Invited Review
Quantitative Imaging of Pavement Cell Morphogenesis and Stomatal Differentiation
2020 Volume 85 Issue 2 Pages 101-106
Details
2020 Volume 85 Issue 2 Pages 101-106
Leaf epidermal tissue senses various environmental factors, such as light and pathogens, and is the site of plant responses to environmental changes. The epidermis is composed of stomata distributed at regular intervals and pavement cells shaped like jigsaw puzzle pieces. These characteristic forms are thought to be the basis of the plant’s excellent ability to respond to the environment. I have investigated the mechanisms of cell morphogenesis and distribution to understand the construction of the epidermal tissue at the cellular level. This review describes the quantitative analyses of cortical microtubules and membrane trafficking involved in the morphogenetic mechanisms of pavement cells and an induction system for clustered stomata that involves a sugar solution-based immersion treatment.
Leaves, photosynthetic organs important for plant growth, are covered by a single layer of planarly spreading epidermal tissue. Epidermal tissue is composed of guard cells that constitute stomata, which perform gas exchange and transpiration in photosynthesis, and pavement cells that fill the flat surface. The two cell types are formed as protodermal cells during the early stage of embryonic development and differentiate into guard cells or pavement cells as individuals mature. Protodermal cells have a simple shape, and immediately after germination, pavement cells in cotyledons have low complexity. Pavement cells form multiple lobe regions during leaf maturation, increasing the cell area and complexity (Panteris et al. 1993, Geitmann and Ortega 2009, Akita et al. 2015). Each pavement cell does not form this shape independently, but the cells are formed so that lobe and intended regions are intricately adjoined between adjacent pavement cells. Using microtubule-associated mutants, molecular genetic studies have shown that Rho of plants (ROPs), a plant-specific Rho GTPase, and ROP-interactive CRIB motif-containing proteins, which interact with ROPs, work together through cytoskeleton control to form cell curvatures (Fu et al. 2005, 2009).
To examine the relationship between pavement cell morphology and cortical microtubule orientation, we used our original image analysis techniques to obtain microscopic images of microtubules visualized by fluorescent proteins and quantitatively analyzed microtubule orientation (Fig. 1) (Akita et al. 2015).
Microscopic image analysis techniques allowed us to quantitatively evaluate cytoskeleton orientation (Yoneda et al. 2007, 2010, Higaki et al. 2010, Higaki 2017), and we modified the method to quantitatively analyze the relationship between the shape of the pavement cell and the orientation of the cortical microtubules. We acquired confocal images of cotyledon epidermal cells of the green fluorescent protein (GFP)-fused TUB6 (GFP-TUB6)-expression line growing 3–12 d after germination (DAG), and then, we quantified the cell shape and cortical microtubule orientation using image processing procedures as described in Fig. 1 (Akita et al. 2015). The maximum intensity projection image was produced from the serial optical section image of the cortical microtubule, which was three-dimensionally imaged in the depth direction (Fig. 1; Maximum intensity projection), and then skeletonized (Fig. 1; Extracted microtubules). As a cell-shape evaluation index, the cell medial axis was determined (Fig. 1; Cell medial axis) by skeletonization using the binarized cell region image (Fig. 1; Cell region image). By improving the method of a previous study targeting the actin filaments of guard cells (Higaki et al. 2010), the angles of cortical microtubules related to the nearest cell medial axis were measured, and the average value was calculated for each cell. As a result, because the cell areas of the 3–5 DAG young cotyledon pavement cells were small, edge effects (Ambrose et al. 2011) were likely to occur, and the cortical microtubules were parallel to the cell medial axis. Mature pavement cells at 11–12 DAG are arranged in small areas, and microtubule self-organization (Ambrose and Wasteneys 2012) appears to have occurred locally, but the relationship with the cell medial axis was reduced.
The important role of cortical microtubules in pavement cell morphogenesis is supported by inhibitor treatment experiments performed during the seedling stage. In cotyledons sown in solutions containing propyzamide, colchicine or oryzalin, which are polymerization inhibitors of tubulin proteins that constitute microtubules, and taxol, an inhibitor of tubulin depolymerization, the pavement cell area was not affected, but cell complexity was reduced (Akita et al. 2015). Both tubulin polymerization and depolymerization, which are responsible for the dynamic instability of microtubules, are important for the morphogenesis of pavement cells (Akita et al. 2015). When using the genetic expression of guard cells in the GAL4 GFP enhancer trap line E1728, which expresses mature guard cell-specific GFP fused with an endoplasmic reticulum-retention signal (Gardner et al. 2009), circular cells were observed in the cotyledons treated with colchicine and oryzalin (Fig. 2, Akita et al. 2015). Also, microtubules were determined to be required for the normal differentiation of stomatal lineage cells into guard cells (Akita et al. 2015).
