Live-cell analysis of plant reproduction: Live-cell imaging, optical manipulation, and advanced microscopy technologies

Authors

  • Daisuke Kurihara,

    Corresponding author
    1. JST, ERATO, Higashiyama Live-Holonics Project, Nagoya, Aichi, Japan
    • Division of Biological Science, Graduate School of Science, Nagoya, Aichi, Japan
    Search for more papers by this author
  • Yuki Hamamura,

    1. Division of Biological Science, Graduate School of Science, Nagoya, Aichi, Japan
    Search for more papers by this author
  • Tetsuya Higashiyama

    Corresponding author
    1. JST, ERATO, Higashiyama Live-Holonics Project, Nagoya, Aichi, Japan
    2. Institute of Transformative Bio-Molecules (ITbM), Nagoya University, Nagoya, Aichi, Japan
    • Division of Biological Science, Graduate School of Science, Nagoya, Aichi, Japan
    Search for more papers by this author

Authors to whom all correspondence should be addressed.

Emails: kuri@bio.nagoya-u.ac.jp; higashi@bio.nagoya-u.ac.jp

Abstract

Sexual reproduction ensures propagation of species and enhances genetic diversity within populations. In flowering plants, sexual reproduction requires complicated and multi-step cell-to-cell communications among male and female cells. However, the confined nature of plant reproduction processes, which occur in the female reproductive organs and several cell layers of the pistil, limits our ability to observe these events in vivo. In this review, we discuss recent live-cell imaging in in vitro systems and the optical manipulation techniques that are used to capture the dynamic mechanisms representing molecular and cellular communications in sexual plant reproduction.

Sexual reproduction (fertilization) in plants

Sexual reproduction is a key process to conserve species and diversity in the life cycle of higher plants. In flowering plants, sexual reproduction requires complicated and multi-step cell-to-cell communications among male and female cells (reviewed in Suzuki 2009, Higashiyama 2010). At the time of pollination, pollen (male gametophyte) and the embryo sac (female gametophyte) are spatially separated (Fig. 1). When a pollen grain adheres to a papilla cell on the stigma, the pollen grain becomes hydrated, then germinates a pollen tube, and then enters the style. The pollen tube is a tip-growing cell, and it grows directionally thorough the stigma and style to enter the ovary. The pollen tube grows and penetrates through the placenta to the ovule. The target embryo sac in the ovule navigates the directional growth of the pollen tube. Consequently, the pollen tube grows a long way to the embryo sac inside the pistil without losing its way. Thus, pollen tube guidance is the mechanism by which female organs navigate the pollen tube from the stigma to the target embryo sac via multi-step controls (reviewed in Okuda & Higashiyama 2010, Takeuchi & Higashiyama 2011).

Figure 1.

Schematic representation of sexual reproduction in flowering plants. (A) A pollen grain adheres to a papilla cell of the stigma, then germinates a pollen tube and enters the style. The pollen tube elongates directionally through the stigma and style to enter the ovary. Finally, the pollen tube arrives at the target embryo sac. (B) After arriving at the embryo sac, two sperm cells within the pollen tube are discharged into the embryo sac. One sperm cell fuses with the egg to form the embryo, and other fuses with the central cell producing the endosperm. (C) After fertilization, the zygote elongates and divides asymmetrically into an apical cell and a basal cell. At the 8-cell embryo stage, the upper tier develops into the shoot and the lower tier into the hypocotyl and root. The embryo then passes through the heart-stage and torpedo stage before the seed reaches maturity.

After arriving at the embryo sac, direct interactions between the male gametophytes and the female gametophyte are required for fertilization. Two sperm cells within the pollen tube are discharged into the embryo sac through an interaction between the pollen tube and the embryo sac. One of these two sperm cells fuses with the egg to form the embryo, and the other fuses with the central cell to produce the endosperm, the nutritive tissue supporting embryo growth. Thus, flowering plants have evolved a unique reproductive process known as double fertilization (Lord & Russell 2002; Berger et al. 2008). The double fertilization process is achieved when all of the above cell-to-cell communication stages are successful among male and female cells.

