Vacuoles Form Thin Tubular Strands Around the Nucleus and Gradually Accumulate at the Basal Region During Zygote Elongation.

To perform live-cell imaging of the vacuole in the polarizing zygote, we developed a dual-color marker that simultaneously labels the vacuolar membrane and nucleus (vacuole/nucleus) and could be imaged using two-photon excitation microscopy (2PEM) (Fig. 1 A and B). We detected the same pattern of polar localization for the vacuole and the nucleus as the fixed ovules of nontransgenic plants (Fig. 1 and SI Appendix, Fig. S1) (5), thus indicating that our system visualized native intracellular patterns.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 1.

Live-cell imaging and quantification of vacuolar dynamics during zygote polarization in wild type (WT). (A and B) 2PEM image of the egg cell (A) and time-lapse observation of the zygote in in vitro cultivated WT ovules (B) expressing the vacuolar/nuclear marker. Maximum-intensity projection (MIP) images are shown. Images are representative of 11 time-lapse images. Numbers indicate the time (h:min) from when the zygote started elongation (“bulging”). Cyan arrowheads indicate nuclei. (Scale bars, 10 µm.) (B, Inset) An enlarged image of the perinuclear region. (C) Illustrations showing a summary of the respective stages. (D) Time course of “apical” (orange stripe) and “basal” (green stripe) lengths, which are the distances from the center of the nucleus to the apical edge and the basal end of the cell, respectively, as shown (Right). Both lengths were measured from −1:00 until 14:00, when the zygote elongation was completed (“mature” zygote). (E) Time course of the apical (orange) and basal (green) vacuolar areas as shown in the illustration.

We then captured the behavior using time-lapse imaging until the first zygotic cell division (Fig. 1 B and C and Movie S1). Before fertilization, the egg cell showed a clear polar organization, with the nucleus at the apical pole and huge vacuoles occupying the basal region. After fertilization, the cell shrunk as we previously reported (6), and vacuolar size was also markedly reduced. The nucleus shifted to the cell center, and the vacuoles were detected both in the apical and basal regions (Fig. 1 D and E). Then, the zygote elongated, and the vacuoles formed thin tubular structures around the nucleus, which migrates toward the apical tip (Fig. 1 B–D). Simultaneously, vacuolar size was also gradually increased. The vacuoles in the apical region remained small while vacuolar size in the basal area was massively increased, resulting in a highly asymmetric distribution of the vacuoles on completion of nuclear migration and zygotic elongation (Fig. 1 B, C, and E). Therefore, after chromosome segregation and cytokinesis, most of the vacuoles were inherited into the basal cell.

In summary, fertilization triggers vacuolar shrinkage, and the vacuoles swell again during zygote elongation. The nucleus becomes surrounded by thin vacuolar tubules and migrates to the apical apex, while the vacuoles gradually accumulate at the basal region.

The sgr2-1 Mutant Is Defective in Asymmetrical Zygotic Division and Embryo Morphology.

To understand the role of vacuoles in zygote polarization, we examined whether the zygote-division asymmetry was affected when vacuolar morphology and/or functions were disturbed. Because zygotic vacuolar regulators were not identified, we observed various vacuole mutants that exhibited defects in other tissues (SI Appendix, Fig. S1A). We found that sgr2-1 zygotes divided in a more symmetrical manner (Fig. 2A and SI Appendix, Fig. S1A), as confirmed by the higher ratio of apical to basal daughter cell lengths than in the wild type (Fig. 2B). In contrast, the total lengths of one-cell-stage embryos were indistinguishable between the wild type and sgr2-1 (Fig. 2C), indicating that cell elongation was unaffected in the sgr2-1 zygote.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 2.

Identification of the sgr2-1 phenotypes in zygote-division asymmetry and embryo morphology. (A) DIC images of cleared one-cell-stage embryos of WT, sgr2-1, and sgr2-1 containing the pSGR2::SGR2i-GFP transgene (sgr2-1 SGR2). Embryos are outlined, and the orange arrowhead indicates the apically remaining large vacuole, which is colored in the enlarged images (Insets). Magenta arrowheads indicate the cell-division planes. (B) The ratio of apical cells divided by basal cell lengths in one-cell-stage embryos, denoted as the asymmetric cell division of the zygote. (C) The total length of one-cell-stage embryos (sum of apical and basal cell lengths), denoted as zygote elongation. Error bars represent the SD (N = 80 for each genotype), and the letters indicate significant differences determined by Tukey–Kramer test; P < 0.01. (D) 2PEM images of eight-cell-stage embryos expressing the vacuolar/nuclear marker. Center plane images are shown. Orange arrowheads indicate the apically remaining large vacuoles, and Insets show the enlarged images of the apical region. (Scale bars, 10 µm.)

