Comparative metabolites proling of Polygonatum cyrtonema Hua seed during sand storage and germination

The seeds of Polygonatum cyrtonema Hua have dormancy phenomenon. Previous studies have shown that sand storage factors effects of the seed dormancy of P. cyrtonema Hua seeds and enhance the seed germination process. Subsequently, metabolic activities and different changes during the sand storage and germination process of P. cyrtonema Hua seed has not been heavily researched. In this study the changes in the metabolites of P. cyrtonema Hua seeds at different sand storage times and germination stages, we used untargeted metabolomics to determine them. Most of the sugar and glycoside contents in seed coat increased after 30 d on the other hand, in peeled seeds increased at 30 d and decreased at 60 d after sand storage treatment. The content of proline and benzoic acid decreased in the seed coat after sand storage. PCA, OPLS-DA and HCA showed that the contents of most metabolites increased after 7 d and decreased after 14 d of seed germination. The process of 7 d to 14 d was the key stage of seed germination of P. cyrtonema Hua. Differential metabolic pathway analysis showed that seed germination was controlled by multiple metabolic pathways. Metabolic correlation revealed the interdependence between seed germination metabolites and metabolic pathways. the mechanism


Results
In this study the changes in the metabolites of P. cyrtonema Hua seeds at different sand storage times and germination stages, we used untargeted metabolomics to determine them. Most of the sugar and glycoside contents in seed coat increased after 30 d on the other hand, in peeled seeds increased at 30 d and decreased at 60 d after sand storage treatment. The content of proline and benzoic acid decreased in the seed coat after sand storage. PCA, OPLS-DA and HCA showed that the contents of most metabolites increased after 7 d and decreased after 14 d of seed germination. The process of 7 d to 14 d was the key stage of seed germination of P. cyrtonema Hua. Differential metabolic pathway analysis showed that seed germination was controlled by multiple metabolic pathways. Metabolic correlation revealed the interdependence between seed germination metabolites and metabolic pathways.

Conclusion
Sand storage can signi cantly increase the rate of seed germination and play a vital role in seed dormancy of P. cyrtonema Hua. There was inherent differences in metabolites during different storage time and germination stages in P. cyrtonema Hua. Our work provides a rst glimpse of the metabolome in seed germination of P. cyrtonema Hua, and provides a valuable informations for revealing the mechanism of breaking seed dormancy. Background P.cyrtonema Hua is a famous medicinal plant,distributed in China. The rhizome of P.cyrtonema Hua has become more popular due to high nutritional and medicinal value, mainly. that are used to treat disorders of the lung, kidney and spleen, and more importantly to prevent aging [1]. The major active ingredients of P. cyrtonema Hua are Polygonatum sibiricum polysaccharides [2]. At present, the reproduction of P. cyrtonema Hua mainly depends on the rhizome, but long-term asexual reproduction causes germplasm degradation and thereby leads to diseases and insect pests related problems. The P. cyrtonema Hua seed reproduction is due to its large number and easy operation in actual production, which are effectively reduce production costs and solve the above problems. However, the low reproduction rate of P. cyrtonema Hua seeds in the natural environment has resulted in the need to use a large number of seeds in production practices to obtain su cient seedlings, which results in a waste of resources and ine ciency. The seeds of P. cyrtonema Hua show comprehensive dormancy [3] which prevents seeds from germinating under unfavorable physical conditions [4,5]. In many wild species, seeds are dormant at maturity and do not germinate until dormancy release after dispersal, even after exposed to optimal conditions. Germination is a complex trait that is in uenced by endogenous and environmental factors [6,7].
There are many methods including chemical, biological, and physical methods to alleviate the seed dormancy. Strati cation treatment is commonly used to break seed dormancy. Cold strati cation has been widely used for breaking the dormancy of seeds germination in the natural habitat in spring, e.g. Carthamus tinctorius L. [8], Acer morrisonense [9]. Phytohormones are signaling molecules for communication between the three seed compartments (seed embryo, endosperm, and seed coat), and play key roles in coordinating appropriate seed formation [10]. Low temperature sand storage is optimal for break the seed dormancy of P. cyrtonema Hua [11]. Nevertheless, in some other plants, room temperature sand storage is more effective [12]. Among them, sand storage (low temperature strati cation treatment ) is the most common method to end the seed dormancy. In the dormancybreaking treatment, seeds are treated with wet sand storage at 4 °C. Few studies on the mechanism of breaking dormancy and changes of metabolites during dormancy breaking are reported.
To reveal the mechanism of breaking dormancy, many studies have focused on different physiological aspects in the process of dormancy. Metabolomics has also been applied to analyze the changes in metabolites of breaking seed dormancy. For example, a gas chromatography-mass spectrometer (GC-MS) has been used to reveal the metabolite's changes of Safron corm from dormancy release to germination [13]. Takahashi analyzed the poor germination of Gentiana tri ora seed by using targeted metabonomics [14]. Recently, metabolomic characterization based on the non-targeted GC-MS of seed germination has been reported [15,16].In present study, the seeds of P. cyrtonema Hua were used as raw materials. GC-MS technique was used to study the changes of metabolites in the process of sand storage and seed germination.

