2016 Volume 41 Issue 2 Pages 121-125
In flowering plants, fertilization of the central cell gives rise to an embryo-nourishing endosperm. Recently, we reported that the endosperm absorbs the adjacent synergid cell through a cell-fusion, terminating the pollen tube guidance by a rapid inactivation of the synergid cell. Although this synergid-endosperm fusion (SE fusion) initiates soon after fertilization, it was still unknown whether the triggers of SE fusion are stimuli during fertilization or other seed developmental processes. To further dissect out the SE fusion process, we investigated the SE fusion in an Arabidopsis mutant defective for MULTICOPY SUPPRESSOR OF IRA1 (MSI1), a subunit of the polycomb repressive complex 2 (PRC2). The mutant msi1 develops autonomous endosperm without fertilization. Time-lapse imaging revealed a rapid efflux of the synergid contents during the autonomous endosperm development, indicating that the initiation of SE fusion is under the control of some of the events triggered by fertilization of the central cell distinct from the discharge of pollen tube contents and plasma membrane fusion.
In animals, cell-fusions are found in various developmental processes including placenta formation, muscle development, and fertilization (Chen et al., 2007). However, only limited numbers of cell-fusion events have been reported in entire developmental processes of flowering plants, probably due to mechanical limitations imposed by the cell wall that prevent the direct contact of plasma membranes from neighboring cells. Recently, we characterized a novel cell-fusion event that takes place after double fertilization in Arabidopsis thaliana (See Fig. S1; Maruyama et al., 2015). The synergid cell is a female accessory cell that plays important roles during double fertilization by secreting pollen tube-attractant peptides (Kanaoka and Higashiyama, 2015; Qu et al., 2015). Most flowering plants have two synergid cells in the ovule (Yadegari and Drews, 2004). At the final stage of pollen tube guidance, one synergid cell degenerates and receives pollen tube contents including two sperm cells. The sperm cells independently fertilize the egg cell and the central cell. The fertilized egg cell becomes a zygote that develops into the embryo, while the fertilized central cell becomes the endosperm that supplies nutrients to the embryo like the placenta in mammals. Within a few hours after double fertilization, the endosperm fuses with the persistent synergid cell that did not receive the pollen tube. The cytoplasm of the persistent synergid is transported into the endosperm by this SE fusion, diluting the pollen tube attractant, and this event rapidly terminates the attraction of the pollen tube.
Interestingly, the induction of SE fusion requires central cell fertilization and can omit egg cell fertilization (Maruyama et al., 2015), indicating that particular stimuli during the gamete fusion or subsequent endosperm development trigger the SE fusion. Fertilization-independent seed (FIS) class polycomb repressive complex 2 (PRC2) is a negative regulator of endosperm differentiation (Köhler et al., 2012). FIS-PRC2 consists of four proteins: MEDEA (MEA) (Grossniklaus et al., 1998), FERTILIZATION INDEPENDENT SEED 2 (FIS2; Chaudhury et al., 1997), FERTILIZATION INDEPENDENT ENDOSPERM (FIE; Ohad et al., 1999), and MULTICOPY SUPPRESSOR OF IRA1 (MSI1) (Köhler et al., 2003; Guitton et al., 2004). Loss of function of the FIS-PRC2 components induces autonomous endosperm development without fertilization, characterized by nuclear proliferations in the central cell.
In this study, we analyzed the dynamics of the synergid cytoplasm during autonomous endosperm development of an Arabidopsis mutant of MSI1 and found that the autonomous endosperm absorbed both synergid cells, demonstrating that SE fusions can be triggered without certain steps of double fertilization.
pMYB98::GFP (Kasahara et al., 2005) and msi1-1 (Köhler et al., 2003) were crossed to generate msi1/+ pMYB98::GFP. Plants were grown in soil at 22°C under continuous light.
Microscopic observationOne, two, or three DAE ovules were dissected from the pistils into 10% glycerol. Images were acquired using an upright microscope (Axio Imager.A2; Zeiss, Jera, Germany) with a 40x objective lens (Plan-APOCHROMAT, NA=0.95; Zeiss) through a cooled CCD camera (AxioCam HRm; Zeiss). A spinning-disk confocal system (CellVoyager CV1000; Yokogawa Electric, Tokyo, Japan) equipped with a 40x dry objective lens (UPLSApo 40×2, WD=0.18 mm, NA=0.95; Olympus, Tokyo, Japan) was also used for time-lapse observation in 1.5 DAE ovules with N5T medium (Gooh et al., 2015). Images of six optical sections with 3-μm thickness were acquired every 30 min through an EMCCD camera (ImagEM C9100-13 512×512; Hamamatsu Photonics).
Image processingFluorescent images with multiple z-planes were processed with CV1000 software (Yokogawa Electric) to create the maximum intensity projection images. ImageJ (https://imagej.nih.gov/ij/) and QuickTime Player 7 Ver. 7.7.1 were used for image adjusted brightness and contrast and movie editing of the time-lapse analyses.
The msi1 mutation causes embryo lethality, and only the heterozygous msi1 mutant (msi1/+) can be recovered (Köhler et al., 2003; Guitton et al., 2004). To examine the behavior of synergid cell contents during autonomous endosperm development, we investigated the msi1/+ that is homozygous for a marker gene expressing cytosolic GFP from a synergid-specific MYB98 promoter (pMYB98::GFP; Kasahara et al., 2005). We removed stamens from flowers at the floral stage 12 (Smyth et al., 1990) and observed the development of fruits and seeds at six days after the emasculation (DAE) in the msi1/+ pMYB98::GFP siblings. Four of seven plants appeared as wild-type (WT) because they did not show any signs of autonomous seed development (Fig. 1A, C). By contrast, the other three displayed growth of the fruits (Fig. 1B) and enlargement of unfertilized ovules (Fig. 1D), which are typically induced by the msi1 mutation (Köhler et al., 2003; Guitton et al., 2004).
