Pollen tube contents from failed fertilization contribute to seed coat initiation in Arabidopsis

Plant seeds are essential for human beings, constituting 70% of carbohydrate resources worldwide; examples include rice, wheat, and corn. In angiosperms, fertilization of the egg by a sperm cell is required for seed formation; therefore, fertilization failure results in no seed formation, except in the special case of apomixis. Initially, plants produce many pollen grains inside the anthers; once the pollen grain is deposited onto the top of the pistil, the pollen tube elongates until it reaches the ovule. Generally, only one pollen tube is inserted into the ovule; however, we previously found that if fertilization by the first pollen tube fails, a second pollen tube could rescue fertilization via the so-called fertilization recovery system (FRS). Our previous reports also demonstrated that failed fertilization results in pollen tube-dependent ovule enlargement morphology (POEM), enlarged seeds, and partial seed coat formation if the pollen tube releases the pollen tube contents into the ovule. However, we have not determined whether all the ovules enlarge or produce seed coats if an ovule accepts the pollen tube contents. Therefore, we conducted a partial seed coat formation experiment taking into account both the FRS and POEM phenomena. Notably, the ratios of failed fertilization and the ovules with partial seed coats matched, indicating that all ovules initiate seed coat formation if the fertilization fails but the pollen tube contents enter the ovule. In addition, we confirmed that the agl62 mutant , defective in early endosperm formation, showed seed coat initiation with and without fertilization, indicating that for a normal seed coat initiation, fertilization is not required; however, for the completion of normal seed coat formation, both normal fertilization and endosperm formation are required. Further molecular evidence is required to understand these phenomena because very few factors related to FRS and POEM have been identified.

In angiosperms, seed formation begins with pollination 1,2 . Once the pollen grain lands on the stigma at the top of the pistil, pollen tubes from the grain elongate toward the ovule. Fusion of the two gametes is required for seed formation. The male gametophyte is the pollen grain and the female gametophyte is the embryo sac 3 . Immediately after arrival at the ovules, the pollen tube bursts and the pollen tube contents (PTC) are released to the female gametophyte 4 .
In a previous study, we reported that once the ovule accepts the PTC inside the female gametophyte, it begins enlargement and seed coat formation, irrespective of fertilization 5,6 . We named this phenomenon pollen tube-dependent ovule enlargement morphology (POEM). We also reported that if fertilization of the ovule fails, a partial seed coat is still produced, even though a complete seed coat cannot be formed. However, we have not confirmed whether all the ovules have the partial seed coat phenotype when ovule fertilization fails but PTC is accepted. To address this question, statistical experiments were conducted that included the fertilization recovery system (FRS), where a second pollen tube rescues the fertilization if fertilization by the first pollen tube fails, which we previously identified 7,8 . We reported that the seed formation ratio of gcs1/+ mutants 9-11 was approximately 65%; the remaining mutants were unable to produce seeds because fertilization of these ovules failed. Therefore, matching of the ratio of the ovules with the partial seed coat phenotype to the seed abortion ratio suggests that there was fertilization failure for all ovules with a partial seed coat. We also conducted experiments to determine whether agl62 mutants 12 had the partial seed coat phenotype. In agl62 seeds, the endosperm cellularizes prematurely, indicating that AGL62 is required for suppression of cellularization during the syncytial phase. During seed development, AGL62 is exclusively expressed in the endosperm. Because agl62 mutants have an abnormal endosperm phenotype after central cell fertilization, these mutants are ideal for investigating the relationship between endosperm formation and seed coat initiation and formation.

Phenotypic analyses
For staining the silique tissue, the WT flowers were emasculated at stage 12c 13 and pollinated with gcs1/+ pollen grains. For agl62 experiments, the agl62 mutant flowers were emasculated at stage 12c and pollinated with WT, gcs1/+, and agl62 pollen grains. The siliques were collected at 3 days after pollination (DAP).
For vanillin staining, the ovules were manually dissected from the ovaries and mounted on slides in 1% (wt/vol) vanillin (4-hydroxy-3-methoxybenzaldehyde; Sigma) in 6 N HCl solution. Slides were analyzed after 20 min of incubation. Samples were analyzed with a Leica DM2500 microscope using differential interference contrast optics. Images were recorded with a Leica DFC 300 FX digital camera at a magnification of 5×, 10× and 20×. The microscopic protocols followed were as previously described 5 .

