This protocol describes a method for the photoconversion of Kaede fluorescent protein in endocardial cells of the living zebrafish embryo that enables the tracking of endocardial cells during atrioventricular canal and atrioventricular heart valve development.
During embryogenesis, cells undergo dynamic changes in cell behavior, and deciphering the cellular logic behind these changes is a fundamental goal in the field of developmental biology. The discovery and development of photoconvertible proteins have greatly aided our understanding of these dynamic changes by providing a method to optically highlight cells and tissues. However, while photoconversion, time-lapse microscopy, and subsequent image analysis have proven to be very successful in uncovering cellular dynamics in organs such as the brain or the eye, this approach is generally not used in the developing heart due to challenges posed by the rapid movement of the heart during the cardiac cycle. This protocol consists of two parts. The first part describes a method for photoconverting and subsequently tracking endocardial cells (EdCs) during zebrafish atrioventricular canal (AVC) and atrioventricular heart valve development. The method involves temporally stopping the heart with a drug in order for accurate photoconversion to take place. Hearts are allowed to resume beating upon removal of the drug and embryonic development continues normally until the heart is stopped again for high-resolution imaging of photoconverted EdCs at a later developmental time point. The second part of the protocol describes an image analysis method to quantify the length of a photoconverted or non-photoconverted region in the AVC in young embryos by mapping the fluorescent signal from the three-dimensional structure onto a two-dimensional map. Together, the two parts of the protocol allows one to examine the origin and behavior of cells that make up the zebrafish AVC and atrioventricular heart valve, and can potentially be applied for studying mutants, morphants, or embryos that have been treated with reagents that disrupt AVC and/or valve development.
The zebrafish is currently one of the most important vertebrate models to study cellular and developmental processes in vivo. This is largely due to the zebrafish’s optical transparency and amenability to genetics, which makes it a powerful model for applying optical techniques involving genetically encoded photoresponsive protein technologies1. Specific to the study of heart development, zebrafish receive sufficient oxygen via diffusion such that even mutants without heartbeat can survive through the first week of development, permitting analyses on the effects of developmental genes and perturbed blood flow on heart morphogenesis not possible in most vertebrates2.
The zebrafish heart tube is formed by 24 hours post fertilization (hpf). Shortly after its formation, the heart tube starts actively beating. By 36 hpf, a clear constriction separates the atrium from the ventricle. This region of constriction is called the atrioventricular canal (AVC), and cells in this region change from a squamous morphology to a cuboidal morphology3. Zebrafish atrioventricular valve morphogenesis starts around 48 h post fertilization. By 5 days post fertilization, two valve leaflets extend into the AVC and prevent the back flow of blood from the ventricle to the atrium during the cardiac cycle4. Tracking cells during AVC and valve formation is challenging as the rapid beating of the heart makes it difficult to follow cells via traditional time-lapse microscopy5,6. This protocol, adapted from Steed et al., 20167, describes a method that uses the Tg(fli1a:gal4FFubs, UAS:kaede)8 zebrafish transgenic line, in which the photoconvertible protein Kaede is expressed in endothelial cells, including the endocardium. The drug 2,3-butanedione-2-monoxime (BDM) is used to temporarily stop the heart beating, allowing accurate photoconversion of EdCs between 36 and 55 hpf, and high-resolution imaging of photoconverted EdCs at specific developmental time points. It has previously been shown that EdCs photoconverted using this method can remain distinguishable from their unconverted neighbors for five days or more after the time of photoconversion7. This protocol also details a method used for image analysis of photoconverted EdCs in embryos younger than 48 hpf, which has successfully been used to follow tissue movements during AVC development (Boselli et al., in press)9. We hope that readers would find this protocol useful for studying AVC and valve development in normal embryos, and in mutants, morphants, or drug-treated embryos. For a more general protocol relating to cell tracking using photoconvertible proteins during zebrafish development, please view the article by Lombardo et al., 201210.
1. Preparing molds and mounting agarose
2. Obtaining Embryos for the Photoconversion
3. Embedding
4. Photoconversion
5. Unmounting photoconverted embryos
6. Imaging photoconverted cells at later embryonic stages
7. Image analysis for embryos under 48 hpf
Note: Tissue movements of the AVC can be difficult to analyze due to its complex three-dimensional structure. To quantify the length of a photoconverted region in the AVC between two non-photoconverted regions or the length of a non-photoconverted region in the AVC between two photoconverted regions, it can be desirable to map the three-dimensional structure onto a two-dimensional map. Images of photoconverted embryos obtained prior to 48 hpf can be analyzed using the following method.
