Here it is shown how to track and quantify developmental processes in C. elegans. The methods presented are based on open-source tools that can be easily implemented. It is demonstrated how to reconstruct 3D cell-shape models, how to manually track subcellular structures, and how to analyze cortical contractile flow.
Quantitatively capturing developmental processes is crucial to derive mechanistic models and key to identify and describe mutant phenotypes. Here protocols are presented for preparing embryos and adult C. elegans animals for short- and long-term time-lapse microscopy and methods for tracking and quantification of developmental processes. The methods presented are all based on C. elegans strains available from the Caenorhabditis Genetics Center and on open-source software that can be easily implemented in any laboratory independently of the microscopy system used. A reconstruction of a 3D cell-shape model using the modelling software IMOD, manual tracking of fluorescently-labeled subcellular structures using the multi-purpose image analysis program Endrov, and an analysis of cortical contractile flow using PIVlab (Time-Resolved Digital Particle Image Velocimetry Tool for MATLAB) are shown. It is discussed how these methods can also be deployed to quantitatively capture other developmental processes in different models, e.g., cell tracking and lineage tracing, tracking of vesicle flow.
With the steady improvements of fluorescent proteins, genome engineering, light microscopy, and computer soft- and hardware, it is now possible to record development of many model organisms at unprecedented spatio-temporal resolution. This allows researchers to ask questions that could not be addressed previously or to revisit known developmental processes in order to search for overlooked aspects. This progress has sparked the field of quantitative developmental biology, which aims at transforming qualitative, informal models into quantitative models by thorough measurements and statistical analyses.
Tracking cells and subcellular structures has made it possible to derive quantitative models of embryonic development, nervous system activity, or cell division1-12. By tracking cell division remnants during early development of the C. elegans embryo, we could recently reveal that they follow a stereotyped path and constitute important polarizing factors13,14.
Here, protocols are presented that make quantitative developmental biology approaches accessible for non-experts. The focus lies on three straight-forward, freely available tools that are implementable in any lab that has access to standard confocal microscopy and computers. These include a protocol to generate 3D cell shapes, a protocol to track cell division remnants, and a protocol to quantitatively describe cortical actomyosin dynamics. The nematode C. elegans is used as an exemplary case, however, the methods and tools discussed here are suited for a variety of questions in other biological models, e.g., cultured cells, tissue explants, organoids or spheroids, other embryos, etc.
Generally, some of the analyses shown here can also be performed with the popular open source tool ImageJ (http://imagej.nih.gov/ij/docs/index.html; or FIJI, the ‘batteries included’ version of ImageJ, http://fiji.sc/Fiji) for which many plugins for different quantitative analyses are available. However, the programs discussed here are designed to tackle specific problems.
Firstly, IMOD, an image processing, modeling and display suite can be used for 3D reconstructions of serial sections from electron or light microscopy15. IMOD also contains tools for viewing the 3D data from any orientation. Secondly, Endrov, a Java program designed to perform image analysis, data processing, and annotation of networks or tracks (among others) on the basis of an extended plugin architecture, with ImageJ plugin compatibility16. It contains over 140 image-processing operations and an extensible user interface in which model and raw data are displayed separately. Its source code can be found at https://github.com/mahogny/Endrov. Thirdly, PIVlab, a MATLAB tool for digital particle image velocimetry that allows the user to quantitatively and qualitatively analyze particle flow fields17. The use of this programs requires a MATLAB license that includes the Image Processing Toolbox (http://mathworks.com). PIVlab is a program designed to quantitatively describe flow. It calculates the velocity distribution, magnitude, vorticity, divergence, or shear within image pairs or series. For this, it cross-correlates small areas of images (called ‘passes’ in the protocol section) of an image pair to derive the most probable particle displacement. This cross-correlation yields a correlation matrix that can be analyzed in the space or frequency domain using either direct cross-correlation or a fast Fourier transformation (FFT), respectively.
