Live Cell Imaging of Male Meiosis in Arabidopsis by a Landmark-based System.

Live cell imaging has tremendously promoted our understanding of cellular and subcellular processes such as cell division. Here, we present a step-by-step protocol for a robust and easy-to-use live cell imaging approach to study male meiosis in the plant Arabidopsis thaliana as recently established. Our method relies on the concomitant analysis of two reporter genes that highlight chromosome configurations and microtubule dynamics. In combination, these reporter genes allowed the discrimination of five cellular parameters: cell shape, microtubule array, nucleus position, nucleolus position, and chromatin condensation. These parameters can adopt different states, e.g., the nucleus position can be central or lateral. Analyzing how tightly these states are associated gives rise to landmark stages that in turn allow a quantitative and qualitative dissection of meiotic progression. We envision that such an approach can also provide valuable criteria for the analysis of cell differentiation processes outside of meiosis.

However, these studies provide little to no information about the underlying dynamics of meiosis. In addition, temporal aspects of meiosis have to be indirectly deduced by the frequency of observed stages, a procedure that is inherently error-prone and can easily misguide the researcher e.g., when two or more populations of meiocytes exist in the same sample undergoing an altered course of meiosis that can be mistaken as one population with cells at different stages Sofroni et al., unpublished). Moreover, short-lived phases, for instance nuclear envelope breakdown, are difficult to catch, and there has been for instance a long discussion in the field whether the nuclear envelop is reformed in Arabidopsis after the first meiotic division.
Live cell imaging of meiosis can complement the analyses of fixed material and build together with these techniques a powerful approach to reach molecular mechanistic insights into this important cell division program. Live cell-imaging of plant meiosis has been previously initiated in maize (Yu et al., 1997;Sheehan and Pawlowski, 2009;Nannas et al., 2016). However, only short phases of meiosis in maize could be recorded with a genetic reporter for microtubules and a chemical stain that highlights DNA. In addition, a recent protocol for live cell imaging of meiosis by light-sheet microscopy has been published (Valuchova et al., 2020). While this set-up is very powerful to follow meiosis in entire flower buds, it still does not reach the subcellular resolution as obtained by confocal laser scanning microscopy (CLSM).
Here, we describe in detail a method to follow meiosis in Arabidopsis anthers based on a recently established procedure by CLSM . Importantly, this method allows keeping the samples alive up to several days allowing the analysis of meiosis in its entirety. Furthermore, the use of A major challenge is the quantitative analysis of the obtained movies. As typical for biological processes, meiosis is a continuous and gradual succession of events. To dissect these movies, we focused on five cellular parameters as visualized by the KINGBIRD  Each of these parameters can adapt different states, see Table 1 (Table 1 Excel  Looking then at the association of these different parameter states revealed that they are not randomly associated but often tightly linked. This gives rise to a biological landmark system where one landmark is a prominent cellular configuration with distinct parameter states. In turn, this landmark system can be used as a map to qualitatively (appearance of the same or new landmarks) and quantitatively (duration of these landmarks) dissect meiotic progression in mutants or different environmental conditions. The use of other meiotic reporters can then be used to refine and/or complement this landmark system. The principle of this analysis can be easily translated to other cellular differentiation processes, of course including other cellular parameters, which need to be identified. 5. Lift the inflorescence from the preparation medium and cleanly cut the stem to a length of circa (ca.) 0.5 cm (with forceps or better using a needle) (Video 4). This will facilitate the uptake of nutrients from the medium and will keep the sample in good condition for a long time.  Therefore, it is advisable to turn on the confocal microscope in advance and position the sample in the room with a few ml of water covering the mounted samples.

Materials and Reagents
2. After approximately one hour you can start the image acquisition.
3. Position the sample on the microscope stage, submerge the water-dipping objective and fill up the Petri dish to the top with autoclaved water (Figures 1C and 1D). It is advisable to thoroughly clean the objective with isopropanol before the acquisition to reduce the risk of bacteria growth during image acquisition. c. Set the Pinhole at 1 Airy Unity for the TagRFP detection, use bidirectional scan function and set the pixel dwell to a value around 2 μs. d. Parameters for image acquisition such as laser intensity, gain and offset have to be adapted to the individual imaging conditions. Among others, they depend on the laser status, as well as on the reporter line used, and thus the level of protein expression. In general, a compromise between sample viability and high-resolution imaging has to be found to obtain the best image quality while maintaining the sample in good conditions. In our case, we set the intensity of the Argon laser between 1% and 4.5%, while the DPSS 561-10 laser was set between 0.3% and 1%. The detector gain to collect GFP and TaqRFP signals was set between 700 and 850, and was set between 650 and 750 for the detection of chloroplast autofluorescence. In all the cases, the offset parameter was 0.

5.
If an x-y drift is detectable, it can be corrected by using the plug-in Stackreg from Fiji (Plugins  Stackreg  Rigid body) (Thévenaz et al., 1998).
6. An example of a time-lapse series after image processing is given in Video 5.  REC8-chromatin is in green and chloroplasts are blue). Each meiocyte is given a number, which will be its ID on the spreadsheet in B. B. Example of a spreadsheet used for analysis as described in the paragraph Data-analysis A2. The numbering of the stage refers to the numbers in Table 1. Each stage is color-coded to facilitate the data organization.
2. Prepare a table using Microsoft Excel that includes 7 columns: frame number, time in min, cell shape, microtubule array, nucleus position, nucleolus position and chromatin state (Figure 2).
The number of rows will be equal to the number of frames acquired for a single cell. The first imaged frame will be counted as Time 0 (Figure 2).
3. Fill in each column with a number that corresponds to the state of the analyzed cellular parameter as annotated in Table 1 and Figure 2

B. Data set analysis
A Python script to perform analysis of the annotated data set created in part A. is available at https://gitlab.com/wurssb/arabidopsis-thaliana---landmark-analysis. This script contains the procedure to extract landmarks from the data and takes the following steps: 1. Prepare the data for analysis by filtering out invalid or duplicate entries. Each entry, consisting of the observed state for each cellular parameter should have an identifier for the sample and the specific cell within the sample.
2. If the data is not sampled at regular intervals, resample to a common interval. However, take care when resampling, as going from short to large intervals might remove short-lived states from the data set while the reverse could introduce (large) errors from interpolating the states. 4. Partially observed states can be either discarded, or interpolated from neighboring states.
Again, take care when interpolating, for example, by setting a maximum time interval that can be safely interpolated depending on the observed processes.  Table 2 (Table 2 Excel