Sec7 gene transport protein marker. This montage, created from a deconvolved and bleach corrected hyperstack, shows all of the optical sections at a single time point before and after editing, to allow isolation of the signal from one of the chosen cisternae. In this figure, several frames from the final movie of the projected Z-stacks are shown with the original projections at the top of the figure and edited projections at the bottom.

Quantification of the green and red fluorescent signals from the chosen Golgi cisternae reveals that the green Vrg4 marker arrives and persists for about 80 seconds, after which, the red Sec7 marker arrives and persists for about 60 seconds, with a brief overlap between the two markers. When performing this procedure, it's important to identify structures that can be unambiguously tracked throughout their lifetimes. The same type of analysis can be performed after making a mutation or some other type of specific perturbation.

This technique has opened the way to studying the dynamic behavior of Golgi cisternae and endoplasmic reticulum exit sites using yeast as a model system.