From gametes to zygote: Mechanistic advances and emerging possibilities in plant reproduction

Introduction

Research in angiosperm sexual reproduction is among the most classical botanical studies with a rich knowledge base. The introduction of molecular tools since the late 1970s, especially the development of genomic, transcriptomics, and gene identification tools, and increasingly sensitive biochemical and cell biological methods have culminated in explosive advances in the molecular and mechanistic understanding of reproduction in flowering plants. With outcrossing being crucial for hybrid vigor, understanding the molecular mechanisms of “self-incompatibility” (SI), which prevents inbreeding, led the study of plant reproduction into the molecular era. These efforts had achieved significant milestones (see Broz and Bedinger 2021; Díaz et al. 2021; Goring et al. 2023; Nasrallah 2023) before the molecular nuts and bolts that enable fertilization emerged. Here, we use SI studies from the 1980s and 1990s as the watershed to begin our discussions of mechanistic advances ranging from the developmental standpoints of gamete production to the cell biological perspectives of pollen-pistil interactions that enable fertilization and post-fertilization zygotic development (see Johnson et al. 2019; Hater et al. 2020; Hafidh and Honys 2021; Zhao et al. 2022a) (Fig. 1). We shall also discuss strategies that maximize reproductive success and how they might be deployed to impact agriculture. While many of our examples come from the model plant Arabidopsis, studies in other model or crop species provide a richness that we also tapped into for our discussion. Box 1 provides a glossary for terms; Box 2 looks toward where future efforts would most benefit from current knowledge.

Figure 1.

A schematic overview of the reproduction process in flowering plants. The eudicot Arabidopsis is used as a model for illustration. A) Male gametogenesis. Two-step meiosis in pollen mother cells produce 4 haploid microspores as a tetrad. Each microspore then undergoes an asymmetrical division to give rise to a large vegetative cell and a small generative cell. The generative cell then undergoes another mitotic division, leading to the production of 2 sperm cells giving rise to tricellular pollen. In some species with bicellular pollen, for example, tobacco, the second mitotic division of the generative cell occurs in the pollen tube. B) Female gametogenesis. Meiosis of the megaspore mother cell producing the functional megaspore. The megaspore undergoes 3 rounds of nuclear division, resulting in a syncytial female gametophyte (embryo sac) composed of 8 nuclei. Cytokinesis follows, resulting in a mature female gametophyte comprising the egg cell, central cell, antipodal cells, and synergid cells, each with specialized functions. C) Pollen germination and tube growth. After landing on a receptive stigma, pollen grains undergo rehydration and activation to initiate germination, leading to the formation of a pollen tube that develops from the vegetative cell. Hydration is almost instantaneous and temporally well-resolved from the later event of pollen tube penetration into the stigmatic surface. D) The pollen tube growth path in the pistil. Sperm cells are immotile and rely on the growth of the pollen tube in the stylar and transmitting tissues to reach the ovules where the target female gametophyte is located. E) Pollen tube guidance to ovules. In the transmitting tract, the pollen tubes grow towards the ovules by navigating through a series of female reproductive tissues. Attractants secreted by the ovules regulate this process. F) Pollen tube reception and double fertilization. Once the pollen tube reaches its destination, it releases its cargo—2 sperm cells—in a process referred to as pollen tube reception. The egg cell is fertilized by 1 sperm cell to form the zygote, and the central cell is fertilized by another sperm cell to initiate the development of the endosperm. Box 1 provides a glossary for nomenclature and terms. Supplementary Table S1 provides a list of components important for the gametogenesis processes, including some not discussed here.

Box 1. A glossary of nomenclature and terms.

Microsporogenesis and Microgametogenesis

• Diploid pollen mother cell (PMC)

• Male germline (MG) cells

  • a generative cell (GC) and its mitotic twin sperm cells (SCs)

• Male gametophyte (pollen)

  • VC (the pollen tube cell)

  • GC and SCs

Megasporogenesis and Megagametogenesis

• Megaspore mother cell (MMC)

• Functional megaspore (FM)

• Embryo sac (ES), female gametophyte (FG);

The polygonum type is monosporic, 8-nucleate and 7-celled at maturity.

• Synergid cells: twin cells at the entrance to the FG; an arriving pollen tube penetrates one of these

• Egg and central cells: female gametes for double fertilization producing the zygote and endosperm of a future seed

• Antipodal cell: for nutrient provisions to the FG

Pre-zygotic events

• Pistillate structures

  • Stigmatic papilla: pollen receptive surface.

  • Transmitting tract: a specialized tissue with highly secretory cells that produce an extracellular matrix to support pollen tube growth.

  • Septum: a partition or wall-like structure/tissue that divides the locules or compartments.

  • Funicular: a stalk-like structure that connects the ovule to the placenta.

  • Ovules: the target organ for pollen tubes that supports the development of the female gametophyte, the precursors to future seeds after fertilization.

  • Micropyle: a small opening or pore in the integuments of an ovule to allow pollen tube entrance.

• Pollen tube, a single highly polarized cell developed from the pollen grain specialized as a vehicle to transport sperm to the female gametophyte for fertilization.

• Pollen tube guidance, the process that the pollen tube is tightly navigated to the target embryo sac along various female tissues:

  • Funiculus guidance, the process that the pollen tube is signaled to exit the septum, then guided toward an ovule along the funiculus, the stalk that connects the ovule to the septum.

  • Ovular guidance, the process by which the ovule itself plays a role in guiding the pollen tube towards it, typically through chemical attractants or other cues that emanate from the ovule.

• Pollen tube reception, the process where the pollen tube, upon entering the micropyle, establishes contact and interaction with the filiform apparatus, triggering the degeneration of the synergid cell and eventually leading to the pollen tube rupture.

• Receptive synergid, the synergid cell that is penetrated by the arriving pollen tube and becomes degenerative.

• Persistent synergid, the synergid cell that is not penetrated by the arriving pollen tube and plays an important “reserved” role for “fertilization recovery.”

• “Polytubey”: the phenotype of supernumerary pollen tube penetration of an ovule/the female gametophyte

• Fertilization recovery: the ability of different tissues within the pistil to activate and support growth and entry of additional pollen tubes into the ovule to bring new sperm cells to salvage double fertilization when the egg and/or the central cell fail to be fertilized.

Gametic fusion and early zygotic events

• Gamete recognition: the process by which male and female gametes recognize and fuse with each other during fertilization.

• Gametic fusion: gametic fusion comprises 2 sequential processes: plasmogamy, the fusion of the cytoplasm of male and female gametic cells, and karyogamy, the fusion of the nuclei of male and female gametes during fertilization.

• Zygotic genome activation (ZGA): the onset of de novo transcription from the zygotic genome, which occurs shortly after fertilization in flowering plants.

Elaborating strategies

• Polyspermy control: ensures single pollen tube entry per ovule, maximizing ovule occupancy, seed yield and protecting progeny health.

• Fertilization recovery: salvage mechanisms when fertilization fails to ensure production of some progeny ensuring species survival.

• Reproductive barriers: intraspecific barriers facilitate out-crossing, reducing inbred suppression; interspecific barriers prevent hybridizations that are likely to result in aborted reproduction, or differentiate between self and pollen from closely related species to favor conspecific fertilization and species preservation.

• Apomixis: generating seeds that contain embryos genetically identical with the mother plants.

Box 2. A look to the future.

• Developmentally, how the multiple factors that control cell fate determination during male and female gametogenesis coordinate to regulate the differentiation of VC, GC, and the female gametophyte remains to be elucidated.

• For the prezygotic phase:

  • The emerging mechanistic understanding for stigma interaction with pollen, pollen tube attraction, reception, prevention of polyspermy and fertilization suggest diverse strategies. A cohesive understanding integrating available knowledge is needed to provide solid foundations for future efforts.

  • The in vivo configurations of various FER-related signaling complexes need to be determined. Is there a predominant configuration, or are various possible configurations assembled dynamically in response to different conditions?

  • With FER and related receptor kinases at the core of multiple strategies, the precise contributions from other positive and negative regulatory components need to be defined.

  • Molecular and biochemical understanding of pre-ovular and funiculus guidance of pollen tubes remains limited. This knowledge will provide useful context to formulate working hypotheses to explore how polyspermy blocks and fertilization recovery are orchestrated.

  • The ability to monitor the pollen tube growth and targeting process inside the pistil (e.g. Mizuta et al. 2023) and with intracellular definitions will illuminate how the pollen tube growth machinery responds to pistillate cues in vivo.

  • Envisioning an all-encompassing mechanism on how fertilization recovery is achieved remains challenging; multiple mechanisms likely exist. Investigating the temporal and spatial correlation between the degradation of the receptive synergid cell and the 2 fertilization events should provide more clues for identifying the initiating signals of persistent synergid cell degradation. Understanding how the central cell senses and reacts to the degeneration of both synergid cells activates the transcription and release of its pool of “backup” pollen tube attractants should provide insights on the hypothesis that accessory cells in the FG had evolved to render more efficient fertilization.

• For the fertilization phase, whether and how the egg cell impacts its own precedence over the central cell for being fertilized requires further investigations. Achieving this will require identifying factors in both the egg cell and central cell that are responsible for gamete recognition, attachment and fusion.

• For post-fertilization, ZGA is an emerging area with the possibility to probe pioneer factors responsible for the first wave of zygotic genome transcription and the initiation of zygotic embryogenesis. Illustration of epigenetic characteristics including DNA methylation, histone methylation, and other modifications in both gametes and early embryos will promote the understanding of mechanisms for ZGA.

• Reproductive barriers maintained at different male-female interaction steps involved distinct strategies, each could also be diverse in mechanisms. Currently, compared with research on postzygotic hybridization barriers (in rice, for example), studies in prezygotic hybridization barriers, especially those with clear molecular mechanisms, are very limited. Proof-of-principle experiments to overcome interspecific barriers at the stigma appear promising.

• Recent advances in generating synthetic apomixis are promising. Further investigations are necessary to devise a synthetic apomixis strategy for hybrid in different crops, ensuring elevated seed-setting rates and improved cloning seed induction rates. Advancements in haploid induction and MiMe processes offer promising avenues for implementing synthetic apomixis in staple crops beyond rice.

Male and female gametogenesis

Flowering plants produce germ cells via the sequential patterning events of sporogenesis and gametogenesis (Fig. 1, A and B) (Ma 2005; Yang et al. 2010; Hater et al. 2020; Hafidh and Honys 2021; Cai et al. 2022). Here, we focus on cell fate determination of male germline (MG) cells, the generative cell (GC) and its mitotic product sperm cells (SCs) in male gametogenesis, and of the female gametes, the egg and the central cell, in female gametogenesis.

Differentiation of MG cells and vegetative cell

Milestone discoveries on MG cell fate determination can be considered 2-fold. One is the importance of the asymmetric division of a microspore to produce the larger vegetative cell (VC) of the male gametophyte (pollen) and the smaller GC (Fig. 1) (Eady et al. 1995; Twell et al. 2002). The other is the role of the R2R3 transcription factor DUO POLLEN 1 (DUO1) in controlling sperm formation from the GC (Rotman et al. 2005; Brownfield et al. 2009). The asymmetric division of the microspore follows the microtubule-dependent migration of its nucleus to the future germ cell pole. Failing to do so results in symmetric mitosis and both daughter cells adopting the VC fate (Eady et al. 1995; Chen and McCormick 1996; Park et al. 1998; Park et al. 2004; Zeng et al. 2009; Oh et al. 2010; Oh et al. 2016; Lee et al. 2017). Loss of the functionally redundant helix-loop-helix transcription factors BONOBO1 and BONOBO2 resulted in the smaller daughter cell from microspore asymmetric division failing to adopt the GC fate, implying a broader genetic network underlies GM cell fate determination (Yamaoka et al. 2018).

