2021 Volume 86 Issue 3 Pages 225-233
During the ontogeny of angiospermous pollen, a small generative cell (GC) detaches from the inner pollen wall (intine) and is engulfed within the cytoplasm of the pollen tube cell (PC). This engulfment process is considered to be a type of endocytosis; however, the type is currently unknown. We investigated morphological changes in the cell membrane and cell wall, and the dynamics of cell wall ingredients and cytoskeletal elements, during GC engulfment by the PC in Liriope muscari pollen. Three-dimensional imaging using field-emission scanning electron microscopy revealed that the GC membrane detached from the intine in all directions. Simultaneously, the GC became spherical, and the PC membrane progressed centripetally while covering the GC detachment area. Finally, the GC was released into the PC cytoplasm, and the exfoliated area was completely covered by the PC membrane. Immediately after microspore division, the cell wall between the GC and PC contained methyl-esterified pectin and callose, which subsequently disappeared, thereby thinning the cell wall. During GC engulfment, actin filaments were distributed throughout the PC, particularly in the GC detachment area. Myosin was distributed around the GC and granules immediately after microspore division, until the completion of GC engulfment. Cytochalasin B prevented GC engulfment. These results suggest that cell wall thinning through the degradation of pectin and callose is important in GC engulfment and that actin filaments are involved, similar to their role in endocytosis in animal cells.
In angiosperms, microspores are produced through meiosis of a pollen mother cell and divide asymmetrically to form pollen, which consists of a small generative cell (GC) and a large pollen tube cell (PC). Subsequently, the GC is engulfed by the PC, forming a “cell-within-a-cell” structure (Borg et al. 2009). The generation of this structure is critical for siphonogamy, which refers to reproduction via pollen tubes. The engulfment process is thought to be a type of endocytosis; however, the details of this process have not been clarified to date.
In animal cells, two types of endocytosis can be observed: phagocytosis, which is the engulfment of bacteria and cell debris within large vesicles, and pinocytosis, which refers to the uptake of liquid droplets within vesicles. Phagocytosis is carried out specifically by phagocytes, such as leukocytes, and is an actin-based endocytic mechanism (Castellano et al. 2001, Kumari et al. 2010). It is inhibited upon treatment with an actin polymerization inhibitor, such as cytochalasin D (Ribes et al. 2010). In contrast, most animal cells continually ingest fluid and solutes via pinocytosis, which is classified as follows: macropinocytosis, caveolae-mediated, clathrin-mediated, and clathrin- and caveolin-independent (Mayor and Pagano 2007, Kumari et al. 2010). Clathrin and caveolin are proteins that play major roles in the formation of coated vesicles. In addition, it is known that clathrin forms diverse cage structures (Morris et al. 2019).
In plants, endocytosis by small vesicles plays important roles in various processes, including nutrient absorption, cell wall construction and maintenance, and the protection against infection (Battey et al. 1999, Baluška et al. 2002, Chen et al. 2011). Although the main endocytic mechanism in plants is considered to be clathrin-dependent, clathrin-independent endocytosis is involved in pollen tube growth and glucose uptake into tobacco BY-2 protoplasts (Bandmann et al. 2012, Onelli and Moscatelli 2013). We previously reported that both clathrin-dependent and clathrin-independent endocytosis occur in pollen tube tip growth in Pinus densiflora (Hiratsuka and Terasaka 1996). However, in plants, in which the majority of cells have cell walls, phagocytosis-like cell engulfment, in which a cell engulfs another cell, has not been reported, except for GC engulfment by the PC. To clarify the engulfment of the GC, Shalag and Hesse (1992) observed chemically fixed Polystatia pubescens pollen using electron microscopy. They reported that the GC gradually separated from the intine, which is the inner wall of the pollen, that the separation region was covered with PC membranes, and that the GC was eventually taken up inside the PC. Zonia et al. (1999) observed Nicotiana tabacum pollen using a fluorescent staining technique. They reported that actin filaments are asymmetrically localized in a ring around the GC and that application of an actin polymerization inhibitor inhibited actin assembly, failing the GC to migrate from the generative pole, mispositioning of the PC nucleus within the cell, and both nuclei having aberrant structures. In addition, they suggested that microtubules, a type of cytoskeleton, are absent during this stage of pollen development, and exposure to oryzalin, a microtubule polymerization inhibitor, had essentially no effect on development. Zhang et al. (2005) reported that in abortive pollen of Oryza sativa, the GC does not migrate or migrates abnormally into the PC.