Each fertilized ovule develops into a seed and the ovary develops into a fruit. The basic body organization of the plant seedling is laid down during embryogenesis, which is characterized by highly reproducible cell division patterns (reviewed in Laux & Jurgens 1997, Lau et al. 2012). After fertilization, the zygote elongates and divides asymmetrically to form a small cytoplasmic apical cell, which is the precursor of the embryo proper, and a large vacuolated basal cell, which develops into the suspensor. After three rounds of cell divisions, the apical cell generates an 8-cell embryo (proembryo), which is connected to the ovule by the suspensor. At the 8-cell embryo stage, the upper tier develops into the shoot and the lower tier into the hypocotyl and root. The embryo then passes through the heart stage and torpedo stage before the seed reaches maturity.

Sexual reproduction in higher plants is a highly dynamic process that requires the spatial and temporal coordination of multiple cells. However, there is still much to learn about the molecular mechanisms and behavior of gametophytic cells and fertilized cells during pollen tube guidance, double fertilization, and embryogenesis. Sexual reproduction processes occur within the pistil. Because pollen tube guidance, double fertilization, and embryogenesis occur in female reproductive organs with several cell layers in the pistil, it has been difficult to observe these important events that occur during sexual plant reproduction. Therefore, the development of in vitro (or semi-in vitro) systems is very important for detailed investigations of plant reproduction processes. Figure 2 shows a schematic representation of in vitro plant reproduction systems. By controlling environmental conditions (medium, temperature, etc.) and components (proteins, cells, and tissues), in vitro systems allow detailed investigations of plant reproduction processes at the cellular and molecular level. In this review, we discuss live-cell imaging of various events in plant sexual reproduction, such as pollen tube guidance, double fertilization, and embryogenesis, obtained using in vitro systems and optical manipulation.

Figure 2.

In vitro analysis of plant reproduction. Schematic representations of: (A) in vitro pollen germination (Taylor & Hepler 1997; Johnson-Brousseau & Mccormick 2004; Boavida & Mccormick 2007), (B) in vitro fertilization (Kranz & Lorz 1993; Antoine et al. 2000; Uchiumi et al. 2007), (C) in vitro embryogenesis (Kurihara et al. 2011), (D) in vitro ovule culture (Sauer & Friml 2004), and (E) semi-in vitro fertilization (Higashiyama et al. 1998; Palanivelu & Preuss 2006; Hamamura et al. 2011).

Live-cell imaging of pollen tube guidance in the Torenia system

During fertilization of flowering plants, the pollen tube grows directionally inside the pistil and delivers the immotile male gametes to the embryo sac. Pollen tube guidance is the mechanism by which female organs navigate the pollen tube from the stigma to the target embryo sac via multi-step controls. In the absence of the embryo sac, the pollen tube cannot reach the ovule; the female sporophytic tissues of the ovule are not sufficient to guide the pollen tube. Thus, the target embryo sac itself is required for pollen tube guidance (Hulskamp et al. 1995; Ray et al. 1997; Shimizu & Okada 2000). Live-cell imaging of pollen tube guidance would be useful to investigate whether the pollen tube is specifically attracted to the embryo sac in vivo; however, this is difficult because the embryo sac is embedded in several cell layers of the pistil.

In flowering plants, the embryo sac is generally located inside the ovule and is covered by thick layers of sporophytic tissues. These sporophytic tissues of the ovule cannot be readily removed to isolate the embryo sac. However, the embryo sac protrudes from the ovule in some plant species (Maheshwari 1950), including Torenia fournieri. In this species, the egg cell, the two synergids, and half of the central cell are located outside the ovule.

A system to study pollen tube guidance in vitro was developed using Torenia (Higashiyama et al. 1998). After optimizing the culture medium, both the ovule and the pollen tube, which was grown semi-in vitro through a cut pistil, were co-cultivated in a thin layer of solid medium. A pollinated style was cut and placed on the medium, and the pollen tubes emerged from the cut end of the style and continued to grow in the medium. When pollen grains were directly germinated on the medium (in vitro condition), no pollen tubes were attracted to the embryo sac. This indicated that female sporophytic tissues are important for the capability of the pollen tube attraction to the embryo sac (Higashiyama et al. 1998). Thus, both pollen tube attraction to the embryo sac and double fertilization occurred in this semi-in vitro Torenia system (Fig. 3A).

Figure 3.

