Live-Cell Imaging and Optical Manipulation of Arabidopsis Early Embryogenesis

Technology

Graphical Abstract Authors Keita Gooh, Minako Ueda, Kana Aruga, ..., Hideyuki Arata, Tetsuya Higashiyama, Daisuke Kurihara

Correspondence [email protected]

In Brief Gooh et al. establish a live-embryo imaging system for Arabidopsis and generate a complete lineage tree from double fertilization in early embryogenesis. They provide a platform for real-time analysis of cell division dynamics and cell fate speciﬁcation using optical manipulation and micro- engineering techniques such as laser irradiation of speciﬁc cells.

Highlights d A live-embryo imaging system for Arabidopsis is established d A complete lineage tree from double fertilization to 64-cell embryo uses this system d Endosperm development is not required for cell patterning during early embryogenesis d Damage to an embryo initial cell induces rapid cell fate conversion in the suspensor

Gooh et al., 2015, Developmental Cell 34, 242–251 July 27, 2015 ª2015 Elsevier Inc. http://dx.doi.org/10.1016/j.devcel.2015.06.008 Developmental Cell Technology

Live-Cell Imaging and Optical Manipulation of Arabidopsis Early Embryogenesis

Keita Gooh,1 Minako Ueda,1,2 Kana Aruga,1 Jongho Park,1,3,4 Hideyuki Arata,1,3 Tetsuya Higashiyama,1,2,3 and Daisuke Kurihara1,3,* 1Division of Biological Science, Graduate School of Science, Nagoya University, Furo-cho, Chikusa-ku, Nagoya, Aichi 464-8602, Japan 2Institute of Transformative Bio-Molecules (ITbM), Nagoya University, Furo-cho, Chikusa-ku, Nagoya, Aichi 464-8602, Japan 3Higashiyama Live-Holonics Project, ERATO, JST, Furo-cho, Chikusa-ku, Nagoya, Aichi 464-8602, Japan 4Present address: Precision and Intelligence Laboratory, Tokyo Institute of Technology, 4259 Nagatsuta-cho, Midori-ku, Yokohama, Kanagawa 226-8503, Japan *Correspondence: [email protected] http://dx.doi.org/10.1016/j.devcel.2015.06.008

SUMMARY to the embryo (Kawashima and Goldberg, 2010; Lau et al., 2012; Wendrich and Weijers, 2013). In addition to cell fate specification Intercellular communications are essential for cell of the apical and basal cells, the fate of a variety of other cell proliferation and differentiation during plant embryo- types is determined during embryogenesis, such as epidermis genesis. However, analysis of intercellular commu- formation when the radial axis is established at the 16-cell em- nications in living material in real time is difficult bryo stage. However, the regulation of pattern formation during owing to the restricted accessibility of the embryo embryogenesis remains unknown. within the flower. We established a live-embryo Flowering plants produce two sperm cells; one fuses with the egg cell to form the zygote and the other fuses with the other imaging system to visualize cell division and cell Arabidopsis thaliana female gamete, the central cell, to form the endosperm. This fate specification in from process is termed double fertilization (Mansfield et al., 1991). zygote division in real time. We generated a cell- The central cell also contributes to early embryogenesis. For division lineage tree for early embryogenesis in example, a central-cell-derived small cysteine-rich peptide, Arabidopsis. Lineage analysis showed that both ESF1, is required for normal pattern formation of embryos (Costa the direction and time course of cell division be- et al., 2014). The endosperm provides nutrition to the embryo (Li tween sister cells differed along the apical-basal and Berger, 2012) and is considered to be an important contrib- or radial axes. Using the Arabidopsis kpl mutant, utor to embryogenesis in developing seeds. In addition, the sus- in which single-fertilization events are frequent, pensor is believed to supply nutrients and growth factors to the we showed that endosperm development is not proembryo and thus play an important role in embryogenesis by required for pattern formation during early embryo- connecting the proembryo to the surrounding tissue (Kawashima and Goldberg, 2010). In spite of these distinct morphological fea- genesis. Optical manipulation demonstrated that tures, little is known about the contribution of the endosperm and damage to the embryo initial cell induces cell fate suspensor to the timing of cell division timing and pattern forma- conversion of the suspensor cell to compensate tion of the embryo. for the disrupted embryo initial cell even after cell Along the apical-basal axis, several factors are expressed in fate is specified. specific regions of the embryo to regulate embryo development (Wendrich and Weijers, 2013). Among the first genes to become activated in the apical cell, DORNRO¨SCHEN (DRN) (also known INTRODUCTION as ENHANCER OF SHOOT REGENERATION1; ESR1) encodes an AP2-type transcription factor and is expressed in the apical In plants, the first division of the fertilized egg cell (zygote) is crit- cell derivatives after the first zygotic division (Cole et al., 2009; ical to establish the apical-basal axis and to produce two Kirch et al., 2003). Given that embryo pattern formation fails in daughter cells with different developmental fates (Mansfield the double mutant of DRN and its closest homolog, DRN-LIKE and Briarty, 1991). The zygote divides asymmetrically to form a (DRNL), resulting in development of an embryo without cotyle- small cytoplasm-rich apical cell and a large vacuolated basal dons, DRN is thought to be a key regulator of embryo formation cell (1-cell embryo stage). This asymmetric zygotic division is (Cole et al., 2009; Kirch et al., 2003). common in plants. The apical cell, which acts as the embryo WUSCHEL RELATED HOMEOBOX (WOX) transcription fac- initial cell, proliferates in a well-organized pattern to generate tors also show region-specific expression and are essential the proembryo, which is the precursor of almost all cells of the for the formation and maintenance of the apical-basal axis plant body (Ueda and Laux, 2012). The basal cell, which acts (Breuninger et al., 2008; Haecker et al., 2004; Ueda et al., as the suspensor initial cell, repeatedly divides horizontally to 2011). After zygote division, WOX2 expression is restricted to generate the filamentous suspensor, which connects the proem- the apical cell derivatives and regulates embryo development. bryo to the surrounding maternal tissues and transports nutrients In contrast, WOX8 expression is restricted to the basal cell

