Downregulation of egg cell-secreted EC1 is accompanied with delayed gamete fusion and polytubey

Abstract

One major player known to be essential for successful gamete interactions during double fertilization in Arabidopsis thaliana is the recently identified family of egg cell-secreted EC1 proteins. Both gamete fusion events are affected in EC1-deficient female gametophytes. Here, we show that the number of ovules with unfused sperm cells is considerably higher than the number of undeveloped seeds in the same ec1-RNAi knockdown lines. We found that some sperm cells are able to fuse with the female gametes even 2 to 3 days after pollination, as reflected by delayed embryo and endosperm development, and by polytubey. We propose that the egg cell secretes EC1 proteins upon sperm arrival to promote rapid sperm activation, thereby accelerating gamete fusion and preventing polytubey.

Fertilization in flowering plants is unique in that 2 female reproductive cells are fertilized by 2 male gametes in a process called double fertilization.^1,2 The growing pollen tube delivers 2 immotile sperm cells into the female gametophyte (FG; embryo sac) where the 2 female reproductive cells (egg cell and central cell) are located. After sperm cell discharge one sperm fuses with the egg cell and gives rise to the embryo, while the second sperm fuses with the central cell to generate the endosperm. Typically, only one pollen tube enters the FG and one pair of sperm approaches the female gametes. In Arabidopsis it takes approximately 5 to 10 minutes after pollen tube discharge for one sperm cell to fuse with the egg cell, and the second sperm to fuse with the central cell,³ indicating that the gamete adhesion and fusion system operates in a very fast and efficient way. At present, only 3 Arabidopsis proteins are known to be essential for gamete fusion by acting on the cell surface.^4-7 The sperm-expressed single-pass transmembrane domain protein GCS1 (GENERATIVE CELL SPECIFIC1)/HAP2 (HAPLESS 2) is thought to act as a fusogen, supported by the fact that HAP2-deficient gametes of Chlamydomonas and Plasmodium berghei are able to adhere but fail to fuse.⁸ Arabidopsis GAMETE EXPRESSED 2 (GEX2), a sperm-expressed single-pass transmembrane protein containing filamin repeat domains, is suggested to contribute to gamete attachment.⁷ The small family of cysteine-rich EC1 proteins, secreted by the egg cell upon sperm arrival, is suggested to control gamete fusion by mediating sperm activation, based on the observed shift of GCS1/HAP2 from the endomembrane system to the sperm cell surface after application of EC1 peptides.⁶

The crucial role of the 5 EC1 genes (EC1.1 to EC1.5) during double fertilization was shown in mutant plants homozygous for T-DNA insertions in 3 EC1 genes (ec1.1, ec1.4, ec1.5) and heterozygous for an RNA interference (RNAi) construct that is directed against the remaining 2 EC1 members (ec1.2/+, ec1.3/+). In ovules of these heterozygous quintuple mutants, termed ec1-RNAi, the sperm cells are delivered into the female gametophyte but the 2 gamete-fusion events are impaired.⁶ Unfused sperm cells are thus visible in 45.1 to 46.7% of the female gametophytes even 30 to 40 h after pollination (HAP) with the fluorescent sperm marker HTR10-mRFP⁹ (Fig. 1A), while in the wild type double fertilization proceeds within 6 to 10 HAP.

Figure 1. Seed abortion phenotype in ec1-RNAi lines. (A) Quantification of the unfused sperm cell phenotype. The number of wild type and ec1-RNAi ovules containing unfused sperm nuclei was quantified 30 to 40 h after pollinating emasculated pistils with the sperm marker line HTR10-mRFP1.⁹ (B) Immature siliques of selfed wild type and ec1-RNAi plants. Undeveloped degenerating ovules (arrows) were frequently detected in the siliques of ec1-RNAi lines. Wild type ovules were normally developed. Bar = 500 µm. (C) Quantification of seed set in developing siliques of the wild type and of 3 independent ec1-RNAi lines. Mean values ± SEM (standard error of the mean) are shown. n = number of siliques (with 40 to 50 ovules, each); PT, pollen tube; WT, wild type.