Between the GC membrane and the PC membrane, there is a cell wall, which has been termed “G-P wall” herein. While it is generally accepted that the callose present in the G-P wall disappears while the GC is still attached to the intine, it remains unclear whether the loss of callose and/or replacement of callose with other wall materials is necessary for GC engulfment (Shalag and Hesse 1992). Further, methyl-esterified and unesterified pectins are distributed in the cell wall of plants, but there are few reports on their presence in the G-P wall (Castro et al. 2013).
In this study, we aimed to clarify the micro-morphological changes in the G-P wall and the dynamics of G-P wall components during pollen development. To this end, we prepared samples by high-pressure freeze-fixation, used field emission scanning electron microscopy (FE-SEM; S-8220, Hitachi, Tokyo, Japan) for observation and 3D reconstruction, and we evaluated the distribution of cell wall components by immunohistochemical staining.
Pollen grains were collected from local Liriope muscari (Decne.) L.H. Bailey plants, between September and October. The stage of pollen development was identified by light microscopy in each experiment.
Light and fluorescence microscopyTo visualize nuclei in pollen grains, the anthers were fixed in Carnoy’s fluid (ethyl alcohol: chloroform: acetic acid=2 : 1 : 1) at 4°C for at least 1 h. Then, the nuclei were stained with 1% aceto-carmine.
To visualize β-1,3-glucan (callose), methyl-esterified pectin, unesterified pectin, and myosin, samples were fixed in 4% paraformaldehyde in phosphate buffer (PB) at 4°C overnight, dehydrated in a graded ethanol series, and embedded in Technovit 7100 resin (Heraeus Kulzer, Wehrheim, Germany). One-micrometer sections of the specimens were placed on a glass slide and blocked with 1% bovine serum albumin in phosphate-buffered saline (PBS) for 1 h. Primary antibody staining was performed at 4°C overnight, whereas secondary antibody staining was performed at room temperature (24–26°C) for 1 h. A monoclonal antibody (mAb) for β-1,3-glucan was obtained from Biosupplies Australia (Parkville, Australia); unesterified pectin (JIM5) and methyl-esterified pectin (JIM7) mAbs were obtained from PlantProbes (Leeds, UK), and antiserum against lily pollen 170-kDa myosin was a generous gift from Dr. E. Yokota (University of Hyogo, Japan).
To visualize actin, samples were fixed with 4% paraformaldehyde in 0.1 M PB, pH 7.2, at 4°C for 1 h and then incubated with rhodamine-phalloidin (Invitrogen, CA, USA) diluted 1 : 40 with PBS. Fluorescence was observed using an LSM 510 laser scanning microscope (Carl Zeiss, Oberkochen, Germany).
Electron microscopy (EM)We used several EM sample preparation methods in this study. The high-pressure freezing (HPF) method is effective in conserving the intracellular structure. The specimens were frozen using a HPF apparatus (HPM 010, BAL-TEC AG, Liechtenstein) and transferred into acetone containing 2% OsO4 for freeze-substitution, which was carried out at −80°C for 4 days, at −40°C for 2 h, and −20°C for 2 h. Then, the specimens were brought up to room temperature and embedded in Spurr’s resin (Polysciences, Warrington, PA, USA).