Cell cycle progression is critical for GC mitosis to produce 2 SCs. Mutations in several cell cycle regulators blocked GC division, resulting in bicellular pollen harboring only a single MG cell (Iwakawa et al. 2006; Nowack et al. 2006; Kim et al. 2008; Liu et al. 2008b; Gusti et al. 2009). The A-type cyclin-dependent kinase CDKA;1 is crucial for the GC progression through S-phase (Iwakawa et al. 2006; Nowack et al. 2006). Defective chromatin assembly causes cell cycle arrest, producing pollen with a single MG cell (Chen et al. 2008). DUO1 is expressed specifically in the MG and controls the expression of the G2/M regulator CYCLIN B1;1 (Rotman et al. 2005; Brownfield et al. 2009), and single MG cell phenotypes are also detected in a group of duo pollen (Durbarry et al. 2005). The single MG cell in pollen with mutations that directly target cell cycle processes differentiates sperm cell character, including the full capacity for fertilization (Chen et al. 2008; Gusti et al. 2009; Aw et al. 2010). Contrarily, the single MG cell in duo1 lacks fertilization capacity (Rotman et al. 2005), consistent with DUO1 controlling the expression of several MG-specific genes, including the gamete fusogen GCS1/HAP2 (GENERATIVE CELL SPECIFIC 1/HAPLESS2) on the sperm cell surface required for fertilization (Mori et al. 2006; Borg et al. 2011).

Epigenetic reprogramming, especially the erasure and remodeling of histone methylation, also plays an important role in the differentiation of the VC and MG cells. The chromatin is diffuse in the VC but highly condensed in the GC/SCs. Some histone variants showed VC- or SC-specific expression (Borg and Berger 2015; Jiang et al. 2020). The VC, GC, and SC nuclei also harbor differential levels of various histone modifications, and the histone3 mark H3K27me3 is essential for the VC fate as its erasure resulted in the VC fate shifting toward the MG cell fate (Houben et al. 2011; Borg et al. 2020; Huang and Sun 2022). The replacement of H3.1 by H3.10, silencing of H3K27me3 writers, and activation of H3K27me3 erasers/demethylases together might underlie how the H3K27me3 marks are globally lost from Arabidopsis SCs (Ingouff et al. 2010; Borg et al. 2020).

The mature pollen grain

The complexity of the surface of the mature pollen grain and the diverse patterns of apertures on pollen grains have long intrigued biologists (Furness and Rudall 2004; Quilichini et al. 2014). The development of increasingly sensitive analytical tools has enabled considerable advances in characterizing the protective sporollenin that sculpts the outer wall of pollen grains (Li et al. 2019). Genetics studies, especially of transcriptional cascades controlled by the transcription factor MS188, have uncovered many key regulatory and biosynthetic genes for the pollen cell wall (see Ariizumi and Toriyama 2011; Wang et al. 2018; Dehors et al. 2019; Ma et al. 2021; Yao et al. 2022; Xue et al. 2023).

Pollen from most species germinates extruding the pollen tube through an aperture, although some species can extrude a tube directly through the pollen wall (Edlund et al. 2016). Arabidopsis pollen has 3 equidistantly spaced apertures. Loss of INAPERTURATE POLLEN 1 (INP1), encoding a membrane protein, resulted in mutant pollen with no aperture, though they remained fertile (Dobritsa and Coerper 2012). INP1 marks pollen cell membrane domains early in pollen development, and they are specified to become lightly covered by the pollen wall and differentiate into 3 spatially restricted pollen apertures around the pollen equator (Dobritsa and Coerper 2012; Lee et al. 2018; Zhou and Dobritsa 2019; Lee et al. 2021; Mazuecos-Aguilera et al. 2021). Another family of proteins, MACARONs, function upstream of INP1 and quantitatively determine the number and patterning of pollen apertures (Zhou et al. 2021a).

INP1 is conserved in cereal crops, such as maize and rice, which typically have a single pollen aperture (Christensen et al. 1997). Although the maize INP1 could functionally substitute for the Arabidopsis INP1, the maize inaperturate pollen is sterile (Lee et al. 2018). In rice, interaction between INP1 and a lectin receptor-like kinase is crucial for aperture formation (Zhang et al. 2020b).

Differentiation of synergid cell, egg cell, and central cell

In polygonum-type embryo sac (ES), development is staged between FG (female gametophyte)1 to FG7, starting from a functional megaspore (FM) (Christensen et al. 1997), which undergoes nuclear division, nuclear migration, and cellularization, producing a mature FG with 7 cells (Fig. 1). Each member cell type in the ES has its function to ensure double fertilization (Box 1). The 2 synergid cells attract and receive pollen tubes (Higashiyama 2002). The 2 sperm cells, 1 fuses with the egg and the other with the central cell for a double fertilization to form the zygote and the endosperm of a future seed. Three antipodal cells may function in nutrient transfer, especially in the ES of species with as many as 100 cells before fertilization (Engell 1994). Consistent with their functional differentiation, member cells in the ES have different transcriptomes (Kumlehn et al. 2001; Le et al. 2005; Sprunck et al. 2005; Okuda et al. 2009; Wuest et al. 2010; Ohnishi et al. 2011; Zhao et al. 2011; Anderson et al. 2013; Yu et al. 2016; Song et al. 2020; Susaki et al. 2021). These enabled detailed functional studies and facilitated the development of marker genes for individual cell types, greatly aiding live-imaging of processes pre- and post-pollen tube arrival (Kasahara et al. 2005; Portereiko et al. 2006; Steffen et al. 2007; Okuda et al. 2009; Sprunck et al. 2012; Chen et al. 2017; Tedeschi et al. 2017; Yu et al. 2021). Cell type–specific transcription factors play crucial roles in determining the synergid, egg, or central cell fate. Among these, the transcription factor MYB98 is crucial for the synergid cell fate determination and maintenance, RKD (RWP-RK domain) proteins control the egg cell fate, AGAMOUS-like 80 promotes central cell fate and inhibits MYB98 (Portereiko et al. 2006; Bemer et al. 2008; Steffen et al. 2008; Zhang et al. 2020c), while MYB119 and BYB64 redundantly restrict the central cell fate (Rabiger and Drews 2013).

Interestingly, female gametes also regulate fate specialization of adjacent cells. For example, a mutation in the egg and central cell–expressed LACHESIS (LIS), a homolog of the yeast mRNA splicing factor PRP4, induced the mis-expression of egg-specific genes in the mutant synergid and central cells, and the antipodal cells in the mutant resembled the central cell (Gross-Hardt et al. 2007). A similar phenotype as lis-1 was also observed in other splicing factor mutants (Moll et al. 2008). LIS transcript levels in the egg cell specifically affected the development of all female gametophytic cells (Völz et al. 2012), further supporting an important role for RNA maturation. Cell ablation studies also supported the idea that the egg cell affects synergid cell specification (Lawit et al. 2013; Susaki et al. 2015).

Nuclear positions before cellularization in the ES are important for cell fate determination. In Arabidopsis, nuclei positions at FG5 already mark the post-cellularization destiny of these nuclei (Christensen et al. 1997; Susaki et al. 2021). When the typical nuclear location in the cell destined to be the egg in FG5 was replaced by 2 nuclei, an extra egg cell formed at the expense of a synergid cell (Gross-Hardt et al. 2007; Pagnussat et al. 2007; Kirioukhova et al. 2011). Disrupting the actin cytoskeleton before cellularization impaired nuclear migration and resulted in the formation of 2 functional egg cells. Both could be fertilized, producing double embryos, thus further supporting that the final positions of the nuclei, rather than nuclear lineages, play critical roles in the specification of the egg cell (Sun et al. 2021). Extra rounds of free nuclear divisions in the maize mutant indeterminate gametophyte 1 ES also resulted in extra synergid, egg, and central cells according to nuclear position in ES (Guo et al. 2004). Auxin and cytokinin are also important for ES cell fate determination (Pagnussat et al. 2009; Yuan et al. 2016; Sun et al. 2021) and might rely on uneven distribution of these hormones along the chalazal to antipodal axis of the ES (Skinner and Sundaresan 2018). Disrupting normal auxin synthesis or distribution pattern impaired egg cell specialization (Pagnussat et al. 2009; Panoli et al. 2015; Sun et al. 2021). Cytokinin signaling, from perception to phosphorylation cascades and downstream response regulators, is also critical for FG development, in particular, central cell specification (Kong et al. 2015; Yuan et al. 2016; Zhu et al. 2022).

Pre-zygotic events

Impressive advances have been made in elucidating key molecular players that enable prezygotic pollen-pistil interactions and render fertilization possible (Fig. 1, C–E) (Johnson et al. 2019). An elaborate network for structural and signaling components for the elongating pollen tube has been established for the polarized cell growth process in vitro (Steer and Steer 1989; Cheung and Wu 2008; Hepler and Winship 2015; Michard et al. 2017; Cameron and Geitmann 2018). Here, we discuss several related signaling modules important for pollen-pistil cooperation and how their studies have advanced the mechanistic understanding of pollen tube growth and ovule targeting in the pistil (Figs. 2 and 3).

Figure 2.

The prezygotic phase of reproduction. A) Aniline blue–stained pollinated pistil showing the pollen tube growth path in an Arabidopsis pistil. It highlights pollen tube targeting of ovules and the typical 1 pollen tube: 1 ovule pattern of pollen (reproduced from Zhong et al. 2022a). B) A prototypical RALF-FER-LRE/LLG tripartite signaling module. C) A list of the tripartite signaling modules playing key roles in the prezygotic phase of the reproductive process. D) FER controls pollen germination and pollen tube penetration of the pistil. The schematic shows the components involved in regulating the early events in pollen hydration and the later events of pollen tube penetration of the stigmatic papilla. E) An abbreviated schematic representation of an elongating pollen tube. Readers are referred to many overall reviews referred to in the text. Listed here are some well-established components that regulate diverse conditions in the pollen tube. Components that regulate Ca²⁺ dynamics include CPK34/17 and CPK2/20 (Myers et al. 2009; Gutermuth et al. 2013), CPK11/24 (Zhao et al. 2013), CPK32 (Zhou et al. 2014), CBL1/9 (Mähs et al. 2013), CBL2/3 together with CIPK12 (Steinhorst et al. 2015), Ca²⁺ exchanger (Demidchik et al. 2018), and P-type IIA Ca²⁺-ATPases (García Bossi et al. 2020). Regulators for the dynamics of other ions, such as K⁺ (Mouline et al. 2002), are also important. Multiple regulators of cytoskeletons are also required (Chen et al. 2003; Lovy-Wheeler et al. 2006; Cheung and Wu 2008; Ye et al. 2009; Cheung et al. 2010; Zhu et al. 2017; Lan et al. 2018; Liu et al. 2018); microtubules are not displayed. Golgi-resident proteins involved in secretion of membrane proteins and cell wall materials are also essential for pollen tube growth (Chen et al. 2015; Tan et al. 2016; Jia et al. 2018). We refer readers to Supplementary Tables S2 to S4 for an extended list of components contributing to diverse functional needs for the pollen tube growth process in vitro and in the pistil. F) FER controls sperm release and suppresses polyspermy. Images show fer-4 ovule phenotypes of pollen tube overgrowth and supernumerary pollen tube entrance. (Reproduced from Duan et al. 2014).

Figure 3.

A schematic summary of key components for pollen tube growth in the pistil, pollen tube guidance and post-targeting events. A) Key factors involved in pre-ovular guidance (blue), pollen tube penetration at the septum (red), and funicular guidance (purple); the collection is from Arabidopsis, lily (chemocyanin, SCAs, LTPs), and tobacco (TTS). B) Key factors involved in micropylar guidance. The collection is mostly from Arabidopsis; AMOR and TfLUREs are from T. fournieri, and ZmEA1 from Z. mays. C) Key factors for polyspermy control, fertilization, and fertilization recovery. (i) A view in the pistil illustrating ovular conditions triggered upon pollen tube arrival (upper ovule and ii) and key molecules that induce the ovular polyspermy block, and an ovule (lower ovule and iii) showing key molecules for gametic fusion (HAP2/GCS1, GEX2 and DMP8/9) and for fertilization recovery (ECS1/2 and SALVAGE1/2). OW, ovule wall; SE, septum; TT, transmitting tract; PTs, pollen tubes; S, style.EC, egg cell; CC, central cell; SC, synergid cell; DSy, degenerated synergid cell; Sp, sperm. Red thunder bolts along the SE, triggers for SE polyspermy block. Not all components here are discussed in the text. Supplementary Tables S3 to S5 provide a more complete list.

