Research ArticlePLANT SCIENCES

Pollen tube contents initiate ovule enlargement and enhance seed coat development without fertilization

See allHide authors and affiliations

Science Advances  28 Oct 2016:
Vol. 2, no. 10, e1600554
DOI: 10.1126/sciadv.1600554


In angiosperms, pollen tubes carry two sperm cells toward the egg and central cells to complete double fertilization. In animals, not only sperm but also seminal plasma is required for proper fertilization. However, little is known regarding the function of pollen tube content (PTC), which is analogous to seminal plasma. We report that the PTC plays a vital role in the prefertilization state and causes an enlargement of ovules without fertilization. We termed this phenomenon as pollen tube–dependent ovule enlargement morphology and placed it between pollen tube guidance and double fertilization. Additionally, PTC increases endosperm nuclei without fertilization when combined with autonomous endosperm mutants. This finding could be applied in agriculture, particularly in enhancing seed formation without fertilization in important crops.

  • Plant sciences
  • plant reproduction
  • pollen tube contents
  • cell biology
  • agriculture


In animals, numerous sperm cells swim toward the egg, and complete fertilization is supported by seminal plasma (1, 2). For example, the SVS2 (seminal vesicle secretion 2) protein localized only in the seminal plasma protects sperms from uterine attack and is required for sperm capacitation in mice (3). In higher plants, the pollen tube content (PTC) also carries two sperm cells toward the embryo sac, where fertilization takes place (4). However, except for being a sperm cell carrier, nothing is known regarding PTC. Here, we identified the PTC functions that are significant for seed development initiation, including endosperm proliferation without fertilization.


As shown in animal reproduction (3), we first considered the function of the PTC, analogous to the seminal plasma. To investigate how PTC affects male-female interaction in plants, we analyzed gene expression profiles in ovules (seed primordia) in bulk. We performed transcriptome analysis on wild-type (WT) ovules without pollination or crossed with WT or gcs1/gcs1 pollen, which are defective in fertilization of sperm cells, resulting in a PTC donor without fertilization (fig. S1 and table S1) (5, 6). As Nagahara et al. (6) reported previously, gcs1/gcs1 sperm cells rarely fertilize egg or central cells; however, the ratio for this has been unknown. Therefore, we confirmed the fertilization ratio by crossing WT ovules and gcs1/gcs1 pollen. As a result, 91.1 ± 3.2% (mean ± SD; n = 11 pistils) of gcs1/gcs1 sperm cells failed fertilization, 4.0 ± 3.2% single-fertilized the egg cell, 4.2 ± 3.7% single-fertilized the central cell, and 0.7 ± 1.5% fertilized both the egg and central cells. Thus, using gcs1/gcs1, we could block fertilization genetically and show the effect triggered by PTC only. First, we confirmed that the expression of embryo and endosperm markers, WOX9 (7) and AGL62 (8), was observed in WT pollination at 12 hours after pollination (HAP), soon after fertilization, but not in gcs1/gcs1 pollination because there was no fertilization. Second, using genome-wide analysis, at 12 HAP, we identified 24 up-regulated genes from each data set of the ovules crossed with WT or gcs1/gcs1 pollen, as compared with the data set of the ovules without pollination. Although most ovules were fertilized at this time in WT pollination, but not in gcs1/gcs1 pollination, 21 of the 24 identified genes overlapped in these sample fractions (Fig. 1A and table S2), suggesting that only PTC release, but not fertilization, accounted for the gene up-regulation. At 24 HAP or later, which is the stage accompanied by embryogenesis, these expression profiles showed increasing differences (Fig. 1A and table S2). The transcriptional similarity of early events at 12 and 24 HAP after crossing with WT and gcs1/gcs1 pollen was also indicated by clustering analysis (Fig. 1B). Furthermore, some of the genes showed similar expression patterns of up-regulation at early times after PTC release in ovules pollinated with WT or gcs1/gcs1 pollen (Fig. 1, C and D, and tables S3 and S4). To avoid the possibility of contamination through unexpected fertilization by gcs1/gcs1 pollen, we performed another transcriptome analysis using ovules that retained two sperm cells from gcs1/gcs1 pollen, indicating that these ovules failed to fertilize (5, 6). As shown in table S5, the second transcriptome profile shared genes with the first profile, suggesting that the first and second transcriptome profiles were not due to fertilization events. Many of these were identified as the genes that are expressed in typical spatial manners in fertilized ovules (table S6). All these data suggest that the PTC release itself triggers typical gene expression soon after pollination. Notably, multiple genes associated with cell expansion, cell division, or seed coat development were significantly expressed even by gcs1/gcs1 pollination at 24 and 48 HAP (Fig. 1E and table S7). Hence, we hypothesized that PTC could affect ovule shape without fertilization through gene induction.