For conventional chemical fixation, samples were fixed with 2.5% glutaraldehyde (Electron Microscopy Sciences, USA) in 0.1 M PB (pH 7.2) at 4°C overnight, rinsed in PB, post-fixed in 2% buffered OsO4 at room temperature for 1 h, dehydrated in an ethanol-acetone series, and embedded in Spurr’s resin at 70°C for 24 h. Then, the samples were sectioned and stained with aqueous uranyl acetate and lead stain solution. Electron micrographs of the stained sections were obtained via FE-SEM, using a high-sensitivity backscattering electron detector, or via transmission electron microscopy (TEM), using a JEM1400 transmission electron microscope (JEOL, Tokyo, Japan).
For immunoelectron microscopy, samples were fixed with 4% paraformaldehyde and 0.5% glutaraldehyde in PB at 4°C overnight, rinsed in PB, dehydrated in an ethanol series, and embedded in LR White Resin (London Resin, Reading, UK) at 55°C for 24 h. Sample sections were stained with primary and colloidal gold-labeled secondary antibodies, using a conventional method. Negative controls, in which the first antibody was omitted, were included.
For the correlative light EM method, Technovit sections stained with a mAb for β-1,3-glucan were observed and imaged under a fluorescence microscope. The same sections were then stained with uranyl acetate and lead stain solution and observed and imaged with FE-SEM. The images were overlaid using Adobe Photoshop CS6 (Adobe Systems Inc., San Jose, CA, USA).
Cytochalasin B treatmentFlower buds were incubated in water containing 10 µM cytochalasin B (Sigma-Aldrich, Taufkirchen, Germany) at room temperature for 24 h.
3D reconstructionDigital FE-SEM images were traced using Adobe Photoshop. 3D images were reconstructed using the TRI/3D SRF III software (Ratoc System Engineering, Tokyo, Japan). Cell volumes and areas where the GC and intine are in contact were calculated using the TRI/3D SRF III software.
In L. muscari, the GC is cut off as a lens-shaped small cell at one end of the pollen through asymmetric cell division of the microspore (Fig. 1a–d). The GC is in contact with the PC on the “internal side” and with the intine on the “external side” (Fig. 1d). Subsequently, the GC becomes spherical, decreasing the contact surface with the intine, and protrudes into the PC (Fig. 1e). Eventually, it completely detaches from the intine and is liberated into the cytoplasm of the PC (Fig. 1f). The GC then takes on a spindle shape.
To clarify the process of GC engulfment by the PC, HPF-fixed samples were observed using FE-SEM. Immediately after microspore division (corresponding to Fig. 1c, d), an ellipsoidal nucleus appears in the lens-shaped GC (Fig. 2a). The G-P wall between the GC and PC membranes is approximately 250 nm thick (arrowheads) and contains substances with high electron density (Fig. 2b arrow). In Fig. 2b, the original contact site of G-P wall with the intine is indicated by an asterisk. Figure 2c and 2e show the pollen at the initial stage of GC engulfment. The separation of the GC membrane from the intine starts at the location indicated by the asterisk and proceeds to the location indicated by the large black arrow (Fig. 2e). To cover the exfoliation area, the membrane of the PC develops centripetally from the original contact site (Fig. 2e asterisk and long arrow). The G-P wall thins to less than half (＜120 nm) of its original thickness, as shown in Fig. 2b, and becomes highly electron-dense (Fig. 2e arrowheads). Between the developing PC membrane and the intine, a layer of approximately 200 nm thickness containing highly electron-dense granules (P-I layer) is newly formed (Fig. 2e white arrow). Numerous microtubules are observed in the vicinity of the developing PC membrane (small arrows). Figure 2d and f show pollen in which GC engulfment has further progressed. As the GC takes on a spherical shape and protrudes into the PC, the contact area with the intine declines. The PC membrane and P-I layer further develop centripetally along the intine (Fig. 2f arrow). The G-P wall becomes even thinner (50 nm; Fig. 2f arrowheads). The GC further protrudes into the PC (Fig. 2g) and only slightly contacts the intine at this point (Fig. 2h arrow). Ultimately, the GC completely detaches from the intine and is released into the PC cytoplasm (Fig. 2i). In this stage, the electron density of the G-P wall is high, and the GC and PC membranes become indistinguishable.