Results and discussion
First, the WT plants as both the female and male parent ( Figure 1A) were crossed and the silique after vanillin staining at 3DAP was observed. The ratio of full seed coat formation was 98.7±1.2% (mean ± SD; n=10 pistils), which was consistent with our previous WT fertilization data 7,8 . By contrast, when the WT plants as the female parent and gcs1/+ as the male parent were crossed, the ratio of full seed coat formation was 68.7±5.8% (n=10 pistils) and the ratio of partial seed coat formation was 32.2±6.5% (n=10 pistils), which also was consistent with our previous gcs1/+ fertilization data. These data suggest that all successfully fertilized ovules produce a full seed coat and all unfertilized but PTC accepted ovules produce a partial seed coat.
Because the agl62 mutant had an abnormal and arrested endosperm formation phenotype after fertilization, this mutant was ideal for investigating the relationship between endosperm formation and seed coat initiation and formation. The agl62/+ plants as the female parent and the WT as the male parent were crossed ( Figure 1) and the silique after vanillin staining at 3DAP was observed. The ratio of full seed coat formation was 97.6±2.1% (n=10 pistils), which was consistent with our previous WT fertilization data. By contrast, when agl62/+ plants as the female parent and agl62/+ as the male parent were crossed, the ratio of full seed coat formation was 74.7±3.9% (n=10 pistils) and the ratio of partial seed coat formation was 25.2±5.4% (n=10 pistils), which was consistent with previous agl62 data 12 . These results suggest that normal endosperm development is required for completion of seed coat formation, irrespective of fertilization. When agl62/+ plants as the female parent and gcs1/+ as the male parent were crossed, the ratio of full seed coat formation was 66.9±6.2% (n=10 pistils) and the ratio of partial seed coat formation was 33.2±5.9% (n=10 pistils), which also was consistent with our previous gcs1/+ fertilization data. These results suggest that agl62/+ abnormal endosperm prevents normal seed coat formation, but these

Amendments from Version 1
We have changed Figure 1E. Blue bars indicate SF (Seed formation ratio). Dark red bars indicate VSW (vanillin stained in whole seed coat). Gray bar indicates SA (Seed abortion ratio). Pink bars indicate VSP (vanillin stained in partial seed coat). ♀WT♂WT seed formation ratio (SF) indicates that a WT silique was crossed with WT pollen and the seed formation ratio was calculated. ♀WT♂WT vanillin stained in whole seed coat (VSW) indicates that a WT silique was crossed with WT pollen and the whole seed coat staining ratio was calculated. ♀agl62♂WT VSW indicates that an agl62/+ silique was crossed with WT pollen and the whole seed coat staining ratio was calculated. ♀WT♂gcs1 SF indicates that a WT silique was crossed with gcs1/+ pollen and the seed formation ratio was calculated. ♀WT♂gcs1 VSW indicates that a WT silique was crossed with gcs1/+ pollen and the whole seed coat staining ratio was calculated. ♀agl62♂gcs1 VSW indicates that an agl62/+ silique was crossed with gcs1/+ pollen and the whole seed coat staining ratio was calculated. ♀WT♂gcs1 seed abortion ratio (SA) indicates that a WT silique was crossed by gcs1/+ pollen and the seed abortion ratio was calculated. ♀WT♂gcs1 (vanillin stained in partial seed coat) indicates that a WT silique was crossed with gcs1/+ pollen and the partial seed coat staining ratio was calculated. ♀agl62♂gcs1 VSP indicates that an agl62/+ silique was crossed with gcs1/+ pollen and the partial seed coat staining ratio was calculated. ♀agl62♂ agl62 VSW indicates that an agl62/+ silique was crossed with agl62/+ pollen and the whole seed coat staining ratio was calculated. ♀agl62♂agl62 VSP indicates that an agl62/+ silique was crossed with agl62/+ pollen and the partial seed coat staining ratio was calculated. Blue bars indicate SF (Seed formation ratio). Dark red bars indicate VSW (vanillin stained in whole seed coat). Gray bar indicates SA (Seed abortion ratio). Pink bars indicate VSP (vanillin stained in partial seed coat). (F) A ♀agl62♂agl62 VSP ovule. The arrowhead indicates the vanillinstained zone. Bar: 50 µm. For normal seed coat initiation, fertilization is not required; however, for completion of normal seed coat formation, both normal fertilization and normal endosperm development are required.

REVISED
ovules still produce a partial seed coat because these ovules had accepted the PTC. In summary, for normal seed coat initiation, fertilization is not required; however, for completion of normal seed coat formation, both normal fertilization and normal endosperm development are required.
This project contains the following underlying data: • # of seeds data (Sheet2 contains the number of seeds stained out of the total number of seeds; Sheet1 the data summary used to produce Figure 1E) • agl62-3v.tif (raw image of stained agl62/+ seeds) • gcs1 vaniline.tif (raw image of stained gcs1/+ seeds) • WT vaniline.tif (raw image of stained wild-type seeds) Data are available under the terms of the Creative Commons Zero "No rights reserved" data waiver (CC0 1.0 Public domain dedication). The author revisions are sufficient for the comments by the reviewers. The revised paper is acceptable for indexing.

Grant information
No competing interests were disclosed. During plant fertilization, many cellular processes are happening, and it is important to integrate the identified phenomena together for comprehensive understanding of plant sexual reproduction. The authors studied the relationships among FRS, POEM, seed coat initiation, and endosperm development, and I think the work reported in this paper is an important piece of information in sexual plant reproduction.
I have a few suggestions: The first paragraph of the introduction, the author stated that "gametophytes" fuse. I believe the 1 2 1.

