During self-pollination, binding of a pollen component, S-LOCUS PROTEIN11 (SP11)/S-LOCUS CYS-RICH PROTEIN (SCR) determinant (Schopfer et al., 1999; Takayama et al., 2000; Shiba et al., 2001), to a stigmatic recognition component, S-RECEPTOR KINASE (SRK; Takasaki et al., 2000; Silva et al., 2001), induces SRK autophosphorylation and generation of a signaling cascade leading to germination arrest (Takayama et al., 2001; Takayama and Isogai, 2005). Additional components involved in the signaling cascade leading to inhibition of self-fertilization include ARC1, an arm repeat-containing protein with E3 ubiquitin ligase activity (Stone et al., 1999, 2003), and the membrane-anchored M-locus protein kinase (Murase et al., 2004). However, the mechanism used by the papilla cell to block self-pollen germination and to facilitate cross-pollen germination remains unclear.

During compatible pollination, the papilla cell is thought to actively transport water, calcium, and boron, factors essential for normal pollen germination, into a pollen grain. Furthermore, pollen tube penetration and elongation into the papilla cell wall likely leads to cell wall loosening or reorganization. However, the mechanisms controlling the provision of water, calcium, and boron to a pollen tube and papilla cell wall reorganization are unclear. A previous electron microscopic study showed that many secretory vesicles were seen at the site of pollen attachment in papilla cells during cross-pollination (Elleman and Dickinson, 1996). A similar phenomenon was seen at the site of infection during plant cell-pathogen interactions and this appears to be regulated by the actin cytoskeleton (Kobayashi et al., 1992, 1994; Skalamera and Heath, 1998; Takemoto et al., 2003; Takemoto and Hardham, 2004). Furthermore, polymerization and depolymerization of actin filaments are involved in tip growth of pollen tubes and root hairs, cytoplasmic streaming, cellular morphogenesis, and cell division (Miller et al., 1999; Staiger, 2000; Hepler et al. 2001; Wasteneys and Galway, 2003; Shimmen and Yokota, 2004). Additionally, the actin cytoskeleton undergoes reorganization in response to external stimuli, such as gravity and hormones (Kandasamy et al., 2001; Sun et al., 2004). Therefore, the physiological changes seen in the papilla cell may be directly or indirectly related to changes in actin cytoskeleton dynamics during pollen germination.

To test this hypothesis, we examined whether the actin cytoskeleton in the papilla cell plays an important role in pollen germination during self- and cross-pollination. We visualized and compared the configurations of actin filaments in stigmatic papilla cells before and after self- and cross-pollination by staining actin filaments with rhodamine-phalloidin and by the transiently expressing GFP fused to the actin-binding domain of mouse talin (mTalin) in a papilla cell. Furthermore, we examined the effects of the cross- and self-pollen coat on actin cytoskeletal structure, as well as the effects of cytochalasin D (CD), an actin-depolymerizing drug, on actin cytoskeletal structure in cross- and self-pollinated papilla cells. Finally, we compared the three-dimensional (3D) configurations of vacuolar structures associated with actin filaments in papilla cells before and after self- and cross-pollination by tomography using ultra-high-voltage electron microscopy (ultra-HVEM).

Figure 1. Cross-pollination induces actin cytoskeleton rearrangements, whereas self-pollination disrupts actin structure. Stacked images of rhodamine-phalloidin stained papilla cells before pollination (A), 30 min (B), or 1 h (C) after cross-pollination are shown. (more ...)

To examine whether the organization of actin filaments changed during cross- and self-pollination, papillar actin microfilaments were observed 30 min, 1 h, and 1.5 h after cross- or self-pollination. Thirty minutes after cross-pollination, a time when most pollen grains begin to hydrate, we observed distinct actin bundles at the apical region (Fig. 1B, arrows); 1 h after cross-pollination, the period when many pollen grains germinate and begin to penetrate the papilla cell wall, many actin bundles were seen both along the longitudinal axis and concentrated at the pollen germination site (Fig. 1C, arrows). In contrast, during self-pollination, actin filaments at the apical region decreased over time and rearrangement of the actin filaments around the nuclei was not apparent 30 min (Fig. 1D) or 1 h (Fig. 1E) after pollination; 1.5 h after pollination, the total amount of actin filaments further decreased (Fig. 1F). These experiments were performed five times with reproducible results. Therefore, by rhodamine-phalloidin staining in stigmatic papilla cells, actin filaments increased during cross-pollination and decreased during self-pollination.