SGR2 (At1g31480) encodes a phospholipase A1-like protein localized on the vacuolar membrane (8, 9). Although the molecular function of SGR2 is still unknown, it has been reported that sgr2-1 severely impairs the dynamic shape changes of vacuoles and thus generates spherical and less mobile vacuoles in inflorescence stems (10). Therefore, the sedimentable plastids (amyloplasts) that are surrounded by the vacuolar membrane cannot smoothly reposition in response to gravitational forces, resulting in reduced gravitropism (10). We noticed that in the unfertilized egg cell of sgr2-1, the huge vacuole normally located at the basal end with the nucleus at the apical top, although the precise vacuolar shape could not be distinguished in the fixed samples (SI Appendix, Fig. S1B). Large vacuoles were detected in the sgr2-1 apical cell and also in the apical cell lineage (embryo proper) after zygotic division in all of the embryo stages examined, whereas only small vacuoles were detected in the wild type (Fig. 2 A and D and SI Appendix, Fig. S1 A and B). The sgr2-1 embryos showed distorted shapes with aberrant cell-division planes (Fig. 2D and SI Appendix, Fig. S1 A–C), and the sgr2-1 seedlings displayed morphological abnormalities such as the presence of three cotyledons (SI Appendix, Fig. S1D) (9). The sgr2-1 phenotypes in the zygote and embryo were restored by the SGR2 transgene (pSGR2::SGR2i-GFP), which complemented the gravitropic defects of sgr2-1 (Fig. 2 A–C and SI Appendix, Fig. S1A) (8). These findings indicated that SGR2 is required for proper vacuole distribution in the zygote, which would be important for zygote-division asymmetry and subsequent embryo development.

The sgr2-1 Mutant Fails to Form Tubular Vacuoles and Has Disturbed Polar Vacuole Distribution and Nuclear Migration.

We then performed time-lapse imaging of sgr2-1 (Fig. 3 and Movie S2). As observed in the fixed samples, the sgr2-1 egg cell polarly positioned the vacuole and the nucleus as in the wild type, although vacuoles were more spherical, as reported in inflorescence stems (Fig. 3A, compared with Fig. 1A and SI Appendix, Fig. S1B) (10). After fertilization, the sgr2-1 zygote showed cell shrinkage, nuclear localization at the cell center, and dispersed vacuoles both in the apical and basal regions. The indistinguishable vacuolar features between sgr2-1 and the wild type were confirmed by 3D reconstructed images (SI Appendix, Fig. S2A) and were also quantified by the similar values of circularity, perimeter, and area of the vacuole (Fig. 3 G–I), which were measured using a vacuolar lumen marker (SI Appendix, Fig. S2 B and C). However, during the zygotic elongation, sgr2-1 did not form a tubular vacuole, and only spherical and spotted vacuoles were observed in the perinuclear region, as shown by the larger value of circularity than in the wild type (Fig. 3G). In addition, the perimeter value in sgr2-1 was smaller than in the wild type (Fig. 3H), indicating reduced vacuolar surface. The apical and basal vacuoles both gradually increased in size (Fig. 3E), and finally vacuolar distribution became more symmetric than in the wild type (Fig. 3F). In contrast to the abnormal vacuole shape and distribution, the total vacuole area was indistinguishable between the wild type and sgr2-1 at all stages (Fig. 3 F and I), indicating that sgr2-1 specifically affected the vacuole morphology and polar vacuole positioning without altering vacuole size. This, together with the observation that the vacuolar lumen marker did not show spotted vacuoles in sgr2-1 (SI Appendix, Fig. S2A), suggested that SGR2 would be required to fuse small empty vacuoles with the large central vacuole, and thus to increase the vacuolar membrane component, not the vacuolar lumen volume.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 3.