4.
The first paragraph of the introduction, the author stated that "gametophytes" fuse. I believe the authors mean "gametes". From the authors previous work in Science Advances, 2016 ), it is clear that PTC initiates seed coat development (transcriptionally and anatomically). It is important to state these details in the introduction to remind the readers. Along with this, please describe what vanillin staining detects (proanthocyanidin accumulation) and why this can be used for seed coat visualization. In Roszak and Kohler, 2011 , the authors reported that not and , but and fis2 mea fie msi1 autonomously developing seeds differentiate seed coat, suggesting that fertilization is not required for seed coat initiation. The authors also described seed coat phenotypes in . Please refer agl62 this publication and highlight and/or discuss new findings/hypotheses in this paper, then this paper will be more impactful. As other reviewers suggested, Figure 1E is difficult to dissect information. I believe "vamiline" in the Y-axis label should be "vanillin".

If applicable, is the statistical analysis and its interpretation appropriate? Yes
Are all the source data underlying the results available to ensure full reproducibility? Yes Are all the source data underlying the results available to ensure full reproducibility? Thank you very much for your critical comments. We will improve our paper based on your suggestion.
Again, thank you very much.

Ryushiro Dora Kasahara
No competing interests were disclosed. Thank you very much for your critical comments. We will improve our paper based on your suggestion.
Again, thank you very much.

Ryushiro Dora Kasahara
No competing interests were disclosed. This paper showed basically the confirmation of the previous report of the POEM paper. However, because this report showed that the relationships between pollen tube contents and seed coat formation statistically, including the data for agl62 mutants, they clearly confirmed that almost all ovules started seed coat formation without fertilization but just required for the pollen tube contents. I believe this report is a sort of important discoveries in the plant science field and will contribute to further research for seed coat formation in angiosperms.

Introduction:
At the introduction part, it is clear what they need to understand and what sorts of approach were taken. I believe their approach conducted here is adequate and leads to a neat statistics for seed coat initiation and completion study. Results and Discussion: I agree with the first reviewer's comment. Authors could improve the paper if they change the colors for Fig. 1. However, I think this change is not mandatory because if authors change the color dramatically, it could be too showy. I thought current style looks fair enough. I also agree with the first reviewer's comment that because the study of agl62 was not the authors study, "our previous agl62 data" sounds strange. But this is just a very minor correction to be done.
Data availability: I think the data title agl61-3v. tif is wrong. Title should be "agl62-3v".

Is the work clearly and accurately presented and does it cite the current literature? Yes
Is the work clearly and accurately presented and does it cite the current literature? Yes

If applicable, is the statistical analysis and its interpretation appropriate? Yes
If applicable, is the statistical analysis and its interpretation appropriate? Yes

Introduction
Authors described constructively what they wanted to clarify. The experimental design and the statistical analysis performed in this study were enough to examine the questions.

Results and Discussion
In the present state, the descriptions and the graph (Fig.1E) were confusing to understand the results. How about using different colors for each SF, VSF, AS, and VSP. Also, if the graph-bars obtained with the same female parent-genotype are grouped and indicated, it would be helpful for easy comparing the results. As a conclusion, the authors claimed that "for completion of normal seed formation, normal endosperm formation is required in addition to normal fertilization." In the previous studies, the aborted seed caused by the single fertilization of the central cell seemed to have developed the seed coat (kokopelli, gex2, dmp9 mutants, etc...). These studies are supportive of the authors' claim, but the endosperm of these mutants was not normal as the wild type seed. Therefore, how about describing "normal endosperm 'development' is required for seed coat formation." As a minor point, authors describe the previous study of agl62 (Kang et al., 2008) as 'our previous....' but no same authors in this paper. Overall, the content of this research note is logically written. If the points described above are re-considered, it would be improved. In the present state, the descriptions and the graph (Fig.1E) were confusing to understand the results. How about using different colors for each SF, VSF, AS, and VSP. Also, if the graph-bars obtained with the same female parent-genotype are grouped and indicated, it would be helpful for easy comparing the results.
I have submitted a new version of the Fig.1. Please take a look.
I have submitted a new version of the Fig.1. Please take a look.
As a conclusion, the authors claimed that "for completion of normal seed formation, normal endosperm formation is required in addition to normal fertilization." In the previous studies, the aborted seed caused by the single fertilization of the central cell seemed to have developed the seed coat (kokopelli, gex2, dmp9 mutants, etc...). These studies are supportive of the authors' claim, but the endosperm of these mutants was not normal as the wild type seed. Therefore, how about describing "normal endosperm 'development' is required for seed coat formation." I changed the word to normal endosperm development.
As a minor point, authors describe the previous study of agl62 (Kang et al., 2008) as 'our previous....' but no same authors in this paper.
I removed "our". from the text.
Again, thank you very much for your critical comments.
Best regards,

Ryushiro Dora Kasahara
No competing interests were disclosed.

Competing Interests:
The benefits of publishing with F1000Research: Your article is published within days, with no editorial bias You can publish traditional articles, null/negative results, case reports, data notes and more The peer review process is transparent and collaborative Your article is indexed in PubMed after passing peer review Dedicated customer support at every stage For pre-submission enquiries, contact research@f1000.com 1