The FERONIA (FER) receptor kinase and related signaling modules

The establishment of the FER-LORELEI (LRE) signaling module (Fig. 2, B and C) provided an invaluable entry point into the mechanistic dissections of the pre-zygotic processes (Cheung et al. 2022). FER-LRE is a key regulator of pollen tube reception at the pollen tube-FG interface, where the pollen tube would penetrate the FG, burst and release sperm for fertilization (Fig. 1E; Box 1). FER is a malectin domain–containing receptor kinase (Escobar-Restrepo et al. 2007; Franck et al. 2018a, 2018b; Yang et al. 2021), and LRE is a glycosylphosphatidylinositol-anchored protein (GPI-AP) (Capron et al. 2008; Tsukamoto et al. 2010). As a coreceptor pair (Li et al. 2015a; Xiao et al. 2019), FER and LRE colocalize in the synergid cell membrane that densely packed the filiform apparatus (Box 1), the synergid cell wall gating the entrance to the female gametophyte. FER is 1 of 17 related receptor kinases in Arabidopsis. It is expressed almost ubiquitously except in pollen and plays profound roles for growth and survival (Li et al. 2016; Malivert and Hamant 2023; Cheung 2024; Liu et al. 2024). LRE is expressed almost exclusively in the FG; LRE-like GPI-AP1 (LLG1) is expressed in sporophytic tissues throughout development, including in the stigma, while LLG2 and LLG3 are pollen-specific (Noble et al. 2022).

Studies of FER during early plant development identified multiple basic parameters of FER-LRE/LLG signaling and have been reviewed extensively (Liao et al. 2017; Doblas et al. 2018; Franck et al. 2018a, 2018b; Cheung et al. 2022; Ortiz-Morea et al. 2022). FER-LRE/LLGs are coreceptors (Li et al. 2015a), and RALF (RAPID ALKALINIZATION FACTOR) peptide regulators are their ligands (Haruta et al. 2014; Blackburn et al. 2020; Abarca et al. 2021). FER binds the major cell wall polysaccharide pectin (Feng et al. 2018; Lin et al. 2022). RAC/ROP GTPases and reactive oxygen species (ROS) are mediators of FER-LLG/LRE signaling (Duan et al. 2010; Li et al. 2015a; Song et al. 2021). Here, we focus on the modules that regulate prezygotic pollen-pistil interactions (Fig. 2, B and C).

Pollen-stigma interaction

Pollen germination and tube penetration into the stigmatic papilla

Compatible pollen germination and tube penetration into the pistil involve an early phase of pollen hydration and germination, that is, extruding a pollen tube, and the later event of pollen tube penetration into the stigmatic papilla (Fig. 1C). In Arabidopsis, hydration occurs within minutes after pollination (MAP) and is completed around 10 MAP (Rozier et al. 2020; Liu et al. 2021; Huang et al. 2023; Lan et al. 2023a). Pollen tube penetration of the stigmatic papilla is notable around 30 MAP; by 3 hours pollen tubes could clearly be traced along the stigmatic cell surface, and, typically, they have grown into the transmitting tissue (Huang et al. 2023; Lan et al. 2023a) (Figs. 1C and 2D). A RALF-FER-LLG1 module mediating RAC/ROP signaled-ROS production in the Arabidopsis stigma provides a redox-controlled gate, which is relaxed by compatible pollen, for example, during self-pollination (Liu et al. 2021). Deposition of wild-type pollen suppresses stigmatic ROS, mitigating the gate to allow compatible pollen hydration and germination, followed by tube growth and penetration into the stigmatic papilla. Loss of each of the components of the signaling module, FER, LLG1, RALF33, RAC/ROP or RBOHs (which produce ROS), resulted in the reduction of stigmatic ROS. This accelerated pollen hydration and germination, hastening pollen tube growth into the stigmatic papilla. FER-related ANJEA, which interacted with FER in vitro, and RALF23 also participated in maintaining the stigmatic gate, suggesting multiple configurations in planta are likely for the stigmatic module (Fig. 2, C and D). In search of the pollen trigger that relaxes the FER-anchored gate, Liu et al. (2021) discovered several PCP-B class pollen coat proteins (PCPs) as signals that disengage the RALF23/33-FER/ANJEA-LLG1-signaled, RAC/ROP-mediated, and ROS-controlled stigmatic gate for pollen hydration and germination.

Lan et al. 2023a proposed a more elaborate FER-dependent stigmatic gate. It engages not only FER and ANJEA but also FER-related receptor kinases HERCULES1 and CURVY1, additional RALFs (RALF1,22 along with RALF23,33; referred to as sRALFs), and Leucine-Rich Extensin (LRXs), cell wall proteins with an extensin domain (Bedinger 2018). FER functioned adequately on its own, whereas ANJEA, HERCULES1, and CURVY1 functioned redundantly to maintain the stigmatic gate. The Leucine-rich-repeat (LRR) domains in LRXs bind FER, RALFs, and other membrane proteins and mediate cell wall integrity (Fabrice et al. 2018; Dünser et al. 2019; Herger et al. 2020; Moussu et al. 2020). Using multiplexed CRISPR/Cas mutagenesis, Lan et al. (2023a) identified multiple pollen-expressed RALFs (RALF10,11,12,13,25,26,30; referred to as pRALFs) as triggers to disengage the sRALF-FER/ANJEA/HERCULES1/CURVY1-dependent and LRX-controlled gate.

Molecular and biochemical mechanisms of FER-dependent stigmatic gating

The stigmatic gate for pollen hydration and germination and subsequent pollen tube penetration into the stigmatic papilla are separable events, one unlocked by PCP-Bs (Liu et al. 2021; Goring et al. 2023; Huang et al. 2023; Nasrallah 2023) and the other by pRALFs (Lan et al. 2023a) (Fig. 2, C and D). PCP-Bs are low-molecular-weight, cysteine-rich PCPs that decorate the Brassica pollen surface, acting at pollen hydration to facilitate germination (Doughty et al. 1998; Wang et al. 2017a; Wang et al. 2023a). Loss of PCP-B peptides, such as PCP-Bγ, induced delays in stigmatic ROS decline, pollen hydration, and germination on wild-type stigmas (Liu et al. 2021). Application of PCP-Bγ was adequate to trigger rapid ROS decline in unpollinated stigmas, suggesting that PCP-Bγ downregulates the stigmatic FER to ROS pathway to facilitate pollen hydration, followed by germination and tube growth into the pistil. Loss of FER compromises cell wall integrity (Feng et al. 2018), which most likely also contributed to fer mutant stigmas being more readily penetrable by pollen tubes (Liu et al. 2021; Zhang et al. 2021; Huang et al. 2023; Lan et al. 2023a).

For the pRALFs-unlocked gate (Lan et al. 2023a), light microscopy, scanning, and transmission electron microscopy documented that pollen from the multiplexed loss-of-function ralf septuplet mutant (referred to as ralf sept) behaved just like wild-type pollen during hydration and germination. The ralf sept pollen tubes failed to penetrate the wild-type stigmatic papilla cells (Lan et al. 2023a), but they efficiently grew into mutant stigmas lacking FER; or combinations of ANJEA, HERCULES1, and CURVY1; or all 4 of the sRALFs (referred to as ralf quad). Mutant stigmas lacking LRX3,4,5 also presented no barrier to ralf sept pollen tubes (Lan et al. 2023a).

However, unlike PCP-B–mediated early events of pollen hydration and germination being dependent on the relaxation of FER-controlled, RBOH produced stigmatic ROS (Liu et al. 2021; Huang et al. 2023), ralf sept mutant tubes remained unable to penetrate the ROS-deficient rboh stigmas (Lan et al. 2023a). However, LRX3,4,5, which interacted with both pRALFs and sRALFs in vitro, are essential for the FER-dependent gate to prevent ralf sept mutant tube penetration. Application of RALF33 restored FER to ROS production in the ralf33 mutant stigma (Liu et al. 2021); it also reimplemented the ralf quad-weakened stigmatic gate, restoring it to be nonpenetrable by ralf sept tubes (Lan et al. 2023a). However, application of RALF33 had no impact on lrx3,4,5 stigmas, which remained penetrable by ralf sept tubes. The most parsimonious interpretation would be that the FER-dependent stigmatic gate controls 2 separable events. First, it gates the early events of pollen hydration and germination via RBOH-produced ROS, and PCP-B unlocks the gate by triggering the decline of stigmatic ROS (Liu et al. 2021). Then, the FER-dependent, LRX-controlled gate mediates papilla cell wall quality to guard against the later event of pollen tube penetration of the stigma, which is unlocked by pRALFs (Lan et al. 2023a).

Molecularly, key-lock mechanisms were proposed with the stigmatic FER-anchored modules as locks that protect the pistil from intrusion until triggered by keys from compatible pollen (Liu et al. 2021; Zhang et al. 2021; Lan et al. 2023a). PCP-Bγ competed with RALF33 for binding to the FER extracellular domain in vitro. On the ralf33 mutant stigma, PCP-Bγ efficiently blocked exogenous RALF33 from restoring stigmatic ROS production. Together, these results led to a model whereby upon pollination, PCP-Bs compete with stigma RALF23/33 to disengage the FER to ROS pathway, resulting in pollination-triggered stigmatic ROS decline, relaxation of the stigmatic gate for the early events of pollen hydration, and germination (Fig. 2D) (Liu et al. 2021; Huang et al. 2023).

For the FER-dependent, LRX-controlled stigmatic gate (Lan et al. 2023a), in vitro experiments indicate that pRALF competes with sRALF for binding to the extracellular domain of FER/CURVY1. When both pRALF and sRALF were simultaneously applied to ralf quad mutant stigmas, pRALF reversed the inhibition of ralf sept pollen tube penetration caused by sRALF treatment. These results suggest that pRALFs compete with sRALFs for FER binding to unlock the FER-dependent gate for pollen tube penetration into the stigmatic cell wall. On the other hand, ralf sept pollen tubes could penetrate the stigma of lrx345, and pRALF/sRALF both bind the LRX, indicating that LRX3,4,5 are essential for the stigma gate that guards against pollen tube penetration. Interestingly, pRALF and sRALF did not compete for stigma LRX3,4,5. With RALFs being disorder peptides (Xiao et al. 2019; Liu et al. 2024), it would be interesting to resolve what characteristics between the sRALFs and pRALFs rendered them competitive interactors with FER and what rendered them equal partners in their interactions with stigma LRXs. Also interesting to examine is whether there are 2 classes of pollen RALFs, with, for example, RALF4 and 19 interacting with pollen LRXs on the pollen tube surface to mediate tube wall integrity (Mecchia et al. 2017; Sede et al. 2018; Moussu et al. 2020, 2023), and others such as RALF10,11,12,13,25,26,30 interacting with stigma LRXs inducing weakening of wall integrity (Lan et al. 2023a). Considering the roles of LRXs in cell wall integrity (Fabrice et al. 2018; Herger et al. 2019), the establishment and opening of the LRX3,4,5-controlled stigma gate are likely achieved through dynamic changes in the cell wall.

PCP-Bγ, in fact, engages a dual process to secure a hospital stigmatic environment for pollen germination. Huang et al. (2023) discovered that as a prelude to reducing FER-controlled ROS, compatible pollination triggered a rapid and FER-dependent increase in stigmatic NO that rapidly reached a maximum (around 5 minutes after pollination). This rise mirrored the reduction in stigmatic ROS and progress of pollen hydration. Application of PCP-Bγ triggered similar NO increases in unpollinated Arabidopsis stigmas. The mirroring impacts of pollination and PCP-Bγ on stigmatic NO and ROS—that is, when NO is low, ROS is high and vice versa—strongly support a causal relationship between compatible pollination-induced stigmatic NO rise and ROS decline, facilitating pollen hydration and germination to start a pollen tube on its journey in the pistil.

Pollen tube growth in the pistil: an overview

The polarized pollen tube growth process

After penetrating the stigmatic papilla and underlying tissue, the pollen tube grows into the transmitting tract surrounded by ovules (Fig. 1). The polarized cell growth process involves multifaceted coordination of extracellular, cell surface, and cytoplasmic components and is dependent on regulated ion dynamics, redox status, and turgor (Fig. 2E) (Steer and Steer 1989; Cheung et al. 2008; Cheung and Wu 2008; Tavares et al. 2011; Hepler and Winship 2015; Michard et al. 2017; Cameron and Geitmann 2018; Hoffmann et al. 2020; Hayashi and Palmgren 2021). Apically located RAC/ROP GTPases (Guo and Yang 2020; Li et al. 2023c) and a dynamically regulated actin cytoskeleton (Cheung and Wu 2008; Cheung et al. 2010) are crucial for tip-focused growth. Pollen tube growth is self-organized, requiring only Ca²⁺ and Boron in minimal culture media. Here, we focus on the pollen tube growth process in the pistil, which is facilitated and regulated by pistillate tissues (Fig. 3).