Results
Effects of sand storage on the seed germinatio n rate of P. cyrtonema Hua The seeds of P. cyrtonema Hua treated with sand storage at 4 °C in winter for 30 d. (Fig.1). The rate of seeds germination was signi cantly higher (3.69 times) after the treatment with sand storage as compared to non-sand storage (control). . These results showed that treatment of sand storage plays an important role in the germination of P. cyrtonema Hua seeds.
Metabolic pro ling of the seed coat and peeled seeds of P. cyrtonema Hua at different storage time The metabolic pro les of the seed coat and peeled seeds during four sand storage stages were investigated using GC-MS. Metabolome-wide expression pro ling showed highest metabolite expression during the treatment of sand storage. Through comparative analysis, 85 and 74 metabolites were identi ed in seed coat and peeled seeds, respectively (Fig.3). In the metabolite pro les were signi cant differences existed between the seed coat and peeled seeds. There are 65 metabolites detected in both the seed coat and peeled seeds, but 20 metabolites were only detected in the seed coat and 9 metabolites were only detected in peeled seeds (Fig.3B). The speci c metabolites of seed coat included four carbohydrates, six alcohols, four organic acids, four nitrogenous compounds, and two other compounds, as well as, the nine metabolites speci c to peeled seeds included two carbohydrates, ve organic acids, and two other metabolites were presented (Fig.3A).
Changes of metabolites in seed coat and peeled seeds of P. cyrtonema Hua in metabolic pathways during different sand storage time Our study sought to obtain a more detailed overview of the abundance of the identi ed substances and, similarities and differences in each storage time of both organs (Fig.4). In starch and sucrose metabolism, trehalose showed a similar trend in seed coat and peeled seeds during sand storage. Trehalose increased at 30 d of sand storage, decreased signi cantly at 60 d, and slightly increased at 90 d. The sucrose, fructose, and maltose in peeled seeds were signi cantly higher than those in other periods when stored in the sand for 30 d. The sucrose and maltose of seed coat decreased signi cantly after 60 d of sand storage. In the TCA cycle, seed coat and peeled seeds had the same metabolites, but the changing trend of metabolites was different in seed coat and peeled seeds. During sand storage, the content of pyruvate, succinate, and malate in peeled seeds showed similar trends, and the metabolites of sand storage at 60 d were signi cantly higher than those in the other three periods, indicating that the TCA cycle was relatively active at this time. Succinate of seed coat showed a signi cant downward trend in the sand storage process and increased slightly at 90 d. Pyruvate of seed coat was signi cantly higher than that in the other three periods during 30 d of sand storage. Malate of seed coat and peeled seeds changed in opposite directions.
These results indicating that, most of the sugar and glycosides content in the seed coat increased at 30 d, indicating that the macromolecular nutrients in the seed coat began to decompose into soluble sugar, and the other sand storage time changed differently. Most of the sugars and glycosides in peeled seeds increased at 30 d of sand storage and decreased at 60 d of sand storage, which indicated that the macromolecular nutrients in seeds began to decompose into soluble sugars in preparation for seeds to break dormancy. After 60 d of sand storage, sugar and glycosides began to degrade, indicating that the seeds have been in a relatively active state, began to consume sugar and glycosides to prepare for seed germination. The changing trend of amino acids in seed coat and peeled seeds were also different. The content of most amino acids in seed coat decreased with the extension of sand storage time, and most of the amino acids in peeled seeds were the highest at 60 d of sand storage. Proline was an important product of dormancy release [20]. The content of proline in the seed coat decreased after sand storage, while the content of proline in peeled seeds increased after 60 d of sand storage. Benzoic acid is only detected in the seed coat, and benzoic acid is considered to be an endogenous inhibitor [21].
Morphological changes of P. cyrtonema Hua seeds at different germination stages Through the dynamic observation of the germination process (Fig.5), it was found that the seeds of P. cyrtonema Hua started to germinate on 14 d, and the seeds began to show small buds, and the corms started outgrowth after the germination of 21 d. Then the corm began to germinate and grow to form radicle and hypocotyl.
Analysis of metabolites during seed germination of P. cyrtonema Hua based on GC-MS GC-MS was used to analyze the metabolites during the germination process of P. cyrtonema Hua. A total of 96 metabolites were isolated and identi ed, including 22 amino acids, 20 sugars and glycosides, 22 organic acids, 13 alcohols and esters, 8 nitrogen compounds, 4 fatty acids, and 7 other compounds.
Firstly, the samples were analyzed by using PCA to detect the distribution of metabolites at different germination stages of P. cyrtonema Hua seeds (Fig.6) The OPLS-DA, a supervised pattern recognition method, was further employed to identify the metabolites during the germination of P. cyrtonema Hua seeds (Fig.7 Table S1).
The dynamic changes of metabolites during the seed germination stages were analyzed by HCA. As shown in Fig.8, the metabolites during seed germination can be divided into three clusters. Cluster I mainly consists of two organic acids (succinic acid and 2, 4, 6-trihydroxybenzoic acid), three amino acids (glycine, 4-aminobutyric acid, β-alanine), and a glycoside compound. The contents of these metabolites decreased gradually during the seed germination process. Cluster II mainly includes 9 amino acids (proline, alanine, glutamic acid, etc.), 10 organic acids (citric acid, mandelic acid, glyceric acid, etc.), 7 sugars and glycosides (maltose, trehalose, moschose, etc.), 4 alcohols (inositol, glycerol, squalinositol, etc.), one fatty acid, four nitrogen-containing compounds, and four other metabolites after seed germination for 7 d, these metabolites contents rstly increase, and most of them decreased after 14 d of germination. Cluster III mainly consists of 12 sugars and glycosides (sucrose, cellobiose, fructose, etc.), 10 amino acids (tryptophan, L-lysine, tyrosine, etc.), 11 organic acids (malic acid, lactic acid, niacin, etc.), Changes of metabolites during seeds germinationof P. cyrtonema Hua Through the analysis of KEGG pathway of differential metabolites, the differential metabolites were enriched to 32, 32, 4, and 9 pathways from 0 d vs. 7 d, 7 d vs. 14 d, 14 d vs. 21 d, 21 d vs. 28 d, respectively. There was only one differential metabolite in 28 d vs. 35 d, and on the other side, there was no differential metabolic pathway (Additional le2: Table S2). The signi cantly enriched metabolic pathway was mainly related to amino acid metabolism, sugar metabolism, inositol phosphate metabolism and citric acid cycle. The metabolic pathways with signi cant enrichment were citric acid cycle, arginine and proline metabolism, starch and sucrose metabolism, etc. (Additional le2: Table S2).
A pair-wise comparison between all the stages was performed. We observed signi cant metabolite changes (p<0.05) between 0 d and 7 d as depicted on the metabolic map (Fig.9A) When the seeds germinated at 21 d, the differential metabolities contents (mainly were amino acids and inositol) were decreased. On the 28 d of germination, the content of most of the differential metabolism decreased signi cantly, except for isoleucine and 4-hydroxypyridine, the content of other differential metabolites decreased signi cantly, including sugar, niacin and sorbitol. In summary, it can be concluded that the seed germination of P. cyrtonema Hua may be affected by a variety of metabolic pathways.