Fig. 1 Transcriptome analysis for ovules with or without fertilization.

(A) Up-regulated genes in ovules crossed with WT (WT) and gcs1/gcs1 (gcs1) pollen compared with those in ovules without pollination (NPT) at 12, 24, and 48 HAP. Number of up-regulated genes for each sample fraction and that of common up-regulated genes in WT (red circle; WT > NPT) and gcs1 (green circle; gcs1 > NPT) are indicated. Numbers of overlapping up-regulated genes are indicated in orange. (B) Cluster analysis of the tested sample fractions. Bar indicates the height of branches. gcs1_12HAP and WT_12HAP are highlighted in pink, and gcs1_24HAP and WT_24HAP are highlighted in blue, showing the similarity between clusters. (C) Early-response genes triggered by PTC release were screened in the two classes indicated and pooled for further expression analyses in (D) (see the Supplementary Materials for details). (D) Temporal expression of the identified genes in NPT, WT, and gcs1. Expression peaks were at 12 HAP in WT and gcs1. (E) Temporal expression of the genes for cell expansion (expansin), cell division (cyclin), and seed coat development in NPT, WT, and gcs1.

To investigate this, we first observed ovule size by crossing WT ovules and hap2-1/+ (9) (allelic to gcs1) pollen. We observed three classes of ovules of different sizes in the same WT silique 3 days after pollination (DAP) (Fig. 2, A and B). The largest ovules (66.5 ± 6.4%, mean ± SD; n = 10 pistils) corresponded to fertilized ovules that received pollen tube(s); the smallest (5.1 ± 1.8%) were virgin ovules that were not inserted with a pollen tube. The remaining unfertilized ovules, previously shown to have the hap2-1 pollen tube insertion(s) (28.4 ± 5.2%) (10), were smaller than fertilized ovules but obviously larger than virgin ones. To investigate whether ovule enlargement resulted from cell expansion, we observed ovule integument cells using a plasma membrane marker line (Fig. 2, C to E, and fig. S2) (11). Cells were expanded when an ovule accepted a WT pollen tube (Fig. 2C), but not in a virgin ovule (Fig. 2D). However, when ovules accepted a gcs1/gcs1 pollen tube, the cells were still expanded (Fig. 2E), indicating that the ovule enlargement resulted from cell expansion. To determine whether cell division in ovule integument contributes to the enlargement, we counted the divided cells after crossing M-phase marker pistils (12) and WT or gcs1/gcs1 pollen (Fig. 2, F to I, and fig. S3). The ovule cell division ratio was high with WT pollination (Fig. 2F) but low without pollination (Fig. 2G). However, the cell division ratio with gcs1/gcs1 pollination was higher (Fig. 2H) than in virgin ovules (Fig. 2G), indicating that ovules underwent cell division without fertilization. As reported previously (13), ovules at 0 DAP had mitotic activity (Fig. 2I). However, at 1 to 3 DAP, the division rates of nonpollinated ovule cells decreased steadily. Ovules with WT pollination showed mitotic activity until 3 DAP to produce seeds. Ovules with gcs1/gcs1 pollination showed mitotic activity until 1 DAP, suggesting that ovules continued cell division as though they had undergone fertilization.

Fig. 2 Discovery of POEM.

(A and B) POEM after crossing WT pistils (♀) with hap2-1/+ pollen (♂). (a) Largest fertilized ovule with a pollen tube (pt). (b) Smallest ovule without a pollen tube. (c) Intermediate ovule with a pollen tube. (C to E) Ovule integument cells observed at 2 DAP using an RPS5Apro::tdTomato-LTI6b marker line. (C) Largest cells (WT) after crossing +/+ pollen, indicating that all the ovules were fertilized. (D) Small cells without pollen tube (Pt–). (E) Intermediate cells with gcs1/gcs1 pollen tube(s) (gcs1), indicating that almost all the ovules were unfertilized. (F to H) Yellow fluorescent protein (YFP) spots using a CycB1;2pro:CycB1;2::NLS-YFP cell division marker line after crossing with WT (F), no pollen (G), and gcs1/gcs1 (H). (I) Number of spots (±SD) observed: 0 DAP, 11.2 ± 1.7 (all); 1 DAP: 25.9 ± 2.9 (W, wild type), 12.5 ± 2.2 (g, gcs1), and 5.3 ± 2.3 (N, no pollen); 2 DAP: 10.1 ± 2.7 (W), 0.9 ± 0.9 (g), and 3.1 ± 1.7 (N); 3 DAP: 5.4 ± 1.7 (W), 0.1 ± 0.2 (g), and 1.1 ± 1.2 (N). (J to L) Ovules at 3 DAP, stained by vanillin in WT (J); unstained virgin ovules (K); partially stained ovules, which were fertilized with gcs1/gcs1 pollen (L). v, vanillin-stained zone. Scale bars, 100 μm (A and J to L); 60 μm (C to E); 40 μm (F to H).