The pistil environment

Pistils are architecturally diverse. Studies in lily, tobacco, and tomato, representing open- and solid-style pistils, produced early insights on pistillate contributions to the pollen tube growth process (Cheung 1995, 1996; Lord and Russell 2002). In tomato, the stigmatic factor LeSTIG supported pollen tube growth by competing with the pollen-expressed LeLAT52 for binding to pollen receptor-like kinases (PRK) LePRK1/2, which activate RAC/ROP GTPases (Muschietti et al. 1998; Tang et al. 2002; Tang et al. 2004; Kaothien et al. 2005; Liu et al. 2020). As the highway for pollen tubes, the transmitting tract is spatially accommodating and, being enriched in polysaccharides, glycolipids, glycoproteins, and adhesive molecules, supportive of the metabolically demanding pollen tube growth process as nutrients and provides cell-cell adhesion (Cheung et al. 1995; Park et al. 2000; Nieuwland et al. 2005; Chae and Lord 2011; Pereira et al. 2016a) (Fig. 3A). The pollen transcriptome changes during growth in the pistil, implying adaptive responses to the pistil environment (Qin et al. 2009). The pistillate tissues prime the pollen tube for responding to directional cues from the ovules. A complex glycan, ethyl-glucuronosyl arabinogalactan (AMOR) from Torenia fournieri ovules, confers pollen tube competency in responding to ovular attractants (Mizukami et al. 2016). Low-molecular-weight, pistil-derived factors are also crucial for pollen germination and tube growth, such as brassinosteroid (Vogler et al. 2014), sulfinated azadecalins (Qin et al. 2011), the amino acid D-serine, which affected pollen tube growth through regulating glutamate receptor-like (GRL) Ca²⁺ channels (Michard et al. 2011; Wudick et al. 2018), and γ-amino butyric acid (GABA) (Palanivelu et al. 2003; Yu et al. 2014). Pistils also express peptides to regulate pollen tube growth, for example, CLE45, protecting them against heat stress through interacting with a pollen-expressed kinase (Endo et al. 2013).

Pistils from Arabidopsis, with their short style and elongated ovary with ovules differentiating from the septum (the epidermis of the transmitting tract), provide excellent specimens to trace pollen tube behavior along its growth pathway (Figs. 1 and 2A). With the vast genetic resource repertoire, the popularized use of semi-in vivo pollen tube growth system (Cheung et al. 1995; Higashiyama et al. 1998; Palanivelu and Preuss 2006) and improvements in histological and live-cell visualization of the pollen tube growth process, studies in Arabidopsis have provided major contributions to our current understanding of the in vivo process discussed below.

Pollen tube growth and exit from the stylar transmitting tissue

The pollen tube extends through cycles of cell wall softening and stiffening at the tip (Hepler et al. 2013). This, combined with turgor exerted from the expanding cell (Hepler and Winship 2015), makes the pollen tube susceptible to rupture. FER-related signaling modules, assembled from pollen counterparts of FER, LRE, and LLG1 and pollen-expressed RALFs (Fig. 2, B and C), are crucial for maintaining pollen tube integrity until it reaches the target FG. The paired ANXUR1 (ANX1)/ANX2 (Boisson-Dernier et al. 2009; Miyazaki et al. 2009) and BUPS1/BUPS2 (BUDDHA'S PAPER SEAL1/2) (Ge et al. 2017) are highly and the most prominently pollen-expressed paralogs of FER. Loss of either pair resulted in male sterility. During pollination, anx1 anx2, bups1 bups2, and bups1 pollen tubes displayed tip defects consistent with precocious rupture and were arrested early in the pistils, precluding arrival at the ovules, thus male sterility. In rice, the BUPS/ANX orthologs also control pollen tube integrity (Liu et al. 2016b). A recent study in maize suggests considerable conservation of a FER-like receptor kinase-anchored cell surface module in control pollen tube integrity during growth (Zhou et al. 2024).

ANXs and BUPSs form heteromers, all 4 receptor kinases interacted with LLG2 and LLG3, and the pollen-expressed RALF4 and RALF19 bound to these coreceptors (Ge et al. 2017; Zhu et al. 2018; Feng et al. 2019; Ge et al. 2019). While the precise in vivo configuration remains unclear, these tripartite complexes are located at the pollen tube tip membrane, where they signal the RAC/ROP-controlled ROS production pathway in the pollen tube and the proper deposition of pectin in the tube wall. ANX1/2 promotes exocytosis through the downstream cytoplasmic kinase MARIS and RBOH oxidases RBOHH/J, generating ROS at the tip (Boisson-Dernier et al. 2013; Kaya et al. 2014; Lassig et al. 2014; Boisson-Dernier et al. 2015). Zhou et al. (2021b) suggested that BUPS1 served as a mechano-sensor of the physical barriers from the female tissues to regulate RAC/ROP-controlled ROS production. Furthermore, GPI-AP COBL11 also modulates pollen tube cell wall integrity via interaction with RALF4/19 (Li et al. 2023b), and protein phosphatases ATUNIS1 and ATUNIS2 function in a distinct but converging pathway as ANX1/2 to negatively regulate tip growth (Franck et al. 2018a, 2018b).

The pollen tube cell wall property, particularly its pectin matrix (Cascallares et al. 2020; Du et al. 2022), is important for pollen tube growth. Pectin is secreted in the methylesterified form at the pollen tube tip and demethylated by subapical cell wall–located pectin methylesterases (PMEs) as the tip advances. The de-esterified pectin is crosslinked by Ca²⁺, mediating a stiffened wall starting from the subapical region. Pectin de-esterification is countered by PME inhibitors (PMEIs). This dynamic pattern facilitates elasticity for growth at the apex and rigidity in distal regions to confine the cell geometry and provide integrity (Fig. 2E). Mutations in PME genes, like VANGUARD and PME48, lead to precocious pollen tube rupture (Jiang et al. 2005; Leroux et al. 2015). Therefore, PMEs, PMEIs, and other pectin-modifying enzymes (Du et al. 2022), FER, ANX/BUPS, and possibly other FER-related receptors are functionally intertwined in impacting physical and biological properties at the cell wall/cell membrane interface (Feng et al. 2018; Herger et al. 2019; Liu et al. 2024).

Extensins are cell wall hydroxyproline-rich glycoproteins. They are crosslinked via tyrosine residues and associate tightly with the cell wall matrix, reinforce the wall, and form a scaffold for pectin assembly (Lamport et al. 2011; Mecchia et al. 2017; Marzol et al. 2018; Moussu et al. 2020; Moussu and Ingram 2023). In Arabidopsis, the extensin-like pollen LRX8/9/10/11 interact with RALF4/19, and loss of these cell wall–linked proteins also causes precocious pollen tube rupture (Mecchia et al. 2017; Fabrice et al. 2018; Sede et al. 2018). Other extensin-like proteins, for example, proline-rich extensin-like receptor kinases, regulate pollen tube growth and impact its cell wall polysaccharide composition (Borassi et al. 2016; Borassi et al. 2021).

LRXs comprise an N-terminal LRR domain that interacts with RALF and a C-terminal pectin-associated extensin domain (Herger et al 2019; Moussu et al. 2020; Moussu et al. 2023). RALF4 inhibition of pollen tube growth depends on the LRXs (Mecchia et al. 2017), suggesting that RALF4-LRX binding stiffens the wall as a brake to control pollen tube growth rate. Interestingly, Moussu et al. (2020) noted that tertiary complexes, that is, receptor kinase-LLG/LRX-RALF, had not been observed. Although pH and redox conditions might impact the interactions between the structurally versatile RALF with neighboring molecules, Moussu et al. suggested that the mutually exclusive binding of RALF peptides to the receptor kinases or LRXs might reflect how RALF peptides regulate 2 parallel, mechanistically distinct but convergent pathways to fine tune pollen tube integrity. Whether a similar scenario could be applicable to the FER-dependent, sRALFs-maintained FER-dependent, LRX-controlled, and pRALFs-unlocked stigmatic gate (Lan et al. 2023a) will be interesting to explore.

RALF-pectin phase-separates in vitro and in vivo to mediate a global regulatory mechanism from the cell surface (Liu et al. 2024). Aligned with the idea of RALF-pectin interaction is the finding that RALF4-LRX8 LRR domain complex interacted with de-esterified pectin through the polycationic C-terminal domain of RALF4, forming a network of puncta with sizes suggestive of heterotetrameric complexes (Moussu et al. 2023). C-terminal–mutated RALF4s were compromised for this interaction; they were also less able to restore normal growth of amiralf4 ralf19 mutant pollen tubes. Instead, amiralf4 ralf19 pollen tubes expressing the C-terminal–mutated RALF4s displayed a wall quality more akin to control mutant tubes than to wild-type RALF4-complement amiralf4 ralf19 tubes. Therefore, the ability for RALF4-LRX8-pectin interaction correlated well with pollen tube wall quality and integrity and tube growth property. Given the hypothesis that RALF peptides regulate 2 parallel but mechanistically distinct but convergent pathways in pollen tubes (Moussu et al. 2020) and that the RALF C-terminal interacts with FER (Xiao et al. 2019), whether and how the C-terminal–mutated RALF4s (Moussu et al. 2023) would interact with ANX/BUPS and regulate their signaling will be interesting to explore.

Proper pollen tube growth rate and integrity rely on many parallel and intertwining factors (Figs. 2E and 3; Supplementary Tables S2 and S3). Some of these factors function through balancing the intracellular turgor pressure and cell wall mechanics during pollen germination and oscillatory growth. Disrupting this balance usually results in pollen tube rupture. For instance, the Cl⁻-gating mechanosensitive-like channel 8 (MSL8) senses membrane tension caused by the osmotic changes during pollen hydration, and loss of its function causes pollen rupture during germination (Hamilton et al. 2015; Hamilton and Haswell 2017). MSL8 intersects exocytosis and possibly the ANX1/2 pathway to achieve osmotic balance and membrane integrity (Zhou et al. 2021b; Wang et al. 2022a; Zhou et al. 2022). At the same time, the apical Ca²⁺ gradient and RAC/ROPs underlie the oscillatory growth through cellular feedback loops (Tunc-Ozdemir et al. 2013; Lichocka et al. 2022), actin-based cytoskeleton assembly/disassembly (Cheung et al. 2010; Qu et al. 2017), and sphingolipid-mediated feedbacks between the turgor pressure and cell wall thickness (Chen et al. 2015).

Pollen tube guidance

Pollen tube emergence from the septum

Architectural and physiochemical properties along the transmitting tract provide an environment for pollen tubes to grow directionally toward their target ovules. Pollen tube guidance (Box 1), further divided into pre-ovular and ovular, is a phenomenon whereby female cues guide a pollen tube to change its growth trajectory, first to reach an ovule, then to target the ovular aperture micropyle and penetrate the FG to deliver sperm (Figs. 1D, 2A and 3) (Higashiyama and Takeuchi 2015; Higashiyama and Yang 2016). The mechanisms for pre-ovular guidance remain unclear, conceivably involving various physicochemical conditions along the pollen tube growth path (Supplementary Table S3) (Crawford et al. 2007; Herrera-Ubaldo et al. 2019). Ovular guidance, studied mainly in dicots in recent years, is further divided into funicular guidance that each pollen tube follows from the septum to reach the ovule, and micropylar guidance, during which an approaching pollen tube responds to attractants produced by the FG and then targets and enters the micropyle.

For funicular guidance (Mizuta and Higashiyama 2018), studies in Arabidopsis showed that the pollen tube engages actively in this event. Loss of 2 pollen-specific and endoplasmic reticulum-located K⁺ transporters CATION/PROTON EXCHANGERS 21/23 (CHX21/23) resulted in mutant pollen tubes bypassing all the ovules to reach the bottom of the pistil. Nevertheless, chx21 chx23 pollen tubes also failed to target the micropyle when exposed directly to ovules (Lu et al. 2011). As K⁺/H⁺ exchangers, CHX21 CHX23 could function, albeit not been investigated yet, through modifying cation balance and pH, conditions critical for proper pollen tube growth. Loss of a GPI-AP, COBLA-like 10, induced phenotypes similar to that of chx21 chx23 (Li et al. 2013). Other factors, such as phytosulfokine (Stührwohldt et al. 2015), mitogen-activated protein kinase cascades (Guan et al. 2014), and the shikimate biosynthetic pathway (Wang et al. 2023b) also impact funicular guidance.

