NtLLG4-mediated unconventional polar exocytosis of NtPPME1 coordinates cell wall rigidity and membrane dynamics to control pollen tube integrity

Abstract

Plant fertilization relies on controlled pollen tube growth that integrates membrane dynamics and cell wall expansion. We previously identified an unconventional exocytic pathway wherein Golgi-derived secretory vesicles (GDSVs) bypass the trans-Golgi network to deliver Nicotiana tabacum pectin methylesterase 1 (NtPPME1), thereby modulating cell wall rigidity. However, the mechanisms linking this pathway with membrane dynamics and signaling remain elusive. Here, we used cryo–focused ion beam–scanning electron microscopy and three-dimensional tomography to identified GDSVs as a distinct vesicle population at the pollen tube tip. We further demonstrated that tobacco LORELEI-like glycosylphosphatidylinositol-anchored protein 4 (NtLLG4), a key signaling molecule controlling membrane dynamics and integrity, functions as a receptor for NtPPME1, regulating its polar exocytosis via GDSVs to control cell wall stiffness. Furthermore, we identified trafficking signals that direct the unconventional exocytosis of NtPPME1 across intracellular organelles. Our findings reveal a crucial mechanism coupling cell wall rigidity with membrane signaling to control pollen tube growth and integrity during fertilization.

Cell wall rigidity couples with membrane signaling and dynamics to regulate pollen tube growth and integrity during fertilization.

INTRODUCTION

In angiosperms, sexual reproduction depends on the directional growth of pollen tubes to deliver two immotile male gametes to the female embryo sac for double fertilization (1, 2). Maintaining pollen tube integrity is critical during its rapid and polarized growth from the stigma surface toward the female gametophyte, ensuring that the pollen tube bursts at the appropriate time to release sperm cells upon arrival (1–3). An intricate signaling network including small peptides, receptor-like kinases (RLKs), actin cytoskeleton, Ca²⁺, pH, and reactive oxygen species (ROS) in the pollen tube tip has been shown to maintain pollen tube integrity (4–11). In addition, the rigidity of pollen tube cell wall must be spatiotemporally modulated to accommodate with membrane dynamics, allowing for active apical membrane expansion while maintaining the cylindrical shape of the pollen tube (12–16). However, the mechanisms coordinating membrane dynamics with cell wall rigidity still remain largely unknown. Apical vesicles, which carry newly synthesized proteins, lipid components, cell wall materials, and extracellular signaling molecules, accumulate in the tip region of growing pollen tubes to support and shape its growth (9, 17–19). Understanding the nature, dynamic behaviors, and functions of these apical vesicles is of great importance to uncover the mechanisms underlying pollen tube growth and integrity.

Identification and studying the nature of apical vesicles in the growing pollen tube tip is challenging due to their remarkably small size and vigorous behavior, which surpass the resolution limits of conventional light microscopy including confocal microscopy (18, 20, 21). Moreover, the lack of specific molecular markers for distinct types of vesicles in the tip region further complicates their differentiation (18, 19, 22). Previous studies using conventional transmission electron microscopy (TEM) have demonstrated that the apical vesicles enriched in the pollen tube tip are morphologically similar with a diameter of ~120 nm (23–29). However, Wang et al., have identified a distinct exocytic pathway in which Golgi-derived secretory vesicles (GDSVs) bypass the trans-Golgi network (TGN) to polarize the secretion of Nicotiana tabacum pectin methylesterase 1 (NtPPME1) at the apical surface of the growing pollen tube, regulating cell wall rigidity (19). This finding suggests that parallel exocytic pathway(s) may coexist and operate alongside the conventional TGN-derived exocytic pathway, facilitating the polar secretion of various cargoes to the pollen tube tip. Actually, GDSVs, which are noncoated vesicles derived from the Golgi apparatus with a diameter of ~40 nm, have been previously characterized on the basis of their morphological features (30). Although NtPPME1 is known to be exocytosed via GDSVs into the apical apoplast to regulate pectic cell wall rigidity, two critical questions remain unanswered: (i) Why have GDSVs not been observed and identified in the pollen tube tip region by conventional TEM imaging, whereas only 120-nm morphologically similar tip vesicles have been observed and identified (23–29)? (ii) Do NtPPME1-containing GDSVs actually traffic through the tip region of growing pollen tubes to reach the apical apoplast? This possibility remains speculative as NtPPME1–green fluorescent protein (GFP) fluorescent puncta are not predominantly localized in the tip region but are instead enriched in distal regions further from the tip, as consistently demonstrated by several previous studies (17, 31). Emerging imaging technologies may provide opportunities to elucidate the populations and dynamics of the pollen tube tip vesicles. Cryo–focused ion beam–scanning electron microscopy (cryo-FIB-SEM) combined with three-dimensional (3D) tomography has emerged as a powerful tool for the visualization of ultrafine intracellular structures in three dimensions. This technique achieves high-resolution visualization by generating a series of continuous sections and SEM images while preserving intracellular structures in their near-native state through rapid freezing. It minimizes the potential risk of morphological alterations in endomembrane trafficking vesicles and organelles by obviating the need for subsequent sample substitution procedures with embedding resins (23, 32–36).

On the other hand, pollen tube integrity during fertilization has been shown to be regulated and sustained by the signaling peptides RALF4 and RALF19, which are derived from the pollen tube (8, 9). Upon the pollen tube’s arrival at the ovule, a related signaling peptide, RALF34, which secreted from female tissues, takes over and induces the rupture of the pollen tube (8, 37). The pollen tube perceives and response to these RALF peptide signals through an apical plasma membrane (PM)–localized receptor–co-receptor complex composed of the RLKs ANXUR1/2 (ANX1/2), Buddha’s Paper Seal 1/2 (BUPS1/2), and pollen tube-specific glycosylphosphatidylinositol-anchored proteins (GPI-APs) LORELEI-LIKE GPI-ANCHORED PROTEIN 2/3 (LLG2/3) (6, 8, 38). The polar localization of this protein complex at the pollen tube apex is mediated by transport and deposition through tip vesicles (6, 38). This membrane-associated RALF-ANX-BUPS-LLG signaling complex regulates downstream Ca²⁺ gradients and ROS levels to ensure pollen tube integrity and polar growth (6, 8, 9, 38, 39). Loss-of-function mutations or reduced expression of RALF4/19, LLG2/3, ANX1/2, or BUPS1/2 consistently result in early pollen tube burst before reaching the ovules (6, 8, 38–40). Meanwhile, the rigidity of the pollen tube cell wall must be spatially regulated to coordinate with membrane dynamics during growth. The apical cell wall must remain sufficiently plastic to allow rapid membrane expansion at the tip or facilitate pollen tube burst when releasing sperm cells, whereas the cell wall in the shank region should be stiff enough to withstand the turgor pressure and maintain the cylindrical shape of the pollen tube (12–15, 41, 42). However, it remains largely uncovered how the underlying mechanisms coordinate cell wall rigidity with membrane signaling to maintain pollen tube integrity.

In this study, we used cryo-FIB-SEM and 3D tomography to identify GDSVs as a distinct type of apical vesicle accumulating at the pollen tube tip. In addition, we systematically dissected the specific trafficking signals that guide NtPPME1 through the GDSV-mediated unconventional polar exocytosis. We further demonstrated that NtLLG4, a member of the RALF-ANX-BUPS-LLG signaling complex involved in PM dynamics and integrity, serves as an unconventional polar exocytic (UPE) receptor for NtPPME1, regulating its polar secretion via GDSVs to modulate cell wall rigidity at the pollen tube tip. Our findings reveal a vital mechanism that synchronizes cell wall rigidity with membrane signaling to preserve pollen tube integrity during fertilization, providing valuable insights into the biological functions of GDSV-mediated unconventional exocytosis in polar cell growth and morphogenesis during plant fertilization.

RESULTS

Cryo–electron tomography reveals GDSV-mediated exocytosis at the pollen tube tip

To elucidate the mechanisms underlying NtPPME1-mediated unconventional polar exocytosis, it is essential to first morphologically identify GDSVs within the pollen tube tip. The tip region is densely populated with numerous tiny vesicles, which have been shown in previous studies to be morphologically similar (18, 19, 23, 29, 43, 44). These vesicles are believed to be involved in the dynamic processes of exocytosis, endocytosis, and recycling, which occur concurrently at the pollen tube tip (18, 19, 45). However, it remains unclear how a single vesicle type can mediate these three distinct intracellular trafficking pathways or whether multiple types of apical vesicles coexist in the pollen tube tip.

To accurately locate and morphologically identify the vesicles in the pollen tube tip, we used cryo-FIB-SEM, an advanced approach distinct from conventional TEM sample preparation and imaging methods. This method allowed us to capture serial sections of the N. tobacum pollen tube tip and generates high-resolution 3D tomography. The general sample preparation and workflow of cryo-FIB-SEM and 3D tomography were illustrated in Fig. 1 (A to D). Mature pollen was germinated in vitro for 2 hours and subsequently vitrified on a TEM grid by plunge-freezing (Fig. 1A). The grid was then mounted into the FIB-SEM cryo-shuttle, and the vitrified pollen tubes were identified using SEM before FIB milling. A germinated pollen tube on the grid was selected (Fig. 1B), with the tip region specifically targeted for FIB sectioning (Fig. 1C). A series of continued sections, each 30 nm in thickness, was obtained from the pollen tube tip through FIB milling and subsequently imaged using SEM (Fig. 1D). Last, the ultrafine 3D structures within the tip region were reconstructed into a 3D tomography, as illustrated in Fig. 1D.