Live-cell imaging of pollen tube guidance using the semi-in vitro Torenia system. (A) Dark-field image of pollen tube attraction in in vitro Torenia system at 8 h after the start of cultivation. Pollen tubes growing through cut style are co-cultivated with ovules in a dish. (B) Montage of representative images showing pollen tube attraction in in vitro Torenia system. Numbers indicate time (mm:ss) from first frame. Also see Movie S1. (C) Pollen tube growing through cut style attracted by LURE2 in gelatin bead (4 n mol/L). Numbers indicate time (mm:ss). Green fluorescence shows 10 kDa Alexa Fluor remaining 15 min after positioning gelatin bead (Fluorescence). Arrowheads indicate position of pollen tube tip in relation to gelatin beads (asterisks) containing recombinant LURE2 and 10 kDa Alexa Fluor (00:00). Scale bars: 30 μm (B) and 50 μm (C). Figures reproduced with permission from American Association for the Advancement of Science (A: Higashiyama et al. 2001) and Springer (B: Higashiyama & Hamamura 2008).

As shown in Figure 3A, attracted pollen tubes were observed frequently in the semi-in vitro Torenia system. However, it was very difficult to observe pollen tube guidance in real time because of the sensitivity of the pollen tube to light. Therefore, time-lapse observation with light microscopy was used for live-cell imaging of pollen tube guidance in the semi-in vitro Torenia system. Movie S1 shows time-lapse imaging at 2-s intervals of pollen tube guidance in T. fournieri. As shown in Figure 3B and Movie S1, the pollen tube grew toward the protruding embryo sac, precisely growing towards the entrance site of the embryo sac. When the ovule attracting the pollen tube was moved by micromanipulation with a glass needle, the pollen tube still moved toward the entrance site of the embryo sac. This behavior of pollen tubes suggested that the entrance site of the embryo sac releases some diffusible attractant, and that the pollen tubes are trapped at the highest concentration of the attractant. Thus, this semi-in vitro Torenia system allows direct observation of pollen tube guidance, and also allows the use of optical manipulation.

The source of the attractant was identified by laser ablation experiments (Higashiyama et al. 2001). When one synergid cell was ablated, the attraction became weaker, but did not completely disappear. However, when two synergid cells were ablated, the embryo sac did not attract pollen tubes. Thus, at least one synergid cell is necessary for pollen tube attraction. These experiments indicated that the diffusible attraction signal is derived from the synergid cell (Higashiyama et al. 2001).

Next, gene expression in the synergid cell was analyzed using EST (expressed sequence tag) data (Okuda et al. 2009). The synergid cells were manually separated as protoplasts with cell-wall degradation enzymes and isolated using a micropipette under a microscope. The cDNA library was constructed from 25 synergid cells. Interestingly, the most abundant proteins in the synergid cells were cysteine-rich polypeptides (CRPs) with a secretion sequence. When genes encoding these CRPs were expressed in Escherichia coli, the purified recombinant proteins showed pollen-tube attraction activity. When a gelatin bead containing recombinant CRP was placed in front of the pollen tube using a micromanipulator, the pollen tube grew towards the bead (Fig. 3C). Thus, two CRPs, designated LURE1 and LURE2, were identified as the pollen tube attractants in T. fournieri (Okuda et al. 2009).

A novel microinjector, the laser-assisted thermal-expansion microinjector (LTM), was developed to downregulate LUREs in Torenia (Okuda et al. 2009). This device uses extremely high pressure to inject substances into cells via a capillary with a fine tip (0.1 μm in diameter). Therefore, the LTM can inject substances into plant cells even with an intact cell wall and high turgor pressure. The LTM was used to microinject antisense oligos for LURE1 and LURE2 into the embryo sac. As a result, the frequency of pollen tube attraction was decreased compared with that of embryo sacs injected with inverted-sequence oligos. These results further indicated that LURE1 and LURE2 are involved in pollen tube attraction by the synergid cells (Okuda et al. 2009).

The pollen tube attractants, LUREs, were identified using the semi-in vitro Torenia system, live-cell imaging, and manipulation of cells and genes under a microscope. To study gene function, generation of loss-of-function mutants by insertional mutagenesis and knockdown by RNA interference are widely used in plant research. However, coupling new model systems with live-cell imaging techniques would greatly enhance understanding these mechanisms, even if it is a non-model organism for which there is limited genome information and/or genetic/molecular manipulation techniques available. By developing new systems and techniques, transient alterations of cell activity and gene-expression patterns coupled with live-cell imaging allow investigation of cellular events, even in non-model organisms.