Ovular guidance

Synergid cell-produced pollen tube attractant LUREs were first discovered from the architecturally unique T. fournieri ovules with exposed ES (Okuda et al. 2009); much mechanistic insight has been advanced since (Higashiyama and Takeuchi 2015; Li et al. 2018; Johnson et al. 2019; Zhou and Dresselhaus 2019). Seven TfLURE homologs from Arabidopsis thaliana (A. thaliana), AtLURE1s, belong to a subclade among hundreds of cysteine-rich proteins (CRPs) (Takeuchi and Higashiyama 2012). They are secreted by the synergids and diffuse to the micropyle and funicular surfaces. AtLURE1 receptors include 2 receptor-like cytoplasmic kinases (RLCKs), LOST IN POLLEN TUBE GUIDANCE 1 and 2 (LIP1/2), and the pollen tube tip-localized PRK6 (Liu et al. 2013; Takeuchi and Higashiyama 2016; Zhang et al. 2017). PRK6 interacts with LIP1/2, which lack an extracellular domain, and the PRK/RLCK module could perceive AtLURE1s and mediate cytoplasmic signaling. Additional cell surface receptors also partake in ATLURE1 perception. For instance, the LRR-RLKs MDIS1-INTERACTING RECEPTOR LIKE KINASE 1/2 (MIK1/2) and MALE DISCOVERER 1/2 (MDIS1/2) form a heterodimer and participate in AtLURE1-regulated growth (Wang et al. 2016).

How do attractant receptors regulate pollen tube reorientation? PRK6 and several other pollen-expressed homologs from different species, for example, the tomato LePRK1 and LePRK2 (Kaothien et al. 2005), interact with ROPGEFs, which activate RAC/ROPs (Berken et al. 2005; Gu et al. 2006). RAC/ROPs regulate cytoskeleton dynamics, cytosolic calcium gradients, and exocytosis, all crucial for polar cell growth (Gu et al. 2005; Lavy et al. 2007). Modeling and experimental studies showed that pollen tube guidance results from reorienting the ROP1-exocytosis signaling network that regulates tip growth (Luo et al. 2017). During semi-in vivo growth and after exposure to AtLURE1, the symmetric distribution of PRK6 on the pollen tube tip rapidly changed to asymmetric and concentrated more on the side of attractant application. This was followed by the pollen tube redirecting its growth trajectory toward the attractant. Together, evidence strongly supports that pollen tube guidance is a result of AtLURE1-triggered repositioning of the LURE receptor-ROPGEF-RAC/ROP signaling module and the downstream growth supporting activity at the apical dome (Takeuchi and Higashiyama 2016; Luo et al. 2017).

The central cell also plays an important role in micropylar guidance in Arabidopsis. CENTRAL CELL GUIDANCE (CCG) is a TFIIB domain-containing central cell-specific transcription factor (Fig. 3). ccg mutant ovules display normal ES development but lack pollen tube attraction ability, resulting in severe sterility (Chen et al. 2007). Central cell-located CCG BINDING PROTEIN 1 (CBP1) interacts with CCG to recruit the Mediator complex and RNA Pol II (Li et al. 2015b). Several genes, including MYB98, crucial for filiform apparatus development, are downregulated in ccg ovules, consistent with the CCG-CBP1 complex functioning in pollen tube guidance through the MYB98 signaling pathway.

The pollen tube itself plays a critical role in responding to ovular guidance. For instance, the pollen tube–localized Ca²⁺channel CNGC18 and GLRs regulate pollen tube guidance by modulating Ca²⁺ oscillations at the pollen tube tip (Michard et al. 2011; Gao et al. 2016). In response to ovular signals, CNGC18 was asymmetrically recruited by MLO5/9/15 to plasma membrane on the side where the pollen tube senses the ovular signals to induce Ca²⁺ influx and direct the pollen tube on the septum surface to the nearest funiculus. Unable to induce this, mlo5/9 pollen tubes penetrated the septum to emerge on the septum surface but twisted into a pile, apparently not able to decode the ovular signals. The pollen tube endoplasmic reticulum serves a depot in synthesizing, assembling, and chaperoning key molecules to the apical dome membrane to mediate growth in response to micropylar guidance (Li et al. 2011; Dai et al. 2014; Xue et al. 2022).

Due to the architecture of the pistil, monocots rely only on ovular guidance (Lausser and Dresselhaus 2010). For instance, the maize transmitting tract terminates at the chalazal end and the pollen tube grows toward the micropylar region following the ovary anatomy and enters by breaking through the inner epidermis of the ovary wall (Zhou and Dresselhaus 2019). RNAi knock-down of the egg apparatus specific peptide, EGG APPARATUS 1 (ZmEA1), induced a severe micropylar targeting defect and significantly reduced seed sets (Márton et al. 2005).

Pollen tube reception

After entering the micropyle, the pollen tube interacts with one of the synergid cells, the receptive synergid, to undergo pollen tube reception, terminating in pollen tube bursting in the FG to release sperm (Hater et al. 2020; Hafidh and Honys 2021) (Fig. 1; Box 1). The process lasts approximately 30 to 60 min, including a brief halt in pollen tube growth, and is accompanied by the degradation of the receptive synergid (Sandaklie-Nikolova et al. 2007; Hamamura et al. 2011; Ngo et al. 2014). The FER-LRE signaling module located at the filiform apparatus (Box 1) of maturing and receptive ovules (Fig. 1) was identified through their roles in pollen tube reception (Huck et al. 2003; Rotman et al. 2003; Escobar-Restrepo et al. 2007; Capron et al. 2008; Tsukamoto et al. 2010; Duan et al. 2014; Li et al. 2015a; Liu et al. 2016a). Loss of FER or LRE alone prevents pollen tube bursting in the FG, resulting in pollen tube overgrowth (Fig. 2F). HERCULES1 and ANJEA participate redundantly in pollen tube reception. In vitro, they complexed with FER (Galindo-Trigo et al. 2020). Five pollen-expressed RALFs (RALF6/7/16/36/37) function as ligands for the FER/ANJEA/HERCULES1-LRE complex (Zhong et al. 2022a). The ralf quintuple mutant pollen tubes failed to burst in wild-type FGs. In vitro, these RALFs interact with each receptor and enhanced interactions between FER and ANJEA/HERK1, suggesting receptor-ligand complexes of varying peptide-receptor kinase-LRE configurations are possible in vivo (Zhong et al. 2022a, 2022b).

A Powdery Mildew Resistance Locus-O (MLO) family protein, NORTIA (NTA)/MLO7, interacts with FER-LRE to participate in pollen tube reception (Kessler et al. 2010). Before pollen tube arrival, NTA/MLO7 is maintained intracellularly in the synergids. Pollen tube arrival induced FER-dependent relocation of NTA/MLO7 to FER-LRE-occupied filiform apparatus. The approaching pollen tube induced the degeneration of the receptive synergid cell is linked to pollen tube rupture. The approaching pollen tube induced FER-, LRE-, and NTA-dependent changes in synergid cytoplasmic Ca²⁺ dynamics (Iwano et al. 2012; Denninger et al. 2014; Ngo et al. 2014; Gao et al. 2022; Ponvert and Johnson 2024). The calcium status/dynamics in the pollen tube is also important for pollen tube rupture. The plasma membrane–localized Ca²⁺-ATPase ACA9 plays a pivotal role for cytosolic Ca²⁺ efflux, maintaining [Ca²⁺] within the pollen tube. aca9 pollen tubes halted upon entering the micropyle, and bursting did not occur (Schiøtt et al. 2004). Moreover, the pollen tube apical [Ca²⁺] gradient was obliterated just before bursting in vitro and required Ca²⁺ channel activity. Sequestration of pistillate [Ca²⁺] inhibited FER-mediated pollen tube bursting in the ovules (Duan et al. 2014). Genes regulating synergid cell death also appeared to be involved (Christensen et al. 1997; Leydon et al. 2015; Mendes et al. 2016; Wang et al. 2017b), consistent with pollen tube penetration and sperm release being tightly linked.

In monocots, K⁺ plays an important role in pollen tube rupture. The maize pollen tube cytoplasmic membrane localized potassium channel KZM1 functions downstream of the synergid cell-secreted peptide EMBRYO SAC4 (ZmES4). The interaction between ZmES4 and KZM1 induced K⁺ influx, the change in osmotic pressure due to water uptake might have led to pollen tube bursts (Amien et al. 2010). In rice, the K⁺ transporter OsHAK1/19/20 interacted directly with pollen-expressed FER homolog RUPTURED POLLEN TUBE (RUPO) to regulate between pollen tube integrity and rupture (Liu et al. 2016b).

Gamete interaction and fertilization

Fertilization and the subsequent zygote development are pivotal stages in the life cycles of both animals and plants, during which haploid male and female gametes fuse to generate a diploid organism. With immotile sperm and 2 gamete fusion events, angiosperm fertilization is complex. Upon sperm release, the 2 gametic fusions usually occur around a 7-minute time window, probably the time needed for gametic cell recognition, attachment, and activation (Hamamura et al. 2011). Then, 1 sperm cell fuses with the egg cell to form the diploid embryo, and the other fuses with the central cell to generate the triploid endosperm. We discussed recent advances, mainly from studies in Arabidopsis, in the understanding of gametic and early post-zygotic events.

Gamete recognition and attachment

In Arabidopsis, the 2 sperm cells in a pollen grain seem isomorphic, i.e. without preferential fusion of either sperm with the egg cell or central cell (Mori et al. 2006; Liu et al. 2008a, 2008b; Ingouff et al. 2009; Hamamura et al. 2011; Cole et al. 2014; Chen et al. 2022). In contrast, the 2 female gametes are dimorphic, showing distinct fertilization preference. Typically sperm-egg fusion occurs first, followed by sperm-central cell fusion several minutes later (Denninger et al. 2014; Hamamura et al. 2014). The use of cell-cycle defective mutants that produced a single sperm-like cell but remaining competent for fertilization showed that the egg cell is preferentially fertilized than the central cell (Li et al. 2022a; Wang et al. 2024), and egg cell–specific aspartic endopeptidases as well as small CRPs might partially contribute to the preferential fertilization of egg cell (Jiang et al. 2022a; Wang et al. 2024).

Current knowledge about gamete recognition and attachment (Fig. 3) has mainly derived from GAMETE EXPRESSED 2 (GEX2), a sperm plasma membrane protein, whose expression is controlled by DUO1 (Brownfield et al. 2009; Borg et al. 2011). GEX2 is expressed with an N-terminal signal sequence, a filamin repeat domain followed by a single-pass transmembrane domain at the C terminal. In vivo gamete attachment assay demonstrated that GEX2 is essential for sperm-egg and sperm-central cell attachment (Mori et al. 2014), consistent with both sperm cells having comparable capacity to adhere to either female gamete for fertilization. Three sperm cell endomembrane-associated cation/H⁺ exchangers (CHXs) are also required for gamete attachment, which failed in the chx17/18/19 mutant (Chen et al. 2022). Disturbed osmotic balance in chx17/18/19 sperm cells and their resultant rupture inside the ES suggest an interplay between turgor and the sperm cell surface.

Gamete activation, fusion, and sperm nucleus migration

After gamete attachment, contributions from the male and female sides are required for the critical events of gamete activation and fusion (Fig. 3). On the female side, a group of egg cell–specific small CRPs, EGG CELL 1 (EC1), are located in vesicles of the unfertilized egg cell. Upon sperm arrival at the sperm-egg fusion sites, the egg cell secretes EC1, and this stimulates the redistribution of the fusogen GCS1/HAP2 from the sperm endomembrane system to its cell surface, a step critical for sperm cell activation and fertilization competence acquisition (Sprunck et al. 2012). ec1-RNAi plants had 2 or 4 unfused sperm cells in about one-half of the ESs, consistent with EC1 is essential for gamete fusion. On the male side, GCS1/HAP2, is an evolutionarily conserved protein for gamete fusion in animals and plants and essential for gamete fusion in algae and angiosperms (Mori et al. 2006; Liu et al. 2008a; Cole et al. 2014; Wang et al. 2022d). The Arabidopsis GCS1/ HAP2 functions by interacting via its N terminus with female gamete-expressed proteins, while its positively charged C terminus facilitates fusion between the plasma membranes. The EC1-triggered GCS1/HAP2 translocation is dependent on 2 sperm-specific DOMAIN OF UNKNOWN FUNCTION 679 membrane proteins (DMP8 and DMP9) with functions conserved across seed plants (Wang et al. 2022d). Sperm from dmp8 dmp9 preferentially fertilized the central cell rather than the egg, consistent with sperm-egg fusion being more dependent on DMP8/9 (Cyprys et al. 2019).