Discussion
Compared with the control group, the sand storage treatment can signi cantly improve the seeds germination rate of P. cyrtonema Hua. This is consistent with the research of Bian, F [22], used of sand storage to release dormancy and improve the germination rate of Taxus yunnanensis. During the postripening development of seeds in the process of sand storage, the internal macromolecular substances and enzyme activities were in a certain dynamic change. In the post-ripening stage, the macromolecular substances (such as starch, etc.) can be decomposed into soluble sugars, such as glucose, sucrose, fructose, etc., and then participate in seed germination and basic physiological respiration needs. During the storage at room temperature, the soluble sugar in the seeds increased at rst and then decreased with the extension of storage time. It is speculated that the starch decomposition in the seeds at the early stage of storage may provide soluble sugar, and with the extension of storage time, the respiration may become weaker, resulting in a decrease during the utilization of soluble sugar [23]. Soluble sugars include most monosaccharides and oligosaccharides, which can be used as mediators for energy storage and transfer in plants, as well as ligands for structural substances and functional molecules such as glycoproteins. From the analysis of the changes of metabolites in the seeds of P. cyrtonema Hua treated with low temperature sand storage at different times, it can be seen that the contents of most sugar and glycosides in the peeled seeds started to increase gradually from 0 d to 60 d of sand storage times, and decreased signi cantly at 90 d. After 60 d, the sugar and glycosides (glucose, sucrose, fructose) began to degrade, indicating that the seed has been in a relatively active state, started to consume carbohydrates to prepare for seed germination, which is consistent with the research of Ma, L, etc [24].
Since the seed was imbibed, its internal metabolism was activated to provide nutrition and energy for seed germination. Through the dynamic observation of the germination process of P. cyrtonema Hua seeds (Fig. 5), it was found that the seeds began to germinate at 14 d, and the seeds started to show small buds, and the corms appeared at 21 d after germination. At the later stage, the corm began to germinate and grow to form radicle and hypocotyl. Sugar, as an important source of carbon source in seeds, provides conditions for seed germination [25]. In this study, during seed germination for 0 to 14 d, the content of disaccharide (trehalose, maltose, sucrose, etc.) began to increase, and these metabolites decreased after 14 d of germination. These disaccharides are soluble sugars and are considered to be rapidly available [26]. The generation of energy may be related to the cell growth in the seed, including the expansion of the protocambium ring and the radicle [27]. After 14 d of seed germination, the content of disaccharide decreased, which indicated that the glycolysis pathway and tricarboxylic acid cycle were affected in these stages. At the same time, we also found that the content of intermediate metabolites (malic acid, succinic acid, citric acid) of tricarboxylic acid cycle decreased at the stage of 14 d to 28 d of germination. During seed germination process, seeds take energy through other metabolic pathways, such as fermentation and amino acid consumption [28]. Amino acid metabolism is essential during seed germination because it must be consistent with the changes of carbohydrates and carbohydrates in seeds in order to drive cell metabolism and promote seed germination [29]. The changing trend of most amino acids in seeds was the same as that of disaccharide. It has been found that γ-aminobutyric acid is related to the production of seed energy [30]. In the process of seed germination, the content of γ-aminobutyric acid was the highest after 0 d of germination, and began to decrease in other stages. Besides, inositol is associated with many aspects of plant physiology, such as carbohydrate metabolism, seed germination, stress response and cell wall formation, as well as cell division [31,32]. Our study showed that the content of inositol in seeds increased at the stage of 0 d to 7 d of germination, and decreased at the beginning of germination (14 d), these indicated that the inositol in seeds was used for seed germination.