Fig. 1. Cryo-FIB-SEM identification of GDSVs in the tobacco pollen tube tip.

(A) Schematic illustration of the preparation of germinated pollen tubes on a gold TEM grid, with an enlarged view showing the tubes on the grid. (B) Simplified illustration of FIB milling and SEM imaging, with an enlarged view highlighting the focus on the germinated pollen tube tip. (C) Representative pollen tube tip after FIB milling. (D) Schematic demonstration of the consecutive 30-nm-thick sections of the pollen tube tip obtained by cryo-FIB-SEM. Subsequently, 3D tomography of pollen tube was generated through stacking of 32 images and computational modeling. (E) Representative SEM image of a pollen tube cryo-FIB section showing well-preserved organelles, including mitochondria, vacuoles, the ER, and two morphologically distinct apical vesicles named type I and type II. (F) Representative image from the pollen tube shank region containing well-preserved organelles. Three Golgi apparatus surrounded by GDSVs are enlarged, with the GDSVs indicated by white arrows. (G) SEM images from different cryo-FIB slices of the pollen tube tip. (H) 3D tomography of type I and type II apical vesicles within the pollen tube tip, reconstructed from a stack of 32 continuous images. (I) Statistical analysis of the diameters of type I and type II apical vesicles were performed using two-tailed paired Student’s t test. For type I vesicles, n = 88 independent replicates were analyzed, and for type I vesicles, n = 185 independent replicates were analyzed (error bars ± SD, ***P < 0.001). (J and K) Representative images and 3D tomography illustrating the dynamics of type I and type II vesicles including their proximity to, contact, and interaction with the PM. Arrows in [(E), (H), and (J)] indicate type I vesicles. Arrowheads in [(E), (H), and (K)] indicate type II vesicles. M, mitochondrion; V, vacuole; ER, endoplasmic reticulum; G, Golgi apparatus.

The intracellular structures in the pollen tube tip region, including organelles such as mitochondria, small vacuoles, and the endoplasmic reticulum (ER) in the subapical region, are well preserved and morphologically clear (Fig. 1E). Notably, we identified two distinct types of noncoated vesicles in the apical region of the pollen tube tip (Fig. 1E and movie S1). An additional representative replicate of cryo-FIB-SEM imaging is provided in movie S2. Among these, type I vesicles exhibit high electron density and have an average diameter of 40 nm (n = 88) (Fig. 1, E to H, and movies S1 and S2). The morphology of the type I vesicles is consistent with the GDSVs near the Golgi apparatus in the shank region (Fig. 1F), which have been identified as mediators of unconventional exocytosis of NtPPME1 bypassing the TGN (19). In contrast, type II vesicles are characterized by a relatively lower electron density and a larger average diameter of 130 nm (n = 185) (Fig. 1, E and G to I, and movies S1 and S2). The morphology of type II vesicles aligns with previous observations of apical vesicles in pollen tubes (23, 29, 43). To further investigate these vesicles, we generated 3D tomography of the apical vesicles that are densely accumulated at the pollen tube tip (Fig. 1, G and H), using serial continued sections of SEM images. This allowed us to visualize the spatial distribution and morphologies of both type I and type II vesicles (Fig. 1H). In addition, we observed that both types of apical vesicles are capable of contacting and interacting with the PM (Fig. 1, J and K, and movie S1). These findings demonstrate that GDSVs are localized in the growing pollen tube tip and have the capacity to interact with and fuse to the PM, supporting their role in NtPPME1-mediated unconventional polar exocytosis.

Domain-specific signals determine NtPPME1 unconventional polar exocytosis

To gain deeper insights into the intracellular trafficking signals that regulate the unconventional polar exocytic process of NtPPME1, we systematically examined the functions of its different domains through a series of truncations, deletions, and point mutations. Using InterPro (https://ebi.ac.uk/interpro/), we predicted the primary structure of NtPPME1, which includes a 23–amino acid signal peptide (SP), following followed by a pro-region, a random coil (RC) region, and a C-terminal pectin methylesterase (PME) domain as illustrated in Fig. 2A. NtPPME1 is classified as the type II PME subfamily, characterized by a unique pre-pro-protein structural configuration that encompasses both the PME inhibitor (PMEI) domain and the PME catalytic domain (46, 47). The pro-region of NtPPME1 has been demonstrated to act as a PMEI, which effectively inhibits the enzymatic activity of PME during its intracellular trafficking. Upon exocytosis of NtPPME1 into the apoplast, its enzymatic activity is activated through the cleavage at the RC region, which separates the pro-region from the PME region (48, 49). Moreover, to further explore the spatial organization of these domains, we used AlphaFold3 for structural modeling (Fig. 2B and fig. S1). The predicted structures of the pro-region and PME domain were determined with high confidence, as reflected by predicted local distance difference test (pLDDT) scores exceeding 85 (high confidence: 70 < pLDDT < 100) (fig. S1). The pro-region primarily consists of multiple α helices, whereas the PME domain is largely composed of β sheets (Fig. 2B). The conserved structural features revealed by these high-confidence models highlight the evolutionary importance of these domains for proper function, supporting the need for accurate structural predictions.

Fig. 2. Dissection of the domain-specific trafficking signals of NtPPME1 during polar unconventional exocytosis.

(A) Schematic illustration of the primary structure of NtPPME1 protein (left) and representative images of a tobacco growing pollen tube expressing NtPPME1-GFP (right). DIC, differential interference contrast. (B) Predicted 3D structure of NtPPME1 by AlphaFold3. (C) Schematic representation of the unconventional exocytosis of NtPPME1 in the growing pollen tube. (D) Schematic of truncated NtPPME1 constructs fused with GFP. (E) Representative images of tobacco growing pollen tubes coexpressing NtPPME1 (ΔPro-region)–GFP with (i) Man1-RFP or (ii) Gonst1-RFP. (F) Representative image of tobacco growing pollen tubes coexpressing NtPPME1 (ΔRC+PME)–GFP and RFP-HDEL. (G) Representative image of tobacco growing pollen tubes coexpressing NtPPME1 (ΔRC)–GFP and RFP-HDEL. (H) Representative images of tobacco growing pollen tubes coexpressing NtPPME1 (ΔPME)–GFP/RFP with (i) RFP-HDEL, (ii) Man1-RFP, (iii) Gonst1-RFP, or (iv) NtPPME1-GFP. Colocalization ratios were calculated using Pearson correlation coefficients or Spearman’s rank correlation coefficients with ImageJ. The r values are in the range of −1 to 1, with 0 indicating no discernible correlation and +1 or −1 indicating strong positive or negative correlations, respectively. The white dashed curved lines in [(E) to (H)] indicate the boundaries of the apical pollen tube tip. White arrows in [(A) and (H)] indicate the apical localization of NtPPME1-GFP at the growing pollen tube tip. SP, signal peptide; pro-region, pro-region domain; RC, random coil; PME, pectin methylesterase domain; N, N terminus; C, C terminus; ER, endoplasmic reticulum; GDSV, Golgi-derived secretory vesicles; TGN, trans-Golgi network. Scale bars in [(A), (E), (F), (G), and (H)], 15 μm.

To assess whether the pro-region, RC region, and PME domain of NtPPME1 contain trafficking signals mediating its unconventional polar exocytic pathway, we generated chimeric proteins by fusing truncated versions of NtPPME1 with GFP and then transiently expressed them in growing tobacco pollen tubes (Fig. 2, C to H, and movies S4 to S11). Deletion of the pro-region, RC region, or PME domain completely abolished the polar localization of NtPPME1 at the pollen tube tip compared to the full-length protein (Fig. 2, E to H). In addition, these deletions showed distinct intracellular localizations and distributions. To precisely determine the subcellular localizations of these truncated variants, we compared them with various organelle markers (Fig. 2, E to H, and movies S4 to S11). Our results found that NtPPME1 (ΔPro-region)–GFP colocalized with both Man1–red fluorescent protein (RFP) and Gonst1-RFP, which serve as markers for the cis-Golgi and trans-Golgi, respectively (Fig. 2E and movies S4 and S5). Deletion of the RC region and PME domain caused NtPPME1 (ΔRC+PME)–GFP to be retained in the ER, as evidenced by its colocalization with RFP-HDEL, an ER marker (Fig. 2F and movie S6). In addition, deletion of the RC region also caused NtPPME1 (ΔRC)–GFP to be retained in the ER (Fig. 2G and movie S7). Although deletion of the PME domain resulted in the loss of apical localization at the pollen tube tip, NtPPME1 (ΔPME)–GFP remained localized to GDSVs due to its colocalization with NtPPME1 intracellular puncta, whereas it remained separate from ER and Golgi markers (Fig. 2H and movies S8 to S11). This suggests that the PME domain may contain a polar exocytic signal essential for its targeting to the pollen tube tip. In addition, the subcellular localizations of NtPPME1 (ΔRC+PME)–GFP, NtPPME1 (ΔPME)–GFP, and NtPPME1 (ΔRC)–GFP suggest that the RC region exclusively contains the export signal necessary for directing the secretion of NtPPME1 from the ER to the Golgi apparatus. Similarly, the localizations of NtPPME1 (ΔPro-region)–GFP and NtPPME1 (ΔPME)–GFP indicate that the pro-region may harbor a Golgi export signal responsible for trafficking from the Golgi to GDSVs. Together, our results reveal that each domain of NtPPME1 contains distinct organelle trafficking signals that collectively drive its unconventional polar exocytosis, bypassing the TGN.

