Live confocal imaging provides biologists with a powerful tool to study development. Here, we present a detailed protocol for the live confocal imaging of developing Arabidopsis flowers.
The study of plant growth and development has long relied on experimental techniques using dead, fixed tissues and lacking proper cellular resolution. Recent advances in confocal microscopy, combined with the development of numerous fluorophores, have overcome these issues and opened the possibility to study the expression of several genes simultaneously, with a good cellular resolution, in live samples. Live confocal imaging provides plant biologists with a powerful tool to study development, and has been extensively used to study root growth and the formation of lateral organs on the flanks of the shoot apical meristem. However, it has not been widely applied to the study of flower development, in part due to challenges that are specific to imaging flowers, such as the sepals that grow over the flower meristem, and filter out the fluorescence from underlying tissues. Here, we present a detailed protocol to perform live confocal imaging on live, developing Arabidopsis flower buds, using either an upright or an inverted microscope.
Most of the plant body forms post-embryonically from groups of stem cells situated at or near the tip — or apical meristems — of the shoots and roots. All the above-ground structures of the adult plant derive from the shoot apical meristem (SAM), which continuously produces lateral organs on its flanks: leaves during the vegetative phase, and flower meristems (FMs) after it transitions to the reproductive phase. FMs in turn develop into flowers. While the Arabidopsis SAM produces lateral organs one at a time, in an iterative, spiral pattern, FMs produce four types of floral organs within four whorls, in a partially synchronous manner, with multiple developmental programs unraveling simultaneously. The genetic networks underlying the specification of the identity of the different floral organs have been partially deciphered (for reviews, see references1,2), but many aspects of flower development, such as floral organ positioning and the definition of boundaries between whorls, remain poorly understood.
Early molecular genetic studies of plant development mostly relied on techniques such as in situ hybridization and GUS reporters to analyze gene expression. While these methods have provided a wealth of information and greatly contributed to our understanding of plant growth and flower development, there are important limitations: they lack good cellular resolution, do not allow for the easy observation of the expression pattern of multiple genes in the same samples, and importantly, can only be applied to dead, fixed tissues. Recent advances in confocal microscopy have overcome these limitations, and provide developmental biologists with a powerful tool to investigate the processes underlying plant morphogenesis. In particular, confocal microscopy allows for the observation of live tissues and organs throughout their formation, which is critical to fully understand a quintessentially dynamic process such as development.
Live confocal imaging has been extensively used to analyze the aerial growth of plants and the production of lateral organs by the SAM (e.g. references3,4,5,6), but with the exception of a few reports (e.g. references7,8,9,10), it has not been widely applied to the study of flower development. Protocols for live confocal imaging of the SAM are available (e.g. references11,12), and provide a good base for how to image the developing flower buds that surround the SAM. However, imaging flower buds presents specific challenges: for instance, flower buds quickly become bigger than the SAM, and at stage 4, sepals start covering the FM (stages as described in reference13), and dimming the fluorescence from underlying tissues (Figure 2A). Here, we provide a detailed protocol explaining how to perform live confocal imaging on developing Arabidopsis flower buds using either an upright or an inverted microscope and a water-dipping lens. We previously published this protocol in reference14.
1. Media and Dishes Preparation
2. Plant Growth
3. Dissection of the Shoot Apex
Figure 1: Preparation of the shoot apices for live confocal imaging. (A–E) Dissection of a shoot apex for imaging. Inflorescence before (A) and after (B) removal of siliques and older flowers. (C–D) Shoot apex immersed in water in the dissecting dish, with an air bubble trapped at the tip (C) and after removal of the air bubble (D). (E) Shoot apex in the dissecting dish, after dissection of flower buds older than stage 5. (F–G) View of shoot apices in the imaging dish, on the stage of an upright (F) and inverted (G) confocal microscope. In (F), a 40X water-dipping lens is positioned above one of the apices, with the tip of the lens immersed in water. In (G), the shoot apex is positioned upside down above the 40X water-dipping lens, with a water column connecting the imaging medium to the tip of the lens. The smaller panel in (F) shows a higher magnification view of the area in the red rectangle, with a shoot apex inserted in imaging medium; red and blue lines indicate the surface of the medium and water, respectively. (H–I) Examples of custom-made devices that allow adding more water at the tip of the water-dipping lens: a silicon rubber sleeve made from a spark plug boot (H), and a makeshift sleeve made from the finger of a powderless latex glove (G). Scale bars = 0.5 cm in (A) and (B), 0.1 cm in (C) and (D), and 100 µm in (E). This figure was initially published in reference14. Please click here to view a larger version of this figure.