After gamete membrane fusion, gamete nuclei migration is the critical next step for the fusion of parental genomes to generate the zygotic genome. In contrast to the microtubule-assisted female pronuclei movement toward the male nucleus in fertilized animal eggs, an F-actin meshwork supports sperm nuclear migration toward the nucleus of female gametes in Arabidopsis in a RAC/ROP regulated process involving formin-regulated actin nucleation and a plant-specific myosin, XI-G (Kawashima et al. 2014; Ohnishi et al. 2014; Peng et al. 2017; Ali et al. 2020; Ali and Kawashima 2021).

Genome-wide zygotic genome activation (ZGA)

Genome-wide ZGA

The fusion of a large egg cell and a small sperm cell generates the fertilized egg cell, the zygote, and the starting point of a new generation in both animals and plants. In animals, ZGA is usually characterized by a minor and a major activation waves (Tadros and Lipshitz 2009). The animal process may start as early as the 1-cell embryo stage or spread out over longer durations, depending on species (Vastenhouw et al. 2019). Starting about 20 years ago and focusing on the eudicot Arabidopsis and monocots rice and maize, the relatively short history of plant ZGA studies has already yielded major insights. In contrast to animals, a comparative transcriptome analysis of egg cells and zygotes revealed that ZGA in eudicots and monocots occurs shortly after fertilization. For example, in Arabidopsis, ∼13% of its genes were already activated before the first zygotic division (Zhao et al. 2017; Dresselhaus and Jürgens 2021; Zhao et al. 2022a; Huang et al. 2023). Notably, among the activated genes are those encoding transcription factors important for zygote and early embryo development, such as WUSCHEL RELATED HOMEOBOX (WOX)2, WOX9, BABY BOOM (BBM), and its homolog PLETHORA 2 (PLT2) (Zhao et al. 2019). In maize, sperm, egg, and zygote transcriptomes at 2 different developmental stages showed that ∼9% of its almost 40,000 annotated genes began to be transcribed in zygotes (Chen et al. 2017). On the other hand, only 181 genes are expressed de novo in rice zygotes (Anderson et al. 2017). Regardless, BBM and WOX9 homologs were activated in maize and rice zygotes, suggesting conserved roles in ZGA between eudicots and monocots. Taken together with comparative transcriptomic data from the brown alga Dictyota (Bogaert et al. 2017), evidence strongly supports that plant ZGA occurs at the zygotic stage, and de novo transcription is required for zygote division (Bogaert et al. 2017; Kao and Nodine 2019; Zhao et al. 2019).

Parental contributions to ZGA

Parental genomes merge upon fertilization. In mammals, the genetic information deposited in the egg cell controls early embryogenesis (Tadros and Lipshitz 2009; Vastenhouw et al. 2019). In plants, the paternal and maternal genomes fused quickly to form the zygotic genome after fertilization (Ingouff et al. 2010; Kimata et al. 2016). In Arabidopsis, genome-wide transcriptome analyses using zygotes and early embryos established that maternal and paternal genomes contribute equally to the transcriptomes of Arabidopsis embryos as early as the elongated zygote stage (Zhao et al. 2019; Zhao et al. 2020). In monocots, parental contributions to the transcriptomes of zygotes and early embryos remain largely unknown. Emerging information from hybrid rice zygote at 9 HAP suggests highly asymmetric parental contribution, with maternal alleles contributing to about 98% of transcripts (Anderson et al. 2017).

Elaborations to maximize reproductive success

Angiosperms have evolved multiple male-female interactive strategies to elaborate beyond the required steps to achieve fertilization to maximize progeny production and their vibrancy. While some of these phenomena are long-known, recent efforts illuminated emerging understandings of some fascinating strategies.

Polyspermy block and the pollen tube to one ovule rule

The entry of multiple sperm cells into a single egg (polyspermy) can lead to anomalies in chromosome numbers in the zygote, resulting in severe consequences in animals (van der Ven et al. 1985), and is generally considered disadvantageous for plants (Spielman and Scott 2008; Bianchi and Wright 2016; Tekleyohans et al. 2017). Contrasting the barriers within the animal female reproductive tract to prevent polyspermy (Bianchi and Wright 2016), angiosperms have evolved multiple mechanisms to preclude multiple pollen tube entrance into the same ovule to avert polyspermy.

A strict 1-to-1 relationship between pollen tubes and ovules is established during pollen tube growth in the pistil when they exit the main growth path to target ovules (Johnson et al. 2019; Cheung et al. 2022) (Fig. 2A). Recent studies have unveiled multiple mechanisms behind this “1 pollen tube to 1 ovule” phenomenon. First, pollen tube overgrowth in loss-of-function fer mutant pistils is frequently accompanied by supernumerary pollen tube penetration of single ovules (Fig. 2E), as well as bundled exits of multiple pollen tubes from the transmitting tract (Huck et al. 2003; Rotman et al. 2003; Duan et al. 2014; Duan et al. 2020; Zhong et al. 2022a). These dual anomalies implicate the relaxation of an early polyspermy block at the septum (Box 1) and a local block at the entry to the FG in the ovule (Fig. 3C). Genetic and biochemical studies suggest that upon the emergence of the first pollen tube from the septum, 5 pollen tube-secreted RALFs (RALF6/7/16/36/37) interact with septum-located receptors FER/ANJEA/HERCULES1 to establish the early polyspermy block (Zhong et al. 2022a). Various multiplexed ralf6/7/16/36/37 mutant pollen tubes failed to trigger the block, resulting in increased levels of multiple pollen tubes exiting as bundles and entering single ovules. Thus, the ovule phenotypes in multiplexed ralf-pollinated wild-type pistils phenocopied ovules from wild-type pollen tube penetrated fer pistils.

The local polyspermy block is triggered upon the entry of the first pollen tube into the ovule and fertilization is assured (Fig. 3C). Thus far, 2 distinct mechanisms involve the inactivation of the pollen tube attractant AtLURE1s (Duan et al. 2020; Yu et al. 2021). The FERONIA-controlled block is mediated by signaling events involving FERONIA, pectin, and pollen tube arrival–triggered nitric oxide (NO) at the filiform apparatus. NO mediates a 2-prong mechanism to disengage the AtLURE1-mediated pollen tube attraction, one to inhibit the activity of the already secreted attractant and the other to prevent its further secretion from the synergids (Duan et al. 2020). Upon fusion with a sperm cell, the egg might also aid in fortifying the polyspermy block by secreting 2 aspartic proteases, ECS1 and ECS2, which would cleave and render any residual attractants inactive (Yu et al. 2021).

Fertilization recovery

Angiosperms have evolved an amazing capacity, referred to as “fertilization recovery” (Box 1), to salvage failed fertilization, for example, because of incompetent sperm or natural calamity has delayed pollination during the peak fertility window. With fertilization recovery, the plant can attract additional pollen tubes to a penetrated but not fertilized ovule or relax the polyspermy control in a dwindling time frame, affording more opportunities to produce offspring (Kasahara et al. 2012; Kasahara et al. 2013; Cheung et al. 2022). Although a naturally occurring process, studies of mutations in A. thaliana that impact the ability of sperm cells to fuse with the female gametes enabled systematic characterization and mechanistic elucidation of this fascinating phenomenon (Beale et al. 2012; Kasahara et al. 2012). When the cdka;1, duo1, duo3 and hap2 mutant pollen pollinated wild-type pistils, they induced significantly increased number of ovules penetrated by multiple pollen tubes, a phenotype referred to as “polytubey” (Beale et al. 2012) (Box 1). A variety of mutations that induce supernumerary pollen tubes in an ovule have since been reported (Supplementary Table S5) (Völz et al. 2013; Maruyama et al. 2015; Pereira et al. 2016b; Nagahara et al. 2020; Yu et al. 2021; Zhong et al. 2022a; Meng et al. 2023).

What triggers fertilization recovery? The principle of fertilization recovery assumes that the synergids, egg, or central cell somehow sense fertilization is threatened, thus the need to salvage the situation (Fig. 3C). While multiple mechanisms likely exist, one obvious requirement is to re-open the early gate created at the septum upon pollen tube exit and the local gate in the ovule generated by the passage of the first pollen tube. Another requirement is to ensure the presence of attractants to guide additional pollen tubes. At the local gate, a pollen tube arrival–triggered FER- and pectin-dependent NO block disengages attractants. Since NO is gaseous, the local gate is easily reversible when NO has dissipated (Duan et al. 2020). Another mechanism emanates from the egg cell. When fertilization failed, the egg cell does not secrete the ECS1 and ECS2 proteases, which degrade attractants, thus preserving them and improving the chances to re-implement pollen tube attraction and fertilization recovery (Yu et al. 2021). How the septum polyspermy block is controlled remains to be demonstrated. Studies of RALF6,7,16,36,37 led to a provocative model that links pollen tube burst to fertilization recovery (Zhong et al. 2022a). It posits that the pollen RALFs are also required not only for establishing but also maintaining the polyspermy block and that under normal conditions, with the first-arriving pollen tube bursting in the FG, the production and secretion of the pollen RALFs would halt. Without the pollen RALFs for maintenance, the septum block re-opens to allow exits of additional pollen tubes to target already penetrated ovules.

Persistent synergid cell represents a crucial component for fertilization recovery and the phenomenon depends on the persistent synergid not being terminally degraded so that it can reactivate attractant production. After the first fertilization attempt fails, the receptive synergid cell degrades (Maruyama and Higashiyama 2016), and the ES relies on the persistent synergid cell to continue secreting attractants to attract additional pollen tubes. It could be argued that the persistent synergid exists for the purpose of fertilization recovery. However, it also poses a risk to the polyspermy block, so it needs to be degraded rapidly upon fertilization. A mechanism that mitigates this threat could be the fusion between the persistent synergid and the endosperm, thus rapidly diluting the persistent cell-produced attractants (Maruyama et al. 2015). Yet fertilization recovery depends on maintaining the attractant factory (Fig. 3C). Therefore, the survival of the persistent synergid cell appears to be an outcome of a delicate balance. Multiple mechanisms for the degradation and survival of the persistent synerdid have emerged. These involve, for example, components of ethylene signaling, though not ethylene per se (Völz et al. 2013; Li et al. 2022a, 2022b, 2022c) and epigenetic regulation (Maruyama et al. 2013). While how these mechanisms integrate into the orchestration of fertilization recovery, they illustrate the connection between the degradation of the persistent synergid cell and the fertilization recovery-induced secondary pollen tube entrance to ovules.

The central cell also plays a crucial role and participates directly in fertilization recovery. When double fertilization fails and both synergids have degenerated after pollen tube entry, the central cell secretes additional peptide pollen tube attractants, SALVAGER1 and SALVAGER2, to attract late-arriving pollen tubes (Dresselhaus and van der Linde 2023; Lan et al. 2023b; Meng et al. 2023). In the absence of functional synergids, as in myb98 mutant ovules, SALVAGERs and its upstream transcription factor CCG are upregulated, and SALVAGERs retained in the central cell before fertilization are directionally secreted to the micropyle (Meng et al. 2023) (Fig. 3B). These observations suggest that the synergid has an inhibitory effect on the transcription and secretion pathways of SALVAGERs, which would normally be switched on for pollen tube attraction and reception after the synergids have been eliminated or are impaired. Compared with the transient effect of synergid-based fertilization recovery, this central cell-controlled fertilization recovery mechanism acts as a main back-up system for fertilization success (Dresselhaus and van der Linde 2023). These observations set the stage for the elucidation of the multifaceted role of the central cell and its interaction with the synergids (Box 2).

