Patterning the Female Gametophyte of Flowering Plants

Over the last decades, the female gametophyte (FG) of flowering plants has attracted much attention. This is on one side due to the fact that this minute structure generates the female gametes, represents the site of fertilization, and marks the inception of new life. In addition, it has become clear that the FG serves as an attractive model system for studies of cell specification, cell-cell communication, and programmed cell death (PCD) (Lituiev et al., 2014; Sprunck and Gross-Hardt, 2011; Heydlauff and Gross-Hardt, 2014; Susaki et al., 2015). The knowledge that we generate is to a large extent dictated by the available techniques and resources. Accordingly, several of the new insights that we gained on development and function of the FG have been triggered by the establishment or adaptation of tools for the analysis of reproductive tissue. In light of these technical advances, we here discuss recent results and concepts that shed light on the complex development and function of flowering plant embryo sacs with an explicit focus on mechanisms underlying the specification of distinct cell identities and cell-cell communication. Most angiosperms, including important crop plants and Arabidopsis (Arabidopsis thaliana), develop a FG of the so-called polygonum type. This few-celled structure integrates and orchestrates various complex processes to ensure reproductive success. To reliably accomplish the different tasks, the FG is based on division of labor. Two female gametes, egg and central cell, get fertilized by one sperm each. While the fertilized egg gives rise to an embryo, the adjacent central cell proliferates to generate the endosperm, which nurtures the developing embryo. In addition, gamete adjoining accessory cells are formed: Two synergids differentiate close to the entrance point of the pollen tube, which is referred to as micropyle. These cells play a key role in the attraction of the pollen tube and mediate sperm cell release. In addition, three central cell adjoining antipodal cells form at the chalazal pole of the FG. While antipodals proliferate and contribute to the endosperm for example in maize (Zea mays), no clear function could yet be ascribed to these cells in the model plant Arabidopsis. The seven cells of the FG are of clonal origin and trace back to a single haploid megaspore, which forms after meiosis (Fig. 1A). The megaspore undergoes a coenocytic phase with three mitotic divisions without cytokinesis (Fig. 1, B–D). In the early eight-nucleate stage, four nuclei lie at either end of the FG (Fig. 1D). Subsequently, one nucleus from the micropylar end and one nucleus from the chalazal pole migrate to the center (Fig. 1, D and E). These nuclei are referred to as polar nuclei and become integrated into the central cell. Cellularization compartmentalizes the syncytium into seven cells, which undergo further morphological changes until maturity (Fig. 1F). Among the changes that can be observed are alterations in nuclear position and size between egg cell and synergids and fusion of the two polar nuclei to a prominent homodiploid nucleus. In addition, antipodals start to shrink and degenerate in a fertilization-independent process, which lasts a couple of days (Kägi et al., 2010; Song et al., 2014; Fig. 1G). Intra- and intercellular communication governs FG patterning. A to G, Graphical illustration of different FG developmental stages. A to D, The functional megaspore undergoes three mitotic divisions without cellularization. E and F, Nuclear migration (E) and cellularization (F) compartmentalize the FG. G, Mature FG with fused polar nuclei. H to K, Schematic representation of FG patterning depicting two distinct layers. The lower layer represents the establishment of the micropylar-chalazal axis (green and yellow, respectively). The second layer represents a FG after cellularization comprising the four distinct cell types superimposed onto the micropylar-chalazal axis. See text for further explanation. aux, Auxin; M, micropylar pole; C, chalazal pole; gam. com., gametic competence; syn, synergid cell; egg, egg cell; cc, central cell; ap, antipodal cell. To study FG development and function, various strategies have been employed with forward genetics being among the most successful approaches. Forward genetic approaches generated a plethora of female gametophytic mutants, the analysis of which has contributed much to our understanding of the molecular mechanisms governing FG development. In this chapter, we will summarize findings that have advanced our understanding of FG patterning and cell specification. The Arabidopsis retinoblastoma homolog RBR (Ebel et al., 2004) and the maize LOB-domain protein INDETERMINATE GAMETOPHYTE (Lin, 1978; Evans, 2007) are necessary to restrict nuclear proliferation in the FG, and the respective mutants give rise to supernumerary nuclei. Notably, the additional nuclei in ig1 mutants differentiate according to their relative position along the micropylar-chalazal axis into extra synergids, egg cells, and central cells (Guo et al., 2004). This cellular readout reveals the presence of polar positioning cues, which recapitulate the nucellar-chalazal axis of the maternal tissue. How is this polarity perceived and established? Several lines of evidence suggest auxin as a micropylar guidance clue. Auxin response reporters reveal a discrete hormone maximum in the sporophytic nucellar tissue during the initial stages of embryo sac formation (Pagnussat