Live-cell imaging of double fertilization in Arabidopsis thaliana

During double fertilization, two dimorphic female gametes (the egg cell and central cell) are fertilized by two immotile and isomorphic sperm cells delivered from the pollen tube. The two sperm cells are arranged in tandem with the leading pollen tube nucleus. After the sperm cells are discharged from the pollen tube, one fertilizes the egg cell to form the embryo and the other fertilizes the central cell to form the endosperm. Despite the fact that double fertilization is a unique event in flowering plants, the dynamics of individual sperm cells after their release into the female tissue remain largely unknown (Berger et al. 2008; Berger et al. 2010).

As described in the previous section, pollen tube attraction was observed in Torenia by differential interference contrast (DIC) images (Higashiyama et al. 1997, 1998, 2000). However, sperm cells could not be observed clearly by DIC images of fertilization in the ovules alone. Therefore, a live-cell imaging system using semi-in vitro double fertilization of Arabidopsis thaliana was developed (Palanivelu & Preuss 2006; Hamamura et al. 2011). A. thaliana is a model plant, and is useful for construction of fluorescent marker lines to visualize specific cells. After hand pollination, the pistil was cut horizontally to collect the stigma and the pistil was dissected to remove the ovules. The stigma was placed horizontally onto culture medium and the ovules were placed immediately on the same medium around the pistil. The pollen tubes germinated on the stigma, grew through the style, and then elongated on the medium. Finally, the pollen tubes entered the ovules on the medium. Thus, both pollen tube attraction to the embryo sac and double fertilization occurred in this semi-in vitro Arabidopsis system. When the pollen grains were directly germinated on the medium (in vitro condition), few pollen tubes (~ 3%) were attracted to the ovules in A. thaliana (Palanivelu & Preuss 2006). Considering that the pollen tubes were required to grow through female sporophytic tissues for the attraction, the cell-to-cell communication between male and female cells is important even for the capability of the pollen tube attraction.

There were some practical problems to solve for successful use of the semi-in vitro fertilization assay. The low fertilization rate presented a major problem, as double fertilization rarely occurred even when the pollen tube entered the ovule. To overcome this problem, the flower samples, temperature, and humidity were optimized during the preparation stages of the semi-in vitro assay. Stamens were sampled from the full-opening flower. The specimens were prepared under the same conditions of temperature and humidity as plant growth room. These optimizations improved the fertilization rate twice (Hamamura Y., unpubl. data, 2007); however, pollen tube discharge and double fertilization occurred suddenly and in a very short time, and it was still difficult to predict their timing. In addition, pollen tube discharge and double fertilization are photo-sensitive phenomena, and so real-time imaging was difficult. To solve these problems, microscopy settings were optimized by using spinning-disk confocal microscope and other equipment as described below. Next, the specific behavior of sperm cell just before pollen tube discharge was searched so that the timing of discharge could be predicted more accurately.

Live-cell imaging of double fertilization in this semi-in vitro Arabidopsis system was performed using a spinning-disk confocal microscope (Hamamura et al. 2011). This method allowed rapid collection of z-stack images with multiple colors, and facilitated imaging without damaging cells because laser irradiation was reduced. Moreover, a piezo-driven objective was used for further rapid collection of z-stack images. Time-lapse imaging of multiple ovules by using an automatic stage-controller and time-lapse imaging allowed efficient observation of double fertilization. The stages from pollen tube discharge to the fusion of sperm nuclei with target female nuclei (karyogamy) could be observed continuously. The nuclei of sperm cells were labeled with HTR10 (a histone H3.3-like protein specifically expressed in sperm cells) fused to monomeric red fluorescent protein (mRFP) (Ingouff et al. 2007). During pollen tube elongation, sperm cells moved together back and forth in the pollen tube. Movie S2 shows time-lapse imaging at 1-min intervals of fertilization in A. thaliana. After pollen tube discharge, the two sperm cells were rapidly transported into the female gametophyte between the egg cell and the central cell (Fig. 4 and Movie S2). After staying together at the apical region of the degenerated synergid cell, the two sperm cell nuclei began to move toward their respective target nuclei in the egg cell and central cell. Live-cell imaging of the entire double fertilization processes in Arabidopsis revealed the behavior of sperm cells from their release from the pollen tube to the fusion of nuclei, a process comprising three steps (Hamamura et al. 2011, 2012).

Figure 4.