242 Developmental Cell 34, 242–251, July 27, 2015 ª2015 Elsevier Inc. derivatives and regulates suspensor development (Breuninger RESULTS AND DISCUSSION et al., 2008; Ueda et al., 2011). In addition, WOX8 regulates embryo development via non-cell-autonomous activation of WOX2 Development of Live-Cell Imaging System for (Breuninger et al., 2008). However, this mechanism remains un- Arabidopsis Embryogenesis clear, and it is unknown how the apical and basal cell fates are Angiosperm embryogenesis in vivo occurs within multiple layers related. of maternal tissues in the flower; therefore, it is difficult to observe and analyze the real-time dynamics in living material DESIGN (Figure 1A). To enable live-cell analysis of early embryogenesis, we developed an in vitro ovule culture system using Arabidopsis The spatiotemporal development of the apical and basal cells, thaliana. First, we examined the effect of medium composition and how their distinct fates are specified, are also poorly under- on ovule growth and embryo development. Nitsch medium sup- stood. Early embryogenesis in vivo occurs within the ovule in the plemented with 5% trehalose resulted in the highest percentage flower. Sauer and Friml (2004) developed an in vitro culture of ovule survival (Figure 1B). Compared with the previously method for fertilized ovules in Arabidopsis, but direct observa- developed in vitro culture medium (Sauer and Friml, 2004), the tion of early stages of embryogenesis was difficult due to the frequency of normal embryo development was significantly low survival rate of young ovules containing a zygote. An effi- higher after incubation for 24, 48, and 72 hr in Nitsch medium cient in vitro culture system for development of the zygote into supplemented with 5% trehalose (Figure 1C). Furthermore, using a mature seed is yet to be established in Arabidopsis. Currently, the latter medium, the embryos were able to develop into adult our knowledge of early embryo development in plants is mostly plants (Figures 1D and 1E). Thus, trehalose was effective for derived from snapshot images of fixed samples. Recently, con- ovule growth and embryo development in A. thaliana. In several struction of a four-dimensional map from a large number of fixed angiosperms, including monocots and dicots, exogenous treha- embryo samples has contributed to an improved understanding lose application confers high desiccation tolerance to somatic of cell division patterns during early embryogenesis in Arabidop- embryos of barley (Ryan et al., 1999) and a greater adaptability sis (Yoshida et al., 2014). However, studies that use fixed sam- to environmental conditions during acclimatization in jojoba ples lack real-time temporal information such as the length of (Llorente et al., 2007), and promotes long-term in vitro vegetative the cell cycle in specific cell types; therefore, how cell fates culture in Torenia fournieri (Yamaguchi et al., 2011). Trehalose are spatiotemporally specified remains largely unknown both accumulation is associated with exposure to abiotic and biotic at the cellular and molecular levels. Moreover, direct interven- stress in many plants (Fernandez et al., 2010). Thus, trehalose tions, such as cell disruption by laser irradiation at a specific may confer enhanced tolerance to stress induced by in vitro con- time point, might reveal cell-to-cell communication in early ditions, but the mechanism by which trehalose exerts its effects embryogenesis, but fixed samples do not permit real-time remains unclear. observation of intercellular communication (Kurihara et al., Recently, we fabricated poly(dimethylsiloxane) (PDMS) micro- 2013). Genetic cell ablation by conditional expression of a cyto- cage arrays to adjust the orientation and fix the position of ovules toxic gene has been previously applied in a study of Arabidopsis (Park et al., 2014). Using a modified PDMS micropillar array, the embryogenesis for such direct interventions (Weijers et al., embryo was clearly and stably observed within the ovule by 2003). However, it is difficult to disrupt individual cells with adjusting the orientation of the embryo close to the cover glass high spatiotemporal precision by means of genetic ablation (Figure 1F). This in vitro ovule culture system enabled us to because of the dependence on the promoter activity. A chemi- perform live-cell imaging with enhanced stability. We obtained cal reagent was used to directly manipulate in vitro zygotic time-lapse recordings of embryogenesis using a spinning-disk embryogenesis in tobacco (Yu and Zhao, 2012), but is difficult confocal microscope, which enabled us to record the cellular to manipulate individual cells with high spatiotemporal precision events within the ovule without cell damage, as we reported pre- using such chemical reagents. Moreover, zygotic embryos of viously (Hamamura et al., 2011, 2014). Using WOX2p::H2B-GFP tobacco show variant cell pattern formation in vitro, which pre- WOX2p::tdTomato-LTI6b plants, the nucleus and plasma mem- sents another difficulty in analyzing cell pattern formation during brane were differentially labeled in the embryo and the suspen- embryogenesis. sor. Cells divided in both the proembryo and the suspensor. Here, we developed the in vitro ovule culture system to allow Normal embryo development, including the transition from the normal ovule growth and embryo development from zygote globular to the heart embryo stages, was observed using the division in Arabidopsis using Nitsch medium supplemented in vitro ovule culture system (Figure 1G; Movie S1). with trehalose. Using this medium, we could trace the cell To explore the relationship between the embryo and the endo- division and the cell fate specification under the microscope sperm in the timing of cell division, we obtained time-lapse re- in real time. To clearly and stably observe the embryo within cordings of nuclear division using Arabidopsis transgenic lines the ovule for long-term period, we also developed the microflui- in which the nuclei were labeled with fluorescent proteins. dic device for fixation of the position of ovules. In addition, we By tracing the nuclei, we generated cell-division lineage trees developed a water-supply system to integrate the laser disrup- for the proembryo, suspensor, and endosperm from the first tion into the in vitro ovule culture system using a femtosecond nuclear division (Figure 2A). Using a female-gametophytic-cell- Ti:sapphire pulse laser for long-term observation. Our devel- specific FGR 8.0 (Vo¨ lz et al., 2013) reporter gene, Arabidopsis oped system allows for analysis of early stages of Arabidopsis embryo and endosperm development were observed after the embryo development in real time by live-cell imaging and opti- first free-nuclear endosperm division (Figure 2B; Movie S1). cal manipulation. The endosperm nuclei divided synchronously, and the zygote