Before pollination, fine actin filaments, which were not visualized by rhodamine-phalloidin staining, were observed at the apical region of the papilla cell. Consistent with the previous rhodamine-phalloidin experiments, however, actin bundles were only rarely seen at the apical region with mTalin-GFP (Fig. 2, A and E). Thirty minutes after cross-pollination, the time at which pollen hydration begins, actin bundles, focused at the pollen grain, appeared in the apical region of the papilla cell (Fig. 2, B and F) and these distinct actin bundles remained focused on the point of pollen grain attachment during pollen hydration (Fig. 2, C and G); 1.5 h after pollination, when the pollen has typically germinated and begun to penetrate into a papilla cell, the direction of the actin bundles changed around the pollen tube (Fig. 2, D and H). Furthermore, prominent actin bundles focused at the penetration site were detected (Fig. 2, T and X). On the contrary, hydration of self-pollen grains only rarely occurred by 1.5 h after pollination. Additionally, under conditions of self-pollination, the number of apical actin filaments decreased at 1 h and prominent actin bundles never formed (Fig. 2, I–P). Finally, the number of fine actin filaments gradually decreased. Thus, the real-time visualization of the dynamics of actin rearrangements following cross- and self-pollination was consistent with data obtained using fixed cells and rhodamine-phalloidin staining, although pollen hydration and germination during cross-pollination were slightly delayed in this system for unknown reasons.

Figure 2. Alterations in the actin cytoskeleton during cross- and self-pollination and after adhesion of cross- or self-pollen coats visualized in cells expressing GFP-mTalin. Bright-field (A–D and T) and fluorescent images (E–H and X) of papilla (more ...)

After cross-pollen coat adhesion, the number of actin filaments gradually increased at the apical region and actin bundles were observed after 20 min (Fig. 2, Q–S). Thus, components of the cross-pollen coat alone were capable of mediating actin polymerization and bundle formation seen in papilla cells during cross-pollination. In contrast, after self-pollen coat adhesion, the number of actin filaments quickly decreased at the apical region and only few filaments were observed after 20 min (Fig. 2, U–W). These data further confirm the importance of the pollen coat in inducing alterations in the actin cytoskeleton in papilla cells during cross- or self-pollination.

Figure 3. CD disrupts the papilla cell actin cytoskeleton, pollen grain hydration, and germination following cross-pollination. A, Control cells untreated with 100 μm CD and stained with rhodamine-phalloidin as in Figure 1. B, Cells treated with 100 μ (more ...)

We next examined whether germination inhibition by CD was dose dependent. We soaked the cut pistil edge in different concentrations of CD for 4 h and examined hydration and germination following pollen grain addition using a micromanipulator (Fig. 2C). When a cross-pollen grain was attached to a papilla cell exposed to 10 μm CD, pollen hydration and germination were not inhibited. However, both pollen hydration and germination were significantly inhibited by 50 μm CD and germination was completely inhibited by 100 μm CD. The ability of CD to inhibit both pollen hydration and germination suggests that actin polymerization is required for these processes.

Figure 4. Vacuolar structure in a papilla cell prior to pollination and after cross- and self-pollination. A, Unpollinated papilla cells were observed using transmission electron microscopy accelerated at 75 kV. Small, tubular vacuoles surround the central vacuole. (more ...)

In contrast to the findings with cross-pollination, we detected no hydrated or germinated pollen grains on the papilla cells at 1 h after self-pollination by light microscopy. When papilla cells were examined by ultra-HVEM tomography following self-pollination, few elongated vacuoles were seen near the plasma membrane and apical vacuoles appeared fragmented (Fig. 4F). Thus, the 3D structure of the vacuolar network following self-pollination substantially differed from that seen either before pollination or after cross-pollination. Self-pollination induces changes in the vacuolar structure rather than maintaining a preexisting structure in papilla cells.

Using both rhodamine-phalloidin staining and real-time monitoring of actin behavior by GFP-mTalin, we precisely visualized the configuration of actin filaments in papilla cells after cross- and self-pollination. In this study, we showed that actin polymerization and bundle formation were promoted by cross-pollination and actin reorganization, most likely depolymerization, occurred after self-pollination. These results are not consistent with a previous report (Dearnaley et al., 1999), but this is likely due to the differences in techniques used to stain actin filaments. In this study, all stigma were treated, frozen, and crushed in liquid nitrogen and permeabilized with 0.05% Tween 20 in Tris-buffered saline containing dithiothreitol for 24 h prior to actin staining. This comprehensive protocol enhances visualization of actin filaments. Differences in sensitivity and/or structure preservation could explain the discrepancies between studies.