4. Staining
NOTE: Cell walls or plasma membranes can be stained with propidium iodide or FM4-64, respectively, to achieve cellular resolution during imaging. Alternatively, reporter lines with fluorescent proteins tagged to the plasma membrane can be used6,16.
5. Sepal Ablation
NOTE: At stage 4, sepal primordia start covering the FM. As they grow, they filter out the fluorescence from underlying tissue, and hinder the imaging process. Sepals can either be removed manually using a metal pin mounted on a pin-vise, or prevented to grow using laser ablation on emerging sepal primordia. Alternatively, it is possible in some cases to use mutants such as apetala1-1, in which sepals are missing or replaced by leaf-like organs that do not cover the FM while the central part of the flower develops normally (Figure 2F)17.
6. Imaging Setup
7. Considerations on the Imaging Parameters
NOTE: How to set up the imaging parameters depends a lot on the confocal system used. Below are suggestions for some of these parameters that can be used with any confocal microscope. For more considerations on the imaging parameters, see the Discussion section and reference14.
8. Visualization of Confocal Data
Figure 2 presents different views of confocal Z-stacks of live Arabidopsis flower buds expressing different fluorescent reporter genes, and stained with either propidium iodide (Figure 2, A-C5 and G) or FM4-64 (Figure 2, E1-F) to provide a clear cellular resolution. Most confocal systems allow for the imaging of two fluorophores with non-overlapping emission spectra such as GFP or YFP together with either propidium iodide or FM4-64 (Figure 2A–2F). The best confocal systems are also able to separate multiple fluorophores with partially overlapping emission spectra in the same samples, such as GFP and YFP, and dsRed and propidium iodide (Figure 2G).
Three examples of time-lapse experiments are presented in Figure 2 and show flower buds developing normally after sepal ablation was performed either with a laser ablation system (Figure 2, C1-C5 and D1-D5) or manually with a pin-vise and a metal pin (Figure 2, E1-E2). Note that the flower bud shown in Figure 2, E1-E2 is from a superman-1 mutant plant, which is why carpel primordia are not observed at the center of the bud at stage 7. Laser ablation of emerging sepal primordia of the flower bud shown in Figure 2, C1-C5 prevents them from covering the FM, which can still be imaged at stage 5 (Figure 2, C5). Conversely, in the case of the flower bud shown in Figure 2, D1-D5, laser ablation only delayed sepal growth, but sepals eventually grew to cover most of the FM by stage 5 (Figure 2, D5). On the contrary, if too much laser power and dwelling time is used during laser ablation, damage spreads to the central part of the flower and affects its growth and sometimes results in the subsequent death of the whole flower bud.