Developmental Cell 34, 242–251, July 27, 2015 ª2015 Elsevier Inc. 243 Figure 1. Live-Cell Imaging of Early Embryogenesis in Arabidopsis using an In Vitro Ovule Culture System (A) Schematic representation of embryogenesis in Arabidopsis. After double fertilization, the embryo and endosperm develop in the ovule. The zygote divides asymmetrically into an apical cell and a basal cell (1-cell embryo stage). The apical cell divides vertically to generate the proembryo, which ultimately gives rise to the adult plant (2-cell embryo stage). The basal cell divides horizontally to produce the suspensor. Early embryo stages are defined by the number of cells in the proembryo. (B) Ovules from WOX2p::H2B-GFP WOX9p::H2B-GFP plants were incubated in the culture media in a growth chamber. Vitality was assessed by detection of H2B-GFP fluorescent signal in the embryo and autofluorescence of the integuments. (C) Ovules of WOX2p::H2B-GFP plants were incubated in the culture media in the inverted confocal microscope system with a stable incubation chamber. Images were taken at 10 min intervals for 72 hr. Normal development was defined as embryo development not arrested during observation. The data are the means ± SD for more than three independent experiments (n = 15–53 ovules per treatment). *p < 0.05; **p < 0.01. Statistical significance was determined using post hoc multiple comparison t tests using the Holm-Sidak method (a = 0.05). (D) Growth of embryos from an early stage to seedlings (indicated by arrowheads) in Nitsch liquid medium supplemented with 5% trehalose. (E) Seedling growth into an adult plant in Nitsch liquid medium supplemented with 5% trehalose (26 days) and on 1/2MS agar medium (32 and 37 days). Numbers indicate days (d) of culturing. Arrowheads indicate seedlings. (F) Micrographs of PDMS micropillar array. Ovules of WOX2p::H2B-GFP (green) and WOX2p::tdTomato-LTI6b (magenta) were successfully attached. (G) Nucleus and plasma membrane of the embryo and suspensor labeled with WOX2p::H2B-GFP (green) and WOX2p::tdTomato-LTI6b (magenta), respectively. The proembryo (pro) is marked by a bracket. Numbers indicate time (hr: min) from the first frame. Scale bars represent 2 mm (D and E), 300 mm (F), and 20 mm (G). See also Movie S1. divided after the fourth endosperm division (Figure 2B). The labeled; Movie S1). Proembryo cell divisions were not strictly timing of the first zygotic division was consistent with that synchronous, even at the 4- to 8- and 8- to 16-cell stages (Fig- observed in vivo (Boisnard-Lorig et al., 2001). Jenik et al. ure 2C, asterisks; Movie S1), consistent with three-dimensional (2005) calculated the doubling time as 9.9 hr for Arabidopsis em- volume analysis of early embryogenesis in Arabidopsis (Yoshida bryos by clearing developing seeds. As shown in Figure S1, the et al., 2014). The duration of cell division between sister cells average times of cell division were 7.2–8.7 hr until the 16-cell differed along the apical-basal axis (apical versus basal, upper stage embryo in in vitro conditions. These results show that tier versus lower tier) or radial axis (epidermis versus inner) (Fig- the timing of cell division in embryos in the in vitro ovule culture ure 2A). Cell volumes of the lower tier and epidermis were higher system was similar to that of embryos in vivo. than those of the upper tier and inner cells, respectively (Yoshida et al., 2014). Geometric asymmetry was correlated with cell fate Cell Fate Specification during Early Embryogenesis specification, including the cell division cycle. Arabidopsis Embryogenesis from the zygote to the globular-stage embryo in embryos may be similar to Xenopus laevis and Caenorhabditis WOX2p::H2B-GFP WOX2p::tdTomato-LTI6b plants is shown in elegans embryos, in which cell-cycle duration is inversely pro- Figure 2C (the nuclei and plasma membrane are differentially portional to cell size (Arata et al., 2015; Masui and Wang, 1998).

244 Developmental Cell 34, 242–251, July 27, 2015 ª2015 Elsevier Inc. Figure 2. Cell-Division Lineage Analysis during Arabidopsis Early Embryogenesis (A) Lineage tree of the zygote and endosperm from double fertilization. Zygote and endosperm division were traced to the sixth nuclear-free division in FGR 8.0 line. All cells were traced to the 32-cell stage in the WOX2p::H2B-GFP line. Each node of the tree indicates a cell division time point. Branch lengths indicate the mean cell-cycle duration (n = 2–7 in the endosperm, n = 4–11 in the proembryo, n = 3–11 in the suspensor for each stage). Timing of the ﬁrst free-nuclear division after fertilization is taken from Boisnard-Lorig et al. (2001). (B) Zygote nucleus (z) and endosperm nuclei (en) labeled with EC1p::NLS-3xDsRed (magenta) and DD22p::YFP (green), respectively. The zygote divided into the apical cell (a) and basal cell (b). (C) Nucleus and plasma membrane of the embryo and suspensor labeled with WOX2p::H2B-GFP (green) and WOX2p::tdTomato-LTI6b (magenta), respectively. Asterisks indicate nuclei that show asynchronous cell division. The proembryo (pro) is marked by a bracket. Numbers indicate time (hr: min) from the ﬁrst frame. (D) Time-lapse images of WOX2::nDsRed (magenta) and WOX8::nVenus (green). (E) Time-lapse images of WOX2::nDsRed (magenta) WOX8::nVenus (green, nucleus localization) fertilized by DRN::erGFP (green, ER localization). Scale bars, 20 mm. See also Movies S1 and S2.

the endosperm to proembryo formation using the kpl mutant, in which sperm oc- casionally fertilize only the egg cell and the central cell fails to develop into the endosperm (Maruyama et al., 2013; Ron et al., 2010). Complete development of the embryo after both single and double fertilization was observed, and the cell division time course was unaffected (Fig- ures 3A–3C; Movie S3). To determine whether such embryos show normal cell To visualize the process of cell fate specification using molec- identity even during late embryogenesis, we studied expression ular markers, the basal-cell embryo marker WOX8 (Ueda et al., of the epidermis-specific marker ATML1 (Takada and Ju¨ rgens, 2011)(WOX8 delta::NLS-3xVenus; hereafter WOX8::nVenus) 2007). The ATML1p::NLS-3xGFP (hereafter ATML1::nGFP) and apical-cell embryo marker DRN (Cole et al., 2009) signal was limited to the proembryo and gradually restricted (DRNp::erGFP; hereafter DRN::erGFP) were applied (Figures to the epidermis at the globular stage after double fertilization 2D and 2E; Movie S2). Expression of WOX8::nVenus was de- (Figure 3D). This ATML1::nGFP expression pattern was also tected in the zygote and in the apical and basal cells of the observed in the absence of endosperm development (Figure 3E). 1-cell embryo. In the 2-cell embryo, the WOX8::nVenus signal Thus, endosperm development was not required for initiation was limited to the suspensor (Figure 2D). In contrast, a of cell fate specification during early embryo development. DRN::erGFP signal was detected in the apical cell after the first Recently, Costa et al. (2014) demonstrated that ESF1 peptides, zygotic division and was limited to the proembryo up to the glob- which are expressed in the central cell before fertilization and ular stage (Figure 2E). Thus, the live-embryo imaging system in the endosperm after fertilization, maternally contribute to em- enabled cell fate specification to be traced from the first zygotic bryo development. In single-fertilized ovules of the cdka;1 division. mutant, normal embryo development occurred in the absence of endosperm development (Costa et al., 2014). The signals Endosperm Development Is Not Required for Cell from the central cell are possibly sufficient for early embryo Patterning during Early Embryogenesis development after fertilization. Laser disruption of the central To determine if early endosperm development is required for cell would clarify the importance of the central cell in embryo normal proembryo formation, we assessed the contribution of development after fertilization.