Because cell expansion and division are initiated without fertilization, we speculated that PTC could initiate seed coat development (13). To investigate this, we stained ovules with vanillin to detect proanthocyanidin, a seed coat marker (14). The largest ovule was stained, but the smallest virgin ovule was not (Fig. 2, J and K). Intermediate ovules with gcs1/gcs1 pollination were partly stained without fertilization (Fig. 2L and fig. S4). These three phenotypes—cell expansion, cell division, and seed coat formation without fertilization—are completely reflected in our transcriptome data, indicating that we have identified a new plant phenomenon from transcriptome analysis. We call this phenomenon pollen tube–dependent ovule enlargement morphology (POEM) because enlargement is observed only when an ovule has accepted one or two pollen tubes.

To confirm that PTC is the trigger of POEM, we first compared the ratio of PTC release with that of POEM. We analyzed lre/lre mutants (1517), which are defective in synergid cells, by blocking pollen tube bursts. The lre/lre mutants sometimes burst PTCs, but the ratio for this was not reported (16, 17). To understand the PTC release ratio, we crossed PTC marker pollen (18) with lre/lre pistils; 49.5 ± 10.8% (n = 12 pistils) of pollen tubes did not release PTCs, whereas 50.5 ± 10.8% of pollen tubes did (Fig. 3, A and B). To assess the POEM ratio, we crossed WT or gcs1/gcs1 pollen with lre/lre mutant pistils; 53.1 ± 11.8% (n = 13 pistils) of ovules showed POEM, but 46.9 ± 11.8% of ovules did not (Fig. 3, C to E), suggesting that the ratios of POEM and PTC release are synchronized. Another possible POEM trigger is synergid cell content because this includes important intercellular substances, such as LURE peptides (19) or calcium ions (20). We previously showed that a pollen tube entering the embryo sac ruptures one of the two synergid cells (10). If synergid cell content triggers POEM, then synergid cell breakdown should facilitate ovule enlargement. To break synergid cells using pollen tubes, we crossed WT or gcs1/gcs1 pollen and lre/lre pistils with FGR8.0 (a synergid cell nucleus marker) (21) background. When we crossed gcs1/gcs1 pollen in 97.1 ± 2.6% (n = 10 pistils) of ovules, at least one synergid cell was lost during pollen tube attack (Fig. 3, F to H), suggesting that almost all ovules have synergid cell content inside the embryo sac. These results strongly suggest that the POEM ratio corresponds to the PTC release ratio but not to the synergid cell disruption ratio (Fig. 3I). Thus, we concluded that the PTC is the trigger for POEM (Fig. 3, J and K).

Fig. 3 PTC triggers POEM.

(A and B) PTC release ratio at 1 DAP following crossing with lre/lre pistil and LAT52pro::GFP pollen (PTC marker): (A) 49.5 ± 10.8% (n = 12) of ovules without and (B) 50.5 ± 10.8% of ovules with PTC release. (C to E) Confocal microscopy images after aniline blue staining and the POEM ratio in each ovule at 2 DAP following crossing lre/lre ovules with gcs1/gcs1 pollen. (C) No pollen tube control. (D) Ovule with pollen tube(s) but no POEM (46.9 ± 11.8%; n = 13). (E) Ovule with pollen tube(s) and POEM (53.1 ± 11.8%; n = 13). (F to H) Survival ratio of synergid cell(s) at 2 DAP following crossing with lre/lre pistil with FGR8.0 background and gcs1/gcs1 pollen. (F) Both synergid cells (84.6%) broken. (G) One synergid cell survived (12.5%). (H) Both synergid cells survived (3.0%) in each ovule. (I) Bar graph indicating ratios with pollen tube burst (50.5%), POEM (gcs1/gcs1; 53.1%), POEM (WT; 49.1%), and synergid cell (SY) death (97.1%). (J and K) Schematic diagrams of POEM. Pollen tube entry into an ovule is not sufficient to trigger POEM (J), but PTC release triggers POEM (K). Scale bars, 50 μm (A to H).