Identify ER-to-Golgi and Golgi export signals for NtPPME1 exocytosis

To further elucidate the specific intracellular trafficking signals that regulate the unconventional polar exocytosis of NtPPME1 in growing pollen tubes, we systematically generated various deletions and point mutations in the C terminus of the RC as the RC region is known to contain the ER-to-Golgi trafficking signal (Figs. 2, F to H, and 3A). Our data revealed that progressive deletions of the last 6 to 16 amino acids at the C terminus of the RC region resulted in increased colocalization with the ER [Fig. 3A (i) and (ii), fig. S2A, and movies S12 and S13]. This suggests that the last 16 amino acids (FVEASNRKLLQISNAK) within the RC region harbor the ER export signal essential for NtPPME1 trafficking to the Golgi. Notably, the 16 amino acids contain dihydrophobic (LL) and FV motifs, previously identified as ER export signals in mammals and yeast (50–52). Mutations in the FV_{218, 219} and LL_{225, 226} motifs, either individually or in combination, resulted in partial or complete colocalization with the ER marker [Fig. 3A (iii), fig. S2B, and movie S14]. These findings confirm that the FV_{218, 219} and LL_{225, 226} within the C-terminal region of the RC domain are crucial for mediating the trafficking of NtPPME1 from the ER to the Golgi apparatus (Fig. 3B).

Fig. 3. Identification of specific trafficking signals mediating NtPPME1 secretion from the ER to GDSV.

(A) Representative images of tobacco growing pollen tubes coexpressing the ER marker RFP-HDEL with (i) NtPPME1 (SP+Pro+RC)–GFP, (ii) SP+Pro+RC(Δ218-233)–GFP, and (iii) SP+Pro+RC(FVLL_AAAA)–GFP. (B) Schematic identification of the ER export determinant trafficking signal FVLL_{218, 219, 225, 226} in the RC region of NtPPME1 during unconventional polar exocytosis. (C) Schematic illustration of the structures of α helix II and the loop II region connecting it with α helix III in the Pro-region of NtPPME1. (D) Representative images of tobacco pollen tubes coexpressing the Golgi marker Man1-RFP with (i) NtPPME1 (Δα helix II + loop II)–GFP, (ii) NtPPME1 (Δα helix II)–GFP, or (iii) NtPPME1 (Δ92MNT94)–GFP. (E) Schematic identification of the Golgi export signal MNT_92–94 in the pro-region of NtPPME1 during unconventional polar exocytosis. Colocalization ratios were calculated using either Pearson correlation coefficients or as Spearman’s rank correlation coefficients with ImageJ. The r values are in the range of −1 to 1, with 0 indicating no discernible correlation and +1 or −1 indicating strong positive or negative correlations, respectively. The white dashed curved lines in [(A) and (D)] indicate the boundaries of the apical pollen tube tip. SP, signal peptide; pro-region, pro-region domain; RC, random coil; ER, endoplasmic reticulum; GDSV, Golgi-derived secretory vesicles; TGN, trans-Golgi network. Scale bars in [(A) and (D)], 15 μm.

Next, we investigated the Golgi export signal within the pro-region of NtPPME1. Given that truncation of the pro-region led to the retention of NtPPME1 (ΔPro-region)–GFP within the Golgi apparatus, this observation suggests that the pro-region contains the Golgi export signal (Fig. 2E). Using AlphaFold3, we predicted that the pro-region is composed of five α helices (I to V) interconnected by short loops (fig. S3A). To pinpoint the Golgi export signal, we generated GFP-tagged NtPPME1 variants with sequential deletions of each α helix and their adjacent loops (Δ45 to 71, Δ72 to 94, Δ95 to 138, Δ139 to 169, and Δ170 to 204) [Fig. 3D (i), fig. S3B, and movie S15]. Deletion of the second α helix and the loop II (72 to 94) resulted in Golgi retention, indicating that this region contains the Golgi export signal [Fig. 3D (i) and movie S15]. In contrast, deletion of the other α helices and their adjacent loops led to punctate distribution, separating from the Golgi apparatus (fig. S3B). Further truncations of α helix II (spanning amino acids 72 to 91) and the loop (MNT_92–94) linking α helices II and III [Fig. 3D (ii) and (iii) and movies S16 and S17] revealed that the MNT_92–94 loop functions as the Golgi export signal [Fig. 3D (iii) and movie S17]. Thus, we identified MNT_92–94 within the pro-region of NtPPME1 as the Golgi export signal responsible for mediating the exit of NtPPME1 from the Golgi apparatus (Fig. 3E).

Identification of key trafficking signals regulating NtPPME1 GDSV-apoplast exocytosis

To elucidate the signals that regulate the unconventional polar exocytosis of NtPPME1 from GDSVs to the apoplast, we generated a series of deletions at the C terminus of the PME domain (Fig. 4A) as the PME domain is believed to contain the GDSV-to-apoplast trafficking signal for NtPPME1 (Fig. 2, A and H). Deletion of the last four amino acids (MMKV) at the C terminus of the PME domain did not affect the apical localization of the NtPPME1 mutant at the pollen tube tip, which remained consistent with the full-length protein [Fig. 4A (i) and movie S18]. However, deletion of the last seven amino acids (EAGMMKV) completely abolished this polar localization [Fig. 4A (ii) and movie S19]. This suggests that the amino acids EAG_549–551 at the C terminus is essential for polar exocytosis to the apoplast. To further validate this finding, we generated GFP-tagged mutants with single, double, and triple amino acid changes in the EAG_549–551 motif [Fig. 4A (iii); fig. S4, A and B; and movie S20]. Both mutation and deletion of the EAG_549–551 motif disrupted the polar localization of NtPPME1 at the pollen tube tip [Fig. 4A (iii) and fig. S4C]. In contrast, single and double amino acid mutations had no effect (fig. S4, A and B). In conclusion, the EAG_549–551 motif in the C-terminal region of the PME domain is critical for mediating the trafficking of NtPPME1 from GDSVs to the apoplast (Fig. 4C).

Fig. 4. Identification of the exocytic determinant signals controlling NtPPME1 unconventional polar exocytosis from GDSVs to apoplast.

(A) Chimeric GFP fusions with progressive C-terminal deletions or key amino acids mutation of the PME region in NtPPME1 and representative images of tobacco pollen tubes expressing these deletion plasmids. (B) Chimeric GFP fusions with various mutations at the four N-glycosylation sites of NtPPME1, and representative images of pollen tubes expressing these mutated constructs. White arrows indicate the apical localization of expressed protein at the growing pollen tube tip. Measurements of the apical fluorescence signal intensity across the pollen tube tip region as indicated by the white dashed lines are normalized and plotted into graphs (n ≥ 3). (C) Schematic identification of the polar exocytic signal of EAG_549–551 and the N-glycosylation site N₄₈₂ in the C-terminal of the PME region of NtPPME1 controls the unconventional polar exocytosis. The white dashed curved lines in [(A) and (B)] indicate the boundaries of the apical pollen tube tip. SP, signal peptide; pro-region, pro-region domain; RC, random coil; PME, pectin methylesterase domain; PM, plasma membrane; ER, endoplasmic reticulum; GDSV, Golgi-derived secretory vesicles; TGN, trans-Golgi network; a.u., arbitrary units. Scale bars in [(A) and (B)], 15 μm.

In addition, glycosylation is a common posttranslational modification that influences protein localization and ensures proper folding (53, 54). To determine whether glycosylation plays a role in NtPPME1 trafficking, we used NetNGlyc (https://services.healthtech.dtu.dk/services/NetNGlyc-1.0/) to predict the N-linked glycosylation sites in NtPPME1 (fig. S5A). Four putative sites were identified at N68, N180, N301, and N482 (fig. S5A). We then generated GFP-tagged point mutations at these positions (N68Q, N180Q, N301Q, and N482Q) and revealed that only the N482Q mutation disrupted apical localization [Fig. 4B (i) and movie S21]. The other mutations, either alone or in combination, had no effect [Fig. 4B (ii), fig. S5B, and movie S22). Furthermore, the N482Q mutation caused NtPPME1 to be retained in GDSVs, as indicated by its colocalization with NtPPME1 intracellular puncta, whereas it remained separate from the Golgi apparatus and TGN markers (fig. S5C). These findings demonstrate that the N-glycosylation site at N482 is crucial for mediating NtPPME1 trafficking from GDSVs to the apoplast (Fig. 4C).