Conclusion
These results indicated that the germination and dormancy process was signi cantly increased after the treatments of sand storage in P. cyrtonema Hua seeds. During the sand storage process, the metabolites in P. cyrtonema Hua seeds and peeled seeds change during different sand storage times. In this process of seed sand storage from 0 d to 90 d, 10 kinds of sugar and glycosides content in the seed coat continued to increase, 7 kinds of them continuously decreased, and another 5 kinds of them increased rst and then decreased, the highest content was at 30 d. The content of most carbohydrates and glycosides in the peeled seeds increased at 30 d after the treatment of sand storage, and gradually decreased after 60 d. The content of most amino acids in the seed coat gradually decreased with the extension of the sand storage times, while the content of most amino acids in the peeled seeds increased rst and then decreased, the highest content was at 60 d. We found that the content of proline,an important product of releasing dormancy decreased in the seed coat after sand storage. Benzoic acid, an endogenous inhibitors, decreased also.
During the germination of P. cyrtonema Hua seeds, from 0 d to 35 d, the content of most amino acids, organic acids, sugars, and alcohols showed a trend of rst increasing and then decreasing, as well as, the content reached the maximum value from 7 d to 14 d. A variety of differential metabolites were screened from the germination process of P. cyrtonema Hua seeds, which were mainly concentrated in the rst 14 days of germination. According to the metabolic map, the germination of P. cyrtonema Hua seeds is affected by various metabolic pathways. Among them, the metabolic pathways of amino acids, sugars and organic acids showed a positive correlation.