After confirming that PTC induces POEM, we further investigated the PTC functions inside ovules, including egg and central cells. We observed egg and central cells after gcs1/gcs1 pollination (Fig. 4). For WT, some ovules had increased numbers of central cell nuclei, but not of endosperm nuclei because of lack of AGL62 expression (fig. S5) without fertilization (Fig. 4, A to D). To investigate whether mea/mea mutants, which develop endosperm nuclei without fertilization (22, 23), could increase autonomous endosperm proportion, we crossed mea/mea ovules with gcs1/gcs1 pollen. mea/mea ovules markedly increased the endosperm amount without fertilization at 2 and 3 DAP (Fig. 4, E to H). In mea/mea, 19.0% of ovules at 2 DAP and 45.7% of ovules at 3 DAP increased the endosperm nuclei number (two or more nuclei) without fertilization (Fig. 4I). We confirmed that the vanillin staining results correspond to increased numbers of endosperm nuclei at 3 DAP (Fig. 4, J and K) after crossing WT or mea/mea ovules with gcs1/gcs1 pollen. After mea/mea pistils were crossed with gcs1/gcs1 pollen, not only all ovules but also the pistils were enlarged (Fig. 4L). Overall, we conclude that PTC can trigger gene induction, resulting in ovule enlargement and seed coat development without fertilization, and can markedly increase endosperm nuclei number without fertilization in the previously identified mutants, which are capable of autonomous endosperm formation (Fig. 4M).


In plants, seed formation without fertilization is called apomixis and is valuable for agriculture because the important genetic traits can be easily fixed in apomictic crops, which then propagate without interference from unfavorable environmental conditions (24). POEM could contribute to creating apomictic crops when combined with autonomous endosperm mutants because PTC can produce enlarged ovules and seed coat initiation without fertilization, although it cannot produce autonomous endosperm or embryo by itself.

Fig. 4 PTCs increase the number of central cell/autonomous endosperm nuclei.

(A to D) WT (C24) virgin ovules with a central cell at 3 DAE (days after emasculation) (100%) (A) and 4 DAE (97.7%) (B). gcs1/gcs1-pollinated ovules with a central cell nucleus (92.2%) (C) and with two nuclei (6.8%) (D) at 2 DAP. (E to H) A central cell nucleus (96.4%) (E) and two nuclei (3.6%) (F) in a mea/mea ovule at 3 DAE. gcs1/gcs1-pollinated mea/mea ovule with two nuclei (16.4%) (G) and four nuclei (6.3%) (H) at 2 DAP. (I) Ratios of the number of central cell/endosperm nuclei. Pollen tube insertion (P+) increased the nuclei ratios (≥2) for no insertion (P−) from 0 to 7.8% (WT, 2 DAP), 2.3 to 12.8% (WT, 3 DAP), 3.6 to 22.6% (mea/mea, 2 DAP), and 6.0 to 51.7% (mea/mea, 3 DAP). Colors indicate number of nuclei. (J and K) Vanillin staining of gcs1/gcs1-pollinated WT (J) and mea/mea (K) pistils at 3 DAP. (J) Few ovules stained. (K) About 50% stained (right) but not without pollination (left). (L) Nonpollinated (left) and gcs1/gcs1-pollinated (right) mea/mea pistils at 3 DAP. (M) PTC induces ovule enlargement, seed coat development, and central cell/endosperm proliferation. Scale bars, 50 μm (A to H); 1 mm (J to L).

In 1910, Bataillon (25, 26) showed that in frogs, physical stimuli could induce segmentation without fertilization. Although not by physical stimuli alone, PTC promotes endosperm division without fertilization, suggesting a function parallel to that in animals wherein germ cells may divide independently of fertilization owing to external stimuli.

We have identified a new plant phenomenon, POEM, by transcriptome analysis and have revealed that POEM is a new step in plant reproduction that occurs between pollen tube guidance and double fertilization. We also note that transcriptome analysis is a powerful tool not only for identifying important genes but also for discovering new biological phenomena. PTC plays a crucial role in enlarging ovules, initiating seed coat formation, and increasing autonomous endosperm without fertilization when combined with autonomous endosperm mutants and could therefore be applied to apomixis in major crops. Further studies on POEM might not only lead to agricultural improvements but also reveal points common to plant and animal reproduction.


