Time-lapse imaging of embryonic neural stem cell division in Drosophila by two-photon microscopy.

1 Time-Lapse Imaging of Embryonic Neural Stem Cell Division in Drosophila by Two-Photon Microscopy Elena Rebollo1 and Cayetano Gonzalez1,2 1Cell Division Group, IRB-Barcelona, Barcelona, Spain 2Institucio´ Catalana de Recerca i Estudis Avancats, Barcelona, Spain ABSTRACT This unit describes a protocol for live imaging of Drosophila embryonic neural stem cells using two-photon microscopy. Compared to traditional one-photon confocal imaging, this technique renders higher-resolution optical sections from deeper within the embryo. It is ideally suited to following embryonic neuroblasts located underneath the neuroepithelial cell layer for several rounds of cell division. INTRODUCTION This unit provides a detailed description of how to prepare and image live Drosophila embryonic neural stem cells using two-photon microscopy. Drosophila embryos constitute an excellent system for live microscopy (Cavey and Lecuit, 2008) and have been used extensively to address the dynamics of different developmental processes. How-ever, relatively little live-imaging data exist on the behavior of embryonic neuroblasts (NBs), the Drosophila neural stem cells, because as NBs delaminate from the embry-onic neuroepithelium towards the interior of the embryo, they become inaccessible to standard microscopy techniques. Two-photon microscopy, which greatly improves resolution at deeper optical sections and causes little cell damage, can be used to circumvent this limitation. By fine-tuning the recording parameters, NBs can be followed through several consecutive divisions, providing an excellent tool to study different aspects of asymmetric cell division and embryonic neurogenesis. The unit begins with a Basic Protocol that describes the method, equipment, and pa-rameters used to image embryonic NBs by two-photon microscopy. It is followed by a Support Protocol for collecting and preparing Drosophila embryos for live microscopy. TIME-LAPSE IMAGING BY TWO-PHOTON MICROSCOPY This protocol describes the basic setup used to obtain a time-lapse series of images of dividing embryonic NBs using a two-photon microscope (see Video 1). Since we routinely use GFP-αTub84B and YFP-Asl (Rebollo et al., 2007) to visualize microtubules and centrioles, a description of how to excite and detect both GFP and YFP signals simultaneously is included. Microscope set-up conditions are given for a Leica TCS-SP5 multiphoton microscope (Leica Microsystems), but any multiphoton system that fits the minimum requirements described above (see Strategic Planning) can be utilized. Materials Dechorionated embryos, mounted and ready for microscopy (see Support Protocol) Leica TCS-SP5 multiphoton LSM system (or equivalent), including: General components: Inverted microscope (DMI 6000 CS) Laser point scanner system 63×/1.4 objective (oil immersion, HCX PL APO) Multiphoton components: Ti:sapphire laser Mai Tai (Spectra Physics) with wavelength range of 710 to 990 nm, pulse frequency of 80 MHz, giving maximum pulses of 1.5 W and 1 psec duration, and power control by EOM (electrooptical modulator) External detector (NDD, non-descanned detection): PMT 9624 (high sensitive meshless type, selected for Leica) Basic barrier filter SP720 (to stop infrared radiation and allow detection of whole GFP and YFP emission spectra) Leica LAS AF acquisition software (implemented for control of infrared laser) Software for visualizing and processing images (Imaris, Bitplane) and for assembling videos (Image J, http://rsb.info.nih.gov/ij/, or Adobe After Effects) Set up equipment 1. Switch on the confocal equipment and the pulse laser to warm them up and ensure that they are operative. It is best to set up the system before dechorionating any embryos, so that the preparation can be observed at the microscope immediately after it is mounted. 2. Make sure that the room temperature is maintained at a maximum of 23◦C throughout the imaging session. Due to laser irradiation, the temperature at the sample will be higher than the room temperature. Embryos will die during recording if the room temperature reaches 26ºC. For a more tight control of temperature, a temperature-controlled chamber around the sample may be used. The recommended chamber for the system described here is the Leica 11531834 (Cooling Thermostat, 230 V, control range –20º to 200ºC, control tolerance ± 0.01ºC). 