The equipment used here is an inverted microscope equipped with a Nipkow ('spinning') disc, an EMCCD, 488 and 561 nm standard solid-state lasers, and 20x air or 40x or 60x plan/apochromat oil- or water-immersion objectives. However, it is also possible to perform time-lapse imaging with other imaging modalities, e.g., point-, line- or sheet-scanning laser-based microscopy, multi-photon microscopy, as well as epi-fluorescence microscopy combined with deconvolution or structured illumination. The advantage of using a Nipkow disc system is the extremely fast image acquisition, especially if a streaming mode (continuous movement and scanning of the object in the z dimension) is available. In addition, to improve resolution, a 1.5-fold magnifying extender in front of the EMCCD can be used.
1. Preparation of C. elegans Embryos for Time-lapse Microscopy using Microbeads
2. Preparation of Adult C. elegans Animals for Time-lapse Microscopy using Nanobeads
NOTE: In general, the protocol for mounting adult animals is an adaptation of the protocol from ref. 18. With this protocol, it is possible to perform long-term time-lapse confocal microscopy of adult animals.
3. Time-lapse microscopy
4. Reconstructing a 3D Cell Shape model with IMOD
NOTE: IMOD software is available from the IMOD web page: http://bio3d.colorado.edu/imod/. Installation of the software is described there. When using Windows OS, a Unix toolkit has to be installed, which can also be found as a package on the IMOD webpage. Moreover, there is an excellently compiled 3dmod introduction available on the IMOD webpage (http://bio3d.colorado.edu/imod/doc/3dmodguide.html).
5. Tracking Fluorescently Labeled Structures with Endrov
NOTE: In order to track cell division remnants, use a strain with a nuclear or plasma membrane marker for cell identification (e.g., H2B, PH-PLC-d1) and a cell division remnant marker (e.g., NMY-2, ZEN-4). The nuclear or plasma membrane marker is important to determine which cell inherits the cell division remnant.
6. Analyzing Cortical Actomyosin Flow with PIVlab
NOTE: PIVlab is a freely available software from: http://pivlab.blogspot.de/; it can be invoked from within the Matlab command line environment by typing PIVlab_GUI. When working with PIVlab, it is recommended to save the image series in a separate folder.
By using protocols 2, 3, and 4, time-lapse imaging of gonads in wild type C. elegans adults is performed (strain OD58 (unc-119(ed3) III; ltIs38[pAA1; pie-1::GFP::PH(PLC1delta1) + unc-119(+)]), expressing a membrane marker from a germline promoter). Focusing on the turn of the gonad, a 3D model of the germ cells is generated from the microscopy data (Figure 2). This model allows to analyze changes in cell size while the cells transit form the distal to the proximal arm, reveals the organization of the rachis and the size of the contacts of individual cells to the rachis (see also http://www.wormatlas.org/hermaphrodite/germ%20line/Germframeset.html).
Next, protocols 1, 3, and 5 are used to perform long-term time-lapse microscopy of C. elegans embryos (strain CHP52 which is a cross of strain RW10226 (unc-119(ed3) III; itIs37[pie-1p::mCherry::H2B::pie-1 3'UTR + unc-119(+)]; stIs10226[his-72 promoter HIS-24::mCherry translational fusion with let-858 3' UTR + unc-119(+)]) and strain LP162 (nmy-2(cp13[nmy-2::gfp + LoxP]) I.). In this strain, it is possible to follow both nuclei (through the mCherry-histone fusion proteins) and cell division remnants (through the genomically modified non-muscle myosin II locus where a GFP has been integrated in frame through the CRISPR/Cas9 technology19) simultaneously when using two-color time-lapse microscopy (Figure 3, left panels). The models obtained by lineaging of both structures in Endrov show the previously described stereotyped pattern of cell division remnant inheritance13,14. Moreover, from the lineaging data, the tracks for each cell and cell division remnant (Figure 3, right panels) and the time of cell division remnant persistence (based on the NMY-2GFP-signal) as well as the correlation to the cell division timing can be obtained (Figure 4, shown for the ABa lineage only). When using a plasma membrane marker in addition, it is also possible to observe the time point of cell division remnant internalization.
