Live-cell Imaging and Quantitative Analysis of Meiotic Divisions in Caenorhabditis elegans Males.

Live-imaging of meiotic cell division has been performed in extracted spermatocytes of a number of species using phase-contrast microscopy. For the nematode Caenorhabditis elegans, removal of spermatocytes from gonads has damaging effects, as most of the extracted spermatocytes show a high variability in the timing of meiotic divisions or simply arrest during the experiment. Therefore, we developed a live-cell imaging approach for in situ filming of spermatocyte meiosis in whole immobilized C. elegans males, thus allowing an observation of male germ cells within an unperturbed environment. For this, we make use of strains with fluorescently labeled chromosomes and centrosomes. Here we describe how to immobilize male worms for live-imaging. Further, we describe the workflow for the acquisition and processing of data to obtain quantitative information about the dynamics of chromosome segregation in spermatocyte meiosis I and II. In addition, our newly developed approach allows us to re-orient filmed spindles in silico, regardless of the initial 3D orientation in the worm, and analyze spindle dynamics in living worms in a statistically robust manner. Our live-imaging approach is also applicable to C. elegans hermaphrodites and should be expandable to other fluorescently labelled nematodes or other fully transparent small model organisms.

for the culture media (i.e., pH, osmolarity and temperature).
In the nematode Caenorhabditis elegans, microscopic investigation of spermatocyte meiosis was previously achieved by labelling of isolated and fixed gonads with antibodies (Howe et al., 2001;Shakes et al., 2009;Peters et al., 2010;Schvarzstein et al., 2013). This method, while powerful for analyzing the molecular composition of male meiotic spindles, cannot be used to study spindle dynamics.
Alternatively, male gonads can be dissected, and the extracted germ cells can be filmed using Normarski optics. Applying this approach ( Figures 1A-1D), it was possible to film cell divisions but the components of the meiotic spindle such as chromosomes and spindle poles were impossible to detect reliably without appropriate fluorescent labels (Shakes et al., 2009;Gleason et al., 2012;Vielle et al., 2016;Hu et al., 2019). Moreover, the imaging time was limited to only a few minutes as the cells regularly arrested before completion of the meiotic divisions. In our hands, about 90% of the spermatocytes arrested prior to the end of the meiotic divisions due to unknown reasons.
We developed a live-imaging approach that enabled us to study spindle dynamics in 3D in living immobilized males ( Figures 1E-1G). With this newly developed approach it is now possible to achieve a parallel imaging of multiple spindles over about 90 min without any damaging effects on spindle dynamics (Fang-Yen et al., 2009; Kim et al., 2013). Another method applies patterned agarose pads created by using Vinyl Records as a mold (Rivera Gomez and Schvarzstein, 2018). However, this method relies on the use of anesthetics to completely immobilize the worm for imaging. Here we intended to avoid the use of anesthetics for animal immobilization, thus excluding any possible side effects of drugs on the dynamics of cell division. A technical problem of this approach that had to be solved was the bleaching of the fluorescent markers during the course of imaging, i.e., a new sample had to be prepared about 1.5 h after the recording started. However, the high number of dividing cells within a single animal enables to collect sufficient data to determine the characteristics of dividing spindles in spermatocyte meiosis.
The challenge was to calculate the properties of dividing spindles over time. Due to the 3D live imaging spindles were not perfectly aligned in the imaging plane but rather randomly oriented in 3D with varying spindle axis angles in each frame. We solved this problem by computationally rotating each spindle in 3D to have a standard orientation for data analysis and then calculating kymographs for each fluorescent channel. By analyzing the kymographs, we could determine the dynamics of the chromosomes and the spindle poles over time with high accuracy. Further, this workflow can be used to image dividing mitotic germ cells in the distal gonad of males and hermaphrodites, which is also a cell type very difficult to image in situ. Our pipeline of imaging acquisition and data processing of cell division in C. elegans males can also be applied for imaging dividing embryos in hermaphrodites of this species and should also be applicable to other small model organisms with fluorescently labeled chromosomes and centrosomes. 3 www.bio-protocol.org/e3785  In this field-of-view, different regions of a gonad are shown. These regions include meiotic prophase, various meiotic divisions of in meiosis I and II, and mature sperm. The inset (red frame) shows an anaphase I spindle. G. Projection of 21 z-slices, 61 min after starting the image acquisition. The inset (red frame) shows two spindles in metaphase II, Scale bar for (E-G), 5 µm. 7. Transfer as many males as possible to a fresh NGM-agar plate and place 5 L4 hermaphrodites on this plate. Aim to have for 4-5 times more male individuals than hermaphrodites per plate.