et al., 2009; Ceccato et al., 2013), and targeted disruption of auxin transport interferes with early steps in embryo sac development (Ceccato et al., 2013). In addition, cell specification in the FG is susceptible to FG-targeted changes in auxin concentration or auxin response (Pagnussat et al., 2009; Panoli et al., 2015). However, there is some uncertainty regarding the presence of auxin within the FG as two of three studies fail to detect auxin response in the haploid embryo sac using reporters commonly applied to monitor auxin response in diploid tissue (Pagnussat et al., 2009; Ceccato et al., 2013; Lituiev et al., 2013). While we cannot exclude the possibility that auxin-decoding mechanisms differ fundamentally between gametophytic and sporophytic tissue, these results suggest that the effect of auxin on FG development might be indirect (Ceccato et al., 2013; Lituiev et al., 2013). Important insights regarding the specification of the chalazal pole come from the analysis of the His kinase CKI1. CKI1 is related to the cytokinin receptors AHK2 to AHK4 but lacks an extracellular cytokinin-binding domain (Hwang and Sheen, 2001; Suzuki et al., 2001; Yamada et al., 2001). Still, ectopic CKI1 expression can activate the cytokinin two-component signaling pathway (Kakimoto, 1996). Mutations in the CKI1 gene interfere with embryo sac development (Pischke et al., 2002; Hejátko et al., 2003) and causes defects in the chalazal portion of the FG: Cell-type-specific reporter genes reveal a chalazal extension of egg and synergid identity reaching out to antipodals and central cell, respectively. Notably, while expressed during all stages of megagametogenesis (Rabiger and Drews, 2013; Yuan et al., 2016), CKI1 expression becomes restricted to the two chalazal nuclei at the four-nucleate stage (Yuan et al., 2016). As a consequence, the two polar nuclei of the central cell recruited from the micropylar and chalazal pole differ with respect to CKI1 expression. This asymmetric inheritance of CKI1 to the central cell might provide a means to install a micropylar-chalazal polarity to the central cell (Fig. 1H). The important role of CKI1 for cell patterning is further demonstrated by ectopic expression of the gene, which is sufficient to induce central cell marker expression at the position of synergids and egg cells and formation of extra endosperms at the expense of embryo (Yuan et al., 2016). Together, these findings indicate that CKI1 not only specifies central cell identity, but also restricts micropylar signatures to synergid and egg cell. Another mutant that can be reconciled with a shift or misinterpretation of the micropylar-chalazal axis is altered meristem program1 (amp1), which differentiates fully functional egg cells at the expense of synergid cells (Kong et al., 2015). Notably, the defect can be complemented by sporophytic AMP1 expression, suggesting the presence of a mobile factor generated in the maternal ovule tissue, which represses egg cell identity in synergids (Kong et al., 2015). wyrd mutants are defective in a putative Arabidopsis ortholog of the inner centromere protein (INCENP) and also exhibit a reprogramming of one or two synergids into egg cells and defects in central cell differentiation (Kirioukhova et al., 2011). Whether WYRD is involved in the initial establishment of cell identities or rather required to maintain an established pattern is currently unclear (Fig. 1I). Similarly, eostre mutants exhibit a second egg cell at the expense of one synergid (Pagnussat et al., 2007). The defect can be traced back to ectopic expression of the BEL1-like homeodomain 1 gene (BLH1) and requires the KNOX gene KNAT3. While BLH1 is not active in wild-type embryo sacs, its ectopic expression could interfere with a functional BELL-KNOX interaction, suggesting a critical role of KNAT3 in the regulation of egg cell identity. By contrast, an expansion of chalazal identities is observed in FGs defective for the MYB transcription factors 119 and 64 (Rabiger and Drews, 2013). MYB64 is expressed in all cells of the FG, whereas MYB119 is restricted to central cell and antipodals. Notably, MYB119 is down-regulated in cki1 mutants. However, given the opposing roles of the genes, with CKI1 promoting chalazal identity while MYB119 is repressing it, this appears to be mainly due to a failure of cki1 to establish MYB119-expressing cells (Fig. 1H). Finally, verdandi mutants, which are defective in a putative transcription factor of the B3 superfamily, provide an intriguing phenotype with both synergids and antipodals exhibiting defects in the activation of cell-type-specific molecular marker. The defect is associated with an infrequent activation of a synergid reporter in antipodal cells and vice versa (Matias-Hernandez et al., 2010). The phenotype potentially indicates (1) a defect in the establishment or interpretation of micropylar-chalazal guidance cues, (2) a common module for the regulation of accessory cells, or (3) traces back to a defect already immanent in the one-nucleated embryo sac, which is perpetuated until the mature stage when the defect becomes manifest (Matias-Hernandez et al., 2010). In addition, several transcription factors have been implicated in the differentiation of FG cells. In mutants defective for MYB98, the filiform apparatus, a cell wall protrusion characteristic to synergids, is not properly formed (Fig. 1J). The defect correlates with a failure to attract pollen tubes and has supported the idea that the filiform apparatus is