Finally, by using protocols 1, 3, and 6, short-term time-lapse microscopy with high temporal resolution (5s intervals between recording a z-stack) of cortical non-muscle myosin II (NMY-2GFP) is performed. The focus here lies on the differences in the dynamics of cortical polarizing flow by comparing this flow between wild type and embryos RNAi-depleted for the Rho GTPase activating protein RGA-320,21. Similar to recent observations22, flow in wild type embryos is predominantly along the long axis of the embryo (Figure 5, left panels), while flow in the rga-3 RNAi embryo is orthogonal to this axis. This is readily apparent from overlaying consecutive time points or from the PIV analysis (Figure 5, bottom panels).
To observe these differences, it is important to choose a sampling interval so that cortical flow particles can be resolved unambiguously for accurate interpretation of data. Smaller time intervals (≤5 sec) are recommended to avoid large displacements and missing vectors. The interrogation window should be large enough to accommodate the size of the particles to be analyzed (cortical granules). Overlap between neighboring interrogation windows allows to reduce the vector spacing and thus increases the number of vectors in the grid.
Figure 1. Assembling a mouth pipette. Left: Parts required to assemble a mouth pipette with filter. Right: The assembled pipette. Please click here to view a larger version of this figure.
Figure 2. A 3D model of the gonad turn region. Top left: 3D projection of a single time point from time-lapse microscopy data. Top middle and right: 3D models of the microscopy data. Bottom: Details of cell-rachis contacts. Please click here to view a larger version of this figure.
Figure 3. Tracking of nuclei and cell division remnants. Left: 3D projections from time-lapse microscopy data. Middle: Snapshots of the model obtained from tracking both nuclei (colored spheres) and cell division remnants (grey spheres). Right: Snapshots from cell division remnant tracking. The paths for the remnants are also shown. Please click here to view a larger version of this figure.
Figure 4. Cell and cell division remnant lineages for the ABa sublineage. Lineages from a representative embryo are shown. Cell divisions are marked by squares. Each tick mark corresponds to 1 min. The cell division remnants were named after the daughters that arise from the division during which the remnant is generated. Please click here to view a larger version of this figure.
Figure 5. Analysis of cortical contractile flow with PIV. Top: 3D projection stills from time-lapse microscopy that depict the start (first row) and end (second row) of the time window used to calculate the vector field with PIVlab. The third row of panels shows the overlay of all time points of the time window. Bottom: Vector fields obtained by PIVlab. Please click here to view a larger version of this figure.
Video 1 – related to Figure 2. A 3D model of the gonad turn region. The video starts with scrolling through the entire stack of planes used to segment the cells. Afterwards a 3D maximum projection of the stack is shown, followed by the 3D segementation model rotating around the x and y axes (with and without the segmented gonad's rachis).
Video 2 – related to Figure 3. Tracking of nuclei and cell division remnants. Left: 3D projections from time-lapse microscopy data. Middle: The model obtained from tracking both nuclei (colored spheres) and cell division remnants (grey spheres). Right: Cell division remnants and their trajectories. Scale bar = 10 µm.
Video 3 – related to Figure 5. Analysis of cortical contractile flow with PIV. 3D maximum projection time-lapse recordings of a representative wild type and rga-3 RNAi embryo used to calculate the vector field with PIVlab. Scale bar = 10 µm.
Through object tracking in development, in particular nuclear tracking, it has been possible to elucidate central patterning mechanisms of C. elegans embryogenesis1,23,24. Expanding this strategy to higher throughput, it has been recently possible to uncover additional patterning rules and to propose a method how to deduce patterning rules de novo10. For many mutants, however, the precise patterning defects are still unknown. The methods described here are tools that can be instrumental in their elucidation. Importantly, it has become clear in recent years that although many other model systems do not follow an invariant development like C. elegans, object tracking and analysis of the quantitative tracking data is a crucial tool to identify developmental mechanisms and mechanisms of disease.
