Measuring Protein Stability in Living Zebrafish Embryos Using Fluorescence Decay After Photoconversion (FDAP)

1. Generating a Photoconvertible Fusion Construct and Injecting Dechorionated Zebrafish Embryos

1. Generate a functional construct containing the protein of interest fused to a green-to-red photoconvertible protein (see Discussion), then use in vitro transcription to generate capped mRNA encoding the fusion protein as in Müller et al., 2012¹⁹.
2. Use pronase to remove the chorions from about 30 zebrafish embryos at the one-cell stage. Alternatively, manually dechorionate embryos using forceps²⁸.
   Note: Embryos must be dechorionated for subsequent imaging. If desired, embryos can be injected through the chorion and dechorionated later, just prior to imaging.
   1. Make a 5 mg/ml stock solution of pronase from Streptomyces griseus in standard zebrafish embryo medium¹⁹. Rock the solution gently at RT for 10 min to allow the protease to dissolve. Aliquot 2 ml into microcentrifuge tubes and freeze at -20 °C.
   2. Transfer one-cell stage embryos to a 5 cm diameter glass or agarose-coated plastic Petri dish containing ~8 ml embryo medium. Add 2 ml of thawed pronase stock solution to the dish and incubate at RT for 5 to 10 min.
   3. Avoid exposing embryos to air or plastic, as contact with either will cause dechorionated embryos to rupture. Fill a 200 ml glass beaker with embryo medium. Transfer the embryos to the beaker by tilting the Petri dish while submerging it in the medium.
   4. After the embryos have settled to the bottom of the beaker, pour out most of the embryo medium, then pour fresh embryo medium into the beaker. The mild swirling of the medium pouring into the beaker causes embryos to lose their weakened chorions.
   5. Repeat step 1.2.4.
3. Transfer the dechorionated embryos to an agarose-coated injection dish²⁹ using a glass Pasteur pipette with a flamed tip. Flaming the pipette tip prevents jagged edges from injuring embryos.
4. Co-inject the mRNA and a 3 kDa Alexa488-dextran conjugate^29,30 (Figure 1B; see Discussion for suggested mRNA and Alexa488-dextran amounts). Inject directly into the center of the cell (not the yolk) to ensure even distribution of mRNA and fluorescent dye once cleavage commences.
   Note: The Alexa488 signal will be used during data analysis to generate compartmental masks in order to distinguish between intracellular and extracellular fluorescence.
5. Transfer injected embryos to a 1–2% agarose-coated well of a six-well plastic dish filled with embryo medium. Incubate in the dark at 28 °C until embryos have reached late sphere stage³¹ (approximately 5 hr post fertilization). Check embryos every one to 2 hr under a stereomicroscope and remove any debris generated by embryos that have died.

2. Mounting Zebrafish Embryos for Photoconversion and Imaging on an Inverted Confocal Microscope

1. Use a stereomicroscope to identify one to five healthy embryos, and use a glass Pasteur pipette with a flamed tip to remove them from the dish.
2. Gently eject the embryos into a microcentrifuge tube containing ~1 ml of melted 1% low melting point agarose in 1x Danieau’s embryo medium (see Materials List) (Figure 2A).
   Note: Agarose should have a temperature between 40 and 42 °C; higher temperatures could damage the embryos.
3. Draw the embryos back into the pipette along with some agarose. Gently eject the agarose and embryos onto the cover glass of a glass-bottom dish (Figure 2B). Ensure that the thickness of the cover glass is compatible with the objective on the confocal microscope.
   1. Re-use the glass pipette if desired. To clean the residual agarose out of the pipette and prevent clogging, quickly pipette embryo medium up and down. Place a 15 ml tube filled with ~5 ml of embryo medium next to the stereomicroscope for this purpose.
4. Use a metal probe to position the embryos so that the animal pole (blastoderm) faces the cover glass. Work quickly since the agarose will solidify in 20–30 sec. Use the stereomicroscope to monitor the embryos’ positions and readjust as necessary until the agarose hardens.
5. Repeat steps 2.1–2.4 until the desired number of embryos has been mounted.
   Note: In a typical experiment, four agarose drops containing four or five embryos each will fit easily on the cover glass. About 16 embryos can be imaged during a single ideal experiment (Figure 2C).
6. When the agarose has solidified, fill the glass-bottom dish with 1x Danieau’s embryo medium.

