Sexual reproduction is an important biological event not only for evolution but also for breeding in plants. It is a well known fact that Charles Darwin (1809–1882) was interested in the reproduction system of plants as part of his concept of ‘species’ and ‘evolution.’ His keen observation and speculation is timeless even in the current post-genome era. In the Darwin anniversary year of 2009, I have summarized recent molecular genetic studies of plant reproduction, focusing especially on male gametophyte development, pollination and fertilization. We are just beginning to understand the molecular mechanisms of the elaborate reproduction system in flowering plants, which have been a mystery for >100 years.
The year 2009 is the bicentennial anniversary of Charles Darwin’s birth, and Darwin tribute reviews have also been published this year in the field of plant science (Fay and Chase 2009, Holland et al. 2009, Hopper and Lambers 2009, Moulia and Fournier 2009, McClure 2009). As is well known, his contribution to plant science involves a wide range of fields, such as plant speciation, orchid biology, phototropism, gravitropism and mating systems. In particular, his great insight into sexual plant reproduction cannot be forgotten. He was fascinated by elaborate plant reproductive systems and the importance of outcrossing, and his pioneer works on self-incompatibility (SI) are still well known among present SI researchers (McClure 2009). Thus, the end of Darwin’s anniversary year is a good time to summarize recent progress in plant reproduction research. Therefore, in this short review, I want to detail what has recently been discovered about plant reproduction, focusing especially on the story of the male gamete up to fertilization in the angiosperm. It will allow us to recall Darwin’s innovative studies as a plant scientist.
Pollen development and male sterility
The male gametophyte in flowering plants develops in the anther of the stamen (Fig. 1). While the animal germline is discriminated from somatic cells in the early developmental stage, pollen mother cells (PMCs) are differentiated from cells which are at the adequate location in the adequate late developmental stage; these are called archesporial cells (Scott et al. 2004). For successful reproduction, timing of production of the male and female gametophytes is important, and male gametophyte development is synchronically regulated in the same anther.
In the early stages of male gametophyte development, PMCs undergo meiosis to form haploid tetrad cells. As in other eukaryotes, a single DNA replication followed by two characteristic divisions (meiosis I and II) results in a reduction in ploidy in plant meiosis. During meiosis I, paired homologous chromosomes undergo homologous recombination and segregate from each other, retaining sister chromatid cohesion. As part of the research into a plant-specific phenomenon, several recent studies have attempted to explain molecular mechanisms of wheat Ph1, which is necessary for correct pairing of homologous chromosomes in the polyploid genome (Griffiths et al. 2006, Sidhu et al. 2008). In the subsequent meiosis II, sister chromatids are separated and, finally, four haploid cells, the tetrad, are produced. Gene regulation of meiosis has been intensively studied in yeast, and many plant orthologs of yeast meiosis genes share similar functions in plants (summarized in Mercier and Grelon 2008). This indicates that mechanisms of recombination and specialized chromosome segregation, which are characteristics of meiosis, are fundamentally common in eukaryotes. However, studies of plant orthologs sometimes provide useful new information for research into meiosis of other organisms (White 2008), and plant-specific meiosis genes were also reported (Nonomura et al. 2004, Mercier and Grelon 2008), suggesting the importance of plant meiosis research.
In a locule of the anther, the tetrad haploid microspores are released and mature to become pollen grains (McCormick 1993, Fig. 1). At first, the uninucleate microspore develops into the bicellular pollen with a larger vegetative cell and a smaller generative cell by an asymmetric mitosis. Then, the generative cell undergoes a second mitosis to form two sperm cells, which are finally used for double fertilization, as discussed below. In Arabidopsis, genetic studies using various mutants were conducted to analyze such male gametophyte development (summarized in Borg et al. 2009), e.g. analysis of fbl17 mutants revealed the control of male germ cell proliferation by the degradation of cell cycle inhibitors (Kim et al. 2008).