As a result of defective gamete fusions, a considerable amount of seed gaps is visible in mature siliques of ec1-RNAi plants.⁶ To quantify the proportion of undeveloped seeds we dissected green siliques of the wild type and of 3 independent ec1-RNAi lines that had been used to estimate the frequency of unfused sperm cells. Like expected, we detected unfertilized, white, and shriveled ovules in the siliques of ec1-RNAi plants (Fig. 1B). The quantification of developed and undeveloped seeds revealed that in the wild type the majority of ovules were fertilized (green) and only 0.5% of the ovules were aborted with no signs of fertilization (Fig. 1C). In siliques of line ec1-RNA_16 the frequency of undeveloped seeds (46%) was almost similar to the percentage of ovules with unfused sperm cells (46.7%; Fig. 1A). However, in lines ec1-RNAi_9 and ec1-RNAi_18 we found only 33.4% and 35.3% of undeveloped seeds, respectively.

Pollination experiments with fusion-defective sperm cells of Arabidopsis recently revealed that successful gamete fusion triggers a block to “polytubey,” i.e., the entry of competing pollen tubes into the FG.^10,11 Only in case of gamete-fusion failure a “fertilization recovery system” of the ovule actively rescues failed fertilization by attracting a second pollen tube which may deliver a second, fusion-capable, sperm pair to maximize the likelihood of successful seed formation.^11,12 Polytubey also occurs in ec1-RNAi ovules and was considered to be due to defective gamete fusion based on the lack of EC1-mediated sperm activation.⁶ The delivery of a second pair of sperm cells will, however, not rescue fertilization in ovules with female fusion-defective gametes such as ec1-RNAi egg cells. Nevertheless, the discrepancy between the number of ovules containing unfused sperm cells and the number of undeveloped ovules suggests that a proportion of EC1-deficient ovules containing unfused sperm cells are able to develop into seeds in the lines ec1-RNAi_9 and ec1-RNAi_18.

To investigate the possibility of delayed fertilization we analyzed cleared seeds dissected from 5 mm long siliques (a mixture of seeds of approximately 2 to 3 DAP) by differential interference contrast (DIC) microscopy (Fig. 2). In the wild type only 1% of the ovules were aborted with collapsed female gametophytes, while 99% of the ovules were fertilized, containing embryos at the 4-celled or the octant-stage and several endosperm nuclei (Fig. 2A and G). Similar embryo and endosperm developmental stages were observed in 39.3 to 47.4% of the ovules of the 3 ec1-RNAi lines (Fig. 2B and G). In ec1-RNAi, delayed fertilization events were indicated by ovules containing zygotes or 2-celled proembryos, along with a couple of endosperm nuclei (1.0 to 2.7%; Fig.  2C, D and G). A considerable proportion of 23.4 to 29.4% ovules did not show obvious signs of fertilization but egg and central cell nuclei were still visible (Fig. 2E and G). However, DIC microscopy of cleared ovules cannot provide any information as to whether these ovules did not attract pollen tubes, or sperm cells were released but did not fuse. Collapsed female gametophytes without any clearly detectable cell structures were detected in 26.5 to 30.0% of the ovules (Fig. 2F and G). Notably, this phenotype was not observed in mature but unfertilized ec1-RNAi ovules.⁶ Since neither pollen tube guidance nor pollen tube perception is affected in ec1-RNAi knock down mutants,⁶ we assume that the collapsed female gametophytes, observed at 2 to 3 DAP, result from unsuccessful fertilization events in EC1-deficient ovules. Furthermore, the lower number of fertilized ovules (47.1% to 50.1%), compared to the number of developed seeds in green siliques (64.7% to 66.6%; Fig. 1C) of the same transgenic lines, suggests that a significant proportion of the ovules with no obvious signs of fertilization (12.3% to 22.8%) will later on develop into seeds.