Developmental Cell 34, 242–251, July 27, 2015 ª2015 Elsevier Inc. 245 Figure 3. Intercellular Communication between the Proembryo and Endosperm in Arabidopsis (A and B) Time-lapse images of WOX2p::H2B-GFP (green) WOX2p::tdTomato-LTI6b (magenta) fertilized by kpl RPS5Ap::H2B-tdTomato (magenta) after double fertilization (A) and single fertilization of the egg cell (B). (C) Lineage tree of the proembryo after double and single fertilization. The wild-type sequence is shown in Figure 2A. All cells were traced from the 2-cell to the 16-cell stages in the WOX2p::H2B- GFP line after pollination with kpl pollen. Branch lengths indicate the mean cell-division duration (n = 4–6). (D) WOX9p::tdTomato-LTI6b (magenta) pistil pollinated with ATML1::nGFP (green). (E) ATML1::nGFP (green) pistil pollinated with kpl RPS5Ap::H2B-tdTomato (magenta). Numbers indicate time (hr: min) from the ﬁrst frame. Scale bars, 20 mm. The proembryo is marked by a bracket. See also Figure S1 and Movie S3.

pulse laser for two-photon imaging (Fig- ure 4A). Pulse laser irradiation bleached the H2B-GFP fluorescent signal from a single cell of an 8-cell embryo located deep within the ovule (Figure S2). We also developed a water-supply system to enable use of an objective lens with a long working distance for long-term two- photon live-imaging with the inverted microscope (Figure 4A). After the first zygotic division, the apical and basal cells showed differences in the pattern and duration of divisions (Figure 2). We first applied the laser disruption method to the apical cell. To monitor cell vitality by visualization of plasma membrane integrity, we used the double-marker transgenic line WOX2p:: H2B-GFP WOX2p::tdTomato-LTI6b. The H2B-GFP fluorescent signal was elimi- nated from the apical cell after laser irradiation (Figure 4B, top; Movie S4). Just after laser irradiation, the plasma membrane was observed in the targeted cell, sug- gesting that the cell was still alive. How- Live-Cell Manipulation of Early Embryogenesis ever, time-lapse imaging after laser irradiation (Figure 4B, Analysis of embryo mutants reveals that intercellular communi- bottom) showed that the target cells did not divide during obser- cation between the proembryo and suspensor is important for vation for more than 24 hr (100%, n = 16). This result suggested embryogenesis (Yadegari and Goldberg, 1997). However, the that pulse laser irradiation resulted in decreased cellular activity spatiotemporal dynamics of the communication remains unclear of the target cell. Laser-irradiated cells were typically shrunken even at the cellular level. Genetic ablation to disrupt proembryo during observation (Figure 4B, asterisk; 13%, n = 16). and suspensor communication is difficult owing to unavailability After laser irradiation of the apical cell, the basal cell first of a specific promoter expressed only in the apical cell or basal divided horizontally (Figure 4B, bottom). The upper daughter cell after zygote division. Therefore, we established a laser irra- cell then divided vertically, comparable to division of the apical diation method for disruption of a single cell in conjunction with cell. The apical-like basal cell similarly divided to give rise to a the in vitro ovule culture system using a femtosecond Ti:sapphire 16-cell embryo (Figure 4B; 69%, n = 16). All vertical divisions

246 Developmental Cell 34, 242–251, July 27, 2015 ª2015 Elsevier Inc. Figure 4. Laser Disruption of the Apical Cell Induced Cell-Fate Conversion of the Basal Cell Derivatives (A) Schematic illustration of the microscope setup. An ovule in a glass-bottom dish was illuminated with near-infrared light from the Ti:sapphire laser for two-photon excitation and disruption and observed under a long-working-distance objective lens. Immersion water was pumped by a syringe pump through a needle to the objective lens. The ﬂuorescence signals were detected using non- descanned GaAsP detectors. (B) Apical cell of a 1-cell-stage embryo was targeted by laser irradiation in WOX2p::H2B-GFP (green) and WOX2p::tdTomato-LTI6b (magenta). The red lightning bolt indicates the laser irradiation point. H2B-GFP ﬂuorescence signals were elimi- nated after laser irradiation of the apical cell (After). The asterisk indicates that the irradiated apical cell was shrunken. The proembryo is marked by a bracket. The uppermost suspensor cell generated an apical embryo-like structure (dashed bracket). Numbers indicate the time from initiation of observations after laser irradiation (hr: min). (C) Time-lapse images of WOX2::nDsRed (magenta) and WOX8::nVenus (green) after laser irradiation of the apical cell. Fluorescence intensity of WOX8:: nVenus in a suspensor cell (arrowhead) is shown. (D) Time-lapse images of WOX2::nDsRed (magenta) WOX8::nVenus (green, nucleus localization) fertilized by DRN::erGFP (green, ER localization) after laser irradiation of the apical cell.