The layer of the anther cells closest to the locule, termed the tapetum, is metabolically active, and plays important roles in pollen maturation (Fig. 1). The involvement of phytohormones, gibberellin and auxin in pollen development was suggested by cell type-specific transcriptome analysis of rice microspore/pollen and tapetum (Hirano et al. 2008). During pollen maturation, the sporophytic tapetum is acting as a nutritive tissue by providing nutrition and materials for pollen wall formation, and disintegrates in the later stage of pollen development (Scott et al. 2004). In Brassica, tapetosomes, unique organelles in the tapetum, contain endoplasmic reticulum-derived vesicles including flavonoids, and oleosin-coated lipid droplets including alkanes (Hsieh and Huang 2007). After disintegration of the tapetum, these flavonoids, alkanes and oleosins are discharged to the surface of the maturing pollen as pollen coats and might play roles in pollen germination and pollen tube growth. Stable intracellular oil bodies with unique oleosin and caleosin were also found in lily mature pollen, which might serve as energy reserves for germination (Jiang et al. 2007, Jiang et al. 2008).
Formation of the surface structure of pollen grains (pollen walls) is an important process in pollen maturation (Scott et al. 2004). The pollen walls consist of inner pectocellulosic intine and outer sporopollenin-based exine. The exine provides the species-specific pollen surface structure and cavities storing the pollen coats, including lipids and proteins, which are mainly supplied from the tapetum. Screening of 12 kaonashi mutants showing abnormal exine structure in Arabidopsis makes it possible to understand more details of the genetic regulation of exine formation (Suzuki et al. 2008), in addition to other Arabidopsis exine mutants reported or reviewed in recent papers (Ariizumi et al. 2008, Guan et al. 2008, Dobritsa et al. 2009, Wilson and Zhang 2009). Transcripts encoding potential proteins having some function in exine synthesis were also observed from the transcriptome analysis of rice anther (Huang et al. 2009).
If we refer to pollen development, an explanation of cytoplasmic male sterility (CMS) is unavoidable, because it is an important agricultural trait used in hybrid breeding. Plant CMS shows a maternally inherited pollen sterility phenotype and is regulated by interaction between mitochondria and nuclei (reviewed in Chase 2007, Fujii and Toriyama 2008b). Defective pollen development caused by mitochondrial genomic organization can be restored by nuclear-encoded Rf genes. In petunia, radish and rice, recently cloned Rf genes encode mitochondria-targeted pentatricopeptide repeat (PPR) proteins, and it is considered that the PPR proteins play a role in post-transcriptional RNA modification of CMS-determining genes in mitochondria. Interestingly, more recently, another rice Rf gene against CW-type CMS cytoplasm (Rf17) was cloned and shown to encode a novel protein, RETROGRADE-REGULATED MALE STERILITY (RMS), whose down-regulation caused fertility restoration (Fujii and Toriyama 2009). Because there are many essential mitochondrial genes in the nuclear genome, retrograde signals from mitochondria to nuclei are undoubtedly important in plants (Fujii and Toriyama 2008b). Similar to the known examples of retrograde regulation in yeast RTG signaling, Drosophila cell cycle-related signaling and Arabidopsis plastid signaling, the plant CMS might be elaborately regulated by typical retrograde signaling (Fujii and Toriyama 2008a, Fujii and Toriyama 2008b).
Pollination and SI
Mature pollen grains are then carried by wind or insects onto the stigma of the pistil, also known as ‘pollination’ (Fig. 1). Pollination is an important gate before fertilization, and involves cell–cell communication between haploid (pollen) and diploid (stigma) cells. It must be noted that haploid pollen is coated by the pollen coat from diploid tapetum cells, as described above, so that the pollination process includes communication between molecules produced from parent diploid cells. These two different kinds of behavior are well reflected in the SI system; some species show gametophytic SI (GSI), while others show sporophytic SI (SSI). In this section, I want to summarize ‘pollination and SI’ based on recent important findings.
As I mentioned in the Introduction, Darwin had a great interest in plant reproduction, especially in pollination, SI and the benefit of outbreeding. He might wonder why so many plant species show self-sterility, even though only advantageous variations are supposed to survive through ‘natural selection’ and the ‘struggle for existence’ based on his theory of evolution. In his book ‘The Effects of Cross and Self-Fertilisation in the Vegetable Kingdom’, published in 1876, he performed a large number of crossing experiments and compared phenotypes of individuals obtained from cross-pollination and self-pollination in various flowering plants (Darwin 1876). Although he somehow misunderstood the evolution of SI as an incidentally acquired character dependent on environmental circumstances, it is noteworthy that he realized the existence of heterosis and the importance of SI from his exhaustive observations.