Figure 2. Delayed fertilization detected in ec1-RNAi ovules at 2 to 3 DAP. (A–F) DIC microscopy analyses of cleared⁶ developing ovules, prepared from 5 mm long siliques. Visible nuclei of the egg cell, zygote, or embryo are artificially colored in red; endosperm or central cell nuclei are artificially colored in blue. (A) Wild type ovule with 4-celled embryo and developing endosperm. (B) Ovule of a heterozygous ec1-RNAi plant showing normally developed embryo and endosperm. (C) Delayed development of fertilized ec1-RNAi ovule. The elongated zygote is visible, with its nucleus in the center. Four endosperm nuclei are visible in this focus plane. (D) Delayed development of fertilized ec1-RNAi ovule. The 2-celled proembryo and 6 endosperm nuclei are visible in this focus plane. (E) ec1-RNAi ovule with detectable central cell and egg cell nuclei. (F) Degenerating ec1-RNAi ovule with collapsed female gametophyte. (G) Quantification of the various phenotypes shown in A-F in the wild type and in 3 independent ec1-RNAi lines. Mean values ± SEM (standard error of the mean) are shown, n, number of siliques (with 40 to 50 ovules, each); 2-celled/2-c, 2-celled proembryo; ccn, central cell nucleus; ecn, egg cell nucleus; emb, embryo; end, endosperm; FG (deg), female gametophyte (degenerated); WT, wild type; zyg, zygote. Bars = 20 μm.

To analyze earlier post-fertilization stages, and to simultaneously image unfused sperm cells, we performed Feulgen staining¹³ and confocal laser scanning microscopy (CLSM) of pistils at 30 HAP (Fig. 3). The highly condensed DNA of the sperm cell nuclei and the more decondensed chromatin of the vegetative cell nucleus can be visualized by this method (Fig. 3A), as well as the nuclei of the synergid cells, the egg cell, and the central cell in ovules of unpollinated wild type and ec1-RNAi pistils at 2 days after emasculation (Fig. 3B and E). In wild type ovules 8 HAP, the 2 sperm cell nuclei can occasionally be detected near the fusion site, shortly before gamete fusion will take place (Fig. 3C). At 30 HAP the 2-celled proembryo and numerous endosperm nuclei are visible in the majority (71%) of pollen tube-targeted wild type ovules (Fig. 3D and J), while 23.5% ovules displayed a zygote and a couple of endosperm nuclei (Fig. 3J). By contrast, CLSM of Feulgen-stained ec1-RNAi ovules 30 HAP showed various phenotypes (Fig. 3F–J). Besides fertilized ovules exhibiting the same stage of development as the majority of the wild-type ovules (26.5%; Fig. 3J) we detected unfertilized ovules with 2 or even 4 sperm nuclei near the 2 female gametes (27.1%; Fig. 3F and J). Importantly, a significant proportion of pollen tube-targeted ovules contained a zygote and a few endosperm nuclei, plus 1 or 2 additional sperm pairs in the area of the synergid cells (8.2 and 7.7%; Fig. 3G, H, and J). The frequency of pollen tube-targeted collapsed female gametophytes 30 HAP (4.7%; Fig. 3I and J) is lower than the frequency of degenerated ovules 48 to 72 HAP (Fig. 2F and G). This supports our assumption that the female gametophyte degenerates over time in case of unsuccessful double fertilization.