believed to be involved in the transfer of nutrients, growth factors, and phytohor- mones to the proembryo for embryo development in plants (Kawashima and Goldberg, 2010). Our results are consistent with the hypothesis that the basal occurred after horizontal division of the original basal cell (100%, cell and its derivatives provide positive regulators for growth n = 11; Figure 5C). To clarify whether basal-cell properties were stimulation and proper cell pattern formation in the proembryo lost and apical-cell properties were attained by the upper, pro- from the 1-cell stage. embryo-like cell, the basal-cell embryo marker WOX8::nVenus and apical-cell embryo marker DRN::erGFP were applied (Fig- Limitations ures 4C and 4D; Movie S5). After horizontal division of the basal There are some technical limitations of live-cell imaging and cell, the WOX8::nVenus signal intensity gradually decreased on optical manipulation for embryogenesis within the ovule in Arabi- the upper side of the two derivative cells and a weak DRN::erGFP dopsis. One limitation is the observation distance of embryo signal was detected in the upper two cells of the suspensor (Fig- within the ovule. As the embryo develops, the ovule expands ures 4C and 4D). The upper suspensor cells then divided verti- and the distance of the embryo from the glass surface increases. cally (Figures 4C and 4D). These observations demonstrated It is difficult to observe the late embryogenesis such as late heart that extra-embryonic cells can be converted into embryonic cells stage and torpedo stage embryos within the ovule even by two- to compensate for disruption of the apical cell. photon microscopy. The second limitation is the phototoxicity to We next applied laser irradiation to the basal cell of a 1-cell the embryo development by laser illumination. It is better to use embryo. After irradiation, the apical cell divided to form a proem- high-sensitivity detectors such as GaAsP detector for reducing bryo of 4–8 cells without division of the basal cell (Figure 5A; the laser intensity. In addition, two-photon imaging with longer Movie S6). Horizontal division was not observed in the proem- wavelength excitation also helps to reduce the laser intensity bryo (0%, n = 13). However, aberrant cell division in the proem- by reducing the autofluorescence (Mizuta et al., 2015). We could bryo at the 2- to 32-cell stages was observed, with 11–16 hr perform live-cell imaging of Arabidopsis embryogenesis for more delays in the duration of cell division compared with that of the than 3 days by two-photon microscopy at 950-nm laser illumina- control embryo and apical-like cells derived from extra-embry- tion with GaAsP detectors. Third limitation is the specificity of onic cells (Figures 5B and 5C; Movie S6). The suspensor is laser disruption in the embryo at a single cell level. The nuclei

Developmental Cell 34, 242–251, July 27, 2015 ª2015 Elsevier Inc. 247 Figure 5. Laser Disruption of the Basal Cell Did Not Induce Cell-Fate Conversion (A) Basal cell of 1-cell-stage embryo targeted by laser irradiation in WOX2p::H2B-GFP (green) and WOX2p::tdTomato-LTI6b (magenta). (B) Suspensor cell of 2-cell-stage embryo targeted by laser irradiation in an ovule of WOX2::nDsRed (magenta) and WOX8::nVenus (green, nuclei localization) fertilized by DRN::erGFP (green, ER localization). Arrowheads indicate laser-irradiated cells. Numbers indicate the time from initiation of observations after laser irradiation (hr: min). Scale bars, 20 mm. See also Figures S1 and S2 and Movies S4, S5, and S6. (C) All cells were traced to 4-cell-stage embryo in WOX2p::H2B-GFP line after laser irradiation to apical or basal cell. The branch lengths indicate the mean of the cell-cycle duration (n = 7–8 [apical irradiation] in the proembryo, n = 4–7 [apical irradiation], n = 4 [basal irradiation] in the suspensor for each stages). Wild-type represents in Figure 2A. The lightning bolt represents the time irradiated by laser. The time of apical or basal cell division were corresponded to that of wild-type. (D) Schematic illustration of cell fate conversion induced by disruption of an apical cell. After apical-cell disruption, the basal cell first divides horizontally (black arrow) to give rise to extra-embryonic cells. After extra-embryonic properties (exemplified by WOX8 expression; green) are lost in the upper cell, embryonic properties are attained (DRN; magenta), and the upper cell divides vertically (blue arrow). As shown at the bottom, in some cases, more than two upper cells divides vertically (blue arrows) after two rounds of longitudinal divisions (black arrows). were crowded from the globular stage embryo. The early embryo (Figure 5D). One important mechanism involved in intercellular from 1-cell to 16-cell stages would be most optimal for laser communication during early embryogenesis is lateral inhibition, disruption. which prevents transformation in the fate of a basal cell (to that of an apical cell) in the presence of an apical cell. Such lateral in- Live-Cell Analysis of Early Embryogenesis hibition controls cell pattern formation via intercellular communi- In this study we undertook live-cell analysis of early embryogen- cation in a variety of plant developmental processes, such as esis in Arabidopsis thaliana. A combination of live-cell imaging root epidermis, trichome, and stomata development (Schiefel- and manipulation of cells or molecules in real time revealed a bein et al., 2014; Torii, 2012). In addition, temporal control of virtually invariant cell division pattern and duration during early cell fate conversion is important. Apical-cell damage results in embryogenesis. Combining the live-cell imaging system with ge- cell fate conversion after basal-cell division (Figure 5D). Conver- netic analysis (fluorescent marker lines and mutants) showed sion of basal cell fate before cell division causes loss of the that endosperm development is not required for pattern forma- suspensor cell, resulting in abnormal proliferation and pattern tion in the proembryo. Moreover, application of optical manipu- formation in the proembryo, as we observed after disruption of lation techniques coupled with the in vitro ovule culture system the basal cell. Thus, embryonic cells may produce an inhibitory enabled examination of the contributions of embryo and endo- signal to extra-embryonic cells, which act as reserve embryonic sperm development to embryogenesis at the cellular level. cells, and to ensure the supply of nutrients. Our live-cell analysis We also revealed the dynamics of cell fate conversion at the revealed the timeline of cell fate conversion at the cellular and molecular level using fluorescent marker lines in Arabidopsis molecular levels. This strategy might be fundamental for various