NtLLG4 acts as a receptor for NtPPME1 exocytosis

Because of the limited understanding of GDSV-mediated unconventional polar exocytosis involved in pollen tube polar growth and cell plate formation during cytokinesis as revealed in previous studies, NtPPME1 remains the only characterized cargo protein associated with this pathway (19, 31, 44). In addition, it is highly speculated that the specific intracellular sorting and packaging of soluble NtPPME1, which undergoes through the unconventional polar exocytosis pathway, is likely to be a receptor-mediated active process. Therefore, to test this hypothesis and investigate which receptor protein mediates the targeting of soluble cargo NtPPME1 to the pollen tube tip, we isolated total proteins from transgenic tobacco pollen tubes expressing NtPPME1-GFP and performed immunoprecipitation (IP) using GFP-Trap, followed by high-performance liquid chromatography–tandem mass spectrometry (HPLC-MS/MS) analysis. The identified proteins were classified into four functional categories on the basis of Eukaryotic Orthologous Groups (KOG) annotation, with the largest group associated with cellular processes and signaling (fig. S6A and table S1). Given that NtPPME1 is a secreted soluble cargo protein containing an SP, we focused our analysis on 165 potential receptor proteins that contain either an SP or a transmembrane domain. Among these, we identified a GPI-AP, LOC107817512, which functions as a chaperone and receptor in maintaining pollen tube integrity in Arabidopsis (fig. S6B) (6, 38). This receptor was designated as N. tabacum LORELEI-like GPI-AP 4 (NtLLG4).

To further investigate the GPI-AP family in N. tabacum, we used the NtLLG4 sequence as a query in BLAST searches, identifying nine members of the NtLLG family (fig. S7A). Comparative analysis with Arabidopsis LLGs (AtLLGs) revealed that three of the nine NtLLGs belong to the AtLORELEI and AtLLG1 clade, which are known to regulate root growth in Arabidopsis (38, 55). The remaining six NtLLGs cluster within the AtLLG2/3 clade, which is involved in regulating pollen tube integrity (fig. S7A) (6, 38, 55). Sequence alignment confirmed a high degree of conservation among the NtLLG family proteins (fig. S7B). Furthermore, gene expression profiling using the National Center for Biotechnology Information (NCBI) database demonstrated that NtLLG4 and NtLLG5 are highly expressed in N. tabacum flowers across various developmental stages, including young, mature, and senescent stages (fig. S7C), consistent with the gene expression patterns of AtLLG2/3 in Arabidopsis (38). These nine NtLLGs were subsequently tested using luciferase complementation assay (LCA) to assess their interactions with NtPPME1. NtLLG1, 2, and 3 could not be amplified from tobacco flowers due to their low expression levels. Our results showed that NtPPME1 can be recognized by several NtLLG family members including NtLLG4, 5, 6, 7, 8, and 9, with the strongest interaction specifically with NtLLG4 (fig. S8).

To further elucidate the spatial interaction between NtPPME1 and NtLLG4, we performed protein structural modeling of both proteins using AlphaFold3 (Fig. 5, A to C). The structural models were predicted with high confidence, as evidenced by a predicted template modeling (pTM) score exceeding 0.9 (pTM > 0.8 represents confident high-quality predictions) (fig. S9). Notably, the protruding surface region of NtLLG4 (highlighted with the black dashed circle) interacts with the concave region of NtPPME1 (highlighted with the white dashed circle) (Fig. 5C). This concave region of NtPPME1 is formed by integration of the pro-region (dark blue) and the PME domain (magenta), suggesting that NtLLG4 interacts with both domains of NtPPME1 (Fig. 5C). Furthermore, colocalization analysis revealed that the punctate signals of NtPPME1-GFP predominantly colocalized with NtLLG4-RFP (Fig. 5D). To explore the interaction in more detail, we conducted a coimmunoprecipitation (co-IP) assay using anti-GFP and anti-Myc antibodies by transiently coexpressing NtPPME1-GFP and NtLLG4-Myc in Nicotiana benthamiana leaves. It confirmed that NtPPME1 interacts with NtLLG4 (Fig. 5E). Moreover, fluorescence resonance energy transfer (FRET) analysis and LCA further validated this interaction in tobacco BY-2 protoplasts and in plants (Fig. 5, F and G). Besides, a yeast two-hybrid (Y2H) assay also confirmed that NtPPME1 interacts with NtLLG4 in the absence of their N-terminal SP (Fig. 5H).

Fig. 5. Identification of NtLLG4 acting as a receptor for NtPPME1 during unconventional polar exocytosis.

(A and B) Structural and surface predictions of the NtLLG4-NtPPME1 protein-protein interaction using AlphaFold3. (C) The protein-protein interaction interfaces between NtLLG4 and NtPPME1, as predicted by AlphaFold3, are highlighted by black and white dashed circles, respectively. (D) Representative images of tobacco BY-2 protoplast coexpressing NtPPME1-GFP and NtLLG4-RFP. Scale bar, 10 μm. Colocalization ratios are calculated using either Pearson correlation coefficients or as Spearman’s rank correlation coefficients with ImageJ. The generated r values are in the range of −1 to 1, with 0 indicating no discernible correlation and +1 or −1 indicating strong positive or negative correlations, respectively. (E) Co-IP analysis demonstrating the interaction between NtPPME1 and NtLLG4. (F) FRET analysis of the protein interaction between NtPPME1 and NtLLG4 in BY-2 protoplasts. (G) LCA to assess the interaction between NtPPME1 and NtLLG4. The N. benthamiana leaves were divided into four sections and infiltrated with different combinations of expression plasmids as indicated. (H) Y2H analysis exploring the roles of NtPPME1 in the interaction between NtPPME1 and NtLLG4. SP, signal peptide; pro-region, pro-region domain; RC, random coil; PME, pectin methylesterase domain N, N terminus; C, C terminus.

To dissect the interaction sites, we generated NtPPME1 truncations targeting either the pro-region or the PME of NtPPME1. Co-IP assays, Y2H analysis, and LCA all demonstrated that NtLLG4 interacts with both the pro-region and PME domain of NtPPME1 (Fig. 6, A to C). These findings suggest that NtLLG4 acts as a specific UPE receptor that binds NtPPME1 during its unconventional polar exocytosis before its release into the apoplast of growing pollen tubes. To further investigate the specific interaction sites between NtPPME1 and NtLLG4, we conducted protein-protein interaction surface analyses of the NtPPME1-NtLLG4 complex using PDBePISA (https://ebi.ac.uk/pdbe/pisa/) (Fig. 6D). The analysis revealed that NtLLG4 establishes electrostatic interactions with the surface of NtPPME1 (Fig. 6D). Specifically, the R41 residue of NtLLG4 fits into an acidic pocket on NtPPME1, engaging residues E165, Q342, Q364, D365, Y368, R454, and N482, which distributed across the pro-region and PME domains [Fig. 6D (i) and (ii)]. In addition, R41 of NtLLG4 forms hydrogen bonds with D365 of NtPPME1 [Fig. 6D (ii)]. Furthermore, N51 of NtLLG4 forms a hydrogen bond with E173 of NtPPME1 via its nitrogen and oxygen atoms, respectively [Fig. 6D (iii) and (iv)]. Mutational analysis using Y2H and LCA showed that mutations in NtPPME1 (EEQQDYRN_AAAAAAAQ) disrupted the interaction between NtPPME1 and NtLLG4 (Fig. 6, E and F). These results confirm that the interaction of NtLLG4 to both the pro-region and PME domain of NtPPME1 is primarily mediated through a network of hydrogen bonds and electrostatic interactions.

Fig. 6. Reveal the structural interaction sites between NtPPME1 and NtLLG4.

(A) Co-IP analysis investigating the roles of the pro-region and PME domains in the interaction between NtPPME1 and NtLLG4. (B) Y2H analysis exploring the roles of the pro-region and PME domains in the interaction between NtPPME1 and NtLLG4. (C) LCA to assess the roles of the pro-region and PME domains in the interaction between NtPPME1 and NtLLG4. (D) Predicted specific protein binding sites between NtPPME1 and NtLLG4. NtPPME1 amino acids are highlighted in orange, whereas NtLLG4 amino acids are shown in blue. The positively and negatively charged surfaces of NtPPME1 are represented in blue and red, respectively. (E) Y2H analysis examining the interaction between the NtPPME1 mutant (mutations at E165, E173, Q342, Q364, D365, Y368, R454, and N482_AAAAAAAQ) and NtLLG4. (F) LCA assessing the interaction between the NtPPME1 mutant (mutations at E165, E173, Q342, Q364, D365, Y368, R454, and N482_AAAAAAAQ) and NtLLG4. SP, signal peptide; pro-region, pro-region domain; RC, random coil; PME, pectin methylesterase domain; N, N terminus; C, C terminus.