Now we know that SI is genetically regulated, although Darwin did not have knowledge of Mendelian genetics at that time. Darwin’s precise description about heterostyly in Primula is famous, but, on the other hand, recent molecular studies of SI have focused on homomorphic incompatibility. Here I summarize the known molecular mechanisms of three types of homomorphic SI systems: S-receptor kinase (SRK)-based SSI systems in Brassicaceae; S-RNase-based GSI systems in Solanaceae, Rosaceae and Plantaginaceae; and S-glycoprotein-based GSI systems in Papaveraceae. For readers who want to understand SI in more detail, a recently published book edited by Vernonica E. Franklin-Tong (Franklin-Tong 2008a) is an excellent resource.
In Brassica SSI, a small cysteine-rich protein, S-locus protein-11/S-locus cysteine-rich (SP11/SCR), in the pollen coat acts as a ligand of SRK, which is localized on the cytoplasmic membrane of the stigma papilla cell. SP11/SCR and an extracellular receptor domain of SRK are highly polymorphic and can interact only when their S haplotypes are the same (i.e. self-pollination), and they transduce ‘self-signals’ in the papilla cells with phosphorylation cascades (reviewed in Takayama and Isogai 2005). Thus, as a result of the SRK-mediated signal transduction, self-pollen is rejected on the stigma papilla cells. On the other hand, non-self-pollen can germinate, and its pollen tube is elongated and penetrates into the pistils for successful fertilization. Even though almost 10 years have passed since the long-anticipated pollen S gene was reported, signaling cascades between self-recognition and rejection of self-pollen are still largely unclear. Gene cloning from a self-compatible Brassica mutant revealed that M-locus protein kinase (MLPK) has a role as a positive regulator of SI and directly interacts with SRK in the plasma membrane (Kakita et al. 2007). The receptor complex activated by phosphorylation on the membrane of papilla cells might transduce the SI signal into the cells, a process which is possibly followed by several steps of phosphorylation cascades and results in the inhibition of self-pollen germination. ARM-repeat-containing protein (ARC1) is known as another positive regulator of SI, and signaling pathways with similar U-box/ARM-repeat-containing E3 ligases are conserved with other kinase proteins in Arabidopsis (Samuel et al. 2008), suggesting that common ubiquitin-mediated protein degradation pathways might be involved downstream of SRK and other receptor-like kinase (RLK) signaling. In the latest interesting findings, Ivanov and Gaude (2009) reported that SRK localizes predominantly to intracellular sorting endosomes with a negative regulator THL1, and Tantikanjana et al. (2009) showed the dual role of SRK in SI and pistil development by using an Arabidopsis rdr6 mutant. Thus, there are now many questions about the complicated SRK-mediated signal transduction leading to inhibition of self-pollination, though the SI recognition itself is simple and easy to understand.
In contrast, the mechanism of arresting self-pollen tube growth is simple to understand in the S-RNase-based GSI. The arrest of the self-pollen tube growth in the style transmitting tissue can be explained by the effect of RNA degradation in the pollen tube catalyzed by RNase activity of the self-S-RNase (reviewed in McClure 2009). However, recent identification of the pollen S determinant, S-locus F-box (SLF/SFB, Sijacic et al. 2004), makes it difficult to understand the co-evolution of pollen and style S determinants and the mechanisms of SI recognition (Newbigin et al. 2008, McClure 2009). In Antirrhinum, sequence divergence of SLFs is too low between S haplotypes, indicating that polymorphisms of SLF alleles have a much shorter evolutionary history than that of the S-RNase alleles, which is in contradiction to the long-held view of co-evolution of the SI genes (Newbigin et al. 2008). From the viewpoint of molecular mechanisms, both of two possible models, the ‘compartmentalization model’ and the ‘S-RNase degradation model’, have some deficiencies for a perfect explanation of the S-RNase-based SI (McClure 2009). Furthermore, two subfamilies of Rosaceae (Prunoideae and Maloideae) possess different characteristics for the existence of competitive interaction, and the copy number of SFB on the S-locus, suggesting the possibility of mechanistic divergence of this GSI (Sassa et al. 2009). Thus, some conflicts arise in understanding the puzzling nature of S-RNase-based SI. If further experiments in the future resolve these conflicts, whole mechanisms will be easily clarified, because there is nothing downstream of S-RNase.