3. Use the computer interface to select a wavelength for infrared excitation. To excite both GFP and YFP at the same time, 960 nm is efficient. Focus microscope on embryo 4. Choose an embryo at the desired stage. To avoid photo-damage, bring the embryo into focus using transmitted light, which is also recommended to identify the devel-opmental stage of the embryo. With practice, the 63× objective can be used for this purpose. 5. Once in focus, rotate the embryo to orient it for optimal image size. Avoid post-acquisition rotations, which are time-consuming. Bear in mind that some software versions deliver digital images that are rotated by 90◦ with respect to the view through the microscope opticals (eyepieces). 6. Select the scanning format. For time-lapse acquisition, it is recommended to select small scanning formats, around 256 × 256 pixels. Larger formats produce more photo damage and slow down the acquisition rate. A 256 × 200 pixel format is recommended for recording embryonic NBs. 7. Select an intermediate laser power ( 50%), start scanning, and adjust laser gain and offset.∼ It is not recommended to use more than 80% laser power, as this will produce too much heat within the specimen. The signal-to-noise ratio can be improved by using the “accumulate” function (which will add up the signal from three consecutive acquisitions) or by reducing the scanning speed to 200 Hz. The resulting signal can be digitally tuned by modifying the detector gain and offset. The under/overexposure option can be used to identify those areas of the image that are not properly exposed. 8. Select a region of interest and apply digital zoom to fit the image to the desired area. The ideal region to record Drosophila NBs is the ventral side of the embryo. Move the embryo until the ventral ectoderm (the more curved side) is in the center of the scanning field. Focus on the surface (neuroepithelial cell layer) and move in the z dimension until NBs appear in the field. Delaminating NBs can be recognized by their position (slightly off the epithelial layer), their shape (pear-like, with a protrusion that connects them to the epithelium), and later by the axis of division (perpendicular to the epithelial layer) and the asymmetry of division (rendering a large new NB and a small basal ganglion mother cell [GMC]). Since embryos contain several rows of delaminating NBs that are oriented from anterior to posterior, many NBs can be recorded at the same time. A 2.5 zoom factor will be enough to capture several NBs within the same field. Higher zoom factors will produce more damage during acquisition. 9. Re-adjust the image digitally using the under/overexposure function. 10. Select the upper and lower limits for the z stack to be acquired. Always acquire one extra section at both ends of the z stack. Make sure a maximum of 1 μm is set between sections. In this way, 30 μm can be easily covered in the z dimension, which allows for imaging NBs at different depths. 11. Select bidirectional scanning to speed up acquisition. 12. Activate the xyzt mode and adjust the interval frequency. The minimum time required for acquiring a stack of 30 sections will be 0.5 min with the specified settings. A 1-min interval is recommended to avoid continuous heating of the sample. Acquire images 13. Start acquisition. During this process, external light sources may considerably increase image noise and must be avoided. Switch off the monitors and cover the microscope with a dark cloth. 14. To visualize the recorded four-dimensional (4-D) stacks and perform some basic image processing 3 (selection, projection, filtering), use the LSM acquisition software or any other software able to handle 4-D stacks, such as Imaris (Bitplane). 15. To process videos from individual NBs, select the z sections corresponding to the cell of interest and project them using the maximum intensity projection function. The movie can be assembled using any software designed for such purposes, such as Image J or Adobe After Effects. EMBRYO PREPARATION Optimal visualization of Drosophila embryos under the microscope requires removal of the chorion while leaving the viteline membrane intact, as the latter is transparent enough for fluorescence imaging and protects the embryo. Hand peeling is chosen instead of chemical dechorionation methods, because very few embryos are needed and chemical methods may affect viability. A brief summary is given below. Protocols for embryo collection and dechorionation—including recipes for apple plates and yeast paste and detailed descriptions of egg-laying cups—have