Live-cell imaging of double fertilization using semi-in vitro Arabidopsis system. Sperm cell dynamics during double fertilization were observed in Arabidopsis thaliana. Sperm cell nuclei (SCN: shown by arrow heads) were labeled with HTR10pro:HTR10-mRFP. Female gametophyte nuclei were labeled with ACT11pro:H2B-GFP. SYN: Synergid cell nucleus. ECN: Egg cell nucleus. CCN: Central cell nucleus. Numbers indicate time (min). Last frame just before rapid movement of sperm cells was designated as time of pollen tube discharge, i.e., 0 min. Scale bar: 10 μm. Also see Movie S2.

In the first step, during the release of sperm cells, the two sperm cells were rapidly transported between the egg cell and the central cell in 8.8 ± 5.5 s at 10 μm/s. The maximum velocity was observed immediately after discharge from the pollen tube. This rapid movement of sperm cells from the pollen tube was possibly due to the plasmoptysis-like flow of the pollen tube cytoplasm (Higashiyama et al. 2000; Hamamura et al. 2011).

In the second step, the sperm cells remained together after their release, and were located between the egg cell and the central cell for approximately 7.4 min (Hamamura et al. 2011). Two-photon microscopy, which allows deep imaging at high resolution compared with single-photon microscopy, was used to examine this process. The boundary region between the central cell and the egg cell was clearly observed by two-photon microscopy. After their discharge from the pollen tube, the nuclei of sperm cells were delivered between the egg cell and the central cell boundary region, and remained in contact with both cells. This immobile phase between the egg and central cells might be required for cell-to-cell communications between male and female gametophytes to determine the fertilization targets. In Arabidopsis, the two sperm cells are isomorphic, and cannot be distinguished morphologically. Therefore, a photo-convertible fluorescent protein, monomeric Kikume Green-Red (reviewed in Lukyanov et al. 2005), was used to distinguish between the two sperm cells by their position in the pollen tube. As the pollen tube elongated, the sperm cells remained in the same positional order in the tube. Photo-conversion analysis showed that both sperm cells in the pollen tube were equally capable of fertilizing the egg cell and the central cell. Thus, the isomorphic sperm cells of Arabidopsis did not show a preference with regard to their fertilization target based on their position in the pollen tube (Hamamura et al. 2011).

In the third step, the sperm cell nuclei suddenly moved to the female gametophyte nuclei after the immobile phase (Hamamura et al. 2011). When sperm cell nuclei reinitiated their movement, the sperm cell nucleus entered the cytosol of the central cell and egg cell, and then moved towards cell each nucleus. This observation suggested that reinitiation of movement of the sperm cell nucleus marks the fusion of cytoplasms, namely fertilization (plasmogamy).

Live-cell imaging of the entire double-fertilization process was achieved using the semi-in vitro Arabidopsis system. These observations suggested that cell-to-cell communications were required for sperm cells to coordinate double fertilization of the two female gametes. When mis-delivery of a sperm cell occurred between the egg cell and the central cell, it seemed that the sperm cells could not normally reinitiate movement toward the female gametes (Hamamura Y., unpubl. data, 2011). These findings raised the possibility that the two sperm cells have equal ability to fertilize each female gamete, and their fertilization targets are likely determined during the immobile phase between the egg cell and the central cell. To reveal these mechanisms of double fertilization, further research should be conducted on cell-to-cell communications between the two male gametes and the two female gametes during this phase by live-cell imaging and optical manipulation.

Live-cell imaging of embryogenesis in Arabidopsis thaliana

After double fertilization, the fertilized egg cell (zygote) forms the embryo and the fertilized central cell forms the endosperm. During embryogenesis, the zygote follows a pattern of cell division to form the body plan of the embryo. The origin of seedling tissues can be traced back to a specific group of cells in the early embryo in A. thaliana (Torres-Ruiz & Jurgens 1994). Although it is suggested that there is spatial regulation of cell division by cell-to-cell communication during pattern formation in the embryo, there is still much to learn about the regulation of cell division during embryo development in flowering plants. The main reason for the current gaps in knowledge about this process is that it is difficult to observe cell division in the embryo and pattern formation in vivo because the embryo is embedded in the maternal tissue. To date, most studies in this area have used fixed materials such as cleared seeds or histological sections of seeds. Investigating cell-to-cell communication during pattern formation in the embryo requires live-cell imaging with transient manipulations of genes, signaling molecules, and cells.