In addition, to determine whether NtLLG4 forms protein complexes with NtANXs and NtBUPSs similar to AtLLG2/3 in Arabidopsis, we first conducted BLAST searches using AtANX1 and AtBUPS1 sequences as queries. We identified four NtANX and three NtBUPS homologs in N. tabacum (fig. S10, A and B). To further elucidate potential interactions among NtLLG4, NtANXs, and NtBUPSs, we performed protein structure modeling and interaction using AlphaFold3 (fig. S10, C to E). Although the pTM scores were below 0.5, the resulting models supported the interactions between AtLLG2/3 and the ectodomains of AtANX1/2 and AtBUPS1/2, consistent with previous Y2H and pull-down assays (6). Likewise, our modeling predicted that NtLLG4 could interact with the ectodomains of NtANX1–4 and NtBUPS1–3 in tobacco (fig. S10E). These findings suggest that NtLLG4 is likely to form protein complexes with multiple NtANX and NtBUPS family members, mirroring the interaction patterns previously reported in Arabidopsis.

NtLLG4 regulates cell wall rigidity for pollen tube integrity

Given the interaction between NtLLG4 and NtPPME1, we next sought to determine the biological functions of NtLLG4 in determining the polar localization of NtPPME1 at the pollen tube tip, and its effects on pollen tube germination and growth. Pollen tubes overexpressing NtLLG4 exhibited markedly reduced lengths compared to wild-type (WT) controls without any other discernible morphological alterations (fig. S11A). Coexpression of NtPPME1-GFP and NtLLG4-RFP in tobacco pollen tubes revealed that both proteins colocalized as punctate structures (Fig. 7A and movie S23). Nevertheless, it is noteworthy that the fluorescence signal of NtPPME1-GFP was absent at the pollen tube apex when NtLLG4 was overexpressed (Fig. 7, A and B). To further explore the function of NtLLG in pollen tube growth, we used RNA interference (RNAi) to silence the NtLLG family, avoiding potentially functional redundancy among family members. Upon knockdown of the NtLLG gene family, the tip localization of NtPPME1-RFP was also abolished comparing with NtLLG scrambled siRNA as the negative control (Fig. 7C, fig. S11B, and movie S24). Either the overexpression of NtLLG4 or NtLLG-RNAi can result in the disruption of apical localization of NtPPME1 at the pollen tube tip. It indicates that an accurate amount of NtLLG receptors is required to facilitate the unconventional polar exocytosis of NtPPME1. It is further supported by previous functional studies of LLG protein in maintaining pollen tube integrity and controlling cell polarity, in which the LLG level in growing pollen tubes needs to be precisely regulated (6, 38). In addition, ~40% of pollen tubes coexpressing NtPPME1-RFP and NtLLG-RNAi exhibited abnormally multiple pollen germination sites and tube growth (Fig. 7, D and E). It indicates critical roles of NtLLG in guarding a single site for pollen germination and proper tube growth. Furthermore, to assess whether the altered localization of NtPPME1 in NtLLG-RNAi pollen tubes is specific, or if other co-receptors are similarly affected, we examined the localization of NtANX3 and NtBUPS1 upon coexpression with NtLLG-RNAi. NtANX3 and NtBUPS1 were selected on the basis of their close evolutionary relationship with Arabidopsis AtANX1/2 and AtBUPS1/2, which have been functionally implicated in regulating pollen tube polarity and cellular integrity (fig. S10A) (8, 38). Moreover, both NtANX3 and NtBUPS1 exhibit high enrichment in pollens. Compared to the scrambled control, NtLLG-RNAi noticeably impaired the polarized PM localization of both NtANX3 and NtBUPS1 at the pollen tube apex (fig. S11, C and D). These results indicate that the apical polarization of ANX and BUPS receptor kinases is dependent on LLG function.

Fig. 7. Coordination between NtPPME1 and NtLLG4 in regulating pollen tube cell wall rigidity.

(A) Representative images of pollen tubes coexpressing NtPPME1-GFP and NtLLG4-RFP. Scale bar, 10 μm. (B) The intensity of NtPPME1-GFP and NtLLG4-RFP at the apical tip was measured along the white dashed line indicated in (A) and plotted on a graph (n = 3). (C) Representative images of pollen tubes coexpressing NtLLG-RNAi and NtPPME1-RFP. The intensity of NtPPME1-RFP and NtLLG-RNAi at the apical tip was measured along the white dashed line (left) and plotted on a graph (n = 3). Scale bar, 10 μm. (D) Representative images of coexpressing NtLLG4-RNAi and NtPPME1-RFP caused the germination and growth of poly-pollen tubes from a single pollen grain. Scale bars, 25 μm (top) and 8 μm (bottom). (E) Statistical analysis of the percentage of different phenotypes observed in (D). The results represent the means ± SEs (n ≥ 125). (F) Immunofluorescence labeling of demethylesterified pectin distribution in various types of pollen tubes. Three different lines (a, b, and c) were drawn across the shank, subapical, and tip regions of the pollen tube, respectively, for signal measurement and analysis. Scale bar, 10 μm. (G to I) Measurement of JIM5 fluorescence in the shank, subapical, and tip regions of the pollen tube, as indicated by the dashed lines (a, b, and c) in (F), and plotted on a graph (n ≥ 4), respectively. (J) Immunolabeling of methylesterified pectin distribution in various types pollen tubes. Scale bar, 10 μm. (K) Measurement of JIM7 fluorescence across the pollen tube tip region, as indicated by the white dashed line in (J), and plotted on a graph (n ≥ 5). The white dashed curved lines in [(A) and (C)] indicate the boundaries of the apical pollen tube tip.

To further reveal the functional coordination between NtLLG4 and NtPPME1 in pollen tube growth, we modulated the expression levels of NtPPME1 and/or NtLLG4 and analyzed the resulting changes in cell wall rigidity using immunofluorescence labeling with JIM5 and JIM7 antibodies. The pollen tube tip cell wall is primarily composed of highly methylesterified pectin, recognized by the JIM7 antibody, which provides the flexibility necessary for the rapid expansion of the pollen tube membrane (28, 56, 57). As the pollen tube elongates, the esterified pectin is progressively demethylesterified by PME, recognized by the JIM5 antibody, leading to a stiffer cell wall that supports the cylindrical structure of the tube (12, 44). Compared to WT pollen tubes, those expressing either NtPPME1-GFP or NtLLG4–yellow fluorescent protein (YFP) showed increased cell wall stiffness in the subapical region as indicated by an extended JIM5 signal (Fig. 7, F to H). In contrast, the JIM5 signal in NtLLG-RNAi pollen tubes remains confined to the shank region, similar to the pattern observed in WT pollen tubes (Fig. 7, F to I). The fluorescence signal intensity of JIM5 was quantitatively assessed across the shank and subapical regions of pollen tubes and plotted on a graph (n ≥ 4) (Fig. 7, G and H). Moreover, in pollen tubes coexpressing NtPPME1–cyan fluorescent protein (CFP) and NtLLG4-YFP, JIM5 fluorescence extended into both the subapical and apical regions, indicating increased rigidity throughout the tip (Fig. 7, F and I). JIM5 fluorescence was quantitatively measured across the apical region of pollen tubes, and plotted on a graph (n = 5) (Fig. 7I). Consistent with the JIM5 labeling results, the JIM7 fluorescence signal at the pollen tube apex was reduced in tubes expressing either NtPPME1-GFP or NtLLG4-YFP and was completely absent in tubes coexpressing both proteins (Fig. 7, J and K). It suggests that coexpression of NtPPME1 and NtLLG4 enhances pectin demethylesterification at the pollen tube tip, contributing to increased cell wall rigidity. Conversely, a greatly increased JIM7 fluorescence signal was observed at the apex of NtLLG-RNAi pollen tubes (Fig. 7, J and K), indicating that NtLLG4 is essential for the proper regulation of pectin methylesterification. Quantitative analysis of JIM7 fluorescence in the apical region was plotted (n ≥ 5) (Fig. 7K). These findings indicate that NtLLG4 acts as a UPE receptor for NtPPME1, collaboratively regulating its polar exocytosis and targeting at the pollen tube apex. Their interaction modulates the rigidity of the pollen tube cell wall, allowing for the spatiotemporal coordination with membrane signaling required to maintain pollen tube integrity during plant fertilization.

Collectively, we propose a schematic model to illustrate the cooperative mechanism of cell wall rigidity and membrane signaling in controlling pollen tube integrity during plant fertilization (Fig. 8). Each domain of NtPPME1 harbors specific trafficking signals that mediate its unconventional polar exocytosis within the growing pollen tube. NtLLG4 serves as a UPE receptor to interact with NtPPME1, ensuring its targeted exocytic secretion to the apoplast. Upon release into the apoplast, NtPPME1 undergoes proteolytic cleavage, separating the pro-region from the PME domain, which then catalyzes pectin demethylesterification to modulate cell wall rigidity. This process ensures the proper morphology and structural integrity of the pollen tube. Furthermore, NtLLG4 is a crucial component of the RALF-LLG-ANX-BUPS signaling complex on the PM, which regulates pollen tube integrity and timely rupture (6, 38). Thus, our study uncovers a coordinated mechanism between cell wall rigidity and membrane signaling that governs pollen tube integrity during plant fertilization.