248 Developmental Cell 34, 242–251, July 27, 2015 ª2015 Elsevier Inc. species, as several studies have described cell fate conversion thickness. After the micropatterns of the photomask were transferred onto the in the suspensor in response to abnormal proembryo develop- SU-8 layer using a mask aligner (MJB4; SU¨ SS MicroTec AG), SU-8 development (Haccius, 1955; Marsden and Meinke, 1985; Vernon and ment was performed using a SU-8 developer (Nippon Kayaku) followed by a post-baking process. For the PDMS (Sylgard184; Dow Corning) casting pro- Meinke, 1994; Weijers et al., 2003). The time lag of cell fate cess, the PDMS mixture was degassed twice to remove air bubbles from the conversion is important for this compensation (Figure 5D). AG- fabricated SU-8 micro mold with a high aspect ratio. After the first degassing dependent histone demethylation controls the timing of was performed during mixing of the PDMS constituents, the PDMS mixture expression of the WUS repressor KNU (a 2-day time lag after was poured onto the SU-8 mold. The second degassing was performed for expression of AG) during flower development. This 2-day time more than 1 hr. The PDMS was cured at 90 C for 60 min in a convection lag, which corresponds with one to two cell-division cycles, is oven. After peeling the PDMS from the SU-8 mold, PDMS devices were cut into smaller sizes to fit inside a glass-bottom dish (D111300; Matsunami Glass). required for coordinated stem cell termination (Sun et al., 2009). Further optical manipulation with the in vitro ovule culture Microscopy system will provide novel insights into the molecular mecha- For live imaging of in vitro embryo development, we used four microscope sys- nisms of intercellular communications in plant embryogenesis. tems: three spinning-disk confocal microscopes and a two-photon excitation microscope. Confocal images were acquired using an inverted fluorescence EXPERIMENTAL PROCEDURES microscope (IX-81; Olympus) equipped with an automatically programmable XY stage (MD-XY30100T-Meta; Molecular Devices), a disk-scan confocal sys- Plant Materials tem (CSU-X1; Yokogawa Electric), 488-nm and 561-nm LD lasers (Sapphire; For all experiments, Arabidopsis thaliana accession Columbia (Col-0) was Coherent), and an EMCCD camera (Evolve 512; Photometrics). Time-lapse im- used as the wild-type. The following transgenic lines have been described pre- ages were acquired with a 403 water-immersion objective lens (UApo 40XW3/ viously: WOX2p::NLS-DsRed2, WOX8 delta::NLS-3xVenus (Ueda et al., 2011), 340, working distance [WD] = 0.25 mm, numerical aperture [NA] = 1.15; DRNp::erGFP (Cole et al., 2009), ATML1p::NLS-3xGFP (Takada and Ju¨ rgens, Olympus) or with a 303 silicone oil immersion objective lens (UPLSAPO30XS, 2007), and kpl RPS5Ap::H2B-tdTomato (Maruyama et al., 2013). WD = 0.80 mm, NA = 1.05; Olympus) mounted on a Piezo focus drive (P-721; Physik Instrumente). We used two band-pass filters, 520/35 nm for GFP, and Cloning, Transformant Generation, and Plant Growth Conditions 600/37 nm for tdTomato. Images were processed with Metamorph (Universal For WOX2p::H2B-GFP and WOX9p::H2B-GFP, the 3.5-kb WOX2 promoter Imaging) to create maximum-intensity projection images and to add color. (Ueda et al., 2011) or the 4.5-kb WOX9 promoter and WOX9 second intron, For higher-resolution live imaging, we used an inverted fluorescence micro- and the full-length coding region of H2B (At1g07790) fused to GFP and the scope (IX-83; Olympus) equipped with an automatically programmable XY NOS terminator, were cloned into the binary vector pMDC99 or pMDC123 stage (BioPrecision2; Ludl Electronic Products), a disk-scan confocal system (Curtis and Grossniklaus, 2003). For WOX2p::tdTomato-LTI6b and WOX9p:: (CSU-W1; Yokogawa Electric), 488-nm and 561-nm LD lasers (Sapphire; tdTomato-LTI6b, the 3.5-kb WOX2 promoter or the 4.5-kb WOX9 promoter, Coherent), and an EMCCD camera (iXon3 888; Andor Technologies). Time- and tdTomato fused to the start codon of LTI6b (At3g05890) with the lapse images were acquired with a 603 silicone oil immersion objective lens

(SGGGG)2 linker, were cloned into the binary vector pMDC99 or pMDC123. (UPLSAPO60XS, WD = 0.30 mm, NA = 1.30; Olympus) mounted on a Piezo The binary vectors were introduced into Agrobacterium tumefaciens strain focus drive (P-721; Physik Instrumente). We used two band-pass filters, EHA105. The floral-dip or inoculation methods were used for Agrobacte- 520/35 nm for GFP, and 593/46 nm for tdTomato. Images were processed rium-mediated Arabidopsis transformation (Clough and Bent, 1998; Narusaka with Metamorph (Universal Imaging) to create maximum-intensity projection et al., 2010). images and to add color. Seeds of A. thaliana were sown on plates containing half-strength Murashige To evaluate the in vitro ovule culture medium, we used an inverted confocal and Skoog (1/2 MS) salts (Duchefa Biochemie B.V.), 0.05% (w/v) MES-KOH microscope system with a stable incubation chamber (CV1000; Yokogawa (pH 5.8), 13 Gamborg’s vitamin solution (Sigma), and 1.5% (w/v) agar. The Electric) equipped with 488-nm and 561-nm LD lasers (Yokogawa Electric), seeds were incubated in a growth chamber at 22C under continuous lighting and an EMCCD camera (ImagEM 1K C9100-14 or ImagEM C9100-13; Hama- after cold treatment at 4C for 2–3 days. Two-week-old seedlings were trans- matsu Photonics). Time-lapse images were acquired with a 403 objective lens ferred to soil (Sakata no Tane; Sakata Seed) and grown at 22C under contin- (UPLSAPO40X, WD = 0.18 mm, NA = 0.95; Olympus). We used two band-pass uous lighting. filters, 520/35 nm for GFP, and 617/73 nm for tdTomato. For laser disruption, we used a laser scanning inverted microscope (A1R In Vitro Ovule Culture MP; Nikon) equipped with a Ti:sapphire femtosecond pulse laser (Mai Tai The liquid medium for in vitro ovule culture contained the Nitsch basal salt DeepSee; Spectra-Physics). Time-lapse images were acquired using a 253 mixture (Duchefa), 5% (w/v) trehalose dihydrate (Wako Pure Chemical), water-immersion objective lens (CFI Apo LWD 253 WI, NA = 1.10, WD = 0.05% (w/v) MES-KOH (pH 5.8), and 13 Gamborg’s vitamin solution (Sigma). 2.00 mm; Nikon) with a handmade water-supply system to enable long-term To optimize of medium components, the inclusion of 1/2MS salts (Duchefa), time-lapse imaging. A Ti:sapphire laser tuned to 950 nm was used for laser 5% (w/v) sucrose, 5% (w/v) glucose, and 400 mgmlÀ1 glutamine was evaluated. disruption and excitation of the fluorescent proteins. Fluorescence signals Immature siliques were selected and placed on double-sided tape attached were detected by the external non-descanned photomultiplier (PMT) or GaAsP to a microscope slide. Under a stereomicroscope, the siliques were opened PMT detectors. We used two dichroic mirrors, DM495 and DM560, and two along the replum with a dissecting needle and the ovules were transferred to band-pass filters, 534/30 nm for GFP or Venus, and 578/105 nm for tdTomato dishes containing in vitro ovule culture medium. The ovules were then trans- or DsRed. Images were processed with NIS-Elements AR 4.10 software ferred to a multi-well glass-bottom dish (D141400; Matsunami Glass) or micro- (Nikon) to create maximum-intensity projection images and to add color. Laser devices in a glass-bottom dish (D111300; Matsunami Glass) as described irradiation was performed by applying a femtosecond-pulse laser at 950 nm to below. The glass-bottom dish was filled with liquid medium and sealed with the nucleus region of the targeted cell. When the targeted cell shrunk immedi- parafilm, then placed on the microscope for observation. ately after laser irradiation, the ovule around the targeted cell was also damaged and became shrunken. Therefore, we adjusted the laser power to Fabrication of PDMS Micropillar Array Device eliminate the fluorescence signals of H2B-sGFP or NLS-DsRed2 without The micropillar array devices were fabricated using microfabricated SU-8 shrinking the targeted cell just after laser irradiation. photoresist (SU-8 3050; MicroChem) on a silicon wafer as described previously (Park et al., 2014). The photomasks were fabricated using a maskless lithog- Statistics raphy system (DL-1000; NanoSystem Solutions). SU-8 photoresist was first Post hoc multiple comparison t tests using the Holm-Sidak method were spin-coated on a silicon wafer at 1,000 rpm. The spin-coating and pre-baking performed using GraphPad Prism version 6.0e for Mac OS X (GraphPad processes were performed twice to generate the SU-8 layer of 200- to 300-mm Software).