Figure 2: Visualization of confocal Z-stacks of live flower buds. A and B2 are transparency views; B1 and B4-G are maximum intensity projections; B3 is an orthogonal slice view. (A–E2) flowers expressing a Venus reporter for the APETALA3 gene (green); cell walls were stained with propidium iodide (red, except for B4, green). (A) Inflorescence; numbers indicate floral stages; sepals in stage 4 and 5 flowers filter out the fluorescence of the Venus reporter, which normally forms a ring. Note that some flower buds appear tilted compared to the inflorescence. (B1–B4) Four different views of the same confocal Z-stack of a stage 5 flower bud: maximum intensity projections (B1 and B4) with Venus signal intensity coded with pixel intensity in the green color (B1) or with fire color code (B4; color code is shown at the bottom of the panel); transparency view (B2) and orthogonal slice view (B3; the XY, XZ and YZ orientation of the slices is indicated in the panel). (C1–C5) 4-day time-lapse of an individual flower bud from stage 3 to stage 5; laser ablations (marked as white traits) performed on day 1 and day 3 were sufficient to prevent the sepals from covering the center of the flower bud at stage 5. (D1–D5) 4-day time-lapse of an individual flower bud from stage 3 to stage 5; laser ablations performed on day 1, 2 and 3 were insufficient to prevent the sepals from covering the center of the flower bud. (E1–E2) Individual flower bud after manual removal of the abaxial and adaxial sepals (E1), and of all sepals (E2); white arrowheads indicate remaining sepals; white asterisks indicate scars resulting from the removal of the sepals. (F) stage 7 apetala1-1 flower expressing a DR5-3xVenusN7 reporter18 (green); plasma membranes were stained with FM4-64 (red); blue arrowheads indicate the leaf-like structures that replace sepals and do not cover the flower bud. (G) Stage 4 flower bud expressing a fluorescent GFP reporter for DORNROSCHEN-LIKE7 (green), a Venus reporter for SUPERMAN (red) and a dsRed reporter for CLAVATA319 (cyan); cells walls were stained with propidium iodide (grey). d: day; st: stage. Scale bars = 25 µm; scale is identical in B1, B2 and B4, in C1-C5 and in D1-D5. This figure was modified from reference14. Please click here to view a larger version of this figure.
Here, we provide a protocol for the imaging of live, developing Arabidopsis flower buds with a confocal microscope and a water-dipping lens. While this can be done with either an upright or an inverted microscope, it is easier and faster to use the former. It is also worth noting that different confocal systems vary significantly in terms of speed, sensitivity, and ability to separate wavelengths. The best confocal microscopes now allow to image several different channels (e.g. GFP, YFP, dsRed and propidium iodide; Figure 2G) in the same samples. Some confocal microscopes are equipped with a spectral detector, which can collect the fluorescence from several fluorophores simultaneously and separate them afterwards. When using a regular detector, however, a set of lasers and filters must be used to excite and separate the fluorescence from different fluorophores. Some fluorophores have fully distinct emission spectra (e.g. CFP and YFP), and can be imaged simultaneously. Other fluorophores have partially overlapping emission spectra (e.g. GFP and YFP), and usually need to be imaged separately, which increases imaging time. Imaging close fluorophores also often requires restricting how much of the spectrum is collected for each channel to prevent fluorescence from one fluorophore from leaking into another channel. This causes a loss of intensity of the collected signal, which can be compensated by an increase in laser power and gain. However, prolonged imaging time and increased laser power may bleach and/or damage the sample. Increased laser power and/or gain may also increase background noise. Optimization of signal intensity, resolution and imaging time for each sample is done through a trial-and-error process, and involves permanent trade-offs.
This protocol explains how to perform live confocal imaging of flower buds developing on the flanks of a dissected shoot apex. It may be argued that the fact that the shoot apex is not connected to the rest of the plant and grows in a medium containing cytokinins might affect SAM and flower development. However, this medium was designed empirically through a trial-and-error process to ensure the dissected shoot apical meristems produce new flower buds at the normal plastochron, and that these flower buds develop normally15. Similarly, sepal dissection does not appear to affect gene expression patterns or the development of the rest of the flower. Other protocols detail how to perform live confocal imaging of the SAM still attached to the rest of the plant (e.g. reference11). Such protocols could be adapted to, and offer a useful alternative for the imaging of developing flowers, particularly when studying cytokinin-related processes. They present several disadvantages however: the stem elongates and twists, and changes the orientation of the flower buds; and while imaging a shoot apex attached to the rest of the plant works well with an upright microscope, it is much more complicated to do so with an inverted microscope.
Immersion lenses have a better numerical aperture (NA) than "dry" lenses (i.e. lenses that are separated from the sample by air), and therefore provide a finer resolution of the sample, which is critical when using a confocal microscope. Using a water-dipping lens thus provides a much better NA than a dry lens, while preventing the shoot apex from dehydrating during the imaging process. While glycerin and oil lenses have an even higher NA than water lenses, they require the use of a coverslip, which is very unpractical with a sample the size of a shoot apex, which also keeps growing during time-lapse experiments. Given the size of the samples (depending on the floral stages, stacks can be over 150 µm thick), it is also important to use a lens with a long working distance. We typically use a 20 or 40X lens with a NA of 1 and a working distance of 1.7-2.5 mm.