Fig. 8. Schematic hypothetical model of the regulatory mechanism underlying NtPPME1 unconventional polar exocytosis and its role in maintaining pollen tube integrity.

This schematic illustrates a proposed model for the regulatory mechanisms by which NtPPME1 mediated the unconventional polar exocytosis and contributes to the maintenance of pollen tube integrity through cell wall rigidity. Each domain of NtPPME1 contains specific organelle trafficking signals that enable its polar exocytosis in the growing pollen tube. During this process, NtLLG4 interacts with NtPPME1, directing its targeted trafficking to the pollen tube apex. Once released into the apoplast, NtPPME1 undergoes proteolytic cleavage, separating the PME domain from the pro-region. The PME domain plays a critical role in regulating cell wall rigidity, which is essential for maintaining the integrity of the growing pollen tube. Concurrently, NtLLG4 functions as a key component of the RALF-LLG-ANX-BUPS complex, which acts as a membrane signaling sensor, controlling membrane dynamics and integrity. Together, the NtPPME1-NtLLG4 interaction integrates membrane signaling with cell wall rigidity to maintain pollen tube integrity.

DISCUSSION

In angiosperms, sexual reproduction relies on the structural integrity of the pollen tube, which serves as a tunnel to deliver two sperm cells and undergoes timely rupture to enable double fertilization (1, 2). The regulation of cell wall rigidity in pollen tubes is critical as it must be precisely coordinated with membrane dynamics to maintain both the structural integrity and allow the necessary rupture during fertilization (1, 2). Previous studies have well established the key role of the membrane-associated RALF-ANX-BUPS-LLG protein complex and its downstream signaling pathways in governing pollen tube polarity and integrity (6, 8, 38). However, the exact mechanisms by which the cell wall cooperates with membrane dynamics to control pollen tube integrity remain largely uncharacterized.

We demonstrated that the pollen-specific protein NtLLG4 functions as a UPE receptor, recognizing NtPPME1 and mediating its unconventional polar exocytosis into the apical apoplast of growing pollen tubes (Figs. 5 to 7). This interaction is critical for maintaining the polar localization of NtPPME1, which, in turn, regulates cell wall rigidity, underlying pollen tube growth and integrity. The role of GPI-APs in polar secretion and protein localization has been well established in various cellular systems, suggesting a conserved mechanism across biological contexts. During pollen tube growth, GPI-APs such as AtLLG2 and AtLLG3 have been identified as chaperones and co-receptors, facilitating the polar secretion of ANX/BUPS to the PM (38). This polar targeting is essential for maintaining membrane dynamics and ensuring pollen tube integrity. Similarly, COBL11, another GPI-AP, interacts with the RALF4-ANX-BUPS protein complex to promote the proper localization and polar distribution of its components, further contributing to the stabilization of the pollen tube structure (40). These findings underscore the pivotal role of GPI-AP–mediated protein trafficking and localization in regulating pollen tube growth and integrity.

The function of GPI-APs extends beyond pollen tubes as these proteins are critical for protein localization in other plant tissues (58, 59). In synergid cells, LORELEI, which is another member of GPI-APs, acts as a chaperone for FERONIA (FER), a homolog of ANX/BUPS, facilitating its trafficking to the PM (58). Likewise, in vegetative tissues, LLG1 mediates the transport of FER from the ER to the PM, reinforcing the conserved function of GPI-APs in ensuring the proper localization of essentially functional proteins (58, 59). Moreover, GPI-AP–mediated protein trafficking is not unique to plants, whereas similar pathways have been identified in yeast and mammals (60), indicating a broadly conserved mechanism across species. Despite these essential roles, receptor proteins that facilitate the unconventional exocytosis of cargo proteins remain unidentified in animal and yeast systems. The identification of the UPE receptor protein NtLLG4, which mediates the unconventional polar exocytosis of NtPPME1, provides insights for the investigation and identification of analogous receptor proteins in mammals and yeast, broadening our understanding of this unconventional secretion pathway. Notably, previous research has shown that JIM7-labeled esterified pectin accumulates prominently at the apex of growing pollen tubes in LLG2/3-RNAi lines and cobl11 mutants (38, 40). This observation suggests that GPI-APs play an essential role in modulating the composition of the pollen tube tip cell wall, particularly in relation to pectin distribution and rigidity. Given the importance of pectin in maintaining cell wall rigidity and integrity, these findings imply that GPI-APs can regulate membrane dynamics; meanwhile, they also contribute to the construction and remodeling of the cell wall during pollen tube growth.

Expanding on these insights, we further reveal that NtLLG4 regulates the polar exocytosis of NtPPME1, a key enzyme involved in regulating the rigidity of pectin during cell wall formation, and contributes to the regulation of the stiffness of pectic cell wall at the pollen tube tip (Fig. 7). Moreover, an optimal level of NtLLG4 is crucial for the accurate targeting and polar exocytosis of NtPPME1 to the pollen tube tip apoplast (Fig. 7). This observation aligns with previous studies demonstrating that the expression level of LLG proteins is a key determinant of pollen tube tip polarity and integrity (6, 38). This regulatory mechanism is crucial for governing pollen tube polar growth and cell integrity during fertilization as it coordinates the interplay between cell wall rigidity and membrane dynamics (Fig. 8). Thus, NtLLG4 emerges as a central molecule that integrates signaling and structural processes required for proper pollen tube growth (38, 40, 58, 59). NtLLG4 is involved in the RALF-ANX-BUPS signaling cascade, mediating downstream ROS and Ca²⁺ signaling to sustain membrane integrity (39). On the other hand, our findings suggest that it acts as a critical receptor for NtPPME1, facilitating its polar exocytosis and modulating cell wall rigidity. This dual function of NtLLG4 in both signaling and cell wall construction underscores its key role in ensuring proper pollen tube polar growth and fertilization.

Tip-focused rapid pollen tube growth is driven by the accumulation of numerous small apical vesicles, including those involved in exocytic secretion, endocytosis, and recycling, all concentrated in the tip region (18, 45, 61, 62). Morphological studies using TEM have consistently identified a predominantly homogeneous population of noncoated vesicles with ~120 nm in diameter within the pollen tube tip (23, 29, 43). A key question that arises is why these various types of apical vesicles, originating from distinct endomembrane trafficking pathways, exhibit such notably similar morphologies. We previously identified that the secretion and tip targeting of NtPPME1 occurs through an unconventional polar exocytosis pathway, in which GDSVs directly mediate the polar exocytosis of NtPPME1 bypassing the TGN in growing pollen tubes (19). It is noteworthy that GDSVs exhibit an approximate diameter of 40 nm, as observed in ultrathin-structure images from TEM and 3D tomography of the Golgi apparatus (30, 63). Therefore, it is highly speculated that GDSVs, despite their smaller size and distinct function, may exhibit unique morphological characteristics that set them apart from other apical vesicles, reflecting their specialized role in mediating the unconventional exocytosis of NtPPME1.

Distinct from conventional TEM sample preparation and imaging approaches, we used cryo-FIB-SEM imaging combined with 3D tomography and identified two distinct populations of vesicles with contrasting morphologies at the pollen tube tip (Fig. 1). Cryo-FIB-SEM is a powerful technique that allows for the observation of cellular structures in their near-native state by direct sample ultrathin section and imaging after quick freezing. It greatly preserves their natural morphology and minimizing potential artifacts and morphological changes typically introduced by chemical fixation or embedding resin infiltration (32–35). Unlike conventional TEM, which often provides 2D images, cryo-FIB-SEM offers 3D high-resolution visualization, allowing for more detailed spatially morphological analysis of complex cellular structures (33, 64). The technique achieves continuous sequential sectioning at resolutions as high as 10 nm, facilitating 3D ultrastructural analysis of spatial distribution and organization of organelle and membrane-bound structure morphology across a wide range of biological specimens (65). Cryo-FIB-SEM has proven particularly valuable for analyzing vesicle morphology at the pollen tube tip, as demonstrated in our study (Fig. 1). It overcomes the technical challenges of preparing pollen tubes for ultrathin sectioning, where the random spatial distribution of tubes complicates precise sectioning of the tip region. The cryo-FIB-SEM allows for targeted sample preparation, enabling precise identification of the pollen tube tip and the acquisition of continuous serial sections through this critical region (Fig. 1). We used advanced cryo-FIB-SEM imaging and 3D tomography to identify and characterize type II vesicles within the pollen tube tip. These vesicles were found to be ubiquitously distributed and exhibited low electron density and an average diameter of 130 nm (Fig. 1). Our results align with previous studies using conventional TEM, which reported average vesicle diameters of 102, 115, and 128 nm in Pyrus pyrifolia, Papaver rhoeas, and Arabidopsis thaliana, respectively (23–25).