involved in the secretion of pollen tube attractants (Punwani et al., 2007). On the basis of a wheat (Triticum aestivum) egg cell transcriptome, the plant-specific RWP-RK-domain-containing (RKD) family of transcription factors was identified. AtRKD1 and AtRKD2 are specifically activated in Arabidopsis egg cells, and, remarkably, their overexpression activates an egg-cell-like transcript profile in callus tissue (Koszegi et al., 2011). These results suggest that RKD1 and RKD2 key to the activation of egg cell identity (Fig. 1J). Due to genetic redundancy of this gene family in Arabidopsis, fully RKD-depleted FGs are not yet available. In light of these complications, recent work in Marchantia polymorpha is very interesting. Compared to Arabidopsis, this simple liverwort exhibits a smaller genome, which reduces complexity. In fact, Marchantia encodes a single RKD copy only, advancing loss-of-function studies. Rövekamp et al. (2016) and Koi et al. (2016) generated RNAi and recombination-induced mutants, respectively. Both studies report defects in egg cell development. However, while the null mutant analyzed by Koi et al. exhibits defects in egg cell differentiation, Rövekamp et al. reported morphologically inconspicuous egg cells, which, however, continued into an embryo-like program without fertilization. The different phenotypes observed likely relate to the different alleles or to different growth conditions and suggest roles for RKD in egg cell differentiation and the repression of a parthenogenetic reproductive pathway. Central cell specification has been shown to rely on the MADS-domain transcription factor AGL80 (Portereiko et al., 2006) and AGL61 (Steffen et al., 2008), which interacts with AGL80. Loss-of-function mutants exhibit developmental defects in the central cell accompanied by a failure to initiate the expression of two out of three tested central cell reporters (Fig. 1J). In addition, reactive oxygen species, which are particularly abundant in the central cell of the FG, play an important role for central cell specification (Fig. 1J). In oiwa mutants defective for the mitochondrial Mn-superoxide dismutase 1, synergids and egg cells exhibit increased reactive oxygen species levels and concomitant activation of central cell characteristic genes (Martin et al., 2013). Orchestrating the differentiation and development of FG cells requires cell-cell communication. It is this interaction between individual cells in the embryo sac together with a remarkable developmental plasticity that has complicated the interpretation of mutant defects. Mutations in the cysteinyl-tRNA synthetase SYCO ARATH result in an increased lifespan of antipodal cells and affect mitochondrial integrity and polar nuclei fusion in the central cell (Kägi et al., 2010). Interestingly, SYCO is only expressed in the central cell, and targeted disruption of central cell mitochondria making use of a dominant version of the ATP/ADP translocator AAC2 (aac2A199D) represses antipodal cell death. Wu et al. have expanded the picture by showing that antipodal lifespan is similarly regulated by the egg cell, as egg-cell-expressed aac2A199D also interferes with antipodal PCD (Wu et al., 2012). These studies thus uncover a noncell autonomous role of mitochondria in the regulation of cell death and suggest that antipodal lifespan and metabolic activity of female gametes are coupled (Fig. 1K). Wu et al. have in addition shown that mitochondrial function is also communicated between egg and central cell: Mutations in GAMETE CELL DEFECTIVE1 (GCD1) affect gamete maturation and antipodal cell PCD. Notably, activation of GCD1 in either egg or central cell advances the maturation in the respective other gamete, while targeted expression of aac2A199D in the egg cell interferes with polar nuclei fusion in the central cell (Wu et al., 2012; Fig. 1K). Maintaining the decision between gametic cells and accessory cells requires the activity of the pre-mRNA splicing factors LACHESIS (LIS), CLOTHO (CLO), and ATROPOS (ATO). In the respective mutants, synergids and central cells exhibit morphological and molecular features characteristic to egg cells, while antipodal cells express central cell attributes (Groß-Hardt et al., 2007; Moll et al., 2008). In contrast to other mutants the extra and cells not get The phenotype of and reveals a gametic of all FG cells and an important role for pre-mRNA splicing in gametic cell identity. Notably, the phenotype and the is not in embryo activity has been shown to be after cellularization in gametic cells (Groß-Hardt et al., and of transcript the mutants et al., 2012). Together, these results have been in light of a according to which female gametes gamete identity in adjacent accessory cells (Groß-Hardt et al., 2007; Fig. 1I). The of all FG cells to be to FG cell identity is remarkable and has complicated the interpretation of mutant In the of accessory cells, this plasticity has been to as a (Groß-Hardt et al., which can of the system (Lituiev et al., The important role of the egg cell in cell specification is by the of the gene in is activated at the micropylar pole of the embryo sac, and of the genes interferes with expression of and a reprogramming of antipodals into central cell et al., 2012; Fig. 1I). The role of the egg cell for the specification of FG cells has also been in single cell studies. is a very to the role of individual cell types and for the important role of the center in the cell of adjacent cells et al., and the key role of synergids for