An example of an embryo photoconverted at 48 hpf and imaged again at 80 hpf is shown in Figure 2, Movies 1 and 2. Exposing Kaede to 405 nm light irreversibly converts from the protein from its fluorescent green form to fluorescent red form, enabling the behavior of cells labeled with either the green or their red form to be followed with respect to their differentially colored neighbors during valve formation. It can be seen that the few photoconverted cells of the 48 hpf AVC later proliferated and later resided predominantly on one side of the valve leaflet.
An example of the results that can be achieved by applying the image analysis method described on a 36 hpf embryo after photoconverting EdCs in the atrium and in the ventricle is shown in Figure 3. The segmentation of the AVC allows the parametrization of the three-dimensional structure that is then projected in two-dimensions for easier visualization and analysis. The method allows one to quantify the distance between the two photoconverted regions. By repeating this process at different developmental times it is possible to study the migration of cells toward the AVC (Boselli et al., in press)9.
Figure 1: Mold used to create agarose wells. A) Front view. B) Back view. C-D) Side views. E-F) Oblique views. Dimensions are shown in millimeters. The .STL file for the mold template is provided in the Supplementary Materials. Please click here to view a larger version of this figure.
Figure 2: Representative results of a Tg(fli1a:gal4FFubs, UAS:kaede) photoconverted embryonic heart. A) The embryonic heart just after photoconversion at 48 hpf. Cytoplasmic Kaede within several cells of the superior side of the AVC has been converted from its green form to its red form. B) The same embryo at 80 hpf. It can be seen that the previously photoconverted cells have proliferated and have come to reside in the inner side of the superior atrioventricular valve. Scale bars: 20 µm. Please click here to view a larger version of this figure.
Figure 3: Representative results of the image analysis process on a 36 hpf embryo. A) The result of the three-dimensional segmentation of the AVC in a reference system that is common among the embryos; the signal from the photoconverted (magenta) and non-photoconverted (green) Kaede is projected onto its surface and marks the photoconverted and non-photoconverted regions. B) The AVC is unfolded according to the parametric position of its point and the signal is projected in two-dimensions for easier visualization. The edges of the non-photoconverted region are labeled in white. C) The length of the non-photoconverted region is plotted as a function of the angular position. Please click here to view a larger version of this figure.
Movie 1: The embryonic heart just after photoconversion at 48 hpf. Related to Figure 2A. The movie goes through the acquired z stack of the 48 hpf embryo shown in Figure 2A z slice by z slice. Only one z plane was chosen during the photoconversion process, but it can be seen here that cells at various depths have been photoconverted. Z slices are spaced 2 µm apart. Right click to download this movie.
Movie 2: The photoconverted embryonic heart at 80 hpf. Related to Figure 2B. The movie goes through the acquired z stack of the 80 hpf embryo shown in Figure 2B z slice by z slice. Z slices are spaced 2 µm apart. Right click to download this movie.
Supplementary file: Related to Figure 1. The .STL file for the mold template. Please click here to download this file.
Timing of photoconversion: Although Kaede remains brightly expressed in EdCs even at 96 hpf, it should be noted that as the embryo grows, laser light diffuses more before it reaches the AVC, making confined photoconversion of Kaede more difficult. At embryonic stages later than 55 hpf, the ballooning of the ventricle and the atrium also means that the violet laser beam used for photoconversion cannot reach AVC cells without first passing through either the atrium or ventricle. This means that beyond 55 hpf, in order to photoconvert EdCs of the AVC, one must also photoconvert additional EdCs in the atrium and ventricle.
BDM and tricaine: Accurate photoconversion of EdCs requires temporarily stopping the heart from beating. BDM (2,3-butanedione monoxime) is an electro-contractile uncoupler that is well known to suppress cardiac muscle contraction12. We do not observe differences between the morphology of embryos that have been subject to this protocol, and embryos that have been growing normally and have not had their hearts stopped for imaging. However, the risk remains that the drug treatment is altering normal development13. Concentrated tricaine can also be used to temporarily stop the heart14. We find that we can effectively stop the heart by placing embryos in embryo medium with 0.32 % tricaine for 1 min, and maintain the heart in its stopped state by transferring the embryos to embryo medium with 0.16 % tricaine. Similar to the BDM treatment, the heart can be stopped this way for 30 min and heart beat can be seen to return within 5 min after returning embryos into normal embryo medium. Higher concentrations of tricaine should be avoided; low concentrations of tricaine for prolonged periods (h/days) has been shown to affect normal embryonic development15,16. Tricaine solution used for stopping the heart is best prepared fresh. If frozen aliquots are used, make sure to keep the solution well protected from light and to defrost the tricaine without heating the solution above 30 °C. Finally, blebbistatin has been reported to stop the heart without severe side effects13. We have not tested this last method in our lab.