Our study uncovered a previously undocumented subpopulation of vesicles, designated as type I vesicles (Fig. 1) and distinguished by their smaller size and high electron density. These vesicles exhibit a remarkable resemblance to GDSVs localized near the Golgi apparatus in the subapical region of pollen tubes (19, 30, 44). The higher electron density of type I than type II vesicles might stem from their selective enrichment with macromolecule comprising NtPPME1 family proteins and apoplast/cell wall–associated proteins, which remain to be further investigated. The rapid sample freezing and afterward immediate sectioning and imaging by using cryo-FIB-SEM without extra steps of sample substitution and resin embedding minimize the potential risk of morphological changes of endomembrane trafficking vesicles and organelles when compared with conventional chemical fixation or advanced high-pressure freezing and freeze substitution methods (23, 32, 35, 36, 66). In addition, cryo-FIB-SEM coupled with 3D tomography generates comprehensive 3D volumes of the pollen tube tip, allowing for the clear observation of different vesicle types and their dynamic behaviors (Fig. 1). Notably, the subcellular localization and distribution pattern of NtPPME1-GFP differs from those of other proteins associated with tip trafficking vesicles, such as SCAMPs and RabA4d (18, 21, 67, 68). Unlike these proteins, NtPPME1-GFP fluorescence signals do not prominently accumulate in the tip region of growing pollen tubes (Fig. 2A). This observation, corroborated by multiple previous independent studies (17, 19, 31, 44), is likely attributable to two possible reasons: (i) The polar exocytosis of NtPPME1 is an exceptionally efficient process characterized by a low incidence of unsuccessful secretion events or recycling within the tip region. Consequently, fluorescence signals from GDSVs carrying NtPPME1-GFP do not accumulate appreciably at the pollen tube tip. This hypothesis is supported by our result demonstrating that disruption of NtPPME1 exocytic secretion markedly increases the fluorescence intensity of a mutated NtPPME1-GFP variant at the pollen tube tip (Fig. 2H). (ii) The polar exocytosis of NtPPME1 and its subsequent release into apoplast, where it plays a critical role in regulating apical cell wall rigidity, occurs as a superfast and transient process. This dynamic behavior is distinct from the trafficking pathways of apical PM-localized or PM-associated proteins (18, 21, 67, 68). Cryo-FIB-SEM provides us an opportunity to gain unprecedented insights into the populations of apical vesicles in the pollen tube tip and greatly enhancing our understanding of the mechanisms underlying pollen tube tip growth and vesicular trafficking.

Polar exocytosis and targeting of proteins, lipids, and cell wall components to the pollen tube tip are essential for supporting rapid tube expansion and growth (8, 19, 69). In contrast to the conventional exocytic pathway, where proteins are transported from the ER through the Golgi and TGN to the PM, NtPPME1 has been found within GDSVs that bypass the TGN. However, the specific protein trafficking determinants that mediate this unconventional polar exocytic pathway remain largely unknown (19). Our study revealed that truncation of either the pro-region or the PME domain of NtPPME1 abolishes its apical localization at the pollen tube tip. Notably, the subcellular localization patterns of the truncated NtPPME1 lacking the pro-region or PME domain differed markedly, suggesting that distinct domains of NtPPME1 may contain specific trafficking signals that guide its trafficking through various intracellular organelles to mediate its unconventional polar exocytosis (Fig. 2). This observation is consistent with a previous study (31). Furthermore, our systematic analysis identified specific trafficking signals to facilitate the export of NtPPME1 from the ER and Golgi apparatus (Fig. 3). These trafficking determinants are conserved with those previously identified in secretory proteins across eukaryotic cells, suggesting that NtPPME1 follows a conserved protein transport mechanism for secretion through the ER and Golgi (70, 71).

In addition, we identified two unique trafficking signals that regulate the unconventional polar transport of NtPPME1 from GDSVs to the pollen tube tip (Fig. 4). In particular, our findings reveal that N-glycosylation plays a crucial role in the unconventional exocytosis of NtPPME1. N-glycosylation is well known for its critical role in the proper localization and function of membrane proteins in the conventional secretory pathway, such as metabotropic glutamate receptors and ion channels (53, 54, 72, 73). Our findings expand this role into unconventional trafficking, showing that N-glycosylation of NtPPME1 facilitates its accurate targeting and secretion via GDSVs, bypassing the TGN (Fig. 4 and fig. S5C). This adds a dimension to our understanding of glycosylation in plant cells, linking it to unconventional polar exocytosis and revealing a broader regulatory mechanism than previously revealed (54, 74). Moreover, our finding contrasts with earlier studies focused primarily on conventional secretion pathways, where glycosylation is key for sorting proteins at the TGN (54, 75). For instance, glycosylation is crucial for the proper sorting of vacuolar sorting receptors within the TGN in Arabidopsis cells (54). However, our results suggest that N-glycosylation plays an equally pivotal role in the function and exocytosis of soluble proteins NtPPME1 that bypass the TGN. This specific role of N-glycosylation highlights its versatility in plant cellular processes and provides a deeper understanding of glycan-mediated protein trafficking.

Together, our study reveals a critical mechanism by which GDSV-mediated unconventional polar exocytosis of NtPPME1 coordinates cell wall rigidity with membrane signaling to maintain pollen tube integrity during fertilization. Moreover, we demonstrated that different populations of apical vesicles exist to meet the distinct demands of pollen tube tip growth. These findings advanced our understanding of unconventional polar exocytosis in plant fertilization and provide insights into the molecular mechanisms governing polar cell growth and morphogenesis. In addition, elucidating the specific trafficking signals involved in NtPPME1 unconventional exocytosis lays the groundwork for future applications in biotechnology, such as optimizing protein production systems.

MATERIALS AND METHODS

Plant materials and growth conditions

Transgenic tobacco plants expressing NtPPME1-GFP were generated by transforming N. tabacum with the construct UBQ10_pro:NtPPME1-GFP. The primers used for construction of the plasmid are detailed in table S2. N. tabacum plants were grown in a greenhouse at 28°C under a light cycle of 12-hour light and 12-hour darkness as described before (76). N. benthamiana seeds were directly sowed in the soil (Jiffy, EN12580) for germination. Three weeks later, each juvenile plant was transferred to an individual flowerpot and grown in a plant cultivate chamber at 28°C, 65% humidity, under a light cycle of 16-hour light and 8-hour darkness with light intensity of 12,000 lux.

Cryo-FIB-SEM and 3D tomography

Pollen grains were germinated in vitro for 2 hours and vitrified on Au grids (Quantifoil, 200-mesh, gold R2/2) plunge-freezing with Leica EM GP1. Before loading the sample, the cryo-stage and cryo anticontaminator were precooled to −170° and −190°C, respectively. Subsequently, the specimen grid was mounted into the FIB-SEM (ZEISS Crossbeam 550) cryo-shuttle in the loading station of the Quorum PP3010Z cryo-transfer system. This shuttle was then loaded into the prechamber of Quorum PP3010Z for sputter coating (5 mA, 60 s) and transferred to the cryo-stage of the ZEISS Crossbeam 550 FIB-SEM. Following transfer, the grid was deposited (30°C, 15 s) with a platinum precursor using the gas injection system to reduce radiation damage from FIB-SEM. After deposition, pollen tubes on the grid were screened and serial sections were acquired using FIB-SEM. The parameters of SEM imaging were 3.0-kV accelerate voltage, 50-pA electron beam current, and a secondary electron detector, whereas 30-kV accelerate voltage and 700-pA ion beam current were used to execute FIB milling, and the slice thickness was 30 nm. For model generation, the contours were drawn manually and meshed with the 3dmod program in the IMOD software package (https://bio3d.colorado.edu/imod/doc/guide.html#ParallelProc).

Transient expression and confocal imaging

Transient expression with growing pollen tubes and tobacco BY-2 protoplasts were carried out essentially as described previously (77, 78). Confocal observation and image collection were performed as previously described. The images from pollen tubes or protoplasts expressing fluorescent-tagged proteins were collected with a laser level of ≤3% to ensure that the fluorescence signal was within the linear range of detection (typically 1 to 2% laser power was used). Time-lapse images of growing pollen tubes were collected with minimal time intervals. The primers used for constructing recombinant plasmids were listed in table S2.

Bioinformatics analysis

Protein sequences were from the NCBI (https://ncbi.nlm.nih.gov/). Protein evolutionary conservation analysis was performed using InterPro (https://ebi.ac.uk/interpro/). The N-linked glycosylation sites of NtPPME1 were predicted using NetNGlyc (https://services.healthtech.dtu.dk/services/NetNGlyc-1.0/). The phylogenetic tree analysis of AtLLGs and NtLLGs were constructed using MEGA version 7.0. Protein conservation of NtLLGs was analyzed with DNAMAN version 9.0. The 3D structures of NtPPME1 and NtLLG4 were predicted using AlphaFold3 (https://golgi.sandbox.google.com/) and visualized with PyMOL version 2.5 (https://pymol.org/). Prediction results were assessed on the basis of the predicted template modeling (pTM) and the interface predicted template modeling (ipTM) values, with the top-ranked model selected for further analysis. Protein-protein interaction interfaces between NtPPME1 and NtLLG4 were predicted using PDBePISA (https://ebi.ac.uk/pdbe/pisa/).