Confocal microscope software offers several ways to visualize confocal data, including three-dimensional reconstructions (Figure 2, B1, B2 and B4) and slice views (Figure 2, B3). While the most commonly used three-dimensional views are maximum intensity projections of the confocal Z-stacks (Figure 2, B1 and B4), some software (e.g. ZEN and Imaris) also offer transparency views that filters out the signal from the deeper parts of the sample (Figure 2, B2), which can be useful when looking at processes taking place, or genes expressed in, the epidermal layer. Signal intensity can either be coded with a single color (Figure 2, B1), using pixel intensity, or using a color code (Figure 2, B4).
Live confocal imaging not only offers valuable qualitative insights into flower development, it also potentially provides developmental biologists with a wealth of quantitative data. Different software (e.g. Imaris, FiJi, MARS-ALT, MorphographX15,20,21) allows for the quantification of the number or the volume of cells expressing one or several reporters, or the level of expression of a reporter in different cells. Such software can also be used to perform automatic cell segmentation, track cell lineages and quantify growth. This different software offers similar tools, but also differs in many ways. MARS-ALT and MorphoGraphX were designed specifically for plants15,20,21, unlike Imaris and FiJi. MorphoGraphX only allows for the segmentation and analysis of the epidermal layer, while Imaris and MARS-ALT allow for the segmentation and analysis of the whole sample. However, MARS-ALT requires the prior imaging of the same sample from three different angles15, and Imaris only requires a single stack. Access to quantitative information is critical to further our understanding of flower development.
The authors have nothing to disclose.
The author wish to thank Prof. Elliot M. Meyerowitz for his support, and comments on the manuscript, as well as Ann Lavanway at Dartmouth College and Dr. Andres Collazo at the Biological Imaging Facility at Caltech for their help in solving technical issues with the live confocal imaging. Nathanaël Prunet's work is supported by the US National Institutes of Health through Grant R01 GM104244 to Elliot M. Meyerowitz.
Forceps | Electron Microscopy Sciences | 72701-D | Dumont Style 5 Inox, for apex dissection |
Pin Vise | Ted Pella, Inc. | 13560 | 80 cm long, for sepal ablation |
Straight Stainless Steel Needles | Ted Pella, Inc. | 13561-50 | 0.1 mm Diameter, 1.2 cm long, for sepal ablation |
Round plastic boxes | Electron Microscopy Sciences | 64332 | transparent, 6 cm diameter, 2 cm deep, to use as dissecting dishes |
Rectangular hinged boxes | Durphy Packaging Co. | DG-0730 | transparent, 2-7/8” long, 2” wide, 1-1/4” deep, to use as imaging dishes with an upright microscope |
Tissue Culture Dishes, Polystyrene, Sterile | VWR | 25382-064 | transparent, 3.5 cm diameter, 1 cm deep,to use as imaging dishes with an inverted microscope |
P10 and P1000 pipettes | with corresponding filter tips | ||
Stereomicroscope | with an 80-90 x maximum magnification | ||
Laser ablation system | for sepal ablation | ||
Laser scanning confocal microscope system | |||
20 or 40 x water-dipping lens | with a NA of 1 and a working distance of 1.7-2.5 mm | ||
Agarose | |||
Sucrose | |||
Murashige and Skoog basic salt mixture | Sigma-Aldrich | M5524 | without vitamins |
Myo-inositol | Sigma-Aldrich | I7508 | vitamin |
Nicotinic acid | Sigma-Aldrich | N0761 | vitamin |
Pyridoxine hydrochloride | Sigma-Aldrich | P6280 | vitamin |
Thiamine hydrochloride | Sigma-Aldrich | T1270 | vitamin |
Glycine | Sigma-Aldrich | G8790 | vitamin |
N6-Benzyladenine | Sigma-Aldrich | B3408 | BAP, a cytokinin |
Propidium iodide solution | Sigma-Aldrich | P4864 | 1 mg/mL solution in water, to stain the cell walls |
FM4-64 | Invitrogen | F34653 | dilute in water to 80 μg/mL, to stain the plasma membranes |