Mounting and unmounting: Correct positioning of the embryos is critical during photoconversion. The mold for making agarose wells facilitates the mounting of embryos such that the anterior-posterior axis of the embryo consistently lies at 25° relative to the surface of the dish. We have found the mold to be applicable for photoconversion at all developmental stages between 36 and 55 hpf. In mounting, the tilt of the embryo in the left-right direction needs to be adjusted according to the embryo's developmental stage and the area one wishes to photoconvert. For example, to photoconvert the AVC at 36 hpf, the embryo should face roughly 15° to the left, while at 55 hpf, the embryo should face roughly 10° to the left. Under the stereomicroscope, you should be able to see the AVC unobscured by the ventricle or the atrium. If, for example, the atrium lies over the AVC, it would not be possible to photoconvert just the AVC or just the atrium. For imaging purposes, the same mold is used to position embryos between 36 and 96 hpf. The mold could be made using a variety of materials using a 3D printer or using traditional mechanical workshop tools. Stainless steel or chemical and heat resistant plastic are both suitable materials. Should the reader wish to create the mold using a 3D printer, please see supplementary information for the .STL file of the mold template.
Adding sufficient LMP agarose to cover the entire embryo is advised to ensure that the embryos do not become dislodged between mounting and photoconversion/imaging, especially if the microscope is not located directly adjacent to the mounting station. In general, provided that the embryos are properly mounted, LMP agarose does not stick to the embryos after unmounting. However, should some LMP agarose remain stuck on the embryo, the LMP agarose can be removed by gently teasing the embryo from the agarose using forceps.
Embryos, especially embryos less than 48 hpf, can be easily damaged during mounting and unmounting. For embryos less than 48 hpf, we recommend fire polishing the end of the glass pipette to prevent damaging embryos during pipetting.
Microscope: This protocol was performed using a Leica SP8 microscope. However, photoconversion is a well-established process and this protocol should be easily convertible to other upright microscopes equipped with lasers capable of producing 405 nm, 488 nm and 561 nm light. For embryos up to 62 hpf, images were acquired using an argon excitation laser and a DPSS 561 laser for green and red Kaede, respectively. For embryos at stages older than 62 hpf, green Kaede can be excited using a multiphoton laser set to 940 nm, which has a greater penetration depth than 488 nm light. In theory, red Kaede should have a two-photon excitation spectra similar to that of DsRed, and could be imaged at 950 nm17.
An objective lens with a high numerical aperture is necessary to efficiently photoconvert EdCs and to acquire high-resolution images. Since the embryos are kept in media, it is important to choose a microscope equipped with a water objective lens.
Photoconverting the region of interest: It is advisable to be conservative when selecting the ROI to photoconvert – the 405 nm laser diffuses as it travels through tissue and can photoconvert cells in an area greater than that selected. If, for example, you wish to label a cell red but would like the cell directly adjacent to remain green, it may be prudent to select only the part of the cell lying further away from the adjacent cell as your ROI. Since Kaede is expressed in the cytoplasm of EdCs, photoconverted protein will diffuse to parts of the cell not selected by the ROI.
Photobleaching and phototoxicity: The Kaede protein, both its green and red forms, is extremely bright and photobleaching is generally not a limiting factor in cell tracking experiments. However, laser light can impact embryonic development and should be minimized, especially during early stages of development. Signs of phototoxicity include developmental delay and, in more serious cases, improper heart looping and malformed valves.
The power of the violet laser (25 %) and the number of iterations the violet laser scans over the ROI (3-6 times) used in this protocol were determined empirically. At these settings, we saw almost complete conversion of Kaede from its green to red form without noticeable developmental delay. The amount of laser power required in other microscopy setups would depend on the laser's intensity. It should be noted that the red form of Kaede can be bleached by intense 405 nm illumination, so longer photoconversion does not necessarily yield more red fluorescence.