The methods shown here are especially well suited in the situation where it would either be too challenging to implement more complex and/or self-generated software solutions, or simply too costly to implement commercial software tools, respectively. Importantly, the tools discussed here enable the researcher to load and analyze any data independently on which instrument they have been recorded. Moreover, if proprietary image acquisition software is used, ImageJ provides the Bio-Formats plugin that can load data in proprietary formats and save them in standard formats (jpeg, tiff).
A critical step in all protocols that rely on time lapse microscopy is to choose the conditions right to obtain sufficient contrast in microscopy images, and, at the same time, to reduce radiation exposure to the specimen. This heavily depends on the type of microscope used and it is generally recommended to use fast scanning techniques such as spinning disc or single plane illumination microscopy. Specifically, for PIV of cortical dynamics, older versions of point-scanning microscopes might be too slow to allow capturing cortical signals at a sufficient resolution and speed to perform valid analyses. It is therefore always crucial to carefully evaluate the radiation exposure and test the viability of the specimen. In addition, although the protocols discussed here use micobeads to mount embryos, it is also possible to use agarose pads, which just takes a little more time to prepare but is less expensive and with some experience equally reproducible. However, it should be pointed out that the combination of agarose pad and nanobeads is the superior method for long-term immobilization of adult worms since it avoids the use of any paralyzing agent, e.g., levamisole.
It should be pointed out that the tools discussed here have their limitations since they do not perform fully automated data analysis. However, for structures like cell division remnants, it is often challenging to identify a molecular marker that only stains this structure, which is crucial for reliable tracking. A major future challenge therefore lies in including features like machine learning and adaptable detection/segmentation algorithms in the easy-to-use and open access software that we discuss here. Some of these features can be found in other software modules, in particular software that can deal with repeated tracking over longer time intervals, or that can compare large sets of tracking data (e.g., StarryNite/AceTree25 or Nucleitracker4D26, which are designed to track nuclei using histone-fusion proteins or spindles27, or Simi BioCell28). These tools may be more complex to implement when aiming at tracking of structures other than nuclei. Notwithstanding the particular structure that should be tracked, these programs are generally very helpful when large amounts of image data will be analyzed. Importantly, there is compatibility between the above mentioned software modules.
When working with other model systems (e.g., fly embryos, mammalian cultured cells, etc.), the methods presented here are also well suited to track orthologous structures as described above. More importantly, independently of the model system under investigation, different subcellular structures, e.g., centrosomes, nucleoli, microparticles, vesicles, etc. can be readily tracked, only the acquisition parameters have to be adjusted to ensure complete capturing of object tracks.
The authors have nothing to disclose.
The authors have nothing to disclose.
Stereo microscope | Motic/VWR | OT4005S | Stereo microscope for dissection and mounting |
Polybead Polystyrene Microspheres, | Polysciences | 18329 | Embryo mounting |
20 µm | |||
Polybead Polystyrene Microspheres, | Polysciences | 876 | Adult animal mounting |
0.1 µm | |||
Microscope slides | VWR | 631-0902 | Adult animal mounting |
Cover glass 18×18 mm | VWR | 631-1331 | Embryo/adult mounting |
Cover glass 24×60 mm | VWR | 631-1339 | Embryo mounting |
Scalpel | VWR | 233-5455 | Embryo dissection |
Silicone tubing | VWR | 228-1501 | Tubing for mouth pipette |
30 mm PTFE membrane filter | NeoLab | Jul-01 | Filter for mouth pipette |
Capillary tubes | VWR | 621-0003 | Pipette tip for mouth pipette |
Vaseline | Roth | E746.1 | Embryo/adult mounting |
Agar | Roth | 5210.5 | Adult animal mounting |
Potassium-di-hydrogenphosphate | Roth | P018.2 | M9 buffer |
Di-sodium- hydrogenphosphate | Roth | P030.2 | M9 buffer |
Sodium chloride | Roth | 3957.1 | M9 buffer |
VisiScope Spinning Disc Confocal System | Visitron Systems | n/a | Confocal microscopy |