Plant materials and growth conditions

Arabidopsis thaliana ecotypes Columbia (Col-0) and C24 were used as WT plants. Testcross experiments were conducted in gcs1/gcs1 (5, 6), hap2-1/+ (9), lre/lre (1517), and mea/mea (22, 23) and in WT plants. The RPS5Apro::tdTomato-LTI6b (11) and CycB1;2pro:CycB1;2::NLS-YFP (12) lines were used for POEM assays. Seeds were sterilized with 5% sodium hypochlorite containing 0.5% Triton X-100 and germinated on plates containing 0.5× Murashige and Skoog salts (pH 5.7) (Wako Pure Chemical), 2% sucrose, Gamborg’s B5 vitamin solution (Sigma), and 0.3% Gelrite (Wako Pure Chemical) in a growth chamber at 21.5°C under 24 hours of light after cold treatment (4°C) for 2 days. Ten-day-old seedlings were transferred to Metro-Mix 350 soil (Sun Gro) and grown at 21.5°C under 24 hours of light.

Phenotypic analyses

For staining of silique tissue, WT flowers were emasculated at stage 12c (27) and pollinated with hap2-1/+ gcs1/gcs1 pollen grains. Siliques were collected at 0, 1, 2, and 3 DAP. After the samples were dissected and viewed with differential interference contrast microscopy, they were rinsed with Milli-Q (Millipore)–purified water and softened with 1 M NaOH for about 16 hours. The samples were directly stained with aniline blue solution [0.1% (w/v) aniline blue and 0.1 M K3PO4] for more than 3 hours. For the analysis of embryo crossed by kpl/kpl mutant pollen (28) and other fluorescence markers, including FGR8.0 (21), we followed the protocols as previously described (29). For gcs1/gcs1 crossing experiment, we did not count accidentally fertilized ovules as previously described (6).

Detection of cell expansion and cell division

For both detection of cell expansion and cell division, ovules with pollination of WT and gcs1/gcs1 and without pollination were analyzed at 1, 2, and 3 DAP. Ovules were also analyzed at the experimental starting point (0 DAP). To visualize cell shapes, an RPS5Apro::tdTomato-LTI6b plant line ubiquitously expressing a plasma membrane–localized tdTomato was imaged by a two-photon microscopy system (11). For detection of cell division, a CycB1;2pro:CycB1;2::NLS-YFP plant line was used (12). A half part of the ovule was z-scanned, from surface to median section, and all the YFP spots were counted for each ovule. Each sample fraction included 15 to 41 ovules. Confocal/two-photon images were acquired using a laser scanning inverted microscope (LSM780-DUO-NLO, Zeiss). The images were processed using the ZEN 2010 software (Zeiss) to create maximum-intensity projection images.

RNA extraction from ovules

Total RNA for RNA sequencing (RNA-Seq) and quantitative reverse transcription polymerase chain reaction was extracted from ovules surgically isolated from pistils, which were pollinated with WT and gcs1/gcs1 pollen or were without pollination, using a pin and a needle. Only ~20 ovules at the upper region (around the top one-third of pistil) were sampled from a pistil. Isolated ovules were pooled into a droplet of 5 μl of RNAlater on the tip of the pestle. After the collection of 150 to 200 ovules within 30 to 60 min, excess RNAlater was removed. The pestle was inserted into a precooled 1.5-ml tube, and tissues were ground with liquid nitrogen. Two hundred microliters of lysis buffer for the RNAqueous-Micro Total RNA Isolation Kit (Life Technologies) and 175 μl of lysis buffer + 25 μl of Plant RNA Isolation Aid (Ambion) were added into the tube, and the samples were stored at −80°C until subsequent RNA extraction procedures. Total RNA extraction was performed according to the manufacturer’s instructions, where total RNA was eluted in 24 μl of elution solution. The RNA was quantified using the Quant-iT RiboGreen RNA Assay Kit (Life Technologies) with the EnSpire Multimode Plate Reader (PerkinElmer) and qualified by the RNA integrity number analysis using the Agilent RNA 6000 Nano Kit (Agilent Technologies).