For detailed observations of the Arabidopsis embryo, embryos were dissected from the ovules. Under a dissecting microscope, the siliques were cut along the replum with a needle and the ovules were transferred into culture medium. The ovules were cut open with forceps and the embryos were transferred into a glass-bottomed dish with a micropipette (Fig. 5A). Using isolated embryos expressing green fluorescent protein (GFP)-α-tubulin, the dynamics of microtubules were observed during mitosis in Arabidopsis torpedo-stage embryos using a spinning disk confocal microscope system. At the onset of mitosis, the preprophase band appeared at the future division plane. After nuclear envelope breakdown (NEBD), kinetochore microtubules were organized at prometaphase and phragmoplasts appeared at anaphase and telophase (Fig. 5B and Movie S3). The dynamics of microtubules in isolated embryos were consistent with those observed in Arabidopsis root cells (Maizel et al. 2011). Live-cell image analysis of cell division could be performed during embryogenesis using the in vitro system with isolated embryos. So far, the dynamics of AtHaspin, the mitotic kinase involved in cell division during mitosis, have been observed in isolated torpedo-stage embryos (Kurihara et al. 2011).

Figure 5.

Live-cell imaging of embryogenesis using in vitro Arabidopsis system. (A) Embryos were isolated from ovary (2–3 days after flowering) and transferred into liquid medium under a stereomicrosope. (B) Microtubule dynamics were observed in torpedo-stage embryo. Microtubules were labeled with RPS5Apro:GFP-TUA1. Numbers indicate time (min) from first frame. Pre-prophase band (5 min; arrowheads), spindle microtubules (30 min; arrow), and phragmoplast (40 min; asterisk) were observed. Also see Movie S3. (C) Live-cell imaging was carried out using embryos isolated after >1 h treatment with control (0.1% dimethylsulfoxide [DMSO]) or 5 μmol/L hesperadin (Aurora kinase inhibitor). Nuclei and chromosomes were labeled with RPS5Apro:H2B-tdTomato. Hesperadin induced early separation of chromosomes before completion of chromosome alignment at equator (Hesperadin; arrowheads). Numbers indicate time (min) from first frame. Also see Movie S4 and Movie S5. (D) Comparison of one- and two-photon imaging of embryo in isolated ovule of Arabidopsis thaliana. Embryo was observed using a 20 × objective lens under both one-photon (488-nm, 637-nm, left) and two-photon (920-nm, right) excitation. Nuclei and chromosomes were labeled with HTR8pro:HTR8-GFP. Scale bars: 50 μm (A and D) and 10 μm (B and C).

This in vitro Arabidopsis system allows real-time manipulation of gene and proteins by inhibitor treatments. Aurora kinases function as key regulators during mitosis in yeast, metazoans, and plants (Demidov et al. 2005; Kawabe et al. 2005; Carmena et al. 2009). Treatment with an aurora kinase inhibitor, hesperadin, delays chromosome alignment and causes lagging anaphase chromosomes without inhibiting mitotic progression in plants (Kurihara et al. 2006, 2008; Demidov et al. 2009). The dynamics of H2B-tdTomato were observed in isolated heart-stage embryos using hesperadin treatment. During prometaphase, chromosomes failed to align on the spindle equator in hesperadin-treated cells (Fig. 5C, Hesperadin), whereas normal alignment was observed in control cells (Fig. 5C, Control). In hesperadin-treated embryos, chromosomes decondensed without aligning on the spindle equator, and did not segregate (Fig. 5C, Hesperadin, arrowheads). Thus, inhibitor assays could be applied successfully using isolated embryos in the in vitro system of Arabidopsis embryogenesis.

Proper cell division occurs only using isolated late embryo, that is, in heart-stage and torpedo-stage embryos, in this in vitro system. We could not observe the further cell division in isolated early embryos, indicating that the development of isolated early embryo is arrested in this in vitro system (Kurihara D., unpubl. data, 2009). The ovule and endosperm are nutrient tissues that support the growth of the embryo via the suspensor during early embryogenesis (Kawashima & Goldberg 2010). Hence, using the isolated ovule with this in vitro system may allow live-cell imaging during the entire process of embryogenesis. Sauer & Friml (2004) developed a system for in vitro culture of the embryo within the ovule in A. thaliana. This in vitro system allowed culturing of embryos on agar medium from the one-cell stage for up to 3 days without aberrations. However, embryos cultured using this method showed a low survival rate, and the method was less suitable for very early stage embryos, such as zygotes. Moreover, live-cell imaging has not yet been performed using this in vitro system. In addition, embryogenesis in Arabidopsis takes a long time (4–5 days from zygote to torpedo-stage embryo). As the embryo and endosperm develop, the ovule expands and the distance of the embryo from the glass surface increases to approximately 100–200 μm. Thus, a system in which there is less cell damage and that allows deep imaging is necessary for long-term live-cell imaging of embryogenesis.