Cross-pollination proceeds in an ordered fashion, beginning with pollen adhesion and eventual hydration, germination, and pollen tube penetration into a papilla cell. Actin bundles formed at the apical region of papilla cells during hydration and these bundles, focused at the site of pollen grain attachment, were present during pollen germination and penetration. Treatment of cells with CD, an inhibitor of actin polymerization, induced actin filament fragmentation and inhibited pollen hydration and germination. Actin bundle formation during cross-pollination and inhibition of cross-pollination by CD suggests that the formation of actin bundles in the papilla cell is indispensable for pollen hydration and germination.

In Brassica, SP11/SCR in the pollen coat of self-pollen grains induces the SI to arrest pollen hydration and germination (Takayama et al., 2000). It is thought that SP11/SCR binds to SRK and induces SRK autophosphorylation to initiate the SI signaling cascade (Kachroo et al., 2001; Takayama et al., 2001). SP11/SCR is secreted by the tapetum cells in the anther and associates with the pollen surface. During pollination, SP11/SCR is transported from the pollen surface to the papilla cell via the pollen coat, leading to penetration of the papilla cell wall (Iwano et al., 2003). In this study, actin reorganization and likely depolymerization was observed 1 h after self-pollination and the pollen coat alone from self-pollen grains induced these configuration changes. When considered with our previous results, these data suggest that actin reorganization and likely depolymerization are involved in the SI response of papilla cells and occur downstream of SRK autophosphorylation.

The situation is reversed during cross-pollination. Actin polymerization was observed 1 h after cross-pollination and the pollen coat alone was responsible for this effect. Furthermore, inhibition of actin polymerization prevented pollen germination. Taken together, these data suggest that the pollen coat initiates a signaling cascade leading to changes in the actin cytoskeleton. Altered actin polymerization may be directly responsible for or coincident with initiation of cross-pollination.

In this study, we visualized the 3D structure of vacuoles in the papilla cell by ultra-HVEM tomography. Vacuolar structures are traditionally visualized by GFP imaging (Mitsuhashi et al., 2000; Saito et al., 2002; Kutsuna et al., 2003; Reisen et al., 2005) and HVEM tomography is typically performed on thin, 0.5-μm sections (Segui-Simarro et al., 2004). However, in this study, we examined 2- to 3-μm-thick sections using ultra-HVEM with acceleration energies approaching 2,000 kV to reconstruct the 3D network of the large vacuole and tubular vacuoles. Using this technique, we determined that the vacuolar network consists of large, globular, and tubular vacuoles located at the apical region of the papilla cells. The conformation of the vacuolar network changed after cross-pollination, extending toward the pollen attachment site. In contrast, the vacuolar structure after self-pollination was substantially different and resembled the vacuolar network in cells treated with CD; an elongated central vacuole and tubular network at the apical region of the papilla cell were rarely observed. Therefore, this correlation between actin bundle organization and vacuolar structure suggests that actin bundles regulate the structure and the positioning of vacuoles in papilla cells after cross- and self-pollination. The vacuole stores many different compounds and inorganic ions and the tonoplast (vacuolar membrane) contains several different transport systems, including proton pumps such as H⁺-ATPase and H⁺-pyrophosphatase, transporters such as Ca²⁺-ATPase, Ca²⁺/H⁺ antiporter, and Na⁺/H⁺ antiporter, and water channels (aquaporins). These transport proteins regulate ion concentration and turgor pressure in plant cells (Maeshima, 2001). Therefore, the actin cytoskeleton may regulate vacuolar structure to control transport of ions and water for pollen germination during cross- and self-pollination.