Live-cell imaging and quantification of vacuolar dynamics during zygote polarization in sgr2-1. (A and B) 2PEM image of the egg cell (A) and time-lapse observation of the zygote in in vitro cultivated sgr2-1 ovules (B) expressing the vacuolar/nuclear marker. MIP images are shown. Images are representative of 10 time-lapse images. Numbers indicate the time (h:min) from the bulging, and cyan arrowheads indicate the nuclei. (Scale bars, 10 µm.) (B, Inset) An enlarged image of the perinuclear region. (C) Illustrations showing a summary of the respective stages. (D and E) Time course of the lengths and vacuolar areas in apical and basal regions, as in Fig. 1 D and E, respectively. Both values were measured from −1:00 until 11:00, when the zygotic elongation was completed. (F) Graph of the sum of apical (orange) and basal (green) vacuolar areas (Left) in the mature zygotes, denoted as total vacuole size, and the ratio of the apical area divided by the basal vacuolar area (gray; Right), denoted as asymmetric vacuole distribution. Error bars represent the SD [N = 11 (WT) and 10 (sgr2-1)], and significant differences from the WT values were determined by Mann–Whitney U test; **P < 0.01; ns, not significant. (G–I) Graphs of the circularity (G), perimeter (H), and area (I) of the vacuolar region at the indicated stages. Elong, elongating zygote; Mat, mature zygote. Error bars represent the SD [N = 14 (WT, Young), 12 (sgr2-1, Young), 14 (WT, Elong), 8 (sgr2-1, Elong), 17 (WT, Mat), and 13 (sgr2-1, Mat)], and significant differences from the values of the WT were determined by Brunner–Munzel test; **P < 0.01. (G–I, Right) Illustrations show the correlation between the vacuolar features and the respective values.

The apical domain of the sgr2-1 zygote was occupied by large spherical vacuoles (Fig. 3E), which prevented the nucleus from reaching the apical cell end (Fig. 3D). Even when the zygote completed cell elongation, the nucleus still remained near the cell center (Fig. 3D). Therefore, the sgr2-1 zygote divided more symmetrically than that of the wild type, inheriting large vacuoles in the apical cell, which was consistent with the fixed-sample observations (Fig. 2 A and B and SI Appendix, Fig. S1 A and B). Despite the severe defect in cell-division asymmetry, cell-division timing was not affected in sgr2-1, as shown by the comparable time from the start of zygote elongation until the onset of nuclear division in the wild type (15 ± 5.6 h; n = 11) and in sgr2-1 (15 ± 3.3 h; n = 10) [not significant (P = 0.50) by Mann–Whitney U test]. These results showed that SGR2 is crucial for the dynamic vacuolar shape change and polar vacuolar distribution, which support accurate nuclear positioning and thus zygote-division asymmetry.

Because the vacuole is a fundamental organelle in plants, we then examined whether the vacuole-dependent mechanism to regulate zygotic asymmetry works in other tissues. As typical events involving cell polarization and/or asymmetric cell division, we focused on root hair initiation at the basal end of epidermal cells (SI Appendix, Fig. S3A) (11), nuclear positioning in the root hair (SI Appendix, Fig. S3 A and B) (12), cell-layer formation in the root meristem (SI Appendix, Fig. S3C) (13), and stomatal shape and patterning (SI Appendix, Fig. S3D) (14). Although SGR2 was expressed in whole seedlings (8) and the vacuole morphology was actually affected (SI Appendix, Fig. S3 A and C), we could not detect any defects in these processes in sgr2-1. These results suggested that the SGR2-dependent regulation of the asymmetric cell division is specific to the zygote.

Tubular Vacuole Formation and Polar Vacuole Distribution Depend on the Longitudinal Array of F-Actin.

To identify the driving force of polar vacuole positioning, we tested the involvement of F-actin, because it is known to associate with vacuolar membranes and regulate vacuolar morphology during various cellular events, such as cell-cycle progression in tobacco culture cells (15) and cell elongation in Arabidopsis roots (16). We have recently reported that in the Arabidopsis zygote, the F-actin array is disorganized upon fertilization and then gradually aligns longitudinally along the apical–basal axis (Fig. 4 A and B, SI Appendix, Fig. S4 A and B, and Movie S3) (6), implying an analogy to the temporal shrinkage and tubular formation of vacuoles. As in the wild type, F-actin in the sgr2-1 zygote was disorganized after fertilization and then gradually reorganized into parallel, dense, and thick bundles (SI Appendix, Fig. S4 C–G and Movie S4). Only the average angle of F-actin fibers (Δθ) was not reduced in the mature zygote of sgr2-1 (SI Appendix, Fig. S4H), indicating the less longitudinal alignment of F-actin cables in the whole cell. However, it seemed due to the large spherical vacuoles occupying the cell center in sgr2-1 (Fig. 4C). Indeed, the F-actin was excluded from the spherical vacuoles in sgr2-1 (Fig. 4G, a), whereas in the wild type the F-actin cable was detected in the basal vacuolar lumens, known as the transvacuolar strand. Moreover, the perinuclear F-actin was highly longitudinal and parallel both in the wild type and sgr2-1 (Fig. 4 B, D, and F). In the perinuclear region, F-actin cables were associated with tubular vacuoles in the wild type but not in sgr2-1 (Fig. 4G, b), showing that sgr2-1 fails to form tubular vacuoles along F-actin.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 4.