Two-photon excitation microscopy allows deeper imaging of tissues than wide-field and one-photon excitation microscopy (reviewed in Helmchen & Denk 2005). Because a visible-light laser produces excitation regions, which generate heat and dispersion around the focus plane, out-of-focus light is rejected by a variable pinhole in one-photon excitation microscopy. In addition, one-photon excitation microscopy is less suitable for analyses of tissues below the surface because strong light scattering occurs at depth in tissues. In contrast, two-photon excitation of fluorescent molecules is restricted to the focus plane, because the two photons are absorbed simultaneously in the two-photon excitation process. Therefore, in two-photon microscopy, all emitted light comes from the focal plane, enabling collection in any direction, including scattered photons. Moreover, because it uses less scattering long-wavelength light, which allows deeper penetration; two-photon excitation microscopy can be used for high-resolution, deep imaging of thin optical sections.

Figure 5D shows a comparison of confocal and two-photon microscopy images of an embryo in an isolated Arabidopsis ovule. In the globular-stage embryo, the nuclei could not be clearly observed by confocal microscopy, but each nucleus was clearly observed in the image obtained by two-photon microscopy. Thus, two-photon excitation microscopy allows clear discrimination of each nucleus in embryo cells, even in isolated ovules.

Optical manipulation of gene expression and protein inactivation

Analyses using these in vitro systems have combined spatio-temporal gene manipulation, cell ablation, and live-cell imaging to explore the regulatory cell-to-cell communications involved in plant sexual reproduction and pattern formation. During early embryogenesis, the polarized zygote divides asymmetrically and generates apical and basal cell fates. WUSCHEL RELATED HOMEOBOX (WOX) transcription factors, WOX2 and WOX8 are co-expressed in zygote, whereas they are expressed specifically in the apical and basal cell lineage after the zygote division, respectively. These expression patterns and knockout analysis of WOX genes shows WOX expressions are important for pattern formation during early embryogenesis (reviewed in Ueda & Laux (2012)). However, the direct effect of the WOX expression is unclear, because it is difficult to spatio-temporally induce the ectopic expression of WOX genes during early embryogenesis.

Recently, the infrared laser-evoked gene operator (IR-LEGO) system was used to induce a heat-shock response in targeted single cells. In this system, an IR laser (wavelength of 1480 nm) was used to induce expression of a transgene under the control of a heat-shock promoter (Kamei et al. 2009). When applied to the lateral root in A. thaliana, IR-LEGO induced GFP expression in targeted single cells (Fig. 6). Heat shock-induced cells subsequently developed normally, suggesting that the IR-LEGO-mediated heat shock induction did not cause significant cell damage. The IR-LEGO system could be applied for gene induction at the single-cell level in plants.

Figure 6.

Infrared laser-evoked gene operator (IR-LEGO)-induced green fluorescent protein (GFP) expression in Arabidopsis lateral root. H2B-GFP expression was observed 2 h after IR laser irradiation (arrows; 11 mW) in lateral root. Irradiation times are shown in left panel (0 h). IR laser irradiation was carried out under 40 × objective lens. Scale bar: 50 μm.

In addition to inducing gene expression at the single-cell level, inactivation of a target protein inside living cells with high spatio-temporal resolution has been achieved by chromophore-assisted light inactivation (CALI) (Tour et al. 2003). Some technologies, such as knockout mutants, or knockdown by antisense oligo or RNA interference, are powerful tools for investigating the functions of the target protein in vivo. However, the disadvantage of these methods is low time resolution because the already expressed protein could not be inhibited. CALI uses a photosensitizer such as KillerRed, which efficiently generates short-lived reactive oxygen spices (ROS) that damage proteins in the immediate vicinity of the chromophore (Bulina et al. 2006). By expressing a KillerRed-fused target protein, CALI can be used for highly specific spatio-temporal manipulation of protein activities or cell ablations to investigate cell-to-cell communications and cell-fate decisions in real time. Several factors involved in pollen tube guidance, pollen perception, and gamete fusion, have been identified by using knockout mutants (reviewed in Kawashima & Berger 2011). CALI using KillerRed-fused these factors enable us to investigate when and where these factors play a key role in cell-to-cell communications.