The morphogenesis of pavement cells requires a volume increase accompanied by the synthesis of new cell walls. Membrane traffic is the mechanism of intracellular transport in eukaryotes and contributes to protein transport between organelles through vesicles, secretion into the extracellular space, and the expansion of the cell membrane. In particular, plant cells require the extracellular transport of cell wall components as well as cell membrane components for cell volume expansion. When we observed mature leaf pavement cells expressing monomeric red fluorescent protein (mRFP) fused with rat sialyltransferase (ST-mRFP; Boevink et al. 1998, Stefano et al. 2012), which is a marker for the trans-Golgi cisternal lumen, the mRFP fluorescence localized in the apoplast (Fig. 3), and normal dot-like Golgi body localizations were also observed. The apoplastic localization of mRFP was confirmed by fluorescence observations of the apoplast extract (Akita et al. 2016). The apoplastic mRFP localization suggested that ST-mRFP was delivered to the plasma membrane through an exocytic pathway, and the plasma membrane-localized ST-mRFP was digested by extracellular proteases (Fig. 4a, b). The curvature of the pavement cell contour and the fluorescence intensity of mRFP were quantitatively examined, and a strong correlation was found between them (Akita et al. 2016). Because the molecular mass of mRFP is approximately 30 kDa, the diameter of mRFP can be roughly assumed to be approximately 5 nm. The diffusion rate of mRFP in the apoplast may be low because the cell wall mesh size is generally 3–4 nm (Lodish et al. 2000). Therefore, our observations suggested a local exocytic pathway to the concaved regions (Fig. 4). Thus, the local exocytosis observed in root hairs and pollen may also occur during the morphogenesis of pavement cells (Akita et al. 2016). An increase in mRFP fluorescence intensity correlated with the curvature was observed at the curved region formed between the two adjacent pavement cells, but not at the three-way junction formed by three pavement cells. The pavement cell curved-regions’ formation process appears to differ between two cells and three cells (Higaki et al. 2016).
To capture the ultra-microstructures involved in membrane traffic to the apoplastic region that was suggested by live imaging, leaves at various stages of maturation were observed using transmission electron microscopy. In pavement cells and the ends of the guard cells, vesicular structures with 30–100-nm diameters were observed in the apoplastic region surrounded by the characteristic invaginations of the plasma membrane which had 200–800-nm diameters and called paramural bodies (PMBs), in chemically and freeze-fixed samples (Akita et al. 2016). PMBs have also been observed in rice during the defense responses to powdery mildew (An et al. 2006). The PMBs may result from multivesicular bodies (MVBs) fusing with the cell membrane during exocytosis (Marchant et al. 1967, An et al. 2006, Samuel et al. 2015). Therefore, we performed immunoelectron microscopy using an anti-mRFP antibody. It revealed that in the same mature plant, not only the apoplastic region but also the electron-dense parts of the cell wall were labeled in 2–3-mm young leaves and in 8–10-mm mid-sized leaves (Akita et al. 2016). These observations suggested that there might be a pathway to transport cytoplasmic components to the apoplastic region by plasma membrane delivery of MVBs during pavement cell morphogenesis (Fig. 4c) (Akita et al. 2016).
Sucrose solution-based immersion system to disturb the stomatal distributionThe electron microscopy did not clarify the pathway by which ST-mRFP was transported to the apoplastic region; however, serendipitously, we attempted to use freeze fixation to preserve the structures of MVBs and PMBs for immunoelectron microscopy. If the sample contained a lot of water, such as vacuoles in plant cells, rapid freezing can lead to the formation of iced nuclei and the destruction of intracellular structures. Thus, as a countermeasure, the sample is pretreated by immersion in a sucrose solution immediately before freezing to reduce the vacuole volume by osmotic pressure. However, I misunderstood that the water immersion in the sucrose solution was performed at the seed stage and sowed the sterilized seeds in the sucrose solution. Microscopic observations to determine if the cotyledon vacuoles germinating in the sucrose solution had contracted revealed unexpected changes. Adjacent guard cells formed stomatal clusters (Fig. 5). In the leaves of Arabidopsis thaliana, stomata are usually separated from each other, and guard cells are separated from each other by pavement cells. To form such a spatial arrangement pattern, during the stepwise differentiation of stomatal lineage cells into guard cells, the former inhibits the differentiation of neighboring cells into guard cells (Bergmann and Sack 2007, Pillitteri and Torii 2012). Although mutants of genes related to stomatal distribution are disturbed during the process and form stomatal clusters (Yang and Sack 1995), the effects of environmental conditions on the stomatal distribution are still unknown. Therefore, we investigated the mechanism responsible for the disruption of the stomatal distribution using the sucrose-immersion treatment.