Transcriptome data analysis

The N. tabacum genome (Ntab-TN90) was downloaded from the NCBI (https://ncbi.nlm.nih.gov/datasets/genome/GCF_000715135.1/) and used as the genome assembly. RNA sequencing (RNA-seq) datasets of N. tabacum (accession numbers: SRR955761, SRR955762, SRR955763, SRR1199069, SRR1199070, SRR1199071, SRR1199072, SRR1199073, SRR1199074, SRR1199121, SRR1199122, SRR1199123, SRR1199124, SRR1199125, SRR1199127, SRR1199128, SRR1199129, SRR1199130, SRR1199132, SRR1199135, SRR1199197, SRR1199198, SRR1199199, SRR1199200, SRR1199202, and SRR1199203) were also obtained from the NCBI (https://ncbi.nlm.nih.gov/sra). The RNA-seq reads were aligned to the Ntab-TN90 reference genome. Gene expression levels were normalized and calculated as fragments per kilobase of transcript per million mapped reads (FPKM) values.

Transient RNAi plasmid construction

To construct the NtLLG-RNAi plasmid under the control of the UBQ promoter, a 258–base pair (bp) conserved sequence from the NtLLG4 gene, representing the conserved region of the NtLLG family, was amplified in two distinct fragments. To generate the NtLLG scrambled siRNA plasmid as a negative control, a 19-bp scrambled sequence derived from the conserved sequence of NtLLG genes was amplified in two complementary fragments. These fragments were subsequently cloned into the hairpin RNAi vector pHANNIBAL to form a hairpin RNA structure. The resulting RNAi construct, containing the UBQ promoter and an octopine synthase terminator, was subcloned into the pBI221-UBQ10_pro:GFP-NOS vector to facilitate transient expression.

Y2H assay

Y2H assays were conducted using the MatchMaker GAL4 Two-Hybrid System 3 (Clontech, 630489) in accordance with the manufacturer’s instructions. The pGBKT7 and pGADT7 vectors were used to carry target cDNAs and then transformed into the Saccharomyces cerevisiae strain AH109. After selection on the SD/-Leu/-Trp medium, single transformant colonies were screened for growth on the SD/-Leu/-Trp/-His/-Ade medium to determine protein interactions. The primers used are listed in table S2.

Luciferase complementation assay

The coding sequence of target genes were amplified and inserted into the pCAMBIA-35S-nLUC and pCAMBIA-35S-cLUC vectors, respectively. The plasmids were transformed into Agrobacterium tumefaciens GV3101. Transformed bacteria were cultured, harvested, and resuspended in a buffer [10 mM MgCl₂, 200 μM acetosyringone, and 10 mM MES (pH 5.7)] to a final concentration of OD₆₀₀ (optical density at 600 nm) = 0.8 to 1.0. The suspension was infiltrated into N. benthamiana leaves using a syringe. The tobacco plants were placed in darkness for 24 hours and then incubated in a 16-hour/8-hour light/dark cycle for 24 hours. Thereafter, the leaves were sprayed with a 1 mM luciferin solution and then kept in darkness for 5 min to quench the fluorescence. A deep cooling charge-coupled device imaging apparatus (Tanon, 5200) was used to capture fluorescence images. The primers used for the plasmid construct are listed in table S2.

FRET assay

For the FRET assay, various constructs, including UBQ10_pro:NtPPME1-CFP, UBQ10_pro:NtLLG4-YFP, UBQ10_pro:NtLLG4-YFP, UBQ10_pro:CFP, UBQ10_pro:YFP, and UBQ10_pro:CFP-YFP were transiently cotransformed into tobacco BY-2 protoplasts. Following 10-hour incubation, FRET analysis was performed using a Leica TCS SP8 confocal system, according to the manufacturer’s instructions, with excitation at 405 nm and 514 nm. The CFP-YFP fusion construct, which includes a short linker between CFP and YFP, served as the positive control, whereas the coexpression of separate CFP and YFP proteins was used as the negative control. Protoplasts expressing the various CFP and YFP fusion proteins were subjected to photobleaching using a 514-nm laser at full power intensity. FRET efficiency was calculated using the formula FRET_eff = (D_post − D_pre)/D_post, where D_post and D_pre represent the CFP intensities before and after acceptor bleaching, respectively.

Protein extraction and HPLC-MS/MS analysis

Germinated genetic transgenic pollen tubes expressing UBQ10_pro:NtPPME1-GFP were collected from ~200 fresh flowers. After 3-hour pollen germination, pollen tubes were frozen in liquid nitrogen and then resuspended in 2 ml of extraction buffer [50 mM tris-HCl (pH 7.5), 150 mM NaCl, 1 mM MgCl₂, 20% glycerol, 0.2% NP-40, and 1× protease inhibitor cocktail]. The samples were centrifuged at 13,000 rpm for 30 min, and the supernatant was collected. For IP, 50 μl of the supernatant was reserved as the input, and the rest was incubated with anti-GFP magnetic beads (Chromotek, gtma) for 2 hours at 4°C. The beads were washed twice with extraction buffer and sent for HPLC-MS/MS analysis.

Co-IP and Western blot analysis

The full-length cDNA of NtPPME1 with its various truncations was tagged with GFP, whereas the full-length cDNA of NtLLG4 was tagged with Myc. These chimeric fusion genes were cloned into pCAMBIA expression vectors and transiently coexpressed in N. benthamiana leaves. For protein extraction, 2 g of frozen leaf tissue was ground in liquid nitrogen and homogenized in 2 ml of IP buffer containing 50 mM tris-HCl (pH 7.4), 150 mM NaCl, 1 mM EDTA, 0.5% NP-40 (v/v; Sigma-Aldrich, 127087-87-0), 10% glycerol (v/v), and 1× protease inhibitor cocktail (Roche, 5892970001). The lysates were then incubated with GFP-Trap magnetic agarose beads (ChromoTek, gtma-100) at 4°C for 4 hours in a top-to-end rotator. Samples were subsequently analyzed by SDS–polyacrylamide gel electrophoresis and Western blot using anti-GFP (Invitrogen, A11122) and anti-Myc (Sangon Biotech, D110006) antibodies. Chemiluminescence was detected with an image analyzer (Tanon, 5200). The primers used for the plasmid construct are listed in table S2.

Immunofluorescence labeling

Fixation of N. tabacum pollen tubes, antibody labeling, and subsequent immunofluorescence analysis were conducted as previously described (21, 79). Briefly, 2-hour germinated pollen tubes, following bombardment, were fixed in a solution containing 3.7% paraformaldehyde, 50 mM Na-phosphate buffer (pH 7.0), 5 mM EGTA, and 0.02% azide. The samples were then incubated overnight at 4°C with JIM5/JIM7 antibodies [1:200 (v/v); Abmart, ZW00005/ZW00007] in blocking buffer 2 (0.25% bovine serum albumin, 0.25% gelatin, 0.05% NP-40, and 0.02% azide in phosphate-buffered saline). After washing with blocking buffer 2, the samples were incubated for 1 hour with anti-rat IgG antibodies [1:5000 (v/v); Life Technologies, A-11077). After additional washes, samples were imaged using a Leica TCS SP8 confocal microscope.

Image processing and analysis

Colocalization analysis between two fluorescent proteins was performed using Fiji software (https://fiji.sc/) with the Pearson-Spearman correlation colocalization plug-in as described previously (80). Results were presented either as Pearson correlation coefficients or as Spearman’s rank correlation coefficients, both of which produce r values in the range of −1 to 1, where 0 indicates no discernible correlation and +1 or −1 indicates strong positive or negative correlations, respectively. Signal intensity analysis was measured using the plot profile plug-in of Fiji.

Statistical analysis

Statistical analysis was performed as described in each figure legend. All experiments were performed at least in triplicate (N ≥ 3; exact values indicated in figure legends), with all data points displayed along with the means ± SD. Raw data and statistical analysis for all graphs are presented in table S3.

Accession numbers

The locus identifiers for the genes mentioned in this article are as follows: NtPPME1 (LOC107768376), NtLLG1 (LOC107768055), NtLLG2 (LOC107769788), NtLLG3 (LOC107800316), NtLLG4 (LOC107817512), NtLLG5 (LOC107783567), NtLLG6 (LOC107783568), NtLLG7 (LOC107797956), NtLLG8 (LOC107762859), NtLLG9 (LOC107814930), AtLORELEI (AT4G26466), AtLLG1 (AT5G56170), AtLLG2 (AT2G20700), AtLLG3 (AT4G28280), NtANX1 (LOC107789667), NtAwNX2 (LOC107830137), NtANX3 (LOC107799099), NtANX4 (LOC107811148), AtANX1 (AT3G04690), AtANX2 (AT5G28680), NtBUPS1 (LOC107822438), NtBUPS2 (LOC107783235), NtBUPS3 (LOC107785095), AtBUPS1 (AT4G39110), and AtBUPS2 (AT2G21480).

Supplementary Materials

The PDF file includes:

Figs. S1 to S11

Legends for tables S1 and S3

Table S2

Legends for movies S1 to S24

Other Supplementary Material for this manuscript includes the following:

Tables S1 and S3

Movies S1 to S24