Plant material
The P. cyrtonema Hua seeds were provided by the Polygonatum planting base of Sansanfulin Chinese Medicinal Materials Co., Ltd.,,China in September 2017, which located in Zongwen Village, Ducun Township, Qingyang County, Chizhou City, Anhui Province in China (30°32' 30.8"N 117°43'54.8"E), and authenticated by Professor Yongping Cai (School of Life Sciences, Anhui Agricultural University). Sand storage and seed germination of P. cyrtonema Hua were carried out in the laboratory of Anhui Agricultural University.
The effect of sand storage on the seed germination rate of P. cyrtonema Hua After rubbing, rinsing, washing and drying the fresh P. cyrtonema Hua fruit, sand storage treatment was carried out at 4 °C in winter, and water was sprayed regularly to maintain humidity. The seeds of P. cyrtonema Hua were taken out, planted in the hole seedling tray, and cultured in a constant-temperature incubator at 26 °C. For this, one hundred seeds per treatment were used in triplicates, and count the seed germination rate after 30 days. Using seeds not in sand storage as the control. SPSS19 software was used for data statistical analysis. The germination rate was calculated using the following formula: Germination rate = a total number of germinated seeds / total number of tested seeds *100%.
Changes of metabolites of P. cyrtonema Hua seed coat and peeled seeds in different sand storage times The seeds of P. cyrtonema Hua were collected on these stages 0 d, 30 d, 60 d, and 90 d in sand storage, respectively. The seed coat and peeled seeds(including seed embryo and endosperm)were separated and frozen with liquid nitrogen and stored in a refrigerator at -80℃ until used for GC-MS to identify metabolites.
Changes of metabolites during the germination of P. cyrtonema Hua seeds The seeds of P. cyrtonema Hua after 60 days of sand storage were soaked in GA 3 for 24 h, and the seeds germinated for 0 d, 7 d, 14 d, 21 d, 28 d and 35 d, respectively. They are frozen with liquid nitrogen and stored in a refrigerator at -80℃ until used for GC-MS to identify metabolites.