Interspecific reproduction barriers

In natural surroundings, besides responding to compatible or incompatible pollen of their own species (intraspecific) via SI (Díaz et al. 2021; Goring et al. 2023; Nasrallah 2023), pistils also must respond to and set up barriers for pollen of other species (interspecific) (Fig. 4). Interspecific reproductive barriers prevent crossing between species that are not likely to produce healthy progeny and ensures species integrity. Overcoming hybridization barriers between interspecific species could facilitate species diversification and promote vigor. Interspecific barriers are diverse and genetically complex and could be implemented on multiple levels. Because of their importance in species preservation and diversification, and in agricultural concerns of the introgression of genes encoding high-quality traits, interspecific reproductive barriers are frequent subjects of reviews, for example, in the Brassicaceae (Kitashiba and Nasrallah 2014; Nasrallah 2023) and the Solanaceae (Baek et al. 2015; Bedinger et al. 2017; Broz and Bedinger 2021). Here, we focus our discussions on recent efforts surrounding the self-compatible A. thaliana and its related Brassicaceae species.

Figure 4.

Intra- and interspecific pollination. B. rapa and A. thaliana are used as representative self-incompatible (SI) and self-compatible (SC) species. For intraspecific pollination, SI species displays incompatibility to self-pollen or pollen with the same S-haplotype (brown X) but are receptive to pollen with distinct haplotype (arrows with no X). Interspecific pollination often follows the SI × SC rule; SI species show unilaterally incompatibility (UI) to pollen of another species (yellow X), while SC species are unilaterally compatible (UC) to interspecific pollen, albeit they often display a notable preference for intraspecific pollen over interspecific pollen (compare yellow tubes and orange tubes in the A. thaliana pistil).

Stigmatic barriers

Membrane located-regulators as stigmatic gatekeepers

Pollen-pistil interaction studies revealed SI-independent and SI-dependent stigmatic gating mechanisms against interspecific hybridization. In Brassicaceae SI, self-pollen is recognized by the ligand–receptor interaction between the pollen-expressed SCR/SP11 and its receptor, the stigma-expressed, plasma membrane-localized SRK, followed by the activation of downstream signaling pathways in papilla cells that ultimately lead to the rejection of self-pollen (Kitashiba and Nasrallah 2014; Nasrallah 2023). SI-independent interspecific incompatibility was identified by Fujii et al. 2019 in a GWAS screen. Using more than 300 A. thaliana lines as recipients of pollen from the distantly related Malcolmia littorea, Fujii et al. identified the stigma-specific gene STIGMATIC PRIVACY 1 (SPRI1) in A. thaliana as important for interspecific incompatibility. M. littorea pollen typically fails to germinate on A. thaliana Col-0 stigmas, but it was able to germinate and penetrate various A. thaliana ecotypes that harbor mutations in the SPRI1 gene or Col-0 T-DNA knockouts of SPRI1, which encodes a 221-amino-acid stigmatic papilla transmembrane protein. However, spri1 mutant expressing a functional SRK retained the typical SI capacity, rejecting pollen expressing the cognate SI pollen factor, SP11/SCR, but did not impact the penetration of M. littorea pollen tubes into the mutant stigma, indicating that SPRI1 is functionally independent of SI. The underlying mechanism of how SPRI1 regulates interspecific incompatibility remains to be explored.

Redox conditions controlled by the FER to ROS pathway in the Brassicaceae stigma (Liu et al. 2021; Zhang et al. 2021) play a major role in establishing an SI-dependent barrier for incompatible pollen intrusion (Franklin-Tong and Bosch 2021; Zhang et al. 2021; Huang et al. 2023). Like self-pollination in Arabidopsis, the self-compatible (SC) Brassica rapa (B. rapa) pollen suppresses B. rapa stigmatic ROS via PCP-B-triggered stigmatic NO increase and the mirrored decline of FER to ROS signaling, allowing pollen hydration and germination to proceed (Liu et al. 2021). On the other hand, SI pollination induced rapid increases in stigmatic ROS, paralleling the SI response of pollen arrest. Suppressing SRK expression on the stigma relaxed the SI pollen response. It abrogated the SI pollen-induced stigmatic ROS response, establishing stigmatic ROS as an SI-dependent barrier for incompatible pollen. The underlying mechanism involves allele-specific SP11/SCR-enhanced SRK-FER interaction, stimulating the FER to ROS pathway and fortifying the intraspecific barrier for SI pollen.

In nature, plants could be nondiscriminately pollinated, including by sympatric species, some could be closely related. A mechanism similar to that in the Arabidopsis stigma also mediates the B. rapa stigmatic barrier for other Brassicaceae species. For stigmas of SI species, an SRK-dependent rise of the FER-controlled stigmatic ROS levels was triggered by interspecific pollen to reject them (Huang et al. 2023; Lan et al. 2023; Song et al. 2023; Wang and Filatov 2023). The FER-controlled ROS stigmatic gate not only blocked pollen from distantly related species but also delayed the hydration and germination of closely related interspecific compatible pollen, resulting in retarded pollen tube penetration of the stigma relative to SC pollen. The tunable stigmatic ROS gate was found to be responsive to PCP-B in a species differential manner, preventing interspecific hybridization and providing a head-start advantage for conspecific pollen to complete self-fertilization before pollen from closely related interspecies have a chance for hybridization, thereby ensuring species integrity.

Cell wall quality as gatekeeper

The cell walls at the pollen-stigma interface are apparently important for species differentiation. Fujii et al. 2023 reported a zinc-finger transcriptional factor, STIGMATIC PRIVACY 2 (SPRI2), as having a controlling role in interspecies incompatibility in A. thaliana. SPRI2 is important in maintaining the expression of many genes related to cell wall organization or biosynthesis, including 2 genes encoding xylan or pectin O-acetyltransferases, TRICHOME BIREFRINGENCE-LIKE (TBL)40 and TBL45, as its direct targets. Loss of SPRI2 or TBLs relaxed the stigmatic barrier, allowing the interspecific pollen from M. littorea and C. rubella to hydrate and germinate and their tubes to penetrate the mutant pistils (Fujii et al. 2023). Furthermore, Lan et al. (2023a) showed that not only pollen from ralf sept mutant Arabidopsis but also pollen from interspecific incompatible species were able to penetrate stigmas from loss of LRX3,4,5 mutants. With LRXs being an integral part of the cell wall matrix (Borassi et al. 2016; Bedinger 2018; Zhao et al. 2018; Herger et al. 2019), the findings from Lan et al. (2023a) directly supports papilla cell wall quality as critical for a pollen tube gating mechanism.

Studies in other incompatibility systems also support that the pollen tube wall property is an important determinant for barrier-crossing capacity. For example, the maize Teosinte crossing barrier1-s (Tcb1-s) genes confer cross-incompatibility between Teosinte and maize. The female Tcb1-s prevents growth and fertilization by interspecific pollen and encodes a pistil-expressed pectin methylesterase (PME), a key enzyme in regulating cell wall stiffness and porosity (Lu et al. 2019). The maize Gametophyte factor 1 (Ga1) locus, which controls intraspecific cross-incompatibility between popcorn and sweet corn, involves at least a silk-specific PME and 2 pollen-expressed PMEs (Moran Lauter et al. 2017; Zhang et al. 2018). The results underscore the importance of properly regulated pectin status in the cell wall in maintaining stigmatic incompatibility barriers.

Reproduction barriers at the level of pollen tube attraction

Earlier studies of attractants from Arabidopsis and maize demonstrated interspecific barriers at the level of pollen tube attraction to the FG. For LURE attractants, recombinant LURE1 from A. thaliana preferentially attracted A. thaliana pollen tubes than those from A. lyarata in vitro. In planta, the expression of the Arabidopsis LURE1 in the synergid cells of T. fournieri guided the heterologous A. thaliana pollen tube to the Torenia FG (Takeuchi and Higashiyama 2012). Expression of the maize attractant EA1 in A. thaliana ovules also augmented the frequency of the maize pollen tube targeting the A. thaliana ovule micropyle (Márton et al. 2012). Studies of AtLUREs further showed that in semi-in vivo assays Capsella rubella (C. rubella) pollen tubes expressing the A. thaliana PRK6 responded to AtLURE1, unlike wild-type C. rubella tubes (Takeuchi and Higashiyama 2016). Implicated in AtLURE1 perception, the Arabidopsis receptor kinase MDIS1, when expressed in C. rubella pollen tubes enhanced Capsella pollen tube targeting to A. thaliana ovules (Wang et al. 2016). These results indicate that both cross-species expression of attractants or their receptors on the pollen tubes could overcome the barrier in pollen tube guidance.

Pollen tube guidance plays an important role in maintaining interspecific isolation. For example, compared with A. thaliana pollen tubes, A. lyrata and the more distantly related Cardamine fluexuosa pollen tubes were increasingly less able to target A. thaliana ovules (Escobar-Restrepo et al. (2007). The contribution from pollen tube attractants is underscored by recent studies of AtLUREs and their related XUIQIUs and TICKETs, all from the CRP810 subclade (Takeuchi and Higashiyama 2012; Meng et al. 2019; Zhong et al. 2019a). The key function of cognate AtLURE1-PRK6 attractant-receptor interaction was determined to be for facilitating the exit of conspecific pollen tubes from the A. thaliana septum where interspecific A. lyrata pollen tubes emerged considerably later than the self-A. thaliana pollen tubes. AtLUREs therefore promote conspecific precedence in fertilization, reducing the chance for interspecific hybridization (Zhong et al. 2019a). Conversely, the self-pollen tubes were significantly delayed in their exits from the atlure1 null mutant transmitting tract, exits of A. lyrata pollen tubes were not affected at all. A member of synergid-secreted non-CRP peptide NPA1 also contribute to species barrier by its species-specific pollen tube attraction activity among sister species in Brassicacea (Wang et al. 2024).

Barrier at the level of pollen tube reception in the ovule

Interspecific barriers also occur in the FG to prevent fertilization by an already penetrated pollen tube from a distant species, for example, in the Rhododendron, interspecific pollen tubes continued growth within the FG, aborting fertilization (Williams et al. 1986). Furthermore, of the tubes attracted to A. thaliana ovules, approximately 50% of A. lyrata and 70% of C. fluexusa failed to burst inside the FG, correlated interspecific barrier strength with species isolation (Escobar-Restrepo et al. 2007). Interestingly, ectopic accumulation of NTA/MLO7 at the filiform apparatus suppressed the interspecific barrier resulting in significantly reduced levels (approximately 25%–30%) of A. lyrata pollen tube overgrowth in these transformed A. thaliana ovules (Ju et al. 2021). A. lyrata and interspecific Olimarabidopsis pumila pollen tubes, which could not burst in A. thaliana FG, nevertheless triggered FER-LRE-dependent changes in A. thaliana synergid cell Ca²⁺ oscillation dynamics, suggesting that downstream signaling of the FER-LRE-NORTIA apparatus plays key roles in maintaining the interspecific barrier (Ponvert and Johnson 2024).

Crossing barriers: emerging capability and continuing challenges

Genetic hybridization can rapidly expand species divergence in the ecosphere and is an important agricultural tool to generate plants with improved performance, especially under suboptimal conditions, or improved quality as a food source. Interspecific barriers for pollen invasion have been observed as weakened by genetic lesions impacting gating mechanism (e.g. Jones et al. 2017; Ju et al. 2021; Huang et al. 2023; Lan et al. 2023a) or the ability to cross barriers could be gained by heterologous expression of interspecific determinants (e.g. Márton et al. 2012; Takeuchi and Higashiyama 2012, 2016; Wang et al. 2016). These demonstrations suggest the possibility to manipulate interspecific barriers allowing hybridization of closely or more distantly related species. Two sets of test-of-principal experiments, one in SI B. rapa (Huang et al. 2023) and the other in SC A. thaliana pistils (Lan et al. 2023a), demonstrated that genetically weakening the stigmatic barrier could indeed be engineered to allow interspecific pollination to proceed. In B. rapa, downregulating the functionally linked SRK, FER, or the FER-controlled RBOH during pollination allowed efficient intraspecific SI pollen tube penetration producing developmentally normal embryos. These manipulations also allowed a limited level of cross-fertilization success by closely related B. oleracea pollen and more distantly related B. vulagris, albeit producing developmentally compromised embryos. In A. thaliana, the FER-RALF-dependent, LRX3,4,5-gated stigmatic barrier for interspecific pollen tube penetration could be relaxed by loss of FER alone, CURVY1, ANJEA and HERCULES1 together, sRALFs, or all LRX3,4,5. Moreover, co-applying a pRALF capable of weakening the wild-type stigma during interspecific pollination provided a mentoring effect for interspecific pollen tube penetration, producing hybrid embryos. Together, test of principal experiments demonstrate that fundamental understanding of how pollen and pistil interact to maintain hybridization barrier could contribute to manipulating these barriers to modulate interspecific isolation.