Actin bundles, which were prominent in the papilla cells after cross-pollination, have also been observed in other plant cells, such as the pollen tube, the root hairs of Hidrocharis, and the stamen hair cells of Tradescantia. These actin bundles were involved not only in cytoplasmic streaming, but also in stabilizing the structure of transvacuolar strands and maintaining cellular morphology (Staiger et al., 1994; Shimmen et al., 1995; Tominaga et al., 2000). In the pollen tube of Lilium longiflorum, formation and structure of actin bundles are regulated by two villins, 135-ABP (for actin-binding proteins) and 115-ABP, in a Ca²⁺/calmodulin-dependent manner (Yokota and Shimmen, 1999; Yokota et al., 2000, 2003; Hussey et al., 2006). Profilins, a family of G-ABPs, are associated with G-actin sequestration and Ca²⁺-regulated actin polymerization and nucleation (Wasteneys and Yang, 2004). Thus, both the configuration of F-actin and actin dynamics are determined by the action of different ABPs and intracellular Ca²⁺ concentration regulates the activity of some of these proteins (Staiger, 2000; Hussey et al., 2006). In this study, actin polymerization and actin bundle formation were observed after cross-pollination, whereas actin reorganization and likely depolymerization were observed after self-pollination. Identification of ABPs, active in papilla cells during cross- and self-pollination, might, therefore, lead us to identify the mechanisms downstream of the SI signaling pathway.

In Papaver rhoeas, which has a gametophytic SI, the pollen S-phenotype is determined by its haploid S-genotype, and the female S-determinant is S-protein, which is expressed and secreted by cells of the stigma. When the pollen tube encounters the S-protein, pollen tube growth is inhibited in an S-haplotype-specific manner and actin cytoskeletal rearrangements in the pollen tube are integral to this process (Geitmann et al., 2000). Increases in intracellular [Ca²⁺] precede actin rearrangements and an artificial increase in [Ca²⁺] induces comparable actin cytoskeletal changes (Franklin-Tong et al., 2002; Snowman et al., 2002; Staiger and Franklin-Tong, 2003). These data suggest that [Ca²⁺] is involved in the control of the actin cytoskeleton. Although the SI response in B. rapa is under sporophytic control, actin cytoskeleton rearrangements might also be triggered by increases in intracellular [Ca²⁺]. During interactions between leguminous plants and Rhizobium and following treatment of root hair cells with Nod factor, a Ca²⁺ influx induces the rapid disappearance of actin from the tip region, suggesting that Ca²⁺ signaling can initiate the rapid depolymerization of actin (Cárdenas et al., 1998, 2003; Whitehead et al., 1998). Dynamic Ca²⁺ fluxes occur during pollen hydration, germination, and penetration following pollination in Arabidopsis (Arabidopsis thaliana; Iwano et al., 2004). It is possible that differences in actin reorganization detected in the papilla cells following cross- and self-pollination are caused by differences in quality or magnitude of Ca²⁺ signaling produced by cross- and self-pollen.

In this study, we showed that actin polymerization and actin bundle formation in the papilla cell are indispensable for pollen hydration and germination. We also showed that during the SI response, actin reorganization and likely depolymerization in the papilla cell prevent pollen germination and that the actin cytoskeleton is involved in controlling function and 3D structure of vacuoles in the papilla cells before and during self- and cross-pollination. Thus, understanding the regulation of actin dynamics in papilla cells is critical for a better understanding of the mechanisms at the heart of the SI response (i.e. the mechanisms that facilitate hydration of cross-pollen grains and inhibit hydration of self-pollen grains).

For GFP imaging, the samples were scanned by argon laser (488-nm excitation). A pistil was mounted on a coverslip prior to pollination, fixed with double-sided tape, and then covered with a piece of 1% agar. After a cross- or self-pollinating grain was mounted on a GFP-expressed papilla cell using a micromanipulator, GFP images were monitored under dry conditions using the microscope system described above.

An ultra HVEM used was operated at 2,000 kV. Images were taken at 2,000× or 3,000× from +60° to −60° at 1° to 2° intervals around two orthogonal axes and collected with a 4K × 4K F415S CCD camera (TVIPS). The series were aligned and the tomographic volumes reconstructed as described (Mastronarde, 1997; McIntosh et al., 2005). The tomograms were displayed and modeled with the 3Dmod software package as described (Otegui et al., 2001; Segui-Simarro et al., 2004; McIntosh et al., 2005). For each tomography experiment, more than 100 different cells from more than 20 different stigmas were observed, and more than 100 electron micrographs were taken. Tomographic analysis of about 20 cells was performed.

• Supplemental Video S1. 3D structure of vacuoles in a papilla cell before pollination.

We are grateful to Dr. Chua of Rockefeller University for his kind donation of pYSC14 and to Mr. Hasegawa, Mr. Takemura, Mrs. Onishi, Mrs. Yamamoto, Mrs. Matsumura, and Mrs. Ichikawa for their technical assistance.