Developmental Cell 34, 242–251, July 27, 2015 ª2015 Elsevier Inc. 249 SUPPLEMENTAL INFORMATION Hamamura, Y., Saito, C., Awai, C., Kurihara, D., Miyawaki, A., Nakagawa, T., Kanaoka, M.M., Sasaki, N., Nakano, A., Berger, F., and Higashiyama, T. Supplemental Information includes Supplemental Experimental Procedures, (2011). Live-cell imaging reveals the dynamics of two sperm cells during dou- two figures, and six movies and can be found with this article online at ble fertilization in Arabidopsis thaliana. Curr. Biol. 21, 497–502. http://dx.doi.org/10.1016/j.devcel.2015.06.008. Hamamura, Y., Nishimaki, M., Takeuchi, H., Geitmann, A., Kurihara, D., and Higashiyama, T. (2014). Live imaging of calcium spikes during double fertiliza- AUTHOR CONTRIBUTIONS tion in Arabidopsis. Nat. Commun. 5, 4722. Jenik, P.D., Jurkuta, R.E., and Barton, M.K. (2005). Interactions between the D.K. conceived the study; K.G. and D.K. designed the experiments; K.G and cell cycle and embryonic patterning in Arabidopsis uncovered by a mutation D.K. carried out most experiments; K.A. carried out laser disruption analysis; in DNA polymerase ε. Plant Cell 17, 3362–3377. J.P. and H.A. fabricated the PDMS microdevice; K.G., K.A., and D.K. analyzed the data; M.U., T.H. and D.K. supervised the project; K.G. and D.K. wrote the Kawashima, T., and Goldberg, R.B. (2010). The suspensor: not just suspend- manuscript; and M.U. and T.H. edited the manuscript. ing the embryo. Trends Plant Sci. 15, 23–30. Kirch, T., Simon, R., Gru¨ newald, M., and Werr, W. (2003). The ACKNOWLEDGMENTS DORNROSCHEN/ENHANCER OF SHOOT REGENERATION1 gene of Arabidopsis acts in the control of meristem ccll fate and lateral organ develop- We thank R. Groß-Hardt, S. Takada, and W. Werr for the plant materials; S. ment. Plant Cell 15, 694–705. Nasu, T. Nishii, and T. Shinagawa for assistance with cloning and generation Kurihara, D., Hamamura, Y., and Higashiyama, T. (2013). Live-cell analysis of of transgenic plants; and Y. Hamamura and N. Kaji for technical advice. Micro- plant reproduction: live-cell imaging, optical manipulation, and advanced scopy was conducted at the Institute of Transformative Bio-Molecules (WPI- microscopy technologies. Dev. Growth Differ. 55, 462–473. ITbM) of Nagoya University and supported by the Japan Advanced Plant Science Network. This research was supported by grants from the Japan Lau, S., Slane, D., Herud, O., Kong, J., and Ju¨ rgens, G. (2012). Early embryo- Science and Technology Agency (ERATO project to T.H. and M.U.) and the genesis in flowering plants: setting up the basic body pattern. Annu. Rev. Plant Ministry of Education, Culture, Sports, Science and Technology, Japan (no. Biol. 63, 483–506. 11025940 to D.K., no. 24113514 and 24770045 to M.U.). Li, J., and Berger, F. (2012). Endosperm: food for humankind and fodder for scientific discoveries. New Phytol. 195, 290–305. Received: March 13, 2015 Revised: May 1, 2015 Llorente, B.E., Juarez, L.M., and Apo´ stolo, N.M. (2007). Exogenous trehalose Accepted: June 8, 2015 affects morphogenesis in vitro of jojoba. Plant Cell Tissue Organ Cult. 89, Published: July 9, 2015 193–201. Mansfield, S.G., and Briarty, L.G. (1991). Early embryogenesis in Arabidopsis REFERENCES thaliana. II. The developing embryo. Can. J. Bot. 69, 461–476. Mansfield, S.G., Briarty, L.G., and Erni, S. (1991). Early embryogenesis in Arata, Y., Takagi, H., Sako, Y., and Sawa, H. (2015). Power law relationship be- Arabidopsis thaliana. I. The mature embryo sac. Can. J. Bot. 69, 447–460. tween cell cycle duration and cell volume in the early embryonic development of Caenorhabditis elegans. Front. Physiol. 5, 529. Marsden, M.P.F., and Meinke, D.W. (1985). Abnormal-development of the suspensor in an embryo-lethal mutant of Arabidopsis thaliana. Am. J. Bot. 72, Boisnard-Lorig, C., Colon-Carmona, A., Bauch, M., Hodge, S., Doerner, P., 1801–1812. Bancharel, E., Dumas, C., Haseloff, J., and Berger, F. (2001). Dynamic ana- lyses of the expression of the HISTONE:YFP fusion protein in Arabidopsis Maruyama, D., Hamamura, Y., Takeuchi, H., Susaki, D., Nishimaki, M., show that syncytial endosperm is divided in mitotic domains. Plant Cell 13, Kurihara, D., Kasahara, R.D., and Higashiyama, T. (2013). Independent control 495–509. by each female gamete prevents the attraction of multiple pollen tubes. Dev. Breuninger, H., Rikirsch, E., Hermann, M., Ueda, M., and Laux, T. (2008). Cell 25, 317–323. Differential expression of WOX genes mediates apical-basal axis formation Masui, Y., and Wang, P. (1998). Cell cycle transition in early embryonic devel- in the Arabidopsis embryo. Dev. Cell 14, 867–876. opment of Xenopus laevis. Biol. Cell 90, 537–548. Clough, S.J., and Bent, A.F. (1998). Floral dip: a simplified method for Mizuta, Y., Kurihara, D., and Higashiyama, T. (2015). Two-photon imaging with Agrobacterium-mediated transformation of Arabidopsis thaliana. Plant J. 16, longer wavelength excitation in intact Arabidopsis tissues. Protoplasma. 735–743. Published online January 15, 2015. http://dx.doi.org/10.1007/s00709-014- Cole, M., Chandler, J., Weijers, D., Jacobs, B., Comelli, P., and Werr, W. 0754-5. (2009). DORNROSCHEN is a direct target of the auxin response factor Narusaka, M., Shiraishi, T., Iwabuchi, M., and Narusaka, Y. (2010). The floral MONOPTEROS in the Arabidopsis embryo. Development 136, 1643–1651. inoculating protocol: a simplified Arabidopsis thaliana transformation method Costa, L.M., Marshall, E., Tesfaye, M., Silverstein, K.A., Mori, M., Umetsu, Y., modified from floral dipping. Plant Biotechnol. 27, 349–351. Otterbach, S.L., Papareddy, R., Dickinson, H.G., Boutiller, K., et al. (2014). Park, J., Kurihara, D., Higashiyama, T., and Arata, H. (2014). Fabrication of Central cell-derived peptides regulate early embryo patterning in flowering microcage arrays to fix plant ovules for long-term live imaging and observa- plants. Science 344, 168–172. tion. Sens. Actuators B 191, 178–185. Curtis, M.D., and Grossniklaus, U. (2003). A gateway cloning vector set for Ron, M., Alandete Saez, M., Eshed Williams, L., Fletcher, J.C., and high-throughput functional analysis of genes in planta. Plant Physiol. 133, McCormick, S. (2010). Proper regulation of a sperm-specific cis-nat-siRNA 462–469. is essential for double fertilization in Arabidopsis. Genes Dev. 24, 1010–1021. Fernandez, O., Be´ thencourt, L., Quero, A., Sangwan, R.S., and Cle´ ment, C. (2010). Trehalose and plant stress responses: friend or foe? Trends Plant Ryan, A.B., Castillo, A.M., Valles, M.P., Sanz, J.M., and Cistue, L. (1999). Sci. 15, 409–417. Desiccated doubled-haploid embryos obtained from microspore culture of barley cv. Igri. Plant Cell Rep. 18, 924–928. Haccius, B. (1955). Experimentally induced twinning in plants. Nature 176, 355–356. Sauer, M., and Friml, J. (2004). In vitro culture of Arabidopsis embryos within Haecker, A., Gross-Hardt, R., Geiges, B., Sarkar, A., Breuninger, H., their ovules. Plant J. 40, 835–843. Herrmann, M., and Laux, T. (2004). Expression dynamics of WOX genes Schiefelbein, J., Huang, L., and Zheng, X. (2014). Regulation of epidermal cell mark cell fate decisions during early embryonic patterning in Arabidopsis thali- fate in Arabidopsis roots: the importance of multiple feedback loops. Front. ana. Development 131, 657–668. Plant Sci. 5,47.