Natural and synthetic apomixis

In angiosperm, over 400 species from different plant phyla can propagate asexually through apomixis (Box 1). Apomixis exists in the gametophytic and sporophytic types. In gametophytic apomixis, apospory is characterized by the direct formation of an ES from somatic ovule cell, while the diplospory refers to the formation of unreduced FG from the MMC by skipping meiosis (Hojsgaard and Hörandl 2019). In both cases, the diploid ES develops into clonal embryo parthenogenetically. Sporophytic apomixis through adventitious embryos is found in some Citrus species and Fortunella genus (Wang et al. 2017c; Wang et al. 2022b). Synthetic apomixis is a newly invented term, defined as the introduction of apomixis into sexual plants through genomic engineering (Xiong et al. 2023). Unlike sexual reproduction, the MMC in synthetic apomicts bypasses meiosis and divides mitotically to form unreduced FG. The unreduced egg cell then undergoes a pseudo-zygotic program to form an embryo without integrating the sperm genome, and the endosperm forms autonomously or differentiates from the fertilized central cell (Xiong et al. 2023; Li et al. 2023a). In the last decade, utilizing the combined genetic system of Mitosis instead of Meiosis (MiMe) and haploid induction (HI) strategies, synthetic apomixis has been achieved in Arabidopsis and rice (Marimuthu et al. 2011; Khanday et al. 2019; Wang et al. 2019; Xie et al. 2019; Vernet et al. 2022; Liu et al. 2023; Wei et al. 2023; Xiong et al. 2023).

Genes involved in MiMe in Arabidopsis and rice

Meiosis is a specialized form of cell division that reduces the ploidy of germ cells (Mercier et al. 2015). Meiosis differs from mitosis in many aspects, including 2 rounds of chromosome segregation (Meiosis I and II) and the segregation of sister chromatids in Meiosis II (Mercier et al. 2015). The core mechanism of MiMe is to inactivate the essential genes required for these unique meiotic processes (dErfurth et al. 2009; dErfurth et al. 2010; Cifuentes et al. 2016; Mieulet et al. 2016; Khanday et al. 2019; Xie et al. 2019; Vernet et al. 2022). In Arabidopsis and rice, homologs of the highly conserved SPO11 with sequence similarity to the A subunit of Archaeal Topoisomerase VI play a key role in the formation of DNA double-strand breaks to initiate meiotic recombination (Stacey et al. 2006; Benyahya et al. 2020; Fayos et al. 2020). Lesions in these SPO11 homologs abolish meiotic recombination (Li et al. 2023a, 2023b, 2023c; Xiong et al. 2023). Additional factors are also required, including PUTATIVE RECOMBINATION INITIATION DEFECT 1, 2, and 3 (PRD1, 2 and 3) and DSB FORMING (DFO) in Arabidopsis (De Muyt et al. 2007; De Muyt et al. 2009; Zhang et al. 2012; Shi et al. 2021; Wang et al. 2022c). In rice, PRD3/HOMOLOGOUS PAIRING ABERRATION IN RICE MEIOSIS1 (PAIR1), CENTRAL REGION COMPONENT 1 (CRC1), P31^comet, and SOLO DANCERS (SDS) are involved in DSB formation during meiotic recombination (Nonomura et al. 2004; Miao et al. 2013; Wu et al. 2015; Ji et al. 2016). Mutations in these factors hinder recombination events (Mercier et al. 2015). RECOMBINATION-DEFICIENT 8 (REC8) encodes an exclusive meiosis-specific cohesin subunit protein and its impairment causes aberrant meiosis I (Chelysheva et al. 2005; Sun et al. 2013). Other cohesion-related components, such as SISTER-CHROMATID COHESION PROTEIN 3 (SCC3), SHUGOSHIN1 (SGO1), and Bub1-related kinase 1 (BRK1), are also required for cohesion functions in meiosis (Chelysheva et al. 2005; Wang et al. 2012; Zamariola et al. 2013).

Genes associated with meiosis progression need to be modified to fully abrogate meiosis II. These include the A-type cyclin TARDY ASYNCHRONOUS MEIOSIS (TAM), CDKs, THREE DIVISION MUTATION 1 (TDM1), and OMISSION of SECOND DIVISION 1 (OSD1) (Dissmeyer et al. 2007; dErfurth et al. 2010; Cifuentes et al. 2016). During meiosis I, TAM interacts with CDKA;1 to suppress the activity of TDM1, a component of the Anaphase Promoting Complex/Cyclosome (APC/C) regulating the exit from meiosis II (Dissmeyer et al. 2007; Cairo et al. 2022). OSD1 is pivotal in initiating meiosis II by governing CDK activity in rice and Arabidopsis (dErfurth et al. 2009; Mieulet et al. 2016). Thus, OSD1, TAM, and TDM1 collectively govern critical transitions within the meiotic cell cycle, and mutations in any of these 3 genes result in the abrogation of meiosis II (dErfurth et al. 2010; Cifuentes et al. 2016).

Gene loci related to HI

HI presents an effective approach to bypass zygote formation. In maize, quir1 and quir8 are gene loci related to HI widely used in agricultural breeding (Wang et al. 2017c; Zhong et al. 2019b). quir1 a mutation in PLA1/NOT LIKE DAD (NLD)/MATRILINEAL (MTL). PLA1/NLD/MTL encodes a phospholipase A1 localized on the endo-plasma membrane (PM) around the 2 sperm cells, also known as the inner vegetative PM for having originated from the PM of the pollen VC (Dong et al. 2013; Gilles et al. 2017; Kelliher et al. 2017). The pla1/mtl/nld haploid inducer line displays a remarkably high rate of sperm DNA fragmentation (Li et al. 2017). Such fragmentation might be caused by perturbed redox status that triggers postfertilization paternal genome elimination, culminating in the generation of maternal haploids (Jiang et al. 2022a). The utilization of PLA1/MTL/NLD-based chromosome fragmentation has been extended to generating haploid-inducing lines in other monocot staple crops, such as wheat, rice, and foxtail millet (Cheng et al. 2021; Liu et al. 2023). The depletion of pollen-specific PHOSPHOLIPASE D3 (ZmPLD3) and PLA1/MTL/NLD also induces a higher HI rate in maize (Li et al. 2021).

Ectopic expression of BBM or its homologs in the egg cell of monocots or dicots induces autonomous transition to embryonic development (Conner et al. 2015; Khanday et al. 2019; Deng et al. 2022). In Pennisetum squamulatum, several BBM-like (BBML) genes in apospory-specific genomic region (ASGR) govern parthenogenesis (Akiyama et al. 2011). Notably, in diverse species, including tobacco, pearl millet, rice, maize, Arabidopsis, Brassica napus, and Solanum lycopersicon, parthenogenesis is induced when BBMLs are ectopically expressed in the egg cells (Conner et al. 2015; Conner et al. 2017; Khanday et al. 2019; Zhang et al. 2020a; Deng et al. 2022). Rice has 4 OsBBM homologs; BBM1,3,4 are exclusively expressed in sperm cells pre-fertilization (Khanday et al. 2019). Ectopic expression of these OsBBMs in rice egg cells induces parthenogenesis (Khanday et al. 2019). Depletion of the gamete-fusion-related DMP8/9 in Arabidopsis or orthologs in multiple plants leads to maternal haploid offspring (Zhong et al. 2019b; Zhong et al. 2020; Li et al. 2022c; Wang et al. 2022b; Zhang et al. 2022a, 2022b; Zhao et al. 2022b; Zhong et al. 2022b; Chen et al. 2023). Except for the fertilization defect, maternal haploid offspring has also been observed in ecs1ecs2 double mutants in Arabidopsis and the rice esc1 mutant (Yu et al. 2021; Jiang et al. 2022b; Jiang and Qu 2023; Zhang et al. 2023).

Combining MiMe and HI in fixation of hybrid rice

Based on the findings discussed above, several strategies have been developed to achieve synthetic apomixis in rice (Khanday et al. 2019; Wang et al. 2019). Using the MiMe mutants, rec8 pair1 osd1, either combining with osmtl or ectopic expression of OsBBM1 in the rice egg cell resulted in clonal hybrid rice seeds (Khanday et al. 2019; Wang et al. 2019). However, these strategies face challenges, such as low seed yields and clonal seed proportion. Employing a single vector for mutating PAIR1, REC8, and OSD1 along with ectopic OsBBM1 expression has subsequently increased the clonal seed proportion to 90% in a hybrid rice variety (Vernet et al. 2022). Another novel strategy combines BBM4 ectopic expression in the egg cell with MiMe, generating clonal hybrid plants with high seed setting, albeit the clonal seed proportion remained low (Wei et al. 2023). Recent advances are promising for further improvements (Box 2).

Concluding remarks

We hope the discussions here provide a perspective of important advances in plant reproduction culminated in the past 30 years or so through the efforts of a huge research community. We hope some of the more definitive conclusions will guide us to explore imminently approachable problems and tackle challenges that might be considered too grand to contemplate even in the past decade (Box 2). For the vast amount of accumulated knowledge that might require further definitions, we hope they will fuel productive discussions and entice new generations of biologists to join plant reproduction research.

Funding

Work from our laboratories was supported by National Natural Science Foundation of China (#32170343 to HL), CAS Project for Young Scientists in Basic Research (YSBR-078 to HL); National Natural Science Foundation of China (#32270360) and The Foundation of Hubei Hongshan Laboratory (#2021hszd010; #2022hszd017 to X-BP and PZ); The Science and Technology Department of Hubei Province (#2022CFA071 to PZ); National Natural Science Foundation of China (#32122014; #32070854 to SZ); Shandong Natural Science Foundation (ZR2020KC017 to Q.D.). U.S. National Science Foundation (MCB-1715764, IOS-1645858, IOS-2101467 to AYC and H-M Wu), National Institute of Food and Agriculture (NIFA)/USDA, the Center for Agriculture, Food, and the Environment under project number MAS00525 (to AYC), and National Institute of Health (R01GM147548-01A1).

Data availability

There is no new data in this paper.

Dive Curated Terms

The following phenotypic, genotypic, and functional terms are of significance to the work described in this paper:

• CNGC18 Gramene: AT5G14870

• CNGC18 Araport: AT5G14870

• SCC3 Gramene: AT2G47980

• SCC3 Araport: AT2G47980

• WOX2 Gramene: AT5G59340

• WOX2 Araport: AT5G59340

• ACA9 Gramene: AT3G21180

• ACA9 Araport: AT3G21180

• BRK1 Gramene: AT2G22640

• BRK1 Araport: AT2G22640

• CLE45 Gramene: AT1G69588

• CLE45 Araport: AT1G69588

• COBL11 Gramene: AT4G27110

• COBL11 Araport: AT4G27110

• GRL Gramene: AT2G35110

• GRL Araport: AT2G35110

• HAP2 Gramene: AT4G11720

• HAP2 Araport: AT4G11720

• MYB119 Gramene: AT5G58850

• MYB119 Araport: AT5G58850

• DUO1 Gramene: AT3G60460

• DUO1 Araport: AT3G60460

• MYB98 Gramene: AT4G18770

• MYB98 Araport: AT4G18770

• ECS1 Gramene: AT1G31450

• ECS1 Araport: AT1G31450

• TFIIB Gramene: AT2G41630

• TFIIB Araport: AT2G41630

• PLT2 Gramene: AT1G51190

• PLT2 Araport: AT1G51190

• RALF19 Gramene: AT2G33775

• RALF19 Araport: AT2G33775

• INP1 Gramene: AT4G22600

• INP1 Araport: AT4G22600

• OSD1 Gramene: AT3G57860

• OSD1 Araport: AT3G57860

• CDKA;1 Gramene: AT3G48750

• CDKA;1 Araport: AT3G48750