Materials and Reagents
The progeny of the hermaphrodites should produce ~50% males. Caution: Melted agarose is hot; protective gloves should be worn. 4. Use a small spoon or spatula to place a drop of the melted 10% agarose on a glass slide ( Figure   2B). 5. Immediately add a second glass slide on top and gently apply pressure for 5 s ( Figure 2C). 6. Repeat this procedure until all glass slides have agarose in between. 6 www.bio-protocol.org/e3785 8. Use a round object (e.g., cap of a permanent marker) to cut out agarose pads of a defined diameter of 10 mm. For this, select areas that do not contain any or only a few trapped air bubbles. Such bubbles might impair worm immobilization or the image quality later. Multiple agarose pads can be cut out of a single agar preparation ( Figure 2D). 9. Grab each agarose pad with fine tweezers and put it to a small container (e.g., empty plate) filled with M9 buffer ( Figure 2E).  2. Aspirate excess of liquid from the side of the agarose pad with a piece of filter paper ( Figure   2F). Be careful not to touch the surface of the agarose pad as also cellulose fibers on the surface of the pad will reduce the immobilization of worms.
3. Add 1 µl of a polystyrene bead solution on the agarose pad ( Figure 2G). 4. Place 5-10 three-day old male worms within the bead solution.
5. Gently add a cover slip ( Figure 2H). It is important to lower the covers slip gently as a dropping of the glass might damage the worms or activate spermatids within the males. A sliding of the mounted cover glass should also be avoided as this might impair the worms as well. If worms are still moving after the assembly procedure, discard the sample and start again with point no.

1.
6. Add a few microliters of M9 buffer from the side under the cover slip. Be careful not to add too much as excess liquid will make the cover slip float and decrease the amount of immobilization. 7. Seal the edges of the cover slip with melted candle wax ( Figure 2I). 8. Check again under a dissection microscope if the worms are immobile. If worms are perfectly immobilized at this point, they will not become mobile again, thus long-term imaging will be possible by using the settings described below. 9. Place the prepared sample on a fluorescence microscope. E. Image pre-processing 1. Import the Z-stacks to ImageJ/Fiji. A user manual for ImageJ/Fiji can be found here.

Convert them to a hyperstack.
3. Correct for bleaching using the method "Exponential fit". 4. Save images.

Crop out and export individual meiotic cells from the meiotic region of the male gonad (
3. Make a note in which frame anaphase starts. 4. Use a 3D median filter with a kernel size of 1 pixel ( Figure 3C) to reduce the noise. 5. Segment individual centrosomes in each meiotic cell with the tool "Blob finder" ( Figure 3D). The settings have to be chosen in accordance to the used imaging setup. As the image quality varied among the live recordings the segmentation settings had to be adjusted within a narrow range.
a. Give the path to both CSV-files containing the coordinates of the center of the centrosomes (PATH).
b. Indicate which fluorescence channel should be processed (CHANNEL).
c. Indicate the frame of the data set that should be used as a start (START_FRAME; typically, this was ten time points before the frame of anaphase onset).
d. Indicate the pixel size in the resampled output data set (LINE_PIXEL_SIZE; typically 0.1 µm was used).
e. Indicate the radius around the spindle axis that is used to calculate the fluorescence pixel values within the kymograph (LINE_RADIUS; typically a '9' was used, which means a radius of 0.9 µm around the spindle axis, given that the pixel size was set to 0.1 µm). 9 www.bio-protocol.org/e3785 f. Indicate the length of extrapolation along the spindle axis after the start and end coordinate (LINE_EXTRAPOLATE, typically 1.0 µm before and after the centrosome centers was added to each kymograph).
g. Define an output filename for the kymograph CSV-file (OUTPUT_FILENAME).
12. Calculate the Gaussian weighted sum of fluorescence (Burger and Burge, 2013) in each plane in 0.1 µm steps along the spindle axis with a radius of 0.9 µm ( Figure 3G). This was repeated for all time points ( Figure 3H) and the kymograph was saved as an image ( Figure 4A) and a csv-file. 13. Analyze the kymograph for each fluorescence channel to determine the position of the peak intensity from both sides of the spindle over time (Figures 4B-4C).
14. Calculate the distance of the peak intensities for each channel over time. This way, the autosome-to-autosome distance ( Figure 4D) and the pole-to-pole distance ( Figure 4E) can be calculated. In addition, the pole-to-autosome distance ( Figure 4F) can also be determined by calculating the distance of the peak intensities from both fluorescence channels on each side of the spindle. profiles are shown (right). D. Analysis of the autosome-to-autosome distance over time for 31 data sets aligned at anaphase onset (t0). Circles represent the mean, shaded area the standard deviation. All mean values are connected by a solid line. E. Analysis of the pole-to-pole distance over time for 31 data sets. F. Analysis of the pole-to-autosome distance over time for 31 data sets.

Data analysis
For each spindle, the pole-to-pole, the autosome-to-autosome and the pole-to-autosome distance was calculated for each time point from the kymograph data. In addition, the time of anaphase onset was determined as time zero (t0) by the first frame showing a clear separation of the autosomes.
Accordingly, spindle dynamics in all data sets were then aligned to t0. By calculating the average distance (± standard deviation) for each time point the dynamic properties of both meiotic spindles were calculated, i.e., the initial and final spindle length, the initial rate of elongation and the duration of spindle elongation or segregation. The solution was filtered afterwards (filter pore size 0.2 μm)