3. Photoconverting and Measuring the Decrease of the Photoconverted Signal

A 25X or 40X water objective is appropriate for the size and refractive index of zebrafish embryos. It is best to use immersion oil with the same refractive index as water rather than actual water, since water will evaporate during the course of the five-hour experiment. Ensure that the immersion oil is designed to be used with a water (not oil) objective.

1. Place a large drop of immersion oil on the objective to ensure that the oil film between the objective and cover glass will not break as the stage moves to different embryo positions during imaging. Securely place the glass-bottom dish onto the stage so that the dish will not shift when the stage moves. If possible, use a heated stage at 28 °C, the optimal temperature for zebrafish development.
2. Define each embryo’s position in the confocal microscope’s software package. Adjust the z-depth for each embryo, and attempt to target roughly the same plane in each embryo.
   Note: About 30 μm from the animal pole is a good depth since at this depth the enveloping layer of the embryo can be avoided, imaging area is maximized, and light scattering is minimal. A single optical slice with a thickness of ~3.3 μm provides sufficient data; there is no need to acquire a z-stack (see Section 5).
3. Collect two signals during the experiment: the “green” signal from the Alexa488-dextran conjugate—which will be used during data analysis to isolate extracellular and intracellular fluorescence—and the “red” signal from the fusion protein after it is photoconverted.
   1. Excite Alexa488 using a 488 nm laser, and collect emitted fluorescence between ~500 and 540 nm. Note: After photoconversion, many green-to-red photoconvertible proteins (e.g., Dendra2) can be excited with a 543 nm laser and emit fluorescence between ~550 and 650 nm. Adjust as necessary based on the photoconvertible protein used.
4. Acquire “pre-photoconversion” images, and configure the confocal microscope’s software to image each of the previously defined positions (from step 3.2) with the appropriate imaging conditions every 10 or 20 min for a five-hour time course (see Section 5 and Discussion).
5. To photoconvert the fusion protein, switch to a 10X objective and expose groups of embryos to UV light from a mercury arc lamp with a ~300–400 nm excitation filter at 100% output for 2 min. Shift the focus along the z-axis to promote uniform photoconversion (see Section 5). Ensure that the immersion oil does not drip onto the 10x objective during photoconversion.
   Note: The shifting of focus during photoconversion could be automated in order to avoid variability among experimenters.
6. Switch back to the 25X or 40X objective immediately after photoconversion. Ensure that the previously defined positions from step 3.2 are still accurate. If the dish shifted during photoconversion, re-define the positions of the embryos.
   1. Start the program created in step 3.4 and allow imaging to continue for 5 hr. Note the time elapsed between photoconversion and the start of imaging for each embryo.
7. Occasionally check on the experiment. Monitor the level of Danieau’s medium and add more if necessary. Restart the software if it has stalled.
8. In order to determine the background fluorescence values that will be used during data analysis to estimate the asymptote of an exponentially decreasing model, include some embryos that have been injected with Alexa488-dextran but not mRNA in the experiment. To determine the instrument noise, which will also be used during subsequent data analysis, acquire an image in the absence of a sample.