Selection of WT and msi1 mutant plants from siblings of msi1/+ pMYB98::GFP. (A and B) Arabidopsis pistils at 6 DAE. (C and D) Inside of the emasculated pistils shown in (A) and (B). Pistil in (A) did not grow, and the ovules in the pistil remained small (C) in WT pMYB98::GFP. By contrast, elongation of pistil (B) and asexual seed development occurred in ~50% of ovules (D) in msi1/+ pMYB98::GFP. Bars, 2 mm.
We then compared the GFP signal in the synergid cells between WT and msi1/+. In WT at three DAE and in msi1/+ at 1 DAE, the GFP signal was restricted in the synergid cells (Fig. 2A–D, G). In contrast, at two and three DAE, an increasing proportion of ovules from msi1/+ exhibited signal distribution in the autonomous endosperm (Fig. 2E, F, H).
Autonomously developed endosperm displayed an ectopic signal of GFP expressed under a synergid-specific gene promoter. (A to C) Merged images of GFP fluorescence and differential interference contrast (DIC) in ovules from a pistil of WT pMYB98::GFP at 1 DAE (A), 2 DAE (B), and 3 DAE (C). (D to F) Merged images of GFP fluorescence and DIC in ovules from a pistil of msi1/+ pMYB98::GFP at 1 DAE (D), 2 DAE (E), and 3 DAE (F). Bars, 20 μm. (G and H) Percentages of viable ovules exhibiting a GFP signal in the central cell dissected from the WT pMYB98::GFP (G) and msi1/+ pMYB98::GFP (H). Error bars are means and standard deviations calculated from the analysis of three pistils.
We have two hypotheses for the GFP signal in the autonomous msi1 endosperm. One is de novo ectopic expression of GFP from the MYB98 promoter caused by msi1 in endosperm. An alternative hypothesis is the diffusion of the GFP signal from synergid cells after their fusion with the autonomous endosperm. To examine these hypotheses, ovules were dissected at 1.5 DAE from WT and msi1/+ plants carrying pMYB98::GFP and subjected to time-lapse imaging for 12 hours by confocal laser scanning microscopy. In contrast with WT ovules that did not show any alteration of GFP signal (97%, n=71, Fig. 3A; Movie S1), we observed an abrupt decrease of GFP signal in the synergid cell (Fig. 3B; Movie S2) in half of the ovules from msi1/+ plants (47%, n=79). As the signal decreased in the synergid cell, the unfertilized central cell often exhibited an increasing GFP signal (Fig. 3B, arrowheads; Movie S2), indicating an efflux of synergid contents into the central cell after SE fusion (Fig. 3C). We also observed cases of consecutive SE fusions in msi1 ovules (Movie S3), an observation that is consistent with the absence of concentrated GFP signal in both synergid cells at 3 DAE (Fig. 2F).
Time-lapse imaging of autonomous synergid–endosperm fusions, related to Movie S1 to S3. (A) Time-lapse images of synergid contents in an ovule from a 1.5 DAE pistil in the WT pMYB98::GFP. (B) Time-lapse images of synergid contents in an ovule from a 1.5 DAE pistil in the msi1/+ pMYB98::GFP. Arrowheads indicate the GFP signal observed in the autonomously developing endosperm. Abbreviations: sy1, synergid cell 1; sy2, synergid cell 2. Bars, 20 μm. (C) Schematic drawings of normal SE fusion after double fertilization (top) and of autonomous SE fusions observed in the msi1 mutant ovule exhibiting autonomous endosperm development (bottom).
In this study, we found that synergid cells fuse with the central cell during autonomous endosperm development in msi1 ovules. The central cell of the mutants of FIS-PRC2 components including MSI1 is thought to mimic functional endosperm because the autonomous endosperm exhibited a similar gene expression pattern to normal endosperm (Guitton and Berger, 2005) and is able to nurse the embryo after single fertilization of the egg cell (Nowack et al., 2007). We previously demonstrated that fertilization of the central cell plays a major role in the initiation of SE fusion (Maruyama et al., 2015). However, central cell fertilization triggers various cellular alterations such as possible calcium spike upon plasma membrane fusion (Hamamura et al., 2014; Denninger et al., 2014) and the transition from quiescent central cell to actively proliferating endosperm. Thus, we could not identify the downstream process responsible for the SE fusion induction. Observations of autonomous SE fusion in the msi1 ovule support that the SE fusion is part of processes that follow fertilization of the central cell.
Synergid cells would be less active after the autonomous SE fusion, because msi1 ovules exhibit reduced pollen tube targeting and an increase of defective pollen tube reception as the timing of pollination delays (Rotman et al., 2008). However, some of the mutant ovules are fertile even after 3.5 DAE. By the three DAE, the SE fusion occurred in almost all of the msi1 ovules (Fig. 2H). Most likely, the synergid nucleus maintains its function, despite the synergid cytoplasm is diluted by SE fusion. Analyses of nuclear disorganization, pollen tube attraction, and gene expression in the autonomously developing ovules would provide novel insights into a mechanism of programmed cell death in the synergid cell.
We thank M. Ueda, M. Kanaoka, T. Sasaki and Y. Kimata for technical advices about imaging, Y. Sato for technical supports for imaging and comments to the manuscript. A part of this work was conducted in Institute of Transformative Bio-Molecules (WPI-ITbM) at Nagoya University, supported by Japan Advanced Plant Science Network. This work was supported by JSPS KAKENHI Grant Numbers 14J08936, 16J02257, 15K14541, and 16H06173, and the Japan Science and Technology Agency (ERATO project to T.H.).