The most exciting topic in recent SI studies is the identification of a pollen S product in the Papaver GSI system, reported by Wheeler et al. (2009). It means that this third set of the pollen/pistil SI factors was successfully identified in Papaver. They are designated as Papaver rhoeas pollen S (PrpS) and Papaver rhoeas stigma S determinant (PrsS). PrpS is a transmembrane protein localized to the pollen tube plasma membrane, and its putative extracellular loop segment interacts with PrsS, which is the stigma-expressed S-glycoprotein. Although the precise protein function of PrpS for activation of the SI response has not been determined, the S haplotype-specific interaction of PrpS and PrsS might trigger the influx of Ca2+ into the shank of the pollen tube, finally inhibiting pollen tube elongation (Franklin-Tong 2008b).
Generally, in compatible pollination, normal growth of the pollen tube requires lipids on the stigma (Wolters-Arts et al. 1998) and a precise cytoplasmic Ca2+ concentration in pollen tube tips (Iwano et al. 2009). In the growing pollen tube, mitochondria and Golgi vesicles are distributed differently, and their movement is regulated differently by microtubule-dependent and actin filament-dependent motors (Romagnoli et al. 2007). Successful pollen germination and pollen tube elongation in the stylar transmitting tissues lead to fertilization, the final step of the story of the male gametophyte (Fig. 1). How are the pollen tubes containing the sperm cells accurately guided to the embryo sac with the egg cell? Here, I now want to describe pollen tube guidance and fertilization.
Pollen tube guidance and fertilization
One of the most impressive findings in plant science reported this year is undoubtedly the pollen tube attractant molecules reported by Okuda et al. (2009). There are many mysterious phenomena in the fertilization process (reviewed in Berger et al. 2008). Since double fertilization was first identified by Sergius Nawaschin and Leon Guignard at the end of the 19th century (Nawaschin 1898, Guignard 1899), the mechanism of how to the pollen tube is attracted to the embryo sac has long been a mystery, because of the difficulty of observing the embryo sac embedded in the ovule (Fig. 1). By using the in vitro fertilization system of Torenia, Higashiyama and colleagues clearly demonstrated that the synergid cell is necessary for the short-range pollen tube guidance before fertilization (Higashiyama et al. 2001) and, after extensive experimentation, the authors finally identified molecules which can attract the pollen tube and are secreted from synergid cells (Okuda et al. 2009). The pollen tube attractants, designated as LURE1 and LURE2, are small cysteine-rich polypeptides (CRPs). It is interesting that both LUREs and SP11/SCR are categorized as CRPs as the signaling molecules involved in plant reproduction. This similarity inspired us to image the molecular recognition mechanism by which LUREs might act as ligands of unidentified RLKs localized on a tip of the pollen tube. Many CRPs were found in the Torenia synergid cells (Okuda et al. 2009), and several RLKs expressed in synergid cells and pollen tubes were recently identified as female and male factors, respectively, controlling pollen tube behavior in Arabidopsis (Escobar-Restrepo et al. 2007, Boisson-Dernier et al. 2009, Miyazaki et al. 2009), suggesting that complex direct communication between female and male cells before fertilization might be carried out via RLK-mediated signaling pathways.
After pollen tube arrival, growth arrest and pollen tube discharge, which are caused by interaction between synergids and pollen tubes as described above, the two released sperm cells migrate and fuse to the egg cell and central cell to accomplish double fertilization. In this step, the GENERATIVE CELL SPECIFIC1 (GCS1) protein localized on the membrane of generative cells is known to have an essential role in gamete attachment and fusion (Mori et al. 2006). GCS1 was first identified in the lily, and homologs were observed in Arabidopsis, other angiosperms, algae and parasites, suggesting its fundamental role in membrane fusion during fertilization. In most plant species, the two sperm cells are morphologically indistinguishable, and there is some debate as to whether or not the decision for each sperm cell to fertilize either the egg cell or the central cell is already accomplished in the pollen tube (Berger et al. 2008). A recent report by Ingouff et al. (2009) proposed that each of the two sperm cells shares an equal ability to fertilize the egg cell, using an Arabidopsis RETINOBLASTOMA RELATED 1 (rbr1) mutant having an abnormal embryo sac with two egg cells.