4. Analyzing the Data Using PyFDAP

1. Visually inspect the time course data sets from each embryo. Discard data sets from embryos that died during imaging, that shifted significantly, that have very low levels of photoconverted signal, or that contain regions of cells that look unusual and have stopped moving and dividing (typical of injured or sick embryos).
   Note: Occasionally, bubbles in the immersion oil or other artifacts will appear in one or two images in an otherwise usable data set. Note any images that contain artifacts; they will be discarded later, and the remaining time points from such a data set can still be analyzed.
2. Use the Python-based software package PyFDAP to analyze the FDAP data. PyFDAP calculates half-lives by determining the average intracellular and extracellular red fluorescence intensity in each image and fitting the data with an exponentially decreasing function⁵⁶ (Figure 3).
   1. Download PyFDAP (see Materials List).
   2. Use PyFDAP to separate intracellular and extracellular photoconverted signal (Figure 3A,B). Use the Alexa488 signal, which is strictly intracellular, to create an intracellular mask. Apply this mask to the corresponding red channel image to prevent intracellular pixels from being considered when calculating average extracellular intensity. To measure average intracellular intensity, invert the mask.
   3. In PyFDAP, display the masked images generated in step 4.2.2. Visually inspect these images and discard data sets in which masks do not accurately distinguish intracellular from extracellular space (this should be rare; note that cell membranes are included in images in which extracellular space has been masked, but they could be removed by altering the thresholding algorithm or by introducing a membrane mask (e.g., using membrane-CFP)). Also discard any single images containing artifacts (e.g., bubbles in the immersion oil) identified in step 4.1.
   4. Use PyFDAP to calculate average extracellular and intracellular fluorescence intensities for each image. PyFDAP calculates these averages by summing the intensities of pixels that fall outside of the mask and dividing by the total number of pixels summed.
   5. Fit the fluorescence data (Figure 3C) with the following exponential function:
      
      where t is time post-photoconversion, c(t) is intensity at a given value of t, c₀ is the intensity at t = 0, k is the clearance rate constant, and y₀ is the asymptote that the function approaches as fluorescence decreases (Figure 1D). y₀ can be constrained based on the measurements in step 3.8¹⁹.
   6. Use PyFDAP to calculate the extracellular and intracellular protein half-lives (τ) from the clearance rate constants (k) using the following relationship:
      

5. Control Experiments to Assess Photobleaching, Inadvertent Photoconversion, and Photoconversion Uniformity

1. Assessing photobleaching
   Note: Photobleaching could cause an artifactual decrease in fluorescence intensity that reflects the bleaching properties of the fluorescent protein in addition to the clearance of the protein of interest.
   1. To assess possible photobleaching, perform one set of FDAP experiments with 10 min intervals between imaging and a second set with 20 min intervals between imaging (Figure 4). Analyze the data from both sets of experiments using PyFDAP as described in Section 4.
   2. Compare the half-lives from the 10 and 20 min interval experiments. Longer half-lives from 20 min interval experiments indicate significant photobleaching. If the half-lives from both experiments are identical, photobleaching is not a significant concern.
   3. Alternatively, assess photobleaching by acquiring a series of ~30 images immediately after photoconversion. A significant decrease in fluorescence intensity indicates significant photobleaching.
   4. If photobleaching is detected, use lower laser power, decrease imaging time, or consider using a more photostable photoconvertible protein³².
2. Assessing inadvertent photoconversion.
   Note: Dendra2 can be photoconverted using 488 nm illumination²⁷. When exciting Alexa488 with the 488 nm laser as described in step 3.3.1, inadvertent photoconversion and therefore an artifactual increase in the apparent half-life of the protein of interest is possible. However, we and others³³ have found that 488 nm illumination is an inefficient method of photoconversion in zebrafish embryos.
   1. Use the control experiment described in step 5.1.1 to detect inadvertent photoconversion. Compare the half-lives from the 10 and 20 min interval experiments. Shorter half-lives from 20 min interval experiments indicate significant inadvertent photoconversion. If the half-lives from both experiments are identical, inadvertent photoconversion is not a significant concern.
   2. If inadvertent photoconversion is detected, use a lower 488 nm laser power and shorter imaging times to avoid inadvertently photoconverting Dendra2.
3. Assessing photoconversion uniformity.
   Note: If photoconversion is biased toward the animal pole of the embryo, the decrease in fluorescence will be influenced by protein diffusion or cell movement into deeper planes (Figure 5A).
   1. To determine whether photoconversion is uniform, express a secreted photoconvertible protein (for experiments with extracellular fusion proteins) or a cytoplasmic photoconvertible protein (for experiments with intracellular fusion proteins). Photoconvert as usual, then acquire a z-stack encompassing most of the blastoderm every 20 min for 80 min.
   2. If photoconversion is biased toward the animal pole, the fluorescence intensity in deeper planes will increase over time due to diffusion or cell movement (Figure 5B). If non-uniform photoconversion is detected, focus deeper into the embryos during photoconversion.