The stomatal distribution was disrupted even with a 1% sucrose solution and germination did not occur with a 10% sucrose solution. Although stomata clustered when immersed in a glucose or fructose solution, the stomatal distribution was maintained at the corresponding concentration of a mannitol solution (Akita et al. 2013). Immersion in a sucrose solution increased the stomatal density but leaves grown on the same 3% sucrose solid medium showed no stomatal distribution effects, although the number of stomata increased. When the GAL4 GFP enhancer trap line E1627, which specifically expresses ER-tagged GFP in stomatal lineage cells (Gardner et al. 2009), was germinated in a 3% sucrose solution, GFP fluorescence was also detected in the jigsaw puzzle piece-shaped cells adjacent to stomata. This suggests that the genes thought to be expressed only in stomatal lineage cells are expressed in other cells after immersion in the sucrose solution (Akita et al. 2013). Furthermore, aniline blue staining showed that the amount of callose deposition on the cell plate partitioning meristemoids tended to be reduced by the sucrose treatment. Thus, the cause of stomatal clustering in response to sucrose-solution immersion is not the abnormal stomatal differentiation frequency but abnormal asymmetric division during stomatal differentiation (Fig. 6).
The immersion system is a simple method of seeding sterilized seeds in a 24-well plate with 1.5 mL of solution (Akita and Hasezawa 2014, Akita and Higaki 2019). In this method, Arabidopsis cotyledons germinate in the sucrose solution, but the first true leaves develop above the water surface. To develop the true leaves in the sucrose solution, a piece of small wire mesh is sterilized in an autoclave and placed into a well together with seeds; consequently, the germinated seedlings can remain in the aqueous solution (Akita et al. 2013). As a high-efficiency experimental system to induce stomatal cluster formation by sucrose-solution immersion, we examined whether guard cells with a disturbed distribution have stomatal opening capabilities, which is a functional feature of guard cells (Akita et al. 2018). The guard cell performs stomatal opening and closing by increasing and decreasing the cell volume, respectively, by transferring water using the difference in osmotic pressure owing to an ion gradient with the adjacent pavement cell. The radial orientation of the cortical microtubules of a guard cell is thought to be important in opening the stomata between guard cell pairs as the cell volume increases. Using GFP-TUB6-expressing strains, we measured the mean angular differences of the cortical microtubules against the stomata (Akita et al. 2018). The sucrose solution-based immersion treatment showed a slightly more disordered orientation than the control treatment, but the radial tendency was maintained (Akita et al. 2018). Furthermore, observations of GFP-EB1-expressing strains labeled with microtubule extension ends revealed that the cortical microtubules did not lose their polarity of ventral to lateral extension in guard cells (Akita et al. 2018). The fungal phytotoxin fusicoccin was used for stomatal opening induction. In leaves treated with a 3% sucrose solution, both spaced and clustered stomata opened in response to fusicoccin (Akita et al. 2018).
Future perspectivesI observed epidermal tissue cells from the horizontal direction. In recent studies, vertical analyses have also been regarded as important, and the anticlinal walls of pavement cells suggest that cortical microtubules, which were considered to be unevenly distributed in the intended region, are also present in the lobe region (Zhang et al. 2011, Belteton et al. 2018). Cell morphogenesis may result in interactions with neighboring cells, and recent approaches using image analyses and simulations have been actively performed (Higaki et al. 2016, 2017). A dynamic simulation, in which the actual pavement cells are reflected in the initial values, indicated that the cause of the pavement cell curvature in the RIC1-deficient mutant might be compression (Higaki et al. 2017). Analyses using a reaction-diffusion equation revealed an unexpected common feature in which the formation of suture lines in the skull could be described by an equation similar to that used for plant pavement cells (Higaki et al. 2016). In the future, by expanding the direction and scale of quantitative microscopic cell imaging, I would like to increase our knowledge of the cells constituting the epidermal tissue.
I am deeply grateful to Dr. Takumi Higaki (Kumamoto University) for his careful guidance and extensive discussion on the direction of the research. I also thank Prof. Seiichiro Hasezawa (Hosei University) for his encouragement and generous support and Prof. Noriko Nagata (Japan Woman’s University) for her helpful advice, especially on electron microscopy. I would like to thank the members of Prof. Seiichiro Hasezawa’s laboratory at The University of Tokyo and the members of Noriko Nagata’s laboratory at Japan Woman’s University for their enormous help. This work was supported by grants from the Japan Society for the Promotion of Science (JSPS) KAKENHI grant number 26891006 and from the Nakatsuji Foresight Foundation Research Grant. We thank Lesley Benyon, PhD, from Edanz Group (www.edanzediting.com/ac) for editing a draft of this manuscript.