To date, various molecular players, whose disruption affects pollen tube guidance, have been reported. In Arabidopsis, the MYB98 transcription factor is necessary to form the filiform apparatus of synergids, which might be essential for secretion of pollen tube attractants (Kasahara et al. 2005). In addition to such synergid-mediated regulation, the CENTRAL CELL GUIDANCE (CCG) nuclear protein in the central cell (Chen et al. 2007), the membrane-bound HAP2 (GCS1) protein in the sperm cells (von Besser et al. 2006) and MAGATAMA3 (MAA3) helicase, which regulates female gametophyte development (Shimizu et al. 2008), also play a role in pollen tube guidance in Arabidopsis. It is expected that an integrative explanation of all the molecular mechanisms of pollen tube guidance will be found.
Species and reproductive barriers
This year is also known for the fact that it is the 150th anniversary of the publication of The Origin of Species (Darwin 1859). In this famous work of Darwin, he refuted the fixed definition of species as distinct creations of God. Categories of species, subspecies and variety are all part of a continuous lineage, and are defined arbitrarily. As is known, the most common biological definition of a species is an isolated group, whose members can cross with each other and produce fertile offspring. However, fertility is also defined arbitrarily and is determined differently by different researchers. In this regard, reproductive barriers might merely be spin-offs of evolution. Nevertheless, reproductive barrier, like interspecific incompatibility is still an attractive theme for researchers of plant and animal reproduction, because breaking the reproductive barriers will give us new important products. A classical Dobzhansky–Muller model explains a simple scenario to produce genetic incompatibility between isolated populations, and Brideau et al. (2006) showed a molecular-level explanation of the Dobzhansky–Muller model in Drosophila. In plants, interestingly, Bikard et al. (2009) reported intraspecific incompatibility in Arabidopsis thaliana caused by the Dobzhansky–Muller-like epistatic interaction. Reciprocal silencing of essential duplicated genes causes genetic incompatibility within species and could contribute to reproductive isolation and speciation. In general, reproductive barriers also exist in the pollination and fertilization steps described here. The relationship between interspecific incompatibility and SI (Hiscock and Dickinson 1993, Murfett et al. 1996) and species preference of the pollen tube attractant from synergid cells (Higashiyama et al. 2006) are both known. In addition to these reports, molecular mechanisms of reproductive barriers, in connection with the molecular evolution of key players of pollination and fertilization, will be uncovered in the near future.
Molecular techniques and bioimaging technology are making remarkable progress day by day in dissecting cell-level gene regulation and in visualizing the dynamic state of organelles, molecules and other materials. Laser microdissection (LM) technologies enable us to perform cell type-specific gene expression profiling in plants (Ohtsu et al. 2007). In fact, in plant reproduction research, LM microarray analysis was applied to the cell-type specific transcriptomes of the male gametophyte and tapetum in rice (Hobo et al. 2008, Suwabe et al. 2008, Watanabe 2008). Whole transcriptome analysis of a single cell using a powerful next-generation sequencing technology (F. Tang et al. 2009) can also be applied to complex plant sexual tissues. In bioimaging analysis, visualizing mitochondria and plastids in the living pollen provides us with useful information (Matsushima et al. 2008, L. Y. Tang et al. 2009), and precise monitoring of actin dynamics in papilla cells and Ca2+ dynamics in pollen tubes has been successful in pollination studies (Iwano et al. 2007, Iwano et al. 2009). By endless scientific trials with these emerging technologies, our molecular understanding of plant reproduction will advance step by step from one level to the next. Although we tend to consider the phenomena of life at a micro level, we should continue research with a broader perspective, as Darwin did.
The Ministry of Education, Culture, Sports, Science, and Technology of Japan (MEXT) [Grants-in-Aid for Special Research on Priority Areas (No. 18075003)].
The author thanks Drs. Takeshi Ishimizu (Osaka University), Keita Suwabe (Mie University) and Masumi Miyano (Tohoku University) for critical reading of the manuscript.
cytoplasmic male sterility
pollen mother cell