250 Developmental Cell 34, 242–251, July 27, 2015 ª2015 Elsevier Inc. Sun, B., Xu, Y., Ng, K.H., and Ito, T. (2009). A timing mechanism for stem cell Weijers, D., Van Hamburg, J.P., Van Rijn, E., Hooykaas, P.J., and Offringa, R. maintenance and differentiation in the Arabidopsis ﬂoral meristem. Genes Dev. (2003). Diphtheria toxin-mediated cell ablation reveals interregional communi- 23, 1791–1804. cation during Arabidopsis seed development. Plant Physiol. 133, 1882–1892. Takada, S., and Ju¨ rgens, G. (2007). Transcriptional regulation of epidermal cell Wendrich, J.R., and Weijers, D. (2013). The Arabidopsis embryo as a miniature fate in the Arabidopsis embryo. Development 134, 1141–1150. morphogenesis model. New Phytol. 199, 14–25. Torii, K.U. (2012). Two-dimensional spatial patterning in developmental sys- Yadegari, R., and Goldberg, R.B. (1997). Embryogenesis in dicotyledonous tems. Trends Cell Biol. 22, 438–446. plants. In Cellular and Molecular Biology of Plant Seed Development, B.A. Ueda, M., and Laux, T. (2012). The origin of the plant body axis. Curr. Opin. Larkins and I.K. Vasil, eds. (Springer Netherlands), pp. 3–52. Plant Biol. 15, 578–584. Yamaguchi, H., Sasaki, K., Shikata, M., Aida, R., and Ohtsubo, N. (2011). Ueda, M., Zhang, Z., and Laux, T. (2011). Transcriptional activation of Trehalose drastically extends the in vitro vegetative culture period and facili- Arabidopsis axis patterning genes WOX8/9 links zygote polarity to embryo tates maintenance of Torenia fournieri plants. Plant Biotechnol. 28, 263–266. development. Dev. Cell 20, 264–270. Yoshida, S., Barbier de Reuille, P., Lane, B., Bassel, G.W., Prusinkiewicz, P., Vernon, D.M., and Meinke, D.W. (1994). Embryogenic transformation of the Smith, R.S., and Weijers, D. (2014). Genetic control of plant development by suspensor in twin, a polyembryonic mutant of Arabidopsis. Dev. Biol. 165, overriding a geometric division rule. Dev. Cell 29, 75–87. 566–573. Yu, M., and Zhao, J. (2012). The cytological changes of tobacco zygote and Vo¨ lz, R., Heydlauff, J., Ripper, D., von Lyncker, L., and Groß-Hardt, R. (2013). proembryo cells induced by beta-glucosyl Yariv reagent suggest the involve- Ethylene signaling is required for synergid degeneration and the establishment ment of arabinogalactan proteins in cell division and cell plate formation. BMC of a pollen tube block. Dev. Cell 25, 310–316. Plant Biol. 12, 126.

Developmental Cell 34, 242–251, July 27, 2015 ª2015 Elsevier Inc. 251