RNA-Seq analysis

Total RNA was treated with RNase-Free DNase I (Life Technologies), according to the manufacturer’s instructions. The TruSeq RNA Sample Preparation Kit (Illumina) was used to construct complementary DNA (cDNA) libraries, according to the manufacturer’s instructions. The single ends of cDNA libraries were sequenced for 36 nucleotides from samples using the Illumina Genome Analyzer IIx (Illumina). The resulting sequence data were deposited in the DDBJ Sequence Read Archive at the DNA Data Bank of Japan ( under accession nos. DRA003909 and DRA004932. The reads were mapped to the Arabidopsis reference (TAIR10) using Bowtie (30) with the following options, “-v 3 -m 1 --all --best --strata,” and the number of reads mapped to each reference was counted. The cluster tree illustrated in Fig. 1B was based on Pearson’s correlation matrix among samples. The correlation between two samples was derived as follows. First, genes without expression in any of the two samples were removed. Second, the logarithmic tag count of genes was used as input for the cor function of R ( The cluster tree was plotted by the hclust function of R. Early-response genes in the prefertilization event were collected by two categorizations, as shown in Fig. 1C; the value of Zero_HAP, a reference, was 0 or >0. For the latter case, the genes of which the value of WT_12HAP divided by the value of ZeroHAP was >3 were further screened. For each case, the genes that fulfilled all the following requirements were screened; the value of WT_24HAP divided by the value of WT_12HAP was <0.5, the value of gcs1_24HAP divided by the value of gcs1_12HAP was <0.5, and the value of gcs1_12HAP divided by the value of WT_12HAP was >0.5. The genes associated with cell division, cyclins, cell expansion, expansins, and seed coat development were manually picked up and analyzed.


Supplementary material for this article is available at

fig. S1. Steps to identify PTC-induced genes.

fig. S2. Expansion of ovule integument cells by pollen tube insertion.

fig. S3. Enlargement of ovules by pollen tube insertion.

fig. S4. Fertilization is not required for POEM.

fig. S5. Expression of the embryo- and endosperm-specific marker genes was absent in the enlarged ovules after fertilization failure.

table S1. All RNA-Seq analysis data on the ovules pollinated with NPT, WT, and gcs1/gcs1.

table S2. List of genes in Fig. 1A.

table S3. Data for early-response genes in cases of Zero_HAP = 0.

table S4. Data for early-response genes in cases of Zero_HAP > 0.

table S5. Confirmation data for early-response genes using unfertilized ovules.

table S6. Spatial expression patterns in the ovules for the early-response genes listed in tables S3 and S4.

table S7. Data for the genes associated with embryo and endosperm development, cell expansion, cell division, and seed coat development as well as internal references.

References (3134)

This is an open-access article distributed under the terms of the Creative Commons Attribution-NonCommercial license, which permits use, distribution, and reproduction in any medium, so long as the resultant use is not for commercial advantage and provided the original work is properly cited.


Acknowledgments: We thank E. Matsumoto, N. Furuichi, and N. Ono for technical assistance. We thank D. Kurihara, M. Ito, and R. Groß-Hardt for the gifts of RPS5Apro::tdTomato-LTI6b, CycB1;2pro:CycB1;2::NLS-YFP, and FGR8.0 marker lines. We thank G. N. Drews for critical discussion for this project. Funding: This work was supported by the Precursory Research for Embryonic Science and Technology (grant 13416724 to R.D.K.; Kasahara Sakigake Project), Japan Science and Technology Agency. This work was also supported by a grant-in-aid (25840106 to R.D.K.) from the Japanese Society for the Promotion of Science (JSPS). S.N. was supported by grant 25009285 from the JSPS Fellowship. A part of this work was supported by the Japan Advanced Plant Science Network and JSPS KAKENHI (grant 16H06465 to T.H.) and the Japan Science and Technology Agency (START no. 15657559 and PRESTO no. 15665754 to M.N.). Author contributions: R.D.K. discovered the POEM phenomenon. R.D.K. and S.N. discovered the enhancement of endosperm division by PTCs. R.D.K., M.N., Y.H., and D.M. performed phenotypic analysis for the POEM phenomenon. R.D.K., M.N., S.N., D.S., Y.H., and D.M. performed RNA extraction from ovules for the new-generation sequencer. M.N. and T.S. conducted the transcriptome analysis. R.D.K. and M.N. designed the experiments. R.D.K., M.N., and T.H. wrote the paper from the input of all authors’ experimental results. Competing interests: The authors declare that they have no competing interests. Data and materials availability: All data needed to evaluate the conclusions in the paper are present in the paper and/or the Supplementary Materials. Additional data related to this paper may be requested from the authors.
View Abstract

Stay Connected to Science Advances

Navigate This Article