Image analysis: The method proposed for the image analysis is a newly developed process to describe the three-dimensional geometry of the AVC and compare it among embryos. It is intended to be applied to monolayer tubular structures, such as the AVC before 48 hpf and potentially the heart tube, the dorsal aorta and others. It is mainly automatic, even though the segmentation on the cutting planes is performed manually by the user and other steps could potentially require manual correction. The manual segmentation represents the most critical part of the process, since knowledge of the structure geometry and training are needed to achieve good results. After practicing, the time-consumption of this step is considerably reduced.
The protocol described here provides an effective way to track endocardial tissue movements during AVC and atrioventricular heart valve development using the Tg(fli1a:gal4FFubs, UAS:kaede) zebrafish transgenic line.
Currently, one of the key limitations of this method is that the violet laser beam is not confined in the axial dimension. This makes targeting of single cells for photoconversion, and hence clonal analyses and single cell tracking, difficult. Two years ago, it was discovered that the photoconvertible protein Dendra2 can be primed converted – the protein can be converted from its green form to its red form by first irradiating the protein with blue light, and then with near-infrared light18. Microscope setups that exploit this property can photoconvert Dendra2 expressing cells in a spatially confined manner in complex 3D tissue structures while avoiding many of the phototoxicity problems associated with intense near-UV irradiation18,19,20. With the recent discovery of the molecular mechanism behind this phenomenon, variants of many green-to-red photoconvertible fluorescent proteins, including Kaede, have been engineered so that are also capable of entering a primed intermediate state21. To our knowledge, zebrafish transgenic lines where a promoter such as fli1a or kdrl drives the expression of protein Dendra2 or other primed convertible fluorescent proteins in endothelial cells are not yet available. However, lines that express Dendra2 ubiquitously are available and may prove useful in certain applications 22,23.
The authors have nothing to disclose.
We would like to thank Anne-Laure Duchemin, Denise Duchemin, Nathalie Faggianelli and Basile Gurchenkov for helping to design and make the mold described in this protocol. This work was supported by FRM (DEQ20140329553), the ANR (ANR-15-CE13-0015-01, ANR-12-ISV2-0001-01), the EMBO Young Investigator Program, the European Community, ERC CoG N°682938 Evalve and by the grant ANR-10-LABX-0030-INRT, a French State fund managed by the Agence Nationale de la Recherche under the frame program Investissements d’Avenir labeled ANR-10-IDEX-0002-02.
Materials | |||
Necessary equipment for raising fish and collecting eggs (see the Zebrafish Book22 for details). | |||
Stereomicroscope | |||
Upright confocal microscope (Equipped with lasers operating at 488 nm, 561 nm, and 405 nm, a tunable multiphoton laser, and a Leica HCX IRAPO L, 25 ×, N.A. 0.95 objective) | Leica | Leica SP8 | |
Heat block | ThermoScientific | 88870001 | |
35 mm x 10 mm tissue culture dish | Falcon | 353001 | |
6 well plate | Falcon | 353046 | |
Petri dish | |||
Forceps | |||
Glass pipette | |||
Mold | |||
Reagents | |||
8 mg/mL Tricaine stock solution | |||
1 M BDM stock solution | |||
UltraPure low melting point agarose | Invitrogen | 16520-050 | |
Agarose | Lonza | 50004 | |
PTU (1-phenyl-2-thiourea) | Sigma Aldrich | P7629 | |
Embryo medium: 30x stock solution | |||
Software | |||
Matlab equipped with the Image Analysis, Curve Fitting, Bio-Formats Toolboxes | MathWorks | ||
8 mg/mL Tricaine stock solution | |||
200 mg of tricaine powder (Ethyl 3-aminobenzoate methanesulfonic acid) | Sigma-Aldrich | E10521 | |
25 mL Danieau | |||
Adjust to pH 7, aliquot and store at -20 °C | |||
1 M BDM stock solution | |||
1 g BDM | Sigma-Aldrich | 112135 | |
10 mL ddH2O | |||
Aliquot and store at -20 °C | |||
Embryo medium: 30x stock solution | |||
50.7 g NaCl | Sigma-Aldrich | 7647-14-5 | |
0.78 g KCl | Sigma-Aldrich | 7447-40-7 | |
1.47 g Magnesium sulfate | Sigma-Aldrich | 7487-88-9 | |
2.1 g Calcium nitrite tetrahydrate | Sigma-Aldrich | 13477-34-4 | |
19.52 g HEPES (4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid) | Sigma-Aldrich | 7365-45-9 | |
Adjust to pH 7.2 and store at room temperature | |||
Mold | |||
PlasCLEAR resin | Asiga | ||
Plus 39 Freeform Pico 3D printer | Asiga |