Next, we performed TEM of HPF samples (Fig. 2j) and FE-SEM of chemically fixated samples (Fig. 2k). In these images, the electron density of the G-P wall is low, and the GC and PC cell membranes are observed. On the other hand, the G-P wall in Fig. 2j shows a uniform thickness, and the G-P wall in Fig. 2k has a non-uniform thickness. A clathrin-coated PC membrane during GC engulfment was not observed with any of the methods.
3D construction of the GCTo visualize the 3D shape of the GC, 56 consecutive 0.5-µm-thick sections were prepared and photographed using FE-SEM. Figure 3a shows a 3D structural image that was constructed by combining all sections. PC cytoplasm (green), PC nucleus (blue), GC cytoplasm (yellow), and GC nucleus (purple) were color-coded for pollen 1 and 2 in Fig. 3a. Figure 3b shows pollen 1 at the beginning of the GC engulfment process (corresponding to Fig. 1c), and Fig. 3c is an image taken in the direction of the arrow in Fig. 3b. The entire volume of pollen 1 was 1.1×104 µm3, the volume of the GC was 6.7×102 µm3, and the volume ratio of the GC to the PC was approximately 1 : 15. The GC had a lens-like shape, with a somewhat elliptical bottom surface area of 1.7×102 µm2. Figure 3d shows pollen 2, in which GC engulfment has progressed (corresponding to Fig. 1e), and Fig. 4e is an image taken in the direction of the arrow in Fig. 3d. The entire volume of pollen 2 was 1.1×104 µm3, the volume of the GC was 7.3×102 µm3, and the volume ratio of the GC to the PC was approximately 1 : 14. The GC had a spherical shape, with a circular bottom surface area of 1.3×102 µm2, indicating that GC detachment from the intine had progressed further than that in pollen 1.
We immunolabelled callose with an anti-callose antibody. Immediately after microspore division, the fluorescence was distributed in a lens shape between the GC and the PC (Fig. 4a). In superimposed fluorescence and FE-SEM images of the same cell, the fluorescence overlapped with the G-P wall (Fig. 4b). In the no-primary-antibody negative controls, no signal was detected (data not shown). Immunoelectron microscopy revealed that callose was distributed only on the G-P wall (Fig. 4c orange), not the intine (Fig. 4c green). During GC engulfment, the callose gradually disappeared from the pollen (Fig. 4d).
Immediately after microspore division, fluorescence-labeled methyl-esterified pectin was observed in the G-P wall and intine (Fig. 4e). During GC engulfment, the fluorescence signal began to disappear starting near the “internal side” vertex (Fig. 4f, g), and at the completion of engulfment, it had completely disappeared from the “internal side” (Fig. 4h). On the other hand, methyl-esterified pectin was present on the intine throughout the development of the GC. To investigate the distribution of methyl-esterified pectin in detail, pollen immediately after microspore division was observed using immunoelectron microscopy. Methyl-esterified pectin was localized in the G-P wall (orange) and intine (green), and it was hardly detected in the cytoplasm of both cells (Fig. 4i). Immunofluorescence analysis did not reveal the presence of demethylated esterified pectin throughout pollen development (data not shown).
Dynamics of the actin-myosin system during GC engulfmentThe distribution of actin fibers (Fig. 5a, b) and 170-kD myosin (Fig. 5c, d) in the pollen was observed using immunofluorescence staining. Figure 5a and 5c show pollen during GC engulfment; Fig. 5b and 5d show pollen after the completion of GC engulfment. Actin fibers were distributed throughout the PC after microspore division. During GC engulfment, actin fibers were distributed particularly around the region where the GC was separating from the intine (arrows in Fig. 5a), and the PC nucleus. After the completion of GC engulfment, actin fibers surrounded the GC and the PC nucleus. Myosin started being distributed around the GC and granules in the PC immediately during pollen development (Fig. 5c, d). To elucidate the involvement of actin in the engulfment process, pollen sampled immediately after microspore division were treated with cytochalasin B, which is an actin polymerization inhibitor, for 24 h. In the absence of treatment (n=180), 80.3% of pollen completed GC engulfment, whereas only 30.9% of pollen could complete engulfment in pollen treated with 10 µM cytochalasin B by this time point (n=180). TEM observation of the pollen treated with cytochalasin B revealed that the outline of the GC was wavy (Fig. 5e arrows) and the development of the PC membrane along the intine was inhibited (Fig. 5f arrow).