been published elsewhere (Wieschaus and Nusslein-Volhard, 1988; Ashburner, 1989; Gonzalez and Glover, 2003; Kiehart et al., 2007; Rothwell and Sullivan, 2007). Also included is a description of how to mount the embryos for live microscopy. A standard protocol based on gluing the embryos to the cover glass and adding an oil layer on top to prevent desiccation is available (Cavey and Lecuit, 2008). However, to improve optical penetration, it is essential to minimize light scattering layers between the optics and the specimen. Therefore, it is better to mount the embryos directly onto the cover glass within a small drop of oil and cover them with an oxygen-permeable Teflon membrane that keeps them immobilized and slightly flattened (Kaltschmidt et al., 2000). Teflon membranes (YSI) are sold in packs of 100 as a component of oxygen monitors. Since they are oxygen-permeable and non-toxic, they can be used as a cheap alternative to Biofolie 25 (Heraeus), normally used for mounting live specimens for microscopy. Under these conditions, embryos can be kept alive and imaged for hours. Materials Male and female Drosophila, 1 to 2 weeks old Apple plates: 3% (w/v) agar plates containing 4% (v/v) apple juice, freshly prepared and kept at 4◦C Freshly made yeast paste Halocarbon oil (Sigma; halocarbon oil 700, cat. no. H8898 or Voltalef oil 10S) Plastic fly cages pierced with small holes to allow breathing Stereomicroscope Fine artist’s brush Double-sided cellophane tape (Scotch tape) Blunt forceps, e.g., standard No. 5 dissection forceps smoothed for this purpose 35-mm culture dish: Petri dish with a coverslip bottom (e.g., Fluoro Dish sterile culture dishes, World Precision Instruments, cat. no. FD#35) Teflon membranes (YSI Incorporated; cat. no. 5793) Vacuum grease (Dow Corning; cat. no. 976V) Collect embryos 1. Place well-fed females and males, ideally between 1 and 2 weeks old, inside a plastic fly cage. Any plastic beaker-like container that securely fits the diameter of the apple plates will work. 2. Seal the cage with an inverted apple plate supplemented with a drop of yeast paste at room temperature. Invert the cage so the plate is at the bottom, and keep in the dark at 25◦C. Females lay better after being kept in such conditions for 1 or 2 days. 3. To start collecting, transfer flies to a clean cage with a new apple plate with fresh yeast. Change the plate every hour to make collections of nearly synchronized embryos. Label the plates with the collection time and allow the embryos to develop for 4.5 hr at 25◦C. Dechorionate embryos and prepare for microscopy The following steps are performed under a stereomicroscope. 4. Using a wet fine artist’s brush, take an embryo from the apple plate and place it on top of a piece of double-sided Scotch tape attached to a glass slide. 5. Remove the chorion by gently rolling the embryo over the tape using the blunt tip of a pair of forceps. The dechorionated embryo will stick to the forceps. 6. Transfer the embryo to a small drop of oil on the coverslip of a culture chamber. Any commercial brand can be used, provided the coverslip has the standard 0.17-mm thickness. To prevent desiccation, make sure that the embryo is fully covered by the oil. 7. Add two drops of vacuum grease at both sides of the oil drop using a fine pipet tip. 8. Cut a piece of Teflon membrane to a size that is just big enough to cover the oil and the vacuum grease drops. Place the membrane on top of the oil and press it down gently with the forceps, so that the embryo is immobilized and slightly flattened against the glass. The embryo is now ready for two-photon microscopy. When mounted as described, the embryo will naturally lay on the lateral surface, so that dorsal and ventral sides remain at the upper and lower edges, which is the ideal orientation for microscopy. It is advisable to process several embryos at once to ensure that at least one will be at the required developmental stage. NB delamination from the epithelium starts at stage 9 of embryogenesis, when embryos are 4.5 hr old. The preparation process should not ∼ take longer than 20 min. Thus, processed embryos should be between 3 hr 50 min and 4 hr 50 min old, right at or about to enter stage 9. This stage is easily recognized by landmarks that can be observed under a stereomicroscope, including the initial formation of the future stomodeal invagination (Wieschaus and NussleinVolhard, 1988). COMMENTARY Background Information Neuroblast division Neuroblasts (NBs) are stem cell−like