pollen tube attraction et al., 2001). et al. cell to the FG of a functional egg cell. This as as a genetic egg cell using the et al., activated an egg-cell-like profile in adjacent synergid cells. The study was by Susaki et al. that the pollen tube function of synergids is egg cell associated with In addition, the observed a shift in synergid polarity with respect to the position of the nucleus. This is particularly as there appears to be a between nuclear position and cell cell and synergids a However, differ with respect to nuclear While synergid nuclei the cell the egg cell nucleus the portion of the cell. In all mutants thus exhibiting a reprogramming of synergids into egg cells (Groß-Hardt et al., 2007; et al., 2007; Moll et al., et al., 2011; et al., the identity is with a shift in nuclear Nuclear position can thus as an important and the that egg cell an identity shift in The mechanisms communication between individual FG cells. How is the of important this was the establishment of a to FG cells. This was for in to study as one readout for cell-cell communication et al., et al. the central cell with of molecular and their that the FG is the sporophytic tissue. of the into the egg cell on their size and the developmental stage of the embryo an initial the of between the central cell and egg cell after while was between central and synergid cell. The knowledge of communication between the female gametophytic cells, in addition, the to into cells that are not for as the egg and synergid cell. By the central cell of with et al. to within synergids, which results in a defect in pollen tube the has been to the model plant Arabidopsis, and of the into the central cell of FGs triggered the of the synergid nucleus et al., 2013). its the in is the of while cell to only, the large central cell could be of and for the and it be to establish that addition or of nuclei. techniques are established in and a for both and applied The few-celled FG is in the ovule by maternal which the of FG that has been to this is the use of mutants that fail to form a FG (Pagnussat et al., et al., 2007; et al., 2007; et al., 2007). While transcript of wild-type mutant has a plethora of genes with important roles in FG development and function, additional techniques to be established to get on individual cell FG of the was the establishment of a that of individual FG cells and 1996). The of the cellular by which, however, has the to affect the of the cells et al., and the use of this it was to several important et al., et al., 2009; Sprunck et al., 2012). This is particularly for Arabidopsis due to the size of the cells. In light of these has become a and 2015). This use of a to out single cell from of plant et al., 2015). While the is to it to the of the Arabidopsis gametophyte as as individual cell for the egg, and central cell et al., In with and the several genes to be in the FG and FG expression of genes implicated in the development of the embryo sac et al., 2010). are as only detect genes, which are on the This is by In fact, of Arabidopsis central cells, and and that the of expressed genes could be et al., 2012). The use of has further important it is not restricted to a of genes, for the of including pre-mRNA splicing it the possibility to study plants that as example is the which is related to Arabidopsis but exhibits a distinct of The of these two species might the for genetic factors associated with In fact, a between and Arabidopsis egg, central cell, and synergid cell several expressed genes et al., 2010; et al., 2014; et al., 2016). The of is the and and the associated This can be either with the of nuclei in cell types and 2011; et al., or with cell or nuclei et al., 2015). Both the of transcript in all cell types for which are available. In addition, these techniques might be particularly attractive when it to a of lifespan of individual cells. The of for was to between and of cells within a However, recent adaptation could that cells could be from a a of et al., 2012). to it was unclear be to individual cells of the FG. To molecular of individual female gametophytic cells by light and provide a and marker (Kägi and can the of gene expression while reporters for of a profile of the processes In the last decades, reporters for of the et al., and cell-type-specific reporter for all FG types have been et al., et al., et al., et al., et al., et al., 2007; et al., 2007; et al., 2009; et al., 2010; et al., 2011). In addition, several marker genes have been established molecular reporter for egg, and central cell or all four FG cells on one et al., 2013; et al., 2013). Notably, these molecular only of cellular and in most we regarding the of the reporter genes on a cellular lifespan To monitor cells in we to the by the of FG cells in stages. It has become that cell differentiation and patterning are regulated by a complex of gametophytic and sporophytic factors and are by cell-cell communication However, it is not to an that the results or the formation of the distinct cell is that important key are regulated and which will be in the due to the of and the of new model In addition, the remarkable of all FG cells to be to other FG cell identity has complicated the interpretation of mutant In the of accessory cells, this plasticity has been to as a (Groß-Hardt et al., which can of the system (Lituiev et al., 2013). of reproductive will a that requires concepts and techniques to in this important are to the and the for the programmed cell death female gametophyte