Electron microscopic images of HPF samples showed that the cell membranes of the GC and PC were smooth, and the samples seemed fixated in a better state than those in previous reports using chemical fixation (Shalag and Hesse 1992, Dinis and Mesquita 1999). Cresti et al. (1987) observed the GC of mature tobacco pollen after rapid freeze fixation and reported that the two cell membranes encasing the GC looked substantially smoother than after chemical fixation. Cryofixation, which is excellent for preserving membrane morphology, is particularly useful in the analysis of membrane dynamics. Disadvantages of this method include the following: only a narrow region is in an optimally frozen state and the resin does not easily penetrate the sample. Therefore, it is not suitable for creating ultra-thin serial sections for TEM observation and 3D analysis. In this study, serial semi-thin sections were prepared and attached to glass slides, followed by FE-SEM observation and 3D reconstruction. During this approach, the electron beam causes hardly any cracks and distortions in the section. FE-SEM mainly detects two types of electrons. Backscattered electrons (BSE) originate from the deeper regions of the sample, and secondary electrons originate from the surface region. In this study, electron micrographs of stained sections were obtained by FE-SEM using a high-sensitivity backscattering electron detector. The amount of reflected electrons generated depends on the material composition of the sample.
After microspore division, the GC was lens-shaped. Then, it gradually became spherical and protruded into the PC, and finally, it was engulfed within the PC. GC engulfment by the PC was initiated by the separation of the external side of the GC from the intine, and the exposed area of the intine was then covered by the developing PC membrane. Shalag and Hesse (1992) reported that in P. pubescens, GC detachment is initiated with the loss of the lenticular shape of the GC. At this stage, the central part of the GC, with a spherical nucleus, protrudes into the PC, while the peripheral part remains thin. In L. muscari, the GC and its nucleus remained lenticular at the beginning of GC engulfment, and the oval nucleus was located near the intine. The 3D images suggested that GC detachment occurred from all directions, and the PC membrane covered the detachment site, like a closing iris diaphragm, while GC engulfment proceeded. Substances with high electron density were observed in the G-P wall, and between the PC membrane and the intine, immediately after microspore division. Although the identity of these substances is unknown, Shalag and Hesse (1992) and Dinis and Mesquita (1999) observed similar structures in P. pubescens and Magnolia×soulangeana, respectively. Based on these findings, it is suggested that these substances widely exist in the G-P wall across species. However, it is not clear whether they are involved in GC engulfment. In the next stage, the PC membrane completely enclosed the detached GC, forming a “cell-within-a-cell” structure. Asymmetric microspore division is a prerequisite for “cell-within-a-cell” structure formation. When the volumes were measured using a 3D image, the volume ratio of the GC to the PC was 1 : 14 to 1 : 15, corroborating that microspore division is extremely asymmetrical. Future studies are needed to clarify the GC-to-PC volume ratio required for the formation of a “cell-within-a-cell” structure in other species.