Conclusions and future perspectives

The live-cell imaging obtained using in vitro systems and optical manipulation techniques in T. fournieri and A. thaliana has captured some of the factors involved in the dynamic mechanisms representing molecular and cellular communications in plant sexual reproduction. As described above, fluorescent proteins are powerful tools for live-cell imaging, as they allow observations of specific cells or cellular components, such as particular organelles, proteins, etc. Bright-field imaging can also provide useful information about dynamic cellular processes. As shown in Movie S1, bright-field microscopy can reveal details of several dynamic events, such as cytoplasmic streaming and organelle movement both in the pollen tube and embryo sac, without labeling. Several reproduction processes, such as female gametogenesis, pollen tube attraction, and double fertilization, have been clearly observed by bright-field microscopy in the naked embryo sac of T. fournieri (Higashiyama et al. 1997, 1998). Moreover, T. fournieri is also a transformable plant (Aida 2012). Therefore, the combination of bright-field and fluorescence microscopy in T. fournieri will be useful for analyses of cell-to-cell communications in female gametogenesis, pollen tube attraction, and double fertilization.

The advantage of in vitro or semi-in vitro systems is that they can analyze the focused phenomena while extracting some factors. However, there are some differences between the in vitro and in vivo environments. In the case of semi-in vitro Torenia and Arabidopsis systems, the pollen tube grows on the agarose plate without the placenta inside the pistil. The dynamics of fluorescent labeled LURE2 were observed on the agarose plate in the semi-in vitro Torenia system (Goto et al. 2011), whereas the dynamics of LURE peptides from the embryo sac to the pollen tube remains unclear in vivo. Although A. thaliana have 98% fertility in vivo, few pollen tube attractions and double fertilizations occurred in the semi-in vitro systems (Palanivelu & Preuss 2006; Kasahara et al. 2012). Our final goal is to visualize the whole-phenomena in the whole-body. Cheung et al. attempted to observe the in vivo pollen germination in intact plants by two-photon microscopy (Cheung et al. 2010). The pistil of whole plant placed near the microscope was mounted on a microchamber and the pollen tubes expressing GFP were observed by two-photon microscopy. After live-cell imaging, the whole plant could be further cultured in the greenhouse (Cheung et al. 2010). Considering the clear images of embryo in an isolated Arabidopsis ovule, two-photon microscopy is also a powerful tool for in vivo imaging.

Microdevices have also been developed for biological applications by microscale technology (Weibel et al. 2007). MEMS (micro electro mechanical systems)-based techniques, which evolved from the semiconductor and microelectronics industries, can be used to directly fabricate 3D structures at length scales ranging from 0.1 to 100 μm (Whitesides et al. 2001). Microscale approaches can be used directly to fabricate devices reflected in the tissue environment and to study cell-to-cell communications and signaling molecules in controlled conditions of cell-microenvironment interactions, even in vitro. Recently, MEMS devices were developed to investigate the behavior of pollen tubes to the physical features of microchannel in vitro using Camellia japonica pollen (Agudelo et al. 2012). In semi-in vitro conditions, Yetisen et al. developed a MEMS device for pollen tube guidance assay to the isolated ovules in A. thaliana (Yetisen et al. 2011). Our group performed semi-in vitro pollen tube guidance assay to the recombinant LURE peptides in T. fournieri with MEMS device (Horade et al. 2012). Advanced collaborations between engineering and biological sciences could enhance our ability to provide physiological models in vivo to further our understanding of fundamental biology.

Microsystem technology can potentially mimic the microenvironment within the pistil, allowing observations of cellular behavior, the dynamics of signaling molecules, and gene activity for pollen tube guidance, double fertilization, and embryogenesis in plants. Together with the spatial and temporal control of gene, protein, and cell activity in real time, the application of these technologies will provide a new view of the mechanisms of cell-to-cell communications and pattern formation in plant sexual reproduction.

Acknowledgments

We thank Satohiro Okuda (Nagoya University) for images shown in Figure 3C. We acknowledge support from Grants-in-Aid for Scientific Research from the Ministry of Education, Culture, Sports, Science and Technology of Japan (No. 11025940 to D.K.).

Ancillary