pro-genitors of the Drosophila nervous system. The first NBs delaminate from the polarized neuroectoderm early in develop-ment, at around stage 9 (Campos-Ortega and Hartenstein, 1997). Delamination is fol-lowed by asymmetric mitosis—with telophase figures aligned perpendicularly to the neuroectoderm—that results in a small differentiating ganglion mother cell (GMC) and a self-renewed NB. GMCs divide once more to produce cells that later differentiate into neurons or glia. Some larval NBs derive from qui-escent embryonic NBs that re-enter the cell cycle. In both developmental stages, asymmetric division of NBs relies on the polarized localization of apical and basal protein complexes that are differentially segregated to the daughter cells by the controlled orientation of the mitotic spindle (Gonzalez, 2007; Knoblich, 2008). Pioneer time-lapse studies carried out in embryonic NBs revealed that the spindle is first assembled parallel to the neuroectoderm, and then rotates approximately 90º during metaphase to align with the apical/basal polarity axis (Kaltschmidt et al., 2000). More re-cent studies based on two-photon microscopy have shown that spindle rotation is limited to delaminating NBs. In later cell cycles, spindles are assembled already aligned with the axis of cortical polarity (Rebollo et al., 2009). This predetermined spindle orientation mode was previously reported in larval NBs (Rebollo et al., 2007; Rusan and Peifer, 2007). Two-photon microscopy When optical sections are obtained deep within a sample using confocal microscopy, light scattering attenuates the fluorescence signal. Furthermore, scattering that occurs in regions away from the point of focus creates undesired fluorescence that will pass through the confocal pinhole to the detector, thereby increasing background. This deterioration in signal-to-noise levels is accentuated when us-ing live fluorescence signals, which are generally less intense than fluorophores used in fixed samples. An alternative for imaging thick specimens is the two-photon system, in which excitation is caused by the simultaneous ab-sorption of two photons with approximately twice the wavelength of the absorption peak of the fluorophore being used. In this manner, excitation is restricted to the focal plane, where the density of photons is high enough to al-low two photons to simultaneously hit the fluorophore. An additional advantage of two-photon microscopy is that longer wavelengths reduce photo damage, thereby improving cell viability. Moreover, since no fluorescence is generated above or below the point of illumination, all detected signal can be used for imaging regardless of whether it has been scattered. More importantly, longer wavelengths suffer less scattering. Thus, the longer wave-lengths allow deeper tissue penetration than can be achieved using the shorter excitation wavelengths used in standard confocal microscopy. Comparative views of a Drosophila embryo taken using two-photon versus one-photon imaging are shown in Figure 1H.2.1. Two-photon live imaging in Drosophila Live microscopy of Drosophila cells and tissues has been widely used to address many different questions related to cell behavior and division during development. It is relatively easy to create transgenic flies expressing different combinations of fusion proteins that render fluorescent chromosomes, kinetochores, centrosomes, microtubules, cell membranes, and so on in different colors and in the tissues of interest. This, together with the particular features of GFP derivatives (such as brightness, low toxicity, and the available palette of different excitation and emission spectra), makes these fluorescent markers invaluable for live 5 microscopy in this system. Two-photon microscopy takes advantage of exactly these same properties. Drosophila embryos expressing fluorescent GFP reporters have been tracked using two-photon microscopy to address collective cell migration and movement during gastrulation (McMahon et al., 2008) and to analyze den-drite elimination in Drosophila dendritic arborizing sensory neurons during metamorphosis (Williams and Truman, 2005). The high-intensity peaks of the pulse laser can also be used to ablate specific cells in the Drosophila embryo while preserving local integrity, which has made it possible to study the regulation of specific morphogenetic movements during development (Supatto et al., 2005). Two-photon excitation has become the imaging choice for highly light-scattering tissues, and many more