Extraction of Metabolites of P. cyrtonema Hua seeds
The protocol for GC-MS metabolite extraction was following that of the previously described procedure [13]. The collected samples (100 mg) were grinded in a mortar using liquid nitrogen, then transferred into 10 mL centrifuge tubes. After adding 1.4 mL of 100% methanol (precooled at -20 °C) and vortexing for 30 s, 60 µL of polar internal standard (0.2 mg·mL − 1 1,2-benzenediol in methanol) was added to the tube.
The samples were vortexed again for 30 s and placed into an ultrasound machine for 30 min at 40 °C, followed by centrifugation for 20 min at 4000 rpm. Then, we transferred the supernatant to a 10 mL centrifuge tube which was followed by the addition of 750 µL of chloroform and 1.4 mL of ddH 2 O and the mixture was vortexed for 30 s before centrifugation at 4000 rpm for 20 min. The supernatant was transferred into a new tube and blow-dried with nitrogen. Subsequently, 60 µL of methoxypyridine (20 mg·mL − 1 ) was added and vortexed for 30 s. After a reaction time of 2 h at 37 °C, 60 µL of N,Obis(trimethylsilyl) tri uoroacetamide (BSTFA) reagent (containing 1% trimethylchlorosilane) was added to the tube. After reacting for 1.5 h at 37 °C, the samples were used for GC-MS analysis.

Detection conditions by GC-MS
The GC-MS data were obtained using an Agilent 7000B (Agilent, Santa Clara, CA, USA) system equipped with an DB-5 MS capillary column (60 m × 0.25 mm, 0.25 µm lm thickness, Agilent J & W Scienti c, Folsom, CA, USA); The injection inlet temperature was 280℃, source temperature of 250℃, an inlet temperature of 250℃. Split ratio 10:1 and injection volume was 1 µL The following temperature program: initial temperature of 40℃ and held for 5 min, increased by 8℃/min to 280℃, and held for 5 min with a constant ow of 1 mL/min helium as the carrier gas. The solvent delay time was 14 min, and m/z range was 33-600. The detection was by an Agilent 5977C Triple Quad with the following settings: electrospray ionization (ESI) source(70 eV); source temperature of 230℃, the quad temperature of 150 °C and m/z range was 35-780.

Data Analysis
To ensure the stability of the assay, we performed six biological replications and analyzed them under identical conditions. The QCs were injected at regular intervals throughout the GC-MS analytical run to provide a set of data from which repeatability could be assessed. All the GC-MS data were processed by XCMS running under the R package, which produced a matrix of features with the associated retention times, accurate masses, and peak areas. All internal peaks were removed from the data set. The resulting data were normalized to the total peak area of each sample in Excel 2010. Principal component analysis (PCA) and orthogonal partial least squares-discriminant analysis (OPLS-DA) models were generated using Simca-P software version 14.0. The differential metabolites were selected based on the combination of the statistically signi cant threshold of variable in uence on projection (VIP) values obtained from the OPLS-DA model and the p-value from a two-tailed Student's t-test of the normalized peak area; metabolites with VIP > 1.0 and p < 0.05, respectively, were selected. For GC-MS analysis, metabolites were identi ed by searching the commercial database NIST11 [17] after their mass spectra were deconvoluted by the Automated Mass Spectral Deconvolution and Identi cation System (AMDIS) [18]. Peaks with a similarity index of more than 70% were tentatively identi ed as metabolites. The identi ed metabolites were mapped to general biochemical pathways according to their annotation in the Kyoto Encyclopedia of Genes and Genomes (KEGG) database. Pathway analysis was performed using MetaboAnalyst 4.0 [19]. The metabolome data of the differential metabolites were used for different analysis and subsequently normalized by sum, log-transformed, and Pareto-scaled thereby compared with the KEGG pathway library of Arabidopsis thaliana. The heatmap was generated with R software. Availability of data and materials All data generated or analysed during this study are included in this published article and its supplementary information les.

Competing interests
The authors declare that they have no competing interests.

Funding
This work was supported by the major project of Anhui Department of Education (03087060) and the Industry-University-Research Cooperation Project of Chizhou City, Anhui Province(11006232).
Author contributions Q.J. and Y.P.C. conceived and designed the experiments; J.T. W.Z., L.X., X.Z., R.R., J.Z., and Y.C. performed the experiments; J.T., W.Z. analyzed the data and wrote the paper; Q.J., and Y.P.C. revised the paper; All authors read and approved the nal manuscript.