Figure 3. Unfused sperm cells are present both in unfertilized and in fertilized ec1-RNAi ovules. (A–I) CLSM images after Feulgen staining. (A) Pollen grains from mature anthers showing fluorescent signals of the 2 sperm cell nuclei (arrowheads) and the vegetative cell nucleus. (B) Wild type ovule 2 d after emasculation (DAE). Fluorescent signals of the 2 synergid nuclei, the egg cell, and the central cell nucleus are visible. (C) Wild type ovule 8 HAP. Sperm cells have been delivered and the sperm nuclei are detectable (arrowheads) near the gamete fusion site. (D) Fertilized wild type ovule, 30 HAP. The nuclei of the apical and the basal cell of the proembryo and 3 endosperm nuclei are detectable in this focus plane. (E) ec1-RNAi ovule, 2 DAE. Nuclei of synergid cells, egg, and central cell are detectable. (F) ec1-RNAi ovule, 30 HAP. Egg and central cell are unfertilized, 2 sperm cell nuclei (arrowheads) are detectable in each synergid cell. (G and H) Fertilized ec1-RNAi ovules, 30 HAP. The nuclei of the developing zygote and the endosperm are visible. Two (G) and 4 (H) unfused sperm cells (arrowheads) are detectable within the region of the synergid cells. (I) Pollen tube-targeted (arrowhead) ovule with collapsed female gametophyte. (J) Quantification of fertilization and sperm cell phenotypes in Feulgen-stained ovules 30 HAP (n = 102 ovules from 5 different lines); Ac, apical cell; bc, basal cell; central cell; ccn, central cell nucleus; ec, egg cell; ecn, egg cell nucleus; en, endosperm nucleus; PT, pollen tube; sy, synergid cell; syn, synergid cell nucleus; vn, vegetative cell nucleus; zyg, zygote. Bars = 20 µm.

Polytubey results in supernumerary sperm cell delivery and increases the risk of multiple fertilizations. Kasahara et al.^11,12 showed that most ovules are targeted by one pollen tube within the first 10 h after pollination, indicating the existence of a blocking system by which ovules avoid polytubey until several hours after the arrival of the first pollen tube. In case of fertilization failure the persistent synergid cell does not degenerate but continues to attract pollen tubes, resulting in approximately 80% of failed ovules accepting a second pollen tube within the next 18 h.^10,11 The presence of supernumerary sperm cells in fertilized ec1-RNAi ovules 30 HAP strongly suggests that gamete fusion took place in a delayed fashion in these ovules, allowing them to overcome the block to polytubey. However, we did not observe polyspermy (fertilization by more than one sperm), indicating that the block to polyspermy is not affected in EC1-deficient egg cells. Previously, we reported that the EC1 gene family is essential for gamete fusion and that synthetic EC1 peptides, applied to sperm cells that had been released from pollen tubes, are able to activate the sperm endomembrane system.⁶ Here, we show that at least some of the sperm cells delivered into EC1-deficient ovules succeed in gamete fusion at a later time. Notably, these sperm cells are able to accomplish fertilization even 2 to 3 DAP. The presence of residual transcripts of EC1.2 and EC1.3 might be one explanation for delayed fertilization, because of the variable knockdown efficiencies in ec1-RNAi lines.⁶ The low amount of EC1 may either decrease the speed of sperm activation, or it may extend the time frame needed to reach a certain EC1 threshold that is necessary to activate the sperm cells. Some sperm cells might, on the other hand, be generally able to slowly acquire fusion competence after they are released from the pollen tube. Importantly, rapid gamete fusion is a prerequisite for the block to polytubey and the fertilization recovery system. A recent study by Maruyama et al.¹⁴ indicates that flowering plants have evolved a dual-control polytubey blocking system, in which egg and central cell fertilization act independently and synergistically to induce a strong polytubey blocking response. In case of single fertilization (i.e., egg or central cell only) polytubey and heterofertilization allowed the ovules to complete double fertilization.¹⁴ We suggest that the egg cell secretes EC1 proteins upon sperm arrival to achieve rapid sperm activation and gamete fusion and contributes to the induction of the dual-control polytubey block.

Glossary

Abbreviations:

CLSM

confocal laser scanning microscopy

DIC

differential interference contrast microscopy

EC1

EGG CELL 1

FG

female gametophyte

GCS1

Generative cell specific 1

HAP

hours after pollination

HAP2

HAPLESS2

RNAi

RNA interference