applications in Drosophila and other systems will be developed in the near future. Critical Parameters and Troubleshooting Two-photon imaging can be directly implemented using a standard confocal laser-scanning microscope. It requires a high peak-power pulse laser to ensure that the density of photons at the focal plane is sufficient to provoke a two-photon event, while the mean power levels remain moderate and do not damage the specimen. In addition, a special barrier filter that stops the infrared light reflected from the sample must be placed before the detector. Since two-photon excitation spectra are typically broader than the corresponding one-photon excitation spectra, it is possible to excite several fluorophores (e.g., GFP and YFP, or GFP and m-RFP) at a time with a single excitation wavelength. The emission spectra of both fluorophores may be detected as a single signal when such fluorophores are used to label cell structures that can be distinguished in other ways (for in-stance, by shape or localization); this requires that the barrier filter cube does not include any additional bandpass filter. In the protocol de-scribed here, microtubules and centrioles are readily distinguishable by structure, so GFP and YFP can be detected as a single signal. When fluorophores cannot be distinguished by the morphology of the labeled structures, spectral separation is required, using two external detectors with the appropriate bandpass filters. Using a unique excitation wavelength for several fluorophores will save time during the acquisition process and avoid cell damage due to extra laser power. Objectives with a high numerical aperture should be used to optimize photon concentration at the focal point. The infrared illumination point is moved along the sample by a computer-controlled scanner, which consists of mirrors located along the epiluminescence path. These mirrors move the laser beam in the xy plane along different z positions, al-lowing for three-dimensional reconstruction of the sample at each selected time point. A beam splitter is required to reflect the excitation infrared light to the objective and transmit the fluorescence emitted by the sample, which will be filtered again by the additional infrared barrier. Since two-photon excitation is restricted to the focal plane, detection pinholes are not needed. Nevertheless, the built-in detection pinhole can be used to reduce noise, although it will be at the expense of signal. Most microscopes have extra software tools to help adjust the image digitally. Due to the use of longer excitation wave-lengths, resolution is always slightly lower in two-photon microscopy than in traditional confocal microscopy. The fluorescence emission intensity of a given fluorophore is also reduced in two-photon microscopy due to the pulsing activity of the laser. The loss of resolution can be partially compensated by reducing the aperture of the confocal pinhole, but at the expense of decreasing signal intensity. To compensate for this, it is possible to use lower scanning speeds or use functions that accumulate or average the signal and there-fore reduce the noise. However, all of these manipulations will result in longer acquisition times and, ultimately, a loss of temporal res-olution. Choosing a smaller scanning format will speed up the acquisition rate and may pro-vide a compromise between spatial and tem-poral resolution. Increasing the repetition rate of the laser will also help improve signal inten-sity, with a limitation imposed by the viability of the specimen due to overheating. The use of objectives that are corrected for infrared light will definitely improve signal detection and resolution. Anticipated Results This protocol generates 4-D acquisitions of dividing embryonic Drosophila NBs that can-not be properly visualized with standard one-photon excitation techniques. The differences in penetration start to be evident at a 20-μm depth. The two-photon system allows one to follow cells throughout the entire z dimension of the embryo and to visualize several rows of NBs located at different depths. In the range of the specified magnification, ten to sixteen dividing NBs can be recorded at a time and over several consecutive divisions. The duration of acquisition and the final number of cell cycles followed will depend on the acquisition parameters. For long-term acquisitions, it is recommended to use longer time intervals (minimum 1 min). Time Considerations The total time needed for embryo collection, development to the desired stage, and dechorionation and mounting for microscopy is approximately 6 hr. For a single embryo, a recording session of 2 hr will cover sev-eral rounds of division. If several consecutive embryo collections are done, two or three em-bryos can be followed in a single day. Process-ing will take much longer, even days, since the data volume produced is heavy and each stack will contain several NBs. Acknowledgements We thank Mónica Roldán for her valuable contribution to the establishment of the basic imaging conditions described in this protocol. Literature Cited Ashburner, M. 1989. Drosophila: A Laboratory Manual. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, N.Y. Campos-Ortega, J.A. and Hartenstein, V. 1997. The Embryonic Development of Drosophila melanogaster. Springer, Berlin. Cavey, M. and Lecuit, T. 2008. Imaging cellular and molecular dynamics in live embryos using flu-orescent proteins. Methods Mol. Biol. 420:219-238. Gonzalez, C. 2007. Spindle orientation, asymmetric division and tumour suppression in Drosophila stem cells. Nat. Rev. Genet. 8:462-472. Gonzalez, C. and Glover, D.M. 2003. Techniques for studying mitosis in Drosophila. In The Cell Cycle: A Practical Approach (P. Fantes and R. Brooks, eds.), pp 143-175. IRL Press, Oxford. Kaltschmidt, J.A., Davidson, C.M., Brown, N.H., and Brand, A.H. 2000. Rotation and asymmetry of the mitotic spindle direct asymmetric cell di-vision in the developing central nervous system. Nat. Cell Biol. 2:7-12. Kiehart, D.P., Crawford, J.M., and Montague, R.A. 2007. Collection, dechorionation, and prepa-ration of Drosophila embryos for quantita-tive microinjection. Cold Spring Harb. Protoc. doi:10.1101/pdb.prot4717. Knoblich, J.A. 2008. Mechanisms of asymmetric stem cell division. Cell 132:583-597. McMahon, A., Supatto, W., Fraser, S.E., and Stathopoulos, A. 2008. Dynamic analysis of Drosophila gastrulation provides insights into collective cell migration. Science 322:1546-1550. Rebollo, E., Sampaio, P., Januschke, J., Llamazares, S., Varmark, H., and Gonzalez, C. 2007. Func-tionally unequal centrosomes drive spindle ori-entation in asymmetrically dividing Drosophila neural stem cells. Dev. Cell 12:467-474. Rebollo, E., Rolda´n, M., and Gonzalez, C. 2009. Spindle alignment is achieved without rotation after the first cell cycle in Drosophila embryonic neuroblasts. Development 36:3393-3397. Rothwell, W.F. and Sullivan, W. 2007. Drosophila embryo dechorionation. Cold Spring Harb. Pro-toc. doi:10.1101/pdb.prot4826. Rusan, N.M. and Peifer, M. 2007. A role for a novel centrosome cycle in asymmetric cell division. J. Cell Biol. 177:13-20. Supatto, W., Debarre, D., Moulia, B., Brouzes, E., Martin, J.L., Farge, E., and Beaurepaire, E. 2005. In vivo modulation of morphogenetic movements in Drosophila embryos with fem-tosecond laser pulses. Proc. Natl. Acad. Sci. U.S.A. 102:1047-1052. Wieschaus, E. and Nusslein-Volhard, C. 1988. Looking at embryos. In Drosophila: A Practi-cal Approach. (D.B. Rogers, ed.), pp. 199-227. Oxford University Press, Oxford. Williams, D.W. and Truman, J.W. 2005. Cellular mechanisms of dendrite pruning in Drosophila: Insights from in vivo time-lapse of remodeling dendritic arborizing sensory neurons. Develop-ment 132:3631-3642. 7 Figure 1H.2.1 Comparative view of two-photon versus one-photon imaging of the ventral neuroectoderm of a Drosophila embryo expressing GFPand YFP-labeled microtubules and centrioles respectively. Two xyz stacks of 86 slices every 0.88 μm were acquired using two-photon and one-photon methods with a delay of 2 min between them. Selected z slices are shown at different depths from the embryo surface (12, 20, 42 and 70 μm). Very little difference is observed at the embryo surface (A) and at 12 μm (B), where dividing epithelial cells are observed (arrows in B). At 20 μm (C), differences in signal detection become very obvious. In delaminated NBs, the apical microtubule aster (arrowhead) and the basal cluster of daughter cells (asterisk) are perfectly distinguishable in the two-photon image, but are hardly visible in the corresponding one-photon image. At 42 μm (D), dividing NBs are still fully visible in the two-photon image (inset), but are undetectable in the corresponding one-photon image. Two-photon imaging can still render high-contrast images at 70 μm (E), where the epithelium is still observable. In contrast, there is a complete lack of signal with the one-photon system. Scale bar, 10 μm. E D C B surfac e A one photon two photon