Structure and dynamics of the G-P wallThe G-P wall, which separates the GC and PC, is formed through microspore division. In L. muscari, callose and methyl-esterified pectin were distributed on the G-P wall immediately after microspore division and disappeared from the G-P wall before completion of the “cell-within-a-cell” structure. In addition, the G-P wall became thinner as GC engulfment progressed. Borg et al. (2009) suggested that callose degeneration in the G-P wall is involved in the GC engulfment process. Using aniline blue staining, Shalag and Hesse (1992) also observed that the callose distributed on the G-P wall immediately after microspore division, disappeared before the GC started to leave its original position. They reportedly could not distinguish the pecto-cellulosic wall by using the periodic acid-thiocarbohydrazide-silver protein method or alcian blue staining. Dinis and Mesquita (1999) reported that the newly formed GC of Magnolia×soulangeana has a wall consisting of callose and other polysaccharides that disappear in later stages, presumably through resorption of the wall material. They also suggested that this wall modification facilitates the deformation of the GC surface and, consequently, GC migration to a deeper position in the PC cytoplasm, along with shape alterations from lenticular over spherical to a spindle. In Larix decidua, both unesterified and methyl-esterified pectins have been observed in the walls between the GC and the sterile cell, and between the prothallial and sterile cells, in mature pollen before pollination, whereas after pollination, they disappeared before the GC was engulfed by the PC (Rafińska et al. 2014). In our study, as GC engulfment progressed, the G-P wall became thinner. It appears that this indicates the disappearance of wall components such as callose and pectin. By the time the GC was engulfed by the PC, the wall thickness of the GC had decreased by approximately one-fifth of the thickness immediately after wall formation. This suggests that callose and pectin were not replaced with other substances. As Dinis and Mesquita (1999) pointed out, the loss of callose and methyl-esterified pectin between the GC and PC, and the accompanying increase in membrane motility may facilitate GC engulfment by the PC and the morphological changes of the GC from the lens- to spherical-shaped. FE-SEM observation of the HPF samples revealed relatively strong BSE signals from the thinned G-P wall, suggesting the presence of substances other than callose and pectin in the G-P wall. In the chemically fixed samples, the level of BSE in the G-P wall was low. This may be due to the loss of wall components in the post-chemical fixation process. Future studies will be needed to identify the components of the high-electron-dense material and clarify their function in the thinned G-P wall.
Role of the cytoskeleton in GC engulfment by the PCActin fibers play important roles in the progression of phagocytosis and macropinocytosis in animal cells (Mayor and Pagano 2007, Kumari et al. 2010). In addition, various types of myosin work together with the actin fibers. For example, myosin X is required for the development and closure of pseudopodia necessary for the uptake of foreign substances in phagocytosis (Castellano et al. 2001). In L. muscari, actin fibers were widely distributed near the developing PC membrane during GC engulfment, as well as around the GC and the PC nucleus. As 170-kD myosin was distributed around the GC and on the intracellular granules of the PC, it is speculated that actin and myosin work together and are involved in the transport of cell membrane components that results in the migration of the GC into the PC cytoplasm. Upon cytochalasin B treatment, the GC membrane became wavy and the adhesion of the PC membrane to the inner wall was inhibited. This suggests that actin fibrils are involved in controlling the direction of PC membrane development during GC engulfment. Using light microscopy, Zonia et al. (1999) observed that treatment of tobacco pollen with cytochalasin D or latrunculin A inhibited the migration of the GC into the PC. Our study was the first to analyze the effect of actin inhibition on GC engulfment by electron microscopy. Similar to animal phagocytosis, GC engulfment by the PC involves actin fibers. Furthermore, our findings suggested for the first time that a dynamically developed PC membrane surrounds the GC and that clathrin is not involved in this process. In this study, numerous microtubules were observed around the PC membrane that was developing to enclose the GC; however, their role is unknown. Zonia et al. (1999) reported that microtubule inhibitors did not affect GC migration in N. tabacum.
In conclusion, GC engulfment by the PC in pollen of L. muscari was found to be a clathrin-independent form of phagocytosis unique to plants that is controlled by G-P wall thinning, and GC softening due to the loss of callose and pectin, and regulation of PC membrane development by the actin-myosin system.
We thank Dr. Tomoko Suzuki (Japan Women’s University), Ms. Nobuko Moritoki (Keio University School of Medicine), and Ms. Ayako Takahashi (Japan Women’s University) for technical assistance during electron microscopy imaging. We also thank Dr. Etsuo Yokota (University of Hyogo) for generously providing antiserum against lily pollen 170-kD myosin and Ms. Yohko Yamada (Japan Women’s University) for useful discussions. We would like to thank Editage (www.editage.com) for English language editing. This work was supported by The Jikei University Research Fund.