Pattern and role of F-actin cables in tubular vacuole formation and polar vacuole positioning. (A–D) 2PEM images of mature zygotes expressing the F-actin/nuclear marker. Center plane (A and C) and MIP images (B and D) are shown. Yellow arrowheads indicate large spherical vacuoles, and brackets indicate the perinuclear regions. (E and F) Graphs of the parallelness (E) and average angle of the fibers against the cell longitudinal axis (Δθ; F) of F-actin in the indicated cell areas of mature zygotes. Error bars represent the SD [N = 12 (WT) and 13 (sgr2-1)], and significant differences from WT values were determined by Brunner–Munzel test; **P < 0.01; ns, not significant. Average values in the whole-cell area were significantly different between WT and sgr2-1, but values in the perinuclear region between the apical edge and the basal end of the nucleus were similar in WT and sgr2-1. Illustrations show the correlation between the respective values and the fiber patterns. (G) 2PEM images of mature zygotes expressing the F-actin/vacuolar lumen marker. Magnified images of the square areas (a and b) are shown (Right). MIP images are shown for the whole-cell area, and single z slices are shown in the magnified area. Images are representative of five and seven samples for WT and sgr2, respectively. (G, a) The transvacuolar strand in the basal vacuoles of WT, which did not appear in sgr2-1. (G, b) The association of tubular vacuoles with F-actin cables in WT, and the lack of detectable perinuclear vacuoles in sgr2-1. (H) 2PEM images of mature zygotes expressing the vacuolar/nuclear marker after exposure to the control DMSO and the polymerization inhibitor of F-actin (1 μM LatB) for 3 h. (H, Insets) Enlarged images of the perinuclear region. (I) Graph of the vacuolar area in the perinuclear region of DMSO- and LatB-treated zygotes for 3 h. Error bars represent the SD [N = 9 (WT) and 7 (sgr2-1)], and significant differences from the values of the DMSO-treated zygotes were determined by two-sided Welch’s t test; **P < 0.01. (J) 2PEM images of one-cell-stage embryos of WT and sgr2-1 expressing the vacuolar/nuclear marker after exposure to DMSO and LatB for 24 h. Yellow arrowheads indicate the large vacuoles in the apical cells, and brackets indicate the lengths of the apical and basal cells. (K) The ratio of apical cells divided by basal cell lengths in one-cell-stage embryos of DMSO- and LatB-treated zygotes for 24 h. Error bars represent the SD [N = 9 (WT, DMSO), 10 (WT, LatB), 10 (sgr2-1, DMSO), and 9 (sgr2-1, LatB)], and letters indicate significant differences determined by Tukey–Kramer test; P < 0.05. (L) Schematic representation of the dynamics and roles of vacuoles and F-actin in zygote polarization and embryo development. (Scale bars, 10 µm.)

We then used the actin polymerization inhibitor, latrunculin B (LatB), which effectively destroyed the F-actin pattern observed in our zygote imaging system (6). After a 3-h LatB treatment, the tubular vacuoles disappeared from the perinuclear region (Fig. 4 H and I). A similar disappearance of the perinuclear vacuoles was caused by another actin polymerization inhibitor, cytochalasin D (CytD), which also disrupted the F-actin pattern in the zygote (SI Appendix, Fig. S4I). Moreover, after a 24-h LatB treatment, the zygote divided in a more symmetric manner, generating highly vacuolated apical cells as in sgr2-1 (Fig. 4 J and K). This LatB effect was not additive to sgr2-1 (Fig. 4K), implying that the F-actin–dependent